Barriers to Clinical Trial Enrollment in Racial and Ethnic Minority Patients with Cancer

cancercontroljournal.org Vol. 23, No. 4, October 2016

H. LEE MOFFITT CANCER CENTER & RESEARCH INSTITUTE, AN NCI COMPREHENSIVE CANCER CENTER

Barriers to Clinical Trial Enrollment in Racial and Ethnic Minority Patients With Cancer

Racial and Ethnic Differences in the Epidemiology and Genomics of Lung Cancer

Black Heterogeneity in Cancer Mortality: US-Blacks, Haitians, and Jamaicans

Genomic Disparities in Breast Cancer Among Latinas

Clinical Considerations of Risk, Incidence, and Outcomes of Breast Cancer in Sexual Minorities

Construction and Validation of a Multi-Institutional Tissue Microarray of Invasive Ductal Carcinoma From Racially and Ethnically Diverse Populations

Determinants of Treatment Delays Among Underserved Hispanics With Lung and Head and Neck Cancers

Geographical Factors Associated With Health Disparities in Prostate Cancer

Disparities in Penile Cancer

Chemoprevention in African American Men With Prostate Cancer

Disparities in Adolescents and Young Adults With Cancer

Tobacco-Related Health Disparities Across the Cancer Care Continuum

Cancer Control is included in Index Medicus/MEDLINE Editorial Board Members

Editor: Conor C. Lynch, PhD Production Team: Assistant Member Lodovico Balducci, MD Veronica Nemeth Tumor Biology Senior Member Editorial Coordinator Program Leader, Senior Adult Oncology Program Kristen J. Otto, MD Sherri Damlo Moffitt Cancer Center Associate Member Medical Copy Editor Head and Neck and Endocrine Oncology Diane McEnaney Deputy Editor: Graphic Design Consultant Julio M. Pow-Sang, MD Michael A. Poch, MD Senior Member Assistant Member Associate Editor of Chair, Department of Genitourinary Oncology Genitourinary Oncology Educational Projects Director of Moffitt Robotics Program Jeffery S. Russell, MD, PhD John M. York, PharmD Moffitt Cancer Center Assistant Member Akita Biomedical Consulting Endocrine Tumor Oncology 1111 Bailey Drive Editor Emeritus: Elizabeth M. Sagatys, MD Paso Robles, CA 93446 John Horton, MB, ChB Assistant Member Phone: 805-238-2485 Professor Emeritus of Medicine & Oncology Pathology - Clinical Fax: 949-203-6115 Saïd M. Sebti, PhD E-mail: [email protected] Moffitt Cancer Center Senior Member Journal Advisory Committee: Drug Discovery For Consumer and Daniel A. Anaya, MD Bijal D. Shah, MD General Advertising Associate Member Assistant Member Information: Gastrointestinal Oncology Malignant Hematology Veronica Nemeth Aliyah Baluch, MD Lubomir Sokol, MD, PhD Editorial Coordinator Assistant Member Senior Member Cancer Control: Journal of the Moffitt Cancer Center Infectious Diseases Hematology/Oncology Dung-Tsa Chen, PhD 12902 Magnolia Drive – MBC-JRNL Hatem H. Soliman, MD Associate Member Tampa, FL 33612 Assistant Member Biostatistics Phone: 813-745-1348 Breast Oncology Fax: 813-449-8680 Hey Sook Chon, MD E-mail: [email protected] Assistant Member Jonathan R. Strosberg, MD Gynecological Oncology Assistant Member Gastrointestinal Oncology Jasreman Dhillon, MD Assistant Member Jennifer Swank, PharmD Pathology - Anatomic Clinical Pharmacist - Medical Oncology/Internal Medicine Jennifer S. Drukteinis, MD Associate Member Sarah W. Thirlwell, MSc, MSc(A), RN Diagnostic Radiology Nurse Director Clement K. Gwede, PhD, RN Supportive Care Medicine Program Associate Member Eric M. Toloza, MD, PhD Health Outcomes and Behavior Assistant Member Sarah E. Hoffe, MD Thoracic Oncology Associate Member Nam D. Tran, MD Radiation Oncology Assistant Member John V. Kiluk, MD Neuro-Oncology Associate Member Jonathan S. Zager, MD Breast Oncology Senior Member Cancer Control is a member of Richard D. Kim, MD Cutaneous Oncology the Medscape Publishers’ Circle®, Associate Member an alliance of leading medical Gastrointestinal Oncology publishers whose content is Bela Kis, MD, PhD featured on Medscape Assistant Member (www.medscape.com). Diagnostic Radiology Rami Komrokji, MD Most issues and supplements of Associate Member Cancer Control are available at Malignant Hematology cancercontroljournal.org

Cancer Control: Journal of the Moffitt Cancer Center (ISSN 1073-2748) is published by H. Lee Moffitt Cancer Center & Research Institute, 12902 Magnolia Drive, Tampa, FL 33612. Telephone: 813-745-1348. Fax: 813-449-8680. E-mail: [email protected]. Internet address: cancercontroljournal.org. Cancer Control is included in Index Medicus ®/MEDLINE® and EMBASE®/ Excerpta Medica, Thomson Reuters Science Citation Index Expanded (SciSearch®) and Journal Citation Reports/Science Edition. Copyright 2016 by H. Lee Moffitt Cancer Center & Research Institute. All rights reserved. Cancer Control: Journal of the Moffitt Cancer Center is a peer-reviewed journal that is published to enhance the knowledge needed by professionals in oncology to help them minimize the impact of human malignancy. Each issue emphasizes a specific theme relating to the detection or management of cancer. The objectives of Cancer Control are to define the current state of cancer care, to integrate recently generated information with historical practice patterns, and to enlighten readers through critical reviews, commentaries, and analyses of recent research studies. Disclaimer: All articles published in this journal, including editorials and letters, represent the opinions of the author(s) and do not necessarily reflect the opinions of the editorial board, the H. Lee Moffitt Cancer Center & Research Institute, Inc, or the institutions with which the authors are affiliated unless clearly specified. The reader is advised to independently verify the effectiveness of all methods of treatment and the accuracy of all drug names, dosages, and schedules. Dosages and methods of administration of pharmaceutical products may not be those listed in the package insert and solely reflect the experience of the author(s) and/or clinical investigator(s). Table of Contents

Letter From the Editor

Better Cancer Treatment for All 322 Lodovico Balducci, MD

Editorial

Highlighting Health Disparities in Racial and Ethnic Minorities and Other Underserved Populations 323 Clement K. Gwede, PhD, RN, Gwendolyn P. Quinn, PhD, and B. Lee Green, PhD

Articles

Barriers to Clinical Trial Enrollment in Racial and Ethnic Minority Patients 327 With Cancer Lauren M. Hamel, PhD, Louis A. Penner, PhD, Terrance L. Albrecht, PhD, Elisabeth Heath, MD, Clement K. Gwede, PhD, RN, and Susan Eggly, PhD

Racial and Ethnic Differences in the Epidemiology and Genomics of Lung Cancer 338 Matthew B. Schabath, PhD, W. Douglas Cress, PhD, and Teresita Muñoz-Antonia, PhD

Black Heterogeneity in Cancer Mortality: US-Blacks, Haitians, and Jamaicans 347 Paulo S. Pinheiro, MD, PhD, Karen E. Callahan, MPH, Camille Ragin, PhD, MPH, Robert W. Hage, MD, PhD, DLO, MBA, Tara Hylton, MPH, and Erin N. Kobetz, PhD, MPH

Genomic Disparities in Breast Cancer Among Latinas 359 Filipa Lynce, MD, Kristi D. Graves, PhD, Lina Jandorf, MA, Charité Ricker, MS, Eida Castro, PsyD, Laura Moreno, Bianca Augusto, Laura Fejerman, PhD, and Susan T. Vadaparampil, PhD

Clinical Considerations of Risk, Incidence, and Outcomes of Breast Cancer 373 in Sexual Minorities Anne E. Mattingly, MD, John V. Kiluk, MD, and M. Catherine Lee, MD

Construction and Validation of a Multi-Institutional Tissue Microarray of 383 Invasive Ductal Carcinoma From Racially and Ethnically Diverse Populations Edward Seijo, MS, Diana Lima, MPH, Egiebade Iriabho, MS, Jonas Almeida, PhD, Jesus Monico, MPH, MS, Margarita Echeverri, PhD, Sylvia Gutierrez, MD, Idhaliz Flores, PhD, Ji-Hyun Lee, PhD, Kate Fisher, MA, William E. Grizzle, MD, PhD, Gabriel L. Sica, MD, PhD, Charles Butler, Chindo Hicks, PhD, Cathy D. Meade, PhD, RN, Stephen Olufemi Sodeke, Krzysztof Moroz, MD, Domenico Coppola, MD, and Teresita Muñoz-Antonia, PhD

October 2016, Vol. 23, No. 4 Cancer Control 319 Table of Contents

Articles, continued

Determinants of Treatment Delays Among Underserved Hispanics With Lung and 390 Head and Neck Cancers Evelinn A. Borrayo, PhD, Katie L. Scott, PhD, Ava R. Drennen, PhD, Tiare MacDonald, PhD, and Jennifer Nguyen, MS

Geographical Factors Associated With Health Disparities in Prostate Cancer 401 Scott M. Gilbert, MD, Julio M. Pow-Sang, MD, and Hong Xiao, PhD

Disparities in Penile Cancer 409 Pranav Sharma, MD, Kamran Zargar-Shoshtari, MD, Curtis A. Pettaway, MD, Matthew B. Schabath, PhD, Anna R. Giuliano, PhD, and Philippe E. Spiess, MD

Chemoprevention in African American Men With Prostate Cancer 415 Nagi B. Kumar, PhD, RD, Julio M. Pow-Sang, MD, Philippe E. Spiess, MD, Jong Y. Park, PhD, Ganna Chornokur, PhD, Andrew R. Leone, MD, and Catherine M. Phelan, PhD, MD

Disparities in Adolescents and Young Adults With Cancer 424 Leidy L. Isenalumhe, MD, Olivia Fridgen, MPH, Lynda K. Beaupin, MD, Gwendolyn P. Quinn, PhD, and Damon R. Reed, MD

Tobacco-Related Health Disparities Across the Cancer Care Continuum 434 Vani Nath Simmons, PhD, Bárbara Piñeiro, PhD, Monica Webb Hooper, PhD, Jhanelle E. Gray, MD, and Thomas H. Brandon, PhD

Departments

Case Series: Human Metapneumovirus Infection in Immunocompromised Patients 442 Sharmeen Samuel, MD, Sowmya Nanjappa, MBBS, MD, Christopher D. Cooper, MD, and John N. Greene, MD

Special Report: Systematic Review and Case Series Report of Acinar Cell Carcinoma of the Pancreas 446 Evan S. Glazer, MD, PhD, Kevin G. Neill, MD, Jessica M. Frakes, MD, Domenico Coppola, MD, Pamela J. Hodul, MD, Sara E. Hoffe, MD, Jose M. Pimiento, MD, Gregory M. Springett, MD, PhD, and Mokenge P. Malafa, MD, PhD Special Report: A Single-Institution Study of Demographics and Outcomes of Adult Patients 455 With Multiple Cancers Thanh P. Ho, MD, Nathan R. Foster, MS, and Aminah Jatoi, MD

320 Cancer Control October 2016, Vol. 23, No. 4 Table of Contents

Departments, continued

Ten Best Readings Related to Health Disparities in Oncology 461 Peer Reviewers, 2016 463 Index for 2016, Volume 23 465

About the art in this issue: Bryan Whitney is a photographer based in New York City. His experimental, architectural, and fine art work have been featured in many magazines, books, and has been exhibited in the United States and Europe. He has been the recipient of a Fulbright lecture grant and currently teaches at the In- ternational Center for Photography. To view more of Bryan’s work or to contact him, please visit www.x-rayphotography.com or www.bryanwhitney.com.

Cover: Calla Lillies

Articles: Cherry Blossoms Fiddle Head Cactus Chrysanthemum Stag Horn Proteus Lily Orchid Birds Nest Rose Yellow Tulips Iris

October 2016, Vol. 23, No. 4 Cancer Control 321 Letter From the Editor

Better Cancer Treatment for All

The best scientific and technological advances may reduce cancer-related mortality only when they are fully available to the people at risk. Availability is a multifaceted construct that includes economical and social resources, education, and cultural adherence. This issue of Cancer Control could not be timelier because it addresses the obstacles that prevents access by different minority groups to cancer diagnoses and treatments. The articles in this issue propose practical solutions to this mounting problem. Minority groups include African Americans, Hispanics, Asian Americans, new immigrants, and gay, lesbians, and transgender persons, among others. When taken together, these select minority groups represent a sizable portion of the US population as well as many Americans living with cancer. Given our globalization and the reduced birth rate among white Americans, these minority groups may embody a majority of patients with cancer in the future. These minority groups deserve to become a main focus of cancer control in our country, because the rate of cancer-related mortality is highest among them.1 In the accompanying editorial, our guest editors describe the burden of cancer affecting selected minorities and found that it is largely due to remediable causes, including poor access to screening and early detection, lifestyle choices, cultural beliefs, and limited socioeconomic resources. In addition, they outline some biological issues that have been largely ignored until recently and instead deserve immediate attention, such as the development of deadly types of breast cancer in young African American women. Cancer control in minorities is an ethical and social imperative. In addition to improving the welfare of the general population, it will likely reduce the cost of medical care and provide a trait d’union among the different ethnic/racial, religious, and gender-identity groups that represent the demographical richness of our country. It is not far fetched to assume that the common fight against the suffering related to cancer may cement the alliance of these different groups and provide a sense of common national identity. Older patients with cancer represent another group with increased rates of morbidity and mortality due in part to the inadequate care they receive.2 This group has been the subject of multiple studies and the focus of medical societies,3 but they have not been included in the present review — appropriately in my opinion. My personal experience with the Middle Eastern Cancer Control group, which was led by a visionary Israeli surgeon and included representatives of 32 countries in the Middle East, taught me that the common goal to relieve pain and discomfort caused by cancer may bring peace and cooperation, even in the most war-torn regions of the world.4,5 The cooperation of different cultures, even those from the poorest countries, is a 2-way street. I learned how important spirituality can be in the delivery of palliative care, even when the technology is primitive and the resources are scarce. I wish we could look to our minority patients as an opportunity to learn to provide better cancer treatment to all patients.

Lodovico Balducci, MD Senior Member, Senior Adult Oncology Program Editor, Cancer Control H. Lee Moffitt Cancer Center & Research Institute Tampa, Florida [email protected]

References 1. Tehranifar P, Goyal A, Phelan JC, et al. Age at cancer diagnosis, amenability to medical interventions, and racial/ethnic disparities in cancer mortality. Cancer Causes Control. 2016;27(4):553-560. 2. Balducci L. Studying cancer treatment in the elderly patient population. Cancer Control. 2014;21(3):215-20. 3. Wildiers H, Heeren P, Puts M, et al. International Society of Geriatric Oncology consensus on geriatric assessment in older patients with cancer. J Clin Oncol. 2014;32(24):2595-2603. 4. Silbermann M, Dweib Khleif A, Balducci L. Healing by cancer. J Clin Oncol. 2010;28(8):1436-1467. 5. Silbermann M, Epner DE, Charalambous, A et al. Promoting new approaches for cancer care in the Middle East. Ann Oncol. 2013;24(suppl 7):vii5-vii10.

322 Cancer Control October 2016, Vol. 23, No. 4 Editorial

Highlighting Health Disparities in Racial and Ethnic Minorities and Other Underserved Populations

More than 50 years ago, Martin Luther King Jr said, ceived racism/discrimination and associated stress in “Of all the forms of inequality, injustice in health care minority and underserved populations. Psychologi- is the most shocking and inhumane.”1 Yet more than cal stress caused by perceived discrimination and ste- 30 years after the report from Heckler2 in 1985 on the reotyping, poverty, and differential access to care can national state of minority health in the United States, produce higher rates of depression and other mor- inequities in cancer care have persisted, widened, and, bidities that in turn impact quality of life and health in some cases, emerged in new areas.3-7 Disparities in outcomes.13,14 health status refer to the variation in rates of disease Articles in this issue of Cancer Control examine occurrence and disabilities between populations de- multiple aspects of health disparities in a variety of ra- fined by demographical, socioeconomical, or geo- cial/ethnic minorities and other underserved patient graphical factors, most of which exist in populations populations. This issue is a collection of articles that referred to as minorities and underserved groups.3-5 broadly address health disparities in cancer care from These groups may be classified by race (eg, black, 2 perspectives: going beyond race and ethnicity (ad- Asian), ethnicity (eg, Hispanic, Ashkenazi Jewish), dressing subethnicity and emerging disparate popu- sexual orientation (eg, gay, lesbian), gender (eg, trans- lations, eg, AYAs, gender and sexual minorities) and gender, gender queer), socioeconomic status (eg, low addressing multiple points across the continuum of income, poor, underserved), or age (eg, seniors, adoles- cancer control and contextual factors (eg, geography). cents and young adults [AYAs]).3-7 These articles explore pathways and mechanisms pro- Evidence suggests that health disparities in the pelling cancer-related disparities, and they offer sum- United States result from a combination of poverty, mary evidence to address such disparities. limited access to quality care, perceived racism/dis- A narrative review of the literature by Dr Hamel crimination, genetics, biology, and environmental and colleagues examines participation in clinical exposure.8,9 A common thread in health disparities trials and takes a close look at the barriers to enroll- specific to cancer is their multidimensionality and ment in clinical trials, with an emphasis on the un- complex nature: covering the lifespan, across the der-representation of racial and ethnic minorities in continuum of cancer care from prevention to survi- research studies. The authors identify barriers across vorship, and transcending socioecological and trans- 3 levels — system, individual, and interpersonal — and cultural contexts.5,8,9 A growing body of literature conclude that many of the interventions utilized to ad- suggests that cancer-related disparities are fueled by dress minority accrual to clinical trials have focused on a multifaceted and concurrent interplay of factors that a single level. They suggest that successful interven- range from biology and genetics to lifestyle and be- tions must be multilevel. havior.5,8-10 As such, intersectionality, a theoretical fo- Dr Schabath and fellow researchers review the lit- cus on the interlocking factors of cultural influence erature to summarize the rates of incidence, mortality, and the multiple identities perceived by patients (eg, a and survival among racial and ethnic minorities with racial/ethnic minority who also identifies as a sexual lung cancer. Identifying racial and ethnic differences in minority), helps explain how these complex realities lung cancer will aid in the creation of effective preven- exacerbate disparities.11 tion and treatment strategies that in turn may reduce It is important to consider reducing health dis- the overall burden of lung cancer. parities in terms of modifiable factors. That is, race Dr Pinheiro and others examine a unique as- and ethnicity do not change, but certain contextual pect of disparities research by focusing on the sub- and behavioral aspects, including cancer risk, screen- ethnic groups within blacks. The authors examined ing behavior, and cancer care needs, can be identified the mortality rates among blacks in the United States and addressed. The core of critical interventions re- (eg, blacks, Haitians, Jamaicans) and compared them quires appropriate attention to the cultural and health with their country of origin. Their research sug- literacy levels of individuals and groups.12-14 Further- gests that US-born blacks have the highest risk of more, as a community of clinicians, researchers, and cancer-related mortality when compared with blacks educators, we can change the way cancer control and from other countries. prevention services are delivered by reducing the per- Dr Lynce and colleagues summarize what is

October 2016, Vol. 23, No. 4 Cancer Control 323 known about Latinas and breast cancer, noting the ex- yond racial disparities to include younger age, lower istence of genetic mutations of high and moderate pen- educational and income levels, marital status, lack of etrance within this ethnic population. They conclude circumcision, and a number of sexual behaviors as that reduced awareness and uptake of genetic counsel- factors that compound disparities in this disease. This ing were observed within this group and that tailored article reminds readers to broaden their view of dis- interventions are needed. parities beyond race and ethnicity. Dr Mattingly and fellow researchers summarize Dr Kumar and colleagues tackle a topic fraught the literature on sexual minority women with breast with clinical and scientific uncertainties: screen- cancer, concluding that limited research exists on this ing, early detection, and prevention of prostate can- population primarily due to lack of collection of sexual cer among high-risk African American men. Their orientation information in databases. work examines biological etiology, biomarkers, and Mr Seijo and coauthors address the lack of banked pathways, specifically underscoring the need for che- tissue among racial and ethnic minorities. Their team moprevention agents and strategies. This synthesis reports on an established infrastructure for a biobank- identifies unique potential pathways that drive excess ing hub in the southeastern United States and Puerto risk and burden of disease and the potential of green Rico to create a web-based database and tissue micro- tea catechins as a chemoprevention agent in African arrays for breast cancer. American men diagnosed with high-grade prostatic in- Dr Borrayo and colleagues examine the determi- traepithelial neoplasia. nants of treatment delays among underserved His- Dr Isenalumhe and fellow researchers examine panics with lung and head and neck cancers. They the literature on AYAs with cancer for whom survival report that women experience longer delays than rates have not improved as they have in younger and men. Other observed factors associated with longer older patients. Disparities in this population range delays include fewer comorbidities, receiving treat- from the unique biology of the cancer itself to delays ment at a public facility, and being underserved (eg, in diagnoses further exacerbated by the psychoso- uninsured). This article identifies actionable factors cial needs of this population (eg, reproductive pres- amenable to intervention to improve outcomes in dis- ervation). Their article focuses on identifying these advantaged Hispanic patients. unique needs and the research advancements made Dr Gilbert and others take an innovative ap- to improve outcomes and quality of life in AYAs. proach to addressing racial and ethnic disparities in Dr Simmons and coauthors attend to racial/ethnic prostate cancer treatment and outcomes by going and socioeconomic disparities in tobacco use, one of beyond the traditional factors of race and socioeco- the leading preventable causes of lung cancer–related nomic factors. Their focus is on the geographic dis- death. Their contribution focuses attention on impor- tribution of disparities (ie, how a place of residence tant disparities that can occur across the cancer con- interacts with race and socioeconomic factors). Using tinuum — from prevention to survivorship. They syn- the Geographic Information System, they illustrate thesize the research that has reduced tobacco-related an example showing locations in the state of Florida where disparities in prostate cancer Race and ethnicity are terms often interchangeably used, although are the most problematic. The they describe 2 different but complex concepts. They are often com- authors identify geographical ingled in describing populations in health care and health disparities differences in treatment across research. In addition, the criticism is growing about the shortcomings racial and socioeconomic stra- of using these classifications in health research. ta, thus availing a better tool for Race is a social classification (not a biological category) that de- public health officials, health scribes a person’s observable physical characteristics and identity systems, health care profes- based on phenotypic genetic characteristics such as anatomical or sionals, and policymakers to bone structure and skin, hair, or eye color. use when addressing health Ethnicity describes the shared identity of a population based on disparities. sociocultural factors, linguistic or religious traits, or national heri- Dr Sharma and coauthors tage, regardless of race that often shape that group’s world-view, life- address a less-common cancer style, and behaviors. and rare disparity in developed Individuals can be of “mixed race or more than 1 race” and can nations, penile cancer, in which have more than 1 ethnicity. Advances in genomic and precision racial differences exist in both medicine are helping to elucidate the genetic and biological basis of rates of disease incidence and health disparities. associated survival outcomes. Their chief findings extend be-

324 Cancer Control October 2016, Vol. 23, No. 4 disparities, pointing to the limited successes of both 4. National Institutes of Health; National Cancer Institute. About cancer health disparities. https://www.cancer.gov/about-nci/organization/crchd/ pharmacotherapy and behavioral counseling for racial about-health-disparities. Accessed October 3, 2016. and ethnic minorities, concluding with a critical call for 5. Alcaraz K, Sly J, Ashing K, et al. The ConNECT Framework: a model for advancing behavioral medicine science and practice to foster health innovative efforts directed at each point along that con- equity. J Behav Med. 2016. Epub ahead of print. tinuum. These authors remind us of the critical role of 6. Quinn GP, Sanchez JA, Sutton SK et al. Cancer and lesbian, gay, bisexual, transgender/transsexual, and queer/questioning (LGBTQ) culture and socioeconomic status in addition to racial populations. CA Cancer J Clin. 2015;65(5):384-400. and ethnic status. 7. Committee on Lesbian, Gay, Bisexual, and Transgender Health Issues and Research Gaps and Opportunities; Board on the Health of The Centers for Disease Control and Prevention Select Populations; Institute of Medicine. The Health of Lesbian, Gay, suggests the key to reducing cancer-related inequities Bisexual, and Transgender People: Building a Foundation for Better Understanding. Washington, DC: National Academies Press; 2011. is to increase screening and early detection, promote 8. Freeman HP. Poverty, culture, and social injustice: determinants of healthy behaviors, and expand access to health care for cancer disparities. CA Cancer J Clin. 2004;54(2):72-77. 3 9. Freeman HP, Chu KC. Determinants of cancer disparities: barriers underserved populations and groups at greatest risk. to cancer screening, diagnosis, and treatment. Surg Oncol Clin N Am. The articles in this issue of Cancer Control help us 2005;14(4):655-669. 10. Dressler WW, Oths KS, Gravlee CC. Race and ethnicity in public identify the ever-expanding definition of what it means health research: models to explain health disparities. Ann Rev Anthropol. to be “at risk” as well as how minority and underserved 2005;34:231-252. 11. Hsieh N, Ruther M. Sexual minority health and health risk factors: populations are disproportionately affected by an ex- Intersection effects of gender, race, and sexual identity. Am J Prev Med. cess burden of cancer. These articles also point us in 2016;50(6):746-755. 12. Kagawa-Singer M, Dressler WW, George SM, et al. The Cultural the direction of key areas for expanding cancer care Framework for Health: An Integrative Approach for Research and Program and new research to help us identify and reduce can- Design and Evaluation. Bethesda, MD; National Institutes of Health; 2016. 13. Smedley BD, Stith AY, Nelson AR, eds; Committee on Understanding cer-related inequities. This issue highlights promising and Eliminating Racial and Ethnic Disparities in Health Care. Unequal strategies for reducing the cancer-related burden and Treatment: Confronting Racial and Ethnic Disparities in Health Care. Washington, DC: National Academies Press; 2003. improving the quality of life of all patients. A height- 14. LaVeist TA, Isaac LA, eds. Race, Ethnicity, and Health: A Public ened understanding of both differences and similari- Health Reader. 2nd ed. San Francisco, CA: Josey-Bass; 2012. ties across age, sex, gender, race, ethnicity, socioeconomic class, and sexual orientation is essential to the delivery of equitable cancer care. It is incumbent upon us to focus on ways to reduce the cancer burden and disparities in rates of cancer incidence, morbidity, and mortality so that no one is left behind.

Clement K. Gwede, PhD, RN Associate Member, Health Outcomes and Behavior Program H. Lee Moffitt Cancer Center & Research Institute Tampa, Florida [email protected]

Gwendolyn P. Quinn, PhD Senior Member, Health Outcomes and Behavior Program H. Lee Moffitt Cancer Center & Research Institute Tampa, Florida [email protected]

B. Lee Green, PhD Vice President, Diversity, Public Relations and Strategic Communications Senior Member, Health Outcomes and Behavior Program H. Lee Moffitt Cancer Center & Research Institute [email protected]

References

1. King ML Jr. Presentation at: Second National Convention of the Medical Committee for Human Rights; Chicago, IL; March 25, 1966. 2. Heckler MM. Report of the Secretary’s Task Force on Black and Minority Health. Washington, DC: US Department of Health and Human Services; US Government Printing Office; 1985. 3. Centers for Disease Control and Prevention. Health disparities in cancer. http://www.cdc.gov/cancer/dcpc/resources/features/cancerhealthdisparities. Accessed September 26, 2016.

October 2016, Vol. 23, No. 4 Cancer Control 325 Health Outcomes and Behavior Program H. Lee Moffitt Cancer Center & Research Institute

The Health Outcomes and Behavior Program contributes to the prevention, detection, and control of cancer through the study of health-related behaviors, health care practices, and health-related quality of life. Work toward this goal involves research across the disease spectrum, from prevention and detection through survivorship or advanced disease. The program’s 4 specific aims are to: (1) understand the determinants of health behaviors that can lead to the prevention and early detection of cancer and develop effective methods of promoting those behaviors, (2) understand and improve the quality of life of patients and family members throughout the disease course, (3) contribute to the evidence base and synthesis of evidence regarding the delivery of cancer care and clinical outcomes, and (4) understand and intervene upon the social, cultural, and behavioral determinants of cancer-related health disparities. Awareness is growing of the disparities that occur across the cancer continuum, including prevention, incidence, diagnosis, treatment, and outcomes. These disparities occur with respect to sex, gender identity, race, ethnicity, socioeconomic status, and other patient characteristics. To understand and address these disparities, program members conduct community-based participatory research throughout the catchment area. Program members have translated their earlier observational research into numerous ongoing, randomized controlled trials of interventions designed to reduce a range of cancer-related health disparities.

DEPARTMENT CHAIR/PROGRAM LEADER Thomas H. Brandon, PhD

VICE DEPARTMENT CHAIR Susan Thomas Vadaparampil, PhD

Faculty Daniel A. Anaya, MD Clement K. Gwede, PhD, RN Carmen S. Rodriguez, PhD, ARNP-BC Lodovico Balducci, MD Ming Ji, PhD Richard G. Roetzheim, MD Alicia Best, PhD Heather S. Jim, PhD Lori A. Roscoe, PhD Margaret Booth-Jones, PhD Julio C. Jiminez Chavez, MD Julian Sanchez, MD Ellen M. Daley, PhD Peter A.S. Johnstone, MD Vani Nath Simmons, PhD Benjamin Djulbegovic, MD, PhD Cecile A. Lengacher, RN, PhD Brent J. Small, PhD Kristine A. Donovan, PhD Dinorah Martinez Tyson, PhD Marilyn Stern, PhD David J. Drobes, PhD Susan C. McMillan, PhD, RN Steve K. Sutton, PhD David Evans, PhD Cathy D. Meade, PhD, RN Tawee Tanvetyanon, MD Martine Extermann, MD, PhD Teresita Muñoz-Antonia, PhD Lora M.A. Thompson, PhD Michael Fradley, MD Michael A. Poch, MD Christine Vinci, PhD Scott M. Gilbert, MD Gwendolyn P. Quinn, PhD Hsiao-Lan Wang, PhD B. Lee Green, PhD Maija Reblin, PhD

For more detailed information about the Health Outcomes and Behavior Program, please visit www.moffitt.org/research-science/research-programs/health-outcomes-and-behavior

www.moffitt.org

326 Cancer Control October 2016, Vol. 23, No. 4 Barriers to racial and ethnic minorities

enrolling in clinical trials exist at system,

individual, and interpersonal levels.

Cherry Blossoms. X-Ray image courtesy of Bryan Whitney. www.x-rayphotography.com.

Barriers to Clinical Trial Enrollment in Racial and Ethnic Minority Patients With Cancer Lauren M. Hamel, PhD, Louis A. Penner, PhD, Terrance L. Albrecht, PhD, Elisabeth Heath, MD, Clement K. Gwede, PhD, RN, and Susan Eggly, PhD

Background: Clinical trials that study cancer are essential for testing the safety and effectiveness of promising treatments, but most people with cancer never enroll in a clinical trial — a challenge exemplified in racial and ethnic minorities. Underenrollment of racial and ethnic minorities reduces the generalizability of research findings and represents a disparity in access to high-quality health care. Methods: Using a multilevel model as a framework, potential barriers to trial enrollment of racial and ethnic minorities were identified at system, individual, and interpersonal levels. Exactly how each level directly or indirectly contributes to doctor–patient communication was also reviewed. Selected examples of implemented interventions are included to help address these barriers. We then propose our own evidence-based intervention addressing barriers at the individual and interpersonal levels. Results: Barriers to enrolling a diverse population of patients in clinical trials are complex and multilevel. Interventions focused at each level have been relatively successful, but multilevel interventions have the greatest potential for success. Conclusion: To increase the enrollment of racial and ethnic minorities in clinical trials, future interventions should address barriers at multiple levels.

Introduction treatments and translating new knowledge into tangi- Clinical trials focused on cancer research are essen- ble benefits for patients; they also represent options for tial for testing the safety and effectiveness of potential novel therapy for cancer.1,2 However, approximately 2% to 3% of all patients with cancer ever enroll in a tri- From the Department of Oncology (LMH, LAP, TLA, EH, SE) and Popu- al.3,4 Estimates of the number of trials that fail to meet lation Studies Disparities Research Program (LMH, LAP, TLA, SE), Wayne State University, Karmanos Cancer Institute, Detroit, Michigan, scientific objectives because of insufficient accrual and the Department of Oncologic Sciences/Population Sciences (CKG), rates range from 22% to 50%.5,6 Low accrual rates jeop- H. Lee Moffitt Cancer Center & Research Institute, Tampa, Florida. ardize the ability of researchers to assess the safety and Submitted January 26, 2016; accepted May 3, 2016. effectiveness of new approaches to cancer care, wastes Address correspondence to Lauren M. Hamel, PhD, Wayne State resources, precludes follow-up studies, and reduces University, Karmanos Cancer Institute, Department of Oncology, Population Studies Disparities Research Program, 4100 John R the ability of our clinical research system to translate Street, MM03CB, Detroit, MI 48201. E-mail: [email protected] research into evidence-based practice.6-8 This work was funded in part by National Institutes of Health/ Underenrollment is an even greater challenge National Cancer Institute grant no. R01CA200718-01. among racial and ethnic minorities — particularly No significant relationships exist between the authors and the companies/organizations whose products or services may be referenced African Americans — despite a requirement by the in this article. National Institutes of Health that members of minority

October 2016, Vol. 23, No. 4 Cancer Control 327 populations be represented in clinical research.1,3,4,9-11 larly African Americans — bear the greatest cancer bur- A systematic review compared the proportion of under- den in the United States, so they should be adequately represented minority participants in phase 3 cancer represented in cancer research.17-19 Under-representa- treatment and prevention clinical trials conducted be- tion of racial and ethnic minorities also limits the gen- tween the periods 1990 to 2000 and 2001 to 2010.12 In the eralizability of research findings.20-22 The Institute of treatment studies conducted between 2001 and 2010 that Medicine has recommended that every individual with reported race/ethnicity, the reviewers found that 82.9% cancer have access to high-quality clinical trials, so we of participants were white, 6.2% were African Ameri- believe that under-representation of racial and ethnic can, 3.3% were Asian, 2.2% were Hispanic, and 0.1% minorities represents a disparity in health care.2 Under- were Native American.12 This is in contrast to studies enrollment of racial and ethnic minorities in clinical tri- conducted between 1990 and 2000, in which 89% of als may therefore contribute to preventable disparities participants were white, 10.5% were African American, in treatment outcomes and survival.1,17,23,24 0.4% were Hispanic, and 0.04% were Asian.12 In other words, even though the proportion of white partici- Purpose pants decreased, whites continued to comprise a large The purpose of this paper is to identify and describe majority of participants in cancer treatment trials, and potential barriers to the enrollment of racial and eth- the proportion of African American participants de- nic minorities in clinical trials at the system, individual creased between the periods 1990 to 2000 and 2001 (health care professional, patient, and family), and in- to 2010. However, during those same periods, the terpersonal levels (eg, doctor–patient relationship). proportion of African American participants in clini- The paper also describes selected examples of evi- cal prevention trials increased (5.5% from 1990–2000, dence-based interventions already implemented to ad- 11.6% from 2001–2010).12 dress some of these barriers at the patient-, physician-, Although several patient populations are under- and doctor–patient interpersonal communication represented in clinical trials, including elderly patients levels. We take this approach because multilevel inter- (≥ 65 years), residents of rural areas, and those with low ventions, as compared with single-level interventions, socioeconomic status, the current review focuses on may have the greatest potential to achieve substantial racial and ethnic minority underenrollment for several and sustained change and to produce additive — and reasons.4,10,13-16 Racial and ethnic minorities — particu- possibly multiplicative — effects.25,26

System Level Availability of clinical trials Supporting infrastructure Cost/payer Strict eligibility criteria Lack of community engagement

Individual Level Interpersonal Level Health Care Professional Perception of the institutional infrastructure Disparities • Awareness of clinical trials Attitude/perception of each other and care received • Perception of the hospital/infrastructure Doctor–patient relationship • Attitude/bias/experience • Perception of patient’s race/ethnicity Patient/Family • Awareness of clinical trials • Eligibility • Personal identity: ethnicity, race, religion • Attitude/bias/experience (eg, suspicion of medical care) • Access (eg, transportation)

Fig 1. — A multilevel model of factors contributing to disparities in clinical trials.

328 Cancer Control October 2016, Vol. 23, No. 4 We propose a multilevel model derived from al, regional, and consortia efforts have implemented general systems theory, which posits that organiza- programs to increase enrollment among racial and eth- tions are comprised of discrete but related levels nic minority populations.8,41,42 For example, the Minori- (Fig 1).27 Our multilevel model demonstrates how ty-Based Community Clinical Oncology Program of the system- and individual-level barriers contribute to National Cancer Institute was designed to increase mi- interpersonal-level barriers to the enrollment of racial nority participation in cancer clinical trials by develop- and ethnic minorities in clinical trials. ing outreach efforts at health care institutions that serve large numbers of minority patients with cancer.32 Mc- System-Level Barriers Caskill-Stevens et al32 reported that the success of this Barriers at the level of health care systems and hos- program was largely due to system- and hospital-level pitals include the limited number of trials nationally factors, such as opening trials that matched the clinical and regionally available; hospital infrastructures that characteristics of patient populations and developing lack the resources to support trials; financial costs to relationships with local physicians and cancer advoca- hospitals and patients; restrictive study designs and cy groups that increased their willingness to enroll or eligibility criteria; and lack of community engagement. refer minority patients to clinical trials. However, their Many of these barriers have a disproportionate effect success was limited by lack of funding for minority out- on minority enrollment because members of minority reach activities, lack of system support to provide staff populations are more likely to receive care at under- and mentoring for minority investigators, and lack of resourced hospital systems where few clinical trials funding to provide protocol-related drugs and related are available, to be underinsured, and to present with services for uninsured patients.32 comorbidities that make them ineligible for trials that The Community Oncology Research Program of may be available.28-30 the National Cancer Institute is a national network of Available clinical trials supported by a well-re- health care professionals, investigators, and organiza- sourced and functional infrastructure are critical to the tions that conducts cancer-related research across the enrollment of minority patients.31 For example, this in- United States.43 The program focuses on determining frastructure includes staff dedicated to enrolling, man- the reasons for the existence of racial and ethnic dis- aging, and tracking participants; data collection and parities in clinical trial participation, as well as ways to management capabilities; and an efficient Institutional increase the participation of racial and ethnic minori- Review Board. Without a functional infrastructure, ef- ties in clinical trials through numerous interventions.43 forts at increasing the enrollment of minorities in other Because its inauguration was in 2014, evaluation data levels are less likely to be successful.31-33 The financial have not yet been published. costs of enrollment to institutions and individual pa- In recent years NRG Oncology convened a work- tients affect all patient populations, but they have a shop on the challenges and opportunities of clinical disproportionate impact on minorities compared with trial enrollment in an attempt to address the issue of nonminorities. Minorities are more likely to be un- minority underenrollment in clinical trials.44 Experts derinsured, to seek care at under-resourced hospitals, in oncology, including members of the Minority-Based and to have concerns about the cost of participating Community Clinical Oncology Program of the National in a clinical trial.29,34 In addition, lack of or inadequate Cancer Institute, and a large patient advocacy group health insurance acts as a barrier to enrollment in clini- were in attendance.44 Ten themes emerged regarding cal trials for several under-represented populations, in- future plans and interventions to increase minority en- cluding racial and ethnic minorities.29,35 rollment in clinical trials; of these, 8 themes focused on As a way to enhance community engagement, the the system level44: infrastructure of clinical trials should exist within a 1. Target recruitment and emphasize the personal- medical institution that has an established, trusted rela- ized nature of trials as a way to improve minority tionship with the members of the community it serves.36 enrollment to specialized studies Building trust within a community requires that an 2. Define study populations at the molecular level institution conduct its research in an open and honest rather than by traditional and less-precise eligibil- manner that involves a complete and accurate descrip- ity criteria tion of the known and potential differences in the risks 3. Make the delivery of cancer care research avail- and benefits for different racial and ethnic populations. able at the community and academic levels so Ways and means to facilitate and encourage open dis- studies are designed to better improve health cussions into research ethics, including the past abuse of outcomes minority patients in research, are needed, as are efforts 4. Improve local infrastructure by providing transla- to create and sustain partnerships with the community tion and navigation services and research-friendly to share ownership of the research.37-40 electronic medical records In response to system-level barriers, several nation- 5. Expand and standardize the collection of demo-

October 2016, Vol. 23, No. 4 Cancer Control 329 graphics and real-time information about accruals For example, health care professionals may not fully for all patient populations agree with or understand the scientific value of trials 6. Strengthen the infrastructure of information in general or the details of specific trials; they may feel technology they do not have adequate system support; or they may 7. Improve recruitment, training, and mentorship of have concerns about practical issues such as strict pro- young investigators, particularly those of racial tocol designs, patient inconvenience, and added work and ethnic minority backgrounds and those with for the health care staff.33,45 Some clinicians find it dif- an interest in cancer disparities ficult to reconcile the roles of physician and researcher, 8. Continue and improve budgetary support from or are concerned about unduly influencing patient de- the US Congress for all clinical cancer research cisions about enrolling in a clinical trial.47 within the National Cancer Institute Some attitudinal factors are specific to the enrollment of minority patients, such as the concern about Individual-Level Barriers harming the therapeutic relationship. For example, a fo- Our multilevel model emphasizes individual factors that cus-group study involving community physicians found may present a barrier to the enrollment of racial and eth- that some physicians were hesitant to discuss clinical tri- nic minorities in clinical trials. The term individuals is als with their African American patients because of their meant to refer to persons directly involved in the clinical perceptions that African Americans are often mistrust- trial enrollment process, including health care profes- ful of physicians and medical institutions.48 sionals, patients, and their family members. Barriers on An additional attitudinal factor that may deter a system level can influence how individuals perceive health care professionals from discussing trials with the enrollment of minority patients in clinical trials and minority patients is implicit bias against members of how they communicate about trials. For example, an in- minority groups. Prior research has shown that implicit vestigation of physician perceptions of clinical trials and bias (comprised of conscious and unconscious biases) trial support available at their institution revealed that among physicians toward members of minority groups most physicians held a favorable attitude about clinical has an impact on clinical interactions with minority pa- trials, both as a source of high-quality patient care and tients.49-52 We are unaware of studies that specifically as a professional benefit to themselves.33 However, ap- investigate whether the implicit bias of health care pro- proximately two-thirds of practicing oncologists per- fessionals is a barrier to the enrollment of minorities ceived a lack of needed infrastructure to conduct trials.33 in clinical trials, but research in other contexts sug- This included a low number of trials open and lack of gests that some physicians may have implicit negative nonphysician staff to support patient participation in attitudes about African Americans that may lead them clinical trials.33 Similarly, the health care system can also to believe they will be poor candidates for clinical tri- influence patient concerns about the cost of participat- als.53-56 For example, research has shown that primary ing in trials and, in turn, their decisions to enroll or not care physicians tend to trust minority patients less than enroll in a clinical trial.34 Thus, these 2 levels are linked they trust white patients.54 Research has also shown in terms of whether and how health care professionals that physicians high in implicit bias may view their and minority patients discuss clinical trials and whether minority patients as less likely to comply with recom- minority patients agree to enroll. mended treatment,57 and that physicians’ diagnoses and treatment of African Americans can be influenced Health Care Professionals by racial stereotypes.55 This body of research suggests Health care professionals play a critical role in facilitat- that physicians may limit offers for trial enrollment to ing or inhibiting the enrollment of a diverse population those patients they perceive as good study candidates of patients into clinical trials.33,45,46 Several factors are so that the studies will be conducted in a timely and related to the reluctance of health care professionals to efficient manner.58 This may mean physicians are less enroll their patients in trials, regardless of whether or likely to offer a clinical trial to patients who are racial not they are members of minority groups themselves. and ethnic minorities than to white patients, and re- A major barrier is the lack of awareness of avail- search has shown this to be the case.46,59 able trials: Health care professionals must be made Recommendations have been made for evidence- aware of national, regional, and local trials currently based approaches to reduce the impact of health care taking patients and the eligibility criteria for each of professional bias in clinical interactions.60 One ap- those trials to discuss such enrollment with their pa- proach is to train health care professionals in the use of tients.45,46 In addition, the attitudes of health care pro- high-quality patient-centered communication, which fessionals about trials and discussing such trials with requires the development of a positive interpersonal patients and their families present a major barrier to relationship.61 Patient-centered communication focuses trial enrollment, and they likely affect the quality of on patient needs and perspectives, and it does not ig- communication during discussions of clinical trials.33,45 nore the racial and ethnic background of the patient,

330 Cancer Control October 2016, Vol. 23, No. 4 because doing so denies an important part of an indi- improve knowledge of, attitudes about, and participa- vidual.61 Results of the NRG Oncology workshop iden- tion in clinical trials among racial and ethnic minor- tified the commitment of physicians and study investi- ity patients has been developed and tested.70,71 A mul- gators to the enrollment of minorities into clinical trials ticenter, randomized trial of a web-based, interactive, as critical to the future of clinical cancer research.44 educational tool was designed to increase knowledge, They also emphasized that physicians must be cultur- decrease attitudinal barriers, and improve preparation ally sensitive and aware of the impact appropriate com- for making decisions about clinical trial enrollment munication and patient trust have on minority enroll- among minority patients with cancer.70 At baseline, ment in clinical trials.44 the level of knowledge about clinical trials, attitudi- Several communication curricula are available, but nal barriers to participation, and preparation for deci- they include little training on discussing clinical trials sion-making among the participants were assessed.70 with patients or with specific patient populations.62 Prior to their initial visit with an oncologist, patients in One exception is an Australian study, which reported the intervention arm watched a set of brief, individu- on a 1-day intensive course for physicians to improve ally tailored videos addressing knowledge and attitu- aspects of their communication related to discussing dinal barriers to clinical trial participation, whereas clinical trials with patients.45 Results showed that phy- those assigned to the control arm received text-based sicians improved in some shared decision-making be- general information about clinical trials from the Na- haviors and were more likely to describe some of the tional Cancer Institute. The study results showed key clinical or ethical aspects of the trial.63 However, that, compared with baseline levels, participants in we are unaware of any specific training curriculum the intervention and control arms both showed im- that emphasizes ways to communicate about clinical proved knowledge (control arm: mean [standard trials with minority patients. deviation {SD}] = 2.5 [3.1]; intervention arm: mean [SD] = 3.2 [3.1]; P < .001) and decreased attitudinal Patients and Families barriers (control arm: mean [SD] = –0.2 [0.4]; interven- Some major barriers to enrolling racial and ethnic mi- tion arm: mean [SD] = –0.3 [0.5]; P < .001).70 However, nority patients in clinical trials are that patients are of- patients in the intervention arm had a significantly ten not aware they are eligible for an existing trial; they greater increase in knowledge and decrease in attitu- are often underinsured and cannot afford the extra ex- dinal barriers compared with patients in the control penses related to trial participation; and they may not arm (P < .001).71 All patients had a significant increase meet the clinical criteria due to comorbid conditions or in their preparedness to consider participation in age restrictions.29,30,64 Similar to health care profession- clinical trials (control arm: mean [SD] = 3.4 [13.5]; in- als, patients hold attitudes and beliefs that may also af- tervention arm: mean [SD] = 4.7 [12.8]; P < .001).70 fect their willingness to participate and the way they Another intervention specifically addressed the communicate during clinic visits in which trials are dis- negative perceptions and attitudes of patients about cussed.65,66 Despite this, research shows that, overall, clinical trials by assessing the effectiveness of a multi- minorities are as likely as whites to consent if they are media, psychoeducational intervention compared with offered a trial.30,59,67,68 For example, an assessment of printed educational materials.71 The intervention was more than 4,000 racially diverse patients with cancer designed to improve patient attitudes and knowledge found no significant association between race and eth- of clinical trials, their ability to effectively make deci- nicity and refusal to participate in a clinical trial or lack sions, their receptivity to receiving more information, of desire to participate in research.30 Similarly, Katz et and their willingness to participate in clinical trials.71 al69 implemented the Tuskegee Legacy Project to ad- The researchers observed that study participants who dress and understand if and how the Tuskegee syphilis received the intervention had more positive attitudes experiment impacted the trial recruitment and reten- toward clinical trials (mean [SD] = 0.21 [0.01]; P = .02), tion of African Americans into biomedical studies. The and, although the 2 interventions were not developed researchers found no difference in self-reported will- and tested with minority patients, the study results ingness among African Americans to participate in bio- showed no significant difference between interven- medical research.69 A systematic review of the available tion by patient demographics.71 Thus, the intervention research on the factors that influence participation in could potentially be adapted to specifically benefit -ra clinical trials among African Americans concluded that cial and ethnic minority patient populations. the strongest inhibitors of participation were low levels Although racial and ethnic minority patients are of knowledge and lack of awareness of clinical trials.66 overall as likely as white patients to agree to enroll The most important facilitators of participation were in a clinical trial, reasons for declining may differ by social support and recommendations from physicians, racial and ethnic background.30,59,66-68 Minority pa- family members, and friends.66 tients — particularly African Americans — may hold Several patient-focused interventions designed to negative race-related attitudes and beliefs that could di-

October 2016, Vol. 23, No. 4 Cancer Control 331 rectly and indirectly influence interactions with their intersection between the impact of patient attitudes health care professional regarding trials and decisions and access to trial participation illustrates the complex about enrollment.38,40,66,72 These attitudes, which have and multilevel nature of this issue and the need for been derived in large part from issues of racism, un- more comprehensive interventional strategies. informed consent, and poor health care for minorities in the United States, include greater mistrust in medi- Interpersonal-Level Barriers cal institutions and health care professionals and a Our model suggests that system- and individual-level greater sense of having been the target of discrimina- barriers to minority enrollment in clinical trials af- tion.20,23,29,38,69,72-78 A systematic review of the research fect whether and how trials are discussed during assessing the multiple barriers to minority enrollment doctor–patient clinical interactions. This suggestion is in clinical trials found that mistrust in research and the based on our work and that of others showing that pa- medical system was the most common barrier to patient and physician attitudes affect the quality of com- tients’ willingness to participate in a clinical trial.29 munication during clinical interactions with African The attitudes of African American patients toward Americans.80-88 Our model also partly explains why medical research were assessed using focus groups of communication during these interactions is often of a African American adult patients from a single urban lower quality and may affect racial and ethnic minori- public hospital.72 Results showed that most were in fa- ties’ decisions about enrollment. vor of medical research as long as they were not treat- For example, implicit bias among health care pro- ed as “guinea pigs.”72 The authors also reported that fessionals (eg, oncologists) negatively affects communi- many of the participants had a limited understanding cation with minority patients.89 An investigation of the of the informed consent process, and many patients as- impact of the implicit racial bias of oncologists on their sumed the consent form was protecting hospitals and clinical interactions with African Americans with can- doctors from any legal responsibility, not to protect pa- cer revealed that oncologists with higher rates of im- tients.72 Another study that assessed and compared the plicit bias against African Americans had shorter interracial differences in factors that influence patient will- actions with African Americans than oncologists with ingness to enroll in clinical research studies found that lower levels of implicit bias.90 In addition, oncologists African Americans were less willing to participate in with higher rates of implicit bias used less patient-cen- medical research if they attributed a high importance tered communication.90 Although this study was not to the race of the physician when seeking routine med- specifically focused on clinical trials, the findings sug- ical care and believed that minorities bear most of the gest that implicit bias among health care professionals risks of medical research.40 may affect clinical interactions with African Americans Interventions designed to increase the recruitment who are eligible or express interest in participating in a of minorities in clinical trials are generally focused on clinical trial.90 communities, rather than on individuals.32,79 One such Racial and ethnic differences in the quality of com- intervention targeted minority recruitment using stra- munication are due to both patient-related commu- tegic planning, which included meetings and confer- nication behaviors (eg, African Americans often ask ences with key stakeholders and minority organiza- fewer questions),82,86 general communication behaviors tions as a means to increase minority enrollment.79 The among physicians, including amount of patient cen- authors examined institutions that employed these teredness, information giving, and shared decision- strategies and those that did not, and they found a sig- making,61,81,86,91 and communication specific to dis- nificant increase in the rate of minority accrual in insti- cussing trials.45,63,92-95 Although few researchers have tutions that implemented strategic planning compared specifically focused on discussions with minority pa- with those that did not.79 This effort suggests that the tients about enrolling in cancer trials, a single study of mistrust of the medical system seen among minority video-recorded interactions was undertaken in which patients might be addressed with culturally sensitive oncologists discussed clinical trials with whites and Af- and open community engagement. rican Americans.96 Using linguistic analyses, the inter- Other identified patient-level barriers to the enroll- actions with African Americans compared with whites ment of racial and ethnic minorities in clinical trials were found to be shorter, the topic of clinical trials was include lack of transportation, inadequate insurance, less frequently mentioned, and, when clinical trials lack of childcare, and poor access to health care.42,66 were mentioned, less time was spent discussing them.96 The NRG Oncology workshop also identified direct-to- Compared with whites, differences were observed in patient communication and advertising as critical for the discussion of some of the key aspects of clinical tri- the future of minority enrollment in clinical trials, em- als: Oncologists and African American patients spent phasizing that this should be done with collaboration less time discussing the purpose of the trial, its risks with key stakeholders such as community groups, sur- and benefits, and alternatives to participating in the vivor advocacy groups, and religious institutios.44 The trial; however, they spent more time discussing the vol-

332 Cancer Control October 2016, Vol. 23, No. 4 untary nature of trials.96 These findings are particularly adapted for specific populations. problematic because very few oncologists are African The clinical interaction is based on a relationship, American, and so racially discordant interactions may so focus on the social interaction between patients and be considered the norm for African American patients physicians from differing social backgrounds may be with cancer.96,97 valuable. Research from the field of social psychology Few patient-focused interventions designed to im- suggests that patients and physicians in racially discor- prove doctor–patient discussions of clinical trials exist, dant clinical interactions should create a sense of com- and none of them focus on minority patient popula- mon purpose, support, and understanding as ways to tions.98 One study assessed an intervention designed reduce bias and increase cooperation and trust.101,102 to improve patient understanding of early-phase clini- This approach formed the basis of a “team” interven- cal trials for which they were eligible and which would tion in a primary care clinic that succeeded in increas- be discussed during a subsequent oncology visit.99 Pri- ing patient trust and adherence following such clinic or to the visit with the oncologist to discuss the trial, visits.103 Thus, this type of approach may reduce the study participants in the intervention group viewed an impact of racial bias in interactions when clinical trials interactive, computer-based presentation about clini- could be discussed. cal trials that included information about several aspects of clinical trials and urged them to discuss the Proposed Multilevel Interventional Model risks and benefits of the trial with their oncologist; Our multilevel model and supporting research identify those assigned to the control group were provided a and describe factors that may inhibit the enrollment of pamphlet that covered similar topics.99 Following the minorities in clinical trials. Authors have challenged clinic visit, participants assigned to the intervention researchers to move beyond reductionist, single-level were more likely than the controls to understand that interventions.25,26,104 We have met this challenge by the purpose of an early-phase trial was related to the proposing a multilevel intervention comprised of evi- safety and dosing of the drug; however, the majority dence-based strategies focused at the patient, oncolo- of participants in both groups reported that the main gist, and patient–oncologist interpersonal levels. Al- purpose of an early-phase trial was “to see if the drug though the proposed intervention does not directly works.”99 In addition, those in the intervention group address the system level, addressing these other 2 lev- were also more likely than the control group to believe els may, in combination, have a greater and longer-last- that the physician talked to them about joining an ear- ing impact on increasing enrollment among minority ly-phase trial because they might benefit from the drug patients (Fig 2). Ideally, an intervention should address compared with controls. Brown et al100 developed and all levels; however, in reality, it may be difficult for all pilot tested another patient-focused intervention that interventions to address all of them. provided eligible patients with a booklet with trial-re- Our proposed intervention builds on previously lated questions they could ask their oncologist during published descriptive and interventional research on the clinic visit. Findings suggested that the study pa- individual attitudes and clinical communication by tients had high informational needs prior to the visit, addressing these 3 levels of barriers. The intervention selected most of the questions to ask their oncologist, places emphasis on the communication about clinical and asked many of the questions during the visit.100 Al- trials that occurs between patients and physicians in though these interventions were not tested in popula- clinical settings. We place the focus on clinical commu- tions with minority patients, they could potentially be nication for 3 reasons:

Individual Level Health Care Professional Training to address negative attitudes and biases to improve com- Greater rates of munication about clinical Interpersonal High-quality informed racial Reduced disparities, trials Level communication about and ethnic minority improved cancer Team building clinical trials participation in treatments Patient/Family clinical trials Question prompts to encourage active patient participation in clinical interaction

Fig 2. — Our proposed model for a system-level intervention.

October 2016, Vol. 23, No. 4 Cancer Control 333 • The quality of doctor–patient communication dur- in the context of clinical trials, although it has never ing clinic visits is considered the most central and been tested among minority patients.100 This question- proximal influence on a patient’s decision to par- prompt list includes 33 questions divided into 11 cate- ticipate in a clinical trial. gories (eg, finding out more about the trial, understand- • Health care is transacted through these interper- ing the trial’s purpose and background, understanding sonal processes among health care organizations, possible risks). Adapting the question-prompt list for health care professionals, and patients and their use in a diverse population of patients will require en- families.105-107 gaging community members and patients to ensure • The quality of communication during clinical in- that it reflects findings from focus groups and other re- teractions with minority patients is less than that search with individuals representative of the patients seen among similar interactions with whites.80-88 likely to be the recipients of the intervention.121 However, even multilevel interventions will have lim- The second aspect of the intervention is intended ited success if they occur in a vacuum; in other words, to enhance the patient-focused intervention by im- they must be implemented within institutions with the proving physician attitudes and communication skills appropriate infrastructure to support trials that match related to discussing trials with minority patients. The the clinical characteristics of an engaged, diverse pa- physician-focused intervention would include 2 com- tient population. ponents, one focused on communication skills and Our proposed multilevel intervention illustrates the other focused on trial-related attitudes. The com- some potential ways to influence the attitudes and communication component would build on prior work on munication skills of minorities related to engaging in training physicians to improve physician communica- discussion about clinical trials with oncologists; the at- tion skills related to discussing trials but would also titudes and communication skills of oncologists when involve increasing physician knowledge about the fact discussing trials with minority patients; and doctor–pa- that most patients, including racial and ethnic minority tient interactions in which trials may be discussed. groups, are willing to participate in trials if their physi- The first aspect of the intervention is a patient-fo- cian offers a trial and does so using high-quality, pa- cused communication intervention called a question- tient-centered communication.45,63,67,92,122 prompt list. A question-prompt list is a list of questions In clinical communication, participants exchange related to the physical and psychosocial aspects of a both informational and relational messages, and the medical condition that patients may wish to ask their training intervention would involve skill-building in physician during a medical visit. Question-prompt lists both.92 Skill-building in informational communication are designed as a simple, inexpensive way to support would include guidelines for discussing information pa- patients in gaining information about their diagnosis tients need to make an informed decision about partici- and treatment and improve doctor–patient communi- pating in a trial based on international guidelines.123,124 cation by encouraging patients to actively participate in Skill-building in relational communication would in- their health care (eg, ask questions, state concerns).108-111 clude explanations and illustrations of communication Active participation has been shown to influence the strategies such as using organizing statements, eliciting amount of information physicians provide, treatments questions and concerns by utilizing strategies such as they recommend, topics discussed, and patient psy- the ask–tell–ask method, using plain language rather chosocial and physical health outcomes.112-116 Question- than technical jargon, assessing understanding by using prompt lists have been developed and tested in several the teach-back method, directly acknowledging and em- medical settings, and, given the research showing that pathically responding to questions and concerns, and racially discordant interactions are often characterized using shared–decision-making principles.125-127 by poor-quality communication, it is surprising that The attitudinal component of the physician-fo- this type of intervention has not been tested in this con- cused intervention would occur following the com- text.98,100,117,118 We have collaborated with community munication component. This aspect of the interven- members, African American patients and their families, tion would be designed to increase the likelihood that and oncologists to develop a question-prompt list for physicians will discuss and offer trials to their minority use in the context of racially discordant oncology treat- patients. Asking people to think about their attitudes ment interactions.119 Preliminary findings from a ran- affects both their attitudes and relevant behavior, and domized trial of the question-prompt list show that its asking people to form a situation-specific plan for a use was feasible and acceptable in this patient popula- type of behavior, such as discussing a trial with a pation, and it was successful in improving their level of ac- tient, increases the likelihood they will engage in the tive participation.120 behavior.128-130 Our proposed intervention translates For the purpose of a multilevel intervention rel- that research into an intervention in which physicians evant to minority enrollment in clinical trials, we sug- are provided with a brief e-mail message prior to a visit gest adapting a question-prompt list developed for use with a patient potentially eligible for a trial. In the mes-

334 Cancer Control October 2016, Vol. 23, No. 4 sage, physicians would be asked to rate the scientific improve outcomes for all. Ann Surg Oncol. 2008;15(7):1818-1819. and clinical benefits of offering a trial to minority pa- 2. Nass SJ, Moses HL, Mendelsohn, eds; Committee on Cancer Clinical Trials; NCI Cooperative Group Program; Institute of Medicine. A National tients and to consider exactly how they will discuss the Cancer Clinical Trials System for the 21st Century: Reinvigorating the trial with the patient. NCI Cooperative Group Program. Washington, DC: National Academies Press; 2010. The interpersonal aspect, which is the third aspect 3. Murthy VH, Krumholz HM, Gross CP. Participation in cancer clinical of the intervention, involves both patients and physi- trials: race-, sex-, and age-based disparities. JAMA. 2004;291(22):2720-2726. 4. Stewart JH, Bertoni AG, Staten JL, et al. Participation in surgical cians. It is based on research that suggests the encour- oncology clinical trials: gender-, race/ethnicity-, and age-based disparities. agement of patients and physicians to create a sense of Ann Surg Oncol. 2007;14(12):3328-3334. 5. Cheng SK, Dietrich MS, Dilts DM. A sense of urgency: evaluating the common purpose, support, and understanding may link between clinical trial development time and the accrual performance of can- assist in reducing bias and increasing cooperation and cer therapy evaluation program (NCI-CTEP) sponsored studies. Clin Cancer 101,102 Res. 2010;16(22):5557-5563. trust. Based on this research, members of our re- 6. Korn EL, Freidlin B, Mooney M, et al. Accrual experience of National search group tested a team intervention in a primary Cancer Institute Cooperative Group phase III trials activated from 2000 to 2007. J Clin Oncol. 2010;28(35):5197-5201. care clinic that successfully increased patient trust and 7. Penner LA, Manning M, Eggly S, et al. Prosocial behavior in cancer adherence following racially discordant clinic visits.103 research: patient participation in cancer clinical trials. In: Graziano B, Schroeder D, eds. Handbook of Prosocial Behavior. Oxford, UK: Oxford A similar team intervention could be provided to phy- University Press; 2014. sicians who have the ability to offer trials to a diverse 8. Minasian LM, Carpenter WR, Weiner BJ, et al. Translating research into evidence-based practice: the National Cancer Institute Community population of patients as well as their patients who are Clinical Oncology Program. Cancer. 2010;116(19):4440-4449. potentially eligible for an ongoing trial. Patients and on- 9. Du W, Gadgeel SM, Simon MS. Predictors of enrollment in lung cancer clinical trials. Cancer. 2006;106(2):420-425. cologists would receive simple instructions about work- 10. Al-Refaie WB, Vickers SM, Zhong W, et al. Cancer trials versus the ing together as a team to achieve a common goal — real world in the United States. Ann Surg. 2011;254(3):438-442. 11. Freedman LS, Simon R, Foulkes MA, et al. Inclusion of women and providing high-quality care to treat cancer — and they minorities in clinical trials and the NIH Revitalization Act of 1993—the would be provided with team items, such as pens and perspective of NIH clinical trialists. Control Clin Trials. 1995;16(5):277-285. 12. Kwiatkowski K, Coe K, Bailar JC, et al. Inclusion of minorities and buttons, with a team logo. This type of intervention may women in cancer clinical trials, a decade later: have we improved? Cancer. overcome — or reduce — some of the effects of patient 2013;119(16):2956-2963. 13. Unger JM, Hershman DL, Albain KS, et al. Patient income level and mistrust of physicians and researchers, physician implic- cancer clinical trial participation. J Clin Oncol. 2013;31(5):536-542. it bias toward minorities, and concerns about harming 14. Baquet CR, Commiskey P, Daniel Mullins C, et al. Recruitment and participation in clinical trials: socio-demographic, rural/urban, and health the therapeutic relationship with these patients. care access predictors. Cancer Detect Prev. 2006;30(1):24-33. 15. Paskett ED, Cooper MR, Stark N, et al. Clinical trial enrollment of rural patients with cancer. Cancer Pract. 2002;10(1):28-35. Conclusions 16. Vanderpool RC, Kornfeld J, Mills L, et al. Rural-urban differences Although racial and ethnic disparities in cancer and in discussions of cancer treatment clinical trials. Patient Educ Couns. 2011;85(2):e69-e74. other conditions have been documented for decades, 17. Siegel R, Ward E, Brawley O, et al. Cancer statistics, 2011: the researchers, health care professionals, and policymak- impact of eliminating socio-economic and racial disparities on premature cancer deaths. CA Cancer J Clin. 2011;61(4):212-236. ers have been unable to eliminate these persistent and 18. American Cancer Society. Cancer Facts & Figures 2016. Atlanta, GA: preventable contributors to poor health.17,23 Under- American Cancer Society. http://www.cancer.org/acs/groups/content/@ research/documents/document/acspc-047079.pdf. Accessed September 13, 2016. enrollment of racial and ethnic minority populations 19. American Cancer Society. Cancer Facts & Figures for African Americans in clinical trials is a health care disparity that results 2016-2018. Atlanta, GA: American Cancer Society. http://www.cancer.org/acs/ groups/content/@editorial/documents/document/acspc-047403.pdf. Accessed from preventable and interlinked policies, practices, September 13, 2016. and barriers at the system, individual, and interper- 20. Banda DR, St. Germain D, McCaskill-Stevens W, et al. A critical re- 23 view of the enrollment of black patients in cancer clinical trials. ASCO Org. sonal levels. Although many interventions designed 2012:153-157. to increase the clinical trial enrollment of racial and 21. Carpenter WR, Tyree S, Wu Y, et al. A surveillance system for monitoring, public reporting, and improving minority access to cancer clinical trials. ethnic minorities and other under-represented groups Clin Trials. 2012;9(4):426-435. have experienced varying levels of success, multilevel 22. Unger JM, Gralow JR, Albain KS, et al. Patient income level and cancer clinical trial participation: a prospective survey study. JAMA Oncol. interventions are likely to be the most effective meth- 2016;2(1):137-139. od for increasing enrollment among a well-informed, 23. Smedley BD, Stith AY, Nelson AR, eds; Committee on Understand- 25 ing and Eliminating Racial and Ethnic Disparities in Health Care; Board on diverse population of patients. When members of Health Sciences Policy; Institute of Medicine. Unequal Treatment: Confront- these groups decide to enroll in a clinical trial based ing Racial and Ethnic Disparities in Health Care. Washington, DC: National Academies Press; 2003. on a supportive and efficient system-level environment 24. Bolen S, Tilburt J, Baffi C, et al. Defining “success” in recruitment and high-quality communication with their health care of underrepresented populations to cancer clinical trials: moving toward a more consistent approach. Cancer. 2006;106(6):1197-1204. professionals, medical researchers will be able to move 25. Taplin SH, Anhang Price R, Edwards HM, et al. Understanding and toward their goal of translating new knowledge into influencing multilevel factors across the cancer care continuum. J Natl Cancer Inst Monogr. 2012;2012(44):2-10. tangible benefits and providing high-quality cancer 26. Clauser SB, Taplin SH, Foster MK, et al. Multilevel intervention re- care for all patients. search: lessons learned and pathways forward. J Natl Cancer Inst Monogr. 2012;2012(44):127-133. 27. Kozlowski SWJ, Klein KJ. A multilevel approach to theory and re- References search in organizations: contextual, temporal, and emergent processes. In: Klein KJ, Koszlowski SWJ, eds. Multilevel Theory, Research, and Methods 1. Newman LA, Roff NK, Weinberg AD. Cancer clinical trials accrual: in Organizations: Foundations, Extensions, and New Directions. San missed opportunities to address disparities and missed opportunities to Francisco: Jossey-Bass; 2000.

October 2016, Vol. 23, No. 4 Cancer Control 335 28. Diehl KM, Green EM, Weinberg A, et al. Features associated with 54. Moskowitz D, Thom DH, Guzman D, et al. Is primary care providers’ successful recruitment of diverse patients onto cancer clinical trials: report trust in socially marginalized patients affected by race? J Gen Intern Med. from the American College of Surgeons Oncology Group. Ann Surg Oncol. 2011;26(8):846-851. 2011;18(13):3544-3550. 55. Moskowitz GB, Stone J, Childs A. Implicit stereotyping and medi- 29. Ford JG, Howerton MW, Lai GY, et al. Barriers to recruiting under- cal decisions: unconscious stereotype activation in practitioners’ thoughts represented populations to cancer clinical trials: a systematic review. Cancer. about African Americans. Am J Public Health. 2012;102(5):996-1001. 2008;112(2):228-242. 56. Sabin JA, Greenwald AG. The influence of implicit bias on treat- 30. Langford AT, Resnicow K, Dimond EP, et al. Racial/ethnic differ- ment recommendations for 4 common pediatric conditions: pain, urinary ences in clinical trial enrollment, refusal rates, ineligibility, and reasons for tract infection, attention deficit hyperactivity disorder, and asthma. Am J decline among patients at sites in the National Cancer Institute’s Community Public Health. 2012;102(5):988-995. Cancer Centers Program. Cancer. 2014;120(6):877-884. 57. Sabin JA, Rivara FP, Greenwald AG. Physician implicit attitudes 31. Bruner DW, Jones M, Buchanan D, et al. Reducing cancer dispari- and stereotypes about race and quality of medical care. Med Care. ties for minorities: a multidisciplinary research agenda to improve patient 2008;46(7):678-685. access to health systems, clinical trials, and effective cancer therapy. J Clin 58. Joseph G, Dohan D. Diversity of participants in clinical trials in an Oncol. 2006;24(14):2209-2215. academic medical center: the role of the ‘Good Study Patient?’. Cancer. 32. McCaskill-Stevens W, McKinney MM, Whitman CG, et al. Increasing 2009;115(3):608-615. minority participation in cancer clinical trials: the minority-based community 59. Simon MS, Du W, Flaherty L, et al. Factors associated with breast clinical oncology program experience. J Clin Oncol. 2005;23(22):5247-5254. cancer clinical trials participation and enrollment at a large academic medical 33. Somkin CP, Altschuler A, Ackerson L, et al. Organizational barri- center. J Clin Oncol. 2004;22(11):2046-2052. ers to physician participation in cancer clinical trials. Am J Manag Care. 60. Penner LA, Blair IV, Albrecht TL, et al. Reducing racial health care 2005;11(7):413-421. disparities: a social psychological analysis. Policy Insights Behav Brain 34. Wong YN, Schluchter MD, Albrecht TL, et al. Financial concerns Sci. 2014;1:204-212. about participation in clinical trials among patients with cancer. J Clin Oncol. 61. Epstein RM, Street RL Jr. Patient-Centered Communication in Cancer 2016;34(5):479 - 487. Care: Promoting Healing and Reducing Suffering. NIH publication no. 07-6225. 35. Sateren WB, Trimble EL, Abrams J, et al. How socio-demographics, Bethesda, MD: National Cancer Institute; 2007. presence of oncology specialists, and hospital cancer programs affect accrual 62. Kissane DW, Bylund CL, Banerjee SC, et al. Communication skills to cancer treatment trials. J Clin Oncol. 2002;20(8):2109-2117. training for oncology professionals. J Clin Oncol. 2012;30(11):1242-1247. 36. Farmer DF, Jackson SA, Camacho F, et al. Attitudes of African 63. Brown RF, Butow PN, Boyle F, et al. Seeking informed consent to American and low socio-economic status white women toward medical cancer clinical trials: evaluating the efficacy of doctor communication skills research. Health Care Poor Underserved. 2007;18(1):85-99. training. Psychooncology. 2007;16(6):507-516. 37. Paskett ED, Reeves KW, McLaughlin JM, et al. Recruitment of mi- 64. Gross CP, Herrin J, Wong N, et al. Enrolling older persons in cancer nority and underserved populations in the United States: the Centers for trials: the effect of socio-demographic, protocol, and recruitment center Population Health and Health Disparities experience. Contemp Clin Trials. characteristics. J Clin Oncol. 2005;23(21):4755-4763. 2008;29(6):847-861. 65. Manne S, Kashy D, Albrecht T, et al. Attitudinal barriers to participa- 38. Skloot R. The Immortal Life of Henrietta Lacks. New York: Broadway tion in oncology clinical trials: factor analysis and correlates of barriers. Eur J Books; 2011. Cancer Care (Engl). 2015;24(1):28-38. 39. Katz RV, Kegeles SS, Kressin NR, et al. Awareness of the Tuske- 66. Rivers D, August EM, Sehovic I, et al. A systematic review of the gee syphilis study and the US presidential apology and their influence factors influencing African Americans’ participation in cancer clinical trials. on minority participation in biomedical research. Am J Public Health. Contemp Clin Trials. 2013;35(2):13-32. 2008;98(6):1137-1142. 67. Wendler D, Kington R, Madans J, et al. Are racial and ethnic minori- 40. Shavers VL, Lynch CF, Burmeister LF. Racial differences in factors ties less willing to participate in health research? PLoS Med. 2006;3(2):e19. that influence the willingness to participate in medical research studies. 68. Svensson K, Ramirez OF, Peres F, et al. Socio-economic determi- Ann Epidemiol. 2002;12(4):248-256. nants associated with willingness to participate in medical research among 41. Weiner BJ, Jacobs SR, Minasian LM, et al. Organizational designs a diverse population. Contemp Clin Trials. 2012;33(6):1197-1205. for achieving high treatment trial enrollment: a fuzzy set analysis of the 69. Katz RV, Kegeles SS, Kressin NR, et al. The Tuskegee Legacy Project: Community Clinical Oncology Program. J Oncol Pract. 2012;8(5):287-291. willingness of minorities to participate in biomedical research. J Health 42. Wujcik D, Wolff SN. Recruitment of African Americans to national Care Poor Underserved. 2006;17(4):698-715. oncology clinical trials through a clinical trial shared resource. J Health 70. Meropol NJ, Wong YN, Albrecht T, et al. Randomized trial of a Care Poor Underserved. 2010;21(1 suppl):38-50. web-based intervention to address barriers to clinical trials. J Clin Oncol. 43. McCaskill-Stevens W, Lyss AP, Good M, et al. The NCI Community 2016;34(5):469-478. Oncology Research Program: what every clinician needs to know. Am Soc 71. Jacobsen PB, Wells KJ, Meade CD, et al. Effects of a brief multime- Clin Oncol Educ Book. 2013. dia psychoeducational intervention on the attitudes and interest of patients 44. Brooks SE, Muller CY, Robinson W, et al. Increasing minority enroll- with cancer regarding clinical trial participation: a multicenter randomized ment onto clinical trials: practical strategies and challenges emerge from controlled trial. J Clin Oncol. 2012;30(20):2516-2521. the NRG Oncology Accrual Workshop. J Oncol Pract. 2015;11(6):486-490. 72. Corbie-Smith G, Thomas SB, Williams MV, et al. Attitudes and be- 45. Brown R, Albrecht TL. Enrollment in clinical trials. In: Kissane D, liefs of African Americans toward participation in medical research. J Gen Bultz BD, Butow PN, et al, eds. Handbook of Communication in Oncology Intern Med. 1999;14(9):537-546. and Palliative Care. Oxford, UK: Oxford University Press; 2010. 73. Scharff DP, Mathews KJ, Jackson P, et al. More than Tuskegee: 46. Howerton MW, Gibbons MC, Baffi CR, et al. Provider roles in the re- understanding mistrust about research participation. J Health Care Poor cruitment of underrepresented populations to cancer clinical trials. Cancer. Underserved. 2010;21(3):879-897. 2007;109(3):465-476. 74. Dovidio JF, Penner LA, Albrecht TL, et al. Disparities and distrust: 47. Henderson GE, Churchill LR, Davis AM, et al. Clinical trials the implications of psychological processes for understanding racial dis- and medical care: defining the therapeutic misconception. PLoS Med. parities in health and health care. Soc Sci Med. 2008;67(3):478-486. 2007;4(11):e324. 75. Penner LA, Dovidio JF, Edmondson D, et al. The experience of 48. Pinto HA, McCaskill-Stevens W, Wolfe P, et al. Physician perspec- discrimination and black-white health disparities in medical care. J Black tives on increasing minorities in cancer clinical trials: an Eastern Coop- Psychol. 2009;35(2):180-203. erative Oncology Group (ECOG) initiative. Ann Epidemiol. 2000;10(8 76. Halbert CH, Armstrong K, Gandy OH Jr, et al. Racial differences in suppl):S78-S84. trust in health care providers. Arch Intern Med. 2006;166(8):896-901. 49. Hagiwara N, Penner LA, Gonzalez R, et al. Racial attitudes, phy- 77. Halbert CH, Weathers B, Delmoor E, et al. Racial differences in sician-patient talk time ratio, and adherence in racially discordant medical medical mistrust among men diagnosed with prostate cancer. Cancer. interactions. Soc Sci Med. 2013;87:123-131. 2009;115(11):2553-2561. 50. Cooper LA, Roter DL, Carson KA, et al. The associations of clini- 78. Boulware L, Cooper L, Ratner L, et al. Race and trust in the health cians’ implicit attitudes about race with medical visit communication and patient care system. Public Health Reports. 2003;118:358-365. ratings of interpersonal care. Am J Public Health. 2012;102(5):979-987. 79. Duda C, Mahon I, Chen MH, et al. Impact and costs of targeted 51. Blair IV, Steiner JF, Fairclough DL, et al. Clinicians’ implicit ethnic/ recruitment of minorities to the National Lung Screening Trial. Clin Trials. racial bias and perceptions of care among black and Latino patients. Ann 2011;8(2):214-223. Fam Med. 2013;11(1):43-52. 80. Jean-Pierre P, Fiscella K, Griggs J, et al. Race/ethnicity-based con- 52. Penner LA, Dovidio JF, West TV, et al. Aversive racism and medi- cerns over understanding cancer diagnosis and treatment plan. J Natl Med cal interactions with Black patients: a field study. J Exp Soc Psychol. Assoc. 2010;102(3):184-189. 2010;46(2):436-440. 81. Street RL Jr, Gordon H, Haidet P. Physicians’ communication and 53. van Ryn M, Burgess D, Malat J, et al. Physicians’ perceptions of perceptions of patients: is it how they look, how they talk, or is it just the patients’ social and behavioral characteristics and race disparities in treat- doctor? Soc Sci Med. 2007;65(3):586-598. ment recommendations for men with coronary artery disease. Am J Public 82. Eggly S, Harper FW, Penner LA, et al. Variation in question asking Health. 2006;96(2):351-357. during cancer clinical interactions: a potential source of disparities in access

336 Cancer Control October 2016, Vol. 23, No. 4 to information. Patient Educ Couns. 2011;82(1):63-68. 111. Henselmans I, de Haes HC, Smets EM. Enhancing patient partici- 83. Cooper LA, Roter DL, Johnson RL, et al. Patient-centered commu- pation in oncology consultations: a best evidence synthesis of patient-tar- nication, ratings of care, and concordance of patient and physician race. geted interventions. Psychooncology. 2013;22(5):961-977. Ann Intern Med. 2003;139(11):907-915. 112. Dimoska A, Butow PN, Dent E, et al. An examination of the initial 84. Johnson RL, Roter D, Powe NR, et al. Patient race/ethnicity and cancer consultation of medical and radiation oncologists using the Cancode quality of patient-physician communication during medical visits. Am J interaction analysis system. Br J Cancer. 2008;98(9):1508-1514. Public Health. 2004;94(12):2084-2090. 113. Cegala DJ, Street RL Jr, Clinch CR. The impact of patient participation 85. Gordon HS, Street RL Jr, Sharf BF, et al. Racial differences in trust on physicians’ information provision during a primary care medical interview. and lung cancer patients’ perceptions of physician communication. J Clin Health Commun. 2007;21(2):177-185. Oncol. 2006;24(6):904-909. 114. Street RL, Jr, Tancredi DJ, Slee C, et al. A pathway linking patient 86. Gordon HS, Street RL Jr, Sharf BF, et al. Racial differences participation in cancer consultations to pain control. Psychooncology. in doctors’ information-giving and patients’ participation. Cancer. 2014;23(10):1111-1117. 2006;107(6):1313-1320. 115. Clayton JM, Butow PN, Tattersall MH, et al. Randomized controlled 87. Siminoff LA, Graham GC, Gordon NH. Cancer communication pat- trial of a prompt list to help advanced cancer patients and their caregiv- terns and the influence of patient characteristics: disparities in information- ers to ask questions about prognosis and end-of-life care. J Clin Oncol. giving and affective behaviors. Patient Educ Couns. 2006;62(3):355-360. 2007;25(6):715-723. 88. Song L, Hamilton JB, Moore AD. Patient-healthcare provider com- 116. Greene J, Hibbard JH. Why does patient activation matter? An munication: perspectives of African American cancer patients. Health Psychol. examination of the relationships between patient activation and health- 2012;31(5):539-547. related outcomes. J Gen Intern Med. 2012;27(5):520-526. 89. Sabin J, Nosek BA, Greenwald A, et al. Physicians’ implicit and 117. Clayton J, Butow P, Tattersall M, et al. Asking questions can help: explicit attitudes about race by MD race, ethnicity, and gender. J Health development and preliminary evaluation of a question prompt list for pallia- Care Poor Underserved. 2009;20(3):896-913. tive care patients. Br J Cancer. 2003;89(11):2069-2077. 90. Penner LA, Dovidio JF, Gonzalez R, et al. The effects of oncolo- 118. Cegala DJ, Chisolm DJ, Nwomeh BC. A communication skills in- gist implicit racial bias in racially discordant oncology interactions. J Clin tervention for parents of pediatric surgery patients. Patient Educ Couns. Oncol. 2016;34(24):2874-2880. 2013;93(1):34-39. 91. Janz NK, Wren PA, Copeland LA, et al. Patient-physician concor- 119. Eggly S, Tkatch R, Penner LA, et al. Development of a question dance: preferences, perceptions, and factors influencing the breast cancer prompt list as a communication intervention to reduce racial disparities in surgical decision. J Clin Oncol. 2004;22(15):3091-3098. cancer treatment. J Cancer Educ. 2013;28(2):282-289. 92. Albrecht TL, Eggly SS, Gleason ME, et al. Influence of clinical com- 120. Eggly S, Hamel LM, Foster T, et al. Using a question prompt list to munication on patients’ decision making on participation in clinical trials. J Clin increase patient active participation in racially discordant cancer interac- Oncol. 2008;26(16):2666-2673. tions. Presented at: 13th International Conference on Communication in 93. Eggly S, Albrecht TL, Harper FW, et al. Oncologists’ recommen- Healthcare; October 25–28, 2015; New Orleans, LA. dations of clinical trial participation to patients. Patient Educ Couns. 121. Brown RF, Cadet DL, Houlihan RH, et al. Perceptions of participa- 2008;70(1):143-148. tion in a phase I, II, or III clinical trial among African American patients with 94. Barton E, Eggly S. Ethical or unethical persuasion? The rhetoric of cancer: what do refusers say? J Oncol Pract. 2013;9(6):287-293. offers to participate in clinical trials. Writ Commun. 2009;26(3):295-319. 122. Bylund CL, Brown R, Gueguen JA, et al. The implementation and 95. Kass N, Taylor H, Fogarty L, et al. Purpose and benefits of early assessment of a comprehensive communication skills training curriculum phase cancer trials: what do oncologists say? What do patients hear? J Empir for oncologists. Psychooncology. 2010;19(6):583-593. Res Hum Res Ethics. 2008;3(3):57-68. 123. Wendler D, Grady C. What should research participants un- 96. Eggly S, Barton E, Winckles A, et al. A disparity of words: racial dif- derstand to understand they are participants in research? Bioethics. ferences in oncologist-patient communication about clinical trials. Health 2008;22(4):203-208. Expect. 2015;18(5):1316-1326. 124. Council for International Organizations of Medical Sciences 97. Hamel LM, Chapman R, Malloy M, et al. Critical shortage of (CIOMS); World Health Organization. International Ethical Guidelines for African American medical oncologists in the United States. J Clin Oncol. Biomedical Research Involving Human Subjects. Geneva: CIOMS; 2002. 2015;33(32):3697-3700. 125. Back AL, Arnold RM, Baile WF, et al. Approaching difficult commu- 98. Street RL Jr, Slee C, Kalauokalani DK, et al. Improving physician- nication tasks in oncology. CA Cancer J Clin. 2005;55(3):164-177. patient communication about cancer pain with a tailored education-coach- 126. Kemp EC, Floyd MR, McCord-Duncan E, et al. Patients prefer the ing intervention. Patient Educ Couns. 2010;80(1):42-47. method of “tell back-collaborative inquiry” to assess understanding of 99. Kass NE, Sugarman J, Medley AM, et al. An intervention to im- medical information. J Am Board Fam Med. 2008;21(1):24-30. prove cancer patients’ understanding of early-phase clinical trials. IRB. 127. Makoul G, Clayman ML. An integrative model of shared decision 2009;31(3):1-10. making in medical encounters. Patient Educ Couns. 2006;60(3):301-312. 100. Brown RF, Bylund CL, Li Y, et al. Testing the utility of a cancer clinical 128. Young AI, Fazio RH. Attitude accessibility as a determinant of ob- trial specific question prompt list (QPL-CT) during oncology consultations. ject construal and evaluation. J Exp Soc Psychol. 2013;49(3):404-418. Patient Educ Couns. 2012;88(2):311-317. 129. Gollwitzer P. The volitional benefits of planning. In: Gollwitzer P, 101. Gaertner S, Dovidio JF. A common ingroup identity: a categori- Bargh J, eds. The Psychology of Action: Linking Cognition and Motivation zation-based approach for reducing intergroup bias. In: Nelson TD, ed. to Behaviour. New York: Guilford Press; 1996. Handbook of Prejudice, Stereotyping, and Discrimination. Thousand 130. Gollwitzer PM. Implementation intentions: strong effects of simple Oaks, CA: Sage; 2009. plans. Am Psychologist. 1999;54(7):493-503. 102. Gaertner S, Dovidio JF, Houlette M. Social categorization. In: Dovidio JF, Hewstone M, Glick P, eds. Handbook of Prejudice, Stereotyping, and Discrimination. London: Sage; 2010. 103. Penner LA, Gaertner S, Dovidio JF, et al. A social psychological approach to improving the outcomes of racially discordant medical interactions. J Gen Intern Med. 2013;28(9):1143-1149. 104. Stange KC, Breslau ES, Dietrich AJ, et al. State-of-the-art and future directions in multilevel interventions across the cancer control continuum. J Natl Cancer Inst Monogr. 2012;2012(44):20-31. 105. Albrecht TL, Penner LA, Cline RJ, et al. Studying the process of clinical communication: issues of context, concepts, and research directions. J Health Commun. 2009;14(suppl 1):47-56. 106. McCormack LA, Treiman K, Rupert D, et al. Measuring patient-centered communication in cancer care: a literature review and the development of a systematic approach. Soc Sci Med. 2011;72(7):1085-1095. 107. Epstein RM, Street RL Jr. Patient-Centered Communication in Cancer Care: Promoting Healing and Reducing Suffering. NIH publication no. 07-6225. Bethesda, MD: National Cancer Institute; US Department of Health and Human Services, National Institutes of Health; 2007. 108. Sansoni JE, Grootemaat P, Duncan C. Question Prompt Lists in health consultations: a review. Patient Educ Couns. 2015. Epub ahead of print. 109. Brandes K, Linn AJ, Butow PN, et al. The characteristics and effectiveness of question prompt list interventions in oncology: a systematic review of the literature. Psychooncology. 2015;24(3):245-252. 110. Dimoska A, Tattersall MH, Butow PN, et al. Can a “prompt list” empower cancer patients to ask relevant questions? Cancer. 2008;113(2):225-237.

October 2016, Vol. 23, No. 4 Cancer Control 337 Lung carcinogenesis is a multifactorial

process that can affect racial and

ethnic groups differently.

Fiddle Head. X-Ray image courtesy of Bryan Whitney. www.x-rayphotography.com.

Racial and Ethnic Differences in the Epidemiology and Genomics of Lung Cancer Matthew B. Schabath, PhD, W. Douglas Cress, PhD, and Teresita Muñoz-Antonia, PhD

Background: Lung cancer is the most common cancer in the world. In addition to the geographical and sex-specific differences in the incidence, mortality, and survival rates of lung cancer, growing evidence suggests that racial and ethnic differences exist. Methods: We reviewed published data related to racial and ethnic differences in lung cancer. Results: Current knowledge and substantive findings related to racial and ethnic differences in lung cancer were summarized, focusing on incidence, mortality, survival, cigarette smoking, prevention and early detection, and genomics. Systems-level and health care professional–related issues likely to contribute to specific racial and ethnic health disparities were also reviewed to provide possible suggestions for future strategies to reduce the disproportionate burden of lung cancer. Conclusions: Although lung carcinogenesis is a multifactorial process driven by exogenous exposures, genetic variations, and an accumulation of somatic genetic events, it appears to have racial and ethnic differences that in turn impact the observed epidemiological differences in rates of incidence, mortality, and survival.

Introduction sis. Approximately 1.2 million cases among men alone Lung cancer is the most common cancer worldwide.1,2 were diagnosed in 2012.2 In the United States, lung In 2012, new diagnoses of lung cancer reached ap- cancer is the second most common cancer in men after proximately 1.8 million worldwide, accounting for 13% prostate cancer and the second most common cancer of the global cancer burden.2 Among men, lung in women after breast cancer.3 In 2016, an estimated cancer remains the most common cancer diagno- 224,390 new cases of lung cancer will be diagnosed.3 Lung cancer is also the leading cause of cancer- related death in the United States and accounts for more From the Departments of Cancer Epidemiology (MBS), Molecular Oncology (WDC), and Tumor Biology Program (TM-A), H. Lee Moffitt deaths than prostate, breast, colorectal, and pancreatic Cancer Center & Research Institute, Tampa, Florida. cancers combined.3 In 2016, an estimated 158,080 deaths Submitted February 19, 2016; accepted May 11, 2016. related to lung cancer will occur.3 From 2008 to 2012, Address correspondence to Matthew B. Schabath, PhD, Department the mortality rate among men was 59.8 per 100,000 of Cancer Epidemiology, Moffitt Cancer Center, 12902 Magnolia Drive, and 37.8 per 100,000 persons for women.3 MRC-CANCONT, Tampa, FL 33612. E-mail: [email protected] No significant relationships exist between the authors and the companies/ Although lung carcinogenesis is a multifactorial organizations whose products or services may be referenced in this article. process driven by exogenous exposures (eg, cigarette This work is supported in part by the James & Esther King Biomedical smoking), inherited genetic variations, and an accu- Research Program (5JK06), National Institutes of Health Grant from mulation of somatic genetic events, it also appears to the National Cancer Institute (U54-CA 163068), and by a cancer center support grant at the H. Lee Moffitt Cancer Center & Research Institute have racial and ethnic differences. However, many of (P30-CA76292). these observed differences attributed to underlying ra-

338 Cancer Control October 2016, Vol. 23, No. 4 cial and ethnic disparities in lung cancer are not fully socioeconomic status. Racial and ethnic differences in understood. Health disparities can also occur as a re- disease burden can reveal specific issues of a particu- sult of systems-level issues, such as access to care, in- lar population or subpopulation. Racial and ethnic dif- surance, and hospital-level factors.4-6 Other potential ferences in rates of incidence, mortality, and survival sources of racial and ethnic disparities are derived of lung cancer are well documented.10,11 The most com- from health care professional–related factors, such as prehensive report in the United States is based on data limited cultural sensitivity, stereotyping, and poor doc- from the Surveillance, Epidemiology, and End Results tor–patient relationships.7 Studies have shown that, (SEER) program, which revealed that blacks have high- even after controlling for these factors, treatment dis- er rates of incidence and mortality than any other ra- parities persist, suggesting that patient-specific factors cial or ethnic group.10 may also contribute to lung cancer differences.8,9 A report from the Centers for Disease Control and In addition to the established geographical and Prevention assessed potential racial/ethnic disparities sex-specific differences in the incidence, mortality, and and geographical differences in incidence rates of lung survival rates of lung cancer, certain racial and ethnic cancer from 1998 to 2006 using data from SEER and differences exist.10,11 Although the terms race and the National Program of Cancer Registries.11 In this ethnicity are often inconsistently applied, a common- report, the annual incidence rate of lung cancer was ly used definition describes race as a social construct highest among blacks, followed by whites, American that incorporates beliefs about language, history, and Indians/Alaska Natives, and Asian/Pacific Islanders.11 culture, forming the basis on which social identity, tra- Hispanics had lower rates of lung cancer incidence ditions, and politics are built.12,13 In the United States, than non-Hispanic whites.11 In the United States, the the race classification scheme used in the 2000 census, highest rate of lung cancer incidence was found in which is often used in biomedical research, includes the South and the lowest rate of incidence was in the 5 major groups: white, black/African American, Asian, West.11 Among whites, the highest rate of lung cancer Native Hawaiian/Pacific Islander, and American Indi- incidence was in the South; the highest incidence rates an/Alaska Native.14 Broadly, this classification scheme among blacks, American Indians/Alaska Natives, and emphasizes the geographical region of origin of an in- Hispanics were in the Midwest; and the highest inci- dividual’s ancestry. Ethnicity is a broader construct that dence among Asian/Pacific Islanders was in the West.11 takes into consideration cultural tradition, common The identification of geographical differences in inci- history, religion, and, oftentimes, a shared genetic heri- dence rates among racial and ethnic populations pres- tage.13 Thus, to avoid confusion, we will use the terms ents novel opportunities for targeted efforts in primary (race or ethnicity) as reported in the originally cited prevention and early detection. work to describe the population studied. Survival Epidemiology Despite considerable improvements in patient survival Incidence during the past several decades for breast and prostate The incidence rates of lung cancer are generally lower cancers, few improvements have been seen in survival among women (583,000 new lung cancer diagnoses rates of lung cancer. Improvements in survival are lack- in 2012).2 The highest incidence rates occur in eastern ing because the majority of lung cancers are at an ad- and Central Europe as well as eastern Asia; by contrast, vanced stage by the time a diagnosis is made and treat- the lowest rates occur in western and Central Africa.2 ment options are limited. Because of its high mortality The geographical variations for incidence and mor- rate, variability in survival is lacking in different world tality differ between men and women, largely attrib- regions.2 In the United States, the 5-year relative sur- uted to geographical differences in cigarette smoking vival rate for all lung cancers is 17%.3 between the 2 sexes.2,15 The highest incidence rates in Among all stages of non–small-cell lung cancer women are observed in North America and northern (NSCLC) and small-cell lung cancer, the combined Europe, whereas the lowest rates are found in western 5-year relative survival rates are 21% and 6%, respec- and Central Africa.2 tively.3,16 The overall prognosis for NSCLC remains In the United States, the incidence rates have been poor, and prognostic factors associated with poor sur- declining in men in the past 20 years and in women vival include late-stage diagnosis, current smoking sta- since the mid-2000s: From 2008 to 2012, they annually tus, advanced age, male sex, poor pulmonary function, decreased by 3.0% among men and 1.9% among wom- presence of cardiovascular disease, nonsquamous cell en.3 During this time period, the incidence rate among histology, and pneumonectomy.17-21 Among individuals men was 76.7 per 100,000 and 54.1 per 100,000 per- with small-cell lung cancer, poor prognosis is associ- sons for women.3 ated with age older than 70 years, male sex, relapsed Race and ethnicity are complex social and cultur- disease, extensive disease, weight loss of more than al constructs, but they are also often associated with 10% of total body weight at diagnosis, and poor per-

October 2016, Vol. 23, No. 4 Cancer Control 339 formance status.22-26 Despite the poor outcomes asso- (18.8%) than females (14.8%), and nearly 17% of adults ciated with a diagnosis of lung cancer, the emergence still continue to smoke cigarettes.41,43 of immunotherapy and immune checkpoint inhibitors By racial and ethnic groups, smoking prevalence has demonstrated durable rates of long-term survival rates were highest among American Indians/Alaska in select patients.27,28 As such, these therapies may lead Natives (29.2%) and multiracial adults (27.9%) and low- to improved survival rates and, in the early stage, pos- est among Asians (9.5%).41,43 For other racial and eth- sibly curable disease.29 nic groups, smoking prevalence rates were 18.2% for Published data have demonstrated that the surviv- whites, 17.5% for blacks, and 11.2% for Hispanics.41,43 al rates of those with lung cancer also differ based on Although the annual incidence is highest among race and ethnicity.4,30-33 Blacks are less likely to receive blacks compared with any other race, the prevalence surgical resection than whites, a finding that likely con- of smoking among blacks is lower than American Indi- tributes to data that show blacks have lower rates of ans/Alaska Natives, multiracial adults, and whites.41,43 survival for early-stage NSCLC.4,30-32 Using SEER data Moreover, significant differences have been reported linked to Medicare claims data during 1995 and 1999, in the association between cigarette smoking and the Shugarman et al31 reported that blacks were 66% less risk of lung cancer among 5 ethnic/racial groups.44 likely to receive timely and appropriate treatment than Among those who smoked no more than 30 cigarettes whites, and black men were least likely to receive re- per day, African Americans and Native Hawaiians had section (22.0% for black men vs 43.7% for white men). significantly greater risks of lung cancer than the other The authors also reported that blacks were 34.0% less groups studied.44 As such, these data provide further likely to receive timely surgery, chemotherapy, or ra- support for ethnic/racial differences, which could be diotherapy for stage III disease and were 51.0% less attributed to innate genetic differences in the smoking- likely to receive chemotherapy in a timely fashion for associated risk of lung cancer. stage IV disease relative to whites.31 Howington et al33 Other established and putative risk factors for reported that American Indians and Alaskan Natives lung cancer include exposure to secondhand smoke have worse survival rates than non-Hispanic whites. (ie, passive smoke), history of chronic obstructive pul- With regard to potential ethnic differences, a study by monary disease, family history of lung cancer, radon Lin et al34 compared rates of stage-appropriate treat- exposure, and occupational exposure (eg, asbestos, ment among blacks, Hispanics, and nonminority pa- arsenic, diesel exhaust, chromium).19,45,46 Evidence is tients and did not observe treatment-related disparities conflicting regarding the impact of exogenous hor- among Hispanics. However, the authors did find that mones on women, diet, and body mass index in as- blacks were less likely than nonminorities to receive sociation with risk of lung cancer.46-53 Moreover, few stage-appropriate treatment, similar to data reported in studies have assessed the potential racial and ethnic previous studies.5,34-36 differences for such risk factors.50

Risk Factors Prevention and Early Detection Tobacco smoking is the most important and preva- Until recently, no screening method was shown to lent risk for lung cancer.37-39 Lung cancer is one of the decrease the mortality rates for NSCLC. The National first chronic diseases to be causally linked to tobacco Lung Screening Trial (NLST) randomized 53,452 indi- smoking, and approximately 90% of lung cancer dividuals who were former or current cigarette smok- agnoses in the United States are attributed to tobacco ers, between the ages of 55 and 74 years, and had a smoking.19,20 Cigarette smoke contains more than 7,000 pack-year smoking history of at least 30 years into ei- chemicals, including more than 60 established carcino- ther low-dose helical computed tomography (LDCT) gens and other toxicants associated with chronic dis- or standard chest radiography — both arms were ad- eases.40 Although approximately 1 in 9 individuals who ministered 3 annual screenings (baseline screening smoke develops lung cancer, the relative risk of lung and 2 follow-up screening sessions approximately cancer in those who are long-term smokers is estimat- 12 months apart).54 After a median follow-up period of ed to be between 10- and 30-fold higher than that of a 6.4 years, a 20% relative reduction in rate of lung can- lifetime never-smoker.19 cer mortality was observed for the LDCT group com- The percentage of cigarette smoking among pared with those assigned to standard chest radiogra- adults in the United States and most Western nations phy.54 The rate of lung cancer detected via screening has been steadily declining during the past several de- (ie, diagnosed on follow-up screening intervals) ac- cades — from 20.9% in 2005 to 16.8% in 2014 — and counted for 58% of all LDCT-detected cases of lung it continues to decline today due to successful tobac- cancer, were 2.7-fold higher in the LDCT arm vs the co-control efforts.2,41 In the United States, smoking chest radiography arm, were associated with a stage rates have steadily declined since the 1960s.42 In 2014, shift from late-stage to more early-stage lung cancer, the prevalence of smoking was higher among males and demonstrated improved rates of 5-year survival

340 Cancer Control October 2016, Vol. 23, No. 4 compared with cancers diagnosed at the baseline prev- 7-day cessation attempts, 6-month continuous absti- alence screening.54 nence from smoking, and the use of smoking cessation An important benefit of early detection is the abil- programs and cessation aids 12 months after screen- ity to detect more cases of early-stage lung cancer. A ing. The authors reported that blacks were more likely meta-analysis revealed that the rate of detection of than whites to have 24-hour and 7-day cessation at- stage I lung cancer was 70% with LDCT screening com- tempts; however, these attempts did not translate to pared with 16% when detected during routine care.55 increased rates of 6-month continuous abstinence Thus, the incidence of lung cancer detected in screen- among blacks.58 Specifically, 1 year after screening, ing and then diagnosed as a result of LDCT have bet- blacks were more likely to report a 24-hour (52.7% for ter surgical outcomes and improved 5-year survival blacks vs 41.2% for whites; P < .01) or 7-day (33.6% for rates than prevalence and routine care–detected lung blacks vs 27.2% for whites; P < .01) cessation attempt.58 cancer diagnosed when patients develop symptoms of Although no statistically significant racial differences late-stage disease.19 Based on the findings from the were found in 6-month continuous abstinence NLST, the US Preventive Services Task Force issued a (5.6% for blacks vs 7.2% for whites), black race was a recommendation in December 2013 for annual screen- statistically significant predictor in multivariable analy- ing for lung cancer using LDCT in adults aged 55 to ses of a higher likelihood of a 24-hour and 7-day cessa- 80 years who have a 30 pack-year smoking history and tion attempt.58 Race was not associated with 6-month, currently smoke or have quit within the past 15 years.54,56 continuous abstinence; by contrast, a positive result Of the 53,452 participants in the NLST, 90.8% on screening for lung cancer was the only significant- were white (n = 48,549), 4.4% were black (n = 2,376), ly predictor of successful 6-month, continuous absti- 1.7% were Asian (n = 895), 2.0% were of other ra- nence.58 At present, knowledge is lacking about the ef- cial groups or reported more than 1 race or ethnic fectiveness of screening for lung cancer with LDCT in group (n = 1,062), and 1.7% were Hispanic or Latino other racial and ethnic groups. (n = 935).54 Although the NLST did not provide eth- The data from the NLST demonstrated that screen- nic-/racial-specific results,54 a posthoc analysis by ing with LDCT can mitigate lung cancer mortality via Tanner et al57 assessed racial differences in outcomes early detection.55 Screening with LDCT appears to be between blacks and whites who participated in the second only to primary prevention (ie, smoking pre- NLST. Although demographics associated with im- vention/cessation) for mitigating lung cancer–related proved survival in lung cancer, such as higher educa- mortality — and the single remaining option for those tion, former smoking status, and fewer comorbidi- who have already quit smoking. The risk of lung can- ties, were less commonly found among blacks in the cer caused by smoking is reduced following smoking NLST, their analysis revealed that screening with LDCT cessation, but this risk always remains elevated com- reduced mortality in all racial groups, but more so in pared with those who are lifetime never-smokers.59 blacks (hazard ratio [HR] 0.61 for blacks vs HR 0.86 for whites).57 Among all racial groups, Tanner et al57 also Genomics noted that study participants who currently smoked The most frequently mutated genes in lung can- had worse rates of lung cancer–specific mortality; cer are TP53 (53.6%), KRAS (16.1%), STK11 (9.8%), however, the risk was 2-fold higher in blacks who EGFR (7.2%), KEAP1 (6.6%), and NFE2L2 (4.5%).60 The were current smokers compared with white current frequency of mutations in oncogenes and tumor-sup- smokers. In addition, the rate of all-cause mortality pressor genes differ across racial and ethnic popula- was 1.35 times higher in blacks vs whites, but blacks tions and are selectively shown in the Table. Advances screened with LDCT had a statistically significant re- in tumor genomic profiling have resulted in a para- duction in all-cause mortality when compared with digm shift, whereby types of lung cancer are charac- whites.57 As such, screening for lung cancer appears terized and classified by genetic alternations in onco- to be beneficial in blacks, but different strategies are genes and tumor-suppressor genes critical to tumor needed to achieve a significant reduction in lung can- growth and survival and can be exploited with specific cer–related mortality in this population.57 Ongoing targeted agents.61 disparities in access to screening and follow-up visits The focus of this paper is on specific oncogenes (in the case of abnormal results) underscore the impor- and tumor-suppressor genes where published studies tance of health care institutions to offer screening and have assessed potential racial and ethnic differences. Al- then quickly resolve any abnormal findings. though other genes have also been studied, including In a separate posthoc analysis of the NLST, Kumar MET, PI3KCA, PTEN, and ROS1, current data are limited et al58 examined racial differences in smoking behav- and suggest that no differences exist in these genes.62 iors among white and black study participants in the Because most types of lung cancer harbor somat- NLST who were current smokers at screening. They ic mutations, alterations, or both, the use of targeted analyzed data from a follow-up survey on 24-hour and therapy is important to improve outcomes.63 Research

October 2016, Vol. 23, No. 4 Cancer Control 341 fective therapy for patients whose tumors harbor ac- Table. — Estimated Frequencies of Select Mutated Genes tivating mutations in the tyrosine kinase domain in Lung Cancer of EGFR.66 Therefore, mutation testing is routinely Gene Frequency, % performed to identify patients harboring targetable EGFR EGFR mutations, given that selection based on clinical By region and pathological characteristics alone is inadequate. Asian populations 30.0–40.0 Overall, approximately 10% of patients with NSCLC in Western countries and 35% of patients in East Asia Western populations 10.0–20.0 have a tumor that exhibits an EGFR mutation.67,68 The By country most frequent EGFR mutations occur in exons 18 to 21, Argentina 19.3 encoding a portion of the EGFR kinase domain; in ad- Colombia 24.8 dition, EGFR mutations are often heterozygous, with Costa Rica 26.0–35.3 the mutant allele also showing gene amplification.69 Ap- Mexico 31.2 proximately 90% of these mutations are exon 19 dele- 70 Panama 24.8 tions or exon 21 L858R point mutations. Irrespective of ethnicity, EGFR mutations are more often found in Peru 67.0 females who are never-smokers and have adenocarcino- KRAS ma histology.67,68 However, EGFR mutations can also be By region found in other histological subtypes of NSCLC, and they Asian populations 3.8–8.0 can be found in former and current smokers.71-73 Western populations 18.0–26.0 In-frame deletions and insertions in exon 19 and STK11 the point mutation L858R in exon 21 are the most common activating EGFR mutations, and their prevalence By region is significantly higher in individuals of Japanese, Ko- Asian populations 3.0–7.0 rean, and Chinese descent than American or European Western populations 9.0–17.0 whites.74 Based on data from the Iressa Pan-Asia Study, BRAF a the frequency of activating EGFR mutations in Asian Overall 1.0–3.5 never-smokers or light smokers with advanced-stage 75 PIK3CAa adenocarcinoma was nearly 60%. By contrast, the Overall 1.0–3.0 rate of frequency was 15% in non-Asian patients based on data from phase 3 trials of retrospective EGFR mu- ALK Fusion 3.0–7.0 tation testing of archival samples.76,77 One study com- By region pared EGFR mutations in individuals of Asian and Rus- Asian populations 2.3–6.7 sian descent with advanced NSCLC and reported that Western populations 1.0–3.0 the frequency of EGFR mutations in adenocarcinoma aNo published reports have assessed racial and ethnic differences. was 49% in East Asians and 18% in Russians, and the frequencies of EGFR mutations in nonadenocarcino- ma lung cancers were 14% and 4%, respectively.78 Sev- has been limited on the impact of race and ethnicity eral studies have explored whether the frequencies of in targeted therapies. The impact and frequency of tu- EGFR mutations differ by race and ethnicity, and oth- mor mutations are not as well characterized in blacks ers have compared the frequencies of EGFR mutations and Hispanics/Latinos to the same extent as they are between blacks and whites with NSCLC; however, the in Asians and whites.62,64 Although evidence suggests results are conflicting.79-82 Specifically, Yang et al79 and that racial and ethnicity differences exist in markers Leidner et al80 found a lower frequency of EGFR muta- of susceptibility, we have chosen to focus on somatic tions among blacks, whereas Riely et al,81 Cote et al,82 mutations because of their clinical implications and and Reinersman et al83 did not find a statistically sig- observed racial and ethnic differences, as well as be- nificant difference between the 2 groups. cause germline variations in lung cancer have yet to be Few studies have addressed the frequency of deemed clinically actionable.62,65 mutations in lung cancer among Hispanics/Latinos. Arrieta et al84 analyzed 1,150 biopsy specimens from EGFR patients with NSCLC who were from Argentina, Co- Epidermal growth factor receptor (EGFR) is a trans- lombia, Peru, and Mexico and found that the combined membrane protein with cytoplasmic kinase activity frequency of EGFR mutations among all 4 countries that facilitates critical growth-factor signaling from the was 32.5%. The frequencies of mutations by country extracellular milieu to the cell.66 EGFR is expressed on were 19.3% in Argentina, 24.8% in Colombia, 31.2% the cell surface. Tyrosine kinase inhibitors can be ef- in Mexico, and 67.0% in Peru.84 The high frequency of

342 Cancer Control October 2016, Vol. 23, No. 4 EGFR mutations in Peru could be attributed to Asian phate–activated protein kinase (AMPK) and other migration.62 The frequency of EGFR mutations ranges members of the AMPK family.101,102 As a multifunc- from 26.0% to 35.3% in Costa Ricans and is 24.8% in tional kinase, STK11 is involved in a broad spectrum of Panamanians.85-87 By contrast, a lower frequency of cellular activity that includes metabolism, polarity, and EGFR mutations has also been reported among His- epithelial-mesenchymal transition, as well as cell-cycle panics/Latinos, and other conflicting reports have regulation, apoptosis, and autophagy.103 Germline mu- found no statistically significant difference between tations in STK11 were first identified in patients with Hispanics/Latinos and whites.62 Peutz-Jeghers syndrome,104 a rare autosomal dominant disorder associated with an increased risk of gas- KRAS trointestinal and other malignancies.105 Studies have The RAS family, which includes HRAS, NRAS, and also reported that STK11 somatic mutations are com- KRAS, encodes for membrane-bound guanosine tri- mon in NSCLC: The prevalence of inactivating mu- phosphate–binding proteins that regulate cell growth, tations can occur in more than 50% of cases (range, differentiation, and apoptosis by interacting with mito- 0.6%–44.4%), thus revealing an important role of gen-activated protein kinase, phosphoinositide 3-kinase, STK11 in lung tumorigenesis.103,106-112 Previously pub- and signal transducer and activator of transcription cas- lished data have reported that STK11 inactivating mu- cades.88,89 When aberrantly activated, KRAS is a potent tations occur in other histology subtypes of lung can- oncogenic driver when a point mutation occurs at codon cer and include 19% of squamous cell carcinomas, 14% 12 or 13 in exon 2 or codon 61 in exon 3.90 These muta- of large cell carcinomas, and 25% of adenosquamous tions result in impaired guanosine-triphosphatase ac- carcinomas.103,109,110,113-115 tivity and a constitutive activation of RAS signaling.91 Similar to KRAS and EGFR, ethnic and racial dif- Mutations in KRAS occur frequently in NSCLC — par- ferences appear in STK11 mutations. Studies in Asian ticularly adenocarcinoma (20%–30%) — and less com- populations (those of Japanese, Korean, and Chinese monly in squamous cell carcinoma (~ 7%).92,93 descent) have reported lower rates of STK11 mutations Although mutationally activated KRAS tumors were than whites (range, 3%–7%).115-118 This observation is originally identified in 1982,78 no successful treatment similar to that seen in KRAS mutations in the setting strategies target these tumors,93 and their impact on sur- of lung cancer, because they frequently co-occur with vival and prognosis in lung cancer is unclear and contro- STK11. Lung tumors in Western populations harbor a versial.94,95 To date, more than 50 studies have evaluated higher frequency of KRAS mutations compared with KRAS mutations on clinical outcomes in lung cancer, the Asian populations.98,119 Asian populations have also results of which are varied and inconsistent.94,95 A meta- been found to express an STK11 germline F354L poly- analysis of 41 studies concluded that KRAS mutations morphism at a frequence of approximately 10%.117 This are associated with a poor prognosis in patients with allele has been called a nonfunctional polymorphism NSCLC, and this was particularly true for those with ad- in lung cancer, but it also affects cell polarity main- enocarcinoma and early-stage NSCLC.95 tenance in an AMPK-dependent manner.120 At pres- Data have also suggested that the rates of frequen- ent, no data have been published on the frequency of cy of KRAS mutations differ by race and ethnicity. Spe- STK11 mutations among blacks and Hispanics/Latinos. cifically,KRAS mutations in the setting of lung cancer are less common in Asians compared with whites.96,97 BRAF The frequency of KRAS mutations ranged from 3.8% BRAF is a proto-oncogene that belongs to a family of to 8.0% in studies of Chinese study participants with serine-threonine protein kinases that also includes NSCLC — a rate lower than that seen in whites (range, ARAF and RAF1. Mutant BRAF has been implicated in 18%–26%).65,80,98 Few studies have assessed KRAS mu- the pathogenesis of several cancers.98 The most com- tational status among blacks, and the published data monly identified BRAF mutation is V600E, which ac- are inconsistent: 3 studies reported no statistically sig- counts for 90% of BRAF mutations in melanoma.121 By nificant difference in frequency of KRAS mutations be- contrast, the frequency of BRAF mutations in NSCLC tween whites and blacks,80,83,99 and some noted that the ranges from 1.0% to 3.5%.122-125 No reports have as- frequency in Hispanics may actually be lower.84,100 sessed racial and ethnic differences of BRAF mutations in patients with lung cancer, likely because this muta- STK11 tion is rare in this population. STK11 encodes a tumor suppressor located on chromosome 19p13.3 that encodes the serine/threonine PIK3CA kinase 11 (STK11) protein. The gene is made up of PIK3CA belongs to a gene family of lipid kinases in- coding exons 1 to 9 and a final, noncoding exon 10. volved in many cellular processes, including cell STK11 regulates cellular energy metabolism and growth, proliferation, differentiation, motility, and cell polarity by initiating adenosine monophos- survival. PIK3CA is mutated in more than 30% of

October 2016, Vol. 23, No. 4 Cancer Control 343 colorectal cancer cases.126 By contrast, mutations in ing with low-dose helical computed tomography can PIK3CA have been found in 1% to 3% of all types of mitigate the risk because it can detect early-stage lung NSCLC.126,127 Similar to BRAF, no reports have been cancer.54,55 Identifying and addressing cultural factors published that have assessed racial and ethnic dif- may also help improve the implementation of racially ferences of PIK3CA mutations in the setting of lung and ethnicity sensitive, population-based interven- cancer, likely because this mutation is rare in this pa- tions. Identifying and eliminating regional and racial tient population. and ethnic differences in lung cancer could contribute to more effective, preventive treatment strategies to ALK help reduce the disproportionate burden of lung can- Anaplastic lymphoma kinase is a tyrosine kinase cer in the United States.43 receptor abnormally expressed by forming a fusion gene with one of several other genes, by gaining ad- References ditional gene copies, or by somatic mutations. The 1. American Cancer Society. Global Cancer Facts & Figures. 3rd ed. EML4-ALK fusion was first documented in 2007 in http://www.cancer.org/acs/groups/content/@research/documents/document/acspc-044738.pdf. Accessed September 1, 2016. NSCLC as a potentially novel oncogenic-driver mutant 2. International Agency for Research on Cancer; World Health kinase.128 Approximately 3% to 7% of all lung tumors Organization. GLOBOCAN 2012: estimated cancer incidence, mortality and prevalence worldwide in 2012. http://globocan.iarc.fr/Pages/fact_sheets_ harbor ALK fusions, and EML4-ALK fusions are usu- cancer.aspx. Accessed August 29, 2016. ally found in light- or never-smokers and are typically 3. American Cancer Society. Cancer Facts & Figures 2016. http:// 128-132 www.cancer.org/acs/groups/content/@research/documents/document/ diagnosed at a young age. acspc-047079.pdf. Accessed August 29, 2016. EML4-ALK rearrangements may differ across 4. Farjah F, Wood DE, Yanez ND, 3rd, et al. Racial disparities among patients with lung cancer who were recommended operative therapy. Arch different racial groups. The frequency of translo- Surg. 2009;144(1):14-18. cation ranges from between 2.3% and 6.7% among 5. Steele CB, Pisu M, Richardson LC. Urban/rural patterns in receipt 133-135 of treatment for non-small cell lung cancer among black and white Medicare Asians. By contrast, EML4-ALK rearrangement is beneficiaries, 2000-2003. J Natl Med Assoc. 2011;103(8):711-718. lower in whites and ranges from 1.0% to 3.0%.128,131,136 6. Esnaola NF, Gebregziabher M, Knott K, et al. Underuse of surgical resection for localized, non-small cell lung cancer among whites and African One study analyzed a cohort of NSCLC samples col- Americans in South Carolina. Ann Thorac Surg. 2008;86(1):220-226. lected from Italy and Spain and found that 7.5% of the 7. van Ryn M. Research on the provider contribution to race/ethnicity disparities in medical care. Med Care. 2002;40(1 suppl):I140-I151. samples expressed EML4-ALK transcripts — a finding 8. Underwood JM, Townsend JS, Tai E, et al. Racial and regional dis- more similar to the data seen in Asians than whites.137 parities in lung cancer incidence. Cancer. 2012;118(7):1910-1918. 9. Yang R, Cheung MC, Byrne MM, et al. Do racial or socioeconomic disparities exist in lung cancer treatment? Cancer. 2010;116(10):2437-2447. Conclusions 10. Howlader N, Noone AM, Krapcho M, eds. SEER Cancer Statistics Review, 1975-2012. Bethesda, MD: National Cancer Institute. Based on Lung cancer is the most common cancer in the world November 2014 SEER data submission, posted to the SEER web site, April and the second most common cancer in both men and 2015. http://seer.cancer.gov/csr/1975_2012. Accessed June 30, 2016. 1-3 11. Centers for Disease Control and Prevention. Racial/ethnic dis- women in the United States. Although lung carcino- parities and geographic differences in lung cancer incidence: 38 states genesis is a multifactorial process driven by exogenous and the District of Columbia, 1998-2006. MMWR Morb Mortal Wkly Rep. 2010;59(44):1434-1438. exposure (eg, cigarette smoking), inherited genetic 12. Sankar P, Cho MK. Genetics. Toward a new vocabulary of human variations, and an accumulation of somatic genetic genetic variation. Science. 2002;298(5597):1337-1338. 13. Witzig R. The medicalization of race: scientific legitimization of a events, it also appears to have racial and ethnic differ- flawed social construct. Ann Intern Med. 1996;125(8):675-679. ences. However, many observed racial and ethnic dis- 14. Grieco EM, Cassidy RC. Overview of race and Hispanic origin. Census 2000 brief. https://www.census.gov/prod/2001pubs/c2kbr01-1.pdf. parities in lung cancer are not fully understood. Accessed August 29, 2016. The genomic diversity of oncogenes and tumor-sup- 15. Islami F, Torre LA, Jemal A. Global trends of lung cancer mortality and smoking prevalence. Transl Lung Cancer Res. 2015;4(4):327-338. pressor genes across racial and ethnic groups poses 16. Harris K, Khachaturova I, Azab B, et al. Small cell lung cancer dou- unique but important challenges for therapeutic op- bling time and its effect on clinical presentation: a concise review. Clin Med Insights Oncol. 2012;6:199-203. portunities to provide personalized medicine. Molecu- 17. Schabath MB, Thompson ZJ, Gray JE. Temporal trends in demo- lar genomic profiling for specific alterations is neces- graphics and overall survival of non-small-cell lung cancer patients at Moffitt Cancer Center from 1986 to 2008. Cancer Control. 2014;21(1):51-56. sary to identify patients likely to benefit from targeted 18. Kachroo S, Tong L, Spitz MR, et al. Trends in prevalence of prog- therapy. As such, identifying novel but actionable tar- nostic factors and survival in lung cancer patients from 1985 to 2004 at a tertiary care center. Cancer Detect Prev. 2008;32(2):101-108. getable mutations exclusive to specific racial and eth- 19. Dela Cruz CS, Tanoue LT, Matthay RA. Lung cancer: epidemiology, nic groups is critical to ensure these populations ben- etiology, and prevention. Clin Chest Med. 2011;32(4):605-644. 20. Birim O, Kappetein AP, van Klaveren RJ, et al. Prognostic factors in efit from such therapies. non-small cell lung cancer surgery. Eur J Surg Oncol. 2006;32(1):12-23. Even after smoking cessation is successfully ac- 21. National Cancer Institute. SEER Survival Monograph: Cancer Survival Among Adults: U.S. SEER Program, 1988-2001, Patient and complished, former smokers remain at significant risk Tumor Characteristics. NIH pub. no. 076215, 2007. https://surveillance. of developing lung cancer.19 Although effective smok- cancer.gov/publications/factsheets/SEER_Cancer_Survivor_Fact_Sheet.pdf. Accessed August 29, 2016. ing cessation programs and early cancer detection will 22. Paesmans M, Sculier JP, Lecomte J, et al. Prognostic factors for reduce the overall lung cancer burden, improvements patients with small cell lung carcinoma: analysis of a series of 763 patients included in 4 consecutive prospective trials with a minimum follow-up of 5 years. to the access of these modalities and the affordability Cancer. 2000;89(3):523-533. of health care are important topics to address.55 Screen- 23. Spiegelman D, Maurer LH, Ware JH, et al. Prognostic factors in

344 Cancer Control October 2016, Vol. 23, No. 4 small-cell carcinoma of the lung: an analysis of 1,521 patients. J Clin Oncol. lung cancer among never, former, and current smokers. J Natl Cancer Inst. 1989;7(3):344-354. 2012;104(10):778-789. 24. Albain KS, Crowley JJ, LeBlanc M, et al. Determinants of improved 51. Wu SH, Liu Z. Soy food consumption and lung cancer risk: a outcome in small-cell lung cancer: an analysis of the 2,580-patient Southwest meta-analysis using a common measure across studies. Nutr Cancer. Oncology Group data base. J Clin Oncol. 1990;8(9):1563-1574. 2013;65(5):625-632. 25. Osterlind K, Andersen PK. Prognostic factors in small cell lung can- 52. Wu QJ, Xie L, Zheng W, et al. Cruciferous vegetables consump- cer: multivariate model based on 778 patients treated with chemotherapy tion and the risk of female lung cancer: a prospective study and a meta- with or without irradiation. Cancer Res. 1986;46(8):4189-4194. analysis. Ann Oncol. 2013;24(7):1918-1924. 26. Schabath MB, Nguyen A, Wilson P, et al. Temporal trends from 53. Song J, Su H, Wang BL, et al. Fish consumption and lung cancer risk: 1986 to 2008 in overall survival of small cell lung cancer patients. Lung Cancer. systematic review and meta-analysis. Nutr Cancer. 2014;66(4):539-549. 2014;86(1):14-21. 54. Aberle DR, Adams AM, Berg CD, et al. Reduced lung-cancer mor- 27. Santarpia M, Giovannetti E, Rolfo C, et al. Recent developments in tality with low-dose computed tomographic screening. N Engl J Med. the use of immunotherapy in non-small cell lung cancer. Expert Rev Respir 2011;365(5):395-409. Med. 2016;10(7):781-798. 55. Slatore CG, Baumann C, Pappas M, et al. Smoking behaviors 28. Tchekmedyian N, Gray JE, Creelan BC, et al. Propelling im- among patients receiving computed tomography for lung cancer screen- munotherapy combinations into the clinic. Oncology (Williston Park). ing. Systematic review in support of the U.S. Preventive Services Task 2015;29(12):pii:214814. Force. Ann Am Thorac Soc. 2014;11(4):619-627. 29. Brahmer JR. Harnessing the immune system for the treatment of 56. US Preventive Services Task Force. Lung cancer: screening, non-small-cell lung cancer. J Clin Oncol. 2013;31(8):1021-1028. December 2013. http://www.uspreventiveservicestaskforce.org/Page/ 30. Margolis ML, Christie JD, Silvestri GA, et al. Racial differences Document/RecommendationStatementFinal/lung-cancer-screening. pertaining to a belief about lung cancer surgery: results of a multicenter Accessed August 29, 2016. survey. Ann Intern Med. 2003;139(7):558-563. 57. Tanner NT, Gebregziabher M, Hughes Halbert C, et al. Racial 31. Shugarman LR, Mack K, Sorbero ME, et al. Race and sex differ- differences in outcomes within the National Lung Screening Trial. ences in the receipt of timely and appropriate lung cancer treatment. Med Implications for widespread implementation. Am J Respir Crit Care Med. Care. 2009;47(7):774-781. 2015;192(2):200-208. 32. Zullig LL, Carpenter WR, Provenzale DT, et al. The association of 58. Kumar P, Gareen IF, Lathan C, et al. Racial differences in tobacco race with timeliness of care and survival among Veterans Affairs health cessation and treatment usage after lung screening: an examination of the care system patients with late-stage non-small cell lung cancer. Cancer National Lung Screening Trial. Oncologist. 2016;21(1):40-49. Manag Res. 2013;5:157-163. 59. Pinsky PF, Zhu CS, Kramer BS. Lung cancer risk by years since 33. Howington JA, Blum MG, Chang AC, et al. Treatment of stage I and quitting in 30+ pack year smokers. J Med Screen. 2015;22(3):151-157. II non-small cell lung cancer: diagnosis and management of lung cancer, 60. Clinical Lung Cancer Genome Project (CLCGP); Network Genomic 3rd ed: American College of Chest Physicians evidence-based clinical Medicine (NGM). A genomics-based classification of human lung tumors. practice guidelines. Chest. 2013;143(5 suppl):e278S-e313S. Sci Transl Med. 2013;5(209):209ra153. 34. Lin JJ, Mhango G, Wall MM, et al. Cultural factors associated with ra- 61. Pao W, Girard N. New driver mutations in non-small-cell lung cancer. cial disparities in lung cancer care. Ann Am Thorac Soc. 2014;11(4):489-495. Lancet Oncol. 2011;12(2):175-180. 35. Hardy D, Liu CC, Xia R, et al. Racial disparities and treatment 62. Cress WD, Chiappori A, Santiago P, et al. Lung cancer mutations and trends in a large cohort of elderly black and white patients with nonsmall use of targeted agents in Hispanics. Rev Recent Clin Trials. 2014;9(4):225-232. cell lung cancer. Cancer. 2009;115(10):2199-2211. 63. Politi K, Herbst RS. Lung cancer in the era of precision medicine. Clin 36. Hardy D, Xia R, Liu CC, et al. Racial disparities and survival for Cancer Res. 2015;21(10):2213-2220. nonsmall-cell lung cancer in a large cohort of black and white elderly patients. 64. Bollig-Fischer A, Chen W, Gadgeel SM, et al. Racial diversity Cancer. 2009;115(20):4807-4818. of actionable mutations in non-small cell lung cancer. J Thorac Oncol. 37. US Department of Health and Human Services. The Health Conse- 2015;10(2):250-255. quences of Smoking: 50 Years Of Progress. A Report of the Surgeon General. 65. Zhou W, Christiani DC. East meets West: ethnic differences in epi- Atlanta, GA: US Department of Health and Human Services; Centers for demiology and clinical behaviors of lung cancer between East Asians and Disease Control and Prevention; National Center for Chronic Disease Caucasians. Chin J Cancer. 2011;30(5):287-292. Prevention and Health Promotion; Office on Smoking and Health; 2014. 66. da Cunha Santos G, Shepherd FA, et al. EGFR mutations and lung 38. Carter BD, Freedman ND, Jacobs EJ. Smoking and mortality: cancer. Annu Rev Pathol. 2011;6:49-69. beyond established causes. N Engl J Med. 2015;372(22):2170. 67. Pao W, Miller V, Zakowski M, et al. EGF receptor gene mutations 39. Pesch B, Kendzia B, Gustavsson P, et al. Cigarette smoking and lung are common in lung cancers from “never smokers” and are associated with cancer: relative risk estimates for the major histological types from a pooled sensitivity of tumors to gefitinib and erlotinib. Proc Natl Acad Sci U S A. analysis of case-control studies. Int J Cancer. 2012;131(5):1210-1219. 2004;101(36):13306-13311. 40. US Department of Health and Human Services. How Tobacco Smoke 68. Paez JG, Janne PA, Lee JC, et al. EGFR mutations in lung cancer: Causes Disease: The Biology and Behavioral Basis for Smoking-Attributable correlation with clinical response to gefitinib therapy. Science. Disease: A Report of the Surgeon General. Atlanta: US Department of 2004;304(5676):1497-1500. Health and Human Services; Centers for Disease Control and Prevention; 69. Soh J, Okumura N, Lockwood WW, et al. Oncogene mutations, National Center for Chronic Disease Prevention and Health Promotion; copy number gains and mutant allele specific imbalance (MASI) frequently Office on Smoking and Health; 2010. occur together in tumor cells. PLoS One. 2009;4(10):e7464. 41. Jamal A, Homa DM, O’Connor E, et al. Current cigarette smoking 70. Ladanyi M, Pao W. Lung adenocarcinoma: guiding EGFR-targeted among adults: United States, 2005-2014. MMWR Morb Mortal Wkly Rep. therapy and beyond. Mod Pathol. 2008;21(suppl 2):S16-S22. 2015;64(44):1233-1240. 71. Shi Y, Li J, Zhang S, et al. Molecular epidemiology of EGFR muta- 42. Centers for Disease Control and Prevention. Cigarette smoking tions in Asian patients with advanced non-small-cell lung cancer of adeno- among adults: United States, 2006. MMWR Morb Mortal Wkly Rep. carcinoma histology - mainland China subset analysis of the PIONEER 2007;56(44):1157-1161. study. PLoS One. 2015;10(11):e0143515. 43. Jamal A, Agaku IT, O’Connor E, et al. Current cigarette smoking 72. Dogan S, Shen R, Ang DC, et al. Molecular epidemiology of EGFR among adults: United States, 2005-2013. MMWR Morb Mortal Wkly Rep. and KRAS mutations in 3,026 lung adenocarcinomas: higher susceptibil- 2014;63(47):1108-1112. ity of women to smoking-related KRAS-mutant cancers. Clin Cancer Res. 44. Haiman CA, Stram DO, Wilkens LR, et al. Ethnic and racial dif- 2012;18(22):6169-6177. ferences in the smoking-related risk of lung cancer. N Engl J Med. 73. Varghese AM, Sima CS, Chaft JE, et al. Lungs don’t forget: com- 2006;354(4):333-342. parison of the KRAS and EGFR mutation profile and survival of collegiate 45. Driscoll T, Nelson DI, Steenland K, et al. The global burden of dis- smokers and never smokers with advanced lung cancers. J Thorac Oncol. ease due to occupational carcinogens. Am J Ind Med. 2005;48(6):419-431. 2013;8(1):123-125. 46. Alberg AJ, Brock MV, Ford JG, et al. Epidemiology of lung can- 74. Shigematsu H, Lin L, Takahashi T, et al. Clinical and biological fea- cer: diagnosis and management of lung cancer, 3rd ed: American College tures associated with epidermal growth factor receptor gene mutations in of Chest Physicians evidence-based clinical practice guidelines. Chest. lung cancers. J Natl Cancer Inst. 2005;97(5):339-346. 2013;143(5 suppl):e1S-e29S. 75. Yang JC, Wu YL, Chan V, et al. Epidermal growth factor receptor 47. Brinton LA, Gierach GL, Andaya A, et al. Reproductive and hor- mutation analysis in previously unanalyzed histology samples and cytology monal factors and lung cancer risk in the NIH-AARP Diet and Health Study samples from the phase III Iressa Pan-ASia Study (IPASS). Lung Cancer. cohort. Cancer Epidemiol Biomarkers Prev. 2011;20(5):900-911. 2014;83(2):174-181. 48. Yao Y, Gu X, Zhu J, et al. Hormone replacement therapy in females 76. Mok TS, Wu YL, Thongprasert S, et al. Gefitinib or carboplatin-pacli- can decrease the risk of lung cancer: a meta-analysis. PLoS One. taxel in pulmonary adenocarcinoma. N Engl J Med. 2009;361(10):947-957. 2013;8(8):e71236. 77. Zhu CQ, da Cunha Santos G, Ding K, Sakurada A, et al. Role of 49. Yang Y, Dong J, Sun K, et al. Obesity and incidence of lung cancer: KRAS and EGFR as biomarkers of response to erlotinib in National Can- a meta-analysis. Int J Cancer. 2013;132(5):1162-1169. cer Institute of Canada Clinical Trials Group Study BR.21. J Clin Oncol. 50. Smith L, Brinton LA, Spitz MR, et al. Body mass index and risk of 2008;26(26):4268-4275.

October 2016, Vol. 23, No. 4 Cancer Control 345 78. Han B, Tjulandin S, Hagiwara K, et al. Determining the prevalence 109. Matsumoto S, Iwakawa R, Takahashi K, et al. Prevalence and of EGFR mutations in Asian and Russian patients (pts) with advanced non- specificity of LKB1 genetic alterations in lung cancers. Oncogene. small-cell lung cancer (aNSCLC) of adenocarcinoma (ADC) and non-ADC 2007;26(40):5911-5918. histology: Ignite Study. Ann Oncol. 2015;26(suppl 1):29-44. 110. Ji H, Ramsey MR, Hayes DN, et al. LKB1 modulates lung cancer 79. Yang SH, Mechanic LE, Yang P, et al. Mutations in the tyrosine ki- differentiation and metastasis. Nature. 2007;448(7155):807-810. nase domain of the epidermal growth factor receptor in non-small cell lung 111. O’Neill GM, Seo S, Serebriiskii IG, et al. A new central scaffold for me- cancer. Clin Cancer Res. 2005;11(6):2106-2110. tastasis: parsing HEF1/Cas-L/NEDD9. Cancer Res. 2007;67(19):8975-8979. 80. Leidner RS, Fu P, Clifford B, et al. Genetic abnormalities of the 112. Gao Y, Xiao Q, Ma H, et al. LKB1 inhibits lung cancer progression EGFR pathway in African American Patients with non-small-cell lung cancer. through lysyl oxidase and extracellular matrix remodeling. Proc Natl Acad J Clin Oncol. 2009;27(33):5620-5626. Sci U S A. 2010;107(44):18892-18897. 81. Riely GJ, Pao W, Pham D, et al. Clinical course of patients with 113. Wingo SN, Gallardo TD, Akbay EA, et al. Somatic LKB1 mutations non-small cell lung cancer and epidermal growth factor receptor exon 19 promote cervical cancer progression. PloS One. 2009;4(4):e5137. and exon 21 mutations treated with gefitinib or erlotinib. Clin Cancer Res. 114. Sanchez-Cespedes M. A role for LKB1 gene in human cancer be- 2006;12(3 pt 1):839-844. yond the Peutz-Jeghers syndrome. Oncogene. 2007;26(57):7825-7832. 82. Cote ML, Haddad R, Edwards DJ, et al. Frequency and type of 115. Koivunen JP, Kim J, Lee J, et al. Mutations in the LKB1 tumour sup- epidermal growth factor receptor mutations in African Americans with pressor are frequently detected in tumours from Caucasian but not Asian non-small cell lung cancer. J Thorac Oncol. 2011;6(3):627-630. lung cancer patients. Br J Cancer. 2008;99(2):245-252. 83. Reinersman JM, Johnson ML, Riely GJ, et al. Frequency of EGFR 116. Onozato R, Kosaka T, Achiwa H, et al. LKB1 gene mutations in and KRAS mutations in lung adenocarcinomas in African Americans. Japanese lung cancer patients. Cancer Sci. 2007;98(11):1747-1751. J Thorac Oncol. 2011;6(1):28-31. 117. Gao B, Sun Y, Zhang J, et al. Spectrum of LKB1, EGFR, and KRAS 84. Arrieta O, Cardona AF, Federico Bramuglia G, et al. Genotyping mutations in Chinese lung adenocarcinomas. J Thorac Oncol. 2010 non-small cell lung cancer (NSCLC) in Latin America. J Thorac Oncol. Aug;5(8):1130-1135. 2011;6(11):1955-1959. 118. Sun YH, Ren Y, Fang ZY, et al. Lung adenocarcinoma from East 85. Corrales-Rodriguez L, Abbas J, Ramos-Esquivel A, et al. Immu- Asian never-smokers is a disease largely defined by targetable oncogenic nohistochemistry as a predictor of EGFR mutations: a retrospective case- mutant kinases. J Clin Oncol. 2010;28(30):4616-4620. control study. J Thorac Oncol. 2014;9(9 suppl 3):S176. 119. Schabath MB, Welsh EA, Fulp WJ, et al. Differential association of 86. Rojas Vigott R, Ugalde Gamboa L, Rojas Solano J. Thoracic on- STK11 and TP53 with KRAS mutation-associated gene expression, pro- cology clinic: our first 200 cases. San José, Costa Rica. J Thorac Oncol. liferation and immune surveillance in lung adenocarcinoma. Oncogene. 2014;9(9 suppl 3):S182. 2016;35(24):3209-3216. 87. Castillo-Fernandez O, Arauz E, Cabreja A, et al. EGFR mutation and 120. Forcet C, Etienne-Manneville S, Gaude H, et al. Functional analysis outcomes in patients with metastatic NSCLC treated with erlotinib at Instituto of Peutz-Jeghers mutations reveals that the LKB1 C-terminal region exerts Oncologico Nacional, Panama. J Thorac Oncol. 2014;9(9 suppl 3):S189. a crucial role in regulating both the AMPK pathway and the cell polarity. 88. Riely GJ, Marks J, Pao W. KRAS mutations in non-small cell lung Hum Mol Genet. 2005;14(10):1283-1292. cancer. Proc Am Thorac Soc. 2009;6(2):201-205. 121. Sánchez-Torres JM, Viteri S, Molina MA, et al. BRAF mutant non- 89. Adjei AA. Blocking oncogenic Ras signaling for cancer therapy. small cell lung cancer and treatment with BRAF inhibitors. Transl Lung J Natl Cancer Inst. 2001;93(14):1062-1074. Cancer Res. 2013;2(3):244-250. 90. Cox AD, Der CJ. Ras history: the saga continues. Small GTPases. 122. Davies H, Bignell GR, Cox C, et al. Mutations of the BRAF gene in 2010;1(1):2-27. human cancer. Nature. 2002;417(6892):949-954. 91. Castellano E, Santos E. Functional specificity of ras isoforms: so 123. Naoki K, Chen TH, Richards WG, et al. Missense mutations of the BRAF similar but so different. Genes Cancer. 2011;2(3):216-231. gene in human lung adenocarcinoma. Cancer Res. 2002;62(23):7001-7003. 92. Pao W, Girard N. New driver mutations in non-small-cell lung cancer. 124. Brose MS, Volpe P, Feldman M, et al. BRAF and RAS mutations in Lancet Oncol. 2011;12(2):175-180. human lung cancer and melanoma. Cancer Res. 2002;62(23):6997-7000. 93. Ding L, Getz G, Wheeler DA, et al. Somatic mutations affect key 125. Marchetti A, Felicioni L, Malatesta S, et al. Clinical features and pathways in lung adenocarcinoma. Nature. 2008;455(7216):1069-1075. outcome of patients with non-small-cell lung cancer harboring BRAF muta- 94. Califano R, Landi L, Cappuzzo F. Prognostic and predictive value of K- tions. J Clin Oncol. 2011;29(26):3574-3579. RAS mutations in non-small cell lung cancer. Drugs. 2012;72(suppl 1):28-36. 126. Samuels Y, Wang Z, Bardelli A, et al. High frequency of mutations 95. Meng D, Yuan M, Li X, et al. Prognostic value of K-RAS mutations of the PIK3CA gene in human cancers. Science. 2004;304(5670):554. in patients with non-small cell lung cancer: a systematic review with meta- 127. Kawano O, Sasaki H, Endo K, et al. PIK3CA mutation status in analysis. Lung Cancer. 2013;81(1):1-10. Japanese lung cancer patients. Lung Cancer. 2006;54(2):209-215. 96. Riely GJ, Kris MG, Rosenbaum D, et al. Frequency and distinctive 128. Soda M, Choi YL, Enomoto M, et al. Identification of the trans- spectrum of KRAS mutations in never smokers with lung adenocarcinoma. forming EML4-ALK fusion gene in non-small-cell lung cancer. Nature. Clin Cancer Res. 2008;14(18):5731-5734. 2007;448(7153):561-566. 97. Sun Y, Ren Y, Fang Z, et al. Lung adenocarcinoma from East Asian 129. Koivunen JP, Mermel C, Zejnullahu K, et al. EML4-ALK fusion gene never-smokers is a disease largely defined by targetable oncogenic mutant and efficacy of an ALK kinase inhibitor in lung cancer. Clin Cancer Res. kinases. J Clin Oncol. 2010;28(30):4616-4620. 2008;14(13):4275-4283. 98. El-Telbany A, Ma PC. Cancer genes in lung cancer: racial disparities: 130. Wong DW, Leung EL, So KK, et al. The EML4-ALK fusion gene is are there any? Genes Cancer. 2012;3(7-8):467-480. involved in various histologic types of lung cancers from nonsmokers with 99. Bauml J, Mick R, Zhang Y, et al. Frequency of EGFR and KRAS wild-type EGFR and KRAS. Cancer. 2009;115(8):1723-1733. mutations in patients with non small cell lung cancer by racial background: 131. Inamura K, Takeuchi K, Togashi Y, et al. EML4-ALK fusion is linked do disparities exist? Lung Cancer. 2013;81(3):347-353. to histological characteristics in a subset of lung cancers. J Thorac Oncol. 100. McQuitty E, Zhang W, Hendrickson H, et al. Lung adenocarcinoma 2008;3(1):13 -17. biomarker incidence in Hispanic versus non-Hispanic white patients. Arch 132. Inamura K, Takeuchi K, Togashi Y, et al. EML4-ALK lung cancers Pathol Lab Med. 2014;138(3):390-394. are characterized by rare other mutations, a TTF-1 cell lineage, an acinar 101. Shaw RJ, Kosmatka M, Bardeesy N, et al. The tumor suppres- histology, and young onset. Mod Pathol. 2009;22(4):508-515. sor LKB1 kinase directly activates AMP-activated kinase and regulates 133. Ren S, Kuang P, Zheng L, et al. Analysis of driver mutations in fe- apoptosis in response to energy stress. Proc Natl Acad Sci U S A. male non-smoker Asian patients with pulmonary adenocarcinoma. Cell 2004;101(10):3329-3335. Biochem Biophys. 2012;64(2):155-160. 102. Cao Y, Li H, Liu H, et al. The serine/threonine kinase LKB1 controls 134. Li C, Fang R, Sun Y, et al. Spectrum of oncogenic driver mutations thymocyte survival through regulation of AMPK activation and Bcl-XL in lung adenocarcinomas from East Asian never smokers. PLoS One. expression. Cell Res. 2010;20(1):99-108. 2011;6(11):e28204. 103. Gao Y, Ge G, Ji H. LKB1 in lung cancerigenesis: a serine/threonine 135. Kim HR, Shim HS, Chung JH, et al. Distinct clinical features and out- kinase as tumor suppressor. Protein Cell. 2011;2(2):99-107. comes in never-smokers with nonsmall cell lung cancer who harbor EGFR 104. Hemminki A, Markie D, Tomlinson I, et al. A serine/threonine kinase or KRAS mutations or ALK rearrangement. Cancer. 2012;118(3):729-739. gene defective in Peutz-Jeghers syndrome. Nature. 1998;391(6663):184-187. 136. Shinmura K, Kageyama S, Tao H, et al. EML4-ALK fusion tran- 105. Lim W, Hearle N, Shah B, et al. Further observations on LKB1/ scripts, but no NPM-, TPM3-, CLTC-, ATIC-, or TFG-ALK fusion transcripts, STK11 status and cancer risk in Peutz-Jeghers syndrome. Br J Cancer. in non-small cell lung carcinomas. Lung Cancer. 2008;61(2):163-169. 2003;89(2):308-313. 137. Martelli MP, Sozzi G, Hernandez L, et al. EML4-ALK rearrangement 106. Sanchez-Cespedes M, Parrella P, Esteller M, et al. Inactivation of in non-small cell lung cancer and non-tumor lung tissues. Am J Pathol. LKB1/STK11 is a common event in adenocarcinomas of the lung. Cancer 2009;174(2):661-670. Res. 2002;62(13):3659-3662. 107. Shah U, Sharpless NE, Hayes DN. LKB1 and lung cancer: more than the usual suspects. Cancer Res. 2008;68(10):3562-3565. 108. Sanchez-Cespedes M. The role of LKB1 in lung cancer. Fam Cancer. 2011;10(3):447-453.

346 Cancer Control October 2016, Vol. 23, No. 4 We must continue to focus attention

and effort to reduce the disproportionate

burden of cancer among blacks living

in the United States.

Cactus. X-Ray image courtesy of Bryan Whitney. www.x-rayphotography.com.

Black Heterogeneity in Cancer Mortality: US-Blacks, Haitians, and Jamaicans Paulo S. Pinheiro, MD, PhD, Karen E. Callahan, MPH, Camille Ragin, PhD, MPH, Robert W. Hage, MD, PhD, DLO, MBA, Tara Hylton, MPH, and Erin N. Kobetz, PhD, MPH

Introduction: The quantitative intraracial burden of cancer incidence, survival and mortality within black populations in the United States is virtually unknown. Methods: We computed cancer mortality rates of US- and Caribbean-born residents of Florida, specifically focusing on black populations (United States, Haiti, Jamaica) and compared them using age-adjusted mortality ratios obtained from Poisson regression models. We compared the mortality of Haitians and Jamaicans residing in Florida to populations in their countries of origin using Globocan. Results: We analyzed 185,113 cancer deaths from 2008 to 2012, of which 20,312 occurred in black populations. The overall risk of death from cancer was 2.1 (95% CI: 1.97–2.17) and 1.6 (95% CI: 1.55–1.71) times higher for US-born blacks than black Caribbean men and women, respectively (P < .001). Conclusions: Race alone is not a determinant of cancer mortality. Among all analyzed races and ethnicities, including Whites and Hispanics, US-born blacks had the highest mortality rates while black Caribbeans had the lowest. The biggest intraracial difference was observed for lung cancer, for which US-blacks had nearly 4 times greater mortality risk than black Caribbeans. Migration from the islands of Haiti and Jamaica to Florida resulted in lower cancer mortality for most cancers including cervical, stomach, and prostate, but increased or stable mortality for 2 obesity-related cancers, colorectal and endometrial cancers. Mortality results in Florida suggest that US-born blacks have the highest incidence rate of “aggressive” prostate cancer in the world, rather than Caribbean men.

Introduction men and women suffer the largest cancer burden, Cancer is the second leading cause of death among with higher mortality and lower survival compared to blacks in the United States. In relative terms, black all major racial-ethnic groups: whites, Asians, Hispan-

From the School of Community Health Sciences University of Nevada Las Vegas, Las Vegas, Nevada (PSP, KEC), the African-Caribbean Cancer Consortium (PSP, CR, RWH), College of Public Health, Temple University and Fox Chase Cancer Center, Philadelphia, Pennsylva- nia (CR), Department of Anatomy, St. George's University, St. George’s, Grenada, West Indies (RWH), Florida Department of Health, Tal- lahassee, Florida (TH), and Sylvester Comprehensive Cancer Center, University of Miami Miller School of Medicine, Miami, Florida (ENK). Submitted November 23, 2015; accepted March 4, 2016. This document is the unedited authors’ version of a submitted work that was accepted for publication. Address correspondence to Paulo S. Pinheiro, MD, PhD, University of Nevada Las Vegas, School of Community Health Sciences, African- Caribbean Cancer Consortium, 4505 South Maryland Parkway, Las Vegas, NV 89154. E-mail: [email protected] No significant relationships exist between the authors and the companies/organizations whose products or services may be referenced in this article. Data provided by the Florida Department of Health, Bureau of Vital Statistics. Any conclusions are the authors’ own and do not necessarily reflect the opinion of the Florida Department of Health. Dr Pinheiro is sponsored by 8 P20 GM103440-11 from the National Institute of General Medical Sciences as a NV/INBRE recipient.

October 2016, Vol. 23, No. 4 Cancer Control 347 ics, and American Indians.1,2 Black males also have the living in the United States is non-existent. Problems greatest cancer incidence.2 These cancer disparities by studying this topic derive from limitations in the avail- race are the result of a complex and entangled web of able incidence and survival data. First, cancer registries differences in risk factors that directly impact not only do not collect demographic data on black ancestries. cancer incidence, but also access to and availability of Secondly, the collection of birthplace is substantially early detection and timely treatment. In turn, these af- incomplete in the United States cancer surveillance fect the capacity to successfully combat cancer, thus programs.12,13 These limitations have impeded unbi- impacting cancer survival. Cancer mortality is a result ased cancer incidence and survival studies of these of the interplay between both incidence and survival. populations.14-16 Birthplace reporting on death certifi- Assessing each of these 3 components — incidence, cates, however, is nearly complete. Thus, mortality survival, and mortality — is critical to understanding data can be a valuable source of information for assess- how disparities between populations arise and how ing the intraracial cancer heterogeneity among blacks they can be reduced or eliminated. in the United States. Black men and women in the United States are of- Here we use cancer mortality data for Florida, ten categorized under a single ostensibly homogenous a state with a 19% immigrant and 17% black popula- umbrella: blacks, United States blacks, African Ameri- tion,17 to analyze cancer outcome heterogeneity. First, cans, or people of African descent. For researchers, the we compare US-born populations of all races with use of different denominations for describing popula- Hispanic Caribbean populations and majority-black tions can lead to confusion. “United States blacks”, com- Caribbean populations born outside the United States. monly used for data collection, may or may not char- Then, we compare US-born blacks with majority-black acterize blacks born outside of the United States, and Caribbean populations. Within the latter, we look spe- “African descent” may include North African popula- cifically at Haitians, Jamaicans, and other West Indian tions. For well-documented historical reasons, the vast populations. Through this analysis we aim to char- majority of black people living in the Western hemi- acterize race-specific patterns of cancer mortality, ex- sphere are genetic admixtures of African, European, plore the intraracial heterogeneity in blacks, and make and/or Native American populations, warranting cau- a contribution that will assist health policy makers in tion when attributing risk to race, a social construct.3 shaping targeted cancer prevention and control efforts Most importantly, however, the use of one aggregate to reduce health inequities in cancer outcomes in the term obscures considerable heterogeneity in health out- United States. comes among the black population in the United States, irrespective of whether these differences are attribut- Materials and Methods able to biological, cultural, or socioeconomic factors.4 Cancer Data Blacks born outside the United States number Cancer mortality data for 5 years, January 1, 2008 3.8 million in the United States, nearly 9% of the black through December 31, 2012, were obtained from the population.5 According to Census projections, this per- Florida Department of Health Vital Statistics.18 We con- centage will reach 17% by 2060.6 Substantial black im- sidered only resident cases in the state. We analyzed migration into the United States is a recent dynamic the most common causes of cancer death: lung and following the Immigration America Act of 1963.7 The bronchus, breast, prostate, colorectal, pancreas, cor- majority of black immigrants come from English and pus uteri (endometrial); and cancers of special interest French speaking nations of the Caribbean, led by Ja- in populations born outside the United States: cervix maica and Haiti, the birthplace of 18% and 15% of Unit- uteri, liver, stomach; and all-sites-combined, which ined States-black immigrants in 2013.6 To the extent that cluded all cases of malignant cancers. Cancer site was health is partially determined by social and cultural coded according to the International Statistical Classifi- factors, including beliefs on disease etiology, types of cation of Diseases 10th revision. family structure and social support, expectations re- We examined the cancer mortality burden at garding medical care, as well as cumulative lifetime ex- 2 levels of comparison. First, all Florida populations posure to cancer risk factors, aggregating these distinct regardless of race and based on birthplace alone: US- Caribbean populations together with blacks who have born populations and the following 2 immigrant Ca- been in the United States for generations may limit the ribbean populations: Hispanic Caribbean, including ability to detect and address culturally driven determi- Cuba, the Dominican Republic and Puerto Rico, and nants of health outcomes among all black heteroge- majority-black Caribbean including Haiti, Jamaica and neous sub-populations.4,8,9 Other West Indies (OWI). Haiti and Jamaica are consid- Two studies have shown lower overall cancer mor- ered separately as they are the 2 largest non-Hispanic tality for blacks born outside of the United States than Caribbean immigrant groups in Florida as well as the for US-born.10,11 Yet a more detailed analysis of the in- United States.7 The OWI category included the follow- traracial cancer heterogeneity in black populations ing territories and island nations: Anguilla, Antigua

348 Cancer Control October 2016, Vol. 23, No. 4 and Barbuda, Aruba, Bahamas, Barbados, Cayman there were 14.6 million US-born (10.2 million whites, Islands, Dominica, Grenada, Guadeloupe, Martinique, 2.4 million blacks, 2 million other races); 1.3 million Montserrat, Saint Kitts and Nevis, Saint Barthelemy, Hispanic Caribbean-born; and 570,000 black Carib- Saint Lucia, Saint Martin, Saint Vincent and the Gren- bean-born (277,000 Haitians, 196,000 Jamaicans and adines, the Netherlands Antilles, the British Virgin 97,000 OWIs).19 Islands, Trinidad and Tobago, Turks and Caicos, and Population denominators for the State of Florida West Indies not otherwise specified. The vast majority were obtained from the 5-year 2008–2012 American of immigrants from these nations are black, including Community Survey using the University of Minnesota 99% of all Haitians immigrants and 96% of all Jamai- interface,20 thus serving the same years as our mortal- cans.7 Thus, while recognizing that a small number of ity data. Population numbers were grouped by birth- immigrants born in these countries may not be black, place and by gender for all groups, and also by race for we will for the purposes of this paper refer to these the US-born. populations as black Caribbean. To strictly adhere to the Caribbean definition as a geographical region dis- Data Analysis tinct from the continental Americas, Belize and Guyana, A very small number of cases of non-Hispanic blacks (although also majority black populations), were ex- with unknown birthplace (n = 145, 0.7% of the total) cluded from the OWI category. were proportionally reassigned by imputation models Second, we examined a subset of the Florida pop- stratified by age, sex, and cancer site into the United ulation, based on race for US-born populations and States, Haiti, Jamaica, OWI, and all other non-US birth- birthplace for those born in the Caribbean. We com- places as described elsewhere.21 pare US-born blacks with the black population born Cancer mortality rates for 2008–2012 were calculat- outside the United States. Caribbean populations com- ed per 100,000 persons, annualized and age-standard- bined (Haiti, Jamaica, and OWI), as well as the 2 larg- ized to the 2000 US Standard Population and the 2000 est populations, Haitians and Jamaicans. For simplicity World Standard Population using eighteen age group purposes in the current study, the Haitian-born, Jamaican- bands, all 5-year except the last, which was 85 and old- born, and those born in other West Indies (OWI), and er. We used mortality and population data for a period Hispanic Caribbean-born populations are referred to of 5 years (2008–2012), accumulating cancer mortality as Haitian, Jamaican, OWI and Hispanic Caribbean. data for a total of 2.85 million person-years for black Ca- The designation US-blacks will exclusively refer to US- ribbean populations. Corresponding 95% confidence born blacks. intervals (CIs) were calculated with gamma intervals Non-Hispanic black populations from Africa or modification.22 Site-specific mortality ratios with 95% elsewhere outside the United States or the Caribbean CIs were computed by using Poisson regression adjust- were excluded from this study. These excluded popula- ed for 5-year age groups starting at age 35, except tions are relatively young with a more recent immigra- for prostate cancer where the groups started at age 50. tion history,6,7 and accounted only for 1% — approxi- Because breast cancer risk factors differ between pre- mately 40 deaths annually — of all cancer deaths in and post-menopausal women and these rates differ by Florida among blacks in 2008–2012. race,23 we used approximate pre- and post-menopausal ages, using 50 as our dividing age, to calculate pre- Population Data and postmenopausal ratios. SAS 9.3 (SAS Institute Inc., The American Community Survey estimates for 2010 Cary, NC, USA) was used for data analysis. indicated that 30% (795,500) of the total 2.6 million The OWI cancer burden in Florida was distributed people of non-Hispanic West Indian ancestry in the between more than 20 different countries and territo- United States, lived in Florida.19 Of the 2.4 million ries of origin, the largest of which were Trinidad and blacks in Florida, 33% were of Caribbean ancestry, Tobago at 34%, Bahamas at 22% and Barbados at 9%. representing the highest proportion of a Caribbean This diverse origin as well as relatively small numbers black-majority population in any American state.19 makes interpretation of the results difficult. Moreover, Since specific ancestry information is unavailable for unlike Haiti and Jamaica, some of these countries have cancer mortality data, we are using birthplace. While a more diverse racial composition, including Trinidad ancestry and birthplace are not exactly overlapping, and Tobago, with a 38% East Indian population.7 None- the overwhelming majority of the Caribbean popula- theless, we present OWI data to demonstrate Carib- tion in the United States aged 30 and above was born bean completeness. outside the United States.6,7 These older age groups are Finally, we calculated age-adjusted mortality rates, also the ones where cancer will be an important cause adjusted to the 2000 World Standard, for Haitian and of death; therefore, birthplace and ancestry are not Jamaican immigrant populations and well as US-born likely to result in substantially different mortality num- blacks living in Florida. bers. Looking at birthplace alone in Florida in 2010, For comparison purposes, we used the best avail-

October 2016, Vol. 23, No. 4 Cancer Control 349 able data from each of these countries, the 2009-2011 av- liver cancers. Both Hispanic and black Caribbean pop- erage cancer mortality rates from the World Health Orga- ulations had lower rates for lung, pancreas and breast nization Mortality Database24 for Jamaica, and the 2012 cancers when compared to the US-born population. Globocan25 estimates for Haiti. This study was approved by the University of Nevada, Table 1. — Average Annual Age-Adjusted* Mortality Rates for Selected Cancers Las Vegas Institutional Review (per 100,000) by Birthplace: Florida, 2008–2012 Board. A data use agreement was obtained from the Florida US-Born Born Outside the United States Vital Statistics Department. United States Hispanic Caribbean** Majority Black All Races All Races Caribbean Results All Races A total of 841,054 deaths Deaths Rate Deaths Rate Deaths Rate were recorded among Florida (95% CI) (95% CI) (95% CI) residents during 2008–2012; Male 185,113 were deaths attribut- Stomach 1,386 3.2 235 4.7 97 6.7| ed to cancers in US-born and (3.0–3.4) (4.2–5.5) (5.4–8.4) Caribbean-born populations Colorectal 7,455 16.9 917 18.2 206 13.8 (Table 1). There were 16,119 (16.6–17.3) (17.1–19.5) (11.9–16.1) cancer deaths in US-blacks, Liver and 3,766 8.4 476 9.3 109 6.9 and 4,113 among black Carib- Intrahepatic (8.1–8.7) (8.5–10.3) (5.6–8.5) beans. People born in Jamai- Bile Duct ca accounted for 43% of these Pancreas 5,392 12.1 540 10.6 128 9.1 deaths, those from Haiti, 38%, (11.7–12.4) (9.7–11.6) (7.5 –11.1) and from OWI countries, 19%. Lung and 28,023 62.3 2,287 44.7 317 21.0 Bronchus (61.6–63.1) (42.8–46.6) (18.6–23.6) For 2008–2012 in Florida, cancer was the leading cause Prostate 7,936 17.9 943 19.1 437 35.5 (17.5 –18.3) (17.9 –20.4) (32.1–39.3) of death for US-born women All combined 89,387 201.7 8,487 169.4 2,020 141.4 as well as for the black Ca- (200.4–203.1) (165.7–173.2) (134.9–148.2) ribbean populations. It ac- Female counted for 27.2% of all deaths Stomach 911 1.7 173 2.6 92 4.6 among Jamaicans and 23.7% (1.6–1.8) (2.2–3.2) (3.7–5.8) among Haitians. Cancer ranked Colorectal 6,335 11.7 855 12.4 243 11.9 second only to heart disease (11.4–12.0) (11.6–13.4) (10.4–13.6) as the leading cause of death Liver and 1,518 2.8 260 3.8 76 3.9 among the US-born male Intrahepatic (2.7–3.0) (3.3–4.4) (3.0–5.0) population as well as males Bile Duct and females born in Hispanic Pancreas 4,874 8.9 537 7.6 129 6.4 Caribbean countries. (8.7–9.2) (7.0 – 8.4) (5.4 –7.8) In comparison to the total Lung and 22,217 41.9 1090 16.1 194 9.8 US-born Florida population, Bronchus (41.4–42.5) (15.2–17.2) (8.4–11.4) both black Caribbeans and Breast 10,819 21.6 1,100 17.5 413 19.1 Hispanic Caribbeans have (21.2–22.1) (16.4–18.6) (17.3–21.2) lower all-sites-combined Pre- 1,141 3.9 100 2.6 97 4.8 menopausal, (3.7–4.1) (2.1–3.2) (3.9–6.0) cancer mortality rates. How- less than 50 ever, there is considerable Post- 9,678 17.8 1,000 14.9 316 14.3 heterogeneity in the rates menopausal, (17.5 –18.2) (14.0–15.9) (12.7–16.1) between them and by cancer 50 or more site (Tables 1-2). For black Cervix Uteri 1,140 2.8 138 2.6 68 3.3 Caribbeans, there were signif- (2.6–3.0) (2.2–3.2) (2.5–4.3) icantly higher mortality rates Corpus and 1,975 3.8 278 4.2 152 7.3 for stomach, prostate and en- Uterus (3.6–4.0) (3.7–4.8) (6.2–8.7) (endometrial) dometrial cancers than the All combined 74,241 141.4 6,890 104.6 2,093 103.6 US-born; among Hispanic (140.4–142.5) (102.1–107.3) (98.7–108.8) Caribbeans, mortality rates were higher for stomach and *2000 US Standard Population **Cuba, Dominican Republic and Puerto Rico

350 Cancer Control October 2016, Vol. 23, No. 4 Table 2. — Average Annual Age-Adjusted* Mortality Rates for Selected Cancers per 100,000 for US-Born and Blacks Born Outside the United States: Florida, 2008–2012 US-Born Born Outside the United States Blacks Black Caribbean** Haitian Jamaican Other West Indian All Races All Races All Races (OWI) All Races Deaths Rate Deaths Rate Deaths Rate Deaths Rate Deaths Rate (95% CI) (95% CI) (95% CI) (95% CI) (95% CI) Male Stomach 292 9.4 97 6.7 45 7.4 39 6.8 <15 5.0 (8.3–10.6) (5.4–8.4) (5.2–10.3) (4.8–11.9) (2.4–14.5) Colorectal 895 28.3 206 13.8 83 12.4 86 14.7 37 15.4 (26.3–30.4) (11.9–16.1) (9.7–15.8) (11.6–20.3) (10.4–25.7) Liver and 436 11.0 109 6.9 69 9.4 28 4.6 <15 4.5 Intrahepatic (9.9–12.2) (5.6–8.5) (7.2–12.2) (3.0–9.6) (2.0–14.0) Bile Duct Pancreas 474 14.7 128 9.1 34 5.0 63 11.5 31 11.6 (13.3–16.2) (7.5 –11.1) (3.4–7.4) (8.8 –17.0) (7.5 –21.4) Lung and 2,536 78.9 317 21.0 105 17.5 153 25.1 59 21.2 Bronchus (75.6–82.3) (18.6–23.6) (14.1–21.6) (21.1–31.2) (15.6–31.7) Prostate 1,218 50.6 437 35.5 142 29.2 194 37.8 101 44.0 (47.6 – 53.7) (32.1–39.3) (24.3–34.7) (32.4–45.2) (35.0 – 57.3) All combined 8,575 273.9 2,020 141.4 753 124.5 883 154.4 384 148.3 (267.6–280.3) (134.9–148.2) (115.1–134.6) (143.9–166.4) (132.2–167.6)

Female Stomach 179 4.0 92 4.6 42 4.9 31 3.7 19 6.4 (3.4–4.6) (3.7–5.8) (3.5–6.9) (2.5–6.8) (3.8–11.8) Colorectal 809 17.7 243 11.9 75 8.6 122 14.2 46 13.4 (16.5–19.0) (10.4–13.6) (6.7–11.1) (11.8–18.1) (9.7–19.5) Liver and 169 3.6 76 3.9 37 4.6 28 3.3 <15 3.4 Intrahepatic (3.0–4.1) (3.0–5.0) (3.2–6.6) (2.2–6.5) (1.6–8.2) Bile Duct Pancreas 532 11.8 129 6.4 53 6.7 47 5.3 29 9.3 (10.8–12.8) (5.4 –7.8) (5.0–9.0) (3.9–8.6) (6.1–15.0) Lung and 1,473 32/3 194 9.8 66 8.6 95 11.1 33 10.3 Bronchus (30.6–34.0) (8.4–11.4) (6.6–11.2) (8.9–14.7) (7.0 –16.1)

Breast 1,514 32.1 413 19.1 158 17.0 187 21.1 68 19.0 (30.4–33.7) (17.3–21.2) (14.4–20.2) (18.2–25.4) (14.6–25.6)

Pre- 298 6.2 97 4.8 50 5.2 38 5.2 <15 2.6 menopausal, (5.5–6.9) (3.9–6.0) (3.9–7.1) (3.7–8.6) (1.2–7.3) less than 50

Post- 1,216 25.9 316 14.3 108 11.8 149 15.9 <68 16.4 menopausal, (24.4–27.4) (12.7–16.1) (9.6–14.6) (13.5–19.2) (12.4–22.7) 50 or more Cervix Uteri 267 5.5 68 3.3 36 3.9 20 2.6 <15 3.6 (4.8–6.2) (2.5–4.3) (2.7–5.7) (1.6–5.8) (1.8–8.5)

Corpus and 373 8.1 152 7.3 49 5.7 75 8.6 28 7.6 Uterus (7.3–9.0) (6.2–8.7) (4.2–7.8) (6.7–12.1) (5.0–12.8) (endometrial)

All combined 7,544 163.2 2,093 103.6 798 95.6 902 105.4 393 121.3 (159.5–167.0) (98.7–108.8) (88.5–103.3) (98.4–113.4) (106.8–137.8)

*2000 US Standard Population **Haiti, Jamaica and OWI combined

October 2016, Vol. 23, No. 4 Cancer Control 351 When narrowing the comparisons to Haitian, tate cancer mortality was highest, at 35.5 per 100,000, Jamaican, and US-black populations in Florida, the followed by lung cancer at 21.0 per 100,000 (Table 2). cancer mortality rate differences are surprisingly even Across the male populations of US-blacks, Jamaicans, greater (Table 2). Haitians have the lowest mortal- Haitians, and OWIs, colorectal was the third leading ity rates for all cancers combined, 125 per 100,000 cause of cancer death, while pancreatic cancer was among males and 96 per 100,000 in females, followed fourth, except among Haitians for whom liver cancer closely by Jamaicans with 154 and 105 per 100,000 was fourth. among males and females, respectively. Contrasting Among US-black women in Florida, the 2 most with these relatively low rates, US-blacks had high common causes of cancer mortality were lung and mortality rates, 274 per 100,000 for males and 163 breast, almost equal at 32.3 and 32.1 per 100,000 re- per 100,000 for females. The risk of cancer death in spectively, followed by colorectal and pancreatic can- US-black males is 2.1 times higher than among black cers. For black Caribbean women, breast cancer was Caribbean males and 1.6 times higher for black Carib- by far the leading cause of cancer mortality at 19.1 per bean females (P < .05). Of all cancers examined, lung 100,000, with colorectal cancer second and lung can- cancer has the greatest risk differential. US-blacks cer only third (Table 2). Still, US-black women had a compared to black Caribbeans have an astonishing 1.7 times greater overall mortality risk for breast can- 4 times greater risk of lung cancer death among males cer than black Caribbean women, although pre-meno- and a 3.5 times greater risk among females (Table 3). pausal breast cancer risk was only 1.3 times higher The leading cause of cancer mortality among US- (Table 3). Endometrial cancer was the fourth leading black males was lung cancer, at 78.9 per 100,000, fol- cause of cancer death for black Caribbean women. No- lowed distantly by prostate cancer at 50.6 per 100.000. tably, cervical cancer rates were relatively low across For black Caribbean males the opposite was true: pros- black Caribbean female populations. Lastly, the cancer mortality rates of Haitian and Jamaican populations residing in Florida when com- Table 3. — Mortality Rate Ratios for Selected Cancers: pared to the rates in their respective countries of origin Florida, 2008–2012 are lower for both males and females for all cancer sites combined (Table 4). This advantage is most marked for US-Born Blacks in Relation to Black Caribbean cervix, stomach, and prostate cancers for both coun- (Reference Population) tries, liver cancer for Haitians and female breast for Jamaicans. On the other hand, for colorectal (male and Rate Ratio 95% CI female), endometrial, and pancreatic cancers, mortal- Male ity rates are higher for Haitians and relatively similar Stomach 1.4 (1.13–1.80) for Jamaicans in Florida in comparison to rates found Colorectal 2.1 (1.81–2.46) in the countries of origin. Also higher in Florida than Liver and Intrahepatic Bile Duct 1.9 (1.55–2.39) in countries of origin are rates for lung cancer among Haitian males and Jamaican females as well as breast Pancreas 1.8 (1.45–2.15) cancer among Haitian females (Table 4). Lung and Bronchus 4.0 (3.51–4.45) Prostate 1.4 (1.29–1.61) Discussion All combined 2.1 (1.97–2.17) This is the first study to address intraracial differences in Female cancer mortality among persons often categorized un- Stomach 0.9 (0.66–1.09) der the US-blacks designation. Using birthplace, we dif- Colorectal 1.5 (1.30–1.73) ferentiate and directly compare black populations born in and outside the United States, presenting the first Liver and Intrahepatic Bile Duct 1.0 (0.77–1.34) quantification of the cancer burden for black Caribbe- Pancreas 1.9 (1.55–2.27) ans. In Florida, cancer mortality rates for US-blacks are Lung and Bronchus 3.5 (2.99–4.03) higher than all analyzed groups. However, black Carib- Breast 1.7 (1.48–1.84) bean populations have the lowest rates, suggesting that Pre-menopausal, less than 50 1.3 (0.99–1.60) racial health disparities in the United States are attribut- Post-menopausal, 50 or more 1.8 (1.55–1.99) able to more than socioeconomic status (SES) and health Cervix Uteri 1.7 (1.30–2.24) care access, for which Caribbeans are likely to be at least equally as disadvantaged as US-blacks. Corpus and Uterus 1.1 (0.93–1.35) (endometrial) The major contributor for the relatively low overall cancer mortality in blacks born outside the United All combined 1.6 (1.55–1.71) States is the difference in lung cancer rates. This is *Poisson regression rate ratios adjusted for age group most likely a reflection of differences in smoking hab-

352 Cancer Control October 2016, Vol. 23, No. 4 represent a healthier sample Table 4. — Mortality Rates* (per 100,000) for Selected Cancers for Haitians and Jamaicans in Country of Origin and in Florida and for US-Blacks: Florida, 2008–2012 of the population than those who remain in the country Haitians US-Blacks Jamaicans US-Blacks of origin.32 Having said that, In country In Florida In Florida In Country In Florida In Florida immigrants also likely carry of origin, of Origin, with them baseline protective Haiti Jamaica factors inherent to their coun- Male tries of origin that are persis- Stomach 8.7 5.1 6.3 9.6 4.9 6.3 tent in reducing risk, and re- Colorectal 4.6 9.2 19.3 9.8 10.7 19.3 sistant to the negative effects Liver and 7.5 7.4 8.8 3.1 4.0 8.8 of acculturation, as has been Intrahepatic shown for Hispanics, particu- Bile Duct larly for Mexican Hispanics.21 Pancreas 2.6 3.6 10.3 3.1 7.2 10.3 Caribbean immigrants in the Lung and 6.6 11.9 55.1 24.1 18.7 55.1 United States tend to maintain Bronchus traditional dietary patterns Prostate 32.3 17.3 28.8 40.1 22.2 28.8 that include less red meat con- All-site 88.4 85.6 186.3 131.8 107.4 186.3 sumption and higher intake of combined whole grains, fruits and veg- Female etables, which are protective Stomach 6.0 3.6 2.6 4.1 2.7 2.6 against colorectal and cancers Colorectal 5.8 6.4 12.3 9.7 9.5 12.3 of the upper gastrointestinal tract.33-36 Also, fertility rates Liver and 5.1 3.2 2.6 2.4 2.2 2.6 Intrahepatic and breastfeeding patterns Bile Duct may help explain breast can- Pancreas 2.3 4.6 8.2 3.1 3.5 8.2 cer differences.23 Several stud- Lung and 6.2 5.6 23.1 6.0 7.5 23.1 ies demonstrated significantly Bronchus higher odds of breastfeeding Breast 11.5 13.4 24.0 21.5 16.0 24.0 initiation among immigrant Cervix uteri 14.6 3.1 4.4 12.8 2.0 4.4 black women than US-born blacks.37-39 Overall, the lower Corpus and 1.4 4.2 6.0 6.6 5.9 6.0 Uterus rates we found for breast and (endometrial) colorectal cancers among All-site 80.3 69.2 117.7 99.7 75.0 117.7 black Caribbeans are all the combined more surprising in light of *Average annual age-adjusted to 2000 World Standard. previous studies document- Country of origin data: Haiti, GLOBOCAN 2012; Jamaica, WHO Mortality Database 200–2011 average. ing lower rates of mammography and colonoscopy among blacks born outside the United its 2 to 4 decades ago. In 1980, the prevalence of smok- States, especially Haitians, in comparison to US-born ing in Haiti (18% males, 6% females) and Jamaica blacks.40-43 Additionally, if they are of low SES, non– (27% males, 7% females) was much lower than in US- US-citizen immigrants are less likely than citizens to blacks (45% males, 31% female).26 Moreover, studies have had consistent health insurance.44 In the case of have documented that blacks born outside the United the undocumented or recently documented, access is States are less likely to smoke than their US-born coun- blocked to any federal assistance programs in Florida terparts.27,28 The risk factor of smoking may also ac- except limited emergency Medicaid.45 count for some of the differences seen in stomach and While most intraracial cancer mortality differences pancreatic cancers, as well as other tobacco-related in our study seem to be determined by environmental cancer sites like bladder, larynx, and oral cavity which factors, some of which are inherently related to place are included in the all sites combined category.29 of birth, it is important to note that some patterns are Our lower cancer mortality findings among Carib- seemingly more related to race and possibly genetics, bean blacks may be partially explained by the Healthy regardless of geography. Endometrial, premenopaus- Immigrant Effect (HIE), whereby immigrants have bet- al breast (defined here as below age 50) and prostate ter overall health outcomes not solely related to observ- cancer rates reveal specific vulnerabilities. Compared able socioeconomic factors such as education and in- to other groups, all black populations (US-blacks, come.10,30,31 In other words, those who immigrate may Haitians, and Jamaicans) have higher mortality for

October 2016, Vol. 23, No. 4 Cancer Control 353 these 3 cancers. Endometrial cancer is a cancer of the en.60 For Caribbeans, Hepatitis B likely accounts for the Western world,46 highly related to obesity,36 whose high rates, especially among Haitians,61 and would not prevalence commonly increases among first genera- result in a marked gender difference. As a result, liver tion immigrants to the United States.21 Studies have cancer mortality rates are higher for US black males shown that endometrial cancer survival and mortality but similar between US-black females and black Carib- outcomes for black women, even after adjustment for bean females in our study. their higher proportion of hysterectomies, are worse Haitians and Jamaicans were clearly distinct in than all other racial groups.47,48 Black women also have their cancer mortality patterns, although the precision an elevated risk of more aggressive tumor subtypes of the statistical comparisons may be impacted by the with poorer prognosis.48,49 relatively small size of the populations in Florida. For premenopausal breast cancer, the finding of Haitians had lower all-sites-combined mortality rates high mortality among all analyzed black populations than Jamaicans as well as all other analyzed popula- relative to the all US races category as well as Hispanic tions. However, for all 3 infection-related cancers, liver, Caribbeans is intriguing and requires further study. cervical and stomach, Haitian mortality rates exceeded Given the lack of detailed incidence data for each of those of Jamaicans. Haitian rates were lower than Ja- these populations, it is impossible to assess whether maicans’ for prostate, breast, endometrial and colorectal this finding can be attributed to the well-documented cancers: all cancers commonly associated with a “West- survival disparity between blacks and Whites in the ern” lifestyle.62 Jamaicans have a longer immigration United States,50 or if it reflects a true increased risk for history in Florida than Haitians;20 therefore it is possi- premenopausal breast cancer among black women. ble that these differences reflect acculturation patterns, For prostate cancer, black African descent has predicted by length of time in the United States, and the been well established as a risk factor.1,51,52 Our results extent of assimilation into US culture, which may be confirm high rates among both US-born and Carib- greater for English-speaking Jamaicans. Additionally, bean-born blacks. However, the significantly higher since the country of Jamaica has been more developed rates we find for US-blacks compared to Caribbean- than Haiti for at least the past 25 years,63 the prevalence born blacks is notable, given the lower prostate can- of baseline immigrant risk factors such as obesity may cer screening rates for the latter compared to the for- already be more “Westernized”.64 mer.40,53,54 While prostate cancer incidence rates are Additional important findings arise from the ex- challenging to compare globally due to increased use amination of differences in mortality rates between of prostate-specific antigen (PSA) screening in some Haitians and Jamaicans in their countries of origin countries,55,56 mortality rates in Florida likely reflect compared to those same populations in Florida. Since the incidence of aggressive prostate cancer. Thus, our mortality is a function of cancer incidence and surviv- study is highly indicative of US-blacks having the high- al, changes in mortality rates may reflect a difference in est aggressive prostate cancer incidence in the world, either or both of these. If the healthcare infrastructure not Caribbeans as previously suggested.57 in Florida is better than in Jamaica and Haiti, we would Three infection-related cancers — cervical, liver expect cancer survival to be higher, and thus mortal- and stomach cancer — are more commonly found in ity rates to be lower among Jamaicans and Haitians in developing countries.46 Strangely, the mortality rates Florida than in their countries of origin. Despite the are similar or even higher for US-blacks compared to fact that survival data accuracy for blacks born outside the black populations born outside the United States. the United States can be problematic,16 some studies For cervical cancer, despite the fact that the pap test have documented this pattern.65-68 On the other hand, screening rates are higher among US-black women if Caribbeans living in Florida have higher or even sta- than Caribbean-born black women23,53,58 which would ble cancer mortality rates compared to their counter- predict earlier stages at diagnosis with better progno- parts in their countries of origin, then these changes sis, the mortality rates for cervical cancer are higher likely reflect true increases in incidence rather than de- in US-blacks than Caribbean-born black women. For creases in survival. liver cancer, it is possible that the epidemiology and In our study, mortality rates for Haitians and Ja- patterns of transmission are different between these maicans in Florida compared to their countries of ori- populations, resulting in different gender-specific inci- gin were predictably lower for all 3 infection-related dence and thus mortality. Hepatitis C, a more common cancers, except for liver cancer in male Jamaicans. liver cancer risk factor than hepatitis B in the United Stomach cancer gains are likely attributable to a higher States, is largely transmitted by sharing contaminated standard of living in the United States coinciding with intravenous drug needles.59 The greater differential in a lower prevalence of infection with Helicobacter liver cancer between males and females in US-blacks pylori.69,70 Likewise, lower liver cancer mortality rates than in Caribbean blacks may be the result of greater in Florida populations compared to their countries of IV drug use among US-black men than US-black wom- origin suggest lower incidence, at least in part due to

354 Cancer Control October 2016, Vol. 23, No. 4 a lower prevalence among US blacks (0.9%) of chronic the benefit of better health care access. The greater hepatitis B compared to prevalence in Jamaica (3.7%) total baseline fertility rate in Haiti than Jamaica or Haiti (13.6%).61,71 Gains for cervical cancer were sub- (3.1 compared to 2.3 children77) as well as higher obe- stantial in both Haitian and Jamaican populations in sity rates in Jamaica64 may also play a role in explaining Florida, with mortality rates reduced to less than 25% why breast cancer rates in Florida are higher among of the rates in their countries of origin. This likely re- Jamaican than Haitian women. flects improved access to health care, including screen- For lung cancer, for Haitian males and Jamaican fe- ing and radiotherapy treatment, which would impact males, rates were significantly higher in Florida than incidence and survival, respectively. The same likely Haiti. Lung cancer has a poor survival prognosis re- applies to the substantial improvement for female gardless of gender, SES, or race, and even location.78 breast cancer among Jamaicans. Therefore, this difference would suggest an increase in We also found an advantage for prostate cancer for incidence rather than a change in survival. Yet among immigrants living in Florida. Mortality rates among Jamaican men, lung cancer rates were much lower in Haitians and Jamaicans are substantially lower in Flori- Florida than in Jamaica, suggesting an intriguing de- da than in the country of origin. This is not remarkable creased incidence. It is possible that these results re- given the complexity of prostate cancer control: includ- flect changes in smoking prevalence when immigrants ing screening, computed tomography and bone scans settle in the United States.79 Given the importance of for staging, surgery, hormone therapy and radiotherapy. smoking as a risk factor for many cancers, these chang- The mortality rates for colorectal cancer in both es need to be carefully monitored. sexes and endometrial cancer in women suggest an in- Our study has the usual limitations of descriptive creased risk in Florida compared to the country of ori- epidemiology. There is a lack of both individual level gin. For Haitians, the rates are significantly higher than risk factor data for all subjects as well as group level in their countries of origin, while for Jamaicans there risk factor data specific to Caribbean-born populations is minimal change. Yet, in the United States, access to in Florida. We also had some relatively small popula- colorectal screening if utilized would result in either tion sizes, which sometimes resulted in a lack of pre- averted cancer by removal of precancerous lesions or cision in our mortality rates, particularly for the small earlier cancer diagnosis.72 Moreover, since there is no OWI group. We attempted to mitigate this by accumu- knowledge of latent forms of either colorectal or en- lating 5 years of mortality experience and annualizing dometrial cancers, which eventually progress to death the rates. Another limitation is that Globocan mortal- if left untreated, access to better health care facilities ity estimates for 2012 for Haitians are modelled, rely- should result in improved treatment, prolonged sur- ing on the last date for which real data were available, vival, and lower mortality in Florida. Therefore, the lack which may be several years old. of decreases seen in mortality for these cancers must be Our study examines black populations in Florida, rooted in an increase in incidence, which is a worrisome which may not be representative of all black popula- finding. Unfortunately, a common risk factor for both tions in the United States. Additionally, in theory, cancers is obesity.1 Data at the population level shows a our mortality numbers could reflect a “salmon bias”, trend of increasing prevalence of obesity and transition whereby small numbers of immigrants who are very ill to a high caloric fatty diet when immigrants settle in the return to their home countries of origin to die, which United States.64,73,74 would artificially lower mortality rates for those born Some of the results did not follow known or ex- outside the United States in our study. However, the pected patterns. Breast cancer has protective risk fac- magnitude of this out-migration for Hispanics has tors associated with fertility patterns23 that have been been found to be very small,80 and the literature for shown to change with migration to the United States, this phenomenon among black immigrants in the including delayed childbearing and a reduction in to- United States is non-existent. tal number of pregnancies.75,76 In our study, for Haitian Relying on birthplace rather than ancestry may females, breast cancer mortality rates were somewhat result in a small degree of misclassification. Haitian, higher in Florida than Haiti. Offsetting a possible gain Jamaican and other West Indian ancestry populations, in survival due to screening and more access to quality which may include second generation Caribbeans, treatment, the increased mortality among Haitians in are larger than the birthplace populations we used for Florida may well derive from an increased incidence of this study. Some of the people in the US-born blacks breast cancer which may be higher due to diminished category in our study are quite possibly of Haitian, fertility and delayed childbearing among other factors. Jamaican or OWI ancestry. However, given the trend On the other hand, for Jamaican women, breast cancer of relatively recent immigration for these populations, mortality rates were much lower in Florida than in this would likely only affect cancer sites with mortality Jamaica. It is possible that Jamaican women in Florida, in relatively younger age groups, such as cervical can- who tend to have higher levels of education,7 may reap cer. Our Florida mortality data provided information

October 2016, Vol. 23, No. 4 Cancer Control 355 on Haitian ancestry, but not any other Caribbean an- knowledge of cancer epidemiology among black pop- cestry. To test for differences between using birthplace ulations, especially regarding incidence data. Better and ancestry, we calculated mortality rates using the understanding of intraracial cancer patterns including Haitian ancestry information and found that the differ- the extent to which acculturation impacts cancer inci- ences in rates were negligible, thus reducing the likeli- dence, survival, and mortality among black US immi- hood of a threat to validity based on misclassification. grants will help in targeted US cancer prevention and Strengths of our study include the very high com- control efforts. pleteness (99%) of birthplace information among all deaths in the state of Florida, allowing for reliable clas- Conclusions sification of the different populations, which is unfor- The US-born black population has the highest cancer tunately never achieved in cancer incidence data. We mortality risk in the nation — even higher than previ- also used detailed denominators based on birthplace ously documented when rates for blacks born outside available from the American Community Survey. the United States are calculated separately. Continued Overall, cancer mortality rates among black attention and effort is required to reduce the dispro- Caribbeans are remarkably low. While there has been portional burden of cancer in US blacks. On the other much research demonstrating health advantages for hand, a record 3.8 million black immigrants, account- Hispanics, our results point to an even greater health ing for 9% of the current black population in the Unit- advantage for black Caribbeans than Hispanic Carib- ed States,5 are currently low risk populations for most beans in Florida. More striking is the markedly worse cancers. It is important to monitor these trends, look- cancer mortality rates for US-blacks when compared ing for ways to help all populations acquire protective to Caribbean blacks. The low mortality among Carib- factors and resist risk factors for cancer. More studies bean blacks has the effect of improving the overall US into the mechanisms that would avoid an unfavorable cancer rates for blacks. Not surprisingly, of the 17 states trend towards worse cancer outcomes in Caribbean with black populations over 1 million, Florida and New blacks are necessary. Our study suggests that obesity, York, states with the highest proportions of black a known risk factor for many US blacks, may be an Caribbean populations, have the lowest cancer mortal- important focus for intervention in Caribbean blacks, ity rates for blacks in the United States.1 especially among females, as immigrants have higher To our knowledge these are the only results avail- colorectal and endometrial cancer mortality rates in able looking at cancer outcomes on a population basis relation to their countries of origin. Targeted screen- that analyze US-born blacks separately, directly com- ing programs and expanding access to care would also paring them to Caribbean-born people residing in the help these immigrant populations. Concurrently, the United States. The mortality rates presented here will public health community must clarify and address the also constitute a baseline for monitoring the cancer reasons for such poor cancer outcomes among US- burden among Caribbean populations and help the born blacks so that the interracial gaps can be signifi- fight against cancer across all majority-black Caribbean cantly narrowed. nations. Cancer literature is somewhat scarce for the Caribbean region due to a lack of reliable data. Phillips References et al. used estimates from Globocan 2002 data to char- 1. American Cancer Society. Cancer Facts & Figures for African Americans 2013-2014. Atlanta, GA: American Cancer Society; 2013. acterize cancer incidence and mortality in the Carib- 2. Kohler BA, Sherman RL, Howlader N, et al. Annual report to the bean nations.81 However, in a quickly evolving world, nation on the status of cancer, 1975-2011, featuring incidence of breast cancer subtypes by race/ethnicity, poverty, and state. J Natl Cancer Inst. these data are possibly now outdated. Fortunately, can- 2015;107(6):djv048. cer registration in Jamaica and other Caribbean nations 3. Benn-Torres J, Bonilla C, Robbins CM, et al. Admixture and pop- 46 ulation stratification in African Caribbean populations. Ann Hum Genet. is showing signs of improvement. 2008;72(pt 1):90-98. The current study highlights once again the need 4. Arthur CM, Katkin ES. Making a case for the examination of ethnicity of Blacks in United States Health Research. J Health Care Poor Underserved. to thoroughly address known racial disparities in can- 2006;17(1):25-36. cer risk and outcomes in the United States. To analyze 5. Colby SL, Ortman JM. Projections of the Size and Composition of the U.S. Population: 2014 to 2060, Current Population Reports. Washington, DC: intraracial differences, it is fundamental to have spe- U.S. Census Bureau; 2014. https://www.census.gov/content/dam/Census/ cific and accurate cancer data, including data on im- library/publications/2015/demo/p25-1143.pdf. Accessed July 15, 2016. 6. Anderson M. A Rising Share of the U.S. Black Population is Foreign migrant groups. In states with large Caribbean pop- Born: 9 Percent are Immigrants; and While Most are From the Caribbean, ulations, such as Florida and New York, it would be Africans Drive Recent Growth. Washington, DC: Pew Research Center; 2015. http://www.pewsocialtrends.org/2015/04/09/a-rising-share-of-the-u-s- beneficial to have health data on this population col- black-population-is-foreign-born/. Access July 18, 2016. lected methodically as has been done for other racial/ 7. Thomas KAJ. A Demographic Profile of Black Caribbean Immi- grants in the United States. Washington, DC: Migration Policy Institute; ethnic subgroups in California and New York City. An 2012. http://www.migrationpolicy.org/research/CBI-demographic-profile- active role of the North American Central Cancer Reg- black-caribbean-immigrants. Accessed July 18, 2016. 8. Singh GK, Siahpush M. Ethnic-immigrant differentials in health be- istries Association and the central cancer registries haviors, morbidity, and cause-specific mortality in the United States: an in Florida and New York would greatly improve the analysis of two national data bases. Hum Biol. 2002;74(1):83-109.

356 Cancer Control October 2016, Vol. 23, No. 4 9. Consedine NS, Magai C, Spiller R, et al. Breast cancer knowledge feeding initiation. Arch Pediatr Adolesc Med. 2005;159(3):255-260. and beliefs in subpopulations of African American and Caribbean women. 39. Singh GK, Kogan MD, Dee DL. Nativity/immigrant status, race/ Am J Health Behav. 2004;28(3):260-271. ethnicity, and socioeconomic determinants of breastfeeding initiation and 10. Singh GK, Siahpush M. All-cause and cause-specific mortality duration in the United States, 2003. Pediatrics. 2007;119:suppl 1:S38-S46. of immigrants and native born in the United States. Am J Public Health. 40. Consedine NS, Tuck NL, Ragin CR, et al. Beyond the black box: 2001;91(3):392-399. a systematic review of breast, prostate, colorectal, and cervical screening 11. Fang J, Madhavan S, Alderman MH. Influence of nativity on cancer among native and immigrant African-descent Caribbean populations. mortality among black New Yorkers. Cancer. 1997;80(1):129-135. J Immigr Minor Health. 2015;17(3):905-924. 12. Pinheiro PS, Bungum TJ. Country of origin and breast cancer sur- 41. Gwede CK, William CM, Thomas KB, et al. Exploring disparities vival. Asia‐Pac J Clin Oncol. 2014;10(3):279. and variability in perceptions and self-reported colorectal cancer screen- 13. Pinheiro PS, Bungum TJ, Jin H. Limitations in the imputation strat- ing among three ethnic subgroups of U. S. Blacks. Oncol Nurs Forum. egy to handle missing nativity data in the Surveillance, Epidemiology, and 2010;37(5):581-591. End Results program. Cancer. 2014;120(20):3261-3262. 42. Magai C, Consedine N, Conway F, et al. Diversity matters: 14. Pinheiro PS, Williams M, Miller EA, et al. Cancer survival among Lati- Unique populations of women and breast cancer screening. Cancer. nos and the Hispanic Paradox. Cancer Causes Control. 2011;22(4):553-561. 2004;100(11):2300-2307. 15. Gomez SL, Glaser SL. Quality of cancer registry birthplace 43. Wilcox ML, Acuña JM, de la Vega PR, et al. Factors affecting com- data for Hispanics living in the United States. Cancer Causes Control. pliance with colorectal cancer screening among households residing in the 2005;16(6):713-723. largely Haitian community of Little Haiti, Miami-Dade County, Florida: an 16. Pinheiro PS, Morris CR, Liu L, et al. The impact of follow-up type observational study. Medicine (Baltimore). 2015;94(18):e806. and missed deaths on population-based cancer survival studies for 44. Derose KP, Escarce JJ, Lurie N. Immigrants and health care: sources Hispanics and Asians. J Natl Cancer Inst Monogr. 2014;2014(49):210-217. of vulnerability. Health Aff (Millwood). 2007;26(5):1258-1268. 17. U.S. Census Bureau. QuickFacts: Florida. http://www.census.gov/ 45. The Florida Health Care Landscape. The Henry J. Kaiser Family quickfacts/table/PST045215/12,00. Accessed July 18, 2016. Foundation – The Kaiser Commission on Medicaid and the Uninsured. 18. Bureau of Vital Statistics, Florida Department of Health, November 2013 Fact Sheet. http://kff.org/health-reform/fact-sheet/the- Jacksonville, Florida. Mortality data 2008-2012, received June 2015. florida-health-care-landscape/. Accessed July 19, 2016. 19. U.S. Census Bureau. American Fact Finder. Selected Population 46. Forman D, Bray F, Brewster DH, et al. Cancer Incidence in Five Profile in the United States: 2009-2011 American Community Survey Continents, Vol. X. IARC Scientific Publications, No. 164. Lyon, France: 3-Year Estimates, Table S0201. http://factfinder.census.gov/faces/table- International Agency for Research on Cancer; 2014. services/jsf/pages/productview.xhtml?src= bkmk. Accessed July 18, 2016. 47. Jamison PM, Noone AM, Ries LA, et al. Trends in endometrial can- 20. Ruggles S, Genadek K, Goeken R, Grover J, and Sobek cer incidence by race and histology with a correction for the prevalence of M. Integrated Public Use Microdata Series: Version 6.0 [Machine-read- hysterectomy, SEER 1992 to 2008. Cancer Epidemiol Biomarkers Prev. able database]. American Community Survey 2012 5-Year Estimates. 2013;22(2):233-241. Minneapolis: University of Minnesota. Accessed June 2015. 48. Cote ML, Ruterbusch JJ, Olson SH, et al. The Growing Burden of 21. Pinheiro PS, Sherman RL, Trapido EJ, et al. Cancer incidence in Endometrial Cancer: A Major Racial Disparity Affecting Black Women. first generation U.S. Hispanics: Cubans, Mexicans, Puerto Ricans, and Cancer Epidemiol Biomarkers Prev. 2015;24(9):1407-1415. new Latinos. Cancer Epidemiol Biomarkers Prev. 2009;18(8):2162-2169. 49. Felix AS, Weissfeld JL, Stone RA, et al. Factors associated 22. Tiwari RC, Clegg LX, Zou Z. Efficient interval estimation for age- with Type I and Type II endometrial cancer. Cancer Causes Control. adjusted cancer rates. Stat Methods Med Res. 2006;15(6):547-569. 2010;21(11):1851-1856. 23. American Cancer Society. Breast Cancer Facts & Figures, 2015- 50. Daly B, Olopade OI. A perfect storm: how tumor biology, genomics, 2016. Atlanta, GA: American Cancer Society; 2015. and health care delivery patterns collide to create a racial survival disparity 24. World Health Organization. WHO Mortality Database. http://apps. in breast cancer and proposed interventions for change. CA Cancer J Clin. who.int/healthinfo/statistics/mortality/whodpms/. Accessed July 19, 2016. 2015;65(3):221-238. 25. Ferlay J, Soerjomataram I, Dikshit R, et al. Cancer incidence and 51. Blake G, Hanchard B, Mitchell K, et al. Jamaica cancer mortality mortality worldwide: sources, methods and major patterns in GLOBOCAN statistics, 1999. West Indian Med J. 2002;51(2):64-67. 2012. Int J Cancer. 2015;136(5):E359-E386. 52. Bunker CH, Patrick AL, Konety BR, et al. High prevalence of screening- 26. Ng M, Freeman MK, Fleming TD, et al. Smoking prevalence and ciga- detected prostate cancer among Afro-Caribbeans: the Tobago Prostate rette consumption in 187 countries, 1980-2012. JAMA. 2014;311(2):183-192. Cancer Survey. Cancer Epidemiol Biomarkers Prev. 2002;11(8):726-729. 27. King G, Polednak AP, Bendel R, et al. Cigarette smoking among na- 53. Consedine NS, Morgenstern AH, Kudadjie-Gyamfi E, et al. Prostate tive and foreign-born African Americans. Ann Epidemiol. 1999;9(4):236-244. cancer screening behavior in men from seven ethnic groups: the fear factor. 28. Taylor KL, Kerner JF, Gold KF, et al. Ever vs never smoking among Cancer Epidemiol Biomarkers Prev. 2006;15(2):228-237. an urban, multiethnic sample of Haitian-, Caribbean-, and U.S.-born blacks. 54. Gonzalez JR, Consedine NS, McKiernan JM, et al. Barriers to the Prev Med. 1997;26(6):855-865. initiation and maintenance of prostate specific antigen screening in Black 29. Siegel RL, Jacobs EJ, Newton CC, et al. Deaths Due to Cigarette American and Afro-Caribbean men. J Urol. 2008;180(6):2403-2408. Smoking for 12 Smoking-Related Cancers in the United States. JAMA 55. Etzioni R, Penson DF, Legler JM, et al. Overdiagnosis due to Intern Med. 2015;175(9): 1574-1576. prostate-specific antigen screening: lessons from U.S. prostate cancer 30. Argeseanu Cunningham S, Ruben JD, Narayan KM. Health of foreign- incidence trends. J Natl Cancer Inst. 2002;94(13):981-990. born people in the United States: A review. Health Place. 2008;14(4):623-635. 56. Baade PD, Youlden DR, Krnjacki LJ. International epidemiology of 31. Singh GK, Hiatt RA. Trends and disparities in socioeconomic and prostate cancer: geographical distribution and secular trends. Mol Nutr behavioural characteristics, life expectancy, and cause-specific mortality of Food Res. 2009;53(2):171-184. native-born and foreign-born populations in the United States, 1979-2003. 57. Glover FE Jr, Coffey DS, Douglas LL, et al. The epidemiology of Int J Epidemiol. 2006;35(4):903-919. prostate cancer in Jamaica. J Urol. 1998;159(6):1984-1987. 32. Kennedy S, McDonald JT, Biddle N. The healthy immigrant effect 58. Tsui J, Saraiya M, Thompson T, et al. Cervical cancer screen- and immigrant selection: Evidence from four countries. SEDAP Research ing among foreign-born women by birthplace and duration in the United Program. Social and Economic Dimensions of an Aging Population. States. J Womens Health (Larchmt). 2007;16(10):1447-1457. SEDAP-II – Canada in the 21st Century: Moving Towards an Older Society. 59. Ryerson AB, Eheman CR, Altekruse SF, et al. Annual report to the Research paper No. 164. January-December, 2006. http://socserv.mcmaster. nation on the status of cancer, 1975‐2012, featuring the increasing inci- ca/sedap/p/sedap164.pdf. Accessed July 19, 2016. dence of liver cancer. Cancer. 2016;122(9):1312-1237. 33. Aune D, Chan DS, Lau R, et al. Dietary fiber, whole grains, and risk 60. Denniston MM, Jiles RB, Drobeniuc J, et al. Chronic hepatitis C of colorectal cancer: systematic review and dose-response meta-analysis virus infection in the United States, National Health and Nutrition Examina- of prospective studies. BMJ. 2011;343:d6617. tion Survey 2003 to 2010. Ann Intern Med. 2014;160(5):293-300. 34. Greenberg MR, Schneider D, Northridge ME, et al. Region of 61. Schweitzer A, Horn J, Mikolajczyk RT, et al. Estimations of world- birth and black diets: the Harlem Household Survey. Am J Public Health. wide prevalence of chronic hepatitis B virus infection: a systematic review 1998;88(8):1199-1202. of data published between 1965 and 2013. Lancet. 2015;386(10003): 35. Odedina FT, Dagne G, LaRose-Pierre M, et al. Within-group differ- 1546-1555. ences between native-born and foreign-born Black men on prostate can- 62. Jemal A, Bray F, Center MM, et al. Global cancer statistics. CA Cancer cer risk reduction and early detection practices. J Immigr Minor Health. J Clin. 2011;61(2):69-90. 2011;13(6):996-1004. 63. United Nations Development Programme Human Development 36. American Cancer Society. Cancer Facts & Figures 2015. Atlanta, GA: Reports. International Human Development Indicators. http://hdr.undp. American Cancer Society; 2015. org/en/countries. Accessed July 19, 2016. 37. Borrell LN, Castor D, Conway FP, et al. Influence of nativity 64. World Health Organization. Global status report on noncommuni- status on breast cancer risk among US black women. J Urban Health. cable diseases 2014. Geneva: World Health Organization; 2014 2006;83(2):211-220. 65. Taioli E, Attong-Rogers A, Layne P, et al. Breast cancer survival in 38. Celi AC, Rich-Edwards JW, Richardson MK, et al. Immigration, women of African descent living in the US and in the Caribbean: effect of race/ethnicity, and social and economic factors as predictors of breast- place of birth. Breast Cancer Res Treat. 2010;122(2):515-520

October 2016, Vol. 23, No. 4 Cancer Control 357 66. Mutetwa B, Taioli E, Attong-Rogers A, et al. Prostate cancer characteristics and survival in males of African Ancestry according to place of birth: data from Brooklyn-New York, Guyana, Tobago and Trinidad. Prostate. 2010;70(10):1102-1109. 67. Ragin C, Mutetwa B, Attong-Rogers A, et al. Geographic and outcome variation among black men diagnosed with prostate cancer. Infect Agent Cancer. 2011;6(suppl 2):S2. 68. Rebbeck TR, Devesa SS, Chang BL, et al. Global patterns of prostate cancer incidence, aggressiveness, and mortality in men of african descent. Prostate Cancer. 2013;2013:560857. 69. Khalifa MM, Sharaf RR, Aziz RK. Helicobacter pylori: a poor man’s gut pathogen? Gut Pathog. 2010;2(1):2. 70. Plummer M, Franceschi S, Vignat J, et al. Global burden of gastric cancer attributable to Helicobacter pylori. Int J Cancer. 2015;136(2):487-490. 71. Wasley A, Kruszon-Moran D, Kuhnert W, et al. The prevalence of hepatitis B virus infection in the United States in the era of vaccination. J Infect Dis. 2010;202(2):192-201. 72. DeSantis CE, Siegel RL, Sauer AG, et al. Cancer statistics for African Americans, 2016: progress and opportunities in reducing racial disparities. CA Cancer J Clin. 2016;66(4):290-308. 73. Singh GK, Siahpush M, Hiatt RA, et al. Dramatic increases in obesity and overweight prevalence and body mass index among ethnic-immigrant and social class groups in the United States, 1976-2008. J Community Health. 2011;36(1):94-110. 74. Bennett GG, Wolin KY, Okechukwu CA, et al. Nativity and cigarette smoking among lower income blacks: results from the Healthy Directions Study. J Immigr Minor Health. 2008;10(4):305-311. 75. Ford K. Duration of residence in the United States and the fertility of U.S. immigrants. Int Migr Rev. 1990;24(1):34-68. 76. Porter P. “Westernizing” women’s risks? Breast cancer in lower- income countries. N Engl J Med. 2008;358(3):213-216. 77. TheWorld Bank. Birth rate, crude (per 1,000 people). The World Bank Group; 2013. http://data.worldbank.org/indicator/SP.DYN.CBRT.IN/ countries. Accessed July 20, 2016. 78. Mariotto AB, Noone AM, Howlader N, et al. Cancer survival: an overview of measures, uses, and interpretation. J Natl Cancer Inst Monographs. 2014;2014(49):145-186. 79. Blue L, Fenelon A. Explaining low mortality among US immigrants relative to native-born Americans: the role of smoking. Int J Epidemiol. 2011;40(3):786-793. 80. Turra CM, Elo IT. The Impact of Salmon Bias on the Hispanic Mor- tality Advantage: New Evidence from Social Security Data. Popul Res Policy Rev. 2008;27(5):515-530. 81. Phillips AA, Jacobson JS, Magai C, et al. Cancer incidence and mortality in the Caribbean. Cancer Invest. 2007;25(6):476-483.

358 Cancer Control October 2016, Vol. 23, No. 4 There is an urgent need to ensure that

existing genomic research considers

the unique needs of US Latinas with

breast cancer.

Chrysanthemum. X-Ray image courtesy of Bryan Whitney. www.x-rayphotography.com.

Genomic Disparities in Breast Cancer Among Latinas Filipa Lynce, MD, Kristi D. Graves, PhD, Lina Jandorf, MA, Charité Ricker, MS, Eida Castro, PsyD, Laura Moreno, Bianca Augusto, Laura Fejerman, PhD, and Susan T. Vadaparampil, PhD

Background: Breast cancer is the most common cancer diagnosed among Latinas in the United States and the leading cause of cancer-related death among this population. Latinas tend to be diagnosed at a later stage and have worse prognostic features than their non-Hispanic white counterparts. Genetic and genomic factors may contribute to observed breast cancer health disparities in Latinas. Methods: We provide a landscape of our current understanding and the existing gaps that need to be filled across the cancer prevention and control continuum. Results: We summarize available data on mutations in high and moderate penetrance genes for inherited risk of breast cancer and the associated literature on disparities in awareness of and uptake of genetic counseling and testing in Latina populations. We also discuss common genetic polymorphisms and risk of breast cancer in Latinas. In the treatment setting, we examine tumor genomics and pharmacogenomics in Latina patients with breast cancer. Conclusions: As the US population continues to diversify, extending genetic and genomic research into this underserved and understudied population is critical. By understanding the risk of breast cancer among ethnically diverse populations, we will be better positioned to make treatment advancements for earlier stages of cancer, identify more effective and ideally less toxic treatment regimens, and increase rates of survival.

Introduction South America. In the United States, 64.0% of Latinos An estimated 55 million individuals living in the Unit- are of Mexican background, 9.6% of Puerto Rican ed States identify as being Hispanic or Latino.1 Latinos background, 3.8% of Salvadoran background, 3.7% of are a culturally and genetically diverse group with ori- Cuban background, 3.2% of Dominican background, gins in Mexico, the Caribbean, Central America, and 2.4% of Guatemalan background, and the remainder

From Georgetown Lombardi Comprehensive Cancer Center (FL, KDG), Washington, District of Columbia, Department of Hematology and Oncology (FL), MedStar Georgetown University Hospital, Washington, District of Columbia, Department of Oncology (KDG), Georgetown University, Washington, District of Columbia, Department of Oncological Science (LJ), Icahn School of Medicine at Mount Sinai, New York, New York, Department of Medicine (CR), University of Southern California Norris Comprehensive Cancer Center, Los Angeles, California, Department of Psychiatry (EC), Ponce Health Sciences University, Ponce, Puerto Rico, Health Outcomes and Behavior Program, H. Lee Moffitt Cancer Center & Research Institute (LM, BA, STV), Tampa, Florida, and the Department of Medicine (LF), University of California, San Francisco, California. Submitted April 4, 2016; accepted September 8, 2016. Address correspondence to Susan T. Vadaparampil, PhD, Health Outcomes and Behavior Program, Moffitt Cancer Center, 12902 Magnolia Drive, MRC-CANCONT, Tampa, FL 33612. E-mail: [email protected] No significant relationships exist between the authors and the companies/organizations whose products or services may be referenced in this article. Funding for this work was supported in part by U54 CA163071 (Ponce Health Sciences University) and U54 CA163068 (Moffitt Cancer Center).

October 2016, Vol. 23, No. 4 Cancer Control 359 are of other origins.2 Although the terms Hispanic ed or unselected for family history or clinical charac- and Latino/Latina are often interchangeably used, we teristics (eg, affected vs unaffected, age at diagnosis), selected the term Latina for the current manuscript as cancer site (eg, breast, ovarian), and type of testing we feel it extends beyond spoken language to reflect (eg, inclusion of large rearrangement testing).22 In the both origin and cultural traditions of women from cohorts of unselected patients with breast cancer, the Latin America. BRCA mutation prevalence was 1.2% to 4.9%, which Breast cancer is the most common cancer diag- was consistent with expected rates.22 nosed among Latinas in the United States and is the BRCA mutations have also been documented in leading cause of cancer-related death in this popula- all residents of Latin American countries where these tion.3 Although the overall prevalence of breast can- genes have been studied, including Argentina, Bra- cer in Latinas is lower than in non-Latina whites, La- zil, Chile, Colombia, Costa Rica, Cuba, Mexico, Peru, tinas tend to be diagnosed at a later stage and have Puerto Rico, Uruguay, and Venezuela.23-54 Most stud- worse prognostic features (eg, triple-negative disease, ies have focused on the spectrum of BRCA muta- ERBB2 [formerly HER2 or HER2/neu]–positive dis- tions.22,55 In a review of BRCA1 and BRCA2 mutations ease).4 A myriad of socioeconomic and cultural fac- in persons living in Latin America and the Caribbean, tors contribute to health disparities in breast cancer 36% of the 33 studies primarily focused on Mexican among Latinas,5-7 but biological factors — particularly or Mexican American patients.22 Of the Mexican study genomics — remain an important but understudied population, the mutation prevalence was between consideration. 4.3% and 23.0%.22 For other Latina subethnic groups, the mutation prevalence estimates of each coun- High and Moderate Penetrance Genes try studied were: Colombia (1.2%–15.6%; 2 studies), Approximately 10% to 15% of breast cancer cases are Costa Rica (4.5%; 1 study), Cuba (2.6%; 1 study), Peru attributed to inherited gene mutations.8 Although (4.9%; 1 study), Uruguay (17%; 1 study), and Venezu- multiple genes confer an inherited risk for cancer,9 ela (17.2%; 1 study).22 These studies provide insight BRCA mutations are the most prevalent and penetrant into areas of future research of BRCA mutation dis- mutations, accounting for the majority of hereditary tribution and frequency based on country of origin, types of breast cancer.10 BRCA mutations result in an the role of specific founder mutations, the contribu- increased lifetime risk of breast cancer of up to ap- tion of large genomic rearrangements to the spectrum proximately 60% to 70% and a lifetime ovarian cancer of mutations across various Latina subethnic groups, risk of up to 40%.11-13 Among Latinas, breast cancer and the consideration of other non-BRCA genes that is often diagnosed at younger ages and with worse increase the risk of breast cancer. prognostic features, including increased rates of tri- Although recurrent mutations were identified ple-negative disease, than their non-Hispanic white within most studies, the specific mutation varied by counterparts.3,14-16 Triple-negative disease and pre- study and country.22 BRCA1 185delAG has also been menopausal breast cancer are both clinical character- documented in Latinas across Latin America and the istics associated with a higher probability of having a United States.40,42-44,47,56-58 One of the 3 Jewish found- BRCA1 or BRCA2 mutation.17,18 er mutations, BRCA1 185delAG is estimated to have arisen about 800 years ago or earlier and is believed Prevalence of BRCA to have been introduced into Latin America about The prevalence of BRCA mutations in the general 650 years ago.59 When this mutation is identified in US population is estimated to be 1 in 400, excluding Latinos, haplotype analysis supports that this muta- women of Ashkenazi Jewish descent in whom the prev- tion is of the same origin as the Jewish founder muta- alence is 1 in 40.19-21 However, less is known about the tion, rather than a separate genetic event.60,61 Pooled prevalence among racial and ethnic minority groups, mutation estimates performed by Porchia et al55 found including Latinas as a whole or by subethnicity that BRCA1 185delAG is the second most prevalent based on country of origin. A review examined the BRCA1 mutation and its frequency is not significantly spectrum of BRCA1 and BRCA2 mutations in Latin different between Mexico and other Latin American America and the Caribbean using studies published countries (P = .70). However, it is worth noting that between the years 1994 and 2015.22 Six of the 33 stud- not all Central and South American countries were ies were conducted among Latinas living in the Unit- represented in their analysis.55 ed States, with the vast majority of participants drawn The most common BRCA1 mutation in the same from clinic-based samples of patients of Mexican ori- meta-analysis was deletion of exons 9 to 12.55 This muta- gin with breast cancer residing in California, Arizona, tion is estimated to have originated nearly 1,500 years and Texas.22 Prevalence estimates of carrying a BRCA ago near Puebla Mexico.48,58 However, to date, it has mutation for this US Latina group ranged from 0.7% to been reported in Mexicans and Mexican Americans 42.0% and varied based on whether cases were select- alone.22,32,55,61,62 The contribution of large genomic re-

360 Cancer Control October 2016, Vol. 23, No. 4 arrangements to BRCA1 in Latin American patients mutations in these genes among Latinas with breast was evaluated in a study of US Latinas and described cancer (Table 1).26,27,54,72-80 the prevalence of rearrangements by racial and eth- One exception is the Brazilian founder mutation nic groups.63 Large rearrangements were significant- in TP53, R337H. Mutations in TP53 cause Li-Fraumeni ly more common in individuals who reported Latin syndrome, which is associated with an elevated risk American ancestry, and the prevalence of rearrange- for a wide spectrum of cancers, including adrenal cor- ments was two-fold higher than in the overall popu- tical carcinoma, soft-tissue and bone sarcomas, brain lation tested.63 This laboratory-based cohort extract- tumors, and breast cancer.81-83 The overall contribu- ed ethnicity data from genetic testing request forms; tion of TP53 mutations to breast cancer is estimated to therefore, no data about subethnicity was available.63 be less than 1%, unless selecting for early-onset breast However, the 2 most frequent BRCA1 rearrangements cancer.84,85 In studies of women diagnosed with breast identified in this study were deletion of exons 9 to 12 cancer at or before the age of 30 or 35 years, 5% to 8% and deletion of exons 1 and 2, likely reflecting the had TP53 mutations.81,86-88 underlying US Latino population in whom the major- TP53 R337H was first identified in individuals ity is of Mexican ancestry.63 In a Puerto Rican study, with childhood adrenal cortical carcinomas living BRCA1 deletion of exons 1 to 2 was seen in nearly in southern Brazil.89 This mutation occurs in 2.4% to 20% of study patients positive for BRCA1.49 Because 8.6% of Brazilian women with breast cancer.78,79,89,90 In Puerto Ricans represent the second largest US Latino a large study, which included 403 patients with breast group after those of Mexican ancestry, these findings cancer diagnosed at 45 years or younger, 12.1% car- support utilizing an assay that includes large rear- ried the TP53 R337H mutation. Although the muta- rangements when testing Latinos.49 The study results tion was significantly more frequent in younger pa- also highlight the importance of understanding more tients compared with those diagnosed at 55 years or granular aspects of ethnicity, such as country of ori- older (P < .001), 5.1% of the older group carried the gin, to ensure that all mutations that contribute sig- mutation.79 To date, no other populations have been nificantly are captured. identified in whom TP53 makes such a significant Dutil et al22 noted that most Latin American stud- contribution to breast cancer. The prevalence of this ies they reviewed identified a higher proportion of mutation in southern Brazil has been estimated to be BRCA1 than BRCA2 mutations, a finding similar to approximately 0.3%.91,92 Additional haplotype analy- reports in other populations. However, studies from ses support the hypothesis that this recurrent muta- 4 different countries (Costa Rica,37 Cuba,35 Puerto tion is a founder mutation from a shared ancestor.79,93 Rico,49 Uruguay64) reported more BRCA2 mutations Historically, the genetic assessment for hereditary than those in BRCA1. While these studies may have breast cancer involved the formation of a differen- been limited by sample size and the mutation-detec- tial diagnosis followed by a syndrome-by-syndrome tion strategies and technologies,35,37,49,64 this finding evaluation through the sequential testing of genes. has been also reported in a single US-based clinical However, the rapid integration of next-generation site and may warrant further exploration.65 sequencing has enabled simultaneous testing of mul- The meta-analysis performed by Porchia et al55 tiple inherited cancer genes, thereby expanding the identified recurrent BRCA2 mutations across all stud- use of multigene panels in clinical testing at a reduced ies with the following pooled prevalence: H372N cost.94 This expansion is reflected in the emerging (0.88%; 95% confidence interval [CI]: 0.24–1.92), body of literature on breast cancer focused on mul- E49X (0.38%; 95% CI: 0.13–0.75), and 3492insT (0.32%; tigene panel findings from the research, clinical, and 95% CI: 0.24–0.53). BRCA2 3492insT has been identified laboratory settings.95-103 in different regions of Spain with a frequency as high These literature cohorts are predominantly as 2.08%.65-71 Although it is possible that this mutation non-Hispanic whites, with Latinas representing was introduced in Latin America by the Spaniards, no less than 1.0% to 7.4% of study participants, thus haplotype studies of this specific mutation were identi- highlighting another area where future research is fied to confirm a shared ancestry rather than a separate needed.95-97,99-102 One study of 475 patients undergoing mutational event.49 multigene panel testing included 228 Latino patients BRCA1 and BRCA2 account for the major- (47.6% of the study population), and it reported that ity of hereditary breast cancer, but other high- and the likelihood of detecting a deleterious mutation was moderate-risk genes also predispose individuals to no different among the ethnic and racial groups rep- breast cancer, including TP53, PTEN, CDH1, STK11, resented.96 Of the patients with breast cancer (n = 197), CHEK2, PALB2, and ATM, among others.26,27,54,72-80 14.8% (n = 28) carried mutations, and, as expected, Limited studies have been performed of non-BRCA BRCA1 and BRCA2 were the most commonly mutated genes in Latina breast-cancer cohorts, leaving much genes; however, 16 mutations were identified in other to be learned about the prevalence and spectrum of genes (CDH1 = 4, CHEK2 = 3, MUTYH = 3, PALB2 = 2,

October 2016, Vol. 23, No. 4 Cancer Control 361 TP53 = 1, RAD50 = 1, RAD51D = 1, BARD1 = 1).96 Of als who receive genetic counseling and subsequently note, the likelihood of identifying more than 1 variant elect to undergo testing. However, growing evidence of uncertain significance in Latinos was significantly identifies disparities in awareness of and access to higher than that of non-Hispanics whites.104 Thus, a genetic counseling among Latinas compared with need exists for further research to better classify rare non-Hispanic white women. Data from health inter- variants, especially given the under-representation of view surveys from 2000, 2005, and 2010 show that La- Latinos in laboratory and research databases. tinas had the lowest level of awareness about genetic testing for inherited cancer risk than all of the other Genetic Counseling and Testing US racial ethnic groups.105-107 Using telephone surveys, Patient- and Health Care–Related Factors Gammon et al108 studied 63 Latinas and 84 non-His- An important step toward understanding the role of panic whites at increased risk for carrying a BRCA1 BRCA and other high- and moderate-risk breast cancer or BRCA2 mutation, examining their awareness, cog- genes in Latinas is to increase the number of individu- nition level, and psychosocial needs related to genetic

Table 1. — Select Non-BRCA Genes Observed in Latina Populations Study Country Cohort Gene Analysis Findings No. of Inclusion Patients Criteria Assumpção78 Brazil 123 Family history of TP53 Site-specific analysis of 2.4% of cases carried the breast cancer TP53 R337H mutation (P = .0442) Family history of ovarian cancer Sporadic breast cancer Carraro54 54 Early-onset breast cancer BRCA1 Coding introns/exons of 22% carried mutations diagnosis < 30 y BRCA2 BRCA1/2, TP53, and mostly in BRCA1/2 site-specific analysis of CHEK2 2% (n = 1) had TP53 CHEK2 (c.1100delC) mutation TP53 Felix26 106 HBOC testing BRCA1/2 PCR of each exon of BRCA1 2.8% carried mutations CHEK2 Site-specific analyses of (BRCA = 2, TP53 = 1) TP53 BRCA2 (c.5946_5946delT; c.156_157insAlu), CHEK2 (c.1100delC; c.444+1G>A; p.I157T), and TP53 (p.R337H) Giacomazzi79 874 Family history of cancer TP53 Single-site analysis for p.R337H identified in (group 1) p.R337H 3.4% (group 1) and Consecutive breast cancer 8.6% (group 2) (group 2) Higher prevalence when diagnosed ≤ 45 y (12.1%) vs 55 y (5.1%; P < .001) Silva27 120 HBOC testing ATM Coding introns/exons of 26% (n = 31) mutations BRCA1/2 BRCA1/2 (BRCA1/2 = 27, CHEK2 = 1, TP53 = 3) BRIP1 Site-specific analysis of CHEK2 (c.1100delC) and CDH1 TP53 (p. R337H) CDKN2A Array comparative genomic CTNNB1– hybridization for CNVs in CHEK2 other 14 genes MLH1 MSH6 NBN PALB2 PTEN RAD50 RAD51 TP53

Continued on next page

362 Cancer Control October 2016, Vol. 23, No. 4 Table 1. — Select Non-BRCA Genes Observed in Latina Populations, continued Study Country Cohort Gene Analysis Findings No. of Inclusion Patients Criteria González- Chile 137 ≥ 2 family members with ATM PCR-based analysis of 5557G>A, IVS38-8T>C, Hormazábal72 (BRCA– = 126, breast cancer coding sequence and exon/ IVS24-9delT associated BRCA+ = 11) intron boundaries of ATM with elevated risk of breast ≥ 2 family members with – ovarian cancer Analysis of ATM 5557G>A, cancer if BRCA Family history of male IVS38-8T>C, IVS24-9delT Identification of composite breast cancer genotype that confers 3.19-fold risk for breast Early-onset breast cancer cancer with no family history Jara80 143 ≥ 2 family members with RAD51D PCR-based analysis of No mutations detected in (BRCA– = 131, breast cancer coding sequence and RAD51D + BRCA = 12) ≥ 2 family members with exon-intron boundaries of c.135G>C associated with ovarian cancer RAD51D elevated breast cancer risk – Family history of male Analysis of RAD51D, if BRCA breast cancer c.135G>C Early-onset breast cancer with no family history Leyton73 436 BRCA+ PALB2 Full gene sequencing in No pathogenic mutations ≥ 2 family members with 100 “high-risk” cases identified breast cancer Analysis of identified variants 3 variants identified a a ≥ 2 family members with (c.1676A>G , c.2993C>T , ovarian cancer c.1861C>A) Single case of early-onset diagnosed ≤ 50 y Calderón- Mexico 94 Familial breast cancer ATM PCR-FLP of 3 specific 5557G>A (13%) 74 Zúñiga Early-onset breast cancer mutations (IVS24-9delT, IVS24-9delT (21% vs IVS38-8T>C, 5557G>A) 8% controls; P = .0122) IVS38-8T>C (< 1%) Bell76 United 362b Early-onset breast cancer CHEK2 169 cases diagnosed ≤ 40 y Data not reported by States had sequencing of coding ethnicity, but reported region of CHEK2 “infrequency” of Specific analysis of c.1100delC among Latinas 1100delC, H143Y, and 8 other CHEK2 variants/mutations Bretsky75 101b Personal history of breast ATM 20 specific ATM missense L546V had modest but cancer mutations or polymorphisms not significant predictor of risk; almost exclusive to African American women (found in 2 Latinas) Damiola77 158b Breast cancer diagnosed MRE11 PCR-based analysis of Data not reported by ≤ 45 y RAD50 coding sequence and exon/ ethnicity intron boundaries of MRE11, NBN MRE11, RAD50, NBN are RAD50, NBN intermediate-risk genes aThese variants play a role in risk of breast cancer. bLatinas were part of a larger multiethnic cohort. CNV = copy number variation, FLP = fragment length polymorphism, HBOC = hereditary breast and ovarian carcinoma, PCR = polymerase chain reaction. counseling and testing. Among those who had not pre- what genetic testing was and had not received physi- viously undergone genetic counseling (53 of the 120), cian referral for genetic testing.109 Findings did differ Latinas were more unaware than their white coun- based on country of origin — an important area to con- terparts of the availability of testing (56.9% vs 34.8%, sider in future work, given the diversity of Hispanic respectively).108 Vadaparampil et al109 reported on a populations across the United States.109 In another re- sample of Latinas with a personal or family history of port, Kaplan et al110 reported differences in awareness breast cancer, all of whom reported an awareness of of genetic testing by race and ethnicity, such that 19.4% genetic risk for breast cancer (ie, family history). How- of Latinas had heard of genetic testing compared with ever, none of the Latinas had a clear understanding of 59.4% of whites, 26.1% of Asian Americans, and 31.0%

October 2016, Vol. 23, No. 4 Cancer Control 363 of black women. ing. Overall, 52% completed genetic testing, and no dif- In a study of more than 2,400 patients complet- ferences by race and ethnicity were observed.117 Wood- ing a family cancer history form, Mays et al111 found son et al118 reported on the utilization of group pretest that, overall, despite low levels of initial awareness, genetic counseling in a community clinic made up of 65 patients (2.7%) met criteria for cancer risk assess- mostly Latinas (62.3%) with breast cancer; the majority ment; of those, 72.3% expressed interest in receiving (86.7%) underwent BRCA genetic testing when offered. genetic counseling. Furthermore, no differences in in- Overall, these studies have focused on the deliv- terest in genetic services were reported across all racial ery of genetics services for cancer to majority Latina co- and ethnic groups.111 Among 1,536 women with non- horts, demonstrating that genetic counseling and test- metastatic breast cancer, Jagsi et al112 found that Latinas ing is likely well-received by Latinas with breast cancer; had a greater desire for genetic counseling than other however, these studies were all conducted in safety-net groups (58.8% of Spanish-speaking Latinas; 36.7% of hospitals or in community, low-resource settings and English-speaking Latinas; 27.1% of non-Latina whites; were aimed at the provision of service to low-income, and 28.1% of blacks). In addition, Lagos et al113 exam- uninsured patients.115-118 Although Latinas continue to ined social, cognitive, and cultural variables among La- be disproportionally uninsured or underinsured, these tinas prior to an appointment for genetic counseling. study findings might not generalize to other health care Fifty low-income, underserved Latinas completed the settings. Thus, further studies are needed across various assessment, and the results demonstrated their readi- clinical settings and in a wider representation of Latinas ness (having the necessary skills for the genetic-coun- with breast cancer to better understand the utilization of seling process), low fatalism, and high rate of self-ef- genetic testing as well as the barriers for referrals. ficacy, and social support.113 However, this study was conducted in women who showed up to their genetic Common Genetic Polymorphisms and counseling appointments, thus representing a unique Risk of Breast Cancer group of women.113 Vadaparampil et al114 studied a Genome-Wide Association Studies group of Puerto Rican women (living in Puerto Rico Progress in the discovery of germline genetic polymor- or central Florida) with a family or personal history of phisms associated with breast cancer risk changed pace breast cancer and found that the vast majority of partic- when technological advances in genotyping made it ipants said they would undergo genetic testing within possible to characterize genome-wide genetic variation the next 6 months if it was available. Barriers included at a relatively low cost.119 In 2007, the first breast cancer the potential physical pain associated with the test. genome-wide association studies were published, and they reported a handful of single nucleotide polymor- Uptake of Services phisms (SNPs) associated with a modest increase in Given lower levels of patient awareness, physician risk.120-122 Since then, more than 100 common variants recommendations may provide a critical approach to that either increase risk for or are protective against increasing the utilization of genetic counseling and developing breast cancer have been discovered and, testing for hereditary risk of breast cancer. However, including replication efforts, data from more than available studies suggest a missed clinical opportu- 120,000 women have been analyzed.120-143 A small pro- nity, because both English- and Spanish-speaking portion of samples included in these major initiatives Latina survivors of breast cancer may be more likely are from minority populations in the United States to have unmet needs for discussion with a health care (eg, Latinas, African Americans),120-143 and the first re- professional about genetic testing for cancer than sults of genome-wide association studies of Latinas their non-Hispanic white counterparts. For example, with breast cancer were published in 2014.144 This lat- Jagsi et al112 reports that minority patients were the ter study represents important but limited progress, most likely to express an unmet need for a discussion considering that the sample size was one-tenth of that about genetic testing, although they also showed a available for genome-wide association studies involv- strong desire for such testing. ing women of European origin.120-122,144 Preliminary studies support the uptake of genetic Until the first genome-wide association studies of counseling when services are offered.115-118 One study breast cancer in women of European origin were pub- of predominantly Latina patients (71.4%) offered genet- lished, the search for risk-predisposing genetic variants ic counseling at a safety-net hospital found that 88.0% was focused on finding polymorphisms within genes kept their appointments.115 Another study of women that, for known or hypothesized involvement in the bi- (69.6% were Latinas) seen in a safety-net hospital set- ology of the disease, were likely to contribute to breast ting reported that 96.4% of them underwent BRCA test- cancer risk.120-122 These studies in US Latinas or Latin ing when it was recommended to them.116 Once Latinas American women typically consisted of the replication were referred, Olaya et al117 found that they are equally of previously associated polymorphisms reported in likely as the general population to complete BRCA test- Europeans, with few of these studies looking for varia-

364 Cancer Control October 2016, Vol. 23, No. 4 tion in samples of Latinas before further testing specific MAP3K1, and R AD51L1) and breast cancer risk.192 polymorphisms for associations in larger samples.145,146 They replicated associations for the polymorphisms Compared with the hundreds of genome-wide as- in RELN, FGFR2, TOX3, and TLR1 and 2q35 and found sociation studies in non-Hispanic white women, we heterogeneity by ancestry for the RELN, 2q35, and identified 13 case-control studies or cohorts that in- TLR1 SNPs.192 A follow-up study reported that the het- clude US Latina or Latin American women.147-160 These erogeneity by ancestry for the 2q35 polymorphism was studies include populations of women of no more than likely due to the association between genetic ancestry, 100 and up to approximately 5,000 women of Latin use of hormone therapy, and breastfeeding.191 Anoth- American origin; combined, the study populations tal- er analysis of the FGFR2 polymorphism in the Mexi- ly approximately 5,000 Latina women with breast can- can study reported an interaction between the FGFR2 cer and 11,000 Latina healthy controls.147-160 polymorphism and alcohol intake.193 The first genome- wide association study of breast cancer in US Latinas Candidate Gene or Pathway Studies also replicated previous associations, with most of the Multiple breast cancer–association studies of candidate SNPs being concordant in terms of direction and mag- genes or pathways have been reported for US Latina nitude of association with those reported in European and Latin American women during the last 20 years. or Asian populations.144 Twenty-three of the 83 vari- Genes or pathways studied have included those re- ants tested had probability values below .05.144 lated to hormone metabolism, hormone receptors, hormone coactivators or supressors,145,161-166 growth Ancestry factors,146,167-173 matrix metalloproteinases,174,175 inflam- Admixture mapping leverages the demographical his- mation and energy balance,149,176-181 metabolism of xe- tory of admixed populations to find genomic regions nobiotic compounds and oxidative stress,150,182-184 DNA that may carry trait-associated variants.196-204 An ad- repair,151,152,154,186 and angiogenesis.187 Results reported mixed population results from the combination of 2 or in these publications should be interpreted with cau- more ancestral groups.200 The principle of admixture tion, given that approximately 60% of the candidate mapping is to identify genomic regions in which cases gene analyses included did not adjust for genetic an- share more of the same genetic ancestry than either cestry, which is a known confounder in genetic associ- population-based controls (case-control analysis) or ation studies in admixed populations.188,189 In addition, compared with the average ancestry of the rest of the no associations in candidate gene or pathways studies, genome among cases (case-only analysis).202 This ap- nor any of the interactions with risk factors, genetic an- proach has identified risk variants or risk regions for cestry, or tumor characteristics, have been replicated in multiple complex traits, including obesity, hyperten- independent samples of Latinas. sion, and cancer.196-199,201,203,204 The incidence of breast cancer varies across different racial and ethnic groups Replication of Identified Single Nucleotide in the United States, and Latinas have lower incidence Polymorphisms rates than non-Latina whites but higher rates than Few studies have included Latinas and tested the as- American Indian women.205 Genetic ancestry has also sociation between SNPs discovered in genome-wide been associated with breast cancer risk in US Latinas association studies of breast cancer conducted in sam- and Mexican women after adjusting for nongenetic risk ples of European or Asian women.190-195 The first study factors, suggesting that a genetic component could be genotyped previously reported SNPs in the 2q35 region responsible for the difference in risk.149,206,207 An ad- and FGFR2, TOX3, and MAP3K1, reporting statisti- mixture mapping study in Latinas reported a statistically significant replications for the polymorphisms cally significant association between a region in the in FGFR2 and 2q35.190 Two different studies published long arm of chromosome 6 (6q25) near ESR1 and risk the results of analyses conducted in the same sample of of breast cancer and a suggestive association on chro- high-risk families from Chile and healthy controls, test- mosome 11.208 Higher Indigenous American ancestry ing associations between previously reported variants at chromosome 6q25 was associated with lower risk of in FGFR2, MAP3K1, and TOX3 and the 2q35 and 8q24 breast cancer.208 This finding was concordant with the regions and breast cancer risk.194,195 They replicated the previous reports of lower rates of risk of breast cancer associations for FGFR2, MAP3K1, TOX3, and 2q35 but among Latinas with high American Indian ancestry not for 8q24.194,195 An analysis conducted in a pooled compared with women with high European ancestry sample of Latina cases and controls from the Four- after adjusting for possible risk factors such as socio- Corners study, San Francisco Bay Area Breast Cancer economic status, number of full-term pregnancies, and Study, and a study in Mexico, investigated the associa- breast feeding.206-208 tion between 10 identified polymorphisms in genome- One included a discovery phase and replication wide association studies (in region 2q35 and in or near in 3 additional studies.144 The study reported ge- RELN, MRPS30, RNF146, FGFR2, TOX3, LSP1, TLR1, nome-wide results that were statistically significant

October 2016, Vol. 23, No. 4 Cancer Control 365 for 2 linked SNPs 56kb upstream of ESR1 (rs140068132 express human epidermal growth factor receptor 2 and rs147157845).144 These SNPs have a frequency (ERBB2; formerly HER2 or HER2/neu) protein.216 High of between 5% and 23% in Latin American popula- levels of ERBB2 expression identify those women who tions and are absent in most all other groups.144 The benefit from treatment with ERBB2-targeted agents, minor allele was protective, with an associated odds which have been shown to increase survival in the ad- ratio (OR) of 0.60 (95% CI: 0.53–0.67) per allele and juvant and metastatic settings.217,218 Most studies with was more protective for ER-negative disease than for ERBB2-targeted therapies have enrolled majority ER-positive disease (OR for ER-negative disease 0.34; populations of non-Hispanic whites, although consis- 95% CI: 0.21–0.54).144 tent evidence demonstrates that a higher proportion of ERBB2-positive tumors exist among Latinas, even af- Treatment ter adjusting for other tumor characteristics (eg, grade, Tumor Genomics stage, ER status) and breast cancer risk factors (eg, num- A growing body of evidence suggests differences in ber of children, alcohol consumption).209 the tumor biology of breast carcinoma across vari- Further tumor characterization has been made ous races and ethnicities. Several studies have evalu- possible due to advances in molecular tumor profil- ated the prevalence of phenotypic subtypes of breast ing. Oncotype DX (Genomic Health, Redwood City, cancer in Latinas compared with other population CA) is a 21-gene breast cancer assay — known as a groups.209-215 Most data have shown a higher propor- recurrence score — that provides prognostic and pre- tion of HR-negative disease types among Latinas when dictive information regarding the benefits of adjuvant compared with non-Hispanic whites (Table 2).4,209-213 chemotherapy in patients with ER-positive tumors. However, those results have not always been concor- Use of Oncotype DX is part of several guidelines from dant, and the differences seen across these studies professional medical organizations, including the Na- could represent small sample sizes, patient age, or un- tional Comprehensive Cancer Network, the American adjusted rates for genetic ancestry. Although studies Society of Clinical Oncology, and the European Society based on data from the California Cancer Registry indi- for Medical Oncology.219-221 The characteristics of this cated a higher proportion of triple-negative tumors,214 assay and the impact of its results on treatment deci- this finding was not confirmed in a Colorado study.209 sions among Latinas with breast cancer are lacking in In a retrospective study performed in Brazil, patients the medical literature. Kalinsky et al222 studied 74 La- in the southern regions with a higher percentage of Eu- tinas and 145 non-Hispanic white women matched for ropean ancestry and higher socioeconomic status pre- age, disease stage, and nodal status, and they observed sented with the highest proportion of luminal tumors, no differences in the overall recurrence score, ER whereas the more aggressive subtypes were seen in or PR status, or ERBB2 expression by Oncotype DX. the northern parts of Brazil, an area with a higher Afri- However, Latinas had a higher expression of CCNB1 can ancestral influence.215 and AURKA, 2 genes that are part of the proliferation Approximately 15% of breast cancer types over- score and heavily weighted in the calculation of the

Table 2. — Comparison of ER and ERBB2 Status in Latina, Black, and White Women Study No. of Patients Latina White Black Chlebowski211 N = 3,800 ER+: 83.0% ER+: 87.0% ER+: 71.0% n = 103 Latinas Dunnwald212 N = 209,276 ER+: 70.2% ER+: 78.4% ER+: 60.5% n = 5,585 Latinas Hausauer213 N = 243,906 ER+: 53.2% ER+: 63.5% ER+: 46.4% n = 15,355 Latinas (unknown ER status: 29.4%) (unknown ER status: 22.3%) (unknown ER status: 29.4%) Hines209 N = 285 ER+: 63.8% ER+: 77.3% — n = 69 Latinas ERBB2 +: 31.9% ERBB2 +: 14.3% TNBC: 17.4% TNBC: 15.1% Li210 N = 124,934 ER+: 68.7% ER+: 78% ER+: 53.4% n = 7,219 Latinas Parise4 N = 143,184 ER+: 74.6 % ER+: 82.7% ER+: 66.5% n = 24,078 Latinas ERBB2 +: 22.2% ERBB2 +: 17.2% ERBB2 +: 20.1% TNBC: 15.9% TNBC: 11.2% TNBC: 24.5% TNBC = triple-negative breast cancer.

366 Cancer Control October 2016, Vol. 23, No. 4 recurrence score. Latinas).241 In general, the researchers found that, com- Multiple trials whose study populations were pared with non-Hispanic whites, minorities had fewer mostly comprised of non-Hispanic white women have associated toxicities and no definitive survival benefit shown that use of Oncotype DX affects treatment rec- with aromatase inhibitors.241 However, these results ommendations and leads to an increase in physician should be cautiously interpreted, because the minority and patient confidence in treatment decisions.223-226 A participants were less adherent to hormonal therapy small study of 96 patients with breast cancer treated in and the study was not powered to detect survival ben- Mexico showed that use of the Oncotype DX changed efit in the subgroups.241 treatment decisions for 32% of patients, a finding suggesting that its use has a meaningful impact on rec- Conclusions ommendations for adjuvant treatment.227 Results from Several genetic and genomic factors are related to the cost-effectiveness analyses indicated that use of the health disparities of breast cancer in Latinas. Increas- Oncotype DX assay was projected to improve rates of ing our knowledge about the contribution of high- and life expectancy when compared with the current stan- moderate-penetrance mutations to the risk of breast dard of care.228 cancer among Latinas overall and for subgroups based on country of origin is an important priority. Although Pharmacogenomics genetic counseling and testing for inherited suscepti- Several factors cause variations in an individual’s re- bility to breast cancer has been clinically available for sponse to drugs, including age, body mass index, nearly 20 years, disparities in awareness, referral to diet, and genetic variation.229-231 SNPs in genes related services, and access persist. Therefore, interventions to to drug-metabolizing enzymes have been recognized address barriers related to low levels of awareness and as important determinants of variability to drug re- lacking physician referrals are critical.105-110 sponse.232 Most studies have not been sufficiently Even though multiple candidate gene studies powered to determine whether specific chemotherapy have been conducted in this population, only identi- agents used to treat breast cancer have different rates fied variants in genome-wide association studies have of effectiveness and toxicities based on race or ethnic- been systematically replicated. Much larger sample ity.233 However, differences in the metabolism of endo- sizes will be required if we expect to discover similar crine therapies have been well documented according results in Latinas as would be identified for the Euro- to race and ethnicity.234-237 pean genome in view of their Indigenous American In a study that evaluated clinical data and blood component. We cannot assume that overlapping vari- samples from patients with breast cancer undergoing ants alone between different ancestral genomes will adjuvant tamoxifen therapy, mostly non-Hispanic white be associated with rate of risk, so our efforts should be women (68%) and Latinas (26%) had significantly higher focused on reducing research disparities by expand- serum levels of tamoxifen and 4-hydroxytamoxifen, ing available resources to include large cohorts and one of the tamoxifen metabolites (P = .02 and P = .007, case-control studies of diverse populations in and out- respectively).234 In 2 other studies, genetic polymorphisms side the United States. in CYP2D6 associated with lower plasma concentrations Latinas remain systematically underrepresented of the active metabolites of tamoxifen were described in pharmacogenomics studies and the current studies in Mexican, Puerto Rican, and Spanish patients.235,236 were not powered to detect outcome differences. They A higher prevalence of this poor metabolizer pheno- are also underrepresented in clinical treatment trials type has also been observed in non-Hispanic whites.237 and other patient-reported outcomes research. Future In an attempt to clarify whether CYP2D6 allele status research should draw from the few models of success influences outcomes from tamoxifen, investigators as- in prior studies that have recruited sufficient numbers sessed data from 2 large prospective trials and found of Latinas.242-245 Lessons can also be learned from suc- that CYP2D6 allele status did not predict clinical ben- cessful examples of recruiting other racial and ethnic efit of adjuvant tamoxifen in terms of risk of recur- minority patients who have survived breast cancer.246 rence.238,239 Therefore, changes in treatment decisions Elements that appear to bolster success include part- based on CYP2D6 allele status alone are not recom- nering with community-based organizations that pro- mended. Differences in the incidence of polymor- vide services to Latinas and the provision of language- phisms of the aromatase gene among different ethnic concurrent clinical care.247-249 The heterogeneity of the groups have also been reported and could potentially US Latina population must also be considered, because lead to different outcomes and toxicities among popu- different cultural influences, levels of awareness of, lations.240 One trial evaluated the benefit of extended and interest in genetic and genomic services appear to hormonal therapy with an aromatase inhibitor after vary by country of origin.109 5 years of tamoxifen treatment in non-Hispanic whites An urgent need exists to ensure that existing ge- (n = 4,708) and minority women (n = 352; 1.5% were nomic research considers the unique needs of this

October 2016, Vol. 23, No. 4 Cancer Control 367 BRCA1 and BRCA2 alleles in Latin America and the Caribbean: a clinical Latina population. As the US population continues to perspective. Breast Cancer Res Treat. 2015;154(3):441-453. diversify with up to one-third identifying as Hispan- 23. Gomes MC, Costa MM, Borojevic R, et al. Prevalence of BRCA1 1 and BRCA2 mutations in breast cancer patients from Brazil. Breast Cancer ics by 2060, extending genetic and genomic research Res Treat. 2006;103(3):349-353. into this underserved and understudied population 24. Esteves V, Thuler L, Amêndola L, et al. Prevalence of BRCA1 and BRCA2 gene mutations in families with medium and high risk of breast and will be critical. By understanding the risk of breast can- ovarian cancer in Brazil. Brazl J Med Biol Res. 2009;42(5):453-457. cer among diverse populations, we will be better posi- 25. Ewald IP, Izetti P, Vargas FR, et al. Prevalence of the BRCA1 founder mutation c. 5266dupin Brazilian individuals at-risk for the hereditary tioned to make advancements in the number of wom- breast and ovarian cancer syndrome. Hered Cancer Clin Pract. 2011;9:12. en diagnosed at earlier stages, identify more effective 26. Felix GE, Abe-Sandes C, Machado-Lopes TM, et al. Germline mutations in BRCA1, BRCA2, CHEK2 and TP53 in patients at high-risk for HBOC: and less toxic treatment regimens, and increase rates of characterizing a Northeast Brazilian population. Human Genome Variation. survival. Meeting these goals will contribute to reduc- 2014;1:14012. 27. Silva FC, Lisboa BC, Figueiredo MC, et al. Hereditary breast and ing the current health disparities in these patients with ovarian cancer: assessment of point mutations and copy number variations breast cancer. in Brazilian patients. BMC Medi Genet. 2014;15(1):1. 28. Gallardo M, Silva A, Rubio L, et al. Incidence of BRCA1 and BRCA2 Acknowledgment: The authors wish to thank mutations in 54 Chilean families with breast/ovarian cancer, genotype– Lindsay Salem for her assistance in the literature re- phenotype correlations. Breast Cancer Res Treat. 2006;95(1):81-87. 29. Sanchez A, Faundez P, Carvallo P. Genomic rearrangements of the view for this manuscript. BRCA1 gene in Chilean breast cancer families: an MLPA analysis. Breast Cancer Res Treat. 2011;128(3):845-853. References 30. Jara L, Ampuero S, Santibáñez E, et al. BRCA1 and BRCA2 mutations in a South American population. Cancer Genet Cytogenet. 1. Colby SL, Ortman JM. Projections of the size and composition of the 2006;166(1):36-45. US population: 2014 to 2060. Curr Pop Rep. 2014:1-13 https://www.census.gov/ 31. Gonzalez-Hormazabal P, Gutierrez-Enriquez S, Gaete D, et al. content/dam/Census/library/publications/2015/demo/p25-1143.pdf. Accessed Spectrum of BRCA1/2 point mutations and genomic rearrangements in September 14, 2016. high-risk breast/ovarian cancer Chilean families. Breast Cancer Res Treat. 2. Pew Research Center. Statistical portrait of hispanics in the United 2011;126(3):705-716. States, 2014. http://www.pewhispanic.org/files/2016/04/Statistical-Portrait-of- 32. Torres D, Rashid MU, Seidel-Renkert A, et al. Absence of the Hispanics-in-the-United-States-2014-final.pdf. Accessed September 14, 2016. BRCA1 del (exons 9–12) mutation in breast/ovarian cancer families outside 3. American Cancer Society. Cancer Facts & Figures for Hispanics/ of Mexican Hispanics. Breast Cancer Res Treat. 2009;117(3):679-681. Latinos 2015-2017. http://www.cancer.org/acs/groups/content/@research/ 33. Torres D, Rashid MU, Gil F, et al. High proportion of BRCA1/2 documents/document/acspc-046405.pdf. Accessed September 14, 2016. founder mutations in Hispanic breast/ovarian cancer families from Colombia. 4. Parise CA, Caggiano V. The Influence of socioeconomic status on Breast Cancer Res Treat. 2006;103(2):225-232. racial/ethnic disparities among the ER/PR/HER2 breast cancer subtypes. 34. Sanabria MC, Muñoz G, Vargas CI. Mutations in the BRCA1 gene J Cancer Epidemiol. 2015;2015:813456. (185delAG and 5382insC) are not present in any of the 30 breast cancer patients 5. Quinn GP, McIntyre JQ, Vadaparampil ST. Challenges in recruit- analyzed from eastern Colombia [in Spanish]. Biomedica. 2009;29(1):61-72. ing Mexican women for cancer genetics research. J Community Genet. 35. Rodriguez RC, Esperon AA, Ropero R, et al. Prevalence of BRCA1 2011;2(1):43 - 47. and BRCA2 mutations in breast cancer patients from Cuba. Familial Cancer. 6. Giuliano AR, Mokuau N, Hughes C, et al. Participation of minorities 2008;7(3):275-279. in cancer research: the influence of structural, cultural, and linguistic factors. 36. Hernandez J, Llacuachaqui M, Palacio GV, et al. Prevalence of Ann Epidemiol. 2000;10(8 suppl):S22-S34. BRCA1 and BRCA2 mutations in unselected breast cancer patients from 7. Lantz PM, Mujahid M, Schwartz K, et al. The influence of race, medellín, Colombia. Hered Cancer Clin Pract. 2014;12(11):11. ethnicity, and individual socioeconomic factors on breast cancer stage at 37. Gutiérrez Espeleta G, Llacuachaqui M, García‐Jiménez L, et al. BRCA1 diagnosis. Am J Public Health. 2006;96(12):2173-2178. and BRCA2 mutations among familial breast cancer patients from Costa Rica. 8. Ellsworth RE, Decewicz DJ, Shriver CD, et al. Breast cancer in the Clin Genet. 2012;82(5):484-488. personal genomics era. Curr Genomics. 2010;11(3):146-161. 38. García-Jiménez L, Gutiérrez-Espeleta G, Narod SA. Descriptive 9. Lindor NM, McMaster ML, Lindor CJ, et al. Concise handbook of epidemiology and molecular genetics of hereditary breast cancer in Costa familial cancer susceptibility syndromes - second edition. J Natl Cancer Rica. Revis Biol Trop. 2012;60(4):1663-1668. Inst Monogr. 2008(38):1-93. 39. Rodríguez AO, Llacuachaqui M, Pardo GG, et al. BRCA1 and BRCA2 10. Euhus DM, Robinson L. Genetic predisposition syndromes and mutations among ovarian cancer patients from Colombia. Gynecol Oncol. their management. Surg Clin North Am. 2013;93(2):341-362. 2012;124(2):236-243. 11. Chen S, Parmigiani G. Meta-analysis of BRCA1 and BRCA2 penetrance. 40. Ruiz‐Flores P, Sinilnikova OM, Badzioch M, et al. BRCA1 and BRCA2 J Clin Oncol. 2007;25(11):1329-1333. mutation analysis of early‐onset and familial breast cancer cases in Mexico. 12. Antoniou A, Pharoah P, Narod S, et al. Average risks of breast and Hum Mutat. 2002;20(6):474-475. ovarian cancer associated with BRCA1 or BRCA2 mutations detected in 41. Calderón-Garcidueñas AL, Ruiz-Flores P, Cerda-Flores RM, et al. Clinical case series unselected for family history: a combined analysis of 22 studies. follow up of Mexican women with early onset of breast cancer and mutations in Am J Hum Genet. 2003;72(5):1117-1130. the BRCA1 and BRCA2 genes. Salud Publica Mex. 2005;47(2):110-115. 13. Litton JK, Ready K, Chen H, et al. Earlier age of onset of BRCA mutation- 42. Vaca-Paniagua F, Alvarez-Gomez RM, Fragoso-Ontiveros V, et al. related cancers in subsequent generations. Cancer. 2012;118(2):321-325. Full-exon pyrosequencing screening of BRCA germline mutations in Mexican 14. Miranda PY, Wilkinson AV, Etzel CJ, et al. Policy implications of early women with inherited breast and ovarian cancer. PloS One. 2012;7(5):e37432. onset breast cancer among Mexican‐origin women. Cancer. 2011;117(2):390-397. 43. Vidal-Millan S, Taja-Chayeb L, Gutierrez-Hernandez O, et al. Mutational 15. Boyle T, McPadden E. Breast cancer presents at an earlier age in Mexican analysis of BRCA1 and BRCA2 genes in Mexican breast cancer patients. Eur J American women. Breast J. 2004;10(5):462-464. Gynaecol Oncol. 2009;30(5):527-530. 16. Lara‐Medina F, Pérez‐Sánchez V, Saavedra‐Pérez D, et al. Triple‐ 44. Villarreal-Garza C, Weitzel J, Llacuachaqui M, et al. The prevalence negative breast cancer in Hispanic patients. Cancer. 2011;117(16):3658-3669. of BRCA1 and BRCA2 mutations among young Mexican women with triple- 17. Claus EB, Schildkraut JM, Thompson WD, et al. The genetic attrib- negative breast cancer. Breast Cancer Res Treat. 2015;150(2):389-394. utable risk of breast and ovarian cancer. Cancer. 1996;77(11):2318-2324. 45. Villarreal‐Garza C, Alvarez‐Gómez RM, Pérez‐Plasencia C, et al. 18. Hartman AR, Kaldate RR, Sailer LM, et al. Prevalence of BRCA Significant clinical impact of recurrent BRCA1 and BRCA2 mutations in mutations in an unselected population of triple‐negative breast cancer. Mexico. Cancer. 2015;121(3):372-378. Cancer. 2012;118(11):2787-2795. 46. Nahleh Z, Otoukesh S, Dwivedi AK, et al. Clinical and pathologi- 19. Anglian Breast Cancer Study Group. Prevalence and penetrance cal characteristics of Hispanic BRCA-associated breast cancers in the of BRCA1 and BRCA2 mutations in a population-based series of breast American-Mexican border city of El Paso, TX. American journal of cancer cancer cases. Br J Cancer. 2000;83(10):1301-1308. research. 2015;5(1):466. 20. Whittemore AS, Gong G, John EM, et al. Prevalence of BRCA1 mutation 47. Torres-Mejía G, Royer R, Llacuachaqui M, et al. Recurrent BRCA1 and carriers among US non-Hispanic Whites. Cancer Epidemi Biomark Prev. BRCA2 mutations in Mexican women with breast cancer. Cancer Epidemi 2004;13(12):2078-2083. Biomark Prevent. 2015;24(3):498-505. 21. Struewing JP, Hartge P, Wacholder S, et al. The risk of cancer associated 48. Abugattas J, Llacuachaqui M, Allende YS, et al. Prevalence of with specific mutations of BRCA1 and BRCA2 among Ashkenazi Jews. N BRCA1 and BRCA2 mutations in unselected breast cancer patients from Engl J Med. 1997;336(20):1401-1408. Peru. Clin Genet. 2015;88(4):371-375. 22. Dutil J, Golubeva VA, Pacheco-Torres AL, et al. The spectrum of 49. Dutil J, Colon-Colon JL, Matta JL, et al. Identification of the preva-

368 Cancer Control October 2016, Vol. 23, No. 4 lent BRCA1 and BRCA2 mutations in the female population of Puerto Rico. women with familial and/or early-onset breast cancer. Salud Publica Mex. Cancer Genet. 2012;205(5):242-248. 2014;56(2):206-212. 50. Delgado La, Fernández G, González A, et al. Hereditary breast cancer 75. Bretsky P, Haiman CA, Gilad S, et al. The relationship between associated with a germline BRCA2 mutation in identical female twins with twenty missense ATM variants andb cancer risk the multiethnic cohort. similar disease expression. Cancer Genet Cytogenet. 2002;133(1):24-28. Cancer Epidemiol Biomark Prevent. 2003;12(8):733-738. 51. Lara K, Consigliere N, Pérez J, et al. BRCA1 and BRCA2 mutations 76. Bell DW, Kim SH, Godwin AK, et al. Genetic and functional in breast cancer patients from Venezuela. Biol Res. 2012;45(2):117-130. analysis of CHEK2 (CHK2) variants in multiethnic cohorts. Int J Cancer. 52. Solano AR, Aceto GM, Delettieres D, et al. BRCA1 And BRCA2 2007;121(12):2661-2667. analysis of Argentinean breast/ovarian cancer patients selected for age 77. Damiola F, Pertesi M, Oliver J, et al. Rare key functional domain and family history highlights a role for novel mutations of putative south- missense substitutions in MRE11A, RAD50, and NBN contribute to breast American origin. Springerplus. 2012;1:20. cancer susceptibility: results from a breast cancer family registry case- 53. Dufloth RM, Carvalho S, Heinrich JK, et al. Analysis of BRCA1 and control mutation-screening study. Breast Cancer Res. 2014;16(3):R58. BRCA2 mutations in Brazilian breast cancer patients with positive family 78. Assumpção JG, Seidinger AL, Mastellaro MJ, et al. Association of history. Sao Paulo Med J. 2005;123(4):192-197. the germline TP53 R337H mutation with breast cancer in southern Brazil. 54. Carraro DM, Folgueira MA, Lisboa BCG, et al. Comprehensive analy- BMC Cancer. 2008;8(1):1. sis of BRCA1, BRCA2 and TP53 germline mutation and tumor characterization: 79. Giacomazzi J, Graudenz MS, Osorio CA, et al. Prevalence of the a portrait of early-onset breast cancer in Brazil. PloS One. 2013;8(3):e57581. TP53 p. R337H mutation in breast cancer patients in Brazil. PloS One. 55. Porchia LM, Gonzalez Mejia ME, Calderilla-Barbosa L, et al. Common 2014;9(6):e99893. BRCA1 and BRCA2 butations among Latin American breast cancer subjects: a 80. Jara L, Acevedo ML, Blanco R, et al. RAD51 135G> C polymor- meta-analysis. J Carcinogen Mutagen. 2015;2015. http://www.omicsonline.org/ phism and risk of familial breast cancer in a South American population. open-access/common-brca1-and-brca2-mutations-among-latin-american- Cancer Genet Cytogenet. 2007;178(1):65-69. breast-cancer-subjects-a-metaanalysis-2157-2518-1000228.pdf. Accessed 81. Gonzalez KD, Noltner KA, Buzin CH, et al. Beyond Li Fraumeni September 15, 2016. Syndrome: clinical characteristics of families with p53 germline mutations. 56. John EM, Miron A, Gong G, et al. Prevalence of pathogenic BRCA1 J Clin Ocol. 2009;27(8):1250-1256. mutation carriers in 5 US racial/ethnic groups. JAMA. 2007;298(24):2869-2876. 82. Olivier M, Goldgar DE, Sodha N, et al. Li-Fraumeni and related syn- 57. Mullineaux LG, Castellano TM, Shaw J, et al. Identification of germline dromes correlation between tumor type, family structure, and TP53 geno- 185delAG BRCA1 mutations in non‐Jewish Americans of Spanish ancestry type. Cancer Res. 2003;63(20):6643-6650. from the San Luis Valley, Colorado. Cancer. 2003;98(3):597-602. 83. Li FP, Fraumeni JF Jr, Mulvihill JJ, et al. A cancer family syndrome 58. Weitzel JN, Clague J, Martir-Negron A, et al. Prevalence and type of in twenty-four kindreds. Cancer Res. 1988;48(18):5358-5362. BRCA mutations in Hispanics undergoing genetic cancer risk assessment in 84. Børresen A-L, Andersen TI, Garber J, et al. Screening for germ line the southwestern United States: a report from the Clinical Cancer Genetics TP53 mutations in breast cancer patients. Cancer Res. 1992;52(11):3234-3236. Community Research Network. J Clin Oncol. 2013;31(2):210-216. 85. Sidransky D, Tokino T, Helzlsouer K, et al. Inherited p53 gene muta- 59. Laitman Y, Feng B-J, Zamir IM, et al. Haplotype analysis of the tions in breast cancer. Cancer research. 1992;52(10):2984-2986. 185delAG BRCA1 mutation in ethnically diverse populations. Eur J Hum 86. Lalloo F, Varley J, Moran A, et al. BRCA1, BRCA2 and TP53 muta- Genet. 2013;21(2):212-216. tions in very early-onset breast cancer with associated risks to relatives. 60. Ah Mew N, Hamel N, Galvez M, et al. Haplotype analysis of a BRCA1: Eur J Cancer. 2006;42(8):1143-1150. 185delAG mutation in a Chilean family supports its Ashkenazi origins. Clin 87. Ginsburg OM, Akbari MR, Aziz Z, et al. The prevalence of germ-line Genet. 2002;62(2):151-156. TP53 mutations in women diagnosed with breast cancer before age 30. 61. Weitzel JN, Lagos V, Blazer KR, et al. Prevalence of BRCA mutations Fam Cancer. 2009;8(4):563-567. and founder effect in high-risk Hispanic families. Cancer Epidemiol Biomark 88. McCuaig JM, Armel SR, Novokmet A, et al. Routine TP53 test- Prevent. 2005;14(7):1666-1671. ing for breast cancer under age 30: ready for prime time? Fam Cancer. 62. Weitzel JN, Lagos VI, Herzog JS, et al. Evidence for common ances- 2012;11(4):607-613. tral origin of a recurring BRCA1 genomic rearrangement identified in high-risk 89. Ribeiro RC, Sandrini F, Figueiredo B, et al. An inherited p53 muta- Hispanic families. Cancer Epidemiol Biomark Prevent. 2007;16(8):1615-1620. tion that contributes in a tissue-specific manner to pediatric adrenal cortical 63. Judkins T, Rosenthal E, Arnell C, et al. Clinical significance of large carcinoma. Proceed Natl Acad Sci. 2001;98(16):9330-9335. rearrangements in BRCA1 and BRCA2. Cancer. 2012;118(21):5210-5216. 90. Gomes M, Kotsopoulos J, de Almeida GL, et al. The R337H mutation 64. Delgado L, Fernández G, Grotiuz G, et al. BRCA1 and BRCA2 germ- in TP53 and breast cancer in Brazil. Hered Cancer Clin Pract. 2012;10(1):3. line mutations in Uruguayan breast and breast–ovarian cancer families. 91. Custódio G, Parise GA, Kiesel Filho N, et al. Impact of neonatal Identification of novel mutations and unclassified variants. Breast Cancer screening and surveillance for the TP53 R337H mutation on early detection Res Treat. 2011;128(1):211-218. of childhood adrenocortical tumors. J Clin Oncol. 2013;31(20):2619-2626. 65. Esteban Cardeñosa E, Bolufer Gilabert P, de Juan Jimenez I, et al. 92. Palmero EI, Schüler-Faccini L, Caleffi M, et al. Detection of R337H, Broad BRCA1 and BRCA2 mutational spectrum and high incidence of recurrent a germline TP53 mutation predisposing to multiple cancers, in asymptom- and novel mutations in the eastern Spain population. Breast Cancer Res Treat. atic women participating in a breast cancer screening program in southern 2010;121(1):257-260. Brazil. Cancer Lett. 2008;261(1):21-25. 66. Miramar M, Calvo M, Rodriguez A, et al. Genetic analysis of BRCA1 93. Garritano S, Gemignani F, Palmero EI, et al. Detailed haplotype and BRCA2 in breast/ovarian cancer families from Aragon (Spain): two novel analysis at the TP53 locus in p. R337H mutation carriers in the population of truncating mutations and a large genomic deletion in BRCA1. Breast Cancer southern Brazil: evidence for a founder effect. Hum Mutat. 2010;31(2):143-150. Res Treat. 2008;112(2):353-358. 94. Blazer KR, Nehoray B, Solomon I, et al. Next-generation test- 67. Infante M, Durán M, Esteban-Cardeñosa E, et al. High proportion ing for cancer risk: perceptions, experiences, and needs among early of novel mutations of BRCA1 and BRCA2 in breast/ovarian cancer patients adopters in community healthcare settings. Genet Test Molec Biomark. from Castilla-Leon (central Spain). J Hum Genet. 2006;51(7):611-617. 2015;19(12):657-665. 68. de la Hoya M, Osorio A, Godino J, et al. Association between 95. Maxwell KN, Wubbenhorst B, D’Andrea K, et al. Prevalence of muta- BRCA1 and BRCA2 mutations and cancer phenotype in Spanish breast/ tions in a panel of breast cancer susceptibility genes in BRCA1/2-negative ovarian cancer families: implications for genetic testing. Int J Cancer. patients with early-onset breast cancer. Genet Med. 2014;17(8):630-638. 2002;97(4):466-471. 96. Tung N, Battelli C, Allen B, et al. Frequency of mutations in individuals 69. Blay P, Santamaría I, Pitiot AS, et al. Mutational analysis of BRCA1 with breast cancer referred for BRCA1 and BRCA2 testing using next‐ and BRCA2 in hereditary breast and ovarian cancer families from Asturias generation sequencing with a 25‐gene panel. Cancer. 2015;121(1):25-33. (Northern Spain). BMC Cancer. 2013;13(1):1. 97. Kapoor NS, Curcio LD, Blakemore CA, et al. Multigene panel test- 70. Bolufer P, Munárriz B, Santaballa A, et al. BRCA1 and BRCA2 ing detects equal rates of pathogenic BRCA1/2 mutations and has a higher mutations in patients with familial breast cancer [In Spanish]. Med Clin. diagnostic yield compared to limited BRCA1/2 analysis alone in patients at 2005;124(1):10-12. risk for hereditary breast cancer. Ann Surg Oncol. 2015;22(10):3282-3288. 71. de Juan Jiménez I, Casado ZG, Suela SP, et al. Novel and recurrent 98. Li J, Meeks H, Feng B-J, et al. Targeted massively parallel sequencing BRCA1/BRCA2 mutations in early onset and familial breast and ovarian of a panel of putative breast cancer susceptibility genes in a large cohort of cancer detected in the Program of Genetic Counseling in Cancer of Valencian multiple-case breast and ovarian cancer families. J Med Genet. 2016;53(1):34-42. Community (eastern Spain). Relationship of family phenotypes with mutation 99. Couch FJ, Hart SN, Sharma P, et al. Inherited mutations in 17 breast prevalence. Fam Cancer. 2013;12(4):767-777. cancer susceptibility genes among a large triple-negative breast cancer cohort 72. González-Hormazábal P, Bravo T, Blanco R, et al. Association of unselected for family history of breast cancer. J Clin Oncol. 2015;33(4):304-311. common ATM variants with familial breast cancer in a South American 100. Desmond A, Kurian AW, Gabree M, et al. Clinical actionability of multi- population. BMC Cancer. 2008;8(1):1. gene panel testing for hereditary breast and ovarian cancer risk assessment. 73. Leyton Y, Gonzalez-Hormazabal P, Blanco R, et al. Association of JAMA Oncol. 2015;1(7):943-951. PALB2 sequence variants with the risk of familial and early-onset breast 101. Doherty J, Bonadies DC, Matloff ET. Testing for hereditary breast cancer in a South-American population. BMC Cancer. 2015;15(1):1. cancer: panel or targeted testing? Experience from a clinical cancer genetics 74. Calderón-Zúñiga Fdel C, Ocampo-Gomez G, Lopez-Marquez FC, et practice. J Genet Counsel. 2015;24(4):683-687. al. ATM polymorphisms IVS24-9delT, IVS38-8T>C, and 5557G>A in Mexican 102. Kurian AW, Hare EE, Mills MA, et al. Clinical evaluation of a multi-

October 2016, Vol. 23, No. 4 Cancer Control 369 ple-gene sequencing panel for hereditary cancer risk assessment. J Clin breast and ovarian cancer risk. PLoS Genet. 2013;9(3):e1003212. Oncol. 2014;32(19):2001-2009. 130. Li J, Humphreys K, Darabi H, et al. A genome-wide association 103. Lincoln SE, Kobayashi Y, Anderson MJ, et al. A systematic comparison scan on estrogen receptor-negative breast cancer. Breast Cancer Res. of traditional and multigene panel testing for hereditary breast and ovarian 2010;12(6):R93. cancer genes in more than 1000 patients. J Molec Diagnost. 2015;17(5):533-544. 131. Murabito JM, Rosenberg CL, Finger D, et al. A genome-wide 104. Ricker C, Culver JO, Lowstuter K, et al. Increased yield of actionable association study of breast and prostate cancer in the NHLBI’s Framingham mutations using multi-gene panels to assess hereditary cancer susceptibility Heart Study. BMC Med Genet. 2007;8(1):1. in an ethnically diverse clinical cohort. Cancer Genet. 2016; 209(4):130-137. 132. Antoniou AC, Wang X, Fredericksen ZS, et al. A locus on 19p13 105. Mai PL, Vadaparampil ST, Breen N, et al. Awareness of cancer modifies risk of breast cancer in BRCA1 mutation carriers and is associated susceptibility genetic testing: the 2000, 2005, and 2010 National Health with hormone receptor-negative breast cancer in the general population. Interview Surveys. Am J Prevent Med. 2014;46(5):440-448. Nat Genet. 2010;42(10):885-892. 106. Vadaparampil ST, Wideroff L, Breen N, et al. The impact of acculturation 133. Cai Q, Long J, Lu W, et al. Genome-wide association study identifies on awareness of genetic testing for increased cancer risk among Hispanics breast cancer risk variant at 10q21. 2: results from the Asia Breast Cancer in the year 2000 National Health Interview Survey. Cancer Epidemiol Biomark Consortium. Hum Molec Genet. 2011;20(24):4991-4999. Prevent. 2006;15(4):618-623. 134. Kim H-c, Lee J-Y, Sung H, et al. A genome-wide association study 107. Wideroff L, Thomas Vadaparampil S, Breen N, et al. Awareness of identifies a breast cancer risk variant in ERBB4 at 2q34: results from the genetic testing for increased cancer risk in the year 2000 National Health Seoul Breast Cancer Study. Breast Cancer Res. 2012;14(2):1-12. Interview Survey. Pub Health Genom. 2004;6(3):147-156. 135. Elgazzar S, Zembutsu H, Takahashi A, et al. A genome-wide as- 108. Gammon AD, Rothwell E, Simmons R, et al. Awareness and prefer- sociation study identifies a genetic variant in the SIAH2 locus associated ences regarding BRCA1/2 genetic counseling and testing among Latinas with hormonal receptor-positive breast cancer in Japanese. J Hum Genet. and non-Latina white women at increased risk for hereditary breast and 2012;57(12):766-771. ovarian cancer. J Genet Counsel. 2011;20(6):625-638. 136. Long J, Cai Q, Shu X-O, et al. Identification of a functional genetic 109. Vadaparampil ST, McIntyre J, Quinn GP. Awareness, perceptions, variant at 16q12. 1 for breast cancer risk: results from the Asia Breast Cancer and provider recommendation related to genetic testing for hereditary breast Consortium. PLoS Genet. 2010;6(6):e1001002. cancer risk among at-risk Hispanic women: similarities and variations by 137. Michailidou K, Beesley J, Lindstrom S, et al. Genome-wide asso- sub-ethnicity. J Genet Counsel. 2010;19(6):618-629. ciation analysis of more than 120,000 individuals identifies 15 new suscep- 110. Kaplan CP, Haas JS, Pérez-Stable EJ, et al. Breast cancer risk re- tibility loci for breast cancer. Nat Genet. 2015;47(4):373-380. duction options: awareness, discussion, and use among women from four 138. Thomas G, Jacobs KB, Kraft P, et al. A multistage genome-wide ethnic groups. Cancer Epidemiol Biomark Prevent. 2006;15(1):162-166. association study in breast cancer identifies two new risk alleles at 1p11. 2 111. Mays D, Sharff ME, DeMarco TA, et al. Outcomes of a systems- and 14q24. 1 (RAD51L1). Nat Genet. 2009;41(5):579-584. level intervention offering breast cancer risk assessments to low-income 139. Siddiq A, Couch FJ, Chen GK, et al. A meta-analysis of genome- underserved women. Fam Cancer. 2012;11(3):493-502. wide association studies of breast cancer identifies two novel susceptibility 112. Jagsi R, Griffith KA, Kurian AW, et al. Concerns about cancer risk loci at 6q14 and 20q11. Hum Molec Genet. 2012;21(24):5373-5384. and experiences with genetic testing in a diverse population of patients 140. Michailidou K, Hall P, Gonzalez-Neira A, et al. Large-scale geno- with breast cancer. J Clin Oncol. 2015;33(14):1584-1591. typing identifies 41 new loci associated with breast cancer risk. Nat Genet. 113. Lagos VI, Perez MA, Ricker CN, et al. Social‐cognitive aspects of 2013;45(4):353-361. underserved Latinas preparing to undergo genetic cancer risk assessment for 141. Chen F, Chen GK, Stram DO, et al. A genome-wide associa- hereditary breast and ovarian cancer. Psycho‐Oncology. 2008;17(8):774-782. tion study of breast cancer in women of African ancestry. Hum Genet. 114. Vadaparampil ST, Quinn GP, Dutil J, et al. A pilot study of knowl- 2013;132(1):39-48. edge and interest of genetic counseling and testing for hereditary breast 142. Haiman CA, Chen GK, Vachon CM, et al. A common variant at and ovarian cancer syndrome among Puerto Rican women. J Comm Genet. the TERT-CLPTM1L locus is associated with estrogen receptor-negative 2011;2(4):211-221. breast cancer. Nat Genet. 2011;43(12):1210-1214. 115. Ricker C, Lagos V, Feldman N, et al. If we build it… will they come?– 143. Zheng W, Long J, Gao Y-T, et al. Genome-wide association study establishing a cancer genetics services clinic for an underserved predominantly identifies a new breast cancer susceptibility locus at 6q25. 1. Nat Genet. Latina cohort. J Genet Counsel. 2006;15(6):505-514. 2009;41(3):324-328. 116. Komenaka IK, Nodora JN, Madlensky L, et al. Participation of low- 144. Fejerman L, Ahmadiyeh N, Hu D, et al. Genome-wide association income women in genetic cancer risk assessment and BRCA 1/2 testing: study of breast cancer in Latinas identifies novel protective variants on 6q25. the experience of a safety-net institution. J Comm Genet. 2015:1-7. Nat Comm. 2014;5. 117. Olaya W, Esquivel P, Wong JH, et al. Disparities in BRCA testing: 145. Haiman CA, Stram DO, Pike MC, et al. A comprehensive haplotype when insurance coverage is not a barrier. Am J Surg. 2009;198(4):562-565. analysis of CYP19 and breast cancer risk: the multiethnic cohort. Hum 118. Woodson AH, Profato JL, Rizvi SH, et al. Service delivery model Molec Genet. 2003;12(20):2679-2692. and experiences in a cancer genetics clinic for an underserved population. 146. Cheng I, Penney KL, Stram DO, et al. Haplotype-based association J Health Care Poor Underserv. 2015;26(3):784-791. studies of IGFBP1 and IGFBP3 with prostate and breast cancer risk: the mul- 119. Macconaill LE, Garraway LA. Clinical implications of the cancer tiethnic cohort. Cancer Epidemiol Biomark Prevent. 2006;15(10):1993-1997. genome. J Clin Oncol. 2010;28(35):5219-5228. 147. Women’s Health Initiative Study Group. Design of the Women’s 120. Easton DF, Pooley KA, Dunning AM, et al. Genome-wide asso- Health Initiative clinical trial and observational study. Control Clin Trials. ciation study identifies novel breast cancer susceptibility loci. Nature. 1998;19(1):61-109. 2007;447(7148):1087-1093. 148. Hays J, Hunt JR, Hubbell FA, et al. The Women’s Health Initiative 121. Stacey SN, Manolescu A, Sulem P, et al. Common variants on chro- recruitment methods and results. Ann Epidemiol. 2003;13(9):S18-S77. mosome 5p12 confer susceptibility to estrogen receptor–positive breast 149. Slattery ML, John EM, Torres-Mejia G, et al. Genetic variation in genes cancer. Nat Genet. 2008;40(6):703-706. involved in hormones, inflammation and energetic factors and breast cancer 122. Hunter DJ, Kraft P, Jacobs KB, et al. A genome-wide association risk in an admixed population. Carcinogenesis. 2012;33(8):1512-1521. study identifies alleles in FGFR2 associated with risk of sporadic post- 150. Ramírez-Patiño R, Figuera LE, Puebla-Pérez AM, et al. Intron 4 menopausal breast cancer. Nat Genet. 2007;39(7):870-874. VNTR (4a/b) polymorphism of the endothelial nitric oxide synthase gene 123. Garcia-Closas M, Couch FJ, Lindstrom S, et al. Genome-wide is associated with breast cancer in Mexican women. J Korean Med Sci. association studies identify four ER negative-specific breast cancer risk loci. 2013;28(11):1587-1594. Nat Genet. 2013;45(4):392-398. 151. Quintero-Ramos A, Gutiérrez-Rubio S, Del Toro-Arreola A, et al. 124. Fletcher O, Johnson N, Orr N, et al. Novel breast cancer suscepti- Association between polymorphisms in the thymidylate synthase gene and risk bility locus at 9q31. 2: results of a genome-wide association study. J Natl of breast cancer in a Mexican population. Genet Mol Res. 2014;13(4):8749-8756. Cancer Inst. 2011;103(5):425-435. 152. Jara L, Dubois K, Gaete D, et al. Variants in DNA double-strand break 125. Gold B, Kirchhoff T, Stefanov S, et al. Genome-wide association repair genes and risk of familial breast cancer in a South American population. study provides evidence for a breast cancer risk locus at 6q22. 33. Proceed Breast Cancer Res Treat. 2010;122(3):813-822. Natl Acad Sci. 2008;105(11):4340-4345. 153. Oliveira CB, Cardoso-Filho C, Bossi LS, et al. Association of CYP1A1 126. Turnbull C, Ahmed S, Morrison J, et al. Genome-wide associa- A4889G and T6235C polymorphisms with the risk of sporadic breast cancer tion study identifies five new breast cancer susceptibility loci. Nat Genet. in Brazilian women. Clinics. 2015;70(10):680-685. 2010;42(6):504-507. 154. Pérez‐Mayoral J, Pacheco‐Torres AL, Morales L, et al. Genetic polymor- 127. Gaudet MM, Kuchenbaecker KB, Vijai J, et al. Identification of a phisms in RAD23B and XPC modulate DNA repair capacity and breast cancer BRCA2-specific modifier locus at 6p24 related to breast cancer risk. PLoS risk in Puerto Rican women. Molec Carcinogen. 2013;52(suppl 1):127-138. Genet. 2013;9(3):e1003173. 155. Kolonel LN, Henderson BE, Hankin JH, et al. A multiethnic cohort 128. Long J, Cai Q, Sung H, et al. Genome-wide association study in in Hawaii and Los Angeles: baseline characteristics. Am J Epidemiol. east Asians identifies novel susceptibility loci for breast cancer. PLoS 2000;151(4):346-357. Genet. 2012;8(2):e1002532. 156. Kolonel LN, Altshuler D, Henderson BE. The multiethnic cohort study: 129. Couch FJ, Wang X, McGuffog L, et al. Genome-wide association exploring genes, lifestyle and cancer risk. Nat Rev Cancer. 2004;4(7):519-527. study in BRCA1 mutation carriers identifies novel loci associated with 157. John EM, Phipps AI, Davis A, et al. Migration history, acculturation,

370 Cancer Control October 2016, Vol. 23, No. 4 and breast cancer risk in Hispanic women. Cancer Epidemiol Biomark 181. Flores-Ramos LG, Escoto-De Dios A, Puebla-Pérez A, et al. Associa- Prevent. 2005;14(12):2905-2913. tion of the tumor necrosis factor-alpha-308G>: a polymorphism with breast 158. John EM, Horn-Ross PL, Koo J. Lifetime physical activity and breast cancer in Mexican women. Genet Mol Res. 2013;12:5680-5693. cancer risk in a multiethnic population the san francisco Bay area breast 182. Pellatt AJ, Wolff RK, John EM, et al. SEPP1 influences breast cancer cancer study. Cancer Epidemiol Biomark Prevent. 2003;12(11):1143-1152. risk among women with greater native american ancestry: the breast cancer 159. John EM, Hopper JL, Beck JC, et al. The breast cancer family reg- health disparities study. PLoS One. 2013;8(11):e80554. istry: an infrastructure for cooperative multinational, interdisciplinary and 183. Baumgartner KB, Schlierf TJ, Yang D, et al. N-acetyltransferase 2 translational studies of the genetic epidemiology of breast cancer. Breast genotype modification of active cigarette smoking on breast cancer risk Cancer Res. 2004;6(4):R375-R389. among hispanic and non-hispanic white women. Toxicol Sci. 2009:kfp199. 160. Slattery ML, Sweeney C, Edwards S, et al. Body size, weight change, 184. Cardenas-Rodriguez N, Lara-Padilla E, Bandala C, et al. CYP2W1, fat distribution and breast cancer risk in Hispanic and non-Hispanic white CYP4F11 and CYP8A1 polymorphisms and interaction of CYP2W1 geno- women. Breast Cancer Res Treat. 2007;102(1):85-101. types with risk factors in Mexican women with breast cancer. Asian Pac J 161. Slattery ML, Sweeney C, Herrick J, et al. ESR1, AR, body size, and Cancer Prevent. 2012;13(3):837-846. breast cancer risk in Hispanic and non-Hispanic white women living in the 185. Gutierrez-Rubio S, Quintero-Ramos A, Duran-Cardenas A, et al. southwestern United States. Breast Cancer Res Treat. 2007;105(3):327-335. 1236 C/T and 3435 C/T polymorphisms of the ABCB1 gene in Mexican 162. Gallegos-Arreola MP, Figuera LE, Flores-Ramos LG, et al. Asso- breast cancer patients. Genet Mol Res. 2015;14(1):1250-1259. ciation of the Alu insertion polymorphism in the progesterone receptor 186. Ramos-Silva A, Figuera L, Soto-Quintana O, et al. Association of the gene with breast cancer in a Mexican population. Arch Med Science. C677T polymorphism in the methylenetetrahydrofolate reductase gene with 2015;11(3):551. breast cancer in a Mexican population. Genet Mol Res. 2015;14(2):4015. 163. Haiman CA, Garcia RR, Hsu C, et al. Screening and association 187. Slattery ML, John EM, Torres‐Mejia G, et al. Angiogenesis genes, testing of common coding variation in steroid hormone receptor co-activator dietary oxidative balance and breast cancer risk and progression: the Breast and co-repressor genes in relation to breast cancer risk: the multiethnic Cancer Health Disparities study. Int J Cancer. 2014;134(3):629-644. cohort. BMC Cancer. 2009;9(1):1. 188. Tsai H-J, Choudhry S, Naqvi M, et al. Comparison of three methods to 164. Boone SD, Baumgartner KB, Baumgartner RN, et al. Associations estimate genetic ancestry and control for stratification in genetic association between CYP19A1 polymorphisms, Native American ancestry, and breast studies among admixed populations. Hum Genet. 2005;118(3-4):424-433. cancer risk and mortality: the Breast Cancer Health Disparities Study. 189. Tang H, Quertermous T, Rodriguez B, et al. Genetic structure, self- Cancer Cause Control. 2014;25(11):1461-1471. identified race/ethnicity, and confounding in case-control association studies. 165. Moreno-Galván M, Herrera-González NE, Robles-Pérez V, et al. Am J Hum Genet. 2005;76(2):268-275. Impact of CYP1A1 and COMT genotypes on breast cancer risk in Mexican 190. Slattery ML, Baumgartner KB, Giuliano AR, et al. Replication of five women: a pilot study. Int J Biol Mark. 2009;25(3):157-163. GWAS-identified loci and breast cancer risk among Hispanic and non- 166. Weston A, Pan C, Bleiweiss IJ, et al. CYP17 genotype and breast Hispanic white women living in the southwestern United States. Breast Cancer cancer risk. Cancer Epidemiol Biomark Prevent. 1998;7(10):941-944. Res Treat. 2011;129(2):531-539. 167. Sarkissyan M, Mishra DK, Wu Y, et al. IGF gene polymorphisms 191. Fejerman L, Stern MC, John EM, et al. Interaction between com- and breast cancer in African-American and Hispanic women. Int J Oncol. mon breast cancer susceptibility variants, genetic ancestry, and nonge- 2011;38(6):1663-1673. netic risk factors in Hispanic women. Cancer Epidemiol Biomark Prevent. 168. Slattery ML, John EM, Stern MC, et al. Associations with growth 2015;24(11):1731-1738. factor genes (FGF1, FGF2, PDGFB, FGFR2, NRG2, EGF, ERBB2) with 192. Fejerman L, Stern MC, Ziv E, et al. Genetic ancestry modifies the breast cancer risk and survival: the Breast Cancer Health Disparities study. association between genetic risk variants and breast cancer risk among Breast Cancer Res Treat. 2013;140(3):587-601. Hispanic and non-Hispanic white women. Carcinogenesis. 2013;34(8):1787-1793. 169. Connor AE, Baumgartner RN, Baumgartner KB, et al. Epidermal 193. Murillo-Zamora E, Moreno-Macías H, Ziv E, et al. Association between growth factor receptor (EGFR) polymorphisms and breast cancer among rs2981582 polymorphism in the FGFR2 gene and the risk of breast cancer in Hispanic and non-Hispanic white women: the Breast Cancer Health Dis- Mexican women. Arch Med Res. 2013;44(6):459-466. parities study. Int J Molec Epidemiol Genet. 2013;4(4):235. 194. Jara L, Gonzalez-Hormazabal P, Cerceño K, et al. Genetic variants in 170. Boone SD, Baumgartner KB, Baumgartner RN, et al. Associations FGFR2 and MAP3K1 are associated with the risk of familial and early-onset between genetic variants in the TGF-β signaling pathway and breast can- breast cancer in a South-American population. Breast Cancer Res Treat. cer risk among Hispanic and non-Hispanic white women. Breast Cancer 2013;137(2):559-569. Res Treat. 2013;141(2):287-297. 195. Elematore I, Gonzalez-Hormazabal P, Reyes JM, et al. Association of 171. Slattery ML, John EM, Torres‐Mejia G, et al. Genetic variation in genetic variants at TOX3, 2q35 and 8q24 with the risk of familial and early- bone morphogenetic proteins and breast cancer risk in hispanic and non‐ onset breast cancer in a South-American population. Molec Biol Rep. hispanic white women: The breast cancer health disparities study. Int J 2014;41(6):3715-3722. Cancer. 2013;132(12):2928-2939. 196. Cheng C-Y, Kao WL, Patterson N, et al. Admixture mapping of 172. Ingles SA, Garcia DG, Wang W, et al. Vitamin D receptor geno- 15,280 African Americans identifies obesity susceptibility loci on chromo- type and breast cancer in Latinas (United States). Cancer Cause Control. somes 5 and X. PLoS Genet. 2009;5(5):e1000490. 2000;11(1):25-30. 197. Jia L, Landan G, Pomerantz M, et al. Functional enhancers at the 173. Connor AE, Baumgartner RN, Baumgartner KB, et al. Associations gene-poor 8q24 cancer-linked locus. PLoS Genet. 2009;5(8):e1000597. between TCF7L2 polymorphisms and risk of breast cancer among Hispanic 198. Reich D, Nalls MA, Kao WL, et al. Reduced neutrophil count in and non-Hispanic White women: the Breast Cancer Health Disparities people of African descent is due to a regulatory variant in the Duffy antigen study. Breast Cancer Res Treat. 2012;136(2):593-602. receptor for chemokines gene. PLoS Genet. 2009;5(1):e1000360. 174. Slattery ML, John E, Torres-Mejia G, et al. Matrix metalloproteinase 199. Nalls MA, Wilson JG, Patterson NJ, et al. Admixture mapping of white genes are associated with breast cancer risk and survival: the Breast Cancer cell count: genetic locus responsible for lower white blood cell count in the Health Disparities study. PloS One. 2013;8(5):e63165. Health ABC and Jackson Heart studies. Am J Hum Genet. 2008;82(1):81-87. 175. Delgado-Enciso I, Cepeda-Lopez FR, Monrroy-Guizar EA, et al. 200. Pritchard JK, Stephens M, Donnelly P. Inference of population Matrix metalloproteinase-2 promoter polymorphism is associated with breast structure using multilocus genotype data. Genetics. 2000;155(2):945-959. cancer in a Mexican population. Gynecol Obstet Invest. 2008;65(1):68-72. 201. Reich D, Patterson N, Ramesh V, et al. Admixture mapping of an 176. Slattery ML, Herrick JS, Torres-Mejia G, et al. Genetic variants in allele affecting interleukin 6 soluble receptor and interleukin 6 levels. Am J interleukin genes are associated with breast cancer risk and survival in Hum Genet. 2007;80(4):716-726. a genetically admixed population: the Breast Cancer Health Disparities 202. Winkler CA, Nelson GW, Smith MW. Admixture mapping comes of age. study. Carcinogenesis. 2014:bgu078. Ann Rev Genom Hum Genet. 2010;11:65-89. 177. Slattery ML, Lundgreen A, Hines L, et al. Energy homeostasis 203. Deo RC, Patterson N, Tandon A, et al. A high-density admixture scan genes and breast cancer risk: The influence of ancestry, body size, and in 1,670 African Americans with hypertension. PLoS Genet. 2007;3(11):e196. menopausal status, the breast cancer health disparities study. Cancer 204. Freedman ML, Haiman CA, Patterson N, et al. Admixture mapping Epidemiol. 2015;39(6):1113-1122. identifies 8q24 as a prostate cancer risk locus in African-American men. 178. Connor A, Baumgartner RN, Kerber RA, et al. ADRB2 G–G haplo- Proceed Natl Acad Sci. 2006;103(38):14068-14073. type associated with breast cancer risk among Hispanic and non-Hispanic 205. Jemal A, Siegel R, Xu J, et al. Cancer statistics, 2010. CA Cancer J white women: interaction with type 2 diabetes and obesity. Cancer Cause Clin. 2010;60(5):277-300. Control. 2012;23(10):1653-1663. 206. Fejerman L, John EM, Huntsman S, et al. Genetic ancestry and risk 179. Connor AE, Baumgartner RN, Baumgartner KB, et al. Associations of breast cancer among US Latinas. Cancer Res. 2008;68(23):9723-9728. between ALOX, COX, and CRP polymorphisms and breast cancer among 207. Fejerman L, Romieu I, John EM, et al. European ancestry is pos- Hispanic and non‐Hispanic white women: the breast cancer health dispari- itively associated with breast cancer risk in Mexican women. Cancer ties study. Molec Carcinogen. 2015;54(12):1541-1553. Epidemiol Biomark Prevent. 2010;19(4):1074-1082. 180. Kaklamani VG, Hoffmann TJ, Thornton TA, et al. Adiponectin path- 208. Fejerman L, Chen GK, Eng C, et al. Admixture mapping identifies a way polymorphisms and risk of breast cancer in African Americans and locus on 6q25 associated with breast cancer risk in US Latinas. Hum Molec Hispanics in the Women’s Health Initiative. Breast Cancer Res Treat. Genet. 2012;21(8):1907-1917. 2013;139(2):461-468. 209. Hines LM, Risendal B, Byers T, et al. Ethnic disparities in breast tumor

October 2016, Vol. 23, No. 4 Cancer Control 371 phenotypic subtypes in Hispanic and non-Hispanic white women. J Women CYP3A5, CYP2C8 and IL-10 genes in Mexican and Spanish women with Health. 2011;20(10):1543-1550. breast cancer. Pharmacogenomics. 2013;14(13):1583-1592. 210. Li CI, Malone KE, Daling JR. Differences in breast cancer stage, treat- 236. González-Tejera G, Gaedigk A, Corey S. Genetic variants of the ment, and survival by race and ethnicity. Arch Intern Med. 2003;163(1):49-56. drug-metabolizing enzyme CYP2D6 in Puerto Rican psychiatry patients: 211. Chlebowski RT, Chen Z, Anderson GL, et al. Ethnicity and breast a preliminary report and potential implications for breast cancer patients. cancer: factors influencing differences in incidence and outcome. J Natl P R Health Sci J. 2010;29(3):299-304. Cancer Inst. 2005;97(6):439-448. 237. Bernard S, Neville KA, Nguyen AT, et al. Interethnic differences in 212. Dunnwald LK, Rossing MA, Li CI. Hormone receptor status, tumor genetic polymorphisms of CYP2D6 in the US population: clinical implications. characteristics, and prognosis: a prospective cohort of breast cancer Oncologist. 2006;11(2):126-135. patients. Breast Cancer Res. 2007;9(1):R6. 238. Cuzick J, Sestak I, Baum M, et al. Effect of anastrozole and tamoxifen 213. Hausauer AK, Keegan T, Chang ET, et al. Recent breast cancer as adjuvant treatment for early-stage breast cancer: 10-year analysis of trends among Asian/Pacific Islander, Hispanic, and African-American women the ATAC trial. Lancet Oncol. 2010;11(12):1135-1141. in the US: changes by tumor subtype. Breast Cancer Res. 2007;9(6):R90. 239. Coates AS, Keshaviah A, Thürlimann B, et al. Five years of letrozole 214. Parise CA, Bauer KR, Brown MM, et al. Breast cancer subtypes as compared with tamoxifen as initial adjuvant therapy for postmenopausal defined by the estrogen receptor (ER), progesterone receptor (PR), and the women with endocrine-responsive early breast cancer: update of study human epidermal growth factor receptor 2 (HER2) among women with invasive BIG 1-98. J Clin Oncol. 2007;25(5):486-492. breast cancer in California, 1999–2004. Breast J. 2009;15(6):593-602. 240. Ma CX, Adjei AA, Salavaggione OE, et al. Human aromatase: gene 215. Carvalho FM, Bacchi LM, Pincerato KM, et al. Geographic differ- resequencing and functional genomics. Cancer Res. 2005;65(23):11071-11082. ences in the distribution of molecular subtypes of breast cancer in Brazil. 241. Moy B, Tu D, Pater J, et al. Clinical outcomes of ethnic minority women BMC Women Health. 2014;14(1):1. in MA.17: a trial of letrozole after 5 years of tamoxifen in postmenopausal 216. Howlader N, Altekruse SF, Li CI, et al. US incidence of breast can- women with early stage breast cancer. Ann Oncol. 2006;17(11):1637-1643. cer subtypes defined by joint hormone receptor and HER2 status. J Natl 242. Allman R, Dite GS, Hopper JL, et al. SNPs and breast cancer risk Cancer Inst. 2014;106(5):dju055. prediction for African American and Hispanic women. Breast Cancer Res 217. Swain SM, Baselga J, Kim S-B, et al. Pertuzumab, trastuzumab, Treat. 2015;154(3):583-589. and docetaxel in HER2-positive metastatic breast cancer. N Engl J Med. 243. Parra A, Karnad AB, Thompson IM. Hispanic accrual on randomized 2015;372(8):724-734. cancer clinical trials: a call to arms. J Clin Oncol. 2014;32(18):1871-1873. 218. Romond EH, Perez EA, Bryant J, et al. Trastuzumab plus adjuvant 244. Gordon NP, Iribarren C. Health-related characteristics and pre- chemotherapy for operable HER2-positive breast cancer. N Engl J Med. ferred methods of receiving health education according to dominant lan- 2005;353(16):1673-1684. guage among Latinos aged 25 to 64 in a large Northern California health 219. Harris LN, Ismaila N, McShane LM, et al; American Society of Clinical plan. BMC Public Health. 2008;8(1):1. Oncology Clinical Practice Guideline. Use of biomarkers to guide decisions on 245. Merchant G, Buelna C, Castañeda SF, et al. Accelerometer- adjuvant systemic therapy for women with early-stage invasive breast cancer. measured sedentary time among Hispanic adults: results from the Hispan- J Clin Oncol. 2016;34(10):1134-1150. ic Community Health Study/Study of Latinos (HCHS/SOL). Prevent Med 220. Senkus E, Kyriakides S, Penault-Llorca F, et al. Primary breast Rep. 2015;2:845-853. cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and 246. Bonner D, Cragun D, Reynolds M, et al. Recruitment of a popula- follow-up. Ann Oncol. 2013;24 Suppl 6:vi7-23. tion-based sample of young black women with breast cancer through a 221. National Comprehensive Cancer Network (NCCN). NCCN Clinical state cancer Registry. Breast J. 2016;22(2):166-172. Pratice Guidelines in Oncology: Breast Cancer. v2.2016. https://www.nccn.org/ 247. Darling M, Gonzalez F, Graves K, et al. Practical tips for establishing professionals/physician_gls/pdf/breast.pdf. Accessed September 15, 2016. partnerships with academic researchers: a resource guide for community- 222. Kalinsky K, Lim EA, Andreopoulou E, et al. Increased expression of based organizations. Prog Community Health Partnersh. 2015;9(2):203-212. tumor proliferation genes in Hispanic women with early-stage breast cancer. 248. Rush CL, Darling M, Elliott MG, et al. Engaging Latina cancer survivors, Cancer Invest. 2014;32(9):439-444. their caregivers, and community partners in a randomized controlled trial: 223. Eiermann W, Rezai M, Kümmel S, et al. The 21-gene recurrence Nueva Vida intervention. Qua Life Res. 2015;24(5):1107-1118. score assay impacts adjuvant therapy recommendations for ER-positive, 249. Detz A, Mangione CM, de Jaimes FN, et al. Language concor- node-negative and node-positive early breast cancer resulting in a risk- dance, interpersonal care, and diabetes self-care in rural Latino patients. adapted change in chemotherapy use. Annals of oncology. 2012:mds512. J Gen Intern Med. 2014;29(12):1650-1656. 224. Partin JF, Mamounas EP. Impact of the 21-gene recurrence score assay compared with standard clinicopathologic guidelines in adjuvant therapy selection for node-negative, estrogen receptor-positive breast cancer. Ann Surg Oncol. 2011;18(12):3399-3406. 225. Geffen D, Abu-Ghanem S, Sion-Vardy N, et al. The impact of the 21-gene recurrence score assay on decision making about adjuvant chemotherapy in early-stage estrogen-receptor-positive breast cancer in an oncology practice with a unified treatment policy. Ann Oncol. 2011:mdq769. 226. Lo SS, Mumby PB, Norton J, et al. Prospective multicenter study of the impact of the 21-gene recurrence score assay on medical oncologist and patient adjuvant breast cancer treatment selection. J Clin Oncol. 2010;28(10):1671-1676. 227. Bargallo JE, Lara F, Shaw‐Dulin R, et al. A study of the impact of the 21‐gene breast cancer assay on the use of adjuvant chemotherapy in women with breast cancer in a Mexican public hospital. J Surg Oncol. 2015;111(2):203-207. 228. Bargalló-Rocha JE, Lara-Medina F, Pérez-Sánchez V, et al. Cost- effectiveness of the 21-gene breast cancer assay in Mexico. Adv Ther. 2015;32(3):239-253. 229. Kalow W, Gunn DR. Some statistical data on atypical cholinesterase of human serum. Ann Hum Genet. 1959;23(3):239-250. 230. Kalow W, Staron N. On distribution and inheritance of atypical forms of human serum cholinesterase, as indicated by dibucaine numbers. CA J Biochem Physiol. 1957;35(12):1305-1320. 231. Evans DAP, Manley KA, McKusick VA. Genetic control of isoniazid metabolism in man. Br Med J. 1960;2(5197):485. 232. Li J, Bluth MH. Pharmacogenomics of drug metabolizing enzymes and transporters: implications for cancer therapy. Pharmgenomics Pers Med. 2011;4:11-33. 233. O’Donnell, Dolan ME. Cancer pharmacoethnicity: ethnic differences in susceptibility to the effects of chemotherapy. Clin Cancer Res. 2009;15(15):4806-4814. 234. Grabinski J, Smith L, Chisholm G, et al. Relationship between CYP2D6 and estrogen receptor alpha polymorphisms on tamoxifen metabolism in adjuvant breast cancer treatment. ASCO Ann Meet Proceed. 2006;24(18 suppl):505. 235. Alcazar-González GA, Calderón-Garcidueñas AL, Garza-Rodríguez ML, et al. Comparative study of polymorphism frequencies of the CYP2D6,

372 Cancer Control October 2016, Vol. 23, No. 4 Various clinical and treatment

factors differently affect sexual

minorities with breast cancer.

Stag Horn. X-Ray image courtesy of Bryan Whitney. www.x-rayphotography.com.

Clinical Considerations of Risk, Incidence, and Outcomes of Breast Cancer in Sexual Minorities Anne E. Mattingly, MD, John V. Kiluk, MD, and M. Catherine Lee, MD

Background: Breast cancer is a leading cause of cancer-related mortality in women. Limited research exists on the impact of sexual orientation on overall risk of and mortality from breast cancer. We sought to summarize the medical literature on breast cancer in sexual minority women and identify possible disparities in this population. Methods: A comprehensive literature search was conducted for English-language studies in peer-reviewed medical journals that referenced breast cancer and sexual minority, lesbian, bisexual, or transgender individuals. Articles published between January 2000 and November 2015 were included. They were reviewed for relevance to breast cancer risk stratification, breast cancer mortality, breast reconstruction, and transgender issues. Results: Behavioral risks, reproductive risks, and risks associated with decreased access to health care may all affect outcomes for sexual minorities with breast cancer. Limited studies have mixed results regarding mortality associated with breast cancer in sexual minorities due to an inconsistent reporting of sexual orientation. Conclusions: Overall, the research examining breast cancer in sexual minority women remains limited. This finding is likely due to limitations in the reporting of sexual orientation within large databases, thus making broader-scale research difficult.

Introduction 3 million women in 2012 were living with breast cancer For American women, breast cancer remains the most in the United States.2 Research has been performed on common cancer and second most common cause of racial disparities in the screening, diagnosis, treatment, cancer-related mortality.1 In 2016, the estimated inci- and survival rates of breast cancer, as well as disparities dence of new breast cancer cases will be 246,660, or due to insurance status and socioeconomic position.3 29.0% of all new cancer cases, and 72,160 cancer-re- Disparity in the risk and mortality rates of breast cancer lated deaths are estimated to be due to breast cancer.1 are likely to continue until we better understand the un- Approximately 1 in 8 women (12.3%) will be diagnosed derlying reasons for the divergence.3,4 with breast cancer during their lifetime, and nearly According to US Census data, 1.2 million individuals live in a household with a same-sex partner.5 The Surveillance, Epidemiology, and End Results (SEER) From the Comprehensive Breast Program, H. Lee Moffitt Cancer Center & Research Institute, Tampa, Florida. Program collects demographical data, including age, Submitted February 19, 2016; accepted May 11, 2016. race, sex, income, education, and geographical loca- Address correspondence to M. Catherine Lee, MD, Comprehensive tion, but it does not collect data on sexual orienta- Breast Program, Moffitt Cancer Center, 109201 North McKinley Drive, tion.2,4 This fact makes research on the disparities in MKC-BR PROG, Tampa, FL 33612. E-mail: [email protected] breast cancer among women who identify as lesbian, No significant relationships exist between the authors and the companies/organizations whose products or services may be referenced bisexual, or sexual minority women difficult. Sexual in this article. minority women are infrequently the focus of stud-

October 2016, Vol. 23, No. 4 Cancer Control 373 ies in breast cancer, and controversy exists on wheth- cult to generalize data as all the respondents were pre- er sexual minority women are at increased risk for menopausal registered nurses, and their work likely breast cancer compared with women in the general affected their interaction with the health care system population.6 and may not have been reflective of other sexual minority women. Methods Breast cancer is much more common in persons The goal of this article is to provide health care profes- as they age, and a premenopausal cohort does not sionals with a discussion framework for their patients give a complete picture of breast cancer either in sex- who are sexual minorities. This is a general overview ual minority or heterosexual women.1 In 2012, Austin of the limited published research on breast cancer in et al9 used the same prospective cohort data as Case sexual minority patients focusing on clinical consider- et al8 to estimate risk of breast cancer in premenopaus- ations. Within the medical literature we identified real women using a modified Rosner–Colditz risk model cent and comprehensive studies to address the follow- and compared sexual minority women with hetero- ing clinical questions: sexual women. Similar to the Gail model, which uses • Are sexual minority women at increased risk for current age, age of menarche, age at first birth, family developing breast cancer? history, previous benign results on breast biopsy, his- • Do sexual minority women have an increased tory of biopsy with atypia, and race to estimate risk of frequency of breast cancer related to a potential breast cancer, the Rosner–Colditz model also includes increase in risk factors? subsequent births following first live birth, type and • Can health care professionals influence patient age of menopause, height, current weight, weight at behavior to mitigate the risk of breast cancer in 18 years of age, and alcohol consumption. Austin et al9 these sexual minority patients? calculated a 1-year breast cancer incidence rate per • Does any identified increased risk lead to a dif- 100,000 person-years and compared sexual minority ference in outcomes for sexual minority women, women with the heterosexual group. specifically breast cancer–related mortality? However, these studies have limitations. The • How do sexual minority women react to breast self-reported nature of sexual orientation makes objec- reconstruction after completing their course of tive study inherently difficult. Oftentimes sexual ori- treatment? entation is not included in large databases of breast • How does the risk of breast cancer change in the cancer research.10 Sexual minority women may be transgender population, specifically when taking less likely to disclose sexual orientation to their health cross-sex hormone therapy? care professional for fear of discrimination.11 As the stigma of sexual minority status has lessened over the Risk Factors decades, it is likely that, as the years go on, more re- Risk factors for breast cancer have been extensively respondents than those who initially self-reported would searched, and studies evaluating risk factors for breast have come forward as a sexual minority in large na- cancer in sexual minority women have been pub- tional surveys of women. For health care profession- lished. However, whether sexual minority women are als, understanding the risk of breast cancer in sexual at an increased risk of breast cancer is still a controver- minority women could help them to educate their pa- sial topic. For example, Cochran et al7 reviewed data tients, thereby improving rates of screening for breast from 7 surveys of women who have sex with women cancer, or to modify behaviors that could have numer- and self-reported bisexual women 18 to 75 years of age ous impacts on patient health in general. and compared their responses to 2 large national sur- From these studies, the risk factors examined can veys of women similar in age (N = 19,000). Case et al8 be categorized as behavioral risks (obesity, alcohol also examined sexual orientation and general health use, tobacco use), reproductive risks (parity, use of risk factors, comparing self-reported sexual minority oral contraceptives), and access to health care (health women to heterosexual female participants in a pro- insurance status, screening).7-9 A summary of selected spective cohort of 116,671 registered nurses 25 to increased risks in sexual minority women are outlined 43 years of age who were followed with biennial sur- in the Table.7-9 veys starting in 1989. Sexual orientation was added to the surveys in 1995, and demographical information Behavioral Risks was collected as well as risk factors for breast can- Obesity is associated with an increased risk of breast cer, other cancers, and cardiovascular disease. Sexual cancer in postmenopausal women and a higher risk of orientation was reported by 98% of respondents, of breast cancer recurrence in premenopausal and post- whom 0.8% identified as lesbian and 0.3% identified as menopausal women.12 Cochran et al7 observed that 28% bisexual — thus making the sexual minority respon- of sexual minority women were classified as obese, sug- dents a smaller subset of the larger cohort.8 It is diffi- gesting that a significantly larger percentage of sexual

374 Cancer Control October 2016, Vol. 23, No. 4 minority women were obese than expected (P < .05). vious smokers than those in the general population Despite this, sexual minority women were significantly (P < .05). Case et al8 found that lesbians and bisexu- less likely to consider themselves obese than women in als reported higher rates of former smoking (60% and the general population (P < .05).7 Compared with het- 50%, respectively), and twice as many sexual minor- erosexuals, lesbians and bisexuals have higher preva- ity women were current smokers. Increased smoking lence rates of obesity (50% and 40%, respectively),8 and in women, including sexual minority women, likely both premenopausal groups have a significantly higher increases their risk for breast cancer as well as other mean body mass index (P < .0001 and P = .0004, respec- medical conditions.1 tively).9 These data suggest that sexual minority women Alcohol has been considered a carcinogen to hu- are more likely than heterosexual women to have chal- mans and has been shown to increase the risk of many lenges with obesity, which is a risk factor for breast cancers, including female breast cancer.14 Cochran et cancer as well as other health problems. al7 did not note any significant differences in current Evidence also suggests that smoking can cause alcohol use among sexual minority women compared breast cancer.13 Cochran et al7 reported that sexual mi- with heterosexual women, but they did find an in- nority women were more likely to be current or pre- creased prevalence of alcohol-related problems, such

Table. — Comparison of Select Studies Evaluating Risk Factors for Breast Cancer in Sexual Minority Women Author Cochran7 Case8 Austin9 Range 1986–1997 1989–1995 1989–2005 Age, y 18–75 25–43 25–42 Type of Heterosexual women = 19,000 Heterosexual women = 89,812 Heterosexual women = 87,392 participant Sexual minority women = 11,886 Lesbian = 694 Lesbian = 665 Bisexual = 317 Bisexual = 309 Risk factor Heterosexual Sexual Heterosexual Lesbian Bisexual Heterosexual Lesbian Bisexual minority Body mass 18.3 kg/m2 27.7 kg/m2 19.8%a 28.1%a 28.2%a 22.04 kg/m2b 23.28 kg/m2b 23.11 kg/m2b index (P < .05) PR: 1.0 PR: 1.5 PR: 1.4 (P < .0001) (P = .0004) (95% CI: 1.4–1.7) (95% CI: 1.1–1.6) Alcohol 4.0%c 12.4%c 4.1%d 9.2%d 11.2%d 3.56 ge 6.18 ge; 5.18 ge; intake (P < .05) PR: 1.0 PR: 1.9 PR: 2.7 (P < .0001) (P < .0001) (95% CI: 1.5–2.3) (95% CI: 2.1–3.5) Current 16.1 21.2 10.6f 18.9f 20.6f smoker, % (P < .05) PR: 1.0 PR: 2.0 PR: 2.2 (95% CI: 1.7–2.3) (95% CI: 1.8–2.7) NA Former 20.1 34.0 23.9g 33.5g 31.5g smoker, % (P < .05) PR: 1.0 PR: 1.6 PR: 1.5 (95% CI: 1.4–1.7) (95% CI: 1.3–1.8) Nulliparity Pregnancyh: Pregnancyh: 22.1% 76.5% 49.5% No. of births: No. of births: No. of births: 66.7% 28.1% PR: 1.0 PR: 3.3 PR: 2.1 1.71 0.38 0.98 Live birthi: (P < .05) (95% CI: 2.9–3.6) (95% CI: 1.8–2.6) (P < .0001) (P < .0001) 56.9% Live birthi: 16.0% (P < .05) History of 79.7 36.2 — — — — — — OCP use, % (P < .05) aBMI ≥ 30.0 kg/m2 compared with < 25.0 kg/m2. bMean BMI premenopause. cHistory of self-reported alcohol-related problems. dAlcohol intake ≥ 15 g/d vs none. eAlcohol intake/d (g) during premenopause. fCurrent smoker vs never smoker. gFormer smoker vs never smoker. hHistory of pregnancy. iHistory of live birth. BMI = body mass index, CI = confidence interval, NA = not applicable, OCP = oral contraceptive pills, PR = prevalence ratio.

October 2016, Vol. 23, No. 4 Cancer Control 375 as history of alcoholism, in sexual minority women cer. At the time of publication, most sexual minority compared with heterosexual women (P < .05). By con- women are unable to marry their partners by law and trast, 2 other studies found significantly increased rates may have had impaired access to a partner’s employ- of alcohol consumption: Twice as many sexual minor- er health insurance.7 Cochran et al7 indicated lower ity women drank at least 15.0 g of alcohol every day, prevalence rates of health insurance coverage (P < .05) and a significantly greater mean daily intake of alco- among sexual minority women compared with stan- hol during premenopause was seen among lesbians dardized estimates, and sexual minority women had (P < .0001) and bisexuals (P < .0001) compared with lower rates of previous mammograms when standard- heterosexual women.8,9 Thus, increased alcohol in- ized for demographics across all age groups (P < .05). take in sexual minority women increases the poten- Current health insurance coverage was associated with tial risk of breast cancer as well as other health prob- a positive history for mammography among sexual mi- lems in this population.1 nority women in their 40s, with 77% of insured sexual minority women vs 64% of uninsured sexual minority Reproductive Risks women having undergone mammography at least once Nulliparity has been associated with an increased risk (P < .05).7 This finding suggests that health insurance of breast cancer, and increased parity can reduce a coverage could improve access to screening mammog- woman’s risk of breast cancer, particularly HR-posi- raphy in sexual minority women, and that sexual mi- tive breast cancer.15 Cochran et al7 demonstrated that nority women are less likely to have health insurance sexual minority women had lower lifetime rates of coverage. However, Case et al8 found no significant pregnancy and live births when compared against difference between sexual minority women compared national data (P < .05 for both). Furthermore, subse- with heterosexual women in regard to history of mam- quent studies demonstrate that nulliparity is 3 times mography. Likely, this is an evolving issue. With the re- more likely in lesbians and twice as likely in bisexual cent legalization of same-sex marriage, access to em- women, with significantly reduced number of births ployer-sponsored health care insurance may change in lesbian and bisexual women compared with het- for sexual minority women. erosexual women (P < .0001 for both).8,9 These data Despite the increased individual risk factors for suggest that sexual minority women are significantly breast cancer described in sexual minority women, more likely to be nulliparous, thus increasing their no definitive studies have demonstrated an increased risk of breast cancer. prevalence of breast cancer in sexual minority wom- If sexual minority women did have children, Case en.7-9 Using the Rosner–Colditz risk model, Austin et al8 found that they were twice as likely to have given et al9 predicted an incidence rate of breast cancer sig- birth before 19 years of age, a factor associated with nificantly elevated in both lesbian and bisexual wom- a decreased risk of breast cancer,16 conferring up to a en (P < .0001 for both), and this rate was consistently 50% risk reduction compared with nulliparous wom- predicted across all age groups. Regression analysis en.17 However, Austin et al9 found no significant differ- found lesbian and bisexual women were more likely to ence in age at first birth between lesbian and bisexual be at higher risk than heterosexual women of similar women compared with heterosexual women. This is age, race, geography, and income. a confounding factor when attempting to elucidate re- However, sexual minority women have not been productive risk factors for breast cancer in sexual mi- shown to have an increased rate of breast cancer when nority women. compared with the general population. Cochran et al7 Use of oral contraceptives is associated with a were unable to link sexual minority women with an decreased risk of HR-positive breast cancer.15 Cochran increased prevalence of breast cancer compared with et al7 noted that sexual minority women reported signif- US women in general. A self-reported history of breast icantly lower rates of oral contraceptive pill use than the cancer among sexual minority women was 0.9%, general female population (P < .05). Although Austin et which was concordant with expected standardized al9 did not specifically look at this factor, they did study estimates, and no statistically significant difference in age of menarche and found a 20% increase rate of men- self-reported prevalence of breast cancer between the arche at 11 years of age or younger in lesbian women; study sample of sexual minority women and US esti- however, subsequent data found no significant differ- mates for women was observed.7 Therefore, sexual mi- ence between lesbian and bisexual women compared nority women may not have an increased incidence or with heterosexual women. Thus, these reproductive prevalence rate of breast cancer despite their increased risk factors deserve further study. predicted risk.

Access to Health Care Modulating Risk Health insurance status is important when discuss- Based on identified individual risk factors, attempts are ing preventive care such as screening for breast can- warranted by health care professionals to modify pa-

376 Cancer Control October 2016, Vol. 23, No. 4 tient behavior in the hopes that doing so will improve ing participation in breast cancer screening for sexual the health of sexual minority patients. It is important minority women is important. Sexual minority women to note that many of the risk factors for breast cancer have consistently lower lifetime prevalence rates of identified in sexual minority women are modifiable screening mammography than expected across all age behavioral risks.7-9,12,14 groups, although this is likely confounded by their de- Bowen et al18 attempted to modulate risk by increased access to health care.7 creasing screening participation in sexual minor- Improving the participation rates of breast cancer ity women using group counseling. This prospective screening among sexual minority women are impor- study randomized sexual minority women to a coun- tant, but opportunities still exist to modulate other seling (n = 69) or control group (n = 81). The interven- modifiable high-risk behaviors identified in sexual tion included psychosocial counseling with education, minority women, such as alcohol intake, obesity, and group discussion, and skills training during 4 weekly smoking.7-9 Health care professionals should be aware 2-hour sessions led by a trained health counselor spe- of these modifiable risk factors associated with breast cializing in sexual minority women. Weekly sessions cancer and other health problems, as they may present had themes that included a personalized breast cancer an educational opportunity to promote healthy behav- risk assessment, discussions of mammography, recom- iors. Clinicians should also be encouraged that inten- mendations for self-examinations and clinical breast sive education is successful at promoting behavioral examination, stress management, progressive muscle change in sexual minority women and improving their relaxation and guided imagery, and social support in participation with breast cancer screening.18 4 main areas (emotional, informational, instrumental, and companionship) focused on devising strategies for Mortality increasing support where needed.18 The goals of reducing risk and modifying behavior At 6 and 24 months, participants were evaluated on should be to lower the mortality rate associated with perceived risk of breast cancer, worry of cancer, men- breast cancer and promote healthy living. Cochran tal health, “outness,” and identification with the lesbian and Mays6 examined the mortality risk of breast can- community.18 At 24 months, the intervention group re- cer among sexual minority women using the National ported a higher rate of breast cancer screening in the Health Interview Survey, which is an annual popula- past 2 years compared with the control group (87% vs tion-based, household interview of married or cohabi- 75%; P < .05). The intervention group also had a signif- tating women 18 to 80 years of age that was conducted icantly higher rate of breast self-examinations at 6 between 1997 and 2003. Of the 155,427 respondents, (P < .01) and 24 months (P < .05) than the control group.18 693 reported having female partners (0.5%).6 Sexual The intervention group changed more over time in minority women were younger (P < .001), had higher perceived risk (P < .001), cancer worry (P < .001), and levels of education (P < .001) and higher family in- mental health (P < .01).18 Regression analysis showed comes (P < .05), were less likely to have health insur- identification with the lesbian community and feelings ance (P < .05), and were more likely to be in a cohabitat- of outness were predictors of participation in screening ing relationship rather than be married (P < .001).6 The mammography and a feeling of outness was also a signif- mortality rate of breast cancer was identified using the icant predictor of performing a breast self-examination. National Death Index: Of the 4,396 identified deaths, With education and group counseling structured 274 (3.1%) were attributed to death from breast cancer.6 for sexual minority women, these researchers demon- Age, race, education, family income, and health insur- strated that trained health counselors can successfully ance coverage were strongly associated with death increase participation of sexual minority women in from any cause (P < .001 for all).6 After adjusting for breast cancer screening.18 However, it is worth noting confounding variables, sexual minority women were that this study has limitations because it studied 1 vol- not at increased risk of overall mortality; however, unteer group in a single community, thus making it dif- sexual minority women had an elevated risk for breast ficult to generalize to all sexual minority women. cancer–related mortality compared with heterosexu- It is encouraging to think that an intensive, de- al women (P < .05).6 Overall, sexual minority women mographical directed counseling intervention can im- were 3 times more likely to die from breast cancer af- prove the participation of sexual minority women in ter adjusting for age, although the researchers did not breast cancer screening.18 However, without a hetero- evaluate the prevalence of breast cancer in the sexual sexual cohort, it is difficult to know whether behavior minority population. The data are limited because they modification is specific to sexual minority women or only include women who were in a defined relation- generalizable to others. Although it is not reasonable ship and, thus, the data cannot be generalized to sexu- to assume that a health care professional would have al minority women without partners. the resources to educate his or her sexual minority pa- In interpreting potential reasons for this dif- tients to such a degree that this study entails, improv- ference in mortality rates, a review of the risk fac-

October 2016, Vol. 23, No. 4 Cancer Control 377 tors previously discussed is important. Is there an Breast Reconstruction increased rate of incidence or prevalence of breast A unique aspect of the treatment for female breast can- cancer not yet identified in sexual minority women cer is the impact that surgery, including breast recon- because of reproductive and behavioral risks that struction, can potentially have on a patient’s self-iden- leads to this increased mortality risk of breast cancer? tity.20 In general, breast cancer treatment is dictated Perhaps the increased mortality risk of breast cancer by tumor biology, particularly with regard to systemic among sexual minority women could be due to later therapy; sexual orientation has no impact on chemo- diagnoses because of decreased access to health care therapeutic or hormonal therapy recommendations.1 due in part to lack of insurance or decreased rates of However, surgery for breast cancer is the cornerstone screening. of breast cancer management in many American wom- Cochran and Mays19 attempted to validate their en.1 Decision-making regarding breast reconstruction findings by reviewing mortality risks associated with in these patients is not well understood, but it has the sexual orientation both in women and in men using potential for disparate care among women.20 data from the General Social Surveys and compar- In particular, breast reconstruction is scrutinized ing mortality status using the same National Death because of the presumed heterosexual focus of the Index. Respondents were interviewed face-to-face breast, as well as an example of health care access.20 in their household and then were self-administered A paucity of literature exists on use of reconstruction questionnaires or computer-assisted self-interviews; options in sexual minority women. It is possible that these responses were matched to National Death In- sexual orientation and self-identity could have signifi- dex records.19 Sexual orientation was classified based cant impact on decisions regarding type of oncologi- on lifetime numbers of male or female sexual part- cal surgery and whether or not a patient would desire ners and the gender of those sexual partners in the reconstruction. Sexual minority women may perceive year prior to the interview. Of the 17,886 respon- that health care professionals are not always consider- dents who completed the interview/questionnaire, ing sexual orientation when making recommendations 853 (4.8%) reported a history of same-sex partners. for surgery or reconstruction. Results were then subdivided by race, education, in- The scientific literature on sexual minority wom- come, geographical region, overall self-rated health en and their decisions to undergo breast reconstruc- status, and generalized self-reported happiness. tion is limited: A single, retrospective, qualitative study Similar to the previous study, sexual minority wom- has explored decisions by sexual minority women for en were younger and reported higher education6; or against reconstruction.20 A total of 15 self-reported however, in this cohort, sexual minority women had sexual minority women diagnosed with nonrecurrent lower household incomes and, geographically, more and nonmetastatic breast cancer who completed inva- sexual minority women lived in western states.19 sive treatment within the past 5 years completed demo- All-cause mortality among women was significantly graphical surveys and a disclosure scale. These partici- associated with older age, non-white race, and lower pants were separated into breast reconstruction and no education (P < .001 for all), but sexual minority status breast-reconstruction groups. Semi-structured inter- was not significantly correlated with all-cause mor- views covered diagnosis, treatment course, treatment tality or breast cancer, even after multivariate surviv- decision-making, sources of social support, and adjust- al analysis (all-cause mortality, P = .72; breast cancer ment; participants identified support persons who were mortality, P = .27).19 Overall, 9.1% of sexual minor- then interviewed covering similar topics. Comparisons ity women died from breast cancer vs 4.4% of hetero- of demographics, clinical stage, receipt of adjuvant ther- sexual women.19 Thus, Cochran and Mays19 were un- apy, and time since diagnosis revealed that the groups able to replicate their 2013 findings,6 and it is unclear were similar. All support persons were female, and the whether sexual minority women have an increased majority of support persons were partners. Mastectomy risk of mortality from breast cancer. was recommended for some, whereas others opted for The later study benefited from a much broader mastectomy to avoid radiotherapy; one-half of partici- definition of sexual orientation (it included lifetime pants underwent breast conservation before undergoing partners), which helped capture people not in cur- mastectomy.20 Of those in the reconstruction group rent same-sex relationships.19 However, sexual ori- (n = 8), 3 had autologous reconstruction, 5 had im- entation remains difficult to objectively measure due plant-based reconstruction, 3 received chemothera- to inherent self-reporting bias in identifying sexual py, 3 received radiotherapy, and 6 received hormon- minority status. Overall, the later study reported al therapy; of those in the no-reconstruction group more same-sex partners (4.8% vs 0.5%), but this is (n = 7), 4 received chemotherapy, 2 received radio- still a relatively low percentage of respondents — therapy, and 4 received hormonal therapy.20 possibly underpowering the study.6,19 Thus, further The researchers discovered that breast size was a research is necessary. consideration among the participants when deciding

378 Cancer Control October 2016, Vol. 23, No. 4 whether to accept or reject reconstruction.20 Sexual It is also worth underscoring that the notion that minority women with small breasts felt more comfort- women may regret undergoing reconstruction after able rejecting reconstruction, whereas sexual minor- mastectomy may not be limited to the sexual minority women with larger breasts feared being unbalanced ity community. Metcalfe et al21 prospectively followed and having a heavy prosthesis. Both groups mentioned patients after mastectomy with and without delayed the importance of the breast for their self-image but breast reconstruction and found that psychosocial emphasized physical functioning and body strength functioning improves over time in patients following over aesthetics. Sexual minority women emphasized mastectomy. During the long-term follow-up, women that their sexual minority status contributed to them with delayed breast reconstruction had significantly not being defined by the presence of their breasts, and higher levels of total distress (P = .01), obsessiveness they noted a discrepancy between their values and (P = .03), and cancer-related distress (P = .02) com- body image and those of mainstream society — this pared with those who underwent mastectomy alone, was especially true when communicating with spe- although no differences were seen in quality of life be- cialists in breast cancer. Sexual minority women who tween the 2 groups.21 Metcalfe et al22 published anoth- decided against reconstruction emphasized strength, er report that demonstrated no difference in body im- long-term health, and survival; they complained less of age for women with mastectomy alone compared with physical problems but did discuss having to adjust to mastectomy and immediate breast reconstruction. their changed body shape; they did not voice doubts or They further found that women with delayed breast re- regrets about their choice. Their partners emphasized construction had higher levels of body stigma (P = .01), body image and their value system as motivators, and body concerns (P = .002), and concerns regarding the they confirmed the same satisfaction with respect to obviousness of mastectomy-related changes in appear- body image. They characterized making the decision ance (“transparency”; P = .001) than mastectomy alone jointly, and they reported attraction and desire being or mastectomy with immediate reconstruction.22 present as before.20 Another study reported on women who under- Sexual minority women who underwent recon- went breast-conserving surgery, mastectomy alone, struction reported physical issues, including numbness and mastectomy with reconstruction, and it found and limited range of motion; 1 woman reversed her re- that women who underwent mastectomy with recon- construction completely.20 The group expressed doubt struction had greater mood disturbance and poorer or regret about their choice and lack of information well-being (P = .002 for both) compared with those about potential complications. Some opted for recon- who underwent mastectomy alone.22 This finding struction to cover the physical effects of the cancer and persisted up to 18 months after surgery. regain a “normal” appearance, although some chose to All of these studies suggest the potential for forego formal nipple reconstruction.20 Their partners selection bias for women who undergo breast recon- confirmed that outcomes of the reconstruction had not struction — in particular, delayed breast reconstruc- been thoroughly explained by the plastic surgeon, and tion — as pursuing that option because of their un- they echoed doubts about reconstruction; their part- derlying distress regarding their body image.20-22 It is ners were characterized as more passive in the deci- difficult to generalize this information to sexual mi- sion-making process, and these couples displayed dis- nority women, but these issues of dissatisfaction may cordance in values and body image.20 Women without not be unique to either population. It is important for a partner at the time of their decision were influenced health care professionals to understand that sexual by their relationship status, and support persons not in minority women may have a different perspective on a partner relationship were more reluctant to express breast reconstruction after mastectomy regardless of reconstructive preferences. their surgical choice for or against reconstruction. Overall, sexual minority women who underwent Boehmer et al22 described sexual minority women reconstruction expressed doubt and regret; by con- as not being defined by having breasts, emphasizing trast, sexual minority women who elected against re- body strength and physical functioning over aesthetics, construction did not, and this finding was echoed by and “otherness,” described as a discrepancy of values their support persons.20 However, this study has limi- and body image among sexual minority women when tations, including its small convenience sample, retro- compared with the general population. Overall, health spective recall bias, and, thus, low applicability to the care professionals must be aware of a patient’s sexual general population.20 This study also has no compari- orientation and include the patient’s support persons son to a heterosexual control, making it difficult to de- or partners in discussions regarding treatment. Health fine the themes as unique to sexual minority women, care professionals should also consider how the pres- but it did outline potential themes that may arise for ence or absence of the breast might affect the patient’s sexual minority women as they embark on their deci- body image, understanding that sexual minority wom- sion for or against reconstructive surgery.20 en may not feel emotionally attached to the physical

October 2016, Vol. 23, No. 4 Cancer Control 379 presence of their breasts. Health care professionals dence of breast cancer in transgender men and women should also discuss physical functioning after breast treated within the VHA. Using encounter and prescrip- reconstruction and not limit the discussion to aesthet- tion data from 1996 to 2013, transgender-related diag- ics alone. Reconstructive options should be clearly and nosis and breast cancer events were identified within comprehensively presented, elucidating realistic ex- the VHA system.25 Medical records were reviewed for pectations for patients. any patients with both a transgender-related diagnosis and a breast cancer event; cross-sex hormone therapy Transgender Persons prescribed by a VHA physician was classified as andro- The term transgender defines a person whose gen/anabolic or estrogen. The incidence rates of breast self-identity does not conform to gender norms, and cancer were compared with Surveillance, Epidemiolo- many patients may choose to use medical or surgical gy, and End Results data from 2007 to 2011 using over- means to live as another gender. The risk of breast can- all observed and expected rates.25 cer is particularly relevant due to cross-sex hormone Of approximately 6 million veterans seen every therapy for many patients who desire to feminize or fiscal year, 5,135 had a transgender-related diagnosis, masculinize their appearance. 1,579 were female at birth, and 3,556 were male at birth.25 In general, research on breast cancer in the trans- No statistically significant difference was observed in gender community is limited, and our understanding breast cancer incidence among women, men, and in the is also limited with regard to how hormonal therapy overall population. For the entire sample, 10 confirmed affects a person’s risk of breast cancer. Sattari et al23 cases of breast cancer were found in transgender veterans summarized the cases of male-to-female transgender (0.2%), 7 in female-to-male veterans, 2 in male-to-female patients with breast cancer published in the literature veterans, and 1 in a natal male veteran with transvestic since 1968. The authors identified 10 cases of breast fetishism.25 The average age at diagnosis was 63.8 years cancer in male-to-female transgender persons, all of and all participating veterans racially self-identified as whom received long-term cross-sex hormone treatment; non-Hispanic white. Breast cancer surgery was per- 7 of the cases had ER status reported (n = 5 negative, formed in 8 transgender veterans; 4 patients died. Of n = 2 positive).23 Thus, breast cancer in general appears those patients not receiving cross-sex hormone thera- to be rare among male-to-female transgender persons, py, breast cancer was diagnosed in 3 transgender men although it is difficult to make any generalizations and 1 transgender woman. For those patients receiving about the risks of cross-sex hormone therapy in this cross-sex hormone therapy breast cancers occurred transgender population. in 3 transgender women, and 3 cases of female breast Understanding the risk of breast cancer in female- cancer occurred in transgender veterans who received to-male transgender persons who take testosterone for cross-sex hormone therapy many years following their masculinization is important, because it is likely dif- diagnosis of breast cancer.25 ferent from the risk of breast cancer in male-to-female More than one-half (52%; n = 2,645) of transgender transgender persons. Circulating testosterone is con- veterans received at least 1 prescription for cross-sex verted to estradiol by aromatase enzyme, and higher hormone therapy, with the average duration of treatment levels of circulating testosterone could increase pe- being 6.5 years and the average age at initial treatment ripheral estrogen. Shao et al24 described 2 cases of fe- being 48.9 years.25 None of the males prescribed cross- male-to-male transgender persons, both of whom took sex hormone therapy developed breast cancer, suggest- testosterone for masculinization. One patient was post- ing that male breast cancer is rare in the transgender menopausal (aged 48 years) when testosterone was population receiving cross-sex hormone therapy. Breast initiated, was also taking an aromatase inhibitor for cancers occurred in 3 transgender female veterans re- symptoms of endometriosis, and was diagnosed with ceiving cross-sex hormone therapy (n = 2 received es- breast cancer 5 years later; the other case patient was trogen products, n = 1 received testosterone products) premenopausal (aged 21 years) when testosterone prior to diagnosis.25 In general, no significant differ- was initiated and then diagnosed with breast cancer ence was observed for transgender patients receiving 6 years later. Both patients had ER-positive breast can- estrogen, testosterone, or cross-sex hormone therapy cer, and both tested negative for a deleterious BRCA when compared with the overall cohort.24 Transgender mutation.24 These cases are worth noting, but they veterans receiving cross-sex hormone therapy had no do not provide conclusive evidence for an increased increased risk of breast cancer, and the overall inci- risk of breast cancer in transgender persons receiving dence of breast cancer in transgender persons was low; cross-sex hormone therapy. however, late presentations in the 3 male patients with A study by the US Veterans Health Administration breast cancer and the high mortality rate seen in trans- (VHA) is the largest and most comprehensive research gender veterans are concerning.25 on breast cancer in the transgender community (as of Clinicians should consider breast and chest exami- publication).25 This research has evaluated the inci- nations and possible mammography, if indicated, in this

380 Cancer Control October 2016, Vol. 23, No. 4 subset of patients. History of chest-contouring surgery cer, many researchers have looked into specific groups or mastectomy does not preclude the development of of women in an effort to identify and reduce disparities breast cancer, as observed in 1 female-to-male trans- in breast cancer among different populations of wom- gender patient in this study.25 This study did not include en. Numerous racial and socioeconomic disparities gene mutation data because they were not readily avail- have already been identified, but research on sexual able, but it is the largest series of transgender patients orientation and disparities in breast cancer is still lim- with breast cancer reported in the literature to date.25 ited and the data contradictory.3 Some evidence sup- Interpretation of the use of cross-sex hormone therapy ports that sexual minority women have higher rates may be limited because the study required that at least of risk factors associated with breast cancer.6,7 Studies 1 prescription was obtained through the VHA system are conflicting on which risk factors are more com- to identify the patient as receiving cross-sex hormone mon in sexual minority women, but such risk factors therapy.24 For example, if a patient received cross-sex consistently noted to be increased in sexual minority hormone therapy outside of the VHA system, then the women include nulliparity, a history of smoking, and therapy may not have been recorded.25 increased alcohol consumption.7-9 Published data also Typically, male breast cancer is hormonally sensi- showed lower rates of health insurance coverage and tive (92%).2,26 As data have emerged and are now as- decreased rates of breast cancer screening in sexual sociating female breast cancer with use of perimeno- minority women.6,7 Despite an increase in risk fac- pausal estrogen-based hormone replacement therapy, tors, no clear increased incidence or prevalence rate concern exists about the potential risks of estrogen for breast cancer in sexual minority women has been therapy in male-to-female transgender persons.27 How- identified,7,8 and research on breast cancer in the trans- ever, male breast cancer is rare in the transgender gender population remains limited. Utilizing educa- population.25 At this time, the evidence is insufficient tion and structured support groups for sexual minor- to support an increased incidence of breast cancer in ity women help to improve rates of participation in transgender male-to-female persons, even when tak- breast cancer screening,18 and health care professionals ing cross-sex hormone therapy. should consider breast and chest examinations as well Risk of breast cancer may be different in the fe- as mammography in transgender patients. male-to-male transgender population compared with Clinicians should also be aware of the sexual ori- the general population. In female-to-male persons re- entation of their patients and the potential associa- ceiving testosterone cross-sex hormone therapy, theo- tion with disparities in breast cancer and differences retical concern of breast cancer is extrapolated from in related outcomes. Recent legislation supporting the data on perimenopausal estrogen-based hormone same-sex marriage may improve the documentation of therapy.27 The theory postulates that the metabolism of sexual orientation, so health care professionals should testosterone could increase peripheral estrogen levels, be aware of their patient’s sexual orientation and in- and, as reported in Shao et al,24 2 such patients both clude support persons in treatment discussions, as developed ER-positive breast cancer. However, only needed, to ensure they understand the patient’s goals 1 female-to-male patient receiving testosterone cross- while counseling them on treatment options, includ- sex hormone therapy in the VHA transgender cohort de- ing breast reconstruction. Consistent documentation of veloped breast cancer.25 Although transgender females gender preferences as part of the medical record could are at an increased risk of breast cancer due to their sex improve the delivery of personalized care. Large-scale at birth compared with male-to-female transgender pa- databases should consider adding sexual orientation tients, the data are lacking for us to conclude whether to their demographical data to allow for more robust an increased risk of breast cancer exists in transgender research. With these considerations in mind, research- female-to-male patients receiving cross-sex hormone ers should consider standardizing terminology regard- therapy compared with the general female population. ing sexual minority patients to allow for more cohesive Moreover, the data do not support an increased inci- dialogue in the scientific literature. dence of breast cancer in transgender persons receiving cross-sex hormone therapy, and, in general, breast can- References cer in transgender persons is rare or under-reported.23 1. American Cancer Society. Cancer Facts & Figures 2016. http://www. cancer.org/acs/groups/content/@research/documents/document/acspc- However, the scientific research related to transgender 047079.pdf. Accessed October 3, 2016. persons is lacking, making conclusions difficult to draw. 2. National Cancer Institute. Surveillance, Epidemiology, and End Results Program. SEER Stat Fact Sheets: Female Breast Cancer. http://seer.cancer. It does highlight the fact that more research is needed. gov/statfacts/html/breast.html. Accessed July 1, 2016. 3. Bigby J, Holmes MD. Disparities across the breast cancer continuum. Cancer Causes Control. 2005;16(1):35-44. Conclusions 4. Brown JP, Tracy JK. Lesbians and cancer: an overlooked health Breast cancer is the most common cancer and second disparity. Cancer Causes Control. 2008; 19(10):1009-1020. 5. Mayer KH, Bradford JB, Makadon MD, et al. Sexual and gender most common cause of cancer-related death in Ameri- minority health: what we know and what needs to be done. Am J Pub Health. can women.1 Given the pervasive nature of breast can- 2008;98(6):989-995.

October 2016, Vol. 23, No. 4 Cancer Control 381 6. Cochran SD, Mays VM. Risk of breast cancer mortality among women cohabiting with same sex partners: findings from the National Health Interview Survey, 1997-2003. J Womens Health (Larchmt). 2012;21(5):528-533. 7. Cochran SD, Mays VM, Bowen D, et al. Cancer-related risk indicators and preventative screening behaviors among lesbians and bisexual women. Am J Public Health. 2001;91(4): 591-597. 8. Case P, Austin SB, Hunter DJ, et al. Sexual orientation, health risk factors, and physical functioning in the Nurses’ Health Study II. J Womens Health (Larchmt). 2004;13(9):1033-1047. 9. Austin SB, Pazaris MJ, Rosner B, et al. Application of the Rosner- Colditz risk prediction model to estimate sexual orientation group disparities in breast cancer risk in a U.S. cohort of premenopausal women. Cancer Epidemiol Biomarkers Prev. 2012;21(12):2201-2208. 10. Meads C, Moore D. Breast cancer in lesbians and bisexual women: a systematic review of incidence, prevalence and risk studies. BMC Public Health. 2013;13:1127. 11. Hart SL, Bowen DJ. Sexual orientation and intentions to obtain breast cancer screening. J Womens Health (Larchmt). 2009;18(2):177-185. 12. Jain R, Strickler HD, Fine E, et al. Clinical studies examining the impact of obesity on breast cancer risk and prognosis. J Mammary Gland Biol Neoplasia. 2013;18(3-4):257-266. 13. Reynolds, P. Smoking and breast cancer. J Mammary Gland Biol Neoplasia. 2013;18(1):15-23. 14. Scoccianti C, Lauby-Secretan B, Bello PY, et al. Female breast cancer and alcohol consumption: a review of the literature. Am J Prev Med. 2014;46(3 suppl 1): S16-S25. 15. Anderson KN, Schwab RB, Martinez ME. Reproductive risk factors and breast cancer subtypes: a review of the literature. Breast Cancer Res Treat. 2014;144(1):1-10. 16. Hunter DJ, Speigelman D, Adami HO, et al. Non-dietary factors as risk factors for breast cancer and as effect modifiers of the association of fat intake and risk of breast cancer. Cancer Causes Control. 1997;8(1):49-56. 17. Kobayashi S, Sugiura H, Ando Y, et al. Reproductive history and breast cancer risk. Breast Cancer. 2012;19(4):302-308. 18. Bowen DJ, Powers D, Greenlee H. Effects of breast cancer risk counseling for sexual minority women. Health Care Women Int. 2006;27(1):59-74. 19. Cochran SD, Mays VM. Mortality risks among persons reporting same-sex sexual partners: evidence from the 2008 General Social Survey- National Death Index data set. Am J Public Health. 2015;105(2): 358-364. 20. Boehmer U, Linde R, Freund KM. Breast reconstruction following mastectomy for breast cancer: the decisions of sexual minority women. Plast Reconstr Surg. 2007;119(2):464-472. 21. Metcalfe KA, Zhong T, Narod SA, et al. A prospective study of mastectomy patients with and without delayed breast reconstruction: long-term psychosocial functioning in the breast cancer survivorship period. J. Surg. Oncol. 2015;111(3):258-264. 22. Metcalfe KA, Semple J, Quan ML, et al. Changes in psychosocial functioning 1 year after mastectomy alone, delayed breast reconstruction, or immediate breast reconstruction. Ann Surg Oncol. 2012;19(1):233-241. 23. Sattari M. Breast cancer in male-to-female transgender patients: a case for caution. Clin Breast Cancer. 2015;15(1):e67-e69. 24. Shao T, Grossbard ML, Klein P. Breast cancer in female-to-male transsexuals: two cases with a review of physiology and management. Clin Breast Cancer. 2011;11(6): 417-419. 25. Brown GR, Jones KT. Incidence of breast cancer in a cohort of 5,135 transgender veterans. Breast Cancer Res Treat. 2015;149(1):191-198. 26. Ruddy KJ, Winer EP. Male breast cancer: risk factors, biology, diagnosis, treatment, and survivorship. Ann Oncol. 2013;24(6):1434-1443. 27. Women’s Health Initiative Study Group. Risks and benefits of estrogen plus progestin in healthy postmenopausal women: principal results from the Women’s Health Initiative randomized controlled trial. JAMA. 2002;288(3):321-333.

382 Cancer Control October 2016, Vol. 23, No. 4 A large racially and ethnically diverse

database was created and TMAs

of IDC were constructed to represent

African American, Hispanic, and

non–Hispanic-white patient populations.

Proteus. X-Ray image courtesy of Bryan Whitney. www.x-rayphotography.com.

Construction and Validation of a Multi-Institutional Tissue Microarray of Invasive Ductal Carcinoma From Racially and Ethnically Diverse Populations Edward Seijo, MS, Diana Lima, MPH, Egiebade Iriabho, MS, Jonas Almeida, PhD, Jesus Monico, MPH, MS, Margarita Echeverri, PhD, Sylvia Gutierrez, MD, Idhaliz Flores, PhD, Ji-Hyun Lee, PhD, Kate Fisher, MA, William E. Grizzle, MD, PhD, Gabriel L. Sica, MD, PhD, Charles Butler, Chindo Hicks, PhD, Cathy D. Meade, PhD, RN, Stephen Olufemi Sodeke, Krzysztof Moroz, MD, Domenico Coppola, MD, and Teresita Muñoz-Antonia, PhD

Background: The scarcity of tissues from racial and ethnic minorities at biobanks poses a scientific constraint to research addressing health disparities in minority populations. Methods: To address this gap, the Minority Biospecimen/Biobanking Geographic Management Program for region 3 (BMaP-3) established a working infrastructure for a “biobanking” hub in the southeastern United States and Puerto Rico. Herein we describe the steps taken to build this infrastructure, evaluate the feasibility of collecting formalin-fixed, paraffin-embedded tissue blocks and associated data from a single cancer type (breast), and create a web-based database and tissue microarrays (TMAs). Results: Cancer registry data from 6 partner institutions were collected, representing 12,408 entries from 8,279 unique patients with breast cancer (years 2001–2011). Data were harmonized and merged, and deidentified information was made available online. A TMA was constructed from formalin-fixed, paraffin-embedded samples of invasive ductal carcinoma (IDC) representing 427 patients with breast cancer (147 African Americans, 168 Hispanics, and 112 non-Hispanic whites) and was annotated according to biomarker status and race/ethnicity. Biomarker analysis of the TMA was consistent with the literature. Conclusions: Contributions from participating institutions have facilitated a robust research tool. TMAs of IDC have now been released for 5 projects at 5 different institutions.

From the Tissue Core Facility (ES), Departments of Thoracic Oncology (DL), Health Outcomes and Behavior (CM), and Pathology (CD), Biostatistics and Bioinformatics Core (KF), and Tumor Biology Program (TM-A), H. Lee Moffitt Cancer Center & Research Institute, Tampa, Florida, Pathology Informatics Core (EI, JA) and Pathology Program for Translational Research in Neoplasia (WEG), University of Alabama at Birmingham, Birmingham, Alabama, Departments of Pathology (JM) and Medicine and Cancer Institute (CH), University of Mississippi Medical Center, Jackson, Mississippi, Center for Minority Health and Health Disparities Research and Education, Xavier University of Louisiana (ME) and Department of Pathology, Tulane University School of Medicine (KM), New Orleans, Louisiana, Department of Anatomy (SG) and Depart- ment of Microbiology (IF), Ponce Health Sciences University, Ponce, Puerto Rico, Department of Internal Medicine, University of New Mexico, Albuquerque, New Mexico (J-H L), Department of Pathology and Laboratory Medicine, Winship Cancer Institute and Emory School of Medicine, Atlanta, Georgia (GLS, CB), and Tuskegee University National Center for Bioethics in Research and Health Care, Tuskegee, Alabama (SOS). Submitted March 3, 2016; accepted May 31, 2016. Address correspondence to Teresita Muñoz-Antonia, PhD, Tumor Biology Program, Moffitt Cancer Center, 12902 Magnolia Drive, MRC-3 WEST, Tampa, FL 33612. E-mail: [email protected] This work was supported in part by the Tissue Core Facility and Bioinformatics and Biostatistics Core at Moffitt Cancer Center, a National Cancer Institute–designated Comprehensive Cancer Center (P30-CA076292).

October 2016, Vol. 23, No. 4 Cancer Control 383 Introduction eastern region, consists of Alabama, Florida, Georgia, The burden of cancer disparities on African Ameri- Louisiana, Mississippi, and Puerto Rico, and it included can and Hispanic populations has been well docu- 9 partner institutions that created a collaborative net- mented.1-8 It is our opinion that discoveries in precision work. A key goal for BMaP-3 was to develop a state- medicine and basic science with the potential to im- of-the-art network that could provide a foundation and prove the prognosis or treatment of cancer in affected infrastructure to adequately and continuously supply individuals might be limited by the scarcity of research high-quality, multiethnic, human biospecimens for biospecimens from racial/ethnic minority populations. cancer research. Since 2015, the National Cancer Insti- The availability of diverse samples is key to inclusive tute has reorganized and consolidated the regions into precision medicine.9 For example, the triple-negative 6 regions, and the institutions previously in region 3 breast cancer subtype, defined by the absence of ER, are now, at the time of publication, in Geographical PR, and ERBB2 (formerly HER2 or HER2/neu) expres- Management of Cancer Health Disparities region 2.15 sion, accounts for 10% to 20% of all breast cancer cases, In 2014, members of the participating BMaP-3 in- with an uneven distribution among ethnicities.10 In Af- stitutions completed a comprehensive assessment tool rican American women, triple-negative breast cancer that revealed a diversity of samples by race/ethnicity is nearly twice as prevalent as it is in European Ameri- and cancer site were available as formalin-fixed, par- cans, with some studies reporting triple-negative sub- affin-embedded (FFPE) tumor samples or fresh-frozen types in up to 40% of all breast cancer cases among samples (from 6 of 9 institutions).16 In addition, 5 of the premenopausal African Americans.10,11 institutions had a tissue advisory/steering committee, Consistent with this observation, epidemiological thus providing the necessary infrastructure and exper- studies have shown substantial differences in rates of tise to build a network approach. This provided the ba- breast cancer mortality between racial groups.5,8,12 For sis for the design of a BMaP-3 pilot study to test the example, premenopausal African American women feasibility of developing and implementing a regional, have a lower 5-year breast cancer–related survival rate web-based database of available biospecimens and than European American women (79.0% vs 91.0%, re- collecting FFPE blocks to create unique tissue microar- spectively), placing them at a higher risk of mortality rays (TMAs) for health disparities in cancer research. despite an overall lower incidence of breast cancer.12 This pilot study represents a regional team, sci- This disparity may be due to the increased frequency ence-oriented approach for constructing and validating of triple-negative breast cancer observed in African multi-institutional breast cancer TMAs from a racially American women who typically present at a later dis- and ethnically diverse population. Using the construc- ease stage and are more likely to have lymph-node me- tion and validation of a breast cancer TMA as an ex- tastases at similar tumor sizes.13 ample, we illustrate how a regional team, science-ori- Study findings have revealed that Hispanic women ented approach can effectively leverage the tissue and have a higher prevalence of ER-negative tumors com- database resources of multiple institutions to facilitate pared with non-Hispanic women (36.2% vs 22.7%; tissue-based studies for a minority cohort — a difficult P = .05) and an unexpectedly high proportion of challenge for single institutions alone. ERBB2-positive tumors compared with non-Hispan- ic women (31.9% vs 14.3%; P < .01).14 These findings, Materials and Methods coupled with more recent data, suggest that consid- The research conducted in this study was performed erable molecular differences in cancers may exist in with tissue samples and patient data unlinked from pa- various ethnic groups.10 Because molecular features tient identifiers. Approval or waiver of informed con- can be used for predicting cancer prognosis and ther- sent was obtained by the Institutional Review Boards apeutic response, research focused on understanding of the 6 institutions contributing FFPE biospecimens the molecular features of cancers occurring in ethnic and data to support this pilot study. minorities is important. Collection and Management of Data Purpose Cancer registry coding discrepancies were harmonized To address the lack of tissue-based studies for ethnic by adopting a standardized coding for each variable minorities with cancer, the Biospecimen/Biobanking and systematically replacing prior coding with cur- Geographic Management Program (BMaP) was ini- rent coding using software R v2.15.2 (R Foundation for tiated by the National Cancer Institute’s Center to Re- Statistical Computing, Vienna, Austria). Patient cases duce Cancer Health Disparities. BMaP coordinated were removed from the data sets for any of the follow- regional efforts focused on health disparities in can- ing reasons: younger than 18 years of age at diagnosis cer research, training, and care and was organized (n = 1), tumor marker and site-specific factor did not into 6 geographical regions within the United States. agree for either ER or PR status (n = 157), had a Span- BMaP region 3 (BMaP-3), which makes up the south- ish surname and unknown “Spanish Hispanic” vari-

384 Cancer Control October 2016, Vol. 23, No. 4 able (n = 17), or assigned a study identification number captured hematoxylin and eosin stained images were that already existed (n = 1). The original site-specific made available to investigators at all participating in- factor 16 was captured after the year 2010; therefore, stitutions. The biomarker status obtained from cancer the site-specific factor was manually derived from the registry data was validated with immunostains for es- tumor marker 1 ER assay, tumor marker 2 PR assay, trogen receptor (ER), progesterone receptor (PR), and and ERBB2 variables, with assignments of either HR+/ erb-b2 receptor tyrosine kinase 2 (ERBB2) using the ERBB2+, HR+/ERBB2–, PR–/ER–/ERBB2+, or PR–/ER–/ following antibodies from Ventana Medical Systems ERBB2–. If the original site-specific factor 16 disagreed (Tucson, AZ): 790-4324 CONFIRM anti-ER (SP1), 790- with the derived site-specific factor, then the entry was 2223 CONFIRM anti-PR (1E2), and 790-100 PATHWAY deemed unreliable and was deleted (n = 19). anti-HER2/Neu (4B5). Antigen retrieval and incubation times were optimized for each antibody as follows: Multi-Institutional, Web-Based Database 60-minute retrieval for ER and PR antibodies; 30-min- To ensure no changes were made to the data set origi- ute retrieval for ERBB2 antibodies; 32-minute incuba- nally merged by the biostatistician, the web-based tion for the ER antibody; and 16-minute incubation for data set was exported and a comparison of the web- the PR and ERBB2 antibodies. based data set against the original merged data set The Ventana Benchmark XT (Tucson, AZ) platform was executed. Counts were also verified between was used for all immunohistochemistry analyses. The both data sets. The aggregated data were posted on Allred scoring system was used to score the ER and PR a website (http://labpages.moffitt.org/bmap3/Can- stains.17 ERBB2 stains were scored in accordance with cer%20Database/Cancer_Database.html [access re- guidelines from the College of American Pathologists, stricted]) in the form of an output of descriptive sta- which defines ERBB2-positive status by immunohis- tistics generated by SPSS (IBM, Armonk, NY). BMaP-3 tochemistry as protein overexpression.18 The HR rules public access database (PAD) has logging features en- for the categorization of cases were: if either ER or PR is abled so that, in the event of improper data modifica- positive, then HR+; if both are negative, then HR–; and if tion, the user account responsible, the data that were both ER and PR are missing, or if one is missing and the changed, and the time of the change can be identified other is negative, then it was labeled as “missing.” and corrected in a timely manner. The governance model for the BMaP-3 PAD consists of 3 hierarchical Results user-access levels (administrator, data manager, and Building an Infrastructure investigator), ranging from full to search access. To ensure equitable governance for issues concerning human biospecimen research sponsored by BMaP-3, a Archival Case Selection and Tissue Microarray Tissue Advisory Board was formed with representatives Construction from all 9 BMaP-3 institutions. A collaboration agree- Each archival tissue sample was categorized according ment, signed by members of each of the 9 partner in- to tumor expression of ER, PR, and ERBB2. For each stitutions, allowed data and tissue sharing between the case, optimal areas for inclusion as cores in the TMA partner institutions regardless of the ability of any indi- were marked on the slides stained with hematoxylin vidual institution to contribute samples. Once the infra- and eosin by pathologists. When sufficient tumor tis- structure was established, the goal was to evaluate the sue was available, 3 or fewer 1-mm cores from each se- feasibility of collecting FFPE tissue blocks and associ- lected donor block were punched and put into 8 TMAs ated data and creating a web-based database and TMAs. 60 × 10 mm in size using Tissue Arrayer (Beecher In- The first step in this pilot study was to choose a tu- struments, Sun Prairie, WI). Each TMA also included mor type for the TMA. The BMaP-3 investigators chose 1 sample of normal breast tissue. Cores from kidney, breast cancer as the highest priority to address health spleen, and tonsil, as well as melanoma, human breast disparities related to cancer in the southeastern United cancer lines (MCF-7 and MDA231), human prostate ade- States and Puerto Rico. This decision was based on a nocarcinoma cells (LNCaP), and human ovarian carcinoma variety of factors, including unpublished observations cells (SK-OV-3), were included as reference cells/tissues. on the scarcity of breast TMAs from diverse populations described in the literature, lack of commercially Immunohistochemical Staining and Evaluation available breast TMAs with associated patient data on All TMA slides were stained with hematoxylin and race and ethnicity, the observed rates of incidence and eosin and processed for whole-slide imaging. Each mortality associated with breast cancer at the 6 con- TMA slide was scanned using ScanScope XT (Aperio, tributing cancer sites, and the unpublished results Vista, CA) with a 20×/0.8 numerical aperture objective of a multi-institutional survey of researcher needs lens at a rate of 10 minutes per slide. Image analyses for TMAs.16 In addition, researchers specializing in for stained TMAs were performed using Nuclear v9.1 breast cancer were surveyed to determine what type (Aperio) to segment nuclei of various intensities. The of breast cancer TMA is most needed for research. As

October 2016, Vol. 23, No. 4 Cancer Control 385 shown in Fig 1, 29 of the 61 respondents (47.5%) ex- HR+/ERBB2–, HR+/ERBB2+, ER–/PR–/ERBB2+). The pressed greatest need for a TMA based on biomarker demographical and clinical characteristics of the data status between different racial/ethnic groups, followed from the breast cancer cases obtained from the 6 insti- by a TMA based on 1 histological type of breast tis- tutions are presented in Table 2. Because cancer reg- sue in different racial/ethnic groups (26%), and a TMA istry standards for ERBB2 were largely adopted from based on 1 tumor, node, and metastasis pathology 2006 onward, many of the archival cases (collected stage group in different racial/ethnic groups (14%). from 2001 to 2005) were categorized as missing for ERBB2 status. Therefore, the Tissue Advisory Board Collection and Management of Cancer decided that the 3 biomarker categories for selecting Registry Data cases would be as follows: (1) triple negative, (2) HR+/ Deidentified American College of Surgeons cancer reg- ERBB2–, and (3) ERBB2+ (HR+/ERBB2+ and ER–/PR–/ istry data from each institution served as the basis for ERBB2+ grouped together). All unique cases of female the web-based database and for the selection of cases to be included in the TMA. Requests for data from cancer registries at each of the 9 partner institutions speci- Table 1. — Data Elements and Their Descriptions fied the following criteria for selecting cases: date range, Data Element Description 2001 to 2011 (to avoid using biospecimens that might Study identifier Unique code used by honest broker for be needed for clinical care), age at diagnosis (≥ 18 years reidentification of age), primary breast cancer only (no metastasis), and Year first seen Year of first patient contact (inpatient or first-course surgery at the institution. In addition, be- outpatient), with the reporting facility for the cause cases from minority populations were a priority diagnosis of the tumor, its treatment, or both for inclusion in the TMA, members at partnering insti- Site primary Code for the primary site of the tumor being tutions were asked to specifically retrieve patient data reported using ICD-O (2nd or 3rd ed.) from African Americans and Hispanics/Latinos. The Age at diagnosis Records the age of the patient in complete 17 data elements requested are described in Table 1. years at his or her last birthday before diagnosis Cancer registry data sets were retrieved from 6 institutions, and a biostatistician harmonized the data Sex Identifies the sex of the patient sets. A total of 12,408 entries from 8,279 unique pa- Race Identifies the primary race of the patient tients with breast cancer pathology were confirmed Spanish/Hispanic Identifies patients of Spanish or Hispanic after merging the data sets. Deidentified information origin origin about breast cancer cases collected from the different Histology First 4 digits of the morphology code that (ICD-O [3rd ed.]) describe the tumor/cell type cancer registries is available in BMaP-3 PAD (https:// apps.mathbiol.org/bmap [access restricted]). Behavior Code for the behavior of the tumor being (ICD-O [3rd ed.])a reported Year of surgery Year in which a surgical procedure was Archival Case Selection and Tissue Microarray performed Construction Surgery of primary Type of surgery performed to breast aimed at The data set was further analyzed for cases with inva- site, at hospital modifying, controlling, removing, or destroying sive ductal carcinoma (IDC; histology selected by the cancerous tissue Tissue Advisory Board) and then cross-tabulated by Systemic/surgery Records the sequencing of systemic treat- race/ethnicity and biomarker status (triple-negative, sequence ment and surgical procedures given as part of first-line therapy Surgery/radiation Records the sequencing of radiation and sequence surgical procedures provided as part of first- TMA based on 1 TNM stage group in different racial/ethnic groups line therapy TMA of 1 histological type of breast Collaborative stage, Combinations of ER, PR, and ERBB2 results tissue in different racial/ethnic groups site-specific factor 16 TMA based on biomarker status between different racial/ethnic groups Tumor marker 1 ER assay Other Tumor marker 2 PR assay 0 5 10 15 20 25 30 35 40 45 50 ERBB2 Interpretation of results (not the same as the Respondents, % laboratory value) for the ERBB2 test used to determine the status of ERBB2 Fig 1. — Results of the breast cancer researcher survey (N = 61). A multi- TNM, pathology, The anatomical extent of disease based institutional, web-based poll was distributed to investigators in region 3 stage group on TNM elements of the tumor following (currently region 2) institutions to assess the research interest regarding surgery different kinds of breast TMAs. Survey responses were anonymous. As a can be seen, the greatest need was for TMA based on biomarker status Behavior is the fifth digit of the ICD-O (3rd ed.) morphology code. between different racial and ethnic groups. ICD-O = International Classification of Diseases for Oncology, TNM = tumor, node, metastasis. TMA = tissue microarray, TNM = tumor, node, metastasis.

386 Cancer Control October 2016, Vol. 23, No. 4 IDC were tabulated according to biomarker status institutions, and 427 blocks from 427 unique patients (site-specific factor 16) and by race/ethnicity (Table 3). with breast cancer were obtained. Tumor expressions Although a large portion of cases were missing data, of ER, PR, and ERBB2 were annotated based on can- this analysis revealed that, in this cohort, African cer registry data in the following manner: (1) HR+ (ER+, American women have a higher incidence of triple- PR+, or both) and ERBB2–, (2) HR+ (ER+, PR+, or both) negative breast cancer compared with white women, and ERBB2+, and (3) HR– (ER– and PR–) and ERBB2–. and Hispanic women have a higher incidence of HR- These 427 FFPEs were used to construct the TMA negative breast cancer compared with non-Hispanic blocks, of which 8 blocks were constructed. TMA women. These data, as well as data on breast cancer slides from each of the blocks were stained for ER, PR, cases not included in the TMA, were incorporated into and ERBB2 to confirm the biomarker status of the sam- the BMaP-3 PAD web-based database. ples. Fig 2 shows a sample representation of tumors FFPE blocks were requested from participating from each of the 3 types of breast TMAs (ERBB2+, ERBB2–, and triple negative) for each stain. Thereafter, the cancer registry data were reviewed, and 4 samples Table 2. — Characteristics of Patients at the originally classified as triple negative were reclassified Contributing Institutions (N = 8,279) as either ERBB2+ or HR+. Variable No. of Patients, n (%) Several discrepancies were found between the cancer registry data and the TMA immunohistochem- Sex istry results. Specifically, for all racial/ethnic groups, Male 59 (0.72) the biomarker status obtained from the cancer regis- Female 8,219 (99.30) try data was 100% in concordance to the staining re- Race sults for ERBB2– cases. However, this was not true for American Indian 2 (0.02) the triple-negative cases (86.8% and 85.7% rates of ac- Asian Indian 1 (0.01) curacy for African Americans/whites and Hispanics, respectively), and concordance was much lower in Asian Other 1 (0.01) the case of ERBB2+ staining (43.6% and 30.6% for Af- Black/African American 2,012 (24.30) rican Americans/whites and Hispanics, respectively). Filipino 2 (0.02) This finding could be due to different factors such as Pacific Islander 3 (0.04) tumor heterogeneity or loss of antigenicity (eg, some Vietnamese 3 (0.04) donor blocks were several years old).19 These stains White 5,904 (71.30) were scanned and quantified, and links to the images Other 80 (0.97) were made available upon request. Detailed information on the samples included on the TMA is posted on Unknown 271 (3.30) http://labpages.moffitt.org/bmap3/Cancer%20Database Ethnicity /Cancer_Database.html. Cuban 43 (0.52) Dominican Republic 7 (0.08) Discussion Mexican 24 (0.29) This pilot study successfully accomplished the follow- Puerto Rican 248 (3.00) ing goals: South/Central American 66 (0.80) • Creation of a regional infrastructure with an organizing framework for communications, processes Spanish (not otherwise specified) 368 (4.40) for memorandums of understanding and Institu- Other Spanish 9 (0.12) tional Review Board approval from multiple insti- Non-Spanish 7,370 (89.00) tutions, and an established Tissue Advisory Board Spanish surname only 2 (0.02) and related guidelines Unknown 142 (1.70) • Processes for retrieval and synchronization of data Age, y from multiple cancer registries into a single data set < 25 23 (0.28) • Development of a collaborative model for optimizing TMA design considerations 25–34 222 (2.70) • Development and implementation of a web-based 35–44 1,126 (13.60) database of cancer registry cases 45–54 2,217 (26.80) All of these accomplishments demonstrate the 55–64 2,267 (27.40) overall feasibility of creating a unique, centralized, 65–74 1,588 (19.20) public resource for multiethnic biobanking and the ≥ 75 836 (10.10) collection of biospecimens. Given the large number of cases included in the web-based database and the

October 2016, Vol. 23, No. 4 Cancer Control 387 Table 3. — Site-Specific Factor 16 for Invasive Ductal Carcinoma by Race and Ethnicity for Women in All Institutions Combined HR+/ERBB2– HR+/ERBB2+ PR–/ER–/ERBB2– PR–/ER–/ Missing Totals ERBB2+ Race, n (%) African American 169 (15) 58 (5) 130 (11) 33 (3) 755 (66) 1,145 White 585 (18) 91 (3) 163 (5) 43 (2) 2,295 (72) 3,177 Other/missing 99 (36) 15 (6) 58 (21) 26 (9) 77 (28) 275 Totals 853 (19) 164 (3) 351 (8) 102 (2) 3,127 (68) 4,597 Ethnicity, n (%) Hispanic 128 (25) 21 (4) 97 (19) 28 (6) 234 (46) 508 Non-Hispanic 711 (18) 139 (3) 251 (6) 73 (2) 2,835 (71) 4,009 Other/missing 14 (17) 4 (5) 3 (4) 1 (1) 58 (73) 80 Totals 853 (19) 164 (3) 351 (8) 102 (2) 3,127 (68) 4,597 amount of TMAs, this model demonstrates that no in- Biomarker Antibodies Status stitution could have accomplished these outcomes ERBB2 PR ER alone. A primary benefit of this approach is that mem- Triple bers of all the partner institutions are eligible to access negative the data and tissue resources regardless of their capabilities to contribute to the database or TMA. The first step in this pilot study was to select a ERBB2+ tumor type for creation of the TMAs. We chose to focus on breast cancer as the highest priority to address health disparities related to cancer in the southeastern United States and Puerto Rico for a variety of reasons.16 ERBB2– This decision was based on1,10,13,14,20-22: • Analysis of the literature on TMAs from patients with different racial/ethnic backgrounds showing a scarcity of breast TMAs from diverse populations Fig 2. — Sample representation of immunostaining according to • Lack of commercially available breast TMAs that biomarker status. TMA slides were immunostained with anti-ER, provide associated data on race and ethnicity anti-PR, and anti-ERBB2 antibodies. • Observed incidence and mortality rates associated ER = estrogen receptor, ERBB2 = erb-b2 receptor tyrosine kinase 2, with breast cancer at each of the 6 contributing PR = progesterone receptor, TMA = tissue microarray. cancer sites by race/ethnicity • Data from the comprehensive assessment tool on breast cancer among Hispanic women compared with FFPE biospecimens for 6 cancer types available for non-Hispanic women. These observations are consis- study at the 9 study institutions tent with the literature.10,14 However, our analysis is • Survey results of researcher needs for TMAs compromised by a large amount of missing data. Hines This pilot study demonstrated the feasibility of col- et al14 showed that breast cancer among Hispanic wom- lecting cancer registry data for more than 12,000 breast en comprises a distinct spectrum of tumor subtypes cancer cases and creating a web-based database. We compared with non-Hispanic white women. In that were successful in creating a racially/ethnically diverse study, a diverse TMA based on biomarkers was creat- TMA from FFPEs representing 427 breast cancer cases. ed, allowing researchers to study potential molecular For validation of the TMAs, we focused on studying differences in triple-negative breast cancer in different the characteristics of IDC among different racial/ethnic racial/ethnic groups.14 groups because the investigators from the BMaP- Although most institutions had compatible bioin- 3 institutions identified these as an important area of formatics systems for the management of biospecimen investigation, given the known differences in incidence data, these systems were used for fresh frozen samples. of triple-negative IDC among diverse populations. No single institution had a direct database for associat- The results of the biomarker analysis showed a ed FFPE tissue data; thus, cancer registry data and pa- higher incidence rate of triple-negative breast cancer thology reports were the sources of data for cases like- among African American women compared with white ly to have tissues stored at the respective institution. women and a higher incidence rate of HR-negative When needed, data would have had to be abstracted

388 Cancer Control October 2016, Vol. 23, No. 4 from pathology reports — which was beyond the re- 2011;127(3):729-738. 3. Partridge EE, Fouad MN, Hinton AW, et al. The deep South network sources of this project. Our findings suggest that the for cancer control: eliminating cancer disparities through community-aca- idea of synchronizing pathology reports with cancer demic collaboration. Fam Community Health. 2005;28(1):6-19. 4. Ward E, Jemal A, Cokkinides V, et al. Cancer disparities by race/ registry data should be pursued because pathology re- ethnicity and socioeconomic status. CA Cancer J Clin. 2004;54(2):78-93. ports offer more complete and accurate information 5. Ademuyiwa FO, Edge SB, Erwin DO, et al. Breast cancer racial disparities: unanswered questions. Cancer Res. 2010;71(3):640-644. on biomarker data than is currently provided by a can- 6. Evens AM, Antillón M, Aschebrook-Kilfoy B, et al. Racial disparities cer registry. in Hodgkin’s lymphoma: a comprehensive population-based analysis. Ann Oncol. 2012;23(8):2128-2137. 7. Fedewa SA, Ward EM, Stewart AK, et al. Delays in adjuvant che- Conclusions motherapy treatment among patients with breast cancer are more likely in African American and Hispanic populations: a national cohort study We have created a model for collaboration. Through 2004-2006. J Clin Oncol. 2010;28(27):4135-4141. the contribution of formalin-fixed, paraffin-embed- 8. Tamimi RM, Hughes ME, et al. Racial and ethnic differences in breast cancer survival: mediating effect of tumor characteristics and sociodemo- ded blocks from the Minority Biospecimen/Biobank- graphic and treatment factors. J Clin Oncol. 2015;33(20):2254-2261. ing Geographic Management Program for region 9. Tian Q, Price ND, Hood L. Systems cancer medicine: towards realiza- tion of predictive, preventive, personalized and participatory (P4) medicine. 3 (BMaP-3) institutions, this project facilitated the J Intern Med. 2012;271(2):111-121. creation of a robust research tool that would not be 10. Carey LA, Perou CM, Livasy CA, et al. Race, breast cancer subtypes, and survival in the Carolina Breast Cancer Study. JAMA. readily available if each institution worked indepen- 2006;295(21):2492-2502. dently. Moreover, each institution benefited from the 11. Kohler BA, Sherman RL, Howlader N, et al. Annual report to the nation on the status of cancer, 1975-2011, featuring incidence of breast collaboration and from the team approach to tackle a cancer subtypes by race/ethnicity, poverty, and state. J Natl Cancer Inst. research/clinical issue common to all team members. 2015;107(6):djv048. 12. U.S. Department of Health and Human Services; National Institutes It is expected that investigators at the partner institu- of Health; National Cancer Institute. Surveillance, Epidemiology, and End tions will utilize the web-based database on an on- Results (SEER) 2010. http://seer.cancer.gov/. Accessed March 24, 2016. 13. McBride R, Hershman D, Tsai WY, et al. Within-stage racial differ- going basis for data analysis. This database is freely ences in tumor size and number of positive lymph nodes in women with available with open access for the BMaP-3 investiga- breast cancer. Cancer. 2007;110(6):1201-1208. 14. Hines LM, Risendal B, Byers T, et al. Ethnic disparities in breast tors. We fully expect that the model and process for tumor phenotypic subtypes in Hispanic and non-Hispanic white women. the work described herein could be applied and ex- J Womens Health (Larchmt). 2011;20(10):1543-1550. 15. National Cancer Institute. Geographical Management of Cancer ported to other multi-institutional networks focused Health Disparities (GMaP). Updated August 1, 2016. http://www.cancer. on a participatory approach to biobanking. gov/about-nci/organization/crchd/inp/gmap. Accessed August 19, 2016. 16. Wells KJ, Lima DS, Meade CD, et al. Assessing needs and assets We also demonstrated the feasibility of creating for building a regional network infrastructure to reduce cancer related tissue microarrays from racially and ethnically diverse health disparities. Eval Program Plann. 2014;44:14-25. 17. Allred DC, Carlson RW, Berry DA, et al. NCCN task force report: es- populations through the BMaP-3 partnership. This trogen receptor and progesterone receptor testing in breast cancer by im- work was made possible by the infrastructure of this munohistochemistry. J Natl Compr Canc Netw. 2009;(7 suppl 6):S1-S21. 18. Wolff AC, Hammond ME, Schwartz JN, et al. American Society partnership, and it can serve as a model for other multi- of Clinical Oncology/College of American Pathologists guideline recom- institutional networks focused on research utilizing mendations for human epidermal growth factor receptor 2 testing in breast cancer. J Clin Oncol. 2007;25(1):118-145. human biospecimens. To date, these tissue microar- 19. Miller RT, Swanson PE, Wick MR. Fixation and epitope retrieval in rays of invasive ductal carcinoma have been released diagnostic immunohistochemistry: a concise review with practical considerations. Appl Immunohistochem Mol Morphol. 2000;8(3):228-235. for 5 projects at 5 different institutions. 20. Wu Y, Sarkissyan M, Elshimali Y, et al. Triple negative breast tumors Acknowledgments: The authors would like to in African-American and Hispanic/Latina women are high in CD44+, low in CD24+, and have loss of PTEN. PLoS One. 2013;8(10):e78259. thank Ivana Sehovic and Zena Sayegh for managing 21. Lindner R, Sullivan C, Offor O, et al. Molecular phenotypes in triple and constructing the BMaP-3 TMA blocks. Marithea negative breast cancer from African American patients suggest targets for therapy. PLoS One. 2013;8(11):e71915. Goberville, PhD, provided editorial support. 22. Winter JL, Stackhouse BL, Russell GB, et al. Measurement of PTEN Substantial support was provided by Region 3 expression using tissue microarrays to determine a race-specific prognostic marker in breast cancer. Arch Pathol Lab Med. 2007;131(5):767-772. GMaP/BMaP partner institutions: Emory Winship Can- cer Center, Moffitt Cancer Center, Morehouse School of Medicine, Ponce Health Sciences University, Tulane University, Tuskegee University, University of Alabama at Birmingham, University of Mississippi Medical Cen- ter, and Xavier University of Louisiana. This work was supported in part by National Cancer Institute grants 3U56 CA118809-14S2; 3U54 CA153509-02S3 and 3U54 CA153509-03S1 (Moffitt Cancer Center).

References

1. Lisovicz N, Johnson RE, Higginbotham J, et al. The Deep South Network for cancer control. Building a community infrastructure to reduce cancer health disparities. Cancer. 2006;107(8 suppl):1971-1979. 2. Ooi SL, Martinez ME, Li CI. Disparities in breast cancer characteristics and outcomes by race/ethnicity. Breast Cancer Res Treat.

October 2016, Vol. 23, No. 4 Cancer Control 389 Data suggest that immediate action

must be taken to improve the time

between diagnosis and treatment

for disadvantaged patients.

Lily. X-Ray image courtesy of Bryan Whitney. www.x-rayphotography.com.

Determinants of Treatment Delays Among Underserved Hispanics With Lung and Head and Neck Cancers Evelinn A. Borrayo, PhD, Katie L. Scott, PhD, Ava R. Drennen, PhD, Tiare MacDonald, PhD, and Jennifer Nguyen, MS

Background: Evidence is lacking to explain the reasons why treatment is delayed among disadvantaged His- panic patients diagnosed with lung and head and neck cancers. Data indicate that treatment delays beyond 46 days increase the risk of death for individuals with these cancers. Methods: A mixed-methods design was used to explore determinants of treatment delays by analyzing data from the medical records of 40 Hispanic patients and data from interviews with 29 Hispanic patients, caregivers, health care professionals, and patient navigators from a safety-net hospital. Results: Of the 40 Hispanic patients, 35% initiated treatment 46 days or more after being diagnosed, but women experienced longer delays than men (average of 48 days). Women with few comorbid diseases (≤ 4) were more likely to experience treatment delays. Institutional-related determinants at publicly funded hospitals appear to delay treatment for patients at the safety-net hospital, and patient-related determinants common to underserved patients (eg, no health insurance coverage) are likely to further contribute to these delays. Conclusions: Delayed treatment is associated with poor outcomes and low rates of survival in patients with lung and head and neck cancers. Therefore, action should be taken to improve the time between diagnosis and the initiation of treatment for disadvantaged Hispanic patients. Introduction From the Department of Psychology (EAB, TM, JN), Colorado State University, Colorado School of Public Health, University In the United States, lung cancer accounts for more of Colorado, Denver, Colorado, Department of Neurology (KLS), deaths than any other cancer in both men and wom- Spectrum Health Medical Group, Grand Rapids, Michigan, and en.1 Among Hispanics, lung cancer is the leading the Department of Behavioral Health (ARD), Salud Family Health Centers, Fort Collins, Colorado. cause of cancer-related death for men and the second- 2 Submitted May 5, 2016; accepted July 28, 2016. leading cause for women. Those with lung cancer are Address correspondence to Evelinn A. Borrayo, PhD, Department also at increased risk for developing secondary neo- of Psychology, Colorado State University, 1876 Campus Delivery, plasms of the head and neck and vice versa.3,4 Lung Fort Collins, CO 80523. E-mail: [email protected] and head and neck cancers are largely induced by to- No significant relationships exist between the authors and the companies/organizations whose products or services may be ref- bacco use, which accounts for 87% and 70% of lung erenced in this article. cancer–related deaths for men and women, respec- This work was supported by a grant from the National Cancer tively.5 Approximately 75% of head and neck cancers Institute at the National Institutes of Health (U01, CA114604). Its are attributed to tobacco use, and a 30-fold increased contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Cancer risk is seen among those who smoke and drink alco- Institute. hol.1,6 Moreover, those with lung and head and neck

390 Cancer Control October 2016, Vol. 23, No. 4 cancers who smoke cigarettes and drink alcohol are who received transitional care had the greatest in- at increased risk for developing comorbid conditions, creases in time to treatment initiation.20 In a follow-up such as chronic bronchitis, emphysema, and cardio- study using the same database, Murphy et al17 found vascular disease.7 Although Hispanics have lower in- that the time to treatment initiation of 46 to 52 days cidences of and mortality rates from lung and head was associated with a significant increased risk of and neck cancers than non-Hispanic whites, dispari- death — a risk that was consistently detrimental be- ties in timely and quality treatment remain for His- yond 60 days. After examining covariates, the re- panics — poverty is typically cited as the explaining searchers also found that the effect of time to treat- factor.1 Mounting evidence demonstrates that worse ment initiation on mortality risk was greater for outcomes (eg, survival, quality of life) for ethnic mi- early-stage (stage I/II) than advanced-stage disease norities with cancer are also best explained by their (stage III/IV) and for radiotherapy and surgery than low socioeconomic status (SES).1,8 Similarly, inequali- for chemoradiotherapy.17 ties exist in properly treating medically underserved To assess the impact of delayed treatment for lung patients (eg, low-income, uninsured) with lung and cancer on an international level, Vinas et al21 con- head and neck cancers.9-11 ducted a systemic review of multinational studies and In a systemic review and meta-analysis of socio- found that long delays were predictive of better surviv- economic inequalities and lung cancer treatment, For- al rates than shorter delays, but patients who experi- rest et al9 found that lower SES was associated with a enced longer delays had fewer symptoms — possibly statistically significant reduced likelihood of receiving explaining a better prognosis. Instead, short delays in any treatment — particularly surgery and chemother- lung cancer treatment had a negative prognostic mean- apy. These SES inequalities could not be accounted for ing because urgent treatment was provided to patients by differences in stage at presentation or by differences with more advanced disease.21 in health care systems (ie, publicly funded vs private Patients with lung and head and neck cancers in- health care systems).9 Choi et al10 conducted a longi- creasingly experience longer delays between their tudinal study to examine disparities as predictors of diagnoses and subsequent treatment.20,21 Delays are survival among patients with head and neck cancers more prevalent for patients with early-stage disease, and found that low income, low education, and ad- thus increasing their disease progression, their wait vanced age were significant independent predictors of times (if candidates for surgical treatment), and their poor survival (ie, 5-year overall, cancer-specific, and level of distress.18,19,22-25 Instead, shorter delays are ex- disease-free survival rates; P < .001). Similarly, Yang perienced by patients with more advanced-stage dis- et al11 found in a population-based study that signifi- ease for whom treatment is urgently initiated without cant independent predictors of worse survival rates clear benefits in their disease progression or rate of among patients with lung cancer were older age and survival.17,21 A paucity of evidence explains what fac- severe poverty. The effects of ethnicity were statistical- tors drive the treatment delays experienced by patients ly significant: A higher percentage of Hispanic patients with lung and head and neck cancers. Murphy et al20 resided in the highest poverty areas, were less likely to contend that the delay in the start of treatment is due receive radiotherapy, and had shorter survival rates if to increases in more sophisticated, pretreatment diag- they smoked compared with their non-Hispanic white nostics, complex multimodal therapies, and transitions counterparts.11 in care for patients treated at academic cancer centers. However, these factors are not as likely to explain treat- Treatment Initiation ment delays for patients treated at community-based More than 10 years ago, the Institute of Medicine rec- hospitals.20 ommended improving the timeliness of treatment by reducing wait times that can potentially result in harm- Community-Based Facilities ful delays for patients.12 Timeliness in the treatment of Few studies have compared the intervals from diag- lung and head and neck cancers is important, because nosis to treatment among academic and public hospi- delaying the start of treatment has been associated tals that treat head and neck cancers; however, among with disease recurrence,13-16 lower rates of survival,17 those that have, significantly longer treatment delays and poor patient outcomes (eg, distress).18,19 have been observed at public hospitals.26,27 In a study Using the US National Cancer Database, Murphy by Yorio et al,26 age, sex, race, and type of insurance et al20 found that the interval from diagnosis to the coverage were not significantly associated with a lon- start of treatment has been increasing since 1985 for ger time to treatment. The single significant factor they Americans diagnosed with head and neck cancers. found for the delay of lung cancer treatment was re- Their examination of predictors found that patients ceiving that care at a public hospital. Patel et al28 sug- with advanced-stage disease who underwent chemo- gest that the factors that affect the timely delivery of radiotherapy, were treated at academic facilities, and cancer treatment are likely institution-related factors

October 2016, Vol. 23, No. 4 Cancer Control 391 that generally apply to public institutions, including variables that procure or impede a patient’s initiation availability of specialty services, adequacy of equip- of treatment (eg, health insurance, income); and pre- ment, timely pathological analysis, and lack of imaging disposing factors include personal propensities to ac- and surgical procedures — this is particularly true for cess treatment (eg, age, sex, language).30,31 patients with head and neck cancers. Measures Patient Factors Data were extracted from the medical records of His- Less is known about patient-related factors that might panic patients with a primary diagnosis of lung cancer contribute to delays in cancer treatment, but they are or head and neck cancer treated at a “safety-net” hospi- likely related to low SES, limited access to transporta- tal in Colorado across a 1-year period. The patient data tion, lack of paid leave from work, language barriers, included sociodemographic variables (eg, age, sex), di- comorbid medical conditions, and homelessness.28 agnosis (eg, cancer type, stages I–IV), cancer treatment (eg, chemotherapy, surgery), number of comorbid dis- Purpose eases (using the Charlson Comorbidity Index), date To our knowledge, no studies have investigated the of cancer diagnosis (by diagnostic biopsy), and date factors at the institutional and patient levels that are when the patient began any type of cancer treatment.32 likely to contribute to treatment delays among un- The interval between the initial diagnosis and the start derserved Hispanic patients with lung and head and of treatment was calculated as the number of days be- neck cancers treated at community-based hospitals, tween the date when patients were diagnosed and the let alone studies that have focused on underserved date when they began their first curative treatment. For Hispanic patients who experience high rates of mor- patients diagnosed at the time of resection and surgery bidity from these cancers. We conducted this explor- was their first treatment, the number of days was cod- atory study to better understand the institution- and ed as 1 day. the patient-level determinants associated with the timely initiation of cancer treatment among under- Data Analysis served Hispanic patients diagnosed with lung and For medical record data, a stepwise, multiple regres- head and neck cancers. sion method was used to evaluate whether predisposing, enabling, and need factors were necessary to pre- Methods dict the range of days between initial diagnosis and Study Design initiation of treatment. Data were centered by convert- A triangulation, mixed-method design (ie, quantitative ing age, cancer stage at diagnosis, and comorbid dis- and qualitative data linked to the literature) was used eases into z scores. Regression assumptions and diag- for breadth and depth of understanding and corrobo- nostics were satisfactory, except for homoscedasticity, ration from multiple angles to determine the initiation which led to the removal of a single influential case. of cancer treatment among the study population.29 We Ethnographical content analysis was utilized to extract collected quantitative data from medical records to as- relevant themes from the semi-structured and focus- sess the magnitude of the influence of patient variables group interviews.33 and diagnostic factors on treatment initiation as well The goal of the analysis was to identify factors that as qualitative research to explore stakeholder perspec- detail the delays in receiving cancer treatment from the tives and other influential factors not captured by the perspective of patients, caregivers, HCPs, and patient data in the medical records. One-on-one interviews navigators. Five analysts were trained and employed were also conducted with patients who had lung can- in this study to follow the content analysis methods cer and head and neck cancers and their health care of Krippendorff.34 All analysts independently coded professionals (HCPs), while focus-group interviews each transcript and then met to corroborate the codes, were conducted with the caregivers and patient naviga- broaden the subject categories, and identify quotes to tors who assisted the study patients. illustrate the themes.

Theoretical Framework Participants The Behavioral Model for Vulnerable Populations was The patients included in this study were treated at a used as the guiding theoretical framework.30,31 The safety-net hospital that serves impoverished community model proposes that need, enabling, and predispos- neighborhoods. ing factors facilitate or block the use of health services Medical Records: Medical records were reviewed by vulnerable populations. Need factors include the for 53 Hispanic patients with a primary diagnosis of vulnerabilities that make patients necessitate medical lung or head and neck cancer; of those, 40 patients treatment (eg, cancer type and stage, comorbid diseas- had medical records of the number of days from diag- es); enabling factors include social or environmental nosis to treatment. Table 1 displays the characteristics

392 Cancer Control October 2016, Vol. 23, No. 4 for all 40 patients. Table 1. — Select Characteristics of Study Patients Patients were an average of 57 years of age (standard de- Men Women (n All Participants F viation [SD] 16.3; range, 23–88); Variable (n = 23) = 17) (N = 40) Statistic 57% were diagnosed with head Mean (SD) and neck cancers and 43% were Age, y 61 (14.5) 52 (17.6) 57 (16.3) 2.95 diagnosed with lung cancer. The Diagnosis to treatment delay, dª 34 (23.1) 48 (42.8) 40 (33.2) 1.76 average number of days from No. of comorbid diseases 5 (4.4) 4 (2.9) 5 (3.8) 0.51 diagnosis to treatment was 40 n (%) n (%) n (%) ² (SD 33.2; range, 1–141); 65% re- X ceived treatment in less than 42 Sociodemographic Characteristics days, 18% between 46 and 59 Place of Birth 0.05 days, and 17% received treat- United States 13 (57) 9 (53) 22 (55) ment after more than 60 days. Latin America 10 (43) 8 (17) 18 (45) The number of comorbidities in Language 0.06 the study patients ranged from 2 to 17 conditions (mean [SD] English 14 (58) 11 (56) 25 (62) = 5 [3.84]), with the most com- Spanish 9 (42) 6 (44) 15 (38) mon being hypertension (58%), Marital Status 4.36 chronic obstructive pulmo- Married 9 (39) 7 (41) 16 (40) nary disease (43%), tobacco use Not married (eg, single, widow) 14 (61) 10 (59) 24 (60) (40%), and depression (33%). Insurance Coverage 0.27 Men and women were signifi- Insured (Medicaid/Medicare) 10 (43) 6 (44) 16 (40) cantly different in their cancer stage at diagnosis: Men were Uninsured (no health insurance) 13 (57) 11 (56) 24 (60) more likely to have advanced Income, $ 4.25 cancer (79%) and women were < 6,144 12 (57) 6 (40) 18 (50) more likely to have less-ad- 6,240–13,908 3 (14) 6 (40) 9 (25) vanced cancer (stages I/II; 69%). 14,248–33,192 6 (29) 3 (20) 9 (25) Moreover, men were significant- Cancer Diagnosis ly more likely to have received Cancer Site 2.07 chemotherapy than women (65% vs 31%, respectively). Head/neck 11 (48) 12 (71) 23 (57) Lung 12 (52) 5 (29) 17 (43) Interviews Cancer Stage at Diagnosis 13.38* In-depth interviews were con- I 4 (17) 11 (69) 15 (38) ducted in 29 participants: 4 with II 1 (4) 0 (0) 1 (3) lung cancer, 5 with head and neck III 3 (13) 3 (19) 6 (15) cancers, 6 caregivers, 7 HCPs, IV 15 (66) 2 (12) 17 (44) and 7 patient navigators. The patients interviewed were an aver- Cancer Treatment age of 62 years of age (SD 9.64; Surgery 2.84 range, 49–75). Five were men Yes 11 (50) 13 (76) 24 (60) and 4 were women. Five patients No 11 (50) 4 (24) 15 (40) were diagnosed with head and Chemotherapy 4.36** neck cancers and 4 with lung Yes 15 (65) 5 (31) 20 (51) cancer; 3 patients were diag- No 8 (36) 11 (69) 19 (49) nosed at stage I/II disease and 6 at stage III/IV disease. Of the Radiotherapy 1.12 6 caregivers, two cared for pa- Yes 16 (73) 9 (56) 25 (66) tients with head and neck can- No 6 (27) 7 (44) 13 (34) cers and 4 cared for patients with Numbers in the group categories do not sum to the total number of participants because of lung cancer. The ages of the care- missing data for certain variables. ªAssessed as the number of days from the time patients received a cancer diagnosis until their givers ranged from 34 to 59 years first treatment. (mean [SD] = 46.67 [10.31]). One *P < .01. was a man and 5 were women. **P < .05.

October 2016, Vol. 23, No. 4 Cancer Control 393 The 7 HCPs included a pulmonologist, medical on- the start of treatment. Although we examined the in- cologist, oncology nurse, oncology pharmacist, oto- teraction effect of all the predictors in the model, the laryngologist, pulmonology surgeon, and radiation interaction alone between sex and comorbidities was oncologist. Five HCPs identified as being white and found to be significant F( [2, 21] = 3.68; P = .04). As seen 2 self-identified as ethnic minorities; 2 were in the Fig, compared with men, women with fewer co- women and 5 were men. Their time at the safe- morbidities experienced the longest delay from their ty-net hospital ranged from 9 months to 10 years (mean initial diagnosis to the start of cancer treatment. [SD] = 5.46 [3.28]). Of the 7 patient navigators, 5 were women and Interviews 2 were men. Six self-identified as Hispanic and 1 as non- Overall, HCPs and patient navigators had more years of Hispanic white. Their employment at the safety-net experiences to draw upon when discussing how vari- hospital ranged 2 to 9 years (mean [SD] = 3.43 [2.57]). ous factors influenced the time from initial diagnosis to treatment initiation among underserved Hispanic pa- Results tients with lung and head and neck cancers. Thus, pa- Medical Records tient and caregiver responses were organized around Predisposing (ie, age, sex), enabling (ie, health in- the responses of their HCPs and navigators. surance), and need (ie, cancer site, stage at diagno- Need Factors: All participating HCPs conveyed sis, comorbid conditions) variables were entered as that lung and head and neck cancers are usually diag- predictors of days from diagnosis to the initiation of nosed at a later stage of disease, thus decreasing sur- treatment. The prediction model contained 4 of the vival rates among patients with these cancers. They 6 predictors and was reached in 4 steps with 2 vari- recognized that reducing delays in treatment is impor- ables removed (ie, age, health insurance). As shown in Table 2, the model was significant F( [4, 36] = 2.58; P < .05, R: 0.482; R²: 0.233 [adjusted R²: 0.142]). The 100 Sex variables of sex and comorbid conditions were signifi- Men cant predictors of the time from diagnosis to treatment Women initiation. Compared with men, women were signifi- 80 cantly more likely to experience more days between diagnosis and the start of treatment. Patients with few- 60 er comorbid conditions also experienced significantly more days from diagnosis to treatment initiation than those with more comorbidities. Specifically, with each 40 additional comorbid condition, a patient’s time from

diagnosis to treatment initiation decreased by nearly Estimated Marginal Mean (Diagnosis to Treatment, d) 20 17 days. Sex accounted for 9% of the variance and comorbidities accounted for 17% of the variance of diagno- 0 sis to treatment initiation (see Table 2). Inspection of 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 the structure coefficients suggests that women and Sum of Charlson Comorbidity Status patients with fewer comorbidities experienced significantly more days between their initial diagnosis and Fig. — Interaction between the sexes and number of comorbidities.

Table 2. — Coefficients for Predictors of Diagnosis to Treatment Delay (N = 40) Model Unstandardized Coefficient Standardized Coefficient Structured Coefficient B Value Standard Error β Level T test sr² (Constant) 24.87 9.35 — 2.55 — — Sex 24.49 12.30 .369 1.99* .09 .52 Cancer site 12.42 12.04 .187 1.03 .02 –.21 Cancer stage at diagnosis 7.88 6.63 .238 1.19 .03 –.27 Comorbid diseases –16.53 6.07 –.523 –2.73** .17 –.77 R = 0.482; R² = 0.233 (adjusted R² = 0.142). Sex coded as 0 = men, 1 = women. Cancer site coded as 0 = head and neck cancer, 1 = lung cancer. Cancer diagnosis stage coded as 1 = stage 0/1, 2 = stage II, 3 = stage III, and 4 = stage IV. *P < .05. **P < .01.

394 Cancer Control October 2016, Vol. 23, No. 4 tant to reduce the high rates of morbidity from these these cancers. The oncology nurse also reported on cancers. The pulmonologist said: the complications of coordinating care among the different specialties: Lung cancer is a progressive disease and cancer grows. My fear with patients is always that this is a particular Oh, yeah, particularly the head and neck patients be- cancer that can be treated with surgery and cured. If it cause they have to go between radiation, medical on- advances to the point where it can no longer be taken cology, and otolaryngology, so their care is more com- care of in time for surgery, then it’s impossible to cure, plicated. and the only thing we can do is give people radiation or drugs to slow it down, but not cure it. The complex coordination of care was also believed to result in negative treatment outcomes, as il- The HCPs acknowledged lengthy delays in lustrated by the radiation oncologist: treatment, similar to what the pulmonology surgeon reported: If the patient at [my institution] had surgery that I’m not aware of, and the patient comes to me, and really there Well, ultimately, the care is good. It’s just the delay from was a long gap, a long interval … I actually had quite one step to the next. They’re getting the appropriate a few patients when they had surgery who, when they treatments and procedures done — it’s just the delay [of came to me, their tumor had recurred already. 8 months]. Related to lack of care coordination was lack of Delays in treatment were viewed as being influ- communication across departments, among HCPs, and enced by systems-level factors. At the systems level, with patients. Communication was seen as lacking be- most patients and caregivers said that a missed diag- tween different constituents and related to treatment nosis of cancer by their primary care physician was the delays and to difficulties tracking patients during fol- factor that contributed to receiving delayed treatment. low-up or transitional periods. The pulmonology sur- When 1 caregiver explained the reasons her relative geon said: with head and neck cancer experienced a lengthy delay of treatment, she said: I think the gaps — the major ones — are the communication between clinics and providers. So, we have a gap Just maybe look to see if it was cancer before. Before they there that results in delays, and sometimes we don’t have thought it was a blister or fever blister. She had it for a all the information transmitted. It’s only what was writ- long time, and it was probably growing. ten down in the chart. If you don’t look in the right place, you might not find it. That’s the biggest issue. HCPs identified lack of coordinated care as a systems-level factor that contributed to delays in treat- Lack of communication between HCPs and pa- ment. Although underserved patients receive afford- tients was also recognized as contributing to delays able care at the safety-net hospital, they recognized in treatment. The patient navigators stated that HCPs that the process of coordinating these patients’ treat- should better communicate with their patients about ment is difficult. HCPs believed that enrolling for the proper expectations of treatment. Patient naviga- services was an obstacle to patients because of the tors believed that when the explanation of the HCP is amount of paperwork needed and the requirements only verbally communicated, without written instruc- that they must meet before the coordination of these tions or drawn diagrams, then patients forget the in- patients’ treatment can begin. The clinical oncology structions given to them — which, in turn, results in pharmacist summarized it as such: patients delaying treatment or not adhering to treatment instructions. Getting them enrolled is the biggest gap. I think that, if Enabling Factors: Patients, caregivers, HCPs, and they can get through that hurdle, everything else goes a patient navigators agreed that the enabling factors that lot smoother [to begin their treatment]. generally apply to procuring or impeding patient access to health care were also contributors to delays in can- HCPs also reported that the coordination of care cer treatment. Factors related to low SES mentioned by for lung and head and neck cancers is complex and the study participants included financial barriers. All the can delay treatment. Medical care from various depart- respondents highlighted that patients at the safety-net ments (eg, medical oncology, surgery, radiation) and hospital generally lacked the finances to cover basic ex- specialty clinics (eg, otolaryngology) must be coordi- penses for housing, food, and transportation, as well as nated prior to the start of treatment as well as during health insurance or, if they had insurance, the ability to the various procedures necessary for the treatment of cover their co-pays. Thus, the access barrier most direct-

October 2016, Vol. 23, No. 4 Cancer Control 395 ly related to low SES was lack of health insurance or of The routine times to get appointments in some clinics sufficient health insurance to cover the cancer treatment can be pretty long. I think that’s why we see 8-month (eg, radiotherapy) they needed. delays. It’s not surprising when we see it. We see it Patients who did have health insurance were, in frequently. general, covered through Medicaid or Medicare; few had sufficient private health insurance. Instead, a ma- Patients with cancer may receive care at a hema- jority of patients, similar to other underserved His- tology/oncology clinic that functions within the overall panic patients, had no health insurance, which signifi- hospital structure, which further contributes to treat- cantly impeded their access to treatment. The medical ment delays, as the oncology nurse explained: oncologist summarized it by saying: The patient can call and get scheduled into the wrong People [who] do not have good health coverage have a clinic. There are issues with having central schedul- hard time accessing the system. There are some things ing that handles all the appointments for all the outpa- we cannot do for people who do not have good cover- tient clinics, as opposed to having someone dedicated age at [this hospital], eg, positron emission tomogra- to oncology who knows the patients, understands how phy, which is used for staging lung cancer to see how our clinic works, and why patients need to be seen in a advanced it is. timely fashion.

Delays in proper cancer staging may then lead to de- At the study hospital, the hematology/oncology lays in timely treatment for these underserved patients. clinic does not offer radiation oncology services but Due to the significant financial barriers that disad- refers patients to a partnering academic cancer center vantaged Hispanic patients at the safety-net hospital face when patients require radiotherapy. Navigating be- to initiate and complete their cancer treatment, state- tween separate treatment clinics was problematic and funded programs are available for select patients, such resulted in further treatment delays, particularly when as the Colorado Indigent Care Program, or some hospi- care coordination was lacking, as the radiation oncolo- tals can assist them by finding resources that could help gist explained: them cover some treatment costs. For example, some patients received assistance with medications they could I just had a patient — he had surgery done, he’s late not afford, as mentioned by the oncology nurse: already, he’s like close to 4 weeks — so I saw the patient. I did the treatment planning the same day, so I want to There is an indigent formulary at this hospital, so that start the treatment really quick. [Results from] comput- means certain folks [with no financial resources] come ed tomography were not there [with the patient file from and get certain drugs. the safety-net hospital].

However, the oncology nurse explained that the Treatment delays that result from difficulties navi- process of finding resources for underserved patients gating the health care system are also patient-depen- makes the treatment of these patients difficult: dent, and they generally originate from patients not understanding their HCPs’ follow-up instructions or We have our hands tied in a lot of situations or have a not understanding how the health care system works. number of hurdles to jump through to provide the best The pulmonologist made the observation: care that we can for patients, and we absolutely do everything that we can to make sure we are providing Part of the problem is the patients don’t understand the best care possible. how to get where they need to be. Sometimes they need help filling out forms to get their tests done in a timely The HCPs believed that the process of assisting fashion. patients was also a lengthy process that contributed to treatment delays. Lack of patient understanding is likely linked to Although the safety-net hospital staff has adminis- lack of care coordination at the systems-level, and dif- trative difficulties in providing care to this underserved ficulties navigating the health care system then contrib- population, these patients also face the hurdles of navi- ute to delays in treatment, as the pulmonology surgeon gating the health care system themselves. Patients en- explained: countered system difficulties throughout the treatment process, from scheduling appointments to receiving I think the major barriers are delays related to the pro- timely and proper treatment. Navigation hurdles that cess of getting from one clinic to the next in our system. delayed treatment tended to exist at the system level, as Some of them may be language-related and some of the pulmonology surgeon indicated: them are just the way our system has been set up.

396 Cancer Control October 2016, Vol. 23, No. 4 Predisposing Factors: Personal, social, cultural, Thus, social support appears to impact patients’ and psychological propensities were noted as predis- timely access and adherence to treatment. posing factors that contribute to a patient’s ability to ac- Lack of health literacy was another predisposing cess cancer treatment. Influential personal factors were factors that played an important role in whether Hispanic related to patient immigration status and included lack men experienced treatment delays. We found that these of language proficiency and lack of legal status. underserved patients have difficulties understanding Most Hispanic patients at the safety-net hospital medical information related to their diagnosis and treat- spoke only Spanish — a barrier to timely treatment: ment, and, consequently, do not follow-up with their treatment regimen, as explained by the medical oncologist: The patient showed up late with the tumor recurred. … I think just for the Spanish patients — I think there are Sometimes patients have no idea that they have to come things … they probably had more difficult than the other back. Even though we tell them we want them back, patient population. Well, once, for whatever reason, they they have no idea why. … And also there is lack of un- missed the appointment; they had [a] language barrier. derstanding from the patients of what is required to be successful when they’re in treatment. Similarly, patients who lacked documentation of their legal immigration status encountered problems Misconceptions about cancer treatment were also accessing treatment care, as the oncology pharmacist reported as stemming from lack of health literacy. For explained: example, when the pulmonologist was asked to expand on how he believed lack of health literacy con- If they [patients] are undocumented and there is a pa- tributed to treatment delays, he said: tient-assistance program available through pharmaceutical companies to provide drugs at low cost or free, Well, so, I think one of the things I’ve heard from a lot undocumented patients don’t qualify. of immigrants when we talked about surgery for lung cancer is that they don’t want to have surgery because Language and legal status barriers impeded ac- they’ve heard that will cause the cancer to spread. cess to care, and, if they eventually accessed resources, their treatment was significantly delayed. Other similar diagnosis and treatment misconcep- Another predisposing factor to timely treatment is tions also appeared to contribute to treatment delays. a patient’s social support system. Patients who have a Mental health propensities were identified as di- caregiver or family members to assist them with navi- rectly and indirectly influencing a patient’s ability to gating the medical system and support them through initiate treatment. Substance abuse was related to treat- treatment are more successful at obtaining timely and ment delays, which the medical oncologist explained: proper treatment, as noted by the otolaryngologist’s report of the important role of a caregiver: Yeah, well, that’s a problem because, number one, they sometimes have to actually quit these other complica- I have one man who lost his larynx from laryngeal tions [before treatment], … and, you know, sometimes cancer, and his sister — his wicked sister — brings him that doesn’t happen. in and she’s just a riot, but she keeps on this guy like you can’t believe. She has him here on time, all the time. Similarly, emotional distress (eg, depression, anxiety, fear) affects patients’ ability to participate in treat- An HCP also emphasized how a social support sys- ment. The otolaryngologist illustrated: tem can help patients overcome other barriers that im- ped timely treatment: People who drink have depression. Uh, either it’s generated by the alcoholism or it’s preceding the alco- My Latino patients who have done well have English- holism — where they drink because they’re depressed. speaking families who have been here, and then they I would think that’s an impediment [to treatment] that bring mom and dad in. begins, you know, with the patient itself.

By contrast, patients without caregivers or fam- Patients and their caregivers also recognized the ily members to care for them have a more difficult role of emotional distress in the patient’s ability to initi- experience: ate treatment and attend follow-up visits, as 1 woman with head and neck cancer explained: A lot of our patients have no support system. So, if they get sick, they have no one to help them. And that’s Yup, it was the first [round of] chemotherapy, after we tough, really tough. found out. I missed my first — I was scared — I missed

October 2016, Vol. 23, No. 4 Cancer Control 397 my first appointment because I was scared. Same with tion was common, either from enrolling patients for radiation. I missed the first appointment. services to coordinating the specialty treatments they need — all of which were believed to result in treat- Discussion ment delays. Our exploratory study findings help to understand the In our study, communication was lacking across systems- and patient-level determinants of treatment various clinics, providers, and patients when coor- delay among underserved Hispanic patients diagnosed dinating care, and it was consistent with the intui- with lung and head and neck cancers. Such contribut- tive notion that the multidisciplinary coordination ing evidence was unavailable prior to this study. The of care is complex and takes time, particularly if literature that does exist suggests that socioeconomic specialty care services are limited.28 Creating and inequalities negatively impact the ability of under- encouraging the participation of HCPs in multi- served patients to receive treatment for lung and head disciplinary tumor boards at public hospitals may and neck cancers, and, when ethnicity is included, im- help improve communication and contribute to bet- poverished Hispanic patients are at a particular disad- ter care coordination among oncologists and other vantage.9-11 HCPs. Use of tumor review boards has been associ- A paucity of evidence explains what factors deter- ated with improved care and increased rates of sur- mine the treatment delays experienced by this patient vival among patients with cancer, although this as- population, but the available data indicate that the time sociation is controversial, and limitations still exist interval between the initial diagnosis and the start of for those living in rural, low-resource areas, as well treatment continues to increase and that delays beyond as in other suboptimal settings.35,36 Undergoing cu- 46 days increase the risk of death.17,20 Treatment delays rative surgery among patients with early-stage lung among patients with lung and head and neck cancers and head and neck cancers is also likely to be im- occur more often in those with early-stage disease,17,21 proved when their HCPs participate in weekly tumor placing them at a higher risk for disease progres- board meetings.37 sion,22-25 disease recurrence,13-16 increased emotional The patients and caregivers interviewed did not distress,18,19 and lower rates of survival.17 Thus, assess- attribute poor treatment outcomes to treatment delays ing the determinants of treatment delays is important, due to lack of care coordination or lack of communica- because such understanding could lead to intervention; rather, they attributed treatment delays to an ini- tions that effectively target the factors that influence tial missed diagnosis of cancer by their primary care treatment inequalities among vulnerable populations, physician. They implied that this initial missed diagno- such as underserved Hispanic patients with lung and sis resulted in a delayed referral to an oncologist. One head and neck cancers. meta-analysis found that delayed referrals to specialists We used the Behavioral Model for Vulnerable predicted a 3-fold mortality risk for patients with head Populations to frame our study design and analysis be- and neck cancers, likely resulting from treatment de- cause it includes domains relevant to understanding lays and advanced disease.38 In our study, patients and the factors that influence access and the utilization of caregivers believed that, after being diagnosed, the health care services among vulnerable populations.30,31 safety-net hospital provided patients with the care that Thus, the Behavioral Model for Vulnerable Populations they needed and viewed any treatment delays as being guided the choice of variables to include in the predic- less consequential. tion model, from the medical records and the ques- The 7 patient navigators interviewed recognized tions to ask patients and their caregivers, HCPs, and that delays in treatment were overall related to lack of patient navigators. care coordination and miscommunication across hospital clinics. They were particularly concerned that Institutional-Related Factors patients with lung and head and neck cancers need The literature and our findings unequivocally point to system-navigation services to decrease delays in treat- the need to decrease the treatment delays for lung and ment, but they noted that patients were not receiving head and neck cancers because the timely initiation of such services at the hospital. treatment is likely to result in improved outcomes and If the systems-level factors are not addressed, survival for patients with these types of cancer. In our our findings suggest that treatment delays are likely study, 35% of patients had treatment delays longer than to continue — and possibly increase disparities in 46 days, at a point when these patients have a signifi- treatment outcomes — among underserved Hispanic cant increased risk of death.17 Series of institution-level patients with lung and head and neck cancers. An factors common to publicly funded hospitals appear to evidence-based solution that can improve communi- be associated with such treatment delays in the under- cation and care coordination may well be a patient- served patients of our study. At the safety-net hospital navigation program for this underserved patient where we conducted the study, lack of care coordina- population.39,40

398 Cancer Control October 2016, Vol. 23, No. 4 Patient-Related Factors cessing care that may result in treatment delays. An As is characteristic of other patients at the safety-net adequate social support system that can assist pa- hospital, the Hispanic patients in our study were medi- tients with language barriers and to navigate the cally underserved, with the majority of participants health care system could facilitate the ability of pa- (60%) being uninsured and all of them being poor tients to obtain proper and timely treatment. A factor (annual income < $36,000).41 Due to lack of variability common to underserved patients is the lack of health in income and health insurance status, this factor did literacy that, in this study, was reported to be asso- not contribute to the prediction of treatment delays in ciated with lack of proper and timely adherence to our study. Instead, sex and number of comorbidities recommended treatment.42 For patients who lack ad- had independent effects on treatment delay, but these equate social support — particularly those with lim- variables were also intertwined. Compared with men ited proficiency in English — or adequate health liter- whose treatment was, on average, delayed by 34 days, acy, the presence of language translators or of patient disparities in treatment delays were more pronounced navigators can help to engage patients and caregivers for women whose treatment was delayed by, on aver- during physician visits, thereby increasing their abil- age, 48 days (above the critical 46-day delay17) from ini- ity to understand and follow the required steps in the tial diagnosis to treatment initiation. Moreover, women cancer treatment process.39,40 with fewer comorbidities (≤ 4) were more likely to have A barrier to individuals with lung and head and increased delays in treatment compared with women neck cancers is substance abuse, which, in our study, who had more comorbidities or men who had a similar was anecdotally linked to delays for patients unable to lower number of comorbid conditions. Compared with initiate treatment if they continued to smoke tobacco men, women were also more likely to be diagnosed or drink alcohol. In particular, substance-abuse prob- with early-stage disease. Thus, our findings suggest lems place underserved Hispanic patients with these that healthier women might be at high risk for treat- types of cancers at a higher risk for emotional distress ment delays and patients whom safety-net hospital and mortality.1,11,43 Because mental health disparities staff should follow-up more closely to avoid disparities are present among underserved Hispanic patients with in poor treatment outcomes,13-16,22-25 distress,18,19 and cancer, affordable mental health services can assist low rates of survival.17 them in addressing these problems and improve treat- Some published data indicate that the time from ment delays.44 first diagnosis to treatment initiation does not significantly vary by comorbidity among patients treated at Limitations academic facilities.17 For patients treated at high-bur- This exploratory study used a convenience sample to den, safety-net hospitals, it is possible that a greater explore some possible determinants of treatment de- number of comorbidities means that these patients are lays among underserved Hispanic patients with lung better integrated into the system and, consequently, and head and neck cancers. A randomly selected experience fewer treatment delays. If this is the case, sample followed longitudinally would strengthen the then we think it appropriate to assume that patients generalizability of the study results to this population who receive care for comorbid conditions are likely to of patients with cancer. The categorization of patients have already enrolled in hospital services, medication as Hispanics also limits generalizations to how deter- programs, and be adept at navigating the health care minants of treatment delays may differ by other eth- system. Thus, having more than 4 comorbid conditions nic groups (eg, country of origin, acculturation levels). could be a proxy for patients having had some insti- Methodologically, we were limited by the data avail- tution-level barriers addressed, possibly resulting in a able in the medical records of the study patients, and more timely initiation of cancer treatment. we were unable to assess their reliability. Lack of legal status in the United States and lack Because the incomes of all of our study partici- of English-language proficiency are determinants pants fell below the poverty level at the beginning of treatment delays that disadvantage medically un- of the study,14 we could not examine the effects of derserved Hispanic patients, although the effects of SES on the number of days from the initial diagno- these factors on treatment delays among these pa- sis to treatment initiation as reported in their medi- tients have not been quantified. As found in our cal records. Instead, in-depth interviews with the study, patients with undocumented status are un- HCPs suggested that low SES contributed to treat- able to qualify for resources that other underserved ment delays via the patients’ lack of health insur- patients are able to access to cover their cancer treat- ance. We also could not explore the effects of health ment, and, if they eventually do, their treatment is insurance coverage, because most patients were unlikely to be significantly delayed. Similarly, our find- insured or lacked proper health insurance through ings indicate that immigrant Hispanic patients who Medicaid or Medicare. However, the Affordable Care do not communicate in English face barriers to ac- Act that expands the coverage of public health in-

October 2016, Vol. 23, No. 4 Cancer Control 399 literature. Radiother Oncol. 2008;87(1):3-16. surance for underserved patients may now improve 17. Murphy CT, Galloway TJ, Handorf EA, et al. Survival impact of their access to timely and proper treatment; how- increasing time to treatment initiation for patients with head and neck cancer in the United States. J Clin Oncol. 2016;34(2):169-178. ever, its effects on cancer disparities remains to be 18. Haisfield-Wolfe ME, McGuire DB, Soeken K, et al. Prevalence demonstrated.45 and correlates of depression among patients with head and neck cancer: a systematic review of implications for research. Oncol Nurs Forum. 2009;36(3):E107-E125. Conclusions 19. Hess CB, Chen AM. Measuring psychosocial functioning in the radiation oncology clinic: a systematic review. Psychooncology. Data suggest that delays in cancer treatment are associ- 2014;23(8):841-854. ated with poor outcomes and lower rates of survival for 20. Murphy CT, Galloway TJ, Handorf EA, et al. Increasing time to treatment initiation for head and neck cancer: an analysis of the National patients with lung and head and neck cancers. Thus, Cancer Database. Cancer. 2015;121(8):1204-1213. immediate action should be taken to improve the time 21. Vinas F, Ben Hassen I, Jabot L, et al. Delays for diagnosis and treatment of lung cancers: a systematic review. Clin Respir J. 2016;10(3):267-271. between diagnosis and treatment initiation for these 22. Waaijer A, Terhaard CHJ, Dehnad H, et al. Waiting times for radio- disadvantaged patients. therapy: consequences of volume increase for the TCP in oropharyngeal carcinoma. Radiother Oncol. 2003;66(3):271-276. Future studies should assess the effects of the de- 23. Jensen AR, Nellemann HM, Overgaard J. Tumor progression in terminants that our interviewees believed influence waiting time for radiotherapy in head and neck cancer. Radiother Oncol. 2007;84(1)5-10. delays in treatment initiation among underserved His- 24. Kowalski LP, Carvalho AL. Influence of time delay and clini- panic patients with cancer treated at a safety-net hos- cal upstaging in the prognosis of head and neck cancer. Oral Oncol. 2001;37(1):94-98. pital. Our findings could be used to guide future re- 25. Laroche C, Wells F, Coulden R, et al. Improving surgical resection search as well as interventions for underserved patient rate in lung cancer. Thorax. 1998;53(6):445-449. 26. Yorio JT, Xie Y, Yan J, et al. Lung cancer diagnostic and treatment populations. intervals in the United States: a health care disparity? J Thoracic Oncol. Acknowledgment: The authors wish to thank Drs 2009;4(11):1322-1330. 27. Munck K, Ali MJ, Murr AH, et al. Impact of socioeconomic status Estevan Flores, Paula Espinoza, Al Marcus, and Tim on the diagnosis to treatment interval (DTI) in Waldeyer’s ring carcinoma. Byers for their generous advice. Laryngoscope. 2005;115(7):1283-1287. 28. Patel UA, Brennan TE. Disparities in head and neck cancer: assessing delay in treatment initiation. Laryngoscope. 2012;122(8):1756-1760. References 29. Creswell JW, Plano Clark VL. Designing and Conducting Mixed Methods Research. 2nd ed. Thousand Oaks, CA: Sage Publications; 2007. 1. American Cancer Society. Cancer Facts and Figures 2016. 30. Gelberg L, Andersen RM, Leake BD. Healthcare access and uti- http://www.cancer.org/acs/groups/content/@research/documents/ lization. The Behavioral Model for Vulnerable Populations: application to document/acspc-047079.pdf. Accessed September 6, 2016. medical care use and outcomes for homeless people. Health Serv Res. 2. American Cancer Society. Cancer Facts & Figures for Hispanics/ 2000;34(6):1273-1302. Latinos: 2015-2017. http://www.cancer.org/acs/groups/content/@research/ 31. Bazargan M, Bazargan SH, Farooq M, et al. Correlates of cervi- documents/document/acspc-046405.pdf. Accessed September 6, 2016. cal cancer screening among underserved Hispanic and African American 3. Ries LAG, Melbert D, Krapcho M, et al, eds. SEER Cancer Statistics women. Prev Med. 2004;39(3):465-473. Review, 1975-2005, National Cancer Institute. Bethesda, MD. http://seer. 32. Singh B, Bhaya M, Stern J, et al. Validation of the Charlson cancer.gov/csr/1975_2005. Based on November 2007 SEER data submission. Comorbidity Index in patients with head and neck cancer: a multi‐institu- Accessed July 29, 2016. tional study. Laryngoscope. 1997;107(11 pt 1):1469-1475. 4. Kuriakose MA, Loree TR, Rubenfeld A, et al. Simultaneously 33. Altheide DL. Ethnographic content analysis. Qualitat Sociol. 1987; presenting head and neck and lung cancer: a diagnostic and treatment 10(1):65-77. dilemma. Laringoscope. 2002;112(1):120-123. 34. Krippendorf K. Content Analysis: An Introduction to its Methodology. 5. Centers for Disease Control and Prevention. Smoking-attributable 2nd ed. Beverly Hills, CA: Sage Publications; 2004. mortality, years of potential life lost, and productivity losses – United 35. Keating NL, Landrum MB, Lamont EB, et al. Tumor boards and the States, 2000-2004. MMWR Morb Mortal Wkly Rep. 2008;57(45):1226-1228. quality of cancer care. J Natl Cancer Inst. 2013;105(2):113-121. 6. Blot WJ, McLaughlin JK, Winn DM, et al. Smoking and drinking in 36. Krasna M, Freeman RK, Petrelli NJ. Re: Tumor boards and the relation to oral and pharyngeal cancer. Cancer Res. 1988;48(11):3282-87. quality of cancer care. J Natl Cancer Inst. 2013;105(23):1839-1840. 7. Centers for Disease Control and Prevention. Cigarette smoking- 37. Kehl KL, Landrum MB, Kahn KL, et al. Tumor board participation attributable morbidity -- United States, 2000. MMWR Morb Mortal Wkly Rep. among physicians caring for patients with lung or colorectal cancer. J Oncol 2003;52(35):842-844. Pract. 2015;11(3):e267-e278. 8. Agency for Healthcare Research and Quality (AHRQ). 2014 National 38. Seoane J, Takkouche B, Varela-Centelles P, et al. Impact of delay Healthcare Quality & Disparities Report. Rockville, MD: AHRQ; 2015. in diagnosis on survival to head and neck carcinomas: a systematic review http://www.ahrq.gov/sites/default/files/wysiwyg/research/findings/nhqrdr/ with meta-analysis. Clin Otolaryngol. 2012;37(2):99-106. nhqdr14/2014nhqdr.pdf. Accessed September 6, 2016. 39. Pasket ED, Harrop PJ, Wells KJ. Patient navigation: an update on 9. Forrest LF, Adams J, Wareham H, et al. Socioeconomic inequalities the state of science. Cancer. 2011;61(4):237-249. in lung cancer treatment: systematic review and meta-analysis. PLoS Med. 40. Raich PC, Whitley EM, Thorland W, et al. Patient navigation im- 2013;10(2):e1001376. proves cancer diagnostic resolution: an individually randomized clinical 10. Choi SH, Terrell JE, Fowler KE, et al. Socioeconomic and other trial in an underserved population. Cancer Epidemiol Biomarkers Prev. demographic disparities predicting survival among head and neck cancer 2012;21(10):1629-1638. patients. PloS one. 2016;1;11(3):e0149886. 41. US Department of Health and Human Services; Office of The 11. Yang R, Cheung MC, Byrne MM, et al. Do racial or socioeconomic Assistant Secretary for Planning and Evaluation. 2013 poverty guidelines. disparities exist in lung cancer treatment? Cancer. 2010;116(10):2437-2447. Published December 1, 2013. http://aspe.hhs.gov/poverty/13poverty.cfm. 12. Committee on Quality of Health Care in America; Institute of Medicine. Accessed September 6, 2016. Crossing the Quality Chasm: A New Health System for the 21st Century. 42. Shaw SJ, Huebner C, Armin J, et al. The role of culture in health Washington, DC: National Academy Press; 2001. literacy and chronic disease screening and management. J Immigr Minor 13. Ang KK, Trotti A, Brown BW, et al. Randomized trial addressing risk Health. 2009;11(6):460-467. features and time factors of surgery plus radiotherapy in advanced head- 43. Glazier KA. Tobacco-induced diseases: an emphasis on mental and-neck cancer. Int J Radiat Oncol Biol Phys. 2001;51(3):571-578. health conditions that may result from cigarette use. N Sch Psychol Bulln. 14. Fortin A, Bairati I, Albert M, et al. Effect of treatment delay on outcome 2010;8(1):56-61. of patients with early-stage head-and neck carcinoma receiving radical 44. Luckett T, Goldstein D, Butow PN, et al. Psychological morbidity radiotherapy. Int J Radiat Oncol Biol Phys. 2002;52(4):929-936. and quality of life of ethnic minority patients with cancer: a systematic 15. Hansen O, Larsen S, Bastholt L, et al. Duration of symptoms: review and meta-analysis. Lancet Oncol. 2011;12(13):1240-1248. impact on outcome of radiotherapy in glottis cancer patients. Int J Radiat 45. Moy B, Polite BN, Halpern MT, et al. American Society of Clinical Oncol Biol Phys. 2005;61(3):789-794. Oncology policy statement: opportunities in the patient protection and 16. Chen Z, King W, Pearcey R, et al. The relationship between waiting Affordable Care Act to reduce cancer care disparities. J Clin Oncol. time for radiotherapy and clinical outcomes: a systematic review of the 2011;29(28):3816-3824.

400 Cancer Control October 2016, Vol. 23, No. 4 Geographical analysis can be used to

respond to select barriers that contribute

to disparities in prostate care.

Orchid. X-Ray image courtesy of Bryan Whitney. www.x-rayphotography.com.

Geographical Factors Associated With Health Disparities in Prostate Cancer Scott M. Gilbert, MD, Julio M. Pow-Sang, MD, and Hong Xiao, PhD

Background: Treatment variation in prostate cancer is common, and it is driven by clinical and clinician factors, patient preferences, availability of resources, and access to physicians and treating facilities. Most research on treatment disparities in men with prostate cancer has focused on race and socioeconomic factors. However, the geography of disparities — capturing racial and socioeconomic differences based on where patients live — can provide insight into barriers to care and help identify outlier areas in which access to care, health resources, or both are more pronounced. Methods: Research regarding treatment patterns and disparities in prostate cancer using the Geographical Information System (GIS) was searched. Studies were limited to English-language articles and research focused on US populations. A total of 43 articles were found; of those, 30 provided information about or used spatial or geographical analyses to assess and describe differences or disparities in prostate cancer and its treatment. Two additional GIS resources were included. Results: The research on geographical and spatial determinants of prostate cancer disparities was reviewed. We also examined geographical analyses at the state level, focusing on Florida. Overall, we described a geographical framework to disparities that affect men with prostate cancer and reviewed existing published evidence supporting the interplay of geographical factors and disparities in prostate cancer. Conclusions: Disparities in prostate cancer are common and persistent, and notable differences in treatment are observable across racial and socioeconomic strata. Geographical analysis provides additional information about where disparate groups live and also helps to map access to care. This information can be used by public health officials, health-systems administrators, clinicians, and policymakers to better understand and respond to geographical barriers that contribute to disparities in care. Introduction From the Departments of Genitourinary Oncology (SMG) and Health Outcomes and Behavior (SMG, JP), H. Lee Moffitt Cancer Prostate cancer is the most common nondermatologi- Center & Research Institute, Tampa, Florida, and the Department cal cancer diagnosed among American men, account- of Pharmaceutical Outcomes and Policy (HX), College of Phar- ing for approximately 180,000 new cases each year in macy, University of Florida, Gainesville, Florida. the United States.1 Although it is typically character- Submitted May 2, 2016; accepted September 20, 2016. Address correspondence to Scott M. Gilbert, MD, Departments ized by a protracted clinical course, prostate cancer is of GU Oncology & Health Outcomes and Behavior, Moffitt Can- also one of the most common causes of cancer-related cer Center, 12902 Magnolia Drive, XCB-GU Program, Tampa, FL death (fourth most common overall following lung, 33612-9416. E-mail: [email protected] colorectal, and breast cancers) and accounts for more No significant relationships exist between the authors and the companies/organizations whose products or services may be ref- than 26,000 deaths related to cancer each year in the erenced in this article. United States alone.2

October 2016, Vol. 23, No. 4 Cancer Control 401 Although most prostate cancers (80%) are detect- and reports were examined and included if they were reed at an early stage and 10-year survival rates surpass lated to geographical and health disparities. 90%, 16% of men present with either locally advanced Studies were limited to research focusing on US or metastatic disease.3 Metastatic prostate cancer is populations and those published in the English lan- an incurable disease despite advances in treatment.3 guage. A total of 43 publications were identified, 30 of Early-stage prostate cancer is also a heterogeneous which were relevant and included in our review. disease defined by variable clinical behavior and outcomes.4 Following local treatment, prostate cancer re- Results curs in up to 40% of men with early-stage disease.5,6 The geographical and prostate cancer factors found in An estimated 2.85 million American men are living selected disparity studies have been summarized in with prostate cancer, making this patient group the the Table.12-25 second largest group of cancer survivors.3 With health care expenditures surpassing an estimated $9.8 bil- Racial and Sociodemographic Disparities lion each year, prostate cancer is also one of the most Disparities in treatment and survival between black and costly cancers in the United States.7 non-black men have long been observed. Research has Although these figures underscore the personal shown that black men are at significantly higher risk of and public health burdens associated with prostate dying from prostate cancer than white men, they are cancer, not all men fare equally. Disparities in risk of less likely to receive curative local therapy for intermedi- aggressive disease and differences in treatment and ate- and high-risk disease than their white counterparts, outcomes are consistently observed.3-11 For example, and small but significant differences in 5-year survival incidence rates among black men (214.5 per 100,000 rates have been documented.26,27 Mahal et al27 reported persons) are significantly higher compared with their that black men are 12% more likely to die from prostate white counterparts (130.0 per 100,000 persons), and cancer than white men (hazard ratio 1.12; 95% confi- death rates from prostate cancer are more than double dence interval [CI]: 1.01–1.25), corresponding to an ab- for black men compared with their white counterparts solute risk of death of 4.5% for blacks (95% CI: 4.1–4.9) (46.3 per 100,000 persons vs 19.8/100,000 persons, compared with 3.4% for whites (95% CI: 3.3–3.6). They respectively).3 Several factors may contribute to ob- also found that black men were 18% less likely to receive served racial disparities, including biological determi- curative treatment than their white counterparts.27 Oth- nants (differences in genetic polymorphisms related to er studies have also reported lower rates of screening testosterone metabolism and the androgen receptor) and appropriate staging rates among black men com- differences in health behaviors, and access to health pared with white men.28,29 Black men less frequently re- care resources.8 Beyond racial factors, men with lower ceive curative therapy, they may receive treatment later social status and educational level are also more fre- after diagnosis, and they are less likely to be treated with quently diagnosed with more aggressive, clinically sig- adjuvant therapy than white men.29 nificant forms of prostate cancer than men from higher Increased risk of mortality has been observed in socioeconomic backgrounds, and prior research indi- black men as well as in men from disadvantaged so- cates that health seeking and literacy issues, as well as cial, educational, and economic backgrounds: these access to health care resources, may partially explain men have higher rates of death from prostate cancer, why some men have poorer outcomes than others.9-11 regardless of race.30 Inequity in mortality related to Certainly social and demographical factors may be sig- prostate cancer is generally ascribed to 1 of 3 factors: nals for other yet-to-be identified sources that drive di- disease factors (eg, differences in biology, risk of more vergent outcomes, or they may be principal factors that aggressive disease in some groups), patient factors explain why some men die from prostate cancer and (eg, health behaviors), and issues endemic to the health others do not. Disparate risk factors and outcomes can care system (eg, differences in access to health care fa- be defined by geographical boundaries, and there may cilities and resources, screening and detection practic- be a geographical component to health disparities. es, treatment variation).31-33 Social, economical, and education disparities may Methods be linked to significant differences in disease manage- In addition to studying broader reviews of racial and so- ment and outcomes.34-39 Several studies have shown that ciodemographic disparities, our search of available litera- men from disadvantaged groups (eg, those with lower ture databases focused on geographical factors associated socioeconomic status) are less likely to receive aggressive with health and outcome disparities, and use of spatial local treatment, such as radical surgery or radiotherapy, analyses, including the Geographical Information Sys- than those of higher socioeconomic status.34-37 Screen- tem (GIS), in disparity research studies related to prostate ings are also less commonly performed in poorer and cancer. Additional, nonindexed prostate cancer treatment less-educated communities, possibly leading to diagno- and outcome reports using the GIS were searched online, ses at a later-stage disease when fewer treatment options

402 Cancer Control October 2016, Vol. 23, No. 4 Table. — Summary of Select Disparity Studies With Geographical and Prostate Cancer Factors Study Geographical Unit Disease Phase Key Finding Aneja12 United States, county Mortality 3.91%–5.45% reduction in mortality in counties with ≥ 1 radiation oncologist compared with counties lacking radiation oncologists Antwi13 Kentucky, Appalachian region Survival Black men had lower survival rates compared with white men (HR 1.41; 95% CI: 1.26–1.65) Cetnar14 Wisconsin, census tract Primary treatment Likelihood of noncurative treatment no different in rural compared with urban areas (OR 0.96; 95% CI: 0.52–1.77) Cobran15 United States, SEER site Treatment Intensity-modulated radiotherapy more commonly used in white than blacks (53% vs 45%; P < .05) Socioeconomic-, geographic-, and tumor-related factors explained preponderance of variation Cook16 United States, census divisions Incidence rates Large geographical differences in black-to-white incidence rate ratios DeChello17 Connecticut and Massachusetts, Incidence rates Significant geographical variation observed in age-adjusted census tract incidence rates among white and black men Garg18 United States, state Screening Significant variation in PSA screening among states Men living in urban areas have lower odds of screening (OR 0.8; 95% CI: 0.7–0.9), low physician density, low PSA screening (OR 0.78; 95% CI: 0.63–0.94) Goovaerts19 Florida, county Incidence of late-stage disease Variation noted in temporal decrease in late-stage diagnosis by county and race Gregorio20 Connecticut, census tract Survival Spatial scan statistical analysis of adjusted survival times identified one significant outlier area in the state with significantly lower survival (death hazard 1.39 > state average; P < .05) Meliker21 Michigan, legislative districts Survival Higher survival rates observed in white men at state level without confirmation of racial disparities at smaller geographical units Onukwugha22 United States, county Follow-up care Significant variation in follow-up physician visits explained by geographical factors Black men living in low socioeconomic and high-crime areas less likely to receive follow-up care Wagner23 Georgia, county Incidence rates “Hot spots” identified in several counties in both black and white men in northern and northwestern-central Georgia Xiao24 Florida, county and census tract Incidence of late-stage disease Late-stage diagnosis related to low census tract median income and lower educational attainment Xiao25 Florida, county and census tract Incidence of late-stage disease Late-stage diagnosis associated with lack of insurance, low household income, rural location CI = confidence interval, HR = hazard ratio, OR = odds ratio, PSA = prostate-specific antigen, SEER = Surveillance, Epidemiology, and End Results. may be available or cure is less likely.38,39 Level of health in treatment.30 In aggregate, these data suggest that low literacy and patient decision-making capacity may also socioeconomic status increases the risk of prostate can- contribute to socioeconomic disparities,40 a hypothesis cer–related death in men with prostate cancer, even af- partly supported by research showing that racial dispari- ter accounting for differences in health, stage at diagno- ties appear to be modified by socioeconomic status.41 sis, and treatment. It is worth noting that socioeconomic This suggests that the gap between white and black men status disparities do not appear to have decreased in residing in higher-income and higher-level education the last 10 years.10 By contrast, average survival rates of neighborhoods is narrower than the gap separating resi- prostate cancer have improved during that same time.3 dents from low socioeconomic status regions. A systematic review analyzing the association be- Geographical Disparities tween socioeconomic status and risk of prostate can- Examining nationwide treatment trends, Harlan cer–related death reported that most research supports et al42 noted significant regional variation across defini- a significant link between socioeconomic status and tive treatment modalities among both white and black prostate cancer–related mortality, even after adjusting men. The age-adjusted proportion of men treated with for factors such as patient comorbidities and differences prostatectomy ranged from 22.5% in Connecticut to

October 2016, Vol. 23, No. 4 Cancer Control 403 47.8% in Utah, although use of radiotherapy varied less further refined and commercialized by Fisher.50 Apply- across Surveillance, Epidemiology, and End Results ing the GIS to visualize and analyze spatial and temporal (SEER) cancer registry sites.42 In addition to confirm- disease patterns has become more prevalent in health ing widespread variation in treatment modality in an- care research in recent years.51 Health conditions such other SEER cohort, Krupski et al36 also documented as cardiovascular disease, hepatitis C virus infection, significant racial and socioeconomic variation in care, birth defects, and various cancers have been displayed with black and low-income men receiving surgery for in a spatial context to better understand these disease localized disease significantly less commonly than states.51 By merging health data and geography, the GIS white, higher-income men. The adjusted likelihood can also help inform decision-making when allocating of surgical therapy was less than one-half (odds ratio resources. Thus, research involving GIS brings together [OR] 0.44; 95% CI: 0.41–0.47) in the lowest utilization several different fields, including medicine, epidemiol- region compared with the highest, and the investiga- ogy, geography, computer sciences, and statistics.50 In tors documented that black men were one-half as likely recent years, GIS has started to extend to areas of epide- to be treated with surgery than whites (OR 0.52; 95% CI: miological and health services research — focusing on 0.48–0.56).36 Regional variation in radiotherapy was health disparities — to examine the availability of health observed: it was typically more common in regions resources, utilization of health care, the geographical with low rates of surgery.36 The results of another study distribution and clustering of cancers, treatment, and examining ecological correlations between variation in rates of mortality. For example, health-related GIS stud- prostate cancer–related deaths and surrogate measures ies have increased the ability of public health officials of access to medical care (eg, prostate-specific antigen to monitor diseases and to identify areas of need. The [PSA] screening uptake, late-stage presentation) sug- National Cancer Institute has an interactive website that gest that between 10% and 30% of observed variations uses the GIS for cancer control purposes.52 in mortality rates may be attributable to access issues.43 In the field of prostate cancer, research using GIS Although regional differences in treatment for pros- approaches has been conducted to investigate diag- tate cancer have been well described, a paucity of re- noses, management, and treatment outcomes across search has focused on the underlying determinants of different geographical levels, such as counties, census geographically linked disparities. Access, distance, and division/tracts, cancer registry areas, and other aggre- travel time to health care facilities are germane when gate geographical scales.12-25,53-68 Cockburn et al53 used compared with concerns related to disparities in treat- the GIS to geocode the historical residential addresses ment and subsequent outcomes among US residents liv- of individuals and measured the association between ing in rural areas.44 One review examined the differenc- ambient exposure to pesticide and incidence rates of es in incidence and mortality rates between residents of prostate cancer in California. They found a positive re- rural and urban areas.45 Few differences were observed lationship between prostate cancer and ambient pesti- in incidence rates, particular in the post–PSA screening cide exposure in and around homes in intensely agri- era, but the findings of the review suggested that men cultural areas.53 Hsu et al56 used a spatial scan statistic living in rural areas are at higher risk of prostate cancer– to detect counties with increased mortality between related death than those living in urban areas, presum- 1980 and 2001 in Texas. DeChello et al17 used the same ably due to issues related to health care accessability.45 method and found geographical variation in incidence Other research has also shown that patients living in ru- rates of prostate cancer among both white and black ral areas, regardless of their racial/ethnic background, men between 1994 and 1998 in Connecticut and Mas- are less likely to undergo PSA screening and more like- sachusetts. In a subsequent analysis of the Connecti- ly to present with more advanced disease, so they may cut data, researchers identified an outlier locale with have fewer curative treatment options.43 Another study a significantly higher mortality rate (adjusted for age, found that awareness of and knowledge about prostate disease, and race) than other parts of the state.20 cancer were lower among rural men,46 and another con- Wagner et al23,62 applied GIS methods to visualize ra- firmed that screening and detection practices are signifi- tios of cancer mortality and incidence and detected cantly less common in rural than urban areas.47 Rates of large geographical disparities among blacks in Georgia mortality and survival are also significantly associated compared with national benchmark data. Antwi et al13 with race/ethnicity, socioeconomic status, and accessi- found that men with prostate cancer living in the Ap- bility to health care among men with prostate cancer.48,49 palachian region within Kentucky have lower rates of survival than those living in non-Appalachian regions. Geographical Information System and Its Role in Differences in health resources and treatment pat- Disparity Research terns have also been examined among men with pros- The GIS is a computer system designed to capture, store, tate cancer, and they have been linked to outcomes and analyze, manage, and display spatially referenced data. possible quality-of-care issues. Aneja and Yu12 mapped It was developed in the 1960s by Tomlinson and then the density of radiation oncologists for all counties in

404 Cancer Control October 2016, Vol. 23, No. 4 the United States and found that a higher density rate temporal trends for whites and blacks were also ob- of radiation oncologists was associated with lower served in rural counties of Florida.19 rates of prostate cancer–related mortality. Cetnar et al14 Another study used boundary analysis and report- compared trends in first-line treatment of prostate can- ed significant regional differences in relation to when cer in rural and urban areas in Wisconsin and found a decline of late-stage cases was seen, suggesting that that a person’s place of residence was not associated the introduction of PSA testing and urbanization in sev- with treatment choice. eral areas of Florida led to fewer late-stage diagnoses GIS projects in other states have also examined than that seen in nonurbanized areas, where uptake of disparities related to geography. Research involving screening practices are likely to have been slower.64 Al- men with prostate cancer has shown that the risk of late-stage prostate cancer diagnoses among black men living in Florida consistently exceeds that of white Florid- 1982 ians, although this gap has decreased in recent years.24,25 White males These studies used geographically weighted regression 2006 to account for the geographical variation in racial and sociodemographic composition, which provides estimates more likely to indicate the actual local impact of 1982 sociodemographic and clinical characteristics on patient Black males outcomes.24,25 Using multiscale join-point regression 2006 analysis to investigate geographical, temporal, and racial/ethnic disparities in the incidence rates of late-stage prostate cancer among men living in Florida, researchers Late-Stage Diagnosis (%) illustrated how the proportion of late-stage incidence can 0.2 0.3 0.5 0.4 0.0 change over time.19,23 They observed striking geographi- 0.1 cal and racial disparities within a single state (Fig 1).19,23 Fig 1. — Temporal patterns of late-stage prostate cancer across Florida Although the percentage of late-stage disease was halved counties. Reprinted from Goovaerts P, Xiao H. Geographical, temporal between 1980 and 2000, nonmetropolitan areas lagged and racial disparities in late-stage prostate cancer incidence across behind urban parts of the state, suggesting that late-stage Florida: a multiscale joinpoint regression analysis. Int J Health Geogr. 2011;10:63. © Goovaerts and Xiao; licensee BioMed Central Ltd. 2011. presentation is a larger problem in rural areas compared Licensed under the terms of the Creative Commons Attribution License with nonrural areas in Florida.19 Larger differences in http://creativecommons.org/licenses/by/2.0.

≥ High-volume facility Travel Time (min) The Northwest Florida CCC is notable for relatively large uncovered (> 45 min to high-volume facilities) County boundary 0–15 rural areas characterized by low income and a high percentage of black residents. Tallahassee is the 15–30 ≥ 20% living in poverty (low income) one area covering a relatively significant area of ≥ 17.0% black (top quartile) 30–45 black residents.

Rural (large and small) > 45

Fig 2. — Service area coverage and average travel times to high-volume prostate cancer treatment facilities in the Northwest Florida CCC area. Reproduced from reference 71. CCC = Cancer Control Collaborative.

October 2016, Vol. 23, No. 4 Cancer Control 405 though this gap has narrowed across most Florida coun- and time to high-volume treatment centers.71 Further- ties and the rate of late-stage diagnoses continues to more, low income and rural areas of Florida also appear decrease, some parts of the state still demonstrate per- to be characterized by limited resources for prostate sistent disparity.64 The decline in number of late-stage cancer care and specialist health care professionals diagnoses of prostate cancer has been slowest in Flor- (eg, urologists, radiation oncologists).71 Treatment facili- ida’s Big Bend region.64 In another study, researchers ties are more sparse in the North-Central and Panhandle compared the impact of comorbidities on late-stage di- regions of Florida, thus resulting in longer travel times agnoses at the local level with the expected state-level for residents in those regions that appear to dispropor- risk and found significant differences in observed-to- tionately affect disadvantaged men (Fig 2); by contrast, expected results in Tallahassee and Pensacola, where the southeastern and western parts of the state are the local OR was smaller than expected, and in Palm characterized by a greater number and higher density Beach, where the OR was larger than expected.65 of treatment facilities, shorter average travel times Racial/ethnic differences in how prostate cancer is (particularly along the coastal areas and in high-income, treated have also been documented in Florida through low-minority areas).71 However, low income and black GIS-based studies.66,67 These results point to the impor- communities are still located beyond the close access tance of more granular, local analyses to evaluate poten- bands even in those regions of Florida, thus indicating tial disparities and gaps, although the stability of estimat- a disproportionate and disparate service coverage for ed incidence, treatment, and mortality rates can become disadvantaged men (Fig 3).71 an issue when using small geographical units (eg, zip codes, census tracts), particularly when cases are The Southwest Florida CCC is characterized by clustering of high-volume facilities and further stratified by race, so- service coverage areas along the Gulf Coast cioeconomic status, or both. with large uncovered inland areas that are It is worth noting that em- low income and rural. pirical research suggests that the census tract is the most reliable unit of analysis for spatial analyses.69,70 Another Florida study examining racial/ethnic and sociodemographic disparities focused on access to health care resources and differences in state- wide treatment patterns.71 The Florida Prostate Cancer Atlas Project71 chronicled geographical variation in the burden of disease (eg, High-volume facilities late-stage diagnosis), distribution and availability of County boundaries health care resources rel- ≥ 20% living in poverty (low income) evant to prostate cancer, ≥ 17.0% black (top quartile) as well as access to those Rural (large and small) resources, as assessed by travel times to treatment Travel Time (min) facilities using cancer regis- 0–15 try data from 1998 through 15–30 2007. The results of the 30–45 project demonstrate that low-income areas and areas > 45 characterized by a high percentage of black residents Fig 3. — Service area coverage and average travel times to high-volume prostate cancer treatment facilities were associated with the in the Southwest Florida CCC area. Reproduced from reference 71. greatest distance to travel CCC = Cancer Control Collaborative.

406 Cancer Control October 2016, Vol. 23, No. 4 mutable factors that contribute to health disparities.52 Responses to actionable geographical health information include outreach and awareness cam- paigns, community-based health service efforts, such as screening events, organizing providers and resources into more accessible networks, and creating local infrastructure in underserved communities Age-Adjusted Incidence Rate to allow better access and ≥ 15.8% Above state ratio by ≥ 20% care to needed health services.52 14.5%–15.8% Above state ratio by 10% to < 20%

11.9%–14.4% Within 10% of state ratio References 10.6%–11.8% Below state ratio by 10% to < 20% 1. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2016. CA Cancer ≤ 10.5% Below state ratio by ≥ 20% J Clin. 2016;66(1):7-30. 2. American Cancer Society. Can- Suppressed (cases < 10) cer Facts & Figures 2016. Atlanta, GA: American Cancer Society. http://www. Unstable rate cancer.org/acs/groups/content/@ research/documents/document/ acspc-047079.pdf. Accessed Sep- Fig 4. — Proportion of prostate cancer diagnoses at an advanced stage across Florida counties tember 21, 2016. (age-adjusted incidence rates, 1998–2007). Reproduced from reference 71. 3. National Cancer Institute. Sur- veillance, Epidemiology, and End Results Program. http://seer.cancer. gov/statfacts/html/prost.html. Accessed March 31, 2016. Differences in resource accessibility may be as- 4. Albertsen PC, Hanley JA, Fine J. 20-year outcomes following sociated with differences in how prostate cancer conservative management of clinically localized prostate cancer. JAMA. 2005;293(17):2095-2101. presents and how it is treated. The Florida Prostate 5. D’Amico AV, Whittington R, Malkowicz B, et al. Biochemical out- Cancer Atlas Project71 observed an overlap between come after radical prostatectomy, external beam radiation therapy, or interstitial radiation therapy for clinically localized prostate cancer. JAMA. low health care resource regions and areas in which 1998;280(11):969-974. the case proportion of late-stage or aggressive dis- 6. Freedland SJ, Humphreys EB, Mangold LA, et al. Risk of prostate cancer-specific mortality following biochemical recurrence after radical ease is highest (Fig 4). This geospatial data indicate prostatectomy. JAMA. 2005;294(4):433-439. significant differences in access to health care, treat- 7. Roehrborn CG, Black LK. The economic burden of prostate cancer. BJU Int. 2011;108(6):806-813. ment, and outcomes that underlie and contribute to 8. Pettaway CA. Racial differences in the androgen/androgen receptor racial/ethnic, rural, and socioeconomic disparities, pathway in prostate cancer. J Natl Med Assoc. 1999;91(12):653-660. 9. Ward E, Jemal A, Cokkinides V, et al. Cancer disparities by race/ and they potentially point to areas in which public ethnicity and socioeconomic status. CA Cancer J Clin. 2004;54(2):78-93. health efforts and resources could be directed to ad- 10. Woods LM, Rachet B, Coleman MP. Origins of socio-economic inequalities in cancer survival: a review. Ann Oncol. 2006;17(1):5-19. dress access, awareness, and treatment gaps.71 11. Bennett CL, Ferreira MR, Davis TC, et al. Relation between literacy, race, and stage of presentation among low-income patients with prostate cancer. J Clin Oncol. 1998;16(9):3101-3104. Conclusions 12. Aneja S, Yu JB. The impact of county-level radiation oncologist Geographical variation in the presentation and density on prostate cancer mortality in the United States. Prostate Cancer Prostatic Dis. 2012;15(4):391-396. management of prostate cancer is well document- 13. Antwi S, Tucker TC, Coker AL, et al. Racial disparities in survival ed. Although biological and disease-related factors after diagnosis of prostate cancer in Kentucky, 2001-2010. Am J Mens Health. 2013;7(4):306-316. most likely explain some of this variation, differ- 14. Cetnar JP, Hampton JM, Williamson AA, et al. Place of residence ences in the health care system, such as availability and primary treatment of prostate cancer: examining trends in rural and nonrural areas in Wisconsin. Urology. 2013;81(3):540-546. of and access to treating physicians and treatment 15. Cobran EK, Chen RC, Overman R, et al. Racial differences in diffu- facilities, are likely to contribute to disparate treat- sion of intensity-modulated radiation therapy for localized prostate cancer. Am J Mens Health. 2015;10(5):399-407. ment and outcomes.51 Geographical level analy- 16. Cook MB, Rosenberg PS, McCarty FA, et al. Racial disparities in ses augment disparities research, providing a lens prostate cancer incidence rates by census division in the United States, 1999-2008. Prostate. 2015;75(7):758-763. through which clinicians, public health officials, 17. DeChello LM, Gregorio DI, Samociuk H. Race-specific geography of prostate cancer incidence. Int J Health Geogr. 2006;5:59. and policymakers can identify underserved areas 18. Garg V, Raisch DW, Selig JP, et al. Health disparities in clinical and make informed decisions about how to address practice patterns for prostate cancer screening by geographic regions

October 2016, Vol. 23, No. 4 Cancer Control 407 in the United States: a multilevel modeling analysis. Prostate Cancer health maintenance in siblings of patients with prostate cancer. Eur Urol. Prostatic Dis. 2013;16(2):193-203. 2006;50(1):64-69. 19. Goovaerts P, Xiao H. Geographical, temporal and racial disparities 47. Koh HK, Judge CM, Ferrer B, et al. Using public health data sys- in late-stage prostate cancer incidence across Florida: a multiscale join- tems to understand and eliminate cancer disparities. Cancer Cause Control. point regression analysis. Int J Health Geogr. 2011;10:63. 2005;16(1):15-26. 20. Gregorio DI, Huang L, DeChello LM, et al. Place of residence effect on 48. Xiao H, Tan F, Goovaerts P, et al. Multilevel factors associated with likelihood of surviving prostate cancer. Ann Epidemiol. 2007;17(7):520-524. overall mortality for men diagnosed with prostate cancer in Florida. Am J 21. Meliker JR, Goovaerts P, Jacquez GM, et al. Breast and prostate Mens Health. 2014;8(4):316-326. cancer survival in Michigan: can geographic analyses assist in under- 49. Xiao H, Tan F, Adunlin G, et al. Factors associated with overall sur- standing racial disparities? Cancer. 2009;115(10):2212-2221. vival prostate cancer in Florida: a multilevel analysis. J Health Care Poor 22. Onukwugha E, Osteen P, Jayasekera J, et al. Racial disparities in Underserved. 2015;26(1):266-277. urologist visits among elderly men with prostate cancer: a cohort analy- 50. Coppock JT, Rhind DW. The history of GIS. In: Longley PA, Good- sis of patient-related and county of residence-related factors. Cancer. child MF, Maguire DJ, et al, eds. Geographical Information Systems. 2014;120(21):3385-3392. Hoboken, NJ: John Wiley & Sons; 1991:21-43. http://www.wiley.com/ 23. Wagner SE, Bauer SE, Bayakly AR, Vena JE. Prostate cancer inci- legacy/wileychi/gis/Volume1/BB1v1_ch2.pdf. Accessed September 21, 2016. dence and tumor severity in Georgia: descriptive epidemiology, racial dis- 51. Krieger N, Chen JT, Waterman PD, et al. Choosing area based parity, and geographic trends. Cancer Cause Control. 2013;24(1):153-166. socioeconomic measures to monitor social inequities in low birth weight 24. Xiao H, Gwede CK, Kiros G, et al. Analysis of prostate cancer and childhood lead poisoning: The Public Health Disparities Geocoding incidence using geographic information system and multilevel modeling. Project (US). J Epidemiol Commun Health. 2003;57(3):186-199. J Natl Med Assoc. 2007;99(3):218-225. 52. National Cancer Institute. Geographic information systems and sci- 25. Xiao H, Tan F, Goovaerts P. Racial and geographic disparities in late- ence for cancer control. https://gis.cancer.gov. Accessed September 21, 2016. stage prostate cancer diagnosis in Florida. J Health Care Poor Underserv. 53. Cockburn M, Mills P, Zhang X, et al. Prostate cancer and ambient 2011;22(4 suppl):187-199. pesticide exposure in agriculturally intensive areas in California. Am J 26. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2015. CA Cancer Epidemiol. 2011;173(11):1280-1288. J Clin. 2015;65(1):5-29. 54. Marusek JC, Cockburn MG, Mills PK, et al. Control selection and 27. Mahal BA, Aizer AA, Ziehr DR, et al. Trends in disparate treatment pesticide exposure assessment via GIS in prostate cancer studies. Am J of African American men with localized prostate cancer across National Prev Med. 2006;30(2 suppl):S109-S116. Comprehensive Cancer Network risk groups. Urology. 2014;84(2):386-392. 55. Cary C, Odisho AY, Cooperberg MR. Variation in prostate cancer 28. Barocas DA, Grubb R III, Black A, et al. Association between race treatment associated with population density of the county of residence. and follow-up diagnostic care after a positive prostate cancer screening Prostate Cancer Prostatic Dis. 2016;19(2):174-179. test in the prostate, lung, colorectal, and ovarian screening trial. Cancer. 56. Hsu CE, Mas FS, Miller JA, et al. A spatial-temporal approach to 2013;119(12):2223-2239. surveillance of prostate cancer disparities in population subgroups. J Natl 29. Zeliadt SB, Potosky AL, Etzioni R, et al. Racial disparity in primary Med Assoc. 2007;99(1):72-80, 85-77. and adjuvant treatment for nonmetastatic prostate cancer: SEER-Medicare 57. Dean LT, Subramanian SV, Williams DR, et al. Getting black men to trends 1991 to 1999. Urology. 2004;64(6):1171-1176. undergo prostate cancer screening: the role of social capital. Am J Mens 30. Klein J, von den Knesebeck O. Socioeconomic inequalities in prostate Health. 2015;9(5):385-396. cancer survival: a review of the evidence and explanatory factors. Soc Sci 58. Wagner SE, Burch JB, Hussey J, et al. Soil zinc content, groundwa- Med. 2015;142:9-18. ter usage, and prostate cancer incidence in South Carolina. Cancer Cause 31. Chornokur G, Dalton K, Borysova ME, et al. Disparities at presenta- Control. 2009;20(3):345-353. tion, diagnosis, treatment, and survival in African American men affected 59. Fogleman AJ, Mueller GS, Jenkins WD. Does where you live play by prostate cancer. Prostate. 2011;71:985-997. an important role in cancer incidence in the U.S.? Am J Cancer Res. 32. Auviven A, Karjalainen S. Possible explanations for social class 2015;5(7):2314-2319. differences in cancer patient survival. IARC Sci Publ. 1997;138:377-397. 60. Lu-Yao G, Albertsen PC, Stanford JL, et al. Screening, treatment, 33. Woods VD, Montgomery SB, Herring RP, et al. Social ecological and prostate cancer mortality in the Seattle area and Connecticut: predictors of prostate-specific antigen blood test and digital rectal exami- fifteen-year follow-up. J Gen Intern Med. 2008;23(11):1809-1814. nation in black American men. J Natl Med Assoc. 2006;98(4):492-504. 61. Mossanen M, Izard J, Wright JL, et al. Identification of underserved 34. Fairley L, Baker M, Whiteway J, et al. Trends in non-metastatic areas for urologic cancer care. Cancer. 2014;120(10):1565-1571. prostate cancer management in the Northern and Yorkshire region of 62. Wagner SE, Hurley DM, Hebert JR, et al. Cancer mortality-to- England, 2000-2006. Br J Cancer. 2009;101(11):1839-1845. incidence ratios in Georgia: describing racial cancer disparities and 35. Hayen A, Smith DP, Patel MI, et al. Patterns of surgical care for potential geographic determinants. Cancer. 2012;118(16):4032-4045. prostate cancer in NSW, 1993-2002: rural/urban and socioeconomic 63. Wan N, Zhan FB, Cai Z. Socioeconomic disparities in pros- variation. Aust N Z J Public Health. 2008;32(5):417-420. tate cancer mortality and the impact of geographic scale. South Med J. 36. Krupski TL, Kwan L, Afifi AA, et al. Geographic and socio- 2011;104(8):553-559. economic variation in the treatment of prostate cancer. J Clin Oncol. 64. Goovaerts P, Xiao H. The impact of place and time on the proportion 2005;23(31):7881-7888. of late-stage diagnosis: the case of prostate cancer in Florida, 1981-2007. 37. Lyratzopoulos G, Barbiere JM, Greenberg DC, et al. Population Spat Spatiotemporal Epidemiol. 2012;3(3):243-253. based time trends and socioeconomic variation in use of radiotherapy and 65. Goovaerts P, Xiao H, Adunlin G, et al. Geographically-weighted radical surgery for prostate cancer in the UK region: continuous survey. regression analysis of percentage of late-stage prostate cancer diagnosed BMJ. 2010;240:c1928. in Florida. Appl Geogr. 2015;62:191-200. 38. Ross LE, Taylor YJ, Howard DL. Trends in prostate-specific antigen 66. Xiao H, Warrick C, Huang Y. Prostate cancer treatment patterns among test use, 2000-2005. Public Health Rep. 2011;126(2):228-239. racial/ethnic groups in Florida. J Natl Med Assoc. 2009;101(9):936-943. 39. Williams N, Hughes LJ, Turner EL, et al. Prostate-specific antigen 67. Xiao H, Tan F, Goovaerts P, et al. Factors associated with time-to- testing rates remain low in UK general practice: a cross-sectional study in treatment of prostate cancer in Florida. J Health Care Poor Underserv. six English cities. BJU Int. 2011;108(9):1402-1408. 2013;24(4 suppl):132-146. 40. Yin HS, Dreyer BP, Vivar KL, et al. Perceived barriers to care and 68. Oliver MN, Smith E, Siadaty M, et al. Spatial analysis of prostate attitudes towards shared decision-making among low socioeconomic sta- cancer incidence and race in Virginia, 1990-1999. Am J Prev Med. tus parents: role of health literacy. Acad Pediatr. 2012;12(2):117-124. 2006;30(suppl):S67-S76. 41. Ziehr DR, Mahal BA, Aizer AA, et al. Income inequality and treat- 69. Krieger N, Waterman P, Chen JT, et al. Zip code caveat: bias due ment of African American men with high-risk prostate cancer. Urol Oncol. to spatiotemporal mismatches between zip codes and US census-defined 2015;33:18e7-e13. geographic areas – the Public Health Disparities Project. Am J Pub Health. 42. Harlan L, Brawley O, Pommerenke F, et al. Geographic, age, and 2002;92(7):1100-1102. racial variation in the treatment of local/regional carcinoma of the prostate. 70. Gregorio DI, Dechello LM, Samociuk H, et al. Lumping or splitting: J Clin Oncol. 1995;13(1):93-100. seeking the preferred areal unit for health geography studies. Int J Health 43. Jemal A, Ward E, Wu X, et al. Geographic patterns of prostate can- Geogr. 2005;4(1):6. cer mortality and variations in access to medical care in the United States. 71. Gilbert SM, Henriksen C, Hutton C, et al. Florida Prostate Cancer Cancer Epidemiol Biomark Prevent. 2005;14(3):590-595. Atlas. https://com-uro-advisory-council.sites.medinfo.ufl.edu/files/2014/07/ 44. Jones AP, Haynes R, Sauerzapf V, et al. Travel times to health Atlas_final_with_cover_final.pdf. Accessed September 21, 2016. care and survival from cancers in Northern England. Eur J Cancer. 2008;44(2):269-274. 45. Obertova Z, Brown C, Holmes M, et al. Prostate cancer and mortality in rural men – a systematic review of the literature. Rural Remote Health. 2012;12(2):2039. 46. Pruthi RS, Tornehl C, Gaston K, et al. Impact of race, age, income, and residence on prostate cancer knowledge, screening behavior, and

408 Cancer Control October 2016, Vol. 23, No. 4 Differences in racial/ethnic, sociodemo-

graphical, and economic disparities

in penile cancer can influence cancer

control and survival-related outcomes.

Birds Nest. X-Ray image courtesy of Bryan Whitney. www.x-rayphotography.com.

Disparities in Penile Cancer Pranav Sharma, MD, Kamran Zargar-Shoshtari, MD, Curtis A. Pettaway, MD, Matthew B. Schabath, PhD, Anna R. Giuliano, PhD, and Philippe E. Spiess, MD

Background: Although penile cancer is a rare malignancy in developed nations, racial and socioeconomic differences exist in the incidence of the disease and its associated survival-related outcomes. Methods: A search of the literature was performed for research published between the years 1990 and 2015. Case reports and non–English-language articles were excluded, instead focusing specifically on large, population-based studies. Results: The incidence of penile cancer is higher in Hispanic and African American men compared with whites and Asians. Men with penile cancer also appear to have a distinct epidemiological profile, including lower educational and income levels, a history of multiple sexual partners and sexually transmitted infections, and lack of circumcision with the presence of phimosis. African American men presented at a younger age with a higher stage of disease and worse survival rates when compared with white men. Rates of cancer-specific mortality increased with age, single marital status, and among those living in regions of lower socioeconomic status. Conclusions: An understanding of sociodemographical differences in the incidence and survival rates of patients with penile cancer can help advance health care policy changes designed to improve access and minimize disparities in cancer care for all men alike.

Introduction However, its incidence is higher in some countries of Penile cancer is rare in the United States, with an Asia, Africa, and South America, where it represents estimated 2,030 new cases in the United States and 10% of all malignancies diagnosed among men.2 340 related deaths estimated to occur this year.1 Patients with penile cancer appear to have a distinct epidemiological profile consisting of lower income and education levels, poor hygiene, a history of From the Departments of Genitourinary Oncology (PS, KZ-S, PES) and Cancer Epidemiology (MBS, ARG), H. Lee Moffitt Cancer Center sexually transmitted infections, phimosis, and multiple & Research Institute, Tampa, Florida, and the Department of sexual partners.3 Risk factors for penile cancer include Urology (CAP), University of Texas MD Anderson Cancer Center, lack of circumcision, presence of phimosis or paraphi- Houston, Texas. mosis, poor hygiene, human papillomavirus (HPV) in- Submitted January 21, 2016; accepted February 26, 2016. fection, presence of genital warts, chronic inflamma- Address correspondence to Philippe E. Spiess, MD, Department of Genitourinary Oncology, Moffitt Cancer Center, 12902 Magnolia Drive, tory conditions of the foreskin, glans penis, or penile WCB-GU PROG, Tampa, FL 33612. E-mail: [email protected] shaft (eg, balanitis xerotica obliterans, lichen sclerosis), Dr Sharma is now affiliated with the Department of Urology at Texas presence of smegma (a whitish substance that accumu- Tech University Health Sciences Center in Lubbock, Texas. lates underneath the foreskin), tobacco abuse, or histo- No significant relationships exist between the authors and the com- 4,5 panies/organizations whose products or services may be referenced ry of psoriasis treatment with ultraviolet light. Tumor in this article. grade and stage, basaloid or sarcomatoid histological

October 2016, Vol. 23, No. 4 Cancer Control 409 features, and the presence and extent of lymph-node (SEER) program of the National Cancer Institute. The (LN) involvement in the groin or pelvis have all been annual, average age-adjusted incidence rate of pe- identified as important prognostic factors for penile nile SCC was 0.81 cases per 100,000 men, and rates cancer in determining long-term, cancer-related surviv- steadily increased with age.8 The incidence of penile al outcomes.6 In addition to disease-specific character- SCC was comparable in whites and African Americans, istics, significant demographical, racial/ethnic, social, but it was approximately twofold lower in Asians and and economic disparities exist with regard to both the Pacific Islanders.8 Rates among Hispanics were 72% incidence and survival rates of penile cancer.7 This is higher compared with non-Hispanics, which may be especially important considering the diverse patient attributable to a decreased prevalence of circumci- population in which this disease presents.2 These find- sion in this ethnic group.11 From 1993 to 2002, white ings also highlight the importance of considering these Hispanics had the highest incidence rates of penile differences when developing health care policies to cancer (1.01 per 100,000 men) followed by Alaska Na- help improve access to standard medical care and re- tives/American Indians (0.77 per 100,000 men) and Af- duce the treatment gap in cancer management. rican Americans (0.62 per 100,000 men).11 Incidence of penile SCC was also elevated in the southern United Methods States and in regions of lower socioeconomic status. A search of the medical literature was performed for However, the study had limitations because the educa- research published between January 1990 and Decem- tion level of participants was not analyzed. ber 2015. Case reports and non–English-language ar- Colón-López et al10 compared the incidence of pe- ticles were excluded, focusing instead on large, pop- nile cancer in Puerto Rico with other racial and ethnic ulation-based studies from national or international groups in the United States. Puerto Rican men had a datasets. A total of 54 studies were initially identified, higher incidence of penile cancer than non-Hispanic 30 of which are included in this review. white men (standardized rate ratio [SRR] 3.33), non- Hispanic African American men (SRR 3.04), and oth- Incidence er US Hispanic men (SRR 2.59).10 This difference was The incidence of penile cancer varies within spe- more pronounced for Puerto Rican men younger than cific racial/ethnic and socioeconomic groups, and a 60 years of age when compared with similarly aged US summary of the reviewed literature analyzing dispari- counterparts.10 Puerto Rican men with the lowest so- ties in penile cancer incidence based on large popula- cioeconomic status had a 70% higher incidence rate of tion-based data sets is shown in Table 1.8-10 penile cancer than Puerto Rican men with the highest Hernandez et al8 examined the burden of inva- socioeconomic status (SRR 1.70).10 Less education sive squamous cell carcinoma (SCC) of the penis in (< 12 years of education) was also statistically associat- the United States from 1998 to 2003 based on data ed with a higher incidence rate of penile cancer among from the Surveillance, Epidemiology, and End Results Puerto Rican men (SRR 2.18).10 However, the overall rate of penile cancer in Puerto Rico declined over time. In other studies, Hispanic and African American Table 1. — Overview of the Disparities in Incidence of Penile men have been reported to have the highest age-ad- Cancer in Select Population-Based Studies justed rates of penile cancer compared with whites, Study Data Set Select Findings Asians, or American Indians.9,12 The highest rates of Colón-López10 Puerto Rico Puerto Rican men had higher penile cancer were observed in Hispanic and African Cancer Registry rates of penile cancer than other American men older than 85 years of age, but penile SEER program groups cancer was rare among males younger than 20 years of Lower socioeconomic status and educational level associated age. Analyses to assess temporal trends from 1995 to with higher incidence rates in 2003 revealed a statistically significant decline in the Puerto Rican men incidence of penile cancer for African Americans and Goodman9 US population- Hispanic and African American men whites. Approximately 60% of penile cancers were di- based registries had highest rates of penile cancer agnosed at a localized stage (≤ pT2) among all racial African American men diagnosed and ethnic groups, although Hispanic and African at advanced stage with more node- positive disease American men tended to be diagnosed at an advanced Highest incidence in the southern stage (≥ pT3) with a higher incidence of node-positive United States disease than whites.9 Racial and ethnic differences in Hernandez8 SEER program Penile cancer rates higher in tumor grade were not found. The incidence of penile Hispanics and those living in regions cancer was highest in the southern United States and of lower socioeconomic status or in the southern United States lowest in the western states. The highest age-adjusted incidence rate of penile SCC was seen in African Amer- SEER = Surveillance, Epidemiology, and End Results. ican men living in the southern United States, and the

410 Cancer Control October 2016, Vol. 23, No. 4 lowest rate was found among Asian-Pacific Islanders tumors when compared with patients who had fewer living in western states.9 The incidence of primary than 6 lifetime female sexual partners.3 However, this malignant penile cancer has decreased significant- trend was not significant when the number of sexual ly over time from 1973 to 2002 based on data from partners was adjusted for age at first coitus and past SEER; however, the incidence of nodal disease has in- history of sexually transmitted infections. HPV-posi- creased over time.13 tive tumors were found in 36% of cases studied, and The overall incidence of penile cancer in the they were characterized by warty or basaloid morphol- United Kingdom (England and Wales) is higher than ogy in addition to a high histological grade in the main Australia and the United States, although the inci- jority of cases.3 dence of penile cancer in all 3 countries was stable over time.14 Although cancer-specific mortality rates Treatment were higher in England and Wales compared with The management of penile carcinoma has evolved Australia and the United States, the incidence ra- in recent years, resulting in a shift in the treatment of tios were similar for all 3 countries.14 In Finland, the the disease.24 Options for the treatment of localized age-adjusted incidence rate of penile cancer has de- disease have expanded to organ-sparing therapy, and creased from the 1960s onward.15 Penile cancer was treatment of advanced disease has developed into a primarily a disease of elderly men during this time multidisciplinary, team-based approach using multi- in Finland, and no systematic geographical variation modal therapies.24 Demographical, racial/ethnic, and or predilection in terms of incidence was observed.15 socioeconomic disparities can influence the treatments The most common predisposing risk factors in Finland offered for patients with penile cancer in different pop- were phimosis (44% of cases) and condylomatous ulations (Table 2).25,26 genital lesions (20% of cases).15 Zhu et al26 used data from SEER to identify individuals diagnosed with penile SCC from 1998 to 2009 Human Papillomavirus Infection and treated with either local tumor excision or par- An estimated 40% to 80% of invasive penile cancers tial or total penectomy for their primary tumor. Of the have been linked to HPV infection.16-18 Warty and ba- 1,292 eligible patients, 24.2% underwent local tumor saloid penile carcinomas have been associated with excision for surgical treatment of penile cancer.26 HPV infection, particularly HPV types 16 and 18, more For stage T1 disease, the rates of local tumor exci- so than typical SCC or verrucous carcinoma of the pe- sion increased from 29% to 40% within 10 years.26 nis.19,20 No significant association between HPV status Following a multivariable analyses, age younger than and pathological tumor stage, grade, or LN status was 65 years, African American descent, tumor size of found in this study.19 less than 3 cm, and pathological stage disease of no The impact of HPV infection on the prognosis of higher than T1 were reported predictors for the use patients with penile cancer is still controversial. Lont of local tumor excision as surgical treatment of the et al21 reported an association between high-risk HPV primary penile tumor as part of a penile-sparing ap- infection and reduced disease-specific mortality from proach.26 The authors concluded that penile-sparing penile cancer (hazard ratio [HR] 0.14), whereas Her- surgery is underutilized in the general population, nandez et al16 observed no association between HPV and significant age and racial/ethnic disparities exist status and rate of survival in patients with penile can- in its use.26 cer. HPV-associated penile malignancies, similar to Other studies have noted that the majority of pa- other nonviral penile cancers, have been associated with lower education level (incidence density ratio 1.32; P = .0002) and income status (incidence density Table 2. — Overview of the Disparities in Treatment for ratio 1.59; P < .0001).22 Current and former smokers Penile Cancer in Select Population-Based Studies were significantly more likely to have an α-HPV infec- Study Data Set Select Findings 23 tion than any other type. Zhu25 SEER pro- Men < 50 y of age more likely to receive In a study by Chaux et al3 of 103 Paraguayans with gram extensive LN dissection (≥ 8 LNs removed) penile cancer, patients with penile cancer were gener- during inguinal LN dissection for high-risk penile cancer ally found to dwell in rural or suburban areas (82%), Zhu26 SEER pro- More likely to receive penile-sparing approach live below the poverty line (75%), have a lower level of gram for surgical treatment of primary tumor: education (91%), and have a significant smoking his- Age < 65 y tory (76%). Phimosis (57%), moderate or poor hygienic African American decent habits (90%), and a history of sexually transmitted in- Tumor size < 3 cm 3 fections (74%) were also observed. Patients with more Pathological stage ≤ T1 than 10 lifetime female sexual partners were nearly LN = lymph node, SEER = Surveillance, Epidemiology, and End Results. 4 times as likely to present with HPV-positive penile

October 2016, Vol. 23, No. 4 Cancer Control 411 tients identified with penile SCC in the SEER data set Table 3. — Overview of the Disparities in Survival Rates of were treated with surgical therapy, and a small per- Penile Cancer in Select Population-Based Studies centage of patients received radiotherapy alone or as adjuvant therapy.27,28 No statistically significant so- Study Data Findings ciodemographic or racial/ethnic differences were re- Set ported with regard to radiotherapy for penile cancer Hernandez8 SEER Penile SCC mortality elevated in: program due to small sample size.28 African American men Zhu et al25 also assessed the impact of patient Hispanic men characteristics on the extent of inguinal LN dissec- Elderly (age > 80 y) tion when it is clinically indicated for patients with Southern US residents high-risk penile SCC. Using SEER data, the authors Low-income areas identified 393 patients who underwent regional LN dis- Less-educated populations section for penile cancer; this study cohort was then Rippentrop27 SEER African American race independent stratified into 2 groups, limited LN dissection (< 8 LNs program predictor of worse cancer-specific survival removed) and extensive LN dissection (≥ 8 LNs re- Sharma31 NCDB African American race independent predictor of worse OS moved).25 The median number of dissected LNs in this Private insurance and higher median series was 15, and 28% of patients underwent limited income independent predictors of better OS 25 LN dissection. The prevalence of extensive LN dissec- Thuret32 SEER Single marital status independent predictor of: tion decreased in tandem with increasing age (81% in program Locally advanced primary tumor 25 men aged < 50 years to 65% in men aged ≥ 70 years). Higher-grade disease On multivariable analysis, age alone was the inde- Increased overall mortality pendent predictor of extensive LN dissection (odds Tyson30 SEER Tumors on penile shaft and overlapping 25 ratio 0.98; P = .01). Inadequate LN dissection with few program lesions increased cancer-specific mortal- LNs removed was observed in some patients with pe- ity vs preputial lesions nile cancer, particularly among elderly men, even de- Verhoeven3 SEER 5-year cancer-specific survival of penile spite evidence that inguinal lymphadenectomy has a program cancer in United States decreased from similar morbidity rate in this age group compared with 1998 to 2011 men younger than 65 years of age.25,26 NCDB = National Cancer Database, OS = overall survival, SEER = Surveillance, Epidemiology, and End Results. Survival Factors affecting health insurance coverage and ac- had a 1.5-fold higher risk of having locally advanced cess to early detection and treatment could influence (pT3/T4, N+, M+) or higher grade (grade 3/4) disease the cancer-specific mortality rate of penile cancers. A at the time of surgery (P < .001) and a 1.3-fold higher summary of the literature analyzing disparities in pe- risk of overall mortality (P = .001); however, marital nile cancer survival rates based on select, large popula- status had no effect on cancer-specific mortality.32 Age, tion-based data sets is shown in Table 3.3,8,27,30-32 socioeconomic status (median family income above Rippentrop et al27 first described the racial and or below the poverty line within the patient’s county socioeconomic disparity in survival from penile car- of residence), and race (white vs African American vs cinoma based on data from SEER. They reported that other) did not correlate with rates of locally advanced African American men presented at a younger age and or higher-grade disease and cancer-specific mortality.32 with a higher stage of disease compared with other However, the study was limited by its small sample size racial groups. The disease-specific risk of death for of African Americans (n = 183). African American men was also nearly twice as high In a multivariable competing-risk model, Thuret than for whites (P = .0023), and marriage was associ- et al33 studied a population of patients with cN0T1 peated with improved disease-specific rates of survival in nile SCC after primary tumor excision and before ini- the localized (P = .0002) and regional stages (P = .001) tial LN dissection and found no statistically significant of disease.27 Multivariable analysis demonstrated that a association between race and age (≥ 70 vs ≤ 69 years) higher stage at diagnosis, age older than 65 years, Af- for cancer-specific mortality after accounting for rican American race, and positive LNs were all signifi- tumor grade. Tyson et al30 also found no significant independent predictors of cancer-specific survival cant association between age and penile SCC-spe- from penile carcinoma.27 cific mortality rate after adjusting for tumor grade Thuret et al32 also used SEER data to evaluate the based on SEER data. Penile tumors on the body of association between marital status, tumor stage and the penis (HR 1.61) and overlapping lesions (HR grade, overall mortality, and cancer-specific mortal- 1.79; P = .01) had increased cancer-specific mortality ity in patients with penile SCC. Multivariable logis- compared with preputial lesions.30 Furthermore, the tic regression models revealed that unmarried men disease-specific 10-year survival rate of those with

412 Cancer Control October 2016, Vol. 23, No. 4 preputial tumors was 89.4% compared with 78.7% for cally significant higher mortality rate when compared tumors at other locations combined (P < .0001).30 with non-Hispanic white men (HR 3.32), non-Hispanic Hernandez et al8 analyzed more than 4,900 cases African American men (HR 2.51), and other US Hispan- of histologically confirmed invasive penile SCC in the ic men (HR 1.89).10 A higher rate of mortality related United States using SEER data, the Centers for Disease to penile cancer was also observed among Puerto Ri- Control and Prevention’s National Program for Cancer can men in the lowest socioeconomic status (HR 1.61) Registries, and the National Center for Health Statis- when compared with Puerto Rican men in the highest tics. The authors found that the rate of mortality due socioeconomic status, although this difference was not to penile SCC increased with age (mortality rate at age statistically significant.10 An increasing trend in mortal- < 50 years: 0.02 vs age > 80 years: 1.80).8 African Amer- ity due to penile cancer was observed among men in ican men were also diagnosed at a significantly young- Puerto Rico due to a possible convergence of early and er age than white or Asian men (median age, 62 vs 68 high levels of sexual activity, low circumcision rates, vs 68 years, respectively) and had a lower incidence-to- and high prevalence rates of sexually transmitted in- mortality ratio than white or Asian men (3.44 vs 5.18 fections in Puerto Rico — particularly among the medi- vs 5.63, respectively; P < .01 for all).8 This was also true cally underserved. for Hispanic men compared with non-Hispanic men Paiva et al34 reported an increased cancer-specif- (median age, 58 vs 69 years; incidence-to-mortality ra- ic mortality rate in patients younger than 40 years of tio: 4.61 vs 5.25; P < .01).8 Rate of mortality due to pe- age at a mean follow-up period of 22.4 months (19%) nile SCC was also increased among those men living in compared with patients aged between 40 and 60 years southern states, in regions of lower socioeconomic sta- (11%) and those older than 60 years of age (13%; tus, and among those living in areas with lower levels P < .05). Younger patients also had an increased like- of educated people.8 lihood of perineural invasion (26% vs 13%) and infil- Sharma et al31 used the National Cancer Data- trative growth pattern (70% vs 50%) on pathological base (NCDB) to study patients with a diagnosis of review, and they had a higher risk of recurrence penile SCC from 1998 to 2011, and they found that (45% vs 25%; P < .05 for all).34 Study findings from African American and Hispanic men were diag- Sharma et al31 similarly showed an increased preva- nosed at a younger age (< 55 years: 38.3% vs 25.9% lence of node-positive disease in patients younger than and 51.0% vs 24.5%; P < .01), and African Americans 55 years of age compared with those older than presented with a higher stage of disease (pT3/T4: 75 years (21.5% vs 13.1%; P < .01) but a decreased prev- 16.6% vs 13.2%; P = .027) and had worse median over- alence of pT3 to T4 disease (12.0% vs 15.4%; P = .007). all survival (OS) rates (68.6 vs 93.7 months; P < .01). Others have also shown younger men (< 50 years) were The estimated median OS rate in the entire cohort more likely to receive more extensive LN dissection was 91.9 months at a median follow-up period of (> 8 LNs removed) than older men (> 70 years; 44.7 months, and survival did not change over the 81% vs 65%; P = .01) with better rates of cancer-spe- study period.31 Hispanic ethnicity was not associated cific survival P( = .006).25 These findings remained with all-cause mortality in the penile cancer popula- true even in the subgroup of study patients with tion.31 Location and education level also did not have node-positive penile cancer (P = .01).25 Older age was a significant relationship with OS. Patients with pri- an independent predictor of OS in the population- vate health insurance and a median income of at least based study using the NCDB on multivariable analysis $63,000 based on zip code presented with a lower (P < .01), but its relationship with cancer-specific mor- stage of disease (pT3/T4: 11.6% vs 14.7%, P = .002; tality is unknown.31 12.0% vs 14.0%, P = .042) and had better median rates Large population-based data have also been used of OS (163.2 vs 70.8 months, P < .01; 105.3 vs 86.4 to demonstrate penile cancer survival trends over time. months, P = .001).31 Multivariable analysis revealed that The NCDB study did not show a correlation between African American race (HR 1.39; 95% confidence inter- the year of diagnosis and OS during the period from val [CI]: 1.21–1.58; P < .01) was independently associat- 1998 to 2011.31 However, Verhoeven et al35 reported ed with worse OS, whereas private insurance (HR 0.79; that the 5-year cancer-specific survival rate of pa- 95% CI: 0.63–0.98; P = .028) and higher median income tients with penile cancer in the United States has sig- of at least $63,000 (HR 0.82; 95% CI: 0.72–0.93; P = .001) nificantly decreased from 72% to 63% based on SEER were independently associated with better OS.31 data, and they also confirmed a significantly increas- Colón-López et al10 compared the mortality rates ing relative risk of death with increasing age (relative of penile cancer in Puerto Rican men with other ra- increase risk of death = 1.15; P < .01). Other groups cial and ethnic groups in the United States and evalu- have also observed an increase in the incidence of ated the influence of socioeconomics and educational regional-stage disease (pN+) over time in US men that level on the rates of mortality observed in Puerto Rico. also paralleled rising age.36 Furthermore, Arya et al37 From 2000 to 2004, Puerto Rican men had a statisti- reported on more than 9,500 men diagnosed with pri-

October 2016, Vol. 23, No. 4 Cancer Control 413 pathologic differences in the incidence of invasive penile cancer in mary penile cancer in England from 1979 to 2009 and the United States, 1995-2003. Cancer Epidemiol Biomarkers Prev. noted that the mortality rates fell by 20% after 1994 2007;16(9):1833-1839. 10. Colón-López V, Ortiz AP, Soto-Salgado M, et al. Penile cancer dis- (from 0.39 to 0.31 per 100,000 men). The 5-year sur- parities in Puerto Rican men as compared to the United States population. vival rate they observed also increased from 61.4% to Int Braz J Urol. 2012;38(6):728-738. 11. Laumann EO, Masi CM, Zuckerman EW. Circumcision in the United 70.2%, and the 5-year survival rate for men diagnosed States. Prevalence, prophylactic effects, and sexual practice. JAMA. between 2006 and 2010 was 77% for men younger than 1997;277(13):1052-1057. 12. Slopnick EA, Kim SP, Kiechle JE, et al. Racial disparities differ for 60 years of age and 53% for men older than 80 years African Americans and hispanics in the diagnosis and treatment of penile of age.37 The 8% difference in 5-year survival rate (66% cancer. Urology. 2016;96:22-28. 13. Barnholtz-Sloan JS, Maldonado JL, Pow-sang J, et al. Incidence vs 74%) between men living in the most affluent vs the trends in primary malignant penile cancer. Urol Oncol. 2007;25(5):361-367. most deprived areas during this time was not statisti- 14. Sewell J, Ranasinghe W, De Silva D, et al. Trends in penile cancer: 37 a comparative study between Australia, England and Wales, and the US. cally significant. The authors propose that the 21% Springerplus. 2015;4:420. increase in incidence of penile cancer since the 1970s 15. Maiche AG. Epidemiological aspects of cancer of the penis in Finland. Eur J Cancer Prev. 1992;1(2):153-158. in England could be explained by changes in sexual 16. Hernandez BY, Goodman MT, Unger ER, et al. Human papillomavirus habits, greater exposure to sexually transmitted infec- genotype prevalence in invasive penile cancers from a registry-based United States population. Front Oncol. 2014;4:9. tions, such as HPV infection, and the decreasing rates 17. Kirrander P, Kolaric A, Helenius G, et al. Human papillomavirus of childhood and neonatal circumcision.37 They also prevalence, distribution and correlation to histopathological parameters in a large Swedish cohort of men with penile carcinoma. BJU Int. suggest that the improvement in survival may be due 2011;108(3):355-359. to advances in medical care, including diagnostic, stag- 18. Miralles-Guri C, Bruni L, Cubilla AL, et al. Human papillomavi- 37 rus prevalence and type distribution in penile carcinoma. J Clin Pathol. ing, and surgical techniques. 2009;62(10):870-878. 19. Gregoire L, Cubilla AL, Reuter VE, et al. Preferential association of human papillomavirus with high-grade histologic variants of penile-invasive Conclusions squamous cell carcinoma. J Natl Cancer Inst. 1995;87(22):1705-1709. Disparities exist in the incidence, treatment, and out- 20. Rubin MA, Kleter B, Zhou M, et al. Detection and typing of human papillomavirus DNA in penile carcinoma: evidence for multiple independent comes of penile carcinoma. Younger patients are often pathways of penile carcinogenesis. Am J Pathol. 2001;159(4):1211-1218. offered more penile-sparing approaches for primary 21. Lont AP, Kroon BK, Horenblas S, et al. Presence of high-risk human papillomavirus DNA in penile carcinoma predicts favorable outcome in survival. tumors but more aggressive treatment such as lymph- Int J Cancer. 2006;119(5):1078-1081. adenectomy for nodal disease. African American men, 22. Benard VB, Johnson CJ, Thompson TD, et al. Examining the association between socioeconomic status and potential human papillomavirus- Hispanics, the uninsured, those who are less educat- associated cancers. Cancer. 2008;113(10 suppl):2910-2918. ed, and those with lower socioeconomic status appear 23. Sichero L, Pierce Campbell CM, Fulp W, et al. High genital prevalence of cutaneous human papillomavirus DNA on male genital skin: the to have lower rates of survival when compared with HPV Infection in Men Study. BMC Infect Dis. 2014;14:677. their insured, educated, white, non-Hispanic counter- 24. Sharma P, Zargar-Shoshtari K, Spiess PE. Current surgical management of penile cancer. Curr Probl Cancer. 2015;39(3):147-157. parts. Marital status, number of sexual partners, and 25. Zhu Y, Gu CY, Ye DW. Population-based assessment of the number human papillomavirus status also seem to play a role of lymph nodes removed in the treatment of penile squamous cell carcinoma. Urol Int. 2014;92(2):186-193. in prognosis. Although the treatment options and rates 26. Zhu Y, Gu WJ, Wang HK, et al. Surgical treatment of primary disease of survival from penile squamous cell carcinoma have for penile squamous cell carcinoma: a Surveillance, Epidemiology, and End Results database analysis. Oncol Lett. 2015;10(1):85-92. improved during the past several decades, further un- 27. Rippentrop JM, Joslyn SA, Konety BR. Squamous cell carcinoma derstanding of the racial/ethnic, socioeconomic, and of the penis: evaluation of data from the surveillance, epidemiology, and end results program. Cancer. 2004;101(6):1357-1363. age-related differences seen in penile malignancies 28. Burt LM, Shrieve DC, Tward JD. Stage presentation, care patterns, may help minimize future disparities in cancer care. and treatment outcomes for squamous cell carcinoma of the penis. Int J Radiat Oncol Biol Phys. 2014;88(1):94-100. 29. Gopman JM, Djajadiningrat RS, Baumgarten AS, et al. Predicting References postoperative complications of inguinal lymph node dissection for penile cancer in an international multicentre cohort. BJU Int. 2015;116(2):196-201. 1. American Cancer Society. Cancer Facts and Figures 2016. http:// 30. Tyson MD, Etzioni DA, Wisenbaugh ES, et al. Anatomic site-specific www.cancer.org/acs/groups/content/@research/documents/document/ disparities in survival outcomes for penile squamous cell carcinoma. Urology. acspc-047079.pdf. Accessed October 4, 2016. 2012;79(4):804-808. 2. Bleeker MC, Heideman DA, Snijders PJ, et al. Penile cancer: epide- 31. Sharma P, Ashouri K, Zargar-Shoshtari K, et al. Racial and economic miology, pathogenesis and prevention. World J Urol. 2009;27(2):141-150. disparities in the treatment of penile squamous cell carcinoma: results from 3. Chaux A, Netto GJ, Rodriguez IM, et al. Epidemiologic profile, sexual the National Cancer Database. Urol Onc. 2016;34(3):122. history, pathologic features, and human papillomavirus status of 103 patients 32. Thuret R, Sun M, Budaus L, et al. A population-based analysis of with penile carcinoma. World J Urol. 2013;31(4):861-867. the effect of marital status on overall and cancer-specific mortality in patients 4. Stern RS, Bagheri S, Nichols K, et al. The persistent risk of genital with squamous cell carcinoma of the penis. Cancer Causes Control. tumors among men treated with psoralen plus ultraviolet A (PUVA) for psoriasis. 2013;24(1):71-79. J Am Acad Dermatol. 2002;47(1):33-39. 33. Thuret R, Sun M, Abdollah F, et al. Competing-risks analysis in patients 5. Tsen HF, Morgenstern H, Mack T, et al. Risk factors for penile with T1 squamous cell carcinoma of the penis. BJU Int. 2013;111(4 pt B):E174-E179. cancer: results of a population-based case-control study in Los Angeles 34. Paiva GR, de Oliveira Araújo IB, Athanazio DA, et al. Penile cancer: County (United States). Cancer Causes Control. 2001;12(3):267-277. impact of age at diagnosis on morphology and prognosis. Int Urol Nephrol. 6. Jayaratna IS, Mitra AP, Schwartz RL, et al. Clinicopathologic char- 2015;47(2):295-299. acteristics and outcomes of penile cancer treated at tertiary care centers 35. Verhoeven RH, Janssen-Heijnen ML, Saum KU, et al. Population- in the Western United States. Clin Genitourin Cancer. 2014;12(2):138-142. based survival of penile cancer patients in Europe and the United States of 7. Pow-Sang MR, Ferreira U, Pow-Sang JM, et al. Epidemiology and America: no improvement since 1990. Eur J Cancer. 2013;49(6):1414-1421. natural history of penile cancer. Urology. 2010;76(2 suppl 1):S2-S6. 36. Barnholtz-Sloan JS, Maldonado JL, Pow-Sang J, et al. Incidence 8. Hernandez BY, Barnholtz-Sloan J, German RR, et al. Burden of inva- trends in primary malignant penile cancer. Urol Oncol. 2007;25(5):361-367. sive squamous cell carcinoma of the penis in the United States, 1998-2003. 37. Arya M, Li R, Pegler K, et al. Long-term trends in incidence, survival Cancer. 2008;113(10 suppl):2883-2891. and mortality of primary penile cancer in England. Cancer Causes Control. 9. Goodman MT, Hernandez BY, Shvetsov YB. Demographic and 2013;24(12):2169-2176.

414 Cancer Control October 2016, Vol. 23, No. 4 The underlying biology of carcinogenesis

should be addressed in African American

men at high risk for prostate cancer.

Rose. X-Ray image courtesy of Bryan Whitney. www.x-rayphotography.com.

Chemoprevention in African American Men With Prostate Cancer Nagi B. Kumar, PhD, RD, Julio M. Pow-Sang, MD, Philippe E. Spiess, MD, Jong Y. Park, PhD, Ganna Chornokur, PhD, Andrew R. Leone, MD, and Catherine M. Phelan, PhD, MD

Background: Recommendations for cancer screening are uncertain for the early detection or prevention of prostate cancer in African American men. Thus, chemoprevention strategies are needed to specifically target African American men. Methods: The evidence was examined on the biological etiology of disparities in African Americans related to prostate cancer. Possible chemopreventive agents and biomarkers critical to prostate cancer in African American men were also studied. Results: High-grade prostatic intraepithelial neoplasia may be more prevalent in African American men, even after controlling for age, prostate-specific antigen (PSA) level, abnormal results on digital rectal examination, and prostate volume. Prostate cancer in African American men can lead to the overexpression of signaling receptors that may mediate increased proliferation, angiogenesis, and decreased apoptosis. Use of chemopreventive agents may be useful for select populations of men. Conclusions: Green tea catechins are able to target multiple pathways to address the underlying biology of prostate carcinogenesis in African American men, so they may be ideal as a chemoprevention agent in these men diagnosed with high-grade prostatic intraepithelial neoplasia.

From the Cancer Epidemiology (NBK, JYP, GC, CMP) and Genitouri- Introduction nary Oncology Programs (NBK, JMP-S, PES), H. Lee Moffitt Cancer Center & Research Institute, Tampa, Florida, and the Departments Despite treatment advances in recent years, prostate of Oncologic Sciences (NBK, PES, JYP, GC, CMP), Urology (PES, ARL), cancer remains the leading cause of cancer-related Molecular Medicine (JYP), Cell, Molecular and Microbiology (JYP), death among men in the United States.1 In 2016, the and Epidemiology and Biostatistics (JYP), University of South Flori- da, Tampa, Florida. American Cancer Society estimates that 180,890 new Submitted March 1, 2016; accepted July 6, 2016. cases of prostate cancer will be diagnosed in the Address correspondence to Nagi B. Kumar, PhD, RD, Department United States, and 26,120 men will die from the dis- of Epidemiology, Moffitt Cancer Center, 12902 Magnolia Drive, ease.1 Racial and geographical differences have been MRC-CAN CONT, Tampa, FL 33612. E-mail: [email protected] observed, including a 40-fold difference in incidence No significant relationships exist between the authors and the companies/organizations whose products or services may be ref- rates between low-risk (Chinese men) and high-risk erenced in this article. populations (African American men).2 In the United The development of this manuscript was supported in part by States, age-standardized incidence rates for prostate the National Cancer Institute (grants nos. 1R21CA179659 and cancer are 272 per 100,000 men for African Americans R01CA128813). The work was made possible by research funding 2 awarded by the National Institute of Health and the National Cancer compared with 164 per 100,000 men for whites. Com- Institute (grant no. R01 CA12060-01A1). pared with white men, African American men with

October 2016, Vol. 23, No. 4 Cancer Control 415 prostate cancer have more aggressive disease, higher ated with inherited genetic risk factors.15 rates of incidence, are diagnosed at a younger age, Genetic risk factors are also significant at a young- present with more advanced disease at diagnosis, and er age, and the attributed risk of inherited susceptibil- have a worse prognosis.3 The mortality rate for African ity is thought to be as high as 40% among men diag- American men with prostate cancer is 3 times higher nosed with prostate cancer at 55 years or younger.13 than white men, although the rate decreased in both Although genome-wide association studies (GWASs) groups between 1990 and 2000.2 and large genetic variation studies have identified With disparate rates of prostate cancer as well as more than 100 prostate cancer risk loci, elucidating aggressive prostate cancer observed in African Ameri- the biological basis for these associations is challeng- can men, the logical approach is to focus efforts on ing.16-18 Identified risk loci include the noncoding vari- prostate cancer screening for early detection and pre- ants, such as those located in the 8q24 region, as well vention.4,5 However, uncertainties still exist about the as polymorphisms in the gene-coding regions that ei- overall value of early detection. Although periodic test- ther alter, or are predicted to alter, protein expression ing for serum concentrations of prostate-specific anti- (eg, HNF1B, TERT, RNASEL, KLK3).16,19-22 gen (PSA) may reduce the mortality of prostate cancer, HNF1B single-nucleotide polymorphisms PSA testing has been linked to increased diagnoses (rs7501939 and rs4430796) have been identified in and overtreatment of clinically insignificant, potential- GWASs of prostate cancer and are associated with a ly indolent tumors that pose little risk of metastasis or risk of prostate cancer in African American men.23-25 In death.5,6 Taking the accumulated evidence into consid- addition, the CTBP2 single-nucleotide polymorphism eration, the US Preventive Services Task Force recom- rs4962416 is associated with risk of prostate cancer in mended against PSA-based prostate cancer screening white men.23-25 HNF1B encodes a transcription factor in asymptomatic men.7 However, these recommenda- protein that forms heterodimers with other members tions note that, in African American men, firm conclu- of the HNF1 family and can influence the gene tran- sions cannot be made to balance the risks and benefits scription. While it was previously believed that HNF1B of such screening because data do not exist to support expression is specific to the liver, its transcripts have a more favorable risk–benefit ratio.7 Two cancer screen- been identified in various tissues, including bone mar- ing trials included a majority of men with European row, and the pancreas, urinary tract, gastrointestinal ancestry, thus largely precluding conclusions specifi- organs, and prostate.26 Mutations in HNF1B cause cally pertaining to men of African descent.6,8 Hence, type 5 maturity-onset diabetes that may be accompa- no specific evidence-based recommendation regard- nied by urinary tract disorders.27,28 Some men with ing prostate cancer screening exists for the high risk HNF1B mutations have malformations in the reproduc- population of African American men, underscoring the tive tract, including epididymal cysts, agenesis of the vas need for chemoprevention strategies targeting African deferens, or infertility due to abnormal spermatozoa.29 American men.2,3 Post-GWASs are suggestive of the interaction between genetic variants and environmental risk Biological Etiology of Disparity factors, but our understanding of this is still inad- The etiology of increased susceptibility of African equate.30,31 Rs7501939 at HNF1B has been shown to American men to prostate cancer has not been eluci- increase the risk of prostate cancer in African Ameri- dated, but it is likely multifactorial involving genetic, can men who are obese but not in African American biological, sociocultural, and lifestyle determinants. men who are not obese or European American men The initiation and progression of prostate cancer in- of any body weight.23 Although these are preliminary volves a complex series of events. During progression, findings, they suggest that weight-loss interventions genetic changes and loss of cellular control are ob- may decrease the risk of prostate cancer in African served as cell phenotypes change from normal to dys- American men who are carriers of the high-risk allele plasia (prostatic intraepithelial neoplasia), severe dys- at rs7501939 of HNF1B. Future studies should further plasia (high-grade prostatic intraepithelial neoplasia examine this observation. [HGPIN]), clinically localized disease, and to metastatic CTBP2 encodes a transcriptional co-repressor disease.9-12 Observations point to the role of genetic activated under stress and can mediate the stress-in- susceptibility factors in human prostate cancer.9-15 Oth- duced migration of tumor cells.32 CTBP2 expression er than older age, African ancestry and family history is detected in the prostate and has been linked to de- of prostate cancer are well-established, nonmodifiable creased PTEN expression and activation of the phos- risk factors for prostate cancer.9-13 The lifetime risk of phatidylinositol 3-kinase pathway, which may support prostate cancer increases 1.5- to 4-fold in men with 1 or or promote the growth of prostate cancer.33 2 first-degree relatives with prostate cancer.14 A Scan- Lindquist et al34 reported a higher prevalence rate dinavian study of twins estimated that 42% of the ob- of mutations in MUC3A and PRIM2 in African Ameri- served rate of prostate cancer susceptibility was associ- can men with prostate cancer when compared with

416 Cancer Control October 2016, Vol. 23, No. 4 their white counterparts. MUC3 is involved in cell sig- nosis and management of prostate cancer in African naling, growth, and survival; MUC3 and PRIM2 are American men. also associated with carcinogenesis.35 The different HGPIN is considered by many to be a prema- patterns of somatic mutations observed between racial lignant lesion of prostate cancer.48-50 Data from some groups may help provide a basis for understanding the studies suggest that the prevalence of HGPIN is greater genomic contributions to aggressive tumors and asso- in African American men and that subgroups of Afri- ciated disparate outcomes in African American men can American men with HGPIN are more prone to the with prostate cancer. development of aggressive, clinically significant can- Evidence is also emerging for racial differences cer.51,52 HGPIN may also be a risk factor for biochemi- in markers of prostate tumor subtypes. For example, cal recurrence following definitive treatment in African gene fusions in TMPRSS2 and ERG are found in pros- American men.51,52 HGPIN is more prevalent in Afri- tate tumors.35-43 However, several studies report that can American men, even after controlling for age, PSA the prevalence of TMPRSS2–ERG fusions in the tumors level, abnormal findings on digital rectal examination, of African American men is lower than that observed and prostate volume.51 Significantly higher prevalence in white men.35-43 Thus, future studies must continue to rates of HGPIN have been reported in African Ameri- pursue this research because such racial differences in can men aged 40 to 49 years compared with white marker expression could reveal mechanisms contribut- men (46% vs 29%, respectively), suggesting that this ing to race disparities. early-onset age range may represent the beginning of Differences exist in tumor biology between African a racial disparity related to prostate cancer; this is be- American men and white men and may be attribut- cause more African American men with HGPIN go on able to race-specific differences in tumor location to develop prostate cancer.52 In concordance with these (ie, anterior vs posterior tumors). African Americans data, Potts et al53 reported that African American men have an increased incidence of prostate cancer (locat- were more likely to be diagnosed with prostatic ined anterior to the peripheral zone in radical prosta- traepithelial neoplasia compared with white men, even tectomy specimens that may potentially have a prog- after adjusting for PSA levels. nostic significance.44 In a study of 1,245 patients who In a study to investigate racial differences in tumor underwent radical prostatectomy, the overall tumor burden (cancer volume, cancer percentage, and tumor locations were anterior in 14%, posterior in 58%, or volume–PSA ratio) in a large cohort of men undergoing both in 28% of cases.44 The incidence of anterior tu- radical prostatectomy, African Americans had higher mors was higher in African American men compared disease burden (estimated tumor volume, percent of with white men, and the rates of positive surgical cancer involvement, estimated tumor volume–PSA margins in anteriorly and posteriorly located tumors ratio) compared with non–African American men.54 were 60% vs 38% in African American men and 48% This association was pronounced in low-grade cancer vs 27% in white men, respectively.44 (Gleason score ≤ 6), thus depicting a complex picture In patients with an abnormal level of serum PSA of relations between race and tumor burden across the and negative findings on sextant prostate biopsies, aggressive spectrum of prostate cancer.54 The frequen- biopsy is recommended for the anterior zone of the cy of HGPIN and the autopsy prevalence of prostate prostate.45 In a study of 398 men, of whom 70% were cancer have been reported to be similar in large stud- African American men, most patients had prostate ies comparing Asian and Western populations, with cancer limited to the peripheral zone or in both the widely differing incidence and mortality rates, suggest- peripheral and anterior to the peripheral zone.46 For ing an environmental influence on the expression of 4% of study patients, prostate cancer was limited to this disease and the possibility of preventing disease the anterior to the peripheral zone — 5% of African progression from HGPIN to prostate cancer through Americans vs 2% of those not of African American pharmacological means.12,48-50,55,56 Targeting precur- descent — but this result was not statistically signifi- sor lesions (HGPIN) that may display simpler genomic cant.46 Another study showed a biologically important aberrations relative to early stages of cancer in these relationship between tumor location and molecu- high-risk groups may offer more promise to develop lar subtype, but racial differences in molecular sub- and test targeted interventions based on the biological types did not persist when tumors were analyzed by differences reported in prostate carcinogenesis among location.47 The study investigated triple-negative dis- African American men. Based on this evidence, Afri- ease and AR signaling. African American men were can American men with biopsy-proven HGPIN are a more likely to be positive for m-SPINK1+ and have subset of men who can be considered to be at high risk triple-negative disease than white men.47 for prostate cancer and, thus, may be ideal candidates Based on this evidence, the racial differences in for chemoprevention interventions. location of the prostate cancer do not appear to be Although PSA and its kinetics are used to monitor significant, nor are they likely to impact the prog- disease progression, controversy exists regarding the

October 2016, Vol. 23, No. 4 Cancer Control 417 efficacy of PSA level as a screening tool due to its low cer cells, suggesting that TCEAL7 is a tumor-suppressor rate of specificity; in addition, PSA levels are increased gene and that varying expression levels of TCEAL7 be- in benign prostatic diseases, such as benign prostatic tween white and African American men may explain hyperplasia (PSA concentrations can be ≤ 50% higher some of the disparities in the growth patterns ob- in some cases).57 African American men continue to served in prostate cancer–derived tumors cells in these have higher PSA levels and Gleason scores than white populations. Population studies comparing Japanese men, despite a narrowing of the differences in patho- and African Americans have observed that short CAG logical stage.58 One study reported that African Ameri- repeats are more frequently associated with higher can men with prostate cancer and a Gleason score of 6 transactivational function in African Americans, pos- produce less PSA than white men.59 African American sibly explaining racial differences in the incidence rate and white men had equal serum PSA and PSA masses of prostate cancer between the 2 populations.11 despite significantly larger prostates in African Amer- Taken together, these data suggest that there may ican men and all other parameters being the same.60 be functional biological differences in the prostate tu- Another report showed that men with higher levels of mors of African American men that may predispose testosterone had higher levels of PSA even after taking them to a biologically more aggressive malignancy. into account other hormones, age, and race/ethnicity.61 However, these observations have not been validated While transcription of the PSA gene is transactivated in clinical trials targeting African American men. by the androgen receptor (AR) with bound androgen, it is unclear whether circulating androgens influence Current Approach circulating PSA levels. Men treated for prostate cancer Chemoprevention refers to the inhibition of prein- with androgen-deprivation therapy experience a de- vasive and invasive cancer and its progression or cline in serum PSA concentration, but this decline may treatments of identifiable precancers.49,77 Chemo- be due to fewer cancer cells remaining to produce PSA prevention efforts require an understanding of the and reduced testosterone available to all PSA-produc- mechanism of carcinogenesis, including signaling cells.41 Despite these drawbacks related to speci- ing, metabolic, and genetic progression pathways. ficity, serum PSA as a continuous variable — as well New technologies in genomics and proteomics have as its doubling time and velocity — has been used in spurred this field of research. Use of this knowledge prostate cancer chemoprevention trials and in clinical to develop pharmacological agents, botanicals, and practice to define risk categories.59,62-67 biologics to reverse or halt the process of carcinogen- Other biological differences in benign tissue esis is called chemoprevention. and prostate tumors have been elucidated in African Agents for chemoprevention include antipro- American and white men alike. Prostate tumors de- motion and antiprogression agents that prevent the rived from African American men overexpress signal- growth and survival of cells already committed to being receptors that may mediate increased prolifera- come malignant.49,77 Finasteride and dutasteride, which tion, including EGFR and AR.68,69 On average, AR has block the conversion of testosterone to dihydrotes- a shorter polyglutamine repeat in African American tosterone, have been evaluated for chemoprevention men and, thus, increased AR activity.11 Reports have of prostate cancer in large phase 3 chemoprevention indicated decreased apoptosis and increased immu- trials.78-80 Although these agents significantly reduced nostaining for antiapoptotic protein B-cell lymphoma the risk of prostate cancer, their use was also associ- 2 (BCL2), suggesting that increased expression of ated with an increased rate of high-grade disease, se- BCL2 may decrease apoptosis and increase rates of verely limiting their clinical adoption and underscoring tumor proliferation in African American men.70,71 Met- the need to identify novel chemoprevention agents for astatic capacity may be higher in prostate tumor cells prostate cancer.79 in African Americans because of the overexpression Botanicals influence multiple biochemical and of several metastasis-related genes (AMFR, CXCR4, molecular cascades that inhibit mutagenesis, prolif- CCR7, and MMP9) as well as KI67 and CAV1 in Afri- eration, induce apoptosis, and suppress the formation can American men.72-74 Prostate tumors obtained from and growth of human cancers, thus modulating several African American men may have decreased rates of hallmarks of carcinogenesis. In addition, these agents tumor suppression. appear promising in their potential to impact the field Decreased TCEAL7 expression has been observed of cancer chemoprevention, because they have a sig- in tumors from African American men compared with nificantly superior safety profile than most available tumors from white men.75 Reams et al75 observed that agents.81-88 Several botanicals have been characterized TCEAL7 expression in prostate cancer and in nonma- and used for hundreds of years, although existing chal- lignant tissue was higher in white men compared with lenges and limitations have hampered progress in this African American men. Chien et al76 demonstrated field.89,90 Multiple botanicals have been identified and that TCEAL7 also inhibits the growth of ovarian can- appear promising for the chemoprevention of prostate

418 Cancer Control October 2016, Vol. 23, No. 4 cancer.91-93 However, the slow pace of growth might US Food and Drug Association and to promote the ac- be attributed to the regulatory protection of classical ceptance of new agents into the marketplace.92,93 formulations and lack of standardization, quality control, a molecular mechanism-based approach in evalu- Green Tea Catechins ation, a population-based normal range of biomarkers, The Table summarizes the clinical, molecular, and bi- laboratory practices, and few translational scientists ological characteristics relevant to prostate cancer in engaged in conducting well-designed trials. However, African American men and compares them with the several valuable lessons have been learned from previ- effects of green tea catechins used to modulate those ous chemoprevention trials.91,94,95 Critical requirements characteristics.11,51,52,58,59,68-75,98-116 for moving botanicals from bench to bedside include Laboratory studies have identified epigallocat- adopting a systematic, molecular-mechanism based echin gallate (EGCG) as a potent chemopreventive approach and utilizing the same ethical and rigorous agent because it affects various key molecular pro- methods such as those used to evaluate other pharma- cesses in prostate carcinogenesis that help induce cological agents. apoptosis and inhibit tumor growth and angiogen- Chemoprevention trials using combinations of esis.117-121 Preclinical studies have demonstrated botanicals have demonstrated that synergy between chemopreventive efficacy in prostate cancer, show- agents can lead to lower doses, improved rates of ef- ing that it has significant activity on prostate cancer ficacy, and fewer or less-severe toxicities.96 An as- cells.99,122-129 Phase 1/2 studies have demonstrated the sessment of the end points of chemoprevention trials bioavailability and tolerance of green tea catechin at whose results have been used to support approval of various ranges of doses of EGCG.64,93,98,100,130-136 an agent for prostate cancer revealed that nearly all Although several mechanisms exist by which have been approved on the basis of intraepithelial neo- EGCG may operate in prostate carcinogenesis, EGCG plasia.97 Intermediate biomarker end points must be selectively inhibits the proteasome activity in intact identified, validated, and obtained using noninvasive human prostate cancer cells and accumulates in IκBα techniques without compromising the safety to men and p27 proteins, thus leading to growth arrest.119-121 in chemoprevention trials. To reduce patient burden, High-grade green tea polyphenol extract, which is a these markers must be obtained from accessible or- mixture of tea catechins with more than 50% of EGCG, gans and during the normal course of clinical surveil- inhibits the proteasomal chymotrypsin-like activity lance. Randomized placebo control trials and the long- with a half maximal inhibitory concentration value term follow-up and monitoring of patients and study of 7 μM.93 The half maximal inhibitory concentration participants are critical to meet the requirements of the value for trypsin-like activity is higher than 100 μM,

Table. — Summary of Relevant Characteristics of Prostate Cancer and the Effects of Green Tea Catechins in African Americans Characteristic of Prostate Cancer Effect of Green Tea Catechins Clinical Increased frequency of HGPIN51,52 Men with HGPIN had significantly reduced disease progression98,101 Higher PSA levels58 Men with HGPIN had lower PSA levels than the control arm98,101 Higher Gleason scores at diagnosis59 Men given short-term daily doses had significant decreases in serum PSA levels100 Molecular Increased Proliferation Decreased Proliferation In vitro models102-104 EGFR overexpression68 In vivo models99,105-107 AR overexpression69 Inhibition of EGFR signaling108 Increased androgen-receptor activity11 Reduced androgen-receptor signaling and activity99,109-111 Decreased Apoptosis Induced Apoptosis Induction Increased expression of BCL2 70,71 Various mechanisms (eg, reduced BCL2 expression)112-114 Increased Metastatic Capacity Inhibition of Invasion and Metastasis Overexpression of AMFR, CXCR4, CCR7, MMP9, KI67, Various mechanisms (eg, MMP9, MMP2 suppression)115 and CAV1 72-74 Decreased Tumor Suppression Reactivation of Silenced Tumor-Suppressor Genes Downregulation of TCEAL7 75 Restores balanced proliferation116 HGPIN = high-grade prostatic intraepithelial neoplasia, PSA = prostate-specific antigen.

October 2016, Vol. 23, No. 4 Cancer Control 419 demonstrating that high-grade green tea polyphenol and disease progression from HGPIN to atypical small extract preferentially inhibits the proteasomal chymo- acinar proliferation or prostate cancer with green tea trypsin-like activities over other activities.93 catechin treatment. Although these are early observa- Kim et al126 investigated the safety and efficacy of tions with small sample sizes, these data are provoca- high-grade green tea polyphenol extract in an animal tive and warrant further evaluation in a well-powered model. Their goal was to reduce the progression of clinical trial targeting African American men with prostate adenocarcinoma in transgenic mice. Mice HGPIN, because they suggest that green tea catechins treated with the green tea extract had significantly have the potential to decrease the risk of progression fewer tumors and decreased tumoral sizes compared to atypical small acinar proliferation and, subsequent- with the animals that did not receive the extract. The ly, prostate cancer.137 However, the study was limited high-grade green tea polyphenol extract also signif- by its small sample size of African American men, so icantly inhibited metastasis in the treated mice in a the researchers were unable to make comparisons dose-dependent manner. Therefore, the data suggest among white men with HGPIN in other blood and tis- that high-grade green tea polyphenol extract is an ef- sue-based biomarkers. fective chemopreventive agent in preventing the pro- Experimental models to address the fundamen- gression of prostate cancer to metastasis in a mouse tal molecular pathways of green tea catechin have yet model.126 These findings provide evidence for the to be validated in green tea catechin–treated tissue safety and chemopreventive effect of this green tea samples from clinical trials.118 In addition to develop- polyphenol extract for preventing metastatic spread ing and refining the fundamental pathways of green in prostate cancer. tea catechin, future studies evaluating these molecu- Based on the results of previous studies, a placebo- lar pathways should define intermediate biomarkers of controlled, randomized clinical trial was conducted prostate carcinogenesis. However, obtaining sufficient of the high-grade green tea extract in 97 men with HG- tissue from prostate biopsies from precursor lesions PIN, atypical small acinar proliferation, or both condi- in prevention trials continues to present a challenge. tions.64,92,119-121 No differences were observed in the Therefore, it will be important to prioritize intermedi- number of prostate cancer cases. The cumulative rate of ate biomarkers available for evaluation based on the prostate cancer plus atypical small acinar proliferation robustness of evidence and relevance to the agent, the among men with HGPIN but without atypical small aci- potential molecular mechanism, and the patient cohort nar proliferation at the start of the study was researched. and stage of disease targeted in future studies.92,93 The data revealed a statistically significant decrease in Green tea catechins target every major clini- this composite end point: 3 of 26 (high-grade green tea cal prostate cancer characteristic reported in African polyphenol extract gorup) vs 10 of 25 (placebo group).126 American men (see Table).11,51,52,58,59,68-75,98-116 They may A decrease in diagnoses of atypical small acinar prolifer- do this by acting through the biological mechanisms ation was also seen among those receiving the green tea especially relevant to African American men, includ- extract, as was a decrease in serum PSA level, compared ing reducing proliferation, inducing apoptosis, and with those assigned to placebo.126 decreasing tumoral metastatic capacity. Thus, it is our Intake of a standardized, decaffeinated green tea opinion that green tea catechins are a promising can- catechin mixture every day for 1 year accumulated in didate for the chemoprevention of prostate cancer in plasma and was well tolerated, and researchers saw African American men. that its consumption significantly reduced the cumulative rates of prostate cancer and atypical small acinar Conclusions proliferation among men with HGPIN who did not Prostate cancer in African American men is a major have atypical small acinar proliferation at baseline.137 public health problem with significant morbidity and Three of the 9 African American men in the study pro- mortality rates. A variety of laboratory approaches are gressed to atypical small acinar proliferation or pros- being applied to unravel the molecular pathogenesis of tate cancer after 1 year, but no such progression was prostate cancer. African American men with a family seen among those assigned to the high-grade green tea history of prostate cancer, diagnosed with high-grade polyphenol extract arm.137 This disease progression prostatic intraepithelial neoplasia prior to age 50 years, rate from HGPIN in the placebo arm was higher than represent an ideal target population for chemopreven- US norms and in white men assigned to the placebo tion. Green tea catechins have been characterized and arm of this trial (20%).62,64,65,137,138 In men with a base- evaluated as agents able to target multiple pathways line diagnosis of atypical small acinar proliferation, that address the underlying biology of prostate carci- the diagnostic rates of cancer at 1 year were similar, re- nogenesis in African American men. Assessing the gardless of treatment arm assignment or race.137 Over- efficacy of this intervention might eventually trans- all, a significant increase in plasma EGCG concentra- late into clinical use in high-risk populations. It is our tion was observed, as was a reduction of serum PSA opinion that the greatest challenge to evaluating prom-

420 Cancer Control October 2016, Vol. 23, No. 4 ising agents for chemoprevention is the recruitment 25. Gudmundsson J, Sulem P, Steinthorsdottir V, et al. Two variants on chromosome 17 confer prostate cancer risk, and the one in TCF2 protects and retention of African American men in randomized against type 2 diabetes. Nat Genet. 2007;39(8):977-983. clinical trials, which will require multiple approaches, 26. The Human Protein Atlas website. http://www.proteinatlas.org/ ENSG00000275410-HNF1B/tissue. Accessed October 10, 2016. actions, and activities for achieving meaningful ad- 27. Bingham C, Hattersley AT. Renal cysts and diabetes syndrome vances. The process is a complex and progressive one, resulting from mutations in hepatocyte nuclear factor-1훽. Nephrol Dial Transplant. 2004;19:2703-2708. including the formation of trusting partnerships with 28. Haldorsen IS, Vesterhus M, Raeder H, et al. Lack of pancreatic target populations and their circles of influence, power body and tail in HNF1B mutation carriers. Diabet Med. 2008;25(7):782-787. 29. Nakayama, K, Nozu, Y. Goto Y, et al. HNF1B alterations associated sharing, priority setting, capacity building, education with congenital anomalies of the kidney and urinary tract. Pediatr Nephrol. and training, the transparent sharing of data and in- 201025(6):1073-1079. 30. Cornelis MC, Agrawal A, Cole JW, et al. The Gene, Environment formation, the development of goals, and insisting on Association Studies consortium (GENEVA): maximizing the knowledge community involvement in the research. obtained from GWAS by collaboration across studies of multiple conditions. Genet Epidemiol. 2010;34(4):364-372. 31. Han KS, Jeong IG, Joung JY, et al. Prevalence of high-grade pros- References tatic intraepithelial neoplasia in prostate gland of Korean men: comparisons between radical prostatectomy and cystoprostatectomy. Urology. 1. American Cancer Society. Cancer Facts and Figures 2016. Atlanta, GA: 2007;70(6):1100-1103. 2016. http://www.cancer.org/acs/groups/content/@research/documents/ 32. Paliwal S, Kovi RC, Nath B, et al. The alternative reading frame document/acspc-047079.pdf. Accessed October 12, 2016. tumor suppressor antagonizes hypoxia-induced cancer cell migration via 2. Centers for Disease Control and Prevention. Prostate cancer rates interaction with the COOH-terminal binding protein corepressor. Cancer by race and ethnicity. Updated June 16, 2016. http://www.cdc.gov/cancer/ Res. 2007;67(19):9322-9329. prostate/statistics/race.htm. Accessed October 12, 2016. 33. Thomas G, Jacobs KB, Yeager M, et al. Multiple loci identified 3. Makridakis NM, Reichardt JK. Molecular epidemiology of hormone- in a genome-wide association study of prostate cancer. Nat Genet. metabolic loci in prostate cancer. Epidemiol Rev. 2001;23(1):24-29. 2008;40(3):310-315 4. Chornokur G, Dalton K, Borysova ME, et al. Disparities at presentation, 34. Lindquist KJ, Paris PL, Hoffmann TJ, et al. Mutational landscape diagnosis, treatment, and survival in African American men, affected by prostate of aggressive prostate tumors in African American men. Cancer Res. cancer. Prostate. 2011;71(9):985-997. 2016;76(7):1860-1868. 5. Powell IJ. Epidemiology and pathophysiology of prostate cancer in 35. Yatsula B, Galvao C, McCrann M, et al. Assessment of F-MuLV- African-American men. J Urol. 2007;177(2):444-449. induced tumorigenesis reveals new candidate tumor genes including 6. Ilic D, O’Connor D, Green S, et al. Screening for prostate cancer: an Pecam1, St7, and Prim2. Leukemia. 2006;20(1):162-165. updated Cochrane systematic review. BJU Int. 2011;107(6):882-891. 36. Berger MF, Lawrence MS, Demichelis F, et al. The genomic com- 7. Moyer VA; US Preventive Services Task Force. Screening for prostate plexity of priary human prostate cancer. Nature. 2011;470:214-220. cancer: U.S. Preventive Services Task Force recommendation statement. 37. Baca SC, Prandi D, Lawrence MS, et al. Punctuated evolution of Ann Intern Med. 2012;157(2):120-134. prostate cancer genomes. Cell. 2013;153(3):666-677. 8. Andriole GL, Crawford ED, Grubb RL, 3rd, et al. Mortality results 38. Alexandrov LB, Nik-Zainal S, Wedge DC, et al. Signatures of muta- from a randomized prostate-cancer screening trial. N Engl J Med. tional processes in human cancer. Nature. 2013;500(7463):415-421. 2009;360(13):1310-1319. 39. Boutros PC, Fraser M, Harding NJ, et al. Spatial genomic het- 9. Lippman SM, Hong WK. Cancer prevention science and practice. erogeneity within localized, multifocal prostate cancer. Nat Genet. Cancer Res. 2002;62(18):5119-5125. 2015;47(7):736-745. 10. Montironi R, Mazzucchelli R, Lopez-Beltran A, et al. Mechanisms of 40. Cooper CS, Eeles R, Wedge DC, et al. Analysis of the genetic disease: high-grade prostatic intraepithelial neoplasia and other proposed phylogeny of multifocal prostate cancer identifies multiple independent preneoplastic lesions in the prostate. Nat Clin Pract Urol. 2007;4(6):321-332. clonal expansions in neoplastic and morphologically normal prostate 11. Suzuki H, Ueda T, Ichikawa T, et al. Androgen receptor involvement in tissue. Nat Genet. 2015;47(4):367-372. the progression of prostate cancer. Endocr Relat Cancer. 2003;10(2):209-216. 41. Clark J, Merson S, Jhavar S, et al. Diversity of TMPRSS2-ERG 12. Zlotta AR, Egawa S, Pushkar D, et al. Prevalence of prostate cancer fusion transcripts in the human prostate. Oncogene. 2007;26(18):2667-2673. on autopsy: cross-sectional study on unscreened Caucasian and Asian 42. Kumar-Sinha C, Tomlins SA, Chinnaiyan AM. Recurrent gene fusions men. J Natl Cancer Inst. 2013;105(14):1050-1058. in prostate cancer. Nat Rev Cancer. 2008;8(7):497-511. 13. Wolters T, Roobol MJ, Steyerberg EW, et al. The effect of study 43. Mosquera JM, Mehra R, Regan MM, et al. Prevalence of TMPRSS2- arm on prostate cancer treatment in the large screening trial ERSPC. Int J ERG fusion prostate cancer among men undergoing prostate biopsy in the Cancer. 2010;126(10):2387-2393. United States. Clin Cancer Res. 2009; 15(14):4706-4711. 14. Cunningham GR, Ashton CM, Annegers JF, et al. Familial aggrega- 44. Tiguert R, Gheiler EL, Tefilli MV, et al. Racial differences and prog- tion of prostate cancer in African-Americans and white Americans. Prostate. nostic significance of tumor location in radical prostatectomy specimens. 2003;56(4):256-262. Prostate. 1998;37(4):230-235. 15. Gibbs M, Stanford JL, Jarvik GP, et al. A genomic scan of families 45. Tiguert R, Gheiler EL, Tefilli MV, et al. Racial differences and prog- with prostate cancer identifies multiple regions of interest. Am J Hum Genet. nostic significance of tumor location in radical prostatectomy specimens. 2000;67(1):100-109. Prostate. 1998 37(4):230-235. 16. Kote-Jarai Z, Olama AA, Giles GG, et al. Seven prostate cancer 46. Feibus AH, Levy J, McCaslin IR, et al. Racial variation in prostate susceptibility loci identified by a multi-stage genome-wide association needle biopsy templates directed anterior to the peripheral zone. Urol Oncol. study. Nat Genet. 2011;43(8):785-791. 2016;34(8):336.e1-6. 17. Eeles RA, Olama AA, Benlloch S, et al. Identification of 23 new 47. Faisal FA, Sundi D, Tosoian JJ, et al. Racial variations in prostate prostate cancer susceptibility loci using the iCOGS custom genotyping array. cancer molecular subtypes and androgen receptor signaling reflect anatomic Nat Genet. 2013;45(4):385-391, 391e1-391e2. tumor location. Eur Urol. 2016;70(1):14-17. 18. Freedman ML, Monteiro AN, Gayther SA, et al. Principles for the 48. Epstein JI, Herawi M. Prostate needle biopsies containing prostatic post-GWAS functional characterization of cancer risk loci. Nat Genet. intraepithelial neoplasia or atypical foci suspicious for carcinoma: implica- 2011;43(6):513-518. tions for patient care. J Urol. 2006;175(3 pt 1):820-834. 19. Haiman CA, Patterson N, Freedman ML, et al. Multiple regions 49. Kelloff GJ, Lieberman R, Steele VE, et al. Chemoprevention of within 8q24 independently affect risk for prostate cancer. Nat Genet. prostate cancer: concepts and strategies. Eur Urol. 1999;35(5-6):342-350. 2007;39(5):638-644. 50. Lieberman R. Prostate cancer chemoprevention: strategies for de- 20. Stevens VL, Ahn J, Sun J, et al. HNF1B and JAZF1 genes, diabetes, signing efficient clinical trials. Urology. 2001;57(4 suppl 1):224-229. and prostate cancer risk. Prostate. 2010;70(6):601-607. 51. Fowler JE, Jr, Bigler SA, Lynch C, et al. Prospective study of cor- 21. Shen MM, Abate-Shen C. Molecular genetics of prostate cancer: relations between biopsy-detected high grade prostatic intraepithelial new prospects for old challenges. Genes Dev. 2010;24(18):1967-2000. neoplasia, serum prostate specific antigen concentration, and race. Cancer. 22. Stegeman S, Amankwah E, Klein K, et al. A large-scale analysis 2001;91(7):1291-1296. of genetic variants within putative miRNA binding sites in prostate cancer. 52. Powell IJ, Bock CH, Ruterbusch JJ, et al. Evidence supports a faster Cancer Discov. 2015;5(4):368-379. growth rate and/or earlier transformation to clinically significant prostate 23. Chornokur G, Amankwah EK, Davis SN, et al. Variation in HNF1B cancer in black than in white American men, and influences racial progression and obesity may influence prostate cancer risk in African American men: a and mortality disparity. J Urol. 2010;183(5):1792-1796. pilot study. Prostate Cancer. 2013;2013:384594. 53. Potts JM, Lutz M, Walker E, et al. Trends in PSA, age and prostate 24. Eeles RA, Kote-Jarai Z, Giles GG, et al. Multiple newly identified loci cancer detection among black and white men from 1990-2006 at a tertiary associated with prostate cancer susceptibility. Nat Genet. 2008;40(3):316-321. care center. Cancer. 2010;116(16):3910-3915.

October 2016, Vol. 23, No. 4 Cancer Control 421 54. Tsivian M, Banez LL, Keto CJ, et al. African-American men with 79. Hamilton RJ, Kahwati LC, Kinsinger LS. Knowledge and use of fin- low-grade prostate cancer have higher tumor burdens: results from the asteride for the prevention of prostate cancer. Cancer Epidemiol Biomarkers Duke Prostate Center. Prostate Cancer Prostatic Dis. 2013;16(1):91-94. Prev. 2010;19(9):2164-2171. 55. Chornokur G, Han G, Tanner R, et al. High grade prostate in- 80. Thompson IM, Goodman PJ, Tangen CM, et al. The influence traepithelial neoplasia (PIN) is a PSA-independent risk factor for prostate of finasteride on the development of prostate cancer. N Engl J Med. cancer in African American men: results from a pilot study. Cancer Lett. 2003;349(3):215-224. 2013;331(2):154-157. 81. Aggarwal BB, Prasad S, Reuter S, et al. Identification of novel anti- 56. Tan PH, Tan HW, Tan Y, et al. Is high-grade prostatic intraepithelial inflammatory agents from Ayurvedic medicine for prevention of chronic neoplasia on needle biopsy different in an Asian population: a clinicopath- diseases: “reverse pharmacology” and “bedside to bench” approach. Curr ologic study performed in Singapore. Urology. 2006;68(4):800-803. Drug Targets. 2011;12(11):1595-1653. 57. Martin BJ, Finlay JA, Sterling K, et al. Early detection of prostate 82. Amin AR, Kucuk O, Khuri FR, et al. Perspectives for cancer prevention cancer in African-American men through use of multiple biomarkers: with natural compounds. J Clin Oncol. 2009;27(16):2712-2725. human kallikrein 2 (hK2), prostate-specific antigen (PSA), and free PSA 83. Aravindaram K, Yang NS. Anti-inflammatory plant natural products (fPSA). Prostate Cancer Prostatic Dis. 2004;7(2):132-137. for cancer therapy. Planta Med. 2010;76(11):1103-1117. 58. Lilja H, Ulmert D, Vickers AJ. Prostate-specific antigen and 84. Gogtay NJ, Bhatt HA, Dalvi SS, et al. The use and safety of non- prostate cancer: prediction, detection and monitoring. Nat Rev Cancer. allopathic Indian medicines. Drug Saf. 2002;25(14):1005-1019. 2008;8(4):268-278. 85. Jia L. Cancer complementary and alternative medicine research at 59. Gee J, Bailey H, Kim K, et al. Phase II open label, multi-center clinical trial the US National Cancer Institute. Chin J Integr Med. 2012;18(5):325-332. of modulation of intermediate endpoint biomarkers by 1alpha-hydroxyvitamin D2 86. Kannappan R, Gupta SC, Kim JH, et al. Tocotrienols fight cancer by in patients with clinically localized prostate cancer and high grade pin. Prostate. targeting multiple cell signaling pathways. Genes Nutr. 2012;7(1):43-52. 2013;73(9):970-978. 87. Reddy BS. Chemoprevention of colon cancer by minor dietary con- 60. Kryvenko ON, Balise R, Soodana Prakash N, et al. African-American stituents and their synthetic analogues. Prev Med. 1996;25(1):48-50. men with Gleason score 3+3=6 prostate cancer produce less prostate specific 88. Wargovich MJ, Jimenez A, McKee K, et al. Efficacy of potential che- antigen than Caucasian men: a potential impact on active surveillance. J Urol. mopreventive agents on rat colon aberrant crypt formation and progression. 2016;195(2):301-306. Carcinogenesis. 2000;21(6):1149-1155. 61. Peskoe SB, Joshu CE, Rohrmann S, et al. Circulating total testosterone 89. Minich DM, Bland JS. A review of the clinical efficacy and safety of and PSA concentrations in a nationally representative sample of men without cruciferous vegetable phytochemicals. Nutr Rev. 2007;65(6 pt 1):259-267. a diagnosis of prostate cancer. Prostate. 2015;75(11):1167-1176. 90. Scott EN, Gescher AJ, Steward WP, et al. Development of dietary 62. Herawi M, Kahane H, Cavallo C, et al. Risk of prostate cancer on phytochemical chemopreventive agents: biomarkers and choice of dose first re-biopsy within 1 year following a diagnosis of high grade prostatic for early clinical trials. Cancer Prev Res (Phila). 2009;2(6):525-530. intraepithelial neoplasia is related to the number of cores sampled. J Urol. 91. Lippman SM, Klein EA, Goodman PJ, et al. Effect of selenium and 2006;175(1):121-124. vitamin E on risk of prostate cancer and other cancers: the Selenium and 63. Taneja SS, Morton R, Barnette G, et al. Prostate cancer diagnosis Vitamin E Cancer Prevention Trial (SELECT). JAMA. 2009;301(1):39-51. among men with isolated high-grade intraepithelial neoplasia enrolled onto 92. Kumar N, Crocker T, Smith T, et al. Prostate cancer chemoprevention a 3-year prospective phase III clinical trial of oral toremifene. J Clin Oncol. targeting men with high-grade prostatic intraepithelial neoplasia (HGPIN) 2013;31(5):523-529. and atypical small acinar proliferation (ASAP): model for trial design and 64. Kumar NB, Pow-Sang J, Egan KM, et al. Randomized, placebo- outcome measures. J Clin Trials. 2012;2(1). controlled trial of green tea catechins for prostate cancer prevention. Cancer 93. Kumar N, Crocker T, Smith T, et al. Prostate cancer chemoprevention Prev Res (Phila). 2015;8(10):879-887. targeting high risk populations: model for trial design and outcome measures. 65. Amin MM, Jeyaganth S, Fahmy N, et al. Subsequent prostate cancer J Cancer Sci Ther. 2012;2011(S3). detection in patients with prostatic intraepithelial neoplasia or atypical 94. The effect of vitamin E and beta carotene on the incidence of lung cancer small acinar proliferation. Can Urol Assoc J. 2007;1(3):245-249. and other cancers in male smokers. The Alpha-Tocopherol, Beta Carotene 66. Bryant RJ, Lilja H. Emerging PSA-based tests to improve screening. Cancer Prevention Study Group. N Engl J Med. 1994;330(15):1029-1035. Urol Clin North Am. 2014;41(2):267-276. 95. Omenn GS, Goodman GE, Thornquist MD, et al. Risk factors for 67. D’Amico AV, Whittington R, Malkowicz SB, et al. Biochemical out- lung cancer and for intervention effects in CARET, the Beta-Carotene and come after radical prostatectomy, external beam radiation therapy, or Retinol Efficacy Trial. J Natl Cancer Inst. 1996;88(21):1550-1559. interstitial radiation therapy for clinically localized prostate cancer. JAMA. 96. Kakarala M, Brenner DE, Korkaya H, et al. Targeting breast stem 1998;280(11):969-974. cells with the cancer preventive compounds curcumin and piperine. Breast 68. Shuch B, Mikhail M, Satagopan J, et al. Racial disparity of epider- Cancer Res Treat. 2010;122(3):777-785. mal growth factor receptor expression in prostate cancer. J Clin Oncol. 97. Parnes HL, Brawley OW, Minasian LM, et al. Phase III prostate can- 2004;22(23):4725-4729. cer chemoprevention trials. Recent Results Cancer Res. 2014;202:73-77. 69. Gaston KE, Kim D, Singh S, et al. Racial differences in androgen 98. Bettuzzi S, Brausi M, Rizzi F, et al. Chemoprevention of human pros- receptor protein expression in men with clinically localized prostate cancer. tate cancer by oral administration of green tea catechins in volunteers with J Urol. 2003;170(3):990-993. high-grade prostate intraepithelial neoplasia: a preliminary report from a 70. deVere White RW, Deitch AD, Jackson AG, et al. Racial differenc- one-year proof-of-principle study. Cancer Res. 2006;66(2):1234-1240. es in clinically localized prostate cancers of black and white men. J Urol. 99. Harper CE, Patel BB, Wang J, et al. Epigallocatechin-3-Gallate 1998;159(6):1979-1982. suppresses early stage, but not late stage prostate cancer in TRAMP mice: 71. Guo Y, Sigman DB, Borkowski A, et al. Racial differences in prostate mechanisms of action. Prostate. 2007;67(14):1576-1589. cancer growth: apoptosis and cell proliferation in Caucasian and African- 100. McLarty J, Bigelow RL, Smith M, et al. Tea polyphenols decrease American patients. Prostate. 2000;42(2):130-136. serum levels of prostate-specific antigen, hepatocyte growth factor, and 72. Kim HS, Moreira DM, Jayachandran J, et al. Prostate biopsies vascular endothelial growth factor in prostate cancer patients and inhibit from black men express higher levels of aggressive disease biomarkers production of hepatocyte growth factor and vascular endothelial growth than prostate biopsies from white men. Prostate Cancer Prostatic Dis. factor in vitro. Cancer Prev Res (Phila). 2009;2(7):673-682. 2011;14(3):262-265. 101. Kumar NB, Pow-Sang J, Egan KM, et al. Effect of polyphenon E 73. Wallace TA, Prueitt RL, Yi M, et al. Tumor immunobiological differ- on progression to prostate cancer after diagnosis of high grade prostatic ences in prostate cancer between African-American and European-American intraepithelial neoplasia or atypical small acinar proliferation in African men. Cancer Res. 2008;68(3):927-936. American men. Presented at: American Association for Cancer Research; 74. Yang G, Addai J, Ittmann M, et al. Elevated caveolin-1 levels in November 9–12, 2014; San Antonio, Texas. African-American versus white-American prostate cancer. Clin Cancer 102. Albrecht DS, Clubbs EA, Ferruzzi M, et al. Epigallocatechin-3-gallate Res. 2000;6(9):3430-3433. (EGCG) inhibits PC-3 prostate cancer cell proliferation via MEK-independent 75. Reams RR, Agrawal D, Davis MB, et al. Microarray comparison ERK1/2 activation. Chem Biol Interact. 2008;171(1):89-95. of prostate tumor gene expression in African-American and Caucasian 103. Paschka AG, Butler R, Young CY. Induction of apoptosis in prostate American males: a pilot project study. Infect Agent Cancer. 2009;4(suppl cancer cell lines by the green tea component, (-)-epigallocatechin-3-gallate. 1):S3. Cancer Lett. 1998;130(1-2):1-7. 76. Chien J, Staub J, Avula R, et al. Epigenetic silencing of TCEAL7 104. Yu HN, Yin JJ, Shen SR. Growth inhibition of prostate cancer cells (Bex4) in ovarian cancer. Oncogene. 2005;24(32):5089-5100. by epigallocatechin gallate in the presence of Cu2+. J Agric Food Chem. 77. Kelloff GJ, Lippman SM, Dannenberg AJ, et al. Progress in che- 2004;52(3):462-466. moprevention drug development: the promise of molecular biomarkers for 105. McCarthy S, Caporali A, Enkemann S, et al. Green tea catechins prevention of intraepithelial neoplasia and cancer--a plan to move forward. suppress the DNA synthesis marker MCM7 in the TRAMP model of prostate Clin Cancer Res. 2006;12(12):3661-3697. cancer. Mol Oncol. 2007;1(2):196-204. 78. Andriole GL, Bostwick D, Brawley OW, et al. The effect of dutaste- 106. Nyska A, Suttie A, Bakshi S, et al. Slowing tumorigenic progression ride on the usefulness of prostate specific antigen for the diagnosis of high in TRAMP mice and prostatic carcinoma cell lines using natural anti-oxidant grade and clinically relevant prostate cancer in men with a previous nega- from spinach, NAO--a comparative study of three anti-oxidants. Toxicol tive biopsy: results from the REDUCE study. J Urol. 2011;185(1):126-131. Pathol. 2003;31(1):39-51.

422 Cancer Control October 2016, Vol. 23, No. 4 107. Sartor L, Pezzato E, Dona M, et al. Prostate carcinoma and administration of polyphenon E in healthy individuals. Clin Cancer Res. green tea: (-)epigallocatechin-3-gallate inhibits inflammation-triggered 2005;11(12):4627-4633. MMP-2 activation and invasion in murine TRAMP model. Int J Cancer. 134. Jatoi A, Ellison N, Burch PA, et al. A phase II trial of green tea in 2004;112(5):823-829. the treatment of patients with androgen independent metastatic prostate 108. Rahmani AH, Al Shabrmi FM, Allemailem KS, et al. Implications of carcinoma. Cancer. 2003;97(6):1442-1446. Green Tea and Its Constituents in the Prevention of Cancer via the Modulation 135. Nguyen MM, Ahmann FR, Nagle RB, et al. Randomized, double- of Cell Signalling Pathway. Biomed Res Int. 2015;2015:925640. blind, placebo-controlled trial of polyphenon E in prostate cancer patients 109. Chuu CP, Chen RY, Kokontis JM, et al. Suppression of androgen before prostatectomy: evaluation of potential chemopreventive activities. receptor signaling and prostate specific antigen expression by (-)-epigallo- Cancer Prev Res (Phila). 2012;5(2):290-298. catechin-3-gallate in different progression stages of LNCaP prostate cancer 136. Pisters KM, Newman RA, Coldman B, et al. Phase I trial of oral green tea cells. Cancer Lett. 2009;275(1):86-92. extract in adult patients with solid tumors. J Clin Oncol. 2001;19(6):1830-1838. 110. Kao YH, Hiipakka RA, Liao S. Modulation of endocrine systems 137. Kumar NB, Pow-Sang J, Spiess PE et al. Effect of polyphenol E and food intake by green tea epigallocatechin gallate. Endocrinology. on progression to prostate cancer after diagnosis of high grade prostatic 2000;141(3):980-987. intraepithelial neoplasia or atypical small acinar proliferation in African 111. Ren F, Zhang S, Mitchell SH, et al. Tea polyphenols down-regulate American men: preliminary finds of a phase II randomized, double-blind, the expression of the androgen receptor in LNCaP prostate cancer cells. placebo-controlled clinical trial. Presented at: 7th Annual American Asso- Oncogene. 2000;19(15):1924-1932. ciation of Cancer Research Science of Cancer Health Disparities in Racial/ 112. Hastak K, Gupta S, Ahmad N, et al. Role of p53 and NF-kappaB in Ethnic Minorities and the Medically Underserved; November 9, 2015; San epigallocatechin-3-gallate-induced apoptosis of LNCaP cells. Oncogene. Antonio, TX. 2003;22(31):4851-4859. 138. Eskew LA, Bare RL, McCullough DL. Systematic 5 region prostate 113. Siddiqui IA, Adhami VM, Afaq F, et al. Modulation of phosphatidylino- biopsy is superior to sextant method for diagnosing carcinoma of the prostate. sitol-3-kinase/protein kinase B- and mitogen-activated protein kinase- J Urol. 1997;157(1):199-202. pathways by tea polyphenols in human prostate cancer cells. J Cell Biochem. 2004;91(2):232-242. 114. Siddiqui IA, Malik A, Adhami VM, et al. Green tea polyphenol EGCG sensitizes human prostate carcinoma LNCaP cells to TRAIL-mediated apoptosis and synergistically inhibits biomarkers associated with angiogenesis and metastasis. Oncogene. 2008;27(14):2055-2063. 115. Kleiner DE, Stetler-Stevenson WG. Matrix metalloproteinases and metastasis. Cancer Chemother Pharmacol. 1999;43 Suppl:S42-51. 116. Nandakumar V, Vaid M, Katiyar SK. (-)-Epigallocatechin-3-gallate reactivates silenced tumor suppressor genes, Cip1/p21 and p16INK4a, by reducing DNA methylation and increasing histones acetylation in human skin cancer cells. Carcinogenesis. 2011;32(4):537-544. 117. Adhami VM, Ahmad N, Mukhtar H. Molecular targets for green tea in prostate cancer prevention. J Nutr. 2003;133(7 suppl):2417S-2424S. 118. Connors SK, Chornokur G, Kumar NB. New insights into the mechanisms of green tea catechins in the chemoprevention of prostate cancer. Nutr Cancer. 2012;64(1):4-22. 119. Kazi A, Daniel KG, Smith DM, et al. Inhibition of the proteasome activity, a novel mechanism associated with the tumor cell apoptosis- inducing ability of genistein. Biochem Pharmacol. 2003;66(6):965-976. 120. Nam S, Smith DM, Dou QP. Ester bond-containing tea polyphenols potently inhibit proteasome activity in vitro and in vivo. J Biol Chem. 2001;276(16):13322-13330. 121. Smith DM, Wang Z, Kazi A, et al. Synthetic analogs of green tea polyphenols as proteasome inhibitors. Mol Med. 2002;8(7):382-392. 122. Adhami VM, Siddiqui IA, Sarfaraz S, et al. Effective prostate cancer chemopreventive intervention with green tea polyphenols in the TRAMP model depends on the stage of the disease. Clin Cancer Res. 2009;15(6):1947-1953. 123. Caporali A, Davalli P, Astancolle S, et al. The chemopreventive action of catechins in the TRAMP mouse model of prostate carcinogenesis is accompanied by clusterin over-expression. Carcinogenesis. 2004;25(11):2217-2224. 124. Gupta S, Hastak K, Ahmad N, et al. Inhibition of prostate carcinogenesis in TRAMP mice by oral infusion of green tea polyphenols. Proc Natl Acad Sci U S A. 2001;98(18):10350-10355. 125. Khan N, Adhami VM, Mukhtar H. Review: green tea polyphenols in chemoprevention of prostate cancer: preclinical and clinical studies. Nutr Cancer. 2009;61(6):836-841. 126. Kim SJ, Amankwah E, Connors S, et al. Safety and chemopreventive effect of Polyphenon E in preventing early and metastatic progression of prostate cancer in TRAMP mice. Cancer Prev Res (Phila). 2014;7(4):435-444. 127. Siddiqui IA, Zaman N, Aziz MH, et al. Inhibition of CWR22Rnu1 tumor growth and PSA secretion in athymic nude mice by green and black teas. Carcinogenesis. 2006;27(4):833-839. 128. Suttie A, Nyska A, Haseman JK, et al. A grading scheme for the assessment of proliferative lesions of the mouse prostate in the TRAMP model. Toxicol Pathol. 2003;31(1):31-38. 129. Yang CS, Lambert JD, Hou Z, et al. Molecular targets for the cancer preventive activity of tea polyphenols. Mol Carcinog. 2006;45(6):431-435. 130. Brausi M, Rizzi F, Bettuzzi S. Chemoprevention of human prostate cancer by green tea catechins: two years later. A follow-up update. Eur Urol. 2008;54(2):472-473. 131. Chow HH, Cai Y, Alberts DS, et al. Phase I pharmacokinetic study of tea polyphenols following single-dose administration of epigallocatechin gallate and polyphenon E. Cancer Epidemiol Biomarkers Prev. 2001;10(1):53-58. 132. Chow HH, Cai Y, Hakim IA, et al. Pharmacokinetics and safety of green tea polyphenols after multiple-dose administration of epigallocatechin gallate and polyphenon E in healthy individuals. Clin Cancer Res. 2003;9(9):3312-3319. 133. Chow HH, Hakim IA, Vining DR, et al. Effects of dosing condition on the oral bioavailability of green tea catechins after single-dose

October 2016, Vol. 23, No. 4 Cancer Control 423 Challenges remain in the field of cancer

care for young adults and adolescents.

Yellow Tulips. X-Ray image courtesy of Bryan Whitney. www.x-rayphotography.com.

Disparities in Adolescents and Young Adults With Cancer Leidy L. Isenalumhe, MD, Olivia Fridgen, MPH, Lynda K. Beaupin, MD, Gwendolyn P. Quinn, PhD, and Damon R. Reed, MD

Background: Cancer care for adolescents and young adults (AYAs) focuses on the care of patients aged 15 to 39 years. Historically, this group has favorable outcomes based on a preponderance of diagnoses such as thyroid cancers and Hodgkin lymphoma. Improvements in outcomes among the AYA population have lagged behind compared with younger and older populations. Methods: We discuss and review recent progress in AYA patient care and highlight remaining disparities that exist, including financial disadvantages, need for fertility care, limited clinical trial availability, and other areas of evolving AYA-focused research. Results: Survival rates have not improved for this age group as they have for children and older adults. Dis- parities are present in the AYA population and have contributed to this lack of progress. Conclusions: Recognizing disparities in the care of AYAs with cancer has led many medical specialty disci- plines to improve the lives of these patients through advocacy, education, and resource development. Research addressing barriers to clinical trial enrollment in this population, quality-of-life issues, and the improvement of survivorship care is also under way.

Introduction tion to improve outcomes and opined that major ac- In 2008, an article published in Cancer Control about ademic centers should develop the framework for a adolescents and young adults (AYAs) addressed this specialized oncology discipline focusing on AYAs.1 emerging, complicated, and multifaceted field.1 The Indeed, published research on AYAs has addressed authors concluded that more data were needed on aspects of their care but has not yet successfully treatment and survivorship in this patient popula- improved upon the entirety of oncology specific to AYAs. Programs devoted to AYAs have developed and specialized their focus to address fertility or psy- From the Adolescent and Young Adult Oncology (LLI, OF, GPQ, DRR) and Health Outcomes and Behavior Programs (GPQ, DRR), Sar- chosocial aspects of care, but no single model has yet coma Department (DRR), H. Lee Moffitt Cancer Center & Research to be universally adopted or standardized as the gold Institute, Tampa, Florida, Department of Pediatric Oncology (LKB), standard for the care of AYAs with cancer. Difficul- Roswell Park Cancer Institute, Buffalo, New York, and Morsani Col- lege of Medicine (GPQ), University of South Florida, Tampa, Florida. ties exist when discussing or acknowledging aspects Submitted February 9, 2016; accepted June 2, 2016. of AYA care, such as reproductive health and sexual- Address correspondence to Damon R. Reed, MD, Sarcoma Department, ity, which was the theme of the annual conference Moffitt Cancer Center, 12902 Magnolia Drive, FOB-1, Tampa, FL 33612. of the Critical Mass Young Adult Alliance in 2015.2 E-mail: [email protected] These topics were presented and discussed within No significant relationships exist between the authors and the companies/organizations whose products or services may be referenced the context of multidisciplinary teams. in this article. Disparities are defined by the National Cancer

424 Cancer Control October 2016, Vol. 23, No. 4 Institute as adverse differences in incidence, out- exist, because AYAs are inexperienced in navigating come, or burden of cancer existing among a popula- the health system or inquiring about clinical trials.4 tion group.3 Although chronological age demarcates They also may experience delays in diagnosis, be- oncology in AYAs, it is this emerging developmental cause malignancies are low on the differential diag- period of adulthood that brings in other aspects of nosis for ill patients who are young.12 disparities. In the context of oncology, AYAs are defined as individuals diagnosed with cancer between Presentation, Diagnosis, and Access 15 and 39 years of age.4 Cancer is the primary cause of to Health Care disease-related death in this population.5 In 2011, an Health Insurance and Financial Hardship estimated 69,212 AYAs were diagnosed with cancer, Cancer diagnosis is typically delayed in AYAs com- equating to 6 times more than children younger than pared with children. The cause is multifactorial and 14 years of age.6 According to a cancer statistics review relates to low suspicion of a malignant diagnosis by of Surveillance, Epidemiology, and End Results pro- both patient and caretaker, but also is related to inad- gram data from 1975 to 1997, improvement in 5-year equate insurance coverage.4 In the United States, those survival rates, as reflected in the average annual per- between the ages of 18 and 24 years and between 25 cent change in US cancer mortality rates, was signifi- and 34 years are more likely than any other age group cantly lower among individuals 15 to 39 years of age to be uninsured.13,14 Lack of health insurance in AYAs compared with those aged 1 to 14 years.4 AYAs have is associated with delays in diagnosis, an increased inferior outcomes compared with children and older likelihood of metastatic disease at presentation, and adults with the same diagnosis.7-11 a decreased likelihood of receiving definitive therapy Among younger AYAs aged 15 to 24 years, leuke- compared with those who are insured.15-18 In a study mia, lymphoma, central nervous system tumors, sar- of more than 39,000 participants aged 20 to 40 years, comas, and germ cell tumors are the most common health insurance coverage was associated with a de- malignancies.6 In AYAs 25 to 39 years of age, leuke- creased likelihood of metastatic disease at presenta- mia and lymphoma are still common, but the inci- tion; those who were uninsured were more likely to dence rates of breast cancer, melanoma, and colorec- be male, nonwhite, and unmarried.15 After adjusting tal cancer (CRC) are increased (Fig).6 for various factors, health insurance coverage was AYAs may experience economic disparities as also significantly associated with lower estimates of they enter the workforce, often in jobs without bene- all-cause mortality.15 In a study looking at the asso- fits.4 In addition, disparities in health education may ciation between health insurance coverage, socioeconomic status, and the risk of advanced Hodgkin 12,000 40,000 lymphoma in AYAs, pub-

35,000 lic insurance coverage for 10,000 low-income or individuals 30,000 with medical need was as- 8,000 25,000 sociated with an increased 6,000 20,000 risk of advanced disease

15,000 on presentation compared 4,000 with patients with private 10,000 insurance.16 Number of Observed Cases 2,000 Number of Observed Cases 5,000 Financial hardships 0 0 are not uncommon in Age Age Age Age Age patients with cancer but 15–19 y 20–24 y 25–29 y 30–34 y 35–39 y they may be more bur- a Leukemia and lymphoma Germ cell Breast Thyroid densome among AYAs.19 CNS Thyroid Leukemia and lymphoma Melanoma of the skin Malignant bone tumors Melanoma of the skin CNS Colon and rectum Some AYAs may have in- Soft tissue and Kaposi sarcoma Otherb Soft tissue and Kaposi sarcoma Cervix and uterus troductory level jobs with Germ cella Otherc limited time off and without the ability to take ex- aIncludes testicular cancer. bIncludes breast, cervical, colon, and other less-prevalent cancers. tended leave, have limited cIncludes malignant bone tumors and other less-prevalent cancers. savings, and may have college debt.19 Fig. — Common types of cancer affecting adolescents and young adults. CNS = central nervous system. Two-thirds of patients Reproduced with permission from the National Cancer Institute; Surveillance, Epidemiology, and End Results. A snapshot of adolescent and young adult cancer. https://www.cancer.gov/research/progress/snapshots aged 21 to 40 years report- /adolescent-young-adult. Accessed October 5, 2016. ed that their cancer ex-

October 2016, Vol. 23, No. 4 Cancer Control 425 perience negatively affected their financial situation.19 Clinical Trial Enrollment Financial hardship in relation to the costs of cancer The percentage of AYAs enrolled in clinical trials is treatment can cause stress, may result in reduced qual- lower than that seen in younger or older patients.21,25-29 ity of life, and may lead to patients terminating their AYAs who are uninsured, older in age, or treated by therapeutic plan.19 nonpediatric oncologists are less likely than older pa- The Affordable Care Act has enabled more than tients to be enrolled in a clinical trial.25 Even within 2.5 million young adults aged 19 to 25 years to gain ac- institutions that have the same clinical trials, enroll- cess to health insurance, although the uninsured rate is ment of AYAs is lower than the enrollment of the pedi- still highest among those aged 15 to 39 years.14,17 After atric population.28,29 Many obstacles contribute to the implementation of the Affordable Care Act Dependent low number of AYAs enrolled in clinical trials, one of Coverage Expansion, detection rates increased for ear- which is lack of knowledge about clinical trials among ly-stage cervical cancer among women aged 21 to 25 AYAs. In 1 study, 62% of AYAs did not know if clini- years secondary to increases in access to health care.20 cal trials were available for their type of cancer.30 Treat- Diagnosing cervical cancer in its early stage affords ment location is also a factor. A total of 40% of AYAs the oncologist the possibility of using fertility-spar- are treated at academic centers.31 Even when treated ing treatments; thus, the proportion of young women at a cancer center, not all clinical trials are available. with cancer able to receive fertility-sparing treatment Some hospitals are not allowed to enroll patients in increased.20 Further analyses are needed to determine Children Oncology Group trials if they are not partici- whether expanded coverage for health insurance is as- pating institutions.32 sociated with other favorable changes in AYAs. In 2012, the Centers for Disease Control and Pre- vention convened a group of oncologists who were Treatment experts in the AYA population to examine barriers to The optimal location of care is a complex topic. Exper- clinical trial enrollment among adolescents. Five issues tise for a given cancer diagnosis may lie within either were identified as the most important barriers33: pediatric or medical oncologists. Pathways could pro- 1. Low referral rates of adolescents with cancer to vide an optimal range of therapeutic options, identify ar- pediatric cancer centers eas for future clinical research for AYAs, and create dis- 2. Limited availability of clinical trials for certain ease-specific champions for AYAs passionate about the cancers care of this patient population. Although the treatment 3. Physician-related barriers limiting clinical trial protocol may be similar for certain cancers, others may accrual vary with respect to outcomes, cost of care, or intensity 4. Institutional barriers impeding collaboration of care.21 It is our opinion that a larger pool of expertise between pediatric and nonpediatric oncologists may exist at an adult facility for melanoma or neuroen- on clinical trials docrine diseases, but patients with sarcoma most likely 5. Unique psychosocial needs of adolescents with receive similar care at both the adult and pediatric can- cancer cer centers due to shared staffing. Care for a patient with Progress has been made by creating AYA pro- leukemia must involve increasing levels of collaboration grams at select cancer centers and by increasing the to provide personalized care using an evidence-based development and availability of clinical trials for approach to treatment.22 AYA patients.34,35 Although the number of programs Patients treated in a National Cancer Institute–des- for AYAs with cancer is increasing in the United ignated comprehensive cancer center have superior States, no formal accreditation, registry, or minimal outcomes compared with those treated at facilities standard exists by which to measure these targeted not designated as such, a fact likely due to the pre- programs; thus, the structures available and the ser- ponderance of trials conducted at these centers.18 vices offered vary among programs.36 In an attempt Barriers impeding access to National Cancer Insti- to standardize treatment and services for AYAs with tute–designated comprehensive cancer centers in- cancer, various organizations have developed treat- clude lack of health care insurance, public insurance, ment guidelines and recommendations for this pa- lower socioeconomic status, and African American or tient population (Table 1).4,35-40 In 2016, the National Hispanic race/ethnicity.22 Approximately 28% of pa- Comprehensive Cancer Network published updated tients older than 15 years of age are treated at commu- treatment guidelines for AYAs with cancer.37 Its rec- nity centers, not tertiary care centers.23 AYAs are the ommendations emphasize the need for a comprehen- least likely of all age groups to be treated at compre- sive, multidisciplinary approach to AYA care, includ- hensive cancer centers and, therefore, they may lack ing supportive and treatment guidelines that aim to opportunities for multidisciplinary care, tumor board improve treatment tolerance, compliance, and clini- discussions, subspecialty care (eg, rare tumors), and cal outcomes.37 the opportunity to enroll in clinical trials.22,24 Although dedicated efforts to design programs

426 Cancer Control October 2016, Vol. 23, No. 4 and vocational outcomes can be challenging. AYAs Table 1. — Select Resources for the Care of AYAs may require psychosocial care throughout the can- With Cancer cer care continuum because, if left unaddressed, such Reference Comment psychosocial issues may lead to poorer outcomes Adolescent and Establish age demarcation of AYA population than patients with access to age-appropriate resourc- Young Adult Summarize epidemiology es. During young adulthood, autonomy and self-re- Oncology Progress Review Group4 Dictate disparities and unique issues that liance are prized and contribute to a person’s sense affect AYAs of accomplishment and maturity.42 AYAs with cancer Ferrari35 AYAs require patient-focused multidisciplinary may have experienced these shortly before medical approach and financial hardships forced them into the care of A need exists for oncologists trained in their parents, away from the working world or from AYA oncology pursuing their education and possibly isolating them Importance of establishing a physical space 36 in the institution for AYAs from their peers without cancer. This isolation can Coccia39 Identify issues specific to AYAs be reinforced when AYAs see their friends achieving National Compre- Explain special considerations related to age-specific milestones, such as graduation, career hensive Cancer cancer management in AYAs success, marriage, and starting a family, because they 37 Network Provide supportive and treatment guidelines may be unable to achieve or experience these same that aim to improve treatment tolerance, life events while undergoing treatment or recovering compliance, and clinical outcomes in survivorship stages.43 Discuss long-term consequences Because of these experiences, the coordination of Promote participation in clinical trials cancer care while maintaining social connectivity is 36 Nass Summarize gaps in cancer care confronted paramount to the successful treatment of AYAs.42 Some by AYAs of the services AYAs require while undergoing treat- Describe potential strategies and actions to improve the quality of care for AYAs ment include financial counseling and insurance assis- Focus on quality-of-life issues and late effects tance, pain management, and mental health and peer Loren40 Discuss fertility preservation with all support (Table 3). The need for these services is not ex- patients of reproductive ages with cancer and clusive to AYAs with cancer; however, the life stage of discuss with parents/guardians of children AYAs exacerbates these associated needs, particularly and adolescents because many AYAs report that these needs remain A need exists for proper documentation 44 of fertility discussion and options provided unmet. AYAs have higher rates of distress at diagno- to patients sis than older matched populations, and they maintain Embryo and oocyte cryopreservation now high rates of distress during survivorship.45-47 Smith standard practice et al48 looked at the unmet needs in AYAs and found AYA = adolescent and young adult. that patients who lacked mental health services had the poorest health-related quality-of-life scores. AYAs have identified family support, religious beliefs, and aimed at increasing clinical trial enrollment in AYAs personal encouragement as the strongest sources of have shown some promise, improvements are more support during a cancer diagnosis and subsequent likely with sustained, systematic approaches to clini- treatment.49 Another study reported that patients who cal trial research in the AYA population.41 Efforts to had adequate health care insurance, a good family sup- improve trial enrollment are ongoing, some of which port system, and were properly informed about pre- use multidisciplinary, collaborative oncology and pe- serving their fertility were the most happy with their diatric specialist teams to study diseases such as leu- cancer care.50 In addition, AYAs who were the most kemia and sarcoma.38 Clinical trial registries, advocacy capable of coping with their diagnosis maintained in- groups, and efforts to increase collaborations and com- volvement in their career, school, or social spheres.47 munication across adult and pediatric cancer centers Relationships can shape the lives of AYAs and may are likely to continue progress in increasing clinical be a helpful source of support throughout the cancer trial enrollment rates in AYAs (Table 2). care continuum. Significant others or life partners may also play roles in decision-making, potential discourse, Psychosocial Needs and Late Effects and act as support systems. The key caregiver or health The wide age range of AYA care includes patients still care proxy assisting in medical decisions may vary under the guardianship of their parents to those in with each AYA, their state of residence, and may not committed relationships and married; thus, the psy- remain consistent throughout the disease trajectory. chosocial needs among this age group widely vary. Research suggests that the most important opinions Quality of life, adjusting to a cancer diagnosis, search- about the medical care of AYAs may not come from ing for meaning or understanding after a diagnosis, family members, as a clinical care team may assume;

October 2016, Vol. 23, No. 4 Cancer Control 427 spouses, and siblings, es- Table 2. — Select Clinical Trials for Adolescents and Young Adults pecially if family members Clinical Trial Type of Study Eligibility Study Objective make special work arrange- Identifier (Age), y ments and sacrifices to be Prospective Feasibility and safety of a pediatric regimen in patients more available to patients un- NCT01406756 16–39 phase 2 with ALL treated by a hematologist/oncologist dergoing treatment.56 Standard- or high-dose combination chemotherapy Randomized Health care profession- NCT02375204 ≥ 14 and stem-cell transplant in males with relapsed or phase 3 als have cited end-of-life care refractory germ cell tumors as the most difficult need Combination chemotherapy in treating newly NCT01406756 Phase 3 1–31 diagnosed ALL to meet in the AYA population.57-59 Several studies have Randomized Combination chemotherapy in treating nonmetastatic, NCT01231906 ≤ 50 phase 3 extracranial Ewing sarcoma looked at the best time to discuss advanced care plan- Randomized Combination chemotherapy ± ganitumab in newly NCT02306161 ≤ 50 phase 2 diagnosed metastatic Ewing sarcoma ning, suggesting that discus- NCT02470091 Phase 2 11–49 Use of denosumab for recurrent/refractory osteosarcoma sions should begin as early as the initial diagnosis and Use of eribulin mesylate in relapse/refractory NCT02097238 Phase 2 12–49 60-62 osteosarcoma as late as hospice referral. Preoperative chemoradiotherapy or preoperative Research also suggests that Randomized NCT02180867 ≥ 2 radiotherapy ± pazopanib in nonrhabdomyosarcoma AYAs who make their final phase 2/3 soft-tissue sarcoma wishes known regain more Comparing ipilimumab to high-dose interferon α-2b NCT01274338 Phase 3 ≥ 12 control of their health care in high-risk stage 3/4 melanoma after surgery situation than patients who Targeted therapy directed by genetic testing in NCT02465060 Phase 2 ≥ 18 do not express their wish- advanced refractory solid tumors or lymphomas es.63,64 In addition, AYAs who ALL = acute lymphoblastic leukemia, AYA = adolescent and young adult, Ph+ = Philadelphia chromosome positive. had a written advanced care planning document reported Table 3. — Select Resources for Adolescents and Adults With Cancer higher rates of active coping and emotional support than Focus Organization Website those without a written in- Fertility preservation Oncofertility Consortium www.oncofertility.northwestern.edu structions for advanced care Financial assistance SAMFund www.thesamfund.org planning.65 One study found Legal assistance Triage Cancer www.triagecancer.org that advanced care planning reduced levels of anxiety and National support and Cancer Care www.cancercare.org familial conflict, because resources Critical Mass www.criticalmass.org such wishes were discussed Patient advocacy and Stupid Cancer www.stupidcancer.org in advance and a health care awareness surrogate was identified.63 Physician education Focus Under Forty www.university.asco.org/focus-under-forty Caregivers of AYAs also ben- Program development Teen Cancer America www.teencanceramerica.org efitted from the conversation about end-of-life planning Program standardization Change it Back www.HCRI.org/programs/change-it-back as they were able, in times of great grief, to realize and carry out the wishes of their instead, AYAs hold opinions of friends and peers in loved one.63 Such events would not likely have hap- higher regard.51 Doing so may increase levels of stress pened without guidance for advanced care planning. or isolate patients in such a way that they feel they are Some meaningful progress in AYA care has in- making decisions on their own.52 volved the formative and focused research to identify Many AYAs are also unsure how or when to share the psychosocial needs of AYAs and the development their cancer diagnosis, impaired fertility, and medical of programs to help address those needs; similarly, debt, among other issues, with a prospective partner.53 AYA-focused organizations strive to provide financial, Self-confidence and vulnerability are important aspects vocational, and insurance assistance to cancer patients, of any relationship formation, but many AYAs struggle survivors, and their friends and families (see Table 3). with negative body issues or low levels of self-esteem, Progress was also made when the American Acad- and report that their experience with cancer has neg- emy of Pediatrics and the Institute of Medicine estab- atively affected their developing romantic relation- lished recommendations to include AYAs in their own ships.53-55 AYAs may also feel burdensome to parents, end-of-life discussions, stating that such discussions

428 Cancer Control October 2016, Vol. 23, No. 4 did not contradict a hopeful, positive outcome.61 The shown to take more sexual risk, such as not using ad- creation of Five Wishes (Aging With Dignity, Tallahas- equate contraceptive measures.77,78 Therefore, contra- see, FL) has allowed health care professionals to use ception should be a key discussion point with patients a more age-appropriate advance care planning tool to and their medical team, because some patients may as- respect the goals and priorities of AYAs at the end of sume they are infertile due to their treatment and think their life.61 This tool has also served as a starting point they have no need for contraception for pregnancy for feedback from AYAs and to continue to develop purposes or they may engage in unprotected sex to the most applicable advance care planning for young verify the fact that their infertility has been impaired.79 adults with cancer and the thoughts they wish to con- It is also worth noting that patients receiving active vey to their loved ones. treatment are more susceptible to sexually transmitted infections because their immune systems are compro- Fertility and Reproductive Health mised.77,80 Pregnancy while undergoing cancer treat- A significant psychosocial concern among AYA patients ment can pose risks to the fetus as well as the patient, is loss of fertility. Chemotherapy, radiotherapy, surgery, and making difficult decisions to terminate a pregnan- and cancer itself can cause temporary or permanent cy may cause increased stress and turmoil for a young damage to the reproductive organs and endocrine or patient already trying to cope with a cancer diagnosis pituitary functioning. The risk of permanent infertility and treatment. depends on patient age (younger patients typically fare Discussions of potential threats to fertility with all better than older patients), type of cancer and its treat- patients of childbearing age with newly diagnosed can- ment, and the dose and intensity of the treatment.66,67 cer are recommended by several medical organizations, In addition, females may experience premature ovarian including as the American Society of Clinical Oncol- insufficiency or failure, causing infertility, premature ogy,40 the National Comprehensive Cancer Network,68 menopause, or both.67 and the American Society for Reproductive Medicine.81 In the midst of a cancer diagnosis, understanding a In addition to discussion about the potential impact of patient’s risk of infertility and possible options for pretreatment or cancer type on fertility, the American So- serving fertility can be daunting and complex. Decision- ciety of Clinical Oncology and other resources recom- making about infertility and fertility preservation can be mend that AYA patients be referred to a reproductive challenging for the AYA patient because multiple barri- endocrinologist or infertility specialist to discuss op- ers exist.67-72 A patient may not have considered his or tions for fertility preservation as soon as possible after her desire for biological children in the future, and such a diagnosis of cancer is received and prior to the start decisions may become compounded by the stress of a of treatment (see Table 1).4,35-40 The established fertility cancer diagnosis.68 The patient or the oncologist may preservation methods are sperm cryopreservation for feel that a delay in treatment to pursue fertility preserva- postpubertal males and oocyte and embryo cryopreser- tion is unwise, and thus the patient may not have time vation for postpubertal females as well as ovarian trans- to pursue preservation.69,70 The financial costs associ- position for patients undergoing pelvic or abdominal ated with fertility preservation are often not covered by radiotherapy.37,40 For prepubertal children, fertility pres- health care insurance, thus making fertility preservation ervation options are experimental and include testicular out of reach for many AYA patients.67 Ethical, religious, or ovarian tissue cryopreservation.40 and other matters should be considered and may im- Studies suggest that patients who do not learn pede the decision-making of an AYA about fertility pres- about potential threats to their fertility may experi- ervation. Discussions about fertility and sexual health ence feelings of remorse and regret during the survi- may be uncomfortable for the AYA patient, particularly vorship period.82-84 Although few patients (particular- if his or her parents are present. Although the parents ly females) opt to use fertility-preservation methods, and loved ones of AYAs with cancer typically want to most patients report satisfaction and appreciation for be helpful and support the patient as much as possible, learning about this information from their health care studies suggest that they may not be appropriate proxy professional.85,86 decision-makers about fertility on behalf of AYA patients.71,72 Progress of Cancer Biology It is also worth noting that AYAs with cancer are as Though many common diagnoses afflicting the AYA sexually active as their peers without cancer.73-76 Thus, population overlap with those seen in younger and health care professionals should be aware of intimacy older patients, new data have emerged showing differ- concerns and changes due to treatment and have tact- ences in disease biology by age.8,87,88 ful and meaningful conversations with AYAs about re- Acute lymphoblastic leukemia (ALL) is the most ported adverse events and solutions, if possible. Equal- common malignancy and leading cause of cancer-rely important is for the oncologist to refer AYA patients lated deaths in AYAs.88 Outcomes of AYAs treated for to contraception counseling, because AYAs have been ALL are worse for those aged 10 years or older.8,88 The

October 2016, Vol. 23, No. 4 Cancer Control 429 treatment varies: In adult-based protocols, transplants are more common, whereas pediatric-based protocols Table 4. — Select Characteristics and Features of Cancers in AYA call for intensive therapy, including asparaginase, and do not rely as heavily on transplants.89-92 A difference Type Description exists in the frequency of cytogenetic abnormalities of Cancer that predispose AYAs to worse outcomes, but it may Leukemia8,10,93-97 Ph+ ALL t(9,22) BCR-ABL also serve as a target for improved therapy.8,10,93,94 3% of children Opportunities for improved outcomes may also ex- 3%–25% of AYAs ist in Philadelphia chromosome–positive ALL when T-cell leukemia tyrosine kinase inhibitors are added to treatment Age 1–9 y: 12.8% (Table 4).7-10,88,93-102 A large prospective trial of 296 pa- Age 10–15 y: 16.9% tients aged 16 to 39 years demonstrated that treatment Age 16–21 y: 24.7% with an intensive pediatric regimen showed signifi- Chromosome 21 (iAMP21) cant rates of improvement in event-free survival (66%) Increase incidence in patients aged > 10 y and overall survival (78%) along with manageable tox- Hypodiploid icities.89,90 Other trials have shown superior outcomes MLL at 11q23 in AYAs treated in pediatric protocols.89 Such ongoing TEL-AML1 translocation (ETV6-RUNX1) collaborations with cooperative groups may lead to 50% in children uniform treatment based on biology, with age being 10% in AYA one among many other variables. Sarcoma101,102 Ewing sarcoma Sarcoma is a common type of cancer in AYAs.103-107 More common in AYAs aged ≥ 15 y than children Overall, these mesenchymal-derived malignancies of Increasing age associated with poor prognosis the soft tissue and bone represent 8% of new diagno- Osteosarcoma ses in patients aged 15 to 39 years.108 Standard chemo- Poor event-free survival associated with age ≥ 18 y therapies for front-line regimens for osteosarcoma, Rhabdomyosarcoma rhabdomyosarcoma, and Ewing sarcoma have been Higher prevalence of alveolar subtype in AYAs, established through cooperative group studies.103-107 conferring a poorer prognosis In these studies, pediatric patients typically comprise Colorectal High incidence of microsatellite instability 10 the majority of the groups. Standard therapies for other cancer High incidence of heritable forms of colorectal soft-tissue sarcomas have been primarily derived from cancer (adenomatous polyposis, nonpolyposis colon cancer) studies of adult populations, in whom these diagnoses 109,110 Breast Higher prevalence of BRAC1/BRAC2 mutations are more common. Oftentimes, no clear consen- cancer99,100 sus exists regarding systemic therapies for metastatic, Present with stage III/IV disease relapsed, or recurrent disease.108 Some retrospective Higher incidence of aggressive pathology reviews of single-institution experiences and coopera- Triple-negative breast cancer tive group studies explore the presentation, biology, Age 15–39 y: 22.8% and therapeutic variables and how they generally neg- Age 40–49 y: 14.3% atively impact the outcomes of AYA patients compared Age ≥ 50 y: 11.7% - with younger patients (see Table 4).7-10,88,93-102,111-113 Ef- ERBB2 overexpressing cancer fort has been made to collaborate through the adult- Age 15–39 y: 10.3% and pediatric-focused cooperative groups and may Age 40–49 y: 6.9% serve as a model for future trials.38 Age ≥ 50 y: 5.9% Breast cancer has an age-related biological differ- Luminal B cancer Age 15–39 y: 17.9% ence that dictates outcomes.99,100,114,115 Breast cancer Age 40–49 y: 14.9% is one of the most frequently diagnosed malignan- Age ≥ 50 y: 10.7% cies among women aged 15 to 39 years and makes up 14% of all cancers diagnosed in AYAs.9 Nearly 7% Melanoma88 Risk factors: congenital melanocytic nevi, xeroderma pigmentosa, dysplastic nevus, prior of all breast cancers are diagnosed in women young- treatment for childhood cancer 8,99 er than 40 years of age. AYAs have a higher inci- Comprehensive staging guidelines not established dent rate of more aggressive phenotypes, such as for children and adolescents triple-negative breast cancer, ERBB2 (formerly known Metastases to sentinel lymph nodes occur more as HER2/neu)-overexpressing cancers, and lumi- frequently in children and AYAs than in adults with same disease stage nal B cancers.9,99,100,114,115 A low proportion of AYAs present with luminal A disease, which confers a bet- ALL = acute lymphoblastic leukemia, ter prognosis.100 In addition, a higher proportion of AYA = adolescent and young adult, + AYAs present with stage III/IV disease.99 AYA pa- Ph = Philadelphia chromosome positive.

430 Cancer Control October 2016, Vol. 23, No. 4 tients with breast cancer have a high prevalence of parities in AYAs have inspired individuals and institu- BRCA1/BRCA2 mutations.9,99,100,114,115 tions to develop programs designed for AYAs to meet CRC is another example of the biological and genet- the specific or overall health care needs of this patient ic differences seen in AYAs compared with older adults. population. In addition, some organizations have ad- The AYA population has high incidence rates of micro- vocated for AYA cancer care as a specific discipline satellite instability and heritable forms of CRC such as within oncology; to improve fertility education and ac- familial adenomatous polyposis and hereditary nonpol- cess to preservation; to address the root causes of the yposis colon cancer.10 It may also be true that the disease lack of survival improvement; and to define the essen- biology of young patients with CRC but without inher- tial elements of care for this population. Today we united syndromes may differ from the type of CRC seen derstand more about what AYA patients need, and we in older adults.10 Overall, compared with older patients, are poised to learn even more about approaches to im- AYAs present with more advanced-stage disease at diag- prove many aspects of care for these patients. We are nosis and more aggressive tumor features.116 optimistic that motivated patients, health care profes- Melanoma can affect individuals at any age, but it sionals, advocacy organizations, and institutions will is more common in adults than children.87 Its basic dis- align to develop improved models for the care of AYAs. ease biology — numerous mutations and a preponderance of BRAF V600E mutations — appears to hold true References across all ages.117 Metastases to sentinel lymph nodes 1. Soliman H, Agresta S. Current issues in adolescent and young adult cancer survivorship. Cancer Control. 2008;15(1):55-62. occur more frequently in children and AYAs than in 2. Young Adult Cancer Alliance. Critical Mass Annual Conference. adults with the same stage of disease.88 With the ap- November 4, 2015–November 6, 2015; Chicago, IL. http://criticalmass.org/ the-annual-conference/2015-conference. Accessed October 18, 2016. proval of ipilimumab, nivolumab, and other inhibitors 3. National Cancer Institute. Cancer health disparities. https://www. and combination therapies, we question whether AYAs cancer.gov/about-nci/organization/crchd/cancer-health-disparities-fact-sheet. Accessed October 5, 2016. were adequately represented and enrolled in recent 4. Adolescent and Young Adult Oncology Program Review Group. successful trials.118 Based on the biology of melanoma, Closing the Gap: Research and Care Imperatives for Adolescents And Young Adults With Cancer. Bethesda, MD: US Department of Health and similar treatment outcomes are expected in the AYA Human Services; National Institutes of Health; National Cancer Institute; population.118 LIVESTRONG Young Adult Alliance; 2006. https://www.cancer.gov/types/ aya/research/ayao-august-2006.pdf. Accessed October 5, 2016. 5. Lewis DR, Seibel NL, Smith AW, et al. Adolescent and Young Adult Remaining Challenges Cancer Survival. J Natl Cancer Inst Monogr. 2014;2014(49):228-235. 6. National Cancer Institute; Surveillance, Epidemiology, and End Re- The management of cancer in AYA patients is a contin- sults. A snapshot of adolescent and young adult cancer. https://www.can- uous process that has experienced progress in recent cer.gov/research/progress/snapshots/adolescent-young-adult. Accessed October 5, 2016. years. We now understand more about what AYA pa- 7. Schafer ES, Hunger SP. Optimal therapy for acute lymphoblastic leu- tients need and are poised to learn even more about kemia in adolescents and young adults. Nat Rev Clin Oncol. 2011;8(7):417-424. 8. Bleyer A, Barr R, Hayes-Lattin B, et al. The distinctive biology of cancer possible approaches to improve many aspects of AYA in adolescents and young adults. Nat Rev Cancer. 2008;8(4):288-298. care. However, doing so requires future research that 9. Keegan TH, Press DJ, Tao L, et al. Impact of breast cancer subtypes on 3-year survival among adolescent and young adult women. emphasizes the biological differences in certain diseas- Breast Cancer Res. 2013;15(5):R95. es to develop targeted treatments for this population. 10. Tricoli JV, Seibel NL, Blair DG, et al. Unique characteristics of adolescent and young adult acute lymphoblastic leukemia, breast cancer, and Lack of health insurance among this age group colon cancer. J Natl Cancer Inst. 2011;103(8):628-635. is still a major concern because being uninsured can 11. Zbuk K, Sidebotham EL, Bleyer A, et al. Colorectal cancer in young adults. Semin Oncol. 2009;36(5):439-450. potentially delay diagnosis and access to treatment as 12. Martin S, Ulrich C, Munsell M, et al. Delays in cancer diagnosis in under- well as to the possibility of inhibiting any transition to insured young adults and older adolescents. Oncologist. 2007;12(7):816-824. 13. Grassley CE, ed. Living Without Health Insurance: Hearing Before surveillance and life-long survivorship care. A large the Committee on Finance, US Senate. Darby, PA: Diane Publishing; 2001. study looking at the unmet needs of AYA patients re- 14. Smith JC, Medalia C. Health Insurance Coverage in the United States: 2013. Washington, DC: US Department of Commerce; Economics ported that 50% of AYA participants had unmet infor- and Statistics Administration; US Census Bureau; 2014. http://www.nber. mation needs about cancer recurrence, cancer treat- org/cps/hi/2014redesign/p60-250.pdf. Accessed October 5, 2016. 15. Aizer AA, Falit B, Mendu ML, et al. Cancer-specific outcomes among ments, and long-term adverse events, and up to 50% of young adults without health insurance. J Clin Oncol. 2014;32(19):2025-2030. AYA patients had unmet information needs on fertility, 16. Smith EC, Ziogas A, Anton-Culver H. Association between insurance 44 and socioeconomic status and risk of advanced stage Hodgkin lymphoma nutrition, and financial support. in adolescents and young adults: Hodgkin’s lymphoma in adolescents and young adults. Cancer. 2012;118(24):6179-6187. 17. Bleyer A, Ulrich C, Martin S. Young adults, cancer, health insurance, Conclusions socioeconomic status, and the Patient Protection and Affordable Care Act. Adolescent and young adult (AYA) patients face dispar- Cancer. 2012;118(24):6018-6021. 18. Rosenberg AR, Kroon L, Chen L, et al. Insurance status and ities based on age that can lead to a delay in diagnosis, risk of cancer mortality among adolescents and young adults. Cancer. an advanced stage of disease at presentation, subopti- 2015;121(8):1279-1286. 19. Bellizzi KM, Smith A, Schmidt S, et al. Positive and negative psycho- mal treatment location, lack of clinical trial availabil- social impact of being diagnosed with cancer as an adolescent or young adult. ity, and an increased burden of quality survivorship Cancer. 2012;118(20):5155-5162. 20. Robbins AS, Han X, Ward EM, et al. Association between the Afford- through the loss of fertility or disruption in social, able Care Act dependent coverage and cervical cancer stage and treatment in personal, or financial growth and stability. These dis- young women. JAMA. 2015;314(20):2189-2191.

October 2016, Vol. 23, No. 4 Cancer Control 431 21. Shaw PH, Reed DR, Yeager N, et al. Adolescent and young adult cancer patients during the first year following diagnosis. Psychooncology. (AYA) oncology in the United States: a specialty in its late adolescence. 2014;23(11):1267-1275. J Pediatr Hematol Oncol. 2015;37(3):161-169. 48. Smith AW, Parsons HM, Kent EE, et al. Unmet support service 22. Wolfson JA, Sun CL, Wyatt LP, et al. Impact of care at comprehensive needs and health-related quality of life among adolescents and young cancer centers on outcome: results from a population-based study. Cancer. adults with cancer: the AYA HOPE study. Front Oncol. 2013;3:75. 2015;121(21):3885-3893. 49. Kyngäs H, Mikkonen R, Nousiainen EM, et al. Coping with the onset 23. Yeager ND, Hoshaw-Woodard S, Ruymann FB, et al. Patterns of of cancer: coping strategies and resources of young people with cancer. care among adolescents with malignancy in Ohio. J Pediatr Hematol Oncol. Eur J Cancer Care (Engl). 2001;10(1):6-11. 2006;28(1):17-22. 50. Zebrack BJ, Mills J, Weitzman TS. Health and supportive care 24. Wolfson J, Sun CL, Kang T, et al. Impact of treatment site in adoles- needs of young adult cancer patients and survivors. J Cancer Surviv. cents and young adults with central nervous system tumors. J Natl Cancer 2007;1(2):137-145. Inst. 2014;106(8):dju166. 51. Zebrack B, Bleyer A, Albritton K, et al. Assessing the health care 25. Parsons HM, Harlan LC, Seibel NL, et al. Clinical trial participation needs of adolescent and young adult cancer patients and survivors. Cancer. and time to treatment among adolescents and young adults with cancer: 2006;107(12):2915-2923. does age at diagnosis or insurance make a difference? J Clin Oncol. 52. Cacioppo JT, Hawkley LC. Social isolation and health, with an emphasis 2011;29(30):4045-4053. on underlying mechanisms. Perspect Biol Med. 2003;46(3 suppl):S39-S52. 26. Fern L, Davies S, Eden T, et al. Rates of inclusion of teenagers and 53. Murphy D, Klosky JL, Reed DR, et al. The importance of assessing young adults in England into National Cancer Research Network clinical trials: priorities of reproductive health concerns among adolescent and young report from the National Cancer Research Institute (NCRI) Teenage and Young adult patients with cancer. Cancer. 2015;121(15):2529-2536. Adult Clinical Studies Development Group. Br J Cancer. 2008;99(12):1967-1974. 54. Nightingale CL, Quinn GP, Shenkman EA, et al. Health-related 27. McTiernan A. Issues surrounding the participation of adolescents with quality of life of young adult survivors of childhood cancer: a review of cancer in clinical trials in the UK. Eur J Cancer Care (Engl). 2003;12(3):233-239. qualitative studies. J Adolesc Young Adult Oncol. 2011;1(3):124-132. 28. Shaw PH, Ritchey AK. Different rates of clinical trial enrollment be- 55. Olsen PR, Harder I. Caring for teenagers and young adults with tween adolescents and young adults aged 15 to 22 years old and children un- cancer: a grounded theory study of network-focused nursing. Eur J Oncol der 15 years old with cancer at a children’s hospital. J Pediatr Hematol Oncol. Nurs. 2011;15(2):152-159. 2007;29(12):811-814. 56. Rechis R; LIVESTRONG Foundation. Session 2: unique risks & 29. Collins CL, Malvar J, Hamilton AS, et al. Case-linked analysis of vulnerabilities of AYA patients across the lifespan: AYA survey results. clinical trial enrollment among adolescents and young adults at a National Published July 2013. https://www.nationalacademies.org/hmd/~/media/Files/ Cancer Institute-designated comprehensive cancer center. Cancer. Activity%20Files/Disease/NCPF/2013-JUL-15/2A_Rechis.pdf. Accessed 2015;121(24):4398-4406. October 5, 2016. 30. Albritton KH, Coccia P. Influencing referral of adolescents and young 57. Durall A, Zurakowski D, Wolfe J. Barriers to conducting advance adults with cancer to sites with higher rates of trial enrollment. Pediatrics. care discussions for children with life-threatening conditions. Pediatrics. 2014;133(suppl 3):S104-S108. 2012;129(4):e975-e982. 31. Albritton KH, Wiggins CH, Nelson HE, et al. Site of oncologic spe- 58. Feudtner C. Collaborative communication in pediatric palliative care: cialty care for older adolescents in Utah. J Clin Oncol. 2007;25(29):4616-4621. a foundation for problem-solving and decision-making. Pediatr Clin North 32. Felgenhauer J, Hooke MC. Regulatory barriers to clinical trial Am. 2007;54(5):583-607. enrollment of adolescent and young adult oncology patients. Pediatrics. 59. Davies B, Sehring SA, Partridge JC, et al. Barriers to palliative care 2014;133(suppl 3):S119-S122. for children: perceptions of pediatric health care providers. Pediatrics. 33. Tai E, Buchanan N, Eliman D, et al. Understanding and addressing 2008;121(2):282-288. the lack of clinical trial enrollment among adolescents with cancer. Pediatrics. 60. Glare PA. Early implementation of palliative care can improve patient 2014;133(suppl 3):S98-S103. outcomes. J Natl Compr Canc Netw. 2013;11(suppl 1):S3-S9. 34. Reed DR, Block RG, Johnson R. Creating an adolescent and young 61. Wiener L, Ballard E, Brennan T, et al. How I wish to be remem- adult cancer program: lessons learned from pediatric and adult oncology bered: the use of an advance care planning document in adolescent and practice bases. J Natl Compr Canc Netw. 2014;12(10):1409-1415. young adult populations. J Palliat Med. 2008;11(10):1309-1313. 35. Ferrari A, Thomas D, Franklin AR, et al. Starting an adolescent and 62. Otsuka M, Koyama A, Matsuoka H, et al. Early palliative intervention young adult program: some success stories and some obstacles to overcome. for patients with advanced cancer. Jpn J Clin Oncol. 2013;43(8):788-794. J Clin Oncol. 2010;28(32):4850-4857. 63. Lyon ME, Jacobs S, Briggs L, et al. Family-centered advance care 36. Nass SJ, Beaupin LK, Demark-Wahnefried W, et al. Identifying and planning for teens with cancer. JAMA Pediatr. 2013;167(5):460. addressing the needs of adolescents and young adults with cancer: sum- 64. Wicks L, Mitchell A. The adolescent cancer experience: loss of con- mary of an Institute of Medicine workshop. Oncologist. 2015;20(2):186-195. trol and benefit finding: Adolescent cancer experience: control and benefit 37. National Comprehensive Cancer Network. Adolescent and Young finding. Eur J Cancer Care (Engl). 2010;19(6):778-785. Adult (AYA) Oncology. V1.2016. https://www.nccn.org/professionals/ 65. Loberiza FR, Swore-Fletcher BA, Block SD, et al. Coping styles, physician_gls/pdf/aya.pdf. Accessed October 5, 2016. health status and advance care planning in patients with hematologic 38. Weiss AR, Nichols CR, Freyer DR. Enhancing adolescent and young malignancies. Leuk Lymphoma. 2011;52(12):2342-2348. adult oncology research within the National Clinical Trials Network: rationale, 66. Moss JL, Choi AW, Fitzgerald Keeter MK, et al. Male adolescent progress, and emerging strategies. Semin Oncol. 2015;42(5):740-747. fertility preservation. Fertil Steril. 2016;105(2):267-273. 39. Coccia PF, Altman J, Bhatia S, et al. Adolescent and young adult 67. Levine JM, Kelvin JF, Quinn GP, et al. Infertility in reproductive- oncology. J Natl Compr Canc Netw. 2012;10(9):1112-1150. age female cancer survivors: infertility in female cancer survivors. Cancer. 40. Loren AW, Mangu PB, Beck LN, et al. Fertility preservation for pa- 2015;121(10):1532-1539. tients with cancer: American Society of Clinical Oncology clinical practice 68. Johnson RH, Kroon L. Optimizing fertility preservation practices for guideline update. J Clin Oncol. 2013;31(19):2500-2510. adolescent and young adult cancer patients. J Natl Compr Canc Netw. 41. Shaw PH, Boyiadzis M, Tawbi H, et al. Improved clinical trial en- 2013;11(1):71-77. rollment in adolescent and young adult (AYA) oncology patients after the 69. Quinn GP, Vadaparampil ST, Bell-Ellison BA, et al. Patient-physician establishment of an AYA oncology program uniting pediatric and medical communication barriers regarding fertility preservation among newly oncology divisions. Cancer. 2012;118(14):3614-3617. diagnosed cancer patients. Soc Sci Med. 2008;66(3):784-789. 42. D’Agostino NM, Penney A, Zebrack B. Providing developmentally 70. Quinn GP, Vadaparampil ST, Lee JH, et al. Physician referral for appropriate psychosocial care to adolescent and young adult cancer sur- fertility preservation in oncology patients: a national study of practice vivors. Cancer. 2011;117(10 suppl):2329-2334. behaviors. J Clin Oncol. 2009;27(35):5952-5957. 43. Wilkins KL, D’Agostino N, Penney AM, et al. Supporting adoles- 71. Crawshaw M. Psychosocial oncofertility issues faced by adolescents and young adults with cancer through transitions: position statement cents and young adults over their lifetime: a review of the research. Hum from the Canadian Task Force on Adolescents and Young Adults With Fertil. 2013;16(1):59-63. Cancer. J Pediatr Hematol Oncol. 36(7):545-551. 72. Quinn GP, Vadaparampil ST. Fertility Preservation and adolescent/ 44. Keegan TH, Lichtensztajn DY, Kato I, et al. Unmet adolescent and young adult cancer patients: physician communication challenges. J Adolesc young adult cancer survivors information and service needs: a population- Health. 2009;44(4):394-400. based cancer registry study. J Cancer Surviv. 2012;6(3):239-250. 73. Vadaparampil ST, Clayton H, Quinn GP, et al. Pediatric oncology nurs- 45. Clerici CA, Massimino M, Casanova M, et al. Psychological referral es’ attitudes related to discussing fertility preservation with pediatric cancer and consultation for adolescents and young adults with cancer treated at patients and their families. J Pediatr Oncol Nurs. 2007;24(5):255-263. pediatric oncology unit. Pediatr Blood Cancer. 2008;51(1):105-109. 74. King L, Quinn GP, Vadaparampil ST, et al. Oncology nurses’ per- 46. Sansom-Daly UM, Wakefield CE, Bryant RA, et al. Online group- ceptions of barriers to discussion of fertility preservation with patients with based cognitive-behavioural therapy for adolescents and young adults cancer. Clin J Oncol Nurs. 2008;12(3):467-476. after cancer treatment: a multicenter randomised controlled trial of Recap- 75. Clayton H, Quinn GP, Lee JH, et al. Trends in clinical practice and ture Life-AYA. BMC Cancer. 2012;12(1):339. nurses’ attitudes about fertility preservation for pediatric patients with can- 47. Zebrack BJ, Corbett V, Embry L, et al. Psychological distress and cer. Oncol Nurs Forum. 2008;35(2):249-255. unsatisfied need for psychosocial support in adolescent and young adult 76. Fallowfield L, Saul J, Gilligan B. Teaching senior nurses how to

432 Cancer Control October 2016, Vol. 23, No. 4 teach communication skills in oncology. Cancer Nurs. 2001;24(3):185-191. Group’s 2013 blueprint for research: bone tumors.. Pediatr Blood Cancer. 77. Murphy D, Klosky JL, Termuhlen A, et al. The need for reproductive 2013;60(6):1009-1015. and sexual health discussions with adolescent and young adult cancer 105. Womer RB, West DC, Krailo MD, et al. Randomized controlled trial patients. Contraception. 2013;88(2):215-220. of interval-compressed chemotherapy for the treatment of localized 78. Maslow BS, Morse CB, Schanne A, et al. Contraceptive use and the Ewing sarcoma: a report from the Children’s Oncology Group. J Clin Oncol. role of contraceptive counseling in reproductive-aged women with cancer. 2012;30(33):4148-4154. Contraception. 2014;90(1):79-85. 106. Malempati S, Hawkins DS. Rhabdomyosarcoma: review of the 79. Benedict C, Shuk E, Ford JS. Fertility issues in adolescent and Children’s Oncology Group (COG) Soft-Tissue Sarcoma Committee ex- young adult cancer survivors. J Adolesc Young Adult. 2016;5(1):48-57. perience and rationale for current COG studies. Pediatr Blood Cancer. 80. Laurence V, Gbolade BA, Morgan SJ, et al. Contraception for teen- 2012;59(1):5-10. agers and young adults with cancer. Eur J Cancer. 2004;40(18):2705-2716. 107. Weiss A, Gill J, Goldberg J, et al. Advances in therapy for pediatric 81. Ethics Committee of American Society for Reproductive Medicine. sarcomas. Curr Oncol Rep. 2014;16:395. Fertility preservation and reproduction in patients facing gonadotropic 108. Amankwah EK, Conley AP, Reed DR. Epidemiology and therapies therapies: a committee opinion. Fertil Steril. 2013;100(5):1224-1231. for metastatic sarcoma. Clin Epidemiol. 2013;5:147-162. 82. Benedict C, Thom B, Kelvin JF. Young adult female cancer survi- 109. Judson I, Verweij J, Gelderblom H, et al. Doxorubicin alone versus vors’ decision regret about fertility preservation. J Adolesc Young Adult intensified doxorubicin plus ifosfamide for first-line treatment of advanced Oncol. 2015;4(4):213-218. or metastatic soft-tissue sarcoma: a randomised controlled phase 3 trial. 83. Crawshaw M, Sloper P. A Qualitative Study of the Experiences of Lancet Oncol. 2014;15(4):415-423. Teenagers and Young Adults When Faced With Possible or Actual Fertility 110. van der Graaf WT, Blay JY, Chawla SP, et al. Pazopanib for metastat- Impairment Following Cancer Treatment. York, UK: University of York; 2006. ic soft-tissue sarcoma (PALETTE): a randomised, double-blind, placebo- http://www.york.ac.uk/inst/spru/pubs/pdf/fertilityExecSumm.pdf. Accessed controlled phase 3 trial. Lancet. 19;379(9829):1879-1886. October 5, 2016. 111. Streby KA, Ruymann FB, Whiteside S, et al. Rhabdomyosarcoma 84. Chapple A, Salinas M, Ziebland S, et al. Fertility issues: the percep- in adolescent and young adults: a 25-year review at Nationwide Children’s tions and experiences of young men recently diagnosed and treated for Hospital. J Adolesc Young Adult Oncol. 2012;1(4):164-167. cancer. J Adolesc Health. 2007;40(1):69-75. 112. Collins M, Wilhelm M, Conyers R, et al. Benefits and adverse 85. Garvelink MM, ter Kuile MM, Bakker RM, et al. Women’s experiences events in younger versus older patients receiving neoadjuvant chemo- with information provision and deciding about fertility preservation in the therapy for osteosarcoma: findings from a meta-analysis. J Clin Oncol. Netherlands: “satisfaction in general, but unmet needs.” Health Expect. 2013;31(18):2303-2312. 2015;18(5):956-968. 113. Haddox CL, Han G, Anijar L, et al. Osteosarcoma in pediatric pa- 86. Letourneau JM, Ebbel EE, Katz PP, et al. Pretreatment fertility tients and young adults: a single institution retrospective review of presen- counseling and fertility preservation improve quality of life in reproductive tation, therapy, and outcome. Sarcoma. 2014;2014: 402509. age women with cancer. Cancer. 2012;118(6):1710-1717. 114. Parise CA, Bauer KR, Caggiano V. Variation in breast cancer sub- 87. American Cancer Society. Cancer Facts & Figures 2016. http:// types with age and race/ethnicity. Crit Rev Oncol Hematol. 2010;76(1):44-52. www.cancer.org/acs/groups/content/@research/documents/document/ 115. Gewefel H, Salhia B. Breast cancer in adolescent and young adult acspc-047079.pdf. Accessed October 18, 2016. women. Clin Breast Cancer. 2014;14(6):390-395. 88. Tricoli JV, Blair DG, Anders CK, et al. Biologic and clinical charac- 116. Teng A, Lee DY, Cai J, Patel SS, et al. Patterns and outcomes teristics of adolescent and young adult cancers: acute lymphoblastic leu- of colorectal cancer in adolescents and young adults. J Surg Res. kemia, colorectal cancer, breast cancer, melanoma, and sarcoma. Cancer. 2016;205(1):19-27. 2016;122(7):1017-1028. 117. Lu C, Zhang J, Nagahawatte P. The genomic landscape of child- 89. Curran E, Stock W. How I treat acute lymphoblastic leukemia in hood and adolescent melanoma. J Invest Dermatol. 2015;135(3):816-823. older adolescents and young adults. Blood. 2015;125(24):3702-3710. 118. Sreeraman Kumar R, Messina JL, Sondak VK, et al. Treating mela- 90. Stock W, Luger SM, Advani AS, et al. Favorable outcomes for noma in adolescents and young adults: challenges and solutions. Clin Oncol older adolescents and young adults (AYA) with acute lymphoblastic Adolesc Young Adults. 2015;5:75-86. leukemia (ALL): early results of U.S. Intergroup trial C10403. Blood. 2014;124(21 suppl):796. 91. Isakoff M, Freyer D, Bleyer A. Young adults with acute lymphoblastic leukemia treated with a pediatric-inspired regimen do not need a bone marrow transplant in first remission. Blood. 2013;121(26):5253-5255. 92. Ram R, Wolach O, Vidal L, et al. Adolescents and young adults with acute lymphoblastic leukemia have a better outcome when treated with pediatric-inspired regimens: Systematic review and meta-analysis. Am J Hematol. 2012;87(5):472-478. 93. Perez-Andreu V, Roberts KG, Xu H, et al. A genome-wide association study of susceptibility to acute lymphoblastic leukemia in adolescents and young adults. Blood. 2015;125(4):680-686. 94. Mullighan CG, Willman CL. Advances in the biology of acute lymphoblastic leukemia—from genomics to the clinic. J Adolesc Young Adult Oncol. 2011;1(2):77-86. 95. Bleyer WA, Barr RD, eds. Pediatric Oncology: Cancer in Adolescents and Young Adults. New York: Springer; 2007. 96. Roberts KG, Li Y, Payne-Turner D, et al. Targetable kinase- activating lesions in Ph-like acute lymphoblastic leukemia. N Engl J Med. 2014;371(11):1005-1015. 97. Pui CH, Yang JJ, Hunger SP, et al. Childhood acute lymphoblastic leukemia: progress through collaboration. J Clin Oncol. 2015;33(27):2938-2948. 98. Janeway KA, Barkauskas DA, Krailo MD, et al. Outcome for adolescent and young adult patients with osteosarcoma: a report from the Children’s Oncology Group. Cancer. 2012;118(18):4597-4605. 99. Keegan TH, DeRouen MC, Press DJ, et al. Occurrence of breast cancer subtypes in adolescent and young adult women. Breast Cancer Res. 2012;14(2):R55. 100. Assi HA, Khoury KE, Dbouk H, et al. Epidemiology and prognosis of breast cancer in young women. J Thorac Dis. 2013;5(suppl 1):S2-S8. 101. Janeway KA, Pappo A. Treatment guidelines for gastrointestinal stromal tumors in children and young adults. J Pediatr Hematol Oncol. 2012;34:S69-S72. 102. Bisogno G, Compostella A, Ferrari A, et al. Rhabdomyosarcoma in adolescents: a report from the AIEOP Soft Tissue Sarcoma Committee. Cancer. 2012;118(3):821-827. 103. Marina N, Bielack S, Whelan J, et al. International collaboration is feasible in trials for rare conditions: the EURAMOS experience. In: Jaffe N, Bruland OS, Bielack S, eds. Pediatric and Adolescent Osteosarcoma. Boston: Springer US; 2010. 104. Gorlick R, Janeway K, Lessnick S, et al. Children’s Oncology

October 2016, Vol. 23, No. 4 Cancer Control 433 Tobacco-related health disparities

exist at each point of

the cancer care continuum.

Iris. X-Ray image courtesy of Bryan Whitney. www.x-rayphotography.com.

Tobacco-Related Health Disparities Across the Cancer Care Continuum Vani Nath Simmons, PhD, Bárbara Piñeiro, PhD, Monica Webb Hooper, PhD, Jhanelle E. Gray, MD, and Thomas H. Brandon, PhD

Background: Use of tobacco is the leading preventable cause of death in the United States. Racial/ethnic minorities and individuals of low socioeconomic status disproportionately experience tobacco-related disease and illness. Unique challenges and circumstances exist at each point in the cancer care continuum that may contribute to the greater cancer burden experienced by these groups. Methods: We reviewed tobacco-related disparities from cancer prevention to cancer survivorship. We also describe research that seeks to reduce tobacco-related disparities. Results: Racial/ethnic minorities and low-income individuals experience unique social and environmental contextual challenges such as greater environmental cues to smoke and greater levels of perceived stress and social discrimination. Clinical practice guidelines support the effectiveness of pharmacotherapy and behavioral counseling for racial and ethnic minorities, yet smoking cessation rates are lower in this group when compared with non-Hispanic whites. Superior efficacy for culturally adapted interventions has not yet been established. Conclusions: To reduce health disparities in this population, a comprehensive strategy is needed with efforts directed at each point along the cancer care continuum. Strategies are needed to reduce the impact of contextual factors such as targeted tobacco marketing and social discrimination on smoking initiation and maintenance. Future efforts should focus on increasing the use of evidence-based cessation treatment methods and studying its effectiveness in these populations. Attention must also be focused on improving treatment outcomes by reducing smoking in diverse racial and ethnic patient populations.

From the Health Outcomes and Behavior Program (VNS, THB) Introduction and Department of Medical Oncology (JEG), H. Lee Moffitt Cancer In the United States, tobacco smoking represents the Center & Research Institute, Tampa, Florida, Department of 1,2 Clinical Psychology and Psychobiology (BP), University of leading preventable cause of death. Nearly one-third Santiago de Compostela, Santiago de Compostela, Spain, and the of all cancer-related deaths are attributable to tobacco Department of Psychology (MWH), Sylvester Comprehensive use.1-3 The health burden of tobacco use is not equally Cancer Center, University of Miami, Miami, Florida. distributed: Racial and ethnic minorities who smoke Submitted April 13, 2016; accepted July 28, 2016. Address correspondence to Vani N. Simmons, PhD, Health Outcomes and have lower levels of education, income, and occu- and Behavior Program, Moffitt Cancer Center, 4115 East Fowler Av- pational status disproportionately experience tobac- enue, Tampa, FL 33617. E-mail: [email protected] co-related disease burden and illness.4 Examples of Dr Brandon receives grants and research support from Pfizer and is a tobacco-related disparities include exposure to tobacco consultant for Voxiva and the Australian General Solicitor. The other authors report no significant relationships with the companies/orga- smoke, current patterns of use and cessation, subse- nizations whose products or services may be referenced in this article. quent health consequences among specific population

434 Cancer Control October 2016, Vol. 23, No. 4 groups, capacity and infrastructure, and access to re- dition to the potential health outcomes of increased sources.4 Attention to tobacco-related health disparities exposure to second-hand smoke, less-restrictive poli- has been growing, and the number of priority popula- cies may serve to maintain smoking behavior. Com- tions has extended beyond racial and ethnic minorities pared with workplaces that do not have smoking re- to include individuals of low socioeconomic status as strictions, workplaces that prohibit smoking result in well as sexual minorities (lesbian, gay, bisexual, and fewer cigarettes smoked, greater cessation attempts, transgender).5 In addition, those in the tobacco-related and more quitting successes among employees.30 Re- health disparities field have moved beyond descriptive search also suggests that lack of self-imposed smoking studies and have begun to develop and test interven- restrictions in the home contributes to greater main- tions to help reduce smoking among these vulnerable tenance of smoking behavior, whereas the adoption populations.6 of smoke-free restrictions results in greater cessation Although tobacco use contributes to the presence ability.31 Individuals who smoke, have a lower socio- and exacerbation of health disparities with respect to economic status, and are African American are less health outcomes such as stroke, diabetes, and coronary likely than non-Hispanic whites to impose smoking heart disease, this paper specifically focuses on the bans in the home and automobile; thus, environmen- association between tobacco use and cancer.7 Unique tal exposure in these settings is an additional barrier challenges and circumstances also exist at different to consider.10-13 Disproportionate exposure to tobacco points along the cancer care continuum that contribute smoke is also a public health issue among residents of to the larger cancer burden experienced by ethnic/racial multiunit housing properties who may be exposed to minorities and underserved populations (Fig).3,4,8-27 second-hand smoke due to shared air between units and poor ventiliation.32 Residents of subsidized hous- Cancer Prevention ing also experience high levels of stress, which can Evidence suggests that cigarette smoking is associated impact smoking behavior.33 with increased risks of developing oropharyngeal, laryngeal, esophageal, tracheal, bronchial, lung, stom- Targeted Marketing ach, liver, pancreatic, kidney, ureteral, cervical, blad- Another key factor that influences the initiation and der, and colorectal cancers as well as acute myeloid maintenance of tobacco smoking among low-income leukemia.1,2 Thus, decreasing rates of tobacco initiation, as well as increasing smoking cessation rates, are im- Key Factors Influencing Disparities at Each Point portant steps for cancer prevention. To further our understanding of tobacco initiation and maintenance Cancer Prevention Cancer Diagnosis Cancer Survivorship among racial/ethnic minorities and individuals of lower socioeconomic status, it is important to consider en- Environmental African Americans Multiple negative influences contribute have highest lung effects of continued vironmental and social contextual in- to smoking, exposure 9 cancer incidence smoking on cancer fluences. to second-hand compared to all outcomes3,22 smoke9-13 other racial and ethnic Environmental Influences groups21 Tobacco-control policies, such as Low representation smoking restrictions, can be effec- Targeted tobacco marketing creates of racially/ethnically tive at reducing tobacco initiation, vulnerability14 Racial and ethnic diverse patients use, and related health care costs.2 minorities more likely with cancer in cessation trials26,27 Despite progress on reducing the to be diagnosed at more advanced-stage amount of exposure to second-hand Perceived stress/ disease23 smoke via smoke-free restrictions discrimination leads to in public places, disparities in work- more smoking15,16 Lack of education, training, and place policies can create or exacer- infrastructure for bate tobacco-related disparities in Potential unequal providers to racial/ethnic minority and under- Validated smoking implementation implement smoking of cancer screening 28 cessation options cessation into served populations. For example, lacking for minorities20 guidelines8,17-19 clinical care4,24,25 those who smoke and are employed in blue-collar or service occupations are more likely to work in environments that permit smoking.29 In ad- Fig. — Tobacco-related health disparities exist across the cancer care continuum.

October 2016, Vol. 23, No. 4 Cancer Control 435 and racial/ethnic minority populations is targeted to- Secondary Prevention bacco marketing.2 The tobacco industry uses targeted To work toward eliminating tobacco-related cancer marketing and promotions to reach ethnic/racial mi- disparities, we must add to our understanding of so- norities, sexual minorities, and low-income commu- cial and environmental factors that influence tobacco nities.2,14 Examples of these industry efforts include use among underserved populations so that primary a greater presence of tobacco billboards in specific prevention strategies to prevent smoking initiation neighborhoods and the targeted marketing of men- may be implemented. Equally important are second- thol cigarettes.14 In a systematic review, the presence ary prevention strategies that seek to develop effective and quantity of tobacco marketing at tobacco retailers smoking cessation interventions to help those who were examined by neighborhood characteristics such have started smoking. Several evidence-based treat- as race, ethnicity, and socioeconomic status.14 The re- ments exist for treating tobacco dependence. Clinical sults suggested that lower-income neighborhoods are practice guidelines from the US Public Health Service exposed to more tobacco marketing, and the amount summarize these treatments and describe a brief in- of marketing of menthol products is higher in both ur- tervention model called the 5 As8,41: ask about tobacco ban neighborhoods and communities with larger num- use, advise to quit smoking, assess interest in quitting bers of African American residents.14 Among neighbor- smoking, assist in quitting smoking via counseling hoods with a greater proportion of Hispanic residents, and pharmacotherapy, and arrange follow-up. These no observed increase in tobacco marketing was seen recommendations apply to a broad population of per- (possibly due to cultural reasons).2,14 sons who smoke and include those from all racial and ethnic groups, individuals with low socioeconomic sta- Perceived Stress and Social Discrimination tus, those with limited formal education, those who are Racial/ethnic minorities and low-income individuals hospitalized, and those who identify as lesbian, gay, report higher levels of perceived stress, which is de- bisexual, or transgender.41 Despite the applicability of fined by the degree to which individuals perceive cir- these guidelines for diverse populations, the unequal cumstances in their life as stressful.15,34 Elevated levels implementation of these guidelines has been docu- of stress may can be attributed to several factors such mented.8,17-19 Among those who smoke, Hispanics and as lower wages, more daily hassles, and stressful life those with low socioeconomic status are less likely to events.15,35 Social discrimination is also a commonly receive advice and assistance from their health care perceived stressor among African Americans and La- professional about smoking cessation compared with tinos.34 Perceived levels of stress and social discrimina- non-Hispanic whites and those with higher socioeco- tion have been associated with cancer burden and to- nomic status, respectively.17-19 Access to evidence-based bacco smoking.2 The prevalence of tobacco smoking is smoking cessation methods is also lacking for individ- higher among individuals who perceive greater rates of uals with limited or no English proficiency.42 The ab- discrimination.16 For example, compared with nonmi- sence of readily available, Spanish-language, evidence- nority individuals, lesbian, gay, bisexual, and transgen- based materials has been cited as a key barrier to the der individuals are more likely to experience stressors ability of primary care physicians to assist with smok- such as discrimination and stigma and have a higher ing cessation.43 smoking prevalence rate (between 24% and 45%).36 In addition, an inverse relationship has been observed Targeted Smoking Cessation between perceived levels of stress and rates of smok- The US Public Health Service guidelines for evi- ing cessation.37 dence-based smoking cessation interventions apply to A primary mechanism through which stress may all racial and ethnic minorities.8 However, important influence smoking behavior is via negative-affect man- differences exist between racial/ethnic minorities and agement.34 Although tobacco smoking may not actually whites that warrant research on the effectiveness of reduce stress (except for nicotine withdrawal-related smoking cessation interventions.44 The smoking preva- stress), if those who smoke believe that smoking reduc- lence rate for non-Hispanic whites (18.2%) is compa- es stress, then that belief will motivate smoking — ie, rable with that seen in African Americans (17.5%) but smoking is often used as a coping strategy for dealing is lower for Hispanics (11.2%).45 However, smoking with stress, and, as stress increases, so does smoking.38 prevalence varies by cultural subgroup; for example, The relationship between negative affect and smoking the smoking prevalence among individuals of Puerto has been well documented.39,40 The greater perceived Rican and Cuban descent exceeds 30%.46 Moreover, ra- level of stress among racial/ethnic minorities and low- cial and ethnic minorities are disproportionately rep- income populations, as well as the robust relationship resented in lower-income categories, where smoking between stress and smoking, remains an important fac- prevalence is highest.45 Several notable differences in tor to assess and consider in the development of target- smoking behavior patterns also exist among different ed interventions for underserved persons who smoke. groups. Although smoking prevalences among African

436 Cancer Control October 2016, Vol. 23, No. 4 Americans and non-Hispanic whites are comparable, Tested modalities for behavioral interventions African Americans start smoking at a later age, are have included both group and individual counsel- more likely to smoke menthol cigarettes, and smoke ing.46,54 In a randomized trial, Webb et al57 found fewer cigarettes per day on average.3,47,48 Despite lower that African Americans who smoked and received daily rates of smoking among Hispanics and African group-based cognitive behavioral therapy to help Americans compared with non-Hispanic whites, racial them quit smoking had higher cessation rates than and ethnic minorities who smoke have greater difficul- African Americans who smoked and were involved in ty quitting smoking.49 a general health education group that controlled for Cultural factors and acculturation must be taken contact and intensity. In addition to support for group into account, because those who are more acculturated therapy, researchers concluded that individual coun- are more likely to smoke and have lower odds of sus- seling (in-person and on the telephone) was effective taining cessation.46 Limited interventional research has for aiding smoking cessation among racial and ethnic also been conducted with American Indians and Alaska minorities.44 Natives, who have the highest rates of tobacco use — in- Some research has focused on the development cluding both ceremonial and commercial tobacco use.5,45 of culturally adapted smoking cessation interventions Benowitz et al50,51 also reports that differences that take into account unique cultural characteristics.44 may exist in the genetics and physiology (eg, how However, little research has been conducted to evalu- fast nicotine is metabolized) among racial and ethnic ate the effectiveness of culturally adapted or culturally groups that warrant the evaluation of specific smok- specific interventions compared with generic interven- ing cessation interventions among these groups, rath- tions. Select studies conducted with African Ameri- er than assuming equivalent rates of efficacy across cans who smoke have been mixed.20,58-62 One study groups. This concept is relevant given that most clini- found evidence of a greater level of interest and atten- cal trials for smoking cessation medications include tion for culturally specific materials, yet readiness to few minorities who smoke.50 quit and cessation attempts were greater for the stan- Narrative and meta-analytic reviews of smoking dard materials.58 By contrast, a study conducted with cessation interventions have been conducted for racial African Americans who smoke found greater rates of and ethnic minorities who smoke.44,52-54 One compre- smoking abstinence at a 12-month follow-up period hensive and narrative review included 64 studies in- among participants who received culturally tailored volving African Americans, Latinos, American Indians, materials than those who received standard materi- Asians, and Pacific Islanders who smoked.44 Pharma- als.59 Other studies have also demonstrated the ben- cotherapy, behavioral interventions, and community efits of culturally tailored interventions for Hispanics interventions for smoking cessation were examined. who smoke.60,61 Overall, researchers have reported that The authors concluded that existing data support the study participants are generally accepting of interven- use of pharmacotherapy treatments for smoking ces- tional materials that include information about cultural sation among African Americans, Latinos, and His- adaptation, yet additional research is needed to de- panics who smoke.44 Specifically, nicotine patches, termine whether such cultural tailoring is superior to nicotine nasal sprays, and bupropion were effective in standard materials to improve smoking cessation rates African Americans who smoke, whereas the nicotine among minorities.20,58,62 patch alone was effective for Hispanics who smoke.44 Nicotine gum has not demonstrated efficacy in 3 stud- Cancer Diagnosis ies with African Americans who smoke, and trials have Higher rates of cancer morbidity and mortality are not been conducted with nicotine lozenges, inhalers, evident among racial/ethnic minorities and those with or combination nicotine-replacement therapy.44 low socioeconomic status.21 Several pathways exist in Regardless of the outcomes of efficacy trials, which the relationship between tobacco and cancer real-world effectiveness and use of pharmacotherapy health disparities can present. A higher risk of tobacco- must be considered. Qualitative research conducted related cancers among persons who smoke and have by Fu et al55 with minorities who smoke (African Ameri- low incomes could, in part, be a function of higher cans, American Indians, and Southeast Asians) revealed smoking prevalence rates. Although the overall preva- that knowledge is lacking about how pharmacotherapy lence of smoking has declined to 16.8%, low-income works as well as the skepticism regarding the effective- individuals who smoke (ie, household income below ness of smoking cessation. These beliefs may explain the poverty level) have a smoking prevalence rate of the low utilization of pharmacotherapy among racial 30.4%.63 Other pathways are less clear. Referred to as and ethnic minorities who smoke.56 Further research on the “smoker’s paradox,” African Americans smoke few- the beliefs and barriers of racial and ethnic differences er cigarettes per day and are less likely to smoke daily, regarding pharmacotherapy may aid in the develop- yet they have higher rates of lung, esophageal, and oral ment of strategies to increase use of cessation methods. cancers than non-Hispanic whites who smoke.64

October 2016, Vol. 23, No. 4 Cancer Control 437 Disparities in Lung Cancer findings underscore the need to understand percep- Lung cancer is the leading cause of cancer-related tions of lung cancer by race/ethnicity to identify po- deaths in the United States, and 80% of these deaths tential barriers to smoking cessation and lung cancer are attributable to cigarette smoking.2 African Ameri- treatment. cans have the highest lung cancer incidence and mortality rates compared with all other racial and ethnic Challenges in Screening groups.3,21 Risk factors for poor cancer outcomes Based on evidence from the National Lung Cancer have been attributed to several factors such as treat- Screening Trial, low-dose computed tomography for ment access and lower socioeconomic status, as well lung cancer screening provides an opportunity to re- as smoking-related factors.22 The disproportionate duce the rate of lung cancer mortality.70 A concern burden in lung cancer risk among African Americans raised by Holford et al71 is that the 30 pack-years’ eli- cannot be explained by a greater number of ciga- gibility criteria required by the US Preventive Servic- rettes smoked.64 However, a distinguishing smokes Task Force and Centers for Medicare & Medicaid ing behavior among African Americans who smoke, Services may result in fewer screenings for African which may aid in explaining disparate disease out- American men who smoke and have fewer pack years. comes, is that African Americans are more likely to African Americans have also been found to be more re- smoke menthol cigarettes than whites.3 As part of the luctant to get screened for lung cancer due to fear of 2009 Family Smoking Prevention and Tobacco Con- the disease.66 Thus, it is possible that unequal imple- trol Act, the Tobacco Products Scientific Advisory mentation of newer screening modalities and height- Committee of the US Food and Drug Administration ened fears of disease may broaden cancer disparities. was tasked with submitting a report on the impact of menthol.65 Findings from that report suggest that Head and Neck Cancers little evidence exists to demonstrate that menthol Tobacco use is a major risk factor for the develop- contributes to greater disease risk; however, menthol ment of head and neck cancers: 75% of all head and cigarettes may have more addictive qualities than neck cancers are attributable to tobacco and alcohol nonmenthol cigarettes, which could contribute to use.72 Molina et al23 examined survival rates of head low rates of smoking cessation.3,65 Menthol also has and neck cancers based on race/ethnicity and so- cooling and anesthetic effects that reduce the harsh- cioeconomic status. Their findings suggested that, ness of cigarette smoke.65 Use of menthol cigarettes compared with non-Hispanic whites, African Ameri- is also associated with increased initiation and pro- cans were diagnosed with head and neck cancers at gression to cigarette smoking and difficulty quitting a younger age and at more advanced-disease stag- smoking.65 Based on these findings, the Tobacco es.23 In addition, the average survival time for Afri- Products Scientific Advisory Committee concluded can American patients was significantly lower than that no longer selling menthol cigarettes would be whites (21 vs 40 months).23 Poor survival outcomes in of a public health benefit.65 However, more research African Americans were observed despite the lack of is needed to further examine the potential role that difference in the frequency of smoking. Living in an menthol plays in the unequal disease burden expe- area of high poverty was found to be an independent rienced by African Americans.3 Specifically, as noted predictor of worse outcomes.73 Poor cancer outcomes by Alexander et al,64 research is needed on examin- among patients from impoverished communities has ing additional pathways such as the role of taste sen- also been attributed to lack of access to health care, sitivity and the effects of menthol on cellular mecha- decreased likelihood of insurance coverage, and ad- nisms and subsequent disease processes. vanced disease at presentation.73 Among Hispanics and Latinos, incidence and mortality rates of head and neck Role of Beliefs About Lung Cancer cancers are lower than both African Americans and Research suggests that beliefs about lung cancer may non-Hispanic whites.74 differ by race/ethnicity. Compared with whites, African Americans have lower perceived notions of risk Cancer Survivorship for lung cancer and are more likely to doubt the rela- Once an individual who smokes has been diagnosed tionship between lung cancer and behavior.66,67 Afri- with cancer, he or she may mistakenly believe that it can Americans also have demonstrated greater skepti- is too late to quit smoking; however, a growing body cism regarding surgical treatment for lung cancer.68,69 of research suggests that patients with cancer who Moreover, the belief that exposure to air during sur- continue to smoke are at risk for several adverse gery can cause the tumor to spread and fatalistic be- health outcomes.1 The 2014 report of the US Surgeon liefs (eg, “If I have lung cancer, I believe it was meant to General concluded that sufficient evidence has accu- be”) may serve as barriers to accepting surgical inter- mulated to infer a causal relationship between indi- vention among racial and ethnic minorities.68,69 These viduals with cancer who smoke cigarettes and cancer

438 Cancer Control October 2016, Vol. 23, No. 4 survivors for both all-cause mortality and cancer-spe- Despite the importance of quitting smoking for cific mortality.1 Evidence also suggests that contin- individuals with cancer, few clinical trials of smoking ued smoking results in an increased risk for second cessation have been conducted in this population. In primary cancers known to be caused by cigarette a systematic review and meta-analysis, 10 randomized smoking, and, although this evidence is insufficient controlled trials and 3 prospective cohort studies were to infer a causal relationship, an association does ex- evaluated.26 Studies included a range of nonpharma- ist between cigarette smoking and risk of cancer re- cological (eg, counseling, self-help) and pharmacologi- currence, poor response to treatment, and increased cal (eg, nicotine-replacement therapy) interventions.26 rates of cancer treatment-related toxicity.1 Because Overall, findings indicated that tobacco cessation interracial and ethnic minorities are more likely to be di- ventions in the oncology setting were not significantly agnosed with late-stage tobacco-related cancers, it different from usual care, suggesting the need for ad- is possible that they could be particularly vulnerable ditional research to enhance smoking cessation among to the adverse effects of continued smoking. Park patients with cancer.26 Although a single smoking ces- et al22 reported that smoking cessation represents a sation trial demonstrated feasibility of recruiting African modifiable risk factor that could result in better treat- American patients, most smoking cessation trials con- ment outcomes in African Americans with lung cancer ducted in the oncology setting have had low represen- by improving rates of treatment efficacy and quality of tation of racially and ethnically diverse patients.26,27 To life. Given the strong associations between continued our knowledge, cessation interventions are not available smoking and cancer outcomes, organizations such as and have not been tested in Spanish-speaking patients. the American Society of Clinical Oncology and Ameri- Thus, as new interventions are developed and tested in can Association for Cancer Research have called for the oncology setting, the inclusion of racial and ethnic smoking cessation assistance to be provided to cancer minorities is imperative to extend the generalizability of survivors.75 The National Comprehensive Cancer Net- findings and reduce tobacco-related cancer disparities. work also established practice guidelines for smoking cessation that call for all patients with cancer to receive Conclusions evidence-based smoking cessation interventions.76 Racial and ethnic minorities experience shared social Implementation of the 5As model for a brief cessa- and environmental contextual factors that impact the tion intervention has been suboptimal in the oncology initiation and maintenance of smoking, and the poten- setting.8,24,25,77 Although the “ask” and “advise” steps tial for tobacco-related health disparities exist at each are often implemented, the remaining 3 steps are not point of the oncology care continuum. Thus, a compre- routinely completed.8,24 A review of members of the hensive strategy for reducing health disparities must American Society of Clinical Oncology reported that call for future research efforts at each of these points. 44% of its members surveyed said they discuss medi- Health disparities exist in the incidence cation options for smoking cessation, and 39% said and mortality of the 2 most strongly associated tobac- they provided smoking cessation treatment.24 These co-related cancers, namely lung and head and neck findings coincide with patient reports of low levels cancers.2,3,21,23,72-74 Future research is needed to elu- of cessation assistance provided by oncology special- cidate the causes of these disparities. As new cancer ists.77 Because racial and ethnic minorities who smoke screening tools become available, attention must be are generally less likely to receive assistance from their paid to prevent the widening of health disparities due physicians in quitting smoking, it is plausible that this to unequal implementation. assistance may also not occur in the oncology set- Overall rates of smoking cessation tend to be low- ting.18,19 Perceptions of inadequate training in tobacco er among racial and ethnic minorities.46,49 Culturally cessation interventions and low rates of self-efficacy adapted interventions are more acceptable44,46; how- in helping patients to quit have been cited barriers to ever, research has not yet supported their effectiveness cessation support by health care professionals.24 These when compared with standard materials.55,61 Future re- findings suggest a need for improved education and search efforts should be directed at increasing the use training of health care professionals to better convey of treatment and greater rates of efficacy in these vul- the rationale for smoking cessation and to develop nerable populations. Individuals with cancer must also strategies to implement smoking cessation into clinical be provided with cessation assistance and interventions care. In the absence of health care professional train- aimed at helping them quit smoking to improve cancer ing, and taking into consideration the constraint of time, outcomes.75,76 Additional education and training are an alternative ask–advise–refer model may be more fea- needed for oncologists so they can begin to understand sible to implement.25 This model involves asking about the impact of continued smoking for patients as well as tobacco use, advising the patient to quit, and then refer- to increase their comfort level for providing smoking ring the patient to existing, evidence-based resources cessation treatment to their patients who need it.25 Due (eg, national tobacco quit line [800-QUIT-NOW]). to the disproportionate disease burden experienced by

October 2016, Vol. 23, No. 4 Cancer Control 439 24. Warren GW, Marshall JR, Cummings KM, et al. Addressing tobacco African Americans and other underserved populations use in patients with cancer: a survey of American Society of Clinical Oncol- who smoke, future trials on smoking cessation inter- ogy members. J Oncol Pract. 2013;9(5):258-262. 25. Schroeder SA. What to do with a patient who smokes. JAMA. ventions must include diverse racial and ethnic patient 2005;294(4):482-487. populations and those cessation tools should be validat- 26. Nayan S, Gupta MK, Strychowsky JE, et al. Smoking cessation 3,26,27 interventions and cessation rates in the oncology population: an updated for use in all vulnerable patient populations. ed systematic review and meta-analysis. Otolaryngol Head Neck Surg. 2013;149(2):200-211. 27. Martinez E, Tatum KL, Weber DM, et al. Issues related to imple- References menting a smoking cessation clinical trial for cancer patients. Cancer 1. US Department of Health and Human Services. The Health Conse- Causes Control. 2009;20(1):97-104. quences of Smoking: 50 Years Of Progress. A Report of the Surgeon Gen- 28. Moolchan ET, Fagan P, Fernander AF, et al. Addressing tobacco- eral. Atlanta, GA: US Department of Health and Human Services; Centers related health disparities. Addiction. 2007;102(suppl 2):30-42. for Disease Control and Prevention; National Center for Chronic Disease 29. Aakko E, Schafer E, Gyarmathy VA, et al. Smoking policies Prevention and Health Promotion, Office on Smoking and Health; 2014. in manufacturing and assembly workplaces, Wisconsin, 1999. WMJ. 2. American Cancer Society. Cancer Facts & Figures 2016. Atlanta, GA: 2001;100(3):67-69. American Cancer Society. http://www.cancer.org/acs/groups/content/@ 30. Glasgow RE, Cummings KM, Hyland A; Community Intervention research/documents/document/acspc-047079.pdf. Accessed August 19, 2016. Trial for Smoking Cessation. Relationship of worksite smoking policy to 3. American Cancer Society. Cancer Facts & Figures for African Americans changes in employee tobacco use: findings from COMMIT. Tob Control. 2016-2018. Atlanta, GA: American Cancer Society. http://www.cancer.org/ 1997;6(suppl 2):S44-S48. acs/groups/content/@editorial/documents/document/acspc-047403.pdf. 31. Pizacani BA, Martin DP, Stark MJ, et al. A prospective study of Accessed August 19, 2016. household smoking bans and subsequent cessation related behaviour: 4. Fagan P, King G, Lawrence D, et al. Eliminating tobacco-related the role of stage of change. Tob Control. 2004;13(1):23-28. health disparities: directions for future research. Am J Public Health. 32. Kraev TA, Adamkiewicz G, Hammond SK, et al. Indoor concentrations 2004;94(2):211-217. of nicotine in low-income, multi-unit housing: associations with smoking 5. US Surgeon General. Tobacco use among U.S. racial/ethnic mi- behaviours and housing characteristics. Tob Control. 2009;18(6):438-444. nority groups--African Americans, American Indians and Alaska Natives, 33. Andrews JO, Mueller M, Newman SD, et al. The association of indi- Asian Americans and Pacific Islanders, Hispanics. Executive summary. vidual and neighborhood social cohesion, stressors, and crime on smok- MMWR Recomm Rep. 1998;47(RR-18):v-xv, 1-16. ing status among African-American women in southeastern US subsidized 6. Okuyemi KS, Reitzel LR, Fagan P. Interventions to reduce tobacco- housing neighborhoods. J Urban Health. 2014;91(6):1158-1174. related health disparities. Nicotine Tob Res. 2015;17(8):887-891. 34. Kendzor DE, Businelle MS, Reitzel LR, et al. The influence of dis- 7. Garrett BE, Dube SR, Winder C, et al. Cigarette smoking — United crimination on smoking cessation among Latinos. Drug Alcohol Depend. States, 2006-2008 and 2009-2010. MMWR Suppl. 2013;62(3):81-84. 2014;136:143-148. 8. Fiore MC, Jaén CR, Baker TB, et al. Clinical Practice Guideline: 35. Skinner ML, Shirtcliff EA, Haggerty KP, et al. Allostasis model Treating Tobacco Use and Dependence: 2008 Update. Rockville, MD: facilitates understanding race differences in the diurnal cortisol rhythm. US Department of Health and Human Services; Public Health Service; 2008. Dev Psychopathol. 2011;23(4):1167-1186. http://bphc.hrsa.gov/buckets/treatingtobacco.pdf. Accessed August 19, 2016. 36. Grady ES, Humfleet GL, Delucchi KL, et al. Smoking cessation 9. Garrett BE, Dube SR, Babb S, et al. Addressing the social deter- outcomes among sexual and gender minority and nonminority smokers in minants of health to reduce tobacco-related disparities. Nicotine Tob Res. extended smoking treatments. Nicotine Tob Res. 2014;16(9):1207-1215. 2015;17(8):892-897. 37. Berg CJ, Thomas JL, Guo H, et al. Predictors of smoking reduction 10. Honjo K, Tsutsumi A, Kawachi I, et al. What accounts for the rela- among Blacks. Nicotine Tob Res. 2010;12(4):423-431. tionship between social class and smoking cessation? Results of a path 38. Landrine H, Klonoff EA. Racial segregation and cigarette smok- analysis. Soc Sci Med. 2006;62(2):317-328. ing among blacks: findings at the individual level. Journal Health Psychol. 11. Muilenburg Legge J, Latham T, Annang L, et al. The home smok- 2000;5(2):211-219. ing environment: influence on behaviors and attitudes in a racially diverse 39. Brandon TH. Negative affect as motivation to smoke. Curr Dir Psychol adolescent population. Health Educ Behav. 2009;36(4):777-793. Sci. 1994;3(2):33-37. 12. Norman GJ, Ribisl KM, Howard-Pitney B, et al. The relationship 40. Kassel JD, Stroud LR, Paronis CA. Smoking, stress, and negative between home smoking bans and exposure to state tobacco control efforts affect: correlation, causation, and context across stages of smoking. Psychol and smoking behaviors. Am J Health Promot. 2000;15(2):81-88. Bull. 2003;129(2):270-304. 13. Mills AL, White MM, Pierce JP, et al. Home smoking bans among 41. Fiore MC, Jaén CR. A clinical blueprint to accelerate the elimination U.S. households with children and smokers. Opportunities for intervention. of tobacco use. JAMA. 2008;299(17):2083-2085. Am J Prev Med. 2011;41(6):559-565. 42. Wetter DW, Mazas C, Daza P, et al. Reaching and treating Spanish- 14. Lee JG, Henriksen L, Rose SW, et al. A systematic review of speaking smokers through the National Cancer Institute’s Cancer Information neighborhood disparities in point-of-sale tobacco marketing. Am J Public Service. A randomized controlled trial. Cancer. 2007;109(2 suppl):406-413. Health. 2015;105(9):e8-e18. 43. Blumenthal DS. Barriers to the provision of smoking cessation ser- 15. Webb Hooper M, Kolar SK. Distress, race/ethnicity and smoking ces- vices reported by clinicians in underserved communities. J Am Board Fam sation in treatment-seekers: implications for disparity elimination. Addiction. Med. 2007;20(3):272-279. 2015;110(9):1495-1504. 44. Cox LS, Okuyemi K, Choi WS, et al. A review of tobacco use 16. Purnell JQ, Peppone LJ, Alcaraz K, et al. Perceived discrimina- treatments in U.S. ethnic minority populations. Am J Health Promot. tion, psychological distress, and current smoking status: results from the 2011;25(5 suppl):S11-S30. Behavioral Risk Factor Surveillance System Reactions to Race module, 45. Jamal A, Homa DM, O’Connor E, et al. Current cigarette smoking 2004-2008. Am J Public Health. 2012;102(5):844-851. among adults - United States, 2005–2014. MMWR Morb Mortal Wkly Rep. 17. Kruger J, Shaw L, Kahende J, et al. Health care providers’ advice 2015;64(44):1233-1240 to quit smoking, National Health Interview Survey, 2000, 2005, and 2010. 46. Merzel CR, Isasi CR, Strizich G, et al. Smoking cessation among Prev Chronic Dis. 2012;9:E130. U.S. Hispanic/Latino adults: findings from the Hispanic Community Health 18. Lopez-Quintero C, Crum RM, Neumark YD. Racial/ethnic disparities Study/Study of Latinos (HCHS/SOL). Prev Med. 2015;81:412-419. in report of physician-provided smoking cessation advice: analysis of the 2000 47. Haiman CA, Stram DO, Wilkens LR, et al. Ethnic and racial dif- National Health Interview Survey. Am J Public Health. 2006;96(12):2235-2239. ferences in the smoking-related risk of lung cancer. N Engl J Med. 19. Sonnenfeld N, Schappert SM, Lin SX. Racial and ethnic differences 2006;354(4):333-342. in delivery of tobacco-cessation services. Am J Prev Med. 2009;36(1):21-28. 48. Trinidad DR, Gilpin EA, Lee L, et al. Do the majority of Asian- 20. Simmons VN, Quinn G, Litvin EB, et al. Transcreation of validated American and African-American smokers start as adults? Am J Prev Med. smoking relapse-prevention booklets for use with Hispanic populations. 2004;26(2):156-158. J Health Care Poor Underserved. 2011;22(3):886-893. 49. Irvin Vidrine J, Reitzel LR, Wetter DW. The role of tobacco in cancer 21. Howlader N, Noone AM, Krapcho M, et al, eds. SEER cancer health disparities. Curr Oncol Rep. 2009;11(6):475-481. statistics review, 1975-2012. http://seer.cancer.gov/csr/1975_2012. 50. Benowitz NL. Smoking cessation trials targeted to racial and eco- Based on November 2014 SEER data submission, posted April 2015. nomic minority groups. JAMA. 2002;288(4):497-499. Accessed August 3, 2016. 51. Benowitz NL, Perez-Stable EJ, Herrera B, et al. Slower metabolism 22. Park ER, Japuntich SJ, Traeger L, et al. Disparities between and reduced intake of nicotine from cigarette smoking in Chinese-Ameri- blacks and whites in tobacco and lung cancer treatment. Oncologist. cans. J Natl Cancer Inst. 2002;94(2):108-115. 2011;16(10):1428-1434. 52. Lawrence D, Graber JE, Mills SL, et al. Smoking cessation inter- 23. Molina MA, Cheung MC, Perez EA, et al. African American and ventions in U.S. racial/ethnic minority populations: an assessment of the poor patients have a dramatically worse prognosis for head and neck can- literature. Prev Med. 2003;36(2):204-216. cer: an examination of 20,915 patients. Cancer. 2008;113(10):2797-2806. 53. Robles GI, Singh-Franco D, Ghin HL. A review of the efficacy of

440 Cancer Control October 2016, Vol. 23, No. 4 smoking-cessation pharmacotherapies in nonwhite populations. Clin Ther. 2008;30(5):800-812. 54. Webb MS. Treating tobacco dependence among African Americans: a meta-analytic review. Health Psychol. 2008;27(3 suppl):S271-S282. 55. Fu SS, Burgess D, van Ryn M, et al. Views on smoking cessation methods in ethnic minority communities: a qualitative investigation. Prev Med. 2007;44(3):235-240. 56. Levinson AH, Perez-Stable EJ, Espinoza P, et al. Latinos report less use of pharmaceutical aids when trying to quit smoking. Am J Prev Med. 2004;26(2):105-111. 57. Webb MS, de Ybarra DR, Baker EA, et al. Cognitive-behavioral therapy to promote smoking cessation among African American smokers: a randomized clinical trial. J Consult Clin Psychol. 2010;78(1):24-33. 58. Webb MS. Culturally specific interventions for African American smokers: an efficacy experiment. J Natl Med Assoc. 2009;101(9):927-935. 59. Orleans CT, Boyd NR, Bingler R, et al. A self-help intervention for African American smokers: tailoring cancer information service counseling for a special population. Prev Med. 1998;27(5 pt 2):S61-S70. 60. Nevid JS, Javier RA. Preliminary investigation of a culturally specific smoking cessation intervention for Hispanic smokers. Am J Health Promot. 1997;11(3):198-207. 61. Rodriguez Esquivel D, Webb Hooper M, Baker EA, et al. Culturally specific versus standard smoking cessation messages targeting Hispan- ics: an experiment. Psychol Addict Behav. 2015;29(2):283-289. 62. Simmons VN, Cruz LM, Brandon TH, et al. Translation and adaptation of smoking relapse-prevention materials for pregnant and postpartum Hispanic women. J Health Commun. 2011;16(1):90-107. 63. Jamal A, Homa DM, O’Connor E, et al. Current cigarette smoking among adults - United States, 2005-2014. MMWR Morb Mortal Wkly Rep. 2015;64(44):1233-1240. 64. Alexander LA, Trinidad DR, Sakuma KL, et al. Why we must continue to investigate menthol’s role in the African American smoking paradox. Nicotine Ton Res. 2016;18(suppl 1):S91-S101. 65. Tobacco Products Scientific Advisory Committee. Menthol Cigarettes and Public Health: Review of the Scientific Evidence and Recommenda- tions. Bethesda, MD: US Food and Drug Administration; 2011. http://www. fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/ TobaccoProductsScientificAdvisoryCommittee/UCM269697.pdf. Accessed August 3, 2016. 66. Lathan CS, Okechukwu C, Drake BF, et al. Racial differences in the perception of lung cancer: the 2005 Health Information National Trends Survey. Cancer. 2010;116(8):1981-1986. 67. Park ER, Ostroff JS, Rakowski W, et al. Risk perceptions among participants undergoing lung cancer screening: baseline results from the National Lung Screening Trial. Ann Behav Med. 2009;37(3):268-279. 68. George M, Margolis ML. Race and lung cancer surgery--a qualitative analysis of relevant beliefs and management preferences. Oncol Nurs Forum. 2010;37(6):740-748. 69. Jonnalagadda S, Bergamo C, Lin JJ, et al. Beliefs and attitudes about lung cancer screening among smokers. Lung Cancer. 2012;77(3):526-531. 70. Aberle DR, Adams AM, Berg CD, et al. Reduced lung-cancer mortality with low-dose computed tomographic screening. N Engl J Med. 2011;365(5):395-409. 71. Holford TR, Levy DT, Meza R. Comparison of smoking history patterns among African American and white cohorts in the United States born 1890 to 1990. Nicotine Tob Res. 2016;18(suppl 1):S16-S29. 72. Blot WJ, McLaughlin JK, Winn DM, et al. Smoking and drinking in relation to oral and pharyngeal cancer. Cancer Res. 1988;48(11):3282-3287. 73. Conway DI, Petticrew M, Marlborough, et al. Socioeconomic inequalities and oral cancer risk: a systematic review and meta-analysis of case-control studies. Int J Cancer. 2008;122(12):2811-2819. 74. Fagan P, Moolchan ET, Lawrence D, et al. Identifying health disparities across the tobacco continuum. Addiction. 2007;102(suppl 2):5-29. 75. Toll BA, Brandon TH, Gritz ER, et al. Assessing tobacco use by cancer patients and facilitating cessation: an American Association for Cancer Research policy statement. Clin Cancer Res. 2013;19(8):1941-1948. 76. National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology: Smoking Cessation. v.2.2015. https://www.nccn.org/ professionals/physician_gls/pdf/smoking.pdf. Accessed August 4, 2016. 77. Simmons VN, Litvin EB, Unrod M, et al. Oncology healthcare providers’ implementation of the 5A’s model of brief intervention for smoking cessation: patients’ perceptions. Patient Educ Couns. 2012;86(3):414-419.

October 2016, Vol. 23, No. 4 Cancer Control 441 Case Series

Human Metapneumovirus Infection in Immunocompromised Patients Sharmeen Samuel, MD, Sowmya Nanjappa, MBBS, MD, Christopher D. Cooper, MD, and John N. Greene, MD

Summary: Human metapneumovirus (HMPV) is a pathogen associated with respiratory tract infection and is related to avian pneumovirus. Typically, children, the elderly, and those who are immunocompromised are the most susceptible to HMPV infection; however, the virus can infect persons of all ages. In otherwise healthy individuals, HMPV infection is generally self-limiting, but immunocompromised individuals can develop fatal complications. We present a case series of 3 severely immunocompromised patients who were infected with HMPV and describe their clinical course. All 3 patients had acute myeloid leukemia, histories of neutropenic fever, and prolonged hospitalization stays. This case series highlights the severe sequelae observed in individuals infected with HMPV, particularly among those who are immunocompromised.

Introduction factor therapy for AML and received multiple doses of In 2001, van den Hoogen et al1 identified human intrathecal methotrexate. metapneumovirus (HMPV) in children with respira- Three months later she was readmitted to the hospi- tory tract infection, because they were able to obtain tal for blast crisis and reinduction chemotherapy. During a genomic sequence of the pathogen via randomly that admission, she developed bacteremia with vanco- primed polymerase chain reaction (PCR). The virus was mycin-resistant enterococcus and Streptococcus mitis. found to be related to avian pneumovirus, a member She was subsequently treated with daptomycin, liposo- of the Metapneumovirus genus.2 HMPV is an emerging mal amphotericin B, cefepime, and acyclovir. pathogen associated with both upper and lower respi- One month later she developed intermittent hemop- ratory tract infections in all age groups, but it is more tysis and mild dyspnea on ambulation that progressed to common in children (age < 5 years), in the elderly hypoxia (oxygen saturation = 90% on room air). She also (age > 60 years), and in those with compromised im- experienced confusion, fever (temperature = 105.5 °F), mune systems.3,4 Although its formal discovery occurred bilateral scattered rhonchi, and crackles on auscultation. 15 years ago, HMPV may have been responsible for Noncontrast computed tomography (CT) of the thorax respiratory tract infections for at least 60 years. 1 was obtained and revealed new focal consolidative infiltrates, including increased peripheral, dense consolida- Case Reports tion of the right middle, left upper, and left lower lobes, This series presents the cases of 3 immunocompro- widespread ill-defined nodules, and ground-glass infil- mised patients treated at the H. Lee Moffitt Cancer trates due to alveolar hemorrhage (Fig). Center & Research Institute (Tampa, FL) who were in- Bronchoalveolar fluid cultures obtained by bron- fected with HMPV. A summary of the case series is choscopy were negative for bacteria and fungi. However, found in the Table. HMPV was detected by PCR from nasopharynx swab sampling. Findings on PCR testing of other viral agents Case 1 (adenovirus, influenza, parainfluenza, and rhinoviruses) A Hispanic woman 47 years of age with acute myeloid were negative. leukemia (AML) had a clinical course complicated by The patient was treated with oral ribavirin (RBV) central nervous system malignancy, bacterial enteritis, 600 mg 3 times a day. After 14 days of therapy, her con- and diffuse alveolar and left-sided retinal hemorrhag- dition clinically improved and oxygen therapy was es. She was taking cladribine, cytarabine, and growth- discontinued. Repeat CT showed improved multifocal pneumonia compared with the findings on the initial From the Departments of Internal Medicine (SS, SN) and Infectious imaging study. Diseases (JNG), H. Lee Moffitt Cancer Center & Research Institute, Tampa, Florida, the Department of Infectious Diseases (CDC), Florida Hospital, Altamonte Springs, Florida. Case 2 Submitted February 17, 2016; accepted May 31, 2016. An African American man 61 years of age with hepati- Address correspondence to John N. Greene, MD, Department of tis B virus infection and non-Hodgkin lymphoma pre- Infectious Diseases, Moffitt Cancer Center, 12902 Magnolia Drive, sented with fever and chills. Laboratory work-up stud- FOB-3 BMT PROG, Tampa, FL 33612. E-mail: [email protected] ies were ordered and the results revealed pancytopenia. No significant relationships exist between the authors and the companies/organizations whose products or services may be referenced Findings on bone marrow biopsy were conclusive of in this article. AML. He was treated with cladribine, cytarabine, and

442 Cancer Control October 2016, Vol. 23, No. 4 Table. — Summary of the Case Series Case No. Underlying Etiology of Immunocompromised Clinical Signs and Treatment Outcome Malignancy State Symptoms Dyspnea AML Oral RBV 600 mg 1 Chemotherapy Fever Survived CNS TID Hemoptysis Cough AML Oral RBV 600 mg 2 Chemotherapy Dyspnea Died NHL TID Fever Cough Oral RBV 600 mg 3 AML Chemotherapy Dyspnea Died TID Fever

AML = acute myeloid leukemia, CNS = central nervous system, NHL = non-Hodgkin lymphoma, RBV = ribavirin, TID = 3 times a day. growth factors and eventually achieved remission. Several months later, bone marrow biopsy was repeated, the results of which showed that the AML had relapsed; thus, cycle 1 of azacitidine therapy was initiated. The treatment course was complicated by pneumonia (presumed to be fungal in origin based on findings seen on CT of his chest). The patient was given daily liposomal amphotericin B to treat pneumonia. On his hospital readmission for the second cycle of azacitidine, the patient developed submandibular cellu- litis, transaminitis, and became infected with type 1 her- pes simplex virus and respiratory syncytial virus. He was treated with intravenous meropenem every day for 6 weeks, weekly liposomal amphotericin B, acyclovir, and lamivudine. However, his clinical course was Fig. — Computed tomography of the chest from the patient in case 1. complicated by neutropenic fever and worsening dys- Findings show focal consolidative infiltrates and dense peripheral con- pnea. Blood cultures grew Enterococcus faecium and solidation in the left upper and lower lobes. multi–drug-resistant S maltophilia. Samples on bron- choscopy were positive for S maltophilia and Candida choalveolar fluid cultures were negative for bacteria and glabrata. He was treated with intravenous daptomycin fungi. However, findings on respiratory viral panel from and minocycline. Blood cultures were repeated and nasopharynx sampling were positive for HMPV. the results were negative after 6 days. The patient was treated with oral RBV 600 mg However, the patient again developed fever (tempera- 3 times a day. Repeat CT of his chest showed extensive ture = 102.5 °F), dry cough, and hypoxia (oxygen satura- bilateral ground-glass nodules consistent with bronchi- tion = 92% on room air). Respiratory viral panel was re- olitis obliterans organizing pneumonia. peated and HMPV was detected in his nasopharynx swab His clinical presentation initially improved after oral specimen. He was treated with oral RBV 600 mg 3 times RBV therapy, but he presented again with high-grade a day. The patient died 4 days after the start of therapy. fever, tachycardia, and hypotension. He was started on meropenem, tobramycin, voriconazole, and acyclovir. Case 3 Blood cultures were obtained and grew Escherichia A Hispanic man 54 years of age with relapsed AML coli. Despite aggressive treatment with antibiotics, the was admitted to the hospital for induction chemother- patient died the following day from septic shock. apy and was treated with cladribine, cytarabine, and growth-factor therapy. One week later while on neu- Discussion tropenic prophylaxis (ciprofloxacin, voriconazole, and HMPV is an enveloped virus with a nonsegmented, acyclovir), the patient developed cough, mild dyspnea negative-sense RNA genome.1 Genetic analysis has elu- (oxygen saturation = 94% on room air), and fever (tem- cidated 2 subgroups of the virus that often circulate to- perature = 103.0 °F). Although findings on CT of the gether in a population.5 Although the virus is related chest revealed bilateral patchy and nodular opacities to avian pneumovirus, HMPV does not cause illness consistent with an infectious process, sputum and bron- in birds.4,6 Study results have shown that the inocula-

October 2016, Vol. 23, No. 4 Cancer Control 443 tion of HMPV in other nonhuman primates can lead to Respiratory tract infections are a significant source significant viral replication restricted to the respiratory of morbidity and mortality in immunocompromised tract with the subsequent development of upper respi- patients, and this is especially true for individuals with ratory tract symptoms.7 HMPV is commonly associated hematological malignancies and in those receiving with respiratory tract infection, causing flulike symp- hematopoietic stem cell transplantation (HSCT).18 Pro- toms, bronchiolitis, and pneumonia. HMPV infection spective and retrospective studies of patients receiving may also be associated with exacerbations of asthma.8 HSCT have provided data on the incidence of upper The data available are limited on the association of and lower respiratory tract infections caused by HMPV HMPV with infections of other organ systems beyond infection and the associated mortality rates.19,20 In these the respiratory tract. In a case report concerning a pre- studies, cohorts of symptomatic and asymptomatic viously healthy child 14 months of age who died of en- patients with HSCT were followed for 4 years, and the cephalitis, HMPV RNA was detected in both brain and virus was identified through PCR testing of nasopha- lung tissue, although HMPV antigens went undetected ryngeal specimens.19,20 The incidence rates of HMPV in these tissues via immunostaining.9 infections in patients receiving HSCT ranged from 3% Several studies have demonstrated the presence to 7% in those with any respiratory disease, whereas of HMPV and absence of other pathogenic organisms the rate increased to between 27% and 41% among among patients with respiratory disease.8,10 Williams patients with lower respiratory disease.19,20 The rate of et al11 examined the nasal-wash specimens from chil- deaths associated with lower respiratory tract infection dren with lower respiratory tract infections during a was between 33% and 40%, and the rate of deaths re- 25-year period. A viral cause other than HMPV was lated to overall infections (upper and lower respiratory identified in 41% of health care visits among 2,009 oth- tracts) was between 0% and 14%.19 In the retrospective erwise healthy infants and children presenting with study, bronchoalveolar lavage specimens were ob- acute symptoms.11 No viral or bacterial cause was iden- tained for the work-up of pulmonary infiltrates from tified in 59% of visits; of the available specimens for 163 patients after receiving HSCT; of these, 3% were which no cause was previously identified, 20% tested positive for HMPV and 20% of patients died due to re- positive for HMPV.11 This finding extrapolates to 12% of spiratory failure.20 These results highlight the potential lower respiratory tract infections in their patient popu- severity of HMPV-related pneumonia in recipients of lation. Thus, Williams et al11 concluded that HMPV in- HSCT. Because many such patients have concurrent fection is an important cause of respiratory tract illness, bacterial and fungal infections, describing the exact particularly during the first few years of life. clinical features of HMPV infection can be difficult. In general, transmission of HMPV is through di- HMPV infection has important associations with rect or close contact via large-particle aerosols, drop- respiratory tract illness in patients with hemato- lets, and fomites.12 Many nosocomial infections have logical malignancies and in those who have received been reported in children and adults.13 Case 2 most lung transplants.21,22 A retrospective evaluation of likely represents a nosocomial infection, because the 55 children (age range, 5 months–19 years) with labora- patient in this case was hospitalized for a prolonged tory-confirmed HMPV infection highlighted a number duration prior to developing symptoms. HMPV infec- of associations. HMPV infection was associated with tion also causes a variable array of clinical symptoms, underlying hematological malignancy (44%), HSCT including self-limiting cough, low-grade fever, rhinor- (15%), solid tumors (16%), and solid organ transplants rhea, severe respiratory failure, and death.14 The incu- (15%). 21 Another report showed that 25% of study pa- bation period has not yet been defined but is thought tients with hematological disease who tested positive to be 5 to 6 days in most cases.15 Little is known about for HMPV infection also had pneumonia.22 A review the pathogenesis of HMPV, but an integrin alpha v beta by Kamboj et al23 of 1,899 study patients with cancer 1 receptor has been identified that facilitates infection showed that nearly 3% were positive for HMPV; more of the epithelial cells of the respiratory tract.16 children were positive than adults. The most common reasons for hospitalization in To diagnose HMPV infection, reverse transcrip- children are bronchiolitis and pneumonia. Although tase (RT)-PCR is the most sensitive and commonly used HMPV has been associated with upper respiratory method available; it detects the virus in nasal and bron- tract infections (rhinopharyngitis, laryngitis, croup), chial secretions and is easy to perform.24 Other less-com- severe lower respiratory tract infection has been re- monly used methods for detecting the virus include di- ported.8 The infection can occur in adults of all ages, rect fluorescent antibody test, serology, and viral culture. although elderly persons with histories of long hos- The incidence of HMPV infection increases during pitalization stays and cardiopulmonary disease are at the winter and spring.25 Patients in cases 1 and 2 in our greater risk for becoming infected.5 The possible role series were infected with HMPV in June 2015, and the of HMPV in the exacerbation of chronic obstructive patient in case 3 was infected in May 2014. Prior infec- pulmonary disease in adults is controversial.17 tion with HMPV does not protect against reinfection,

444 Cancer Control October 2016, Vol. 23, No. 4 4 tract disease. Nat Med. 2001;7(6):719-724. which can occur throughout life. 2. Fauquet CM, Mayo MA, Maniloff J, et al, eds; Virology Division Inter- A live recombinant vaccine for HMPV is still in pre- national Union of Microbiological Societies. Virus Taxonomy: Classification 26 and Nomenclature of Viruses. Eighth Report of the International Committee clinical trials. However, we must learn more about its on Taxonomy of Viruses. Oxford, UK: Elsevier Academic Press; 2005. pathogenesis, genetics, and serological properties to 3. Kahn JS. Epidemiology of human metapneumovirus. Clin Microbiol Rev. 2006;19(3):546-557. achieve a safe and effective vaccine against HMPV. 4. Falsey AR, Walsh EE. Viral pneumonia in older adults. Clin Infect Treatment for HMPV infection is mainly support- Dis. 2006;42(4): 518-524. 5. Boivin G, Mackay I, SlootsTP, et al. Global genetic diversity of human ive, and no data exist on any antiviral agent shown to metapneumovirus fusion gene. Emerg Infect Dis. 2004;10(6):1154-1157. have significant activity against the virus. No guide- 6. Njenga MK, Lwamba HM, Seal BS. Metapneumovirus in birds and humans. Virus Res. 2003;91(2):163-169. lines are available for treatment, but some centers use 7. Boivin G, DeSerres G, Côté S, et al. Human metapneumovirus in- RBV and intravenous immunoglobulin; however, the fections in hospitalized children. Emerg Infect Dis. 2003;9(6):634-640. 27 8. Mullins JA, Erdman, DD, Weinberg GA, et al. Human metapneumo- usefulness of RBV in HMPV infection is controversial. virus infection among children hospitalized with acute respiratory illness. Some animal models have shown that RBV reduces Emerg Infect Dis. 2004;10(4):700-705. 28 9. Schildgen O, Glatzel T, Geikowski T, et al. Human metapneumovi- viral replication in HMPV-infected mice, and use of rus RNA in encephalitis patient. Emerg Infect Dis. 2005;11(3):467-470. RBV to successfully treat the infection in humans has 10. Williams JV, Wang CK, Yang CF et al. The role of hMPV in upper 29 respiratory tract infections in children: a 20 year experience. J Infect Dis. been previously described. Other study data show 2006;193(3):387-395. that RBV does not reduce mortality in immunocom- 11. Williams JV, Harris PA, Tollefson SJ, et al. Human metapneumovirus and lower respiratory tract disease in otherwise healthy infants and promised patients and is not effective against the vi- children. N Engl J Med. 2004;350(5):443-450. rus.30 Although case 1 in our series is an example of 12. Williams JV. Human metapneumovirus: an important cause of respiratory disease in children and adults. Curr Infect Dis Rep. 2005;7(3):204-210. successful treatment against HMPV using oral RBV, 13. van Dijk NM, Linssen, CF, Bussink M, et al. Critical illness after hori- this finding does not establish a beneficial effect of zontal nosocomial transmission of human metapneumovirus in the haemo- tology ward. Neth J Crit Care. 2012;16(1):22-25. http://njcc.nl/sites/default/ RBV against HMPV infection. files/NJCC%2001%20casereport-vDijk.pdf. Accessed August 24, 2016. 14. Xepapadaki P, Psarras S, Bossios A, et al. Human metapneumovirus as a causative agent of acute bronchiolitis in infants. J Clin Virol. Conclusions 2004;30(3):267-270. Discovered 15 years ago, human metapneumovirus 15. Centers for Disease Control and Prevention. Outbreaks of human 1 metapneumovirus in two skilled nursing facilities-West Virginia and Idaho, (HMPV) can cause respiratory tract infections. Al- 2011-2012. MMWR Morb Mortal Wkly Rep. 2013;62(46):909-913. though data have been published on the prevalence of 16. Cseke G, Maginnis MS, Cox RG, et al. Integrin alphabeta1 pro- motes infection by human metapneumovirus. Proc Natl Acad Sci USA. the virus in high-risk groups (eg, children < 5 years of 20093;106(5):1566-1571. age, persons > 60 years of age with comorbidities [eg, 17. Hamelin ME, Côté S, Laforge J, et al. Human metapneumovirus infection in adults with community acquired pneumonia and exacerbation of cardiopulmonary diseases], immunocompromised in- chronic obstructive pulmonary disease. Clin Infect Dis. 2005;41(4):498-502. dividuals), important information on the immunologi- 18. Renaud C, Campbell AP. Changing epidemiology of respiratory viral infections in hematopoietic cell transplant recipients and solid organ cal properties and pathogenesis of the virus is still un- transplant recipients. Curr Opin Infect Dis. 2011;24(4):333-343. known.5 This case series presents the clinical courses 19. Milano F, Campbell AP, Guthrie KA, et al. Human rhinovirus and coronavirus detection among allogeneic hematopoietic stem cell trans- of 3 severely immunocompromised patients infected plantation recipients. Blood. 2010;115(10):2088-2094. with HMPV and describes the impact of the virus on 20. Englund JA, Boeckh M, Kuypers J, et al. Brief communication: fatal human metapneumovirus infection in stem cell transplant recipients. Ann their disease course. Infection with HMPV can add to Intern Med. 2006;144(5): 344-349. the severity of an underlying malignancy, especially in 21. Chu HY, Renaud C, Ficken E, et al. Respiratory tract infections due to human metapneumovirus in immunocompromised children. J Pediatric immunocompromised patients, and can be a contribut- Infect Dis Soc. 2014;3(4):286-293. ing factor toward the death of such patients. Therefore, 22. Godet C, Le Goff J, Beby-Defaux A, et al. Human metapneumovirus pneumonia in patients with hematological malignancies. J Clin Virol. it is important to prevent this viral infection in immu- 2014;61(4):593-596. nocompromised individuals. Measures to control such 23. Kamboj M, Gerbin M, Huang CK, et al. Clinical characterization of human metapneumovirus infection among patients with cancer. J Infect. infection, including avoiding contact with large-particle 2008;57(6):464-471. aerosols and droplets, practicing proper hand hygiene, 24. Côté S, Abed Y, Boivin G. Comparative evaluation of real-time PCR assays for detection of human metapneumovirus. J Clin Microbiol. and, when possible, limiting the visitation of persons 2003;41(8):3631-3635. who are ill, can all help to prevent the spread of HMPV. 25. Caracciolo S, Minni C, Colombrita D, et al. Human metapneumovirus infection in young children hospitalized with acute respiratory tract disease: A need exists for recommendations to guide treatment virologic and clinical features. Pediatr Infect Dis J. 2008;27(5):406-412. for HMPV infection in high-risk patient groups, such as 26. Buchholz UJ, Nagashima K, Murphy BR, et al. Live vaccines for human metapneumovirus designed by reverse genetics. Expert Rev Vaccines. the severely immunocompromised patients presented 2006;5(5):695-706. in this case series. Development of an HMPV vaccine is 27. Hirsch HH, Martino R, Ward KN, et al. Fourth European Conference on Infections in Leukaemia (ECIL-4): guidelines for diagnosis and treat- a reasonable goal for the prevention of this pathogen. ment of human respiratory syncytial virus, parainfluenza virus, metapneu- Acknowledgment: We gratefully acknowledge the movirus, rhinovirus, and coronavirus. Clin Infect Dis. 2013;56(2):258-266. 28. Hamelin ME, Prince GA, Boivin G. Effect of ribavirin and glucocor- help of Peter P. Karpawich, MD, from the Detroit Medi- ticoid treatment in a mouse model of human metapneumovirus infection. cal Center in Michigan for reviewing this case series. Antimicrob Agents Chemother. 2006;50(2):774-777. 29. Kitanovski L, Kopriva S, Pokorn M, et al. Treatment of severe human metapneumovirus pneumonia in an immunocompromised child with References oral ribavirin and IVIG. J Pediatr Hematol Oncol. 2013;35(7):e311-e313. 30. Debur MC, Vidal LR, Stroparo E, et al. Human metapneumovirus 1. van den Hoogen BG, de Jong JC, Groen J, et al. A newly discov- infection in hematopoietic stem cell transplant recipients. Transpl Infect ered human pneumovirus isolated from young children with respiratory Dis. 2010;12(2):173-179.

October 2016, Vol. 23, No. 4 Cancer Control 445 Special Report

Systematic Review and Case Series Report of Acinar Cell Carcinoma of the Pancreas Evan S. Glazer, MD, PhD, Kevin G. Neill, MD, Jessica M. Frakes, MD, Domenico Coppola, MD, Pamela J. Hodul, MD, Sara E. Hoffe, MD, Jose M. Pimiento, MD, Gregory M. Springett, MD, PhD, and Mokenge P. Malafa, MD, PhD

Background: Acinar cell carcinoma of the pancreas is a rare malignancy representing less than 1% of all pancreatic malignancies. Methods: We report on a case series of 21 patients with acinar cell carcinoma of the pancreas treated at a high-volume quaternary center. A systematic review of the medical literature was performed that described typical therapeutic management approaches for acinar cell carcinoma of the pancreas and reported on disease control and survival rates. Data for the case series were obtained from a prospective database. Results: In our systematic review of 6 articles, study patients had a median age of 61 years, 66% were male, 52% had stage I/II disease, and 55% of lesions were located in the pancreatic head. The rates of median survival were approximately 47 months after resection with adjuvant therapy, 38 months for nonmetastatic, locally unresectable disease, and 17 months for metastatic disease treated with chemotherapy. Combination fluoropyrimidine-based chemotherapy regimens had better rates of disease control than other therapies. Our case series included 21 study patients, 14 of whom required resection and 7 who had metastatic disease. The rates of median survival were 40.2 ± 31.9 months in those who underwent surgery and were treated with adjuvant therapy and 13.8 ± 11.3 months for patients with metastatic disease. Conclusions: Multidisciplinary treatment for acinar cell carcinoma of the pancreas should be considered due to the rarity of the disease and its lack of high-level therapeutic data. Progress in the molecular analysis of this tumor may improve outcomes through the use of personalized therapy based on underlying tumor mutations.

Introduction tive that it is a distinct entity. Acinar cell carcinoma of the pancreas is a rare exocrine The rarity of acinar cell carcinoma of the pancre- malignancy representing less than 1% of all pancreatic as, lack of high-quality evidence, and its diagnostic neoplasms.1 The exocrine pancreas is primarily com- difficulty have all limited consensus among authors posed of ductal cells and acinar cells. Acinar cells are regarding the most appropriate management of this enzyme-secreting cells that release amylase, lipases, disease.5-7 Although patients who have undergone re- and proteases. Debate exists regarding whether acinar section for acinar cell carcinoma of the pancreas may cell carcinoma of the pancreas is unique from pancre- appear to have higher survival rates than patients with atic ductal adenocarcinoma2,3; however, acinar cell car- pancreatic ductal adenocarcinoma3,4 — whose median cinoma of the pancreas responds to treatment differ- survival rate is between 4 and 5 years1 — the limited ently than pancreatic ductal adenocarcinoma and its and selection-biased data make assertions difficult. rates of overall survival are not similar4 — data sugges- Nonsurgical treatment options for patients with acinar cell carcinoma of the pancreas include chemotherapy and radiotherapy. Use of various chemothera- From the Departments of Gastrointestinal Oncology (ESG, PJH, peutic agents, including capecitabine monotherapy JMP, MPM), Pathology (KGN, DC, GMS), and Radiation Oncology (JMF, SEH), H. Lee Moffitt Cancer Center & Research Institute, and folinic acid/fluorouracil/oxaliplatin (FOLFOX), Tampa, Florida. have been reported in the literature.8,9 Whether radio- Submitted February 18, 2016; accepted May 31, 2016. therapy improves survival rates in patients with acinar Address correspondence to Mokenge P. Malafa, MD, PhD, cell carcinoma of the pancreas remains unclear.10-12 Department of Gastrointestinal Oncology, Moffitt Cancer Center, Case reports of the use of radiotherapy to treat acinar 12902 Magnolia Drive, FOB-2 GI PROG, Tampa, FL 33612. E-mail: [email protected] cell carcinoma of the pancreas have noted an associa- Dr Glazer is now affiliated with the Division of Surgical Oncology tion with extended rates of survival.8,13 in the Department of Surgery at the University of Tennessee Health The patterns of genetic alterations and specific Sciences Center, Memphis, Tennessee. driver mutations in patients with acinar cell carci- No significant relationships exist between the authors and the companies/organizations whose products or services may be refer- noma of the pancreas are not well described in the enced in this article. literature. Studies have demonstrated lower rates of

446 Cancer Control October 2016, Vol. 23, No. 4 EGFR and KRAS mutations in acinar cell carcinoma of ies were identified as having high-level evidence. the pancreas than in pancreatic ductal adenocarcino- A total of 132 manuscripts were identified dur- ma.14 Some have suggested that pancreatic ductal ad- ing the search, and 10 studies were included based enocarcinoma may develop from acinar cells that un- on our inclusion/exclusion criteria.20 Six studies were dermutate p53 without evidence of developing acinar then identified and selected for summary analysis of cell carcinoma of the pancreas.15 In addition, nearly response rates to chemotherapy and radiotherapy as one-half of acinar cell carcinoma of the pancreas tu- well as to estimate the average median overall survival mors have alterations in APC.16 rate.3,21-25 Four studies contained incomplete data or Unlike pancreatic ductal adenocarcinoma, KRAS mu- nonstandard treatment approaches and were excluded tation is infrequently a driving force in acinar cell carcino- from the summary data analyses but were included in ma of the pancreas; however, chromosomal imbalances the systematic review. are common in acinar cell carcinoma of the pancreas.14 Whole-exome sequencing has revealed that mutations in Case Series BRCA2 and the FAT family are common in patients with We reviewed our records for patients with acinar cell acinar cell carcinoma of the pancreas, but these muta- carcinoma of the pancreas treated between 2004 and tions are not yet readily targetable.17 p53 expression has 2014 at the H. Lee Moffitt Cancer Center & Research been demonstrated as both a positive and negative prog- Institute (Tampa, FL). Demographical, clinical, and nostic indicator in acinar cell carcinoma of the pancreas.18 pathological data were obtained from a retrospective More than 70% of tumors of the acinar cell carcinoma database. Charts were reviewed to confirm data. The of the pancreatic type demonstrated a decrease or com- Institutional Review Board at our institution approved plete loss of DCC expression on immunohistochemis- this research. try.14 Bergmann et al14 also suggested that DCC variation is an early genetic change in acinar cell carcinoma of the Statistical Analyses pancreas, whereas others have suggested that it is a late For the systematic review, summary clinical data from change in pancreatic ductal adenocarcinoma.19 each study were averaged in a weighted fashion to determine the overall value for each characteristic Purpose (demographic, treatment, or survival data). STATA sta- The purpose of this systematic review is to delineate tistical software was utilized for analysis (StataCorp, the most appropriate treatment and expected out- College Station, Texas). comes in patients diagnosed with acinar cell carcino- For the case series, data analyses were also per- ma of the pancreas. Our article focuses on adult pa- formed with STATA (StataCorp), using χ2, Kaplan- tients because genetic mutations in childhood tumors Meier, and multivariate linear-regression techniques. suggest alternative etiologies. Statistical significance was set toα being equal to .05. Uncertainties are standard deviations of the mean un- Methods less otherwise specified. Literature and Systematic Review We searched the medical literature for studies describ- Results ing typical therapeutic management approaches for A total of 68 patients identified from the systematic re- acinar cell carcinoma of the pancreas and reported view had a median age of 61 years, and 66% were men on disease control and survival rates. Studies were (Table 1). These results are similar to other nonsystem- excluded during the initial search if they investigated fewer than 4 patients, focused on mixed carcinoma, were not written in English, did not include clinical Table 1. — Characteristics of Patients With Acinar Cell Carcinoma of the Pancreas in Select Studies data, were review articles, or were a technical manuscript on a surgical procedure alone. Case reports with Study No. of Median Male, Stage Pancreatic fewer than 4 patients were also excluded due to lack Study Age, y % I/II, Head, % Patients % of ability (too few patients to develop a systematic ap- 25 proach given the rarity of the malignancy). For studies Butturini 9 53 67 57 67 that included previously published data, the most com- Holen21 39 60 77 46 44 plete and most recent data were used for our system- Lowery24 40 65 73 55 55 atic analysis. When studies listed patient-level data for Matos3 17 60 Unknown 59 76 multiple histologies, only data concerning acinar cell Seki22 4 63 75 25 62 carcinoma of the pancreas were used. Rates of disease Seth23 14 57 33 86 64 control were calculated using complete response in Current 21 64 76 33 66 combination with partial response and stable disease series (based on the definition used in each study). No stud-

October 2016, Vol. 23, No. 4 Cancer Control 447 atic analyses in which the male-to-female patient ra- the acinar pattern and, typically, lack of staining for tio is at least 2:1.3,18,21-26 In our systematic review, 52% neuroendocrine neoplasm markers (neural cell ad- of patients with acinar cell carcinoma of the pan- hesion molecule, chromogranin, synaptophysin) can creas had resectable disease (stage I/II; see Table confirm the diagnosis of acinar cell carcinoma of the 1).3,21-25 Nonspecific symptoms, such as nausea, abdom- pancreas.14 inal pain, and weight loss, were observed in most patients.18 We also found that 55% of patients with acinar Radiotherapy cell carcinoma of the pancreas had tumors located in Our systematic review revealed a modest response the pancreatic head. The data derived from the system- rate to radiotherapy in patients with localized aci- atic review also demonstrated that most of the lesions nar cell carcinoma of the pancreas. In several stud- described were generally well circumscribed without ies, a “major response” rate was seen in 25% to 35% significant peritumoral stroma.3,25 of these patients.21,23 However, if stable disease was included in the definition of response (ie, Diagnosis disease control rate), significantly higher response The diagnosis of acinar cell carcinoma of the pancre- rates were seen in highly selected studies when ra- as is confirmed by biopsy, such as fine needle aspira- diotherapy was added (Table 2).3,21-25 Radiotherapy is tion, or resection of the pancreatic tumor — with the often used to “down-stage” or convert a tumor from latter being more reliable than the other methods.27 being borderline resectable to resectable, and, in gen- The highly cellular nodules of monotonous tumor eral, is provided with both conventional fractionation cells have little to no stroma and lack a desmoplastic and hypofractionated stereotactic body radiotherapy response (Fig 1). Therefore, tumor cells may be dif- (SBRT).25 Lack of progression following neoadjuvant ficult to identify with fine needle aspiration alone but radiotherapy can help select patients who may benefit reliably seen on a resected specimen.1 Tumor cells from resection. are arranged in solid sheets with spaces between the acinar structures and small lumens.18 Cytologi- Chemotherapy cally, these cells exhibit basal polarization. Fig 1 also Significant heterogeneity was seen in the systematic demonstrates that nuclei can range from moderate to review regarding specific chemotherapeutic proto- marked atypia and contain a single, prominent nucle- cols; however, most study patients received com- olus. In the apical region, the cells contain moderate bination gemcitabine- or fluoropyrimidine-based to abundant eosinophilic cytoplasm rich in zymogen chemotherapy protocols.23,25 Indeed, Holen et al21 granules. and Lowery et al24 concurred on the favored use of Other groups have utilized immunohistochem- combination fluoropyrimidine-based therapies and istry to confirm the diagnosis, and researchers have radiotherapy in patients with acinar cell carcinoma demonstrated that an antibody against a portion of of the pancreas. However, data regarding disease the B-cell chronic lymphocytic leukemia/lymphoma control rate (complete response + partial response + 10 protein has utility in diagnosing acinar cell carci- stable disease) are mixed (see Table 2).3,21-25,28 Com- noma of the pancreas.18 In addition, immunohisto- bination fluoropyrimidine-based chemotherapies chemistry staining for trypsin and chymotrypsin in were shown to be associated with higher rates of

A B

Fig 1A–B. — (A) Acinar cell carcinoma of the pancreas in a man aged 83 y and (B) pancreatic ductal adenocarcinoma in a woman aged 65 y. Hematoxylin and eosin stain, ×20.

448 Cancer Control October 2016, Vol. 23, No. 4 disease control (Table 3).3,21-25 Therefore, FOLFOX or Surgery folinic acid/fluorouracil/irinotecan (FOLFIRI) is usu- Overall, surgical resection remains the mainstay of ther- ally administered to patients with high performance apy for nonmetastatic disease, although adjuvant thera- status, whereas patients with a lower performance py is common. Most patients with acinar cell carcinoma status usually receive gemcitabine/protein-bound of the pancreas undergo negative margin resection.3,21 paclitaxel.3,8,9,21-25 From our systematic review, we found that patients who underwent resection had an overall median survival rate of approximately 47 months (Table 4).3,21-25 The contri- Table 2. — Disease Control Rates of Chemotherapy bution of adjuvant chemotherapy and radiotherapy has and Radiotherapy in Select Series of Acinar Cell Carcinoma not been well investigated in the literature. Resection of the Pancreas and adjuvant chemotherapy were associated with the Study Type of Treatment No. of Disease longest rate of median overall survival (see Table 4).3,21-25 Patients Control a Rate, % Treatment at Our Institution Butturini25 Gemcitabine-based 7 29 We identified 21 patients in a case series; 7 patients chemotherapy had metastatic disease based on imaging at the time Holen21 Fluoropyrimidine-based 22 41 of initial presentation (Table 5). Most patients were chemotherapy RT alone 1 100 Fluoropyrimidine-based Table 4. — Systematic Review Summary of Survival chemotherapy 4 100 Differences Between Select Patients With Resectable and RT Nonresectable Acinar Cell Carcinoma of the Pancreas Lowery24 Gemcitabine-based 20 35 Disease Primary Treatment Median Rate chemotherapy Status of Overall Matos3 Mixed chemotherapy 4 50 Survival, mos Mixed chemotherapy Metastatic Chemotherapy 17 2 100 RT Unresectable Chemotherapy and radiotherapy 38 Seki22 Gemcitabine-based 1 0 Resectable Resection and adjuvant therapy 47 chemotherapy Data from references 3 and 21 to 25. Fluoropyrimidine-/gemcitabine- 3 33 based chemotherapy Seth23 Mixed chemotherapy 4 100 Table 5. — Select Patient Characteristics From Current RT Patient Series Current Gemcitabine-based 6 50 series chemotherapy Patient Characteristic No. of Patients (%) Fluoropyrimidine-based Sex 6 67 chemotherapy Male 16 (76) Mixed chemotherapy 5 100 Female 5 (24) RT American Joint Committee on Cancer Stage Study patients not receiving chemotherapy are not included. aComplete response + partial response + stable disease. II 7 (33) RT = radiotherapy. III 5 (24) IV 9 (43) Site Table 3. — Summary of Select Disease Control Rates For the Nonsurgical Therapy of Acinar Cell Carcinoma Head 8 (38) of the Pancreas Neck/body 7 (33) Treatment No. of Disease Tail 6 (29) Patients Control Surgical Intervention Rate,a % Whipple 6 (29) Any chemotherapy 79 55 Distal pancreatectomy 3 (14) Any radiotherapy 16 100 Total pancreatectomy 2 (10) Gemcitabine-based chemotherapy 34 28 Fluoropyrimidine-based chemotherapy 35 60 Unresectable 3 (14) No surgery 7 (33) aComplete response + partial response + stable disease. Data from references 3 and 21 to 25. Mean age ± standard deviation: 63.5 ± 14 y.

October 2016, Vol. 23, No. 4 Cancer Control 449 A least 13 months after diagnosis. 100 No surgery Of the patients who underwent surgery, 1 patient Surgery also received neoadjuvant chemotherapy and 1 patient 75 P < .001 also received neoadjuvant radiotherapy. Eight patients 50 received adjuvant chemotherapy, and 4 patients re-

25 ceived adjuvant radiotherapy. More surgical patients received chemotherapy than nonsurgical patients

Survival Probability, % 0 (93% vs 57%; P < .05), with 82% of those receiving a 0 20 40 60 80 gemcitabine-based chemotherapy protocol. Five pa- Months tients (24%) received radiotherapy, and all were surgical No. at Risk patients: 1 in a neoadjuvant and 4 in adjuvant settings. No surgery 7 0 0 0 0 Utilizing Kaplan-Meier analysis, we found rates Surgery 14 8 3 1 0 of survival to be statistically associated with stage of B disease (P = .005), surgery (Fig 2A; log-rank P < .001), 100 and radiotherapy (Fig 2B; log-rank P = .008), but not 75 chemotherapy (Fig 2C; log-rank P = .09). Patients who underwent resection followed by adjuvant radiother- 50 No RT RT apy had the longest survival (Fig 3). On multivariate 25 P = .008 Cox regression analysis, higher disease stage was associated with a shorter rate of survival (hazard ratio

Survival Probability, % 0 0 20 40 60 80 [HR] 6.70, P =.003) and adjuvant radiotherapy was associated with longer survival (HR 0.05, P = .03), but Months No. at Risk neither R0 margin nor adjuvant chemotherapy was No RT 16 4 1 0 0 statistically associated with improved rates of survival RT 5 4 2 1 0 (0.70 and 0.30, respectively; P = .2 for both). C 100 No chemo Discussion Chemo Few large trials have studied acinar cell carcinoma of 75 P = .09 the pancreas because it is such a rare tumor. We were 50 only able to identify 6 studies that fully met the criteria for our systematic review. Our institutional experience 25 with 21 patients is the second largest case series in the

Survival Probability, % 0 literature. Because acinar cell carcinoma of the pancre- 0 20 40 60 80 as is treated similarly to pancreatic ductal adenocarci- Months noma, it is interesting to compare the outcomes. Analy- No. at Risk sis of Surveillance, Epidemiology, and End Results data No chemo 4 1 1 0 0 Chemo 17 7 2 1 0

Fig 2A–C. — (A) Kaplan-Meier survival estimates based on surgery, 100 (B) RT, and (C) chemotherapy. Patients with acinar cell carcinoma of the pancreas had significantly improved rates of survival following surgery 75 (P < .001) and RT (P = .008) but not chemotherapy (P = .09). RT = radiotherapy. 50 No RT RT 25 P = .018

men. The average age was 63.5 ± 14.1 years, and 38% Survival Probability, % 0 of lesions were located in the pancreatic head. No pa- 0 20 40 60 80 tients had stage I disease. The median follow-up pe- Months riod was 1.2 years (range, 1 month to 6.5 years), and No. at Risk 62% of study patients were initially diagnosed else- No RT 9 4 1 0 0 where. Of the 21 patients, 14 were candidates for cu- RT 5 4 2 1 0 rative resection. Of those, 11 underwent R0 or R1 resections and 3 patients received R2 resection. At last Fig 3. — Kaplan-Meier survival estimates among patients undergoing surgery and receiving RT. Surgical candidates with acinar cell carcinoma follow-up, 43% of patients were alive; rates of actual of the pancreas had significant improvement in their rates of survival 1- and 5-year survival were 67% and 5%, respective- probability when treated with RT. ly. All patients who received radiotherapy lived for at RT = radiotherapy.

450 Cancer Control October 2016, Vol. 23, No. 4 demonstrated higher rates of survival in acinar cell nar cell.33 An alternative pathway for carcinogenesis is carcinoma of the pancreas compared with pancreatic described by KRAS-mediated acinar-ductal metaplasia ductal adenocarcinoma for both resected and unre- and the subsequent development of pancreatic ductal sected cases,1 and analysis of the National Cancer Da- adenocarcinoma.31,34 In this setting, KRAS is hypothe- tabase demonstrated 5-year survival rates for patients sized as being the driving force for the development of with resected acinar cell carcinoma of the pancreas pancreatic ductal adenocarcinoma, as most commonly of 52%, 40%, 23%, and 17% for stages I, II, III, and IV, seen in the humans, whereas non–KRAS-mutated exo- respectively.29 Positive nodal disease was not associ- crine pancreatic carcinomas develop into acinar cell ated with decreased rates of survival in patients with carcinoma of the pancreas.35-37 Assuming this is true, acinar cell carcinoma of the pancreas, but patients with then KRAS wild-type status would be a critical finding high-grade acinar cell carcinoma of the pancreas had a in distinguishing acinar cell carcinoma of the pancreas 2-fold worse rate of 5-year survival (P = .003).29 Regard- from pancreatic ductal adenocarcinoma. ing margin status, negative microscopic margins were Mutation of DCC may also serve as an early ge- associated with 2-year median survival rates compared netic change in acinar cell carcinoma of the pancreas; with 1-year median survival rates for patients with by contrast, it is a late change in pancreatic ductal ad- R1-/R2-positive margins (P < .02).29 enocarcinoma. Furthermore, the rate of disease control Debate exists as to whether pancreatic ductal ad- was higher in patients with acinar cell carcinoma of the enocarcinoma is a radiosensitive tumor, but acinar pancreas than in those with pancreatic ductal adeno- cell carcinoma of the pancreas does appear to be ra- carcinoma. This finding suggests that acinar cell carci- diosensitive based on the multivariate analysis and noma of the pancreas and pancreatic ductal adenocar- Kaplan-Meier survival curves presented here along cinoma are 2 distinct malignancies arising from similar with other published data.30 Given the findings here cells but with important and early points of divergence and in the literature, it appears that, although acinar (KRAS vs DCC). Other groups have demonstrated that cell carcinoma of the pancreas and pancreatic ductal acinar cell cystadenoma, a benign and rarely prema- adenocarcinoma are similar exocrine pancreatic malig- lignant pancreatic lesion, may have genetic alteration nancies, they are distinct entities.29,30 patterns more similar to pancreatic ductal adenocar- This systematic review and case series illuminate cinoma than to acinar cell carcinoma of the pancreas, 3 important findings. As expected, patients able to toler- suggesting that some genetic overlap may be present.38 ate surgery live significantly longer than patients with At our institution, we favor the initial use of unresectable disease.3,21-25 Patients who undergo radio- FOLFOX followed by FOLFIRI as second-line treatment therapy in combination with surgery also live signifi- for patients with acinar cell carcinoma of the pancre- cantly longer than those not receiving radiotherapy.3,21-25 as and moderate to high performance status (East- The biological mechanism of this result is unknown, but ern Cooperative Oncology Group [ECOG] 0–1) either we hypothesize it is because acinar cell carcinoma of the in the adjuvant setting or as systemic therapy (Fig 4). pancreas is a radiosensitive tumor. This survival advan- Use of folinic acid/fluorouracil/irinotecan/oxali- tage of radiotherapy was seen despite 11 of 14 patients platin (FOLFIRINOX) as first-line therapy for meta- in our case series receiving surgery (R0/R1 resections), static acinar cell carcinoma of the pancreas has been further suggesting that radiotherapy is valuable regard- described in the literature, but whether the benefits of less of margin status. We did not find an association be- such an aggressive regimen outweighs the risks is un- tween use of chemotherapy and survival rate. This may clear.39 Because rates of disease control are higher with be due to the low number of patients in the nonchemo- the use of combination, fluoropyrimidine-based che- therapy group, but lack of high-level evidence remains motherapies, we also consider FOLFIRINOX in rela- problematic given the overall poor prognosis of acinar tively young, very highly active individuals with ECOG cell carcinoma of the pancreas. performance status 0 (but rarely seen in our clinical The cell of origin for acinar cell carcinoma of the practice). Among patients with moderate ECOG per- pancreas is presumed to be the acinar cell, whereas formance status 1 to 2, gemcitabine combined with the cell of origin for pancreatic adenocarcinoma is the protein-bound paclitaxel is our preferred chemother- duct epithelium.31-33 However, controversy is ongoing apy regimen, with single-agent gemcitabine used for regarding this assumption because it is based on the patients with poor performance status not wishing to phenotype of pancreatic ducts in normal tissue, pan- enter hospice or palliative care. creatic intraepithelial neoplastic lesions, and carcino- In the neoadjuvant setting, our clinical approach ma.31 For example, injury to benign cells in the setting for pancreatic cancer typically involves treating pa- of malignancy (ie, enzyme-mediated cell destruction tients with SBRT who have borderline resectable or due to malignant local duct obstruction) may create a locally advanced disease. Therapy involves 5 days of setting in which malignancy may appear to originate in fiducial marker–based radiotherapy applied to gross a duct cell when in fact it actually originated in an aci- disease (median dose of 30 Gy and simultaneously

October 2016, Vol. 23, No. 4 Cancer Control 451 50 Gy to the area of vessel involvement/abutment) to toxicity.30,42 Based on data from the National Cancer help facilitate a negative margin resection.40,41 In pa- Database, R0 resection has been shown to improve tients with acinar cell carcinoma of the pancreas, this rates of overall survival on multivariate analysis, so general treatment approach could be considered given neoadjuvant radiotherapy should be considered when the high R0 resection rates and minimal rates of late a likelihood exists of positive margins.29 Because patients with locally advanced tumors are unlikely to be candidates for resection, care must be taken to avoid late radiation-associat- Diagnosis Made on Biopsy (eg, FNA) ed toxicities to the surrounding gastrointestinal organs. As such, intensity-modulated radiothera- Complete Staging py can be considered instead of Local/regional: • Endoscopic ultrasonography to evaluate superior mesenteric vein/artery and portal vein SBRT for locally advanced acinar • CT protocol for the pancreas (cut 3 mm, arterial/portal/venous phase) cell carcinoma of the pancreas. Distant: • CT of the thorax with contrast and PET-CT Limitations Performance status Due to the rarity of this malignancy, we did not identify any prospective studies or large case- control studies to help guide clini- Resectable Borderline Locally Metastatic cal management. Likewise, oncol- Pancreaticoduodenectomy Resectable Advanced Systemic ogists are more familiar with the or distal pancreatectomy Neoadjuvant Systemic chemotherapy treatment protocols for pancreatic chemotherapy chemotherapy + FOLFOX ductal adenocarcinoma, because SBRT + FOLFOX resectiona this disease is more common, RT and, as such, these regimens are often utilized for acinar cell carcinoma of the pancreas despite the Adjuvant Adjuvant fact that evidence is lacking. The Chemotherapy Radiotherapy retrospective nature of this review FOLFOX Positive margins also limits our interpretation of SBRT-naïve the results. Fundamentally, this limitation is also due to the rarity of the malignancy, thus limiting Progression the initiation of prospective trials. Another important limitation relates to tumor response rates in Second-Line Therapy retrospective studies. Oftentimes, FOLFIRI the time to progression is based on when a patient is scheduled for his or her next imaging, not a pre- Third-Line Therapy defined time point, as seen in a Gemcitabine/protein-bound prospective study. As such, unless paclitaxel a patient is symptomatic, the time to progression may be partially related to an institutional bias. This Palliative Care is a difficult to quantify but important limitation.

Fig 4. — Algorithm for the treatment of acinar cell carcinoma of the pancreas. This treatment plan Future Research is based on expert opinion and focuses on the importance of negative-margin resection coupled Three critical inquiries must be ex- with RT. plored to improve rates of survival CT = computed tomography, FNA = fine needle aspiration, FOLFIRI = folinic acid/fluorouracil/irinotecan, FOLFOX = folinic acid/fluorouracil/oxaliplatin, PET = positron emission tomography, among individuals with acinar cell RT = radiotherapy, SBRT = stereotactic body radiotherapy. carcinoma of the pancreas. Spe- aIf no progression on repeat staging. cific mechanisms (ie, gene muta-

452 Cancer Control October 2016, Vol. 23, No. 4 analysis of a distinct and diagnostically challenging neoplasm. Modern tions) that lead to acinar cell carcinoma of the pancreas Pathol. 2014;27(suppl 2):447A. must be better elucidated to (theoretically) incorporate 6. Bergmann F, Aulmann S, Welsch T, et al. Molecular analysis of pancreatic acinar cell cystadenomas: Evidence of a non-neoplastic nature. small molecule inhibitors into therapy. As demonstrat- Oncol Lett. 2014;8(2):852-858. ed by the overall higher survival rates seen in acinar 7. Hartwig W, Denneberg M, Bergmann F, et al. Acinar cell carcinoma of the pancreas: is resection justified even in limited metastatic disease? cell carcinoma of the pancreas (see Table 4), acinar cell Am J Surg. 2011;202(1):23-27. carcinoma of the pancreas may likely be molecularly 8. Lee JL, Kim TW, Chang HM, et al. Locally advanced acinar cell carcinoma of the pancreas successfully treated by capecitabine and con- driven in a distinct way from pancreatic ductal adeno- current radiotherapy: report of two cases. Pancreas. 2003;27(1):e18-e22. carcinoma.3,21-25 The radiosensitivity of acinar cell car- 9. Chaudhary P, Ranjan G, Chaudhary A, et al. Acinar cell carcinoma: a rare pancreatic malignancy. Clin Pract. 2013;3(2):e18. cinoma of the pancreas must also be investigated and 10. Wang T, X SM, Kong D, et al. Effect of ionizing radiation on acinar modulated to optimize therapy. Our data suggest that morphogenesis of human prostatic epithelial cells under three-dimension- al culture conditions. Neoplasma. 2012;59(3):269-281. acinar cell carcinoma of the pancreas is a radiosensi- 11. Tateishi Y, Sasabe E, Ueta E, et al. Ionizing irradiation induces tive tumor. Similar to many other malignancies, active apoptotic damage of salivary gland acinar cells via NADPH oxidase 1-dependent superoxide generation. Biochem Biophys Res Commun. chemotherapeutics also need identification. Whether 2008;366(2):301-307. current chemotherapy options add any improved like- 12. Peter B, Van Waarde MA, Vissink A, et al. Radiation-induced cell proliferation in the parotid and submandibular glands of the rat. Radiat lihood of long-term survival remains unclear based on Res. 1994;140(2):257-265. published data. 13. Ogbonna OH, Garcon MC, Syrigos KN, et al. Mixed acinar-neuroendocrine carcinoma of the pancreas with neuroendocrine predominance. Case Rep Med. 2013;2013:705092. Conclusions 14. Bergmann F, Aulmann S, Sipos B, et al. Acinar cell carcinomas of the pancreas: a molecular analysis in a series of 57 cases. Virchows Archiv. Acinar cell carcinoma of the pancreas is a rare pancre- 2014;465(6):661-672. atic malignancy with a prognosis slightly better than 15. Baer R, Cintas C, Dufresne M, et al. Pancreatic cell plasticity and cancer initiation induced by oncogenic Kras is completely dependent on ductal adenocarcinoma. At Moffitt Cancer Center, our wild-type PI 3-kinase p110alpha. Genes Dev. 2014;28(23):2621-2635. standard approach is surgical resection followed by ad- 16. Furlan D, Sahnane N, Bernasconi B, et al. APC alterations are frequently involved in the pathogenesis of acinar cell carcinoma of the pan- juvant fluorouracil-based chemotherapy, radiotherapy, creas, mainly through gene loss and promoter hypermethylation. Virchows or both (see Fig 4). The utility of chemotherapy is un- Arch. 2014;464(5):553-564. 17. Furukawa T, Sakamoto H, Takeuchi S, et al. Whole exome se- certain; however, when it is used, a fluorouracil-based quencing reveals recurrent mutations in BRCA2 and FAT genes in acinar regimen is often proposed because of its higher rates cell carcinomas of the pancreas. Sci Rep. 2015;5:8829. 18. La Rosa S, Sessa F, Capella C. Acinar cell carcinoma of the pan- of disease control when compared with gemcitabine or creas: overview of clinicopathologic features and insights into the molecu- other chemotherapeutic agents. (However, this treat- lar pathology. Front Med (Lausanne). 2015;2:41. 19. Hilgers W, Song JJ, Haye M, et al. Homozygous deletions inactivate ment plan is based on expert opinion and individual- DCC, but not MADH4/DPC4/SMAD4, in a subset of pancreatic and biliary ized for patient performance status.) Negative micro- cancers. Genes Chromosomes Cancer. 2000;27(4):353-357. 20. Moher D, Liberati A, Tetzlaff J, et al. Preferred reporting items for scopic margins are the most important prognostic systematic reviews and meta-analyses: the PRISMA statement. PLoS factor, because survival for patients with positive mi- Med. 2009;6(7):e1000097. 21. Holen KD, Klimstra DS, Hummer A, et al. Clinical characteristics croscopic margins is generally as poor as those seen and outcomes from an institutional series of acinar cell carcinoma of the with positive gross margins. pancreas and related tumors. J Clin Oncol. 2002;20(24):4673-4678. 22. Seki Y, Okusaka T, Ikeda M, et al. Four cases of pancreatic acinar Multidisciplinary management of pancreatic acinar cell carcinoma treated with gemcitabine or S-1 as a single agent. Jpn J Clin cell carcinoma should be considered due to the lack of Oncol. 2009;39(11):751-755. 23. Seth AK, Argani P, Campbell KA, et al. Acinar cell carcinoma of the high-level data in treating this patient population. In pancreas: an institutional series of resected patients and review of the current addition, referral to a quaternary center is appropriate literature. J Gastrointest Surg. 2008;12(6):1061-1067. 24. Lowery MA, Klimstra DS, Shia J, et al. Acinar cell carcinoma of given the rarity of this malignancy. Despite the rarity the pancreas: new genetic and treatment insights into a rare malignancy. of this disease, further investigation is needed to per- Oncologist. 2011;16(12):1714-1720. 25. Butturini G, Pisano M, Scarpa A, et al. Aggressive approach to aci- sonalize care to improve long-term survival in these nar cell carcinoma of the pancreas: a single-institution experience and a patients. literature review. Langenbecks Arch Surg. 2011;396(3):363-369. 26. Lack EE, Cassady JR, Levey R, et al. Tumors of the exocrine pan- Acknowledgments: The authors thank Rasa creas in children and adolescents. A clinical and pathologic study of eight Hamilton for editorial support and Kim Murphy for ad- cases. Am J Surg Pathol. 1983;7(4):319-327. 27. Kim HJ, Kim YK, Jang KT, et al. Intraductal growing acinar cell car- ministrative support. cinoma of the pancreas. Abdominal Imaging. 2013;38(5):1115-1119. 28. Sudo K, Ishihara T, Hirata N, et al. Randomized controlled study References of gemcitabine plus S-1 combination chemotherapy versus gemcitabine for unresectable pancreatic cancer. Cancer Chemother Pharmacol. 1. Wisnoski NC, Townsend CM Jr, Nealon WH, et al. 672 patients with 2014;73(2):389-396. acinar cell carcinoma of the pancreas: a population-based comparison to 29. Schmidt CM, Matos JM, Bentrem DJ, et al. Acinar cell carcinoma pancreatic adenocarcinoma. Surgery. 2008;144(2):141-148. of the pancreas in the United States: prognostic factors and comparison to 2. Kong B, Michalski CW, Erkan M, et al. From tissue turnover to ductal adenocarcinoma. J Gastrointest Surg. 2008;12(12):2078-2086. the cell of origin for pancreatic cancer. Nat Rev Gastroenterol Hepatol. 30. Moningi S, Dholakia AS, Raman SP, et al. The role of stereotactic 2011;8(8):467-472. body radiation therapy for pancreatic cancer: a single-institution experi- 3. Matos JM, Schmidt CM, Turrini O, et al. Pancreatic acinar cell carci- ence. Ann Surg Oncol. 2015;22(7):2352-2358. noma: a multi-institutional study. J Gastrointest Surg. 2009;13(8):1495-1502. 31. Shi C, Hong SM, Lim P, et al. KRAS2 mutations in human pancre- 4. Wang Y, Wang S, Zhou X, et al. Acinar cell carcinoma: a report atic acinar-ductal metaplastic lesions are limited to those with PanIN: im- of 19 cases with a brief review of the literature. World J Surg Oncol. plications for the human pancreatic cancer cell of origin. Mol Cancer Res. 2016;14(1):172. 2009;7(2):230-236. 5. Basturk O, Adsay V, Hruban RH, et al. Pancreatic acinar cell car- 32. Morohoshi T, Held G, Kloppel G. Exocrine pancreatic tumours and cinomas with prominent neuroendocrine differentiation: clinicopathologic their histological classification. A study based on 167 autopsy and 97 surgi-

October 2016, Vol. 23, No. 4 Cancer Control 453 cal cases. Histopathology. 1983;7(5):645-661. 33. Bockman DE. Cells of origin of pancreatic cancer: experimental animal tumors related to human pancreas. Cancer. 1981;47(6 suppl):1528-1534. 34. Aichler M, Seiler C, Tost M, et al. Origin of pancreatic ductal adenocarcinoma from atypical flat lesions: a comparative study in transgenic mice and human tissues. J Pathol. 2012;226(5):723-734. 35. Morris JP IV, Wang SC, Hebrok M. KRAS, Hedgehog, Wnt and the twisted developmental biology of pancreatic ductal adenocarcinoma. Nat Rev Cancer. 2010;10(10):683-695. 36. Shi G, DiRenzo D, Qu C, et al. Maintenance of acinar cell organization is critical to preventing Kras-induced acinar-ductal metaplasia. Oncogene. 2013;32(15):1950-1958. 37. Wood LD, Klimstra DS. Pathology and genetics of pancreatic neoplasms with acinar differentiation. Semin Diagn Pathol. 2014;31(6):491-497. 38. Jones M, Zheng Z, Wang J, et al. Impact of next-generation sequencing on the clinical diagnosis of pancreatic cysts. Gastrointest Endosc. 2016;83(1):140-148. 39. Schempf U, Sipos B, Konig C, et al. FOLFIRINOX as first-line treatment for unresectable acinar cell carcinoma of the pancreas: a case report. Zeitschrift Gastroenterol. 2014;52(2):200-203. 40. Mellon EA, Hoffe SE, Springett GM, et al. Long-term outcomes of induction chemotherapy and neoadjuvant stereotactic body radiotherapy for borderline resectable and locally advanced pancreatic adenocarcinoma. Acta Oncol. 2015;54(7):979-985. 41. National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology: Pancreatic Adenocarcinoma. v I.2016. https:// www.nccn.org/professionals/physician_gls/pdf/pancreatic.pdf. Accessed August 9, 2016. 42. Herman JM, Chang DT, Goodman KA, et al. Phase 2 multi-institutional trial evaluating gemcitabine and stereotactic body radiotherapy for patients with locally advanced unresectable pancreatic adenocarcinoma. Cancer. 2015;121(7):1128-1137.

454 Cancer Control October 2016, Vol. 23, No. 4 Special Report

A Single-Institution Study of Demographics and Outcomes of Adult Patients With Multiple Cancers Thanh P. Ho, MD, Nathan R. Foster, MS, and Aminah Jatoi, MD

Background: Few studies have described the demographics and outcomes in patients with multiple cancers. Methods: We reviewed the health records of patients with at least 3 invasive solid-tumor malignancies (lung, colorectal, pancreatic, prostate, breast, or ovarian) diagnosed between 1995 and 2014 at our institution. Results: A total of 163 patients were identified. The median age at first cancer diagnosis was 61 years. Sixteen (10%) had a family history of cancer, and 7 (4%) had a cancer-predisposing syndrome. Most had early-stage malignancies and underwent a series of surgical resections. The median survival from the first cancer diagnosis was 14 years (95% confidence interval: 12–16). Forty-three percent developed 4 or more cancers. In univariate analyses, older age, male sex, smoking history, concurrent cancer, and stage IV disease were associated with worse survival; by contrast, breast cancer as the first cancer diagnosis yielded better survival. In multivariate analyses, older age, smoking history, and stage IV disease were associated with worse survival. Conclusions: A small group of patients with no ostensible risk factors develop multiple early-stage cancers, receive curative therapy (often surgery), and live for years. These patients merit further study to ascertain the broader prevalence rates of individuals with multiple invasive cancers and to better understand factors that contribute to cancer risk.

Introduction noses compared with their counterparts who develop Adult patients with cancer are living longer.1 Zeng et al2 de novo cancers.7,8 However, to our knowledge, these reported on patients with cancer aged 50 to 64 years studies and others have not explored the presence of at the time of their diagnosis between 2005 and 2009 multiple cancers in the absence of predisposing syn- and observed that, when compared with similar pa- dromes or therapy-related secondary cancers. tients diagnosed during an earlier time interval, the hazard ratios (HRs) for patients with colorectal, breast, Purpose liver, and prostate cancers were 0.57 (95% confidence This study explored demographics and outcomes in interval [CI]: 0.55–0.60), 0.48 (95% CI: 0.45–0.51), 0.61 patients who developed multiple cancers. Some pa- (95% CI: 0.57–0.69), and 0.32 (95% CI: 0.30–0.36), re- tients contend with multiple malignancies through- spectively. These findings suggest that patients with out their lives, thus displaying a poorly characterized cancer are living longer and that cancer can, at times, predisposition to cancer and a need for ongoing can- be viewed as a long-term disease.2 cer surveillance. Reporting on the characteristics and Few studies have focused in general terms on outcomes of patients with multiple cancers might help patients who live for many years following a first di- oncologists better counsel patients about prognoses agnosis of cancer and then continue to be diagnosed and realistic treatment goals in the setting of multiple with multiple cancers over time.3,4 Previous studies malignancies. have described multiple cancers in adult patients with predisposing syndromes, showing, for example, that Methods individuals with Lynch syndrome or BRCA1/BRCA2 Overview mutations have a more favorable prognosis after can- This single-institution study focused on patients cer diagnoses.5,6 Others have described patients with with cancer seen at the Mayo Clinic (Rochester, therapy-related malignancies who have worse prog- MN) from 1995 to 2014, encompassing a 20-year interval that saw national improvements in cancer outcomes.2 The Mayo Clinic Tumor Registry, which From the Departments of Medicine (TPH), Health Sciences Research (NRF), and Oncology (AJ), Mayo Clinic, Rochester, Minnesota. relied in part on International Classification of Submitted March 10, 2016; accepted May 3, 2016. Diseases codes, pathology reports, and other clinical Address correspondence to Aminah Jatoi, MD, Department of Oncology, sources to input data, was searched for patients who Mayor Clinic, 200 First Street SW, Rochester, Minnesota 55905. met study criteria. E-mail: [email protected] Eligible patients had to be diagnosed between No significant relationships exist between the authors and the companies/organizations whose products or services may be referenced 1995 and 2014 with at least 3 invasive malignancies of in this article. at least 1 of the following types: lung, colorectal, pan-

October 2016, Vol. 23, No. 4 Cancer Control 455 creatic, prostate, breast, or ovarian. These malignan- The median age of the study participants was cies were chosen because they are among the most 61 years (range, 24–87). Sixteen (10%) had cancer in at common and lethal in the United States and have least 2 first-degree relatives, and 7 (4%) had a known 5-year relative survival rates of 16%, 59%, 7%, 95%, syndrome that predisposed them to cancer. Other 85%, and 41%.9 Inclusion criteria were determined a patient demographics appear in Table 1. priori. No major changes occurred in the Mayo Clinic Tumor Registry or in institutional infrastructure within Cancer Characteristics the interval above; hence, data retrieval uniformly oc- Cancer of the breast, colorectum, and lung were the curred over time. The Institutional Review Board at the 3 most common cancer types in the study group Mayo Clinic approved this study. (Table 2). Seventy patients (43%) had more than 3 cancers. In 10 patients, the first cancer diagnosis did not Medical Record Review fall within the 1995 to 2014 time interval, but these Each patient’s health records were reviewed and ab- patients were eligible for the study because they de- stracted for date of birth, sex, date of each cancer di- veloped 3 invasive cancers during this interval. The agnosis, stage of each cancer, smoking history, family median interval from the first ever to the second can- history of cancer (cancer in ≥ 2 first-degree relatives), cer diagnosis was 2.6 years (range, 0–45), and the documentation of genetic predisposition to cancer, median interval from the first ever to the third can- cancer treatment, vital status, and date of death or last cer diagnosis was 6.2 years (range, 0–58; Fig 1). Most follow-up visit. Because definitions of predisposing patients had relatively early-stage malignancies, and syndromes to cancer have evolved over time to include the majority also underwent surgical treatment for all clinical criteria and, in some instances, molecular crite- 3 cancers. Fewer than one-half of these specific can- ria, the methodology congruent with a real-time con- cers were treated with chemotherapy or radiotherapy. temporary standard was used in this study. If a date was not clearly documented, then a mid-month or mid- year date was used. If the health record lacked clear Table 1. — Characteristics of Study Patients (N = 163) documentation of an item, then an unknown category Characteristic Study Patients, was used. n (%) Median age at first cancer diagnosis, y (range) 61 (24–87) Data Analyses Sex Results are reported with means and percentages; Male 75 (46) all analyses are considered exploratory. Because pa- Female 88 (54) tients were heavily selected based on 6 targeted can- Cancer in ≥ 2 first-degree relatives cer types, this study was not designed to report on Yes 16 (10) the prevalence rates of patients with multiple cancers. No 144 (88) Kaplan-Meier curves were constructed to assess survival, and data were censored as appropriate. Cox pro- Genetic syndrome portional hazard models were used to explore whether Yes 7 (4) certain clinical variables were associated with survival. No 12 (7) Variables were chosen based on their clinical relevance Unknown 144 (88) to cancer outcomes and based on a desire for parsi- Smoking history mony to ensure stability of statistical models; in this Yes 118 (74) context, age at cancer diagnosis, sex, smoking history, No 42 (26) concurrent cancer diagnosis, stage IV disease (any ma- Concurrent cancer lignancy), and cancer type at first diagnosis were cho- Yes, 1 13 (8) sen. HRs and 95% CIs were reported. A 2-tailed P value Yes, 2 17 (10) of less than .05 was considered statistically significant. No 133 (82) All analyses were performed with JMP Statistical Soft- ≥ 4 cancersa ware (SAS Institute, Carey, NC). Yes 70 (43) No 93 (57) Results Percentages may not sum to 100% because of rounding or missing data. Baseline Demographics aOther cancers included multiple cases of colorectal cancer, multiple Nine patients were excluded because of lack of clarity cases of lung cancer, multiple cases of breast cancer, 6 cases of lymphoma, 5 cases of bladder cancer, 4 cases of melanoma, 4 cases as to whether 1 or more of their cancers represented of prostate cancer, 3 cases of head and neck cancers, 2 cases of small a new malignancy, a metastatic site, or recurrence of bowel cancer, 2 cases of ureteral cancer, 2 cases of ovarian cancer, and a previous cancer. The remaining 163 patients are the 1 each of chronic myelogenous and lymphocytic leukemias, and tubal, focus of this report. esophageal, pancreatic, testicular, and cervical cancers.

456 Cancer Control October 2016, Vol. 23, No. 4 Table 2. — Cancer Diagnoses Among Study Patients (N = 163) Diagnosis Study Patients, Diagnosis Study Patients, Diagnosis Study Patients, n (%) n (%) n (%) First Cancer Second Cancer Third Cancer Cancer Type Cancer Type Cancer Type Breast 43 (26) Breast 39 (24) Breast 28 (17) Colorectal 40 (25) Colorectal 52 (32) Colorectal 44 (27) Lung 38 (23) Lung 55 (34) Lung 78 (48) Ovarian 3 (2) Ovarian 0 Ovarian 1 (1) Pancreas 1 (1) Pancreas 1 (1) Pancreas 3 (2) Prostate 26 (16) Prostate 13 (8) Prostate 6 (4) Other 12 (7) Other 3 (2) Other 3 (2) Stage Stage Stage I 72 (47) I 87 (55) I 80 (55) II 43 (28) II 40 (25) II 30 (21) III 31 (20) III 26 (16) III 19 (13) IV 7 (5) IV 6 (4) IV 17 (12) Surgery Surgery Surgery Yes 147 (90) Yes 145 (90) Yes 126 (78) No 15 (9) No 17 (10) No 36 (22) Systemic Therapy Systemic Therapy Systemic Therapy Yes 69 (43) Yes 64 (40) Yes 62 (39) No 93 (57) No 97 (60) No 99 (61) Radiotherapy Radiotherapy Radiotherapy Yes 53 (33) Yes 34 (21) Yes 42 (26) No 107 (66) No 126 (78) No 117 (73)

A B 50 60

50 40

40 30

Years 20 Years

10 10

0 0

No. of Patients No. of Patients

Fig 1A–B. — Time interval between cancer diagnoses. These bar graphs show time on the vertical axis and each bar is representative of the number of patients. (A) The interval from the first to the second cancer diagnosis was 2.6 years (range, 0–45 years). (B) The median interval from the first to the third cancer was 6.2 years (range, 0–58 years).

October 2016, Vol. 23, No. 4 Cancer Control 457 Cancer Outcomes of cancer, and 2 had a documented genetic predispo- A total of 92 patients (56%) had died. The median fol- sition to cancer. The median age at first cancer di- low-up of patients still living was 11.4 years (range, agnosis was 64 years (range, 33–80). Thirty-one pa- 0.3–58.0). In 62 (67%) of the deceased patients, the tients (74%) were diagnosed with stage I disease. The cause of death was cancer; in 6 (7%), a nonmalignant remaining stage distribution included 5 with stage II, cause was determined; and, in 24 (26%), the cause was 5 with stage III, 1 with stage IV, and 1 whose stage unknown. The overall median survival for the entire was unknown. A total of 41 patients (95%) were treat- group was 14 years from the time of first cancer diag- ed with surgery for the first cancer, 26 (60%) were nosis (95% CI: 12–16; Fig 2). treated with chemotherapy for the first cancer, and In univariate analyses, older age, male sex, a 25 (58%) were treated with radiotherapy for the first smoking history, a concurrent cancer diagnosis, and stage IV disease among any of the 100 3 malignancies yielded a statistically significant association with worse survival 80 (Table 3). Breast cancer as the first diagnosis of cancer was associated with improved sur- 60 vival. However, in multivariate analyses, older age, smoking history, and any diagnosis 40 of stage IV disease yielded a Patients Alive, % statistically significant association with worse survival 20 (see Table 3). Among the 43 patients who had breast cancer as 0 their first cancer, all were 0 2 4 6 8 10 12 14 16 18 20 22 24 women. A total of 29 (69%) Years From First Cancer Diagnosis had a smoking history (median of 35 pack-years), Fig 2. — The overall median survival was 14 years. The 95% confidence intervals of 12 and 16 years are 5 (12%) had a family history shown as the parallel dotted lines.

Table 3. — Results of the Cox Proportional Hazards Models Univariate Analyses Multivariate Analyses (N = 163, n = 92 deaths) (n = 136, n = 72 deaths)a Variable HR (95% CI) P Valueb HR (95% CI) P Valueb Age (1-y increase) 1.06 (1.04–1.09) < .0001 1.09 (1.06–1.13) < .0001 Sex (male vs female) 1.99 (1.31–3.04) .0010 0.92 (0.47–1.82) .8020 Smoking history (yes vs no)c 1.79 (1.07–3.19) .0240 2.53 (1.36–5.04) .0020 Concurrent cancers (yes vs no) 1.98 (1.13–3.28) .0190 1.54 (0.76–2.90) .2180 Stage IV (vs not stage IV)d 1.84 (1.09–3.02) .0240 2.02 (1.09–3.65) .0250 First diagnosis of cancer .0040 .4300 Breast 0.48 (0.25–0.87) 0.75 (0.31–1.84) Colorectal 1.43 (0.83–2.43) 0.96 (0.46–1.95) Lung 1.06 (0.60–1.83) 1.42 (0.69–2.87) Other (ref) — — aThe multivariate model included 72 deaths as a result of missing data. Separate models that included select variables above yielded similar findings, thus demonstrating the stability of the model. bLikelihood-ratio P value. cMissing 3 observations and 2 events (n = 160, n = 90 events). dMissing 24 observations and 18 events (n = 139, n = 74 events). CI = confidence interval, HR = hazard ratio.

458 Cancer Control October 2016, Vol. 23, No. 4 cancer. Second cancers within this group included tors remain important regardless of whether a patient breast cancer in 25 (58%), colorectal cancer in 9 (21%), has been diagnosed with 1 cancer type or several. and lung cancer in 9 (21%). Third cancers included More than 40% of the patients in our study were breast cancer in 16 (37%), colorectal cancer in 6 (14%), diagnosed with 4 or more cancers. Although old age lung cancer in 18 (42%), ovarian cancer in 1 (2%), pan- is a risk factor for cancer, the patients in the current creatic cancer in 1 (2%), and another type of cancer in study were not elderly when diagnosed with their first 1 (2%). Twenty-eight patients remain alive at the time malignancy (median age, 61 years). Moreover, a small of this report; the rate of overall median survival was percentage had a known syndrome that predisposed 16 years. Among the 15 deceased patients, the cause of them to cancer or a compelling reason to suspect that death was related to cancer in 9 cases. their history of recurrent cancers was the result of prior cancer therapy, although our understanding of Discussion genetic predisposition to cancer has grown over the This study focused on patients diagnosed with 3 or years. However, many of these patients were treated more common, invasive, and sometimes lethal malig- surgically — a therapy not generally known to spawn nancies. It was not intended to report on the prevalence secondary cancers. Thus, perhaps other, undiscovered rates of patients with multiple cancers because our eli- factors are at play to explain the development of multi- gibility criteria were deliberatively selective; however, ple cancers. This potential for uncharacterized risk fac- the scenario of multiple cancers appears to be rare, par- tors for cancer — including genetic factors — requires ticularly when considering the number of cancer diag- further investigation. noses every year at our institution — 868 patients were diagnosed with breast cancer in 2014 at the Mayo Clinic Limitations alone (unpublished data) — and when acknowledging The retrospective design of our study is a limitation and that cancer remains a major cause of death around the led to missing data in some instances, as is the case in world.10 We have identified a small cohort of patients any retrospective study. However, by using a retrospec- who illustrate this clinical scenario. This study con- tive design, we were able to assess 20 years of data with- tributes to a small but growing body of literature that in a short time frame. Another limitation is our small reports on the development of multiple cancers in a sample size, which is the result of our single-institution group of patients who have not been specifically select- study design. However, our single-institution focus ened based on a genetic or otherwise known predisposi- abled us to acquire detailed information on the dates, tion for having a higher risk of cancer.4,11 cancer stage, cancer therapy, and, in some cases, causes Despite the presumed rarity of this clinical sce- of death. The single-institution nature of our study may nario, this study provides important insights. These have also allowed for a greater degree of homogeneity patients often develop an early-stage malignancy, un- among data elements, thereby enabling us to identify dergo surgical resection, and then live for many years conclusions more readily. prior to developing a subsequent malignancy that can be treated in a similar fashion. The median survival for Conclusions the study group exceeded 10 years, underscoring the Patients diagnosed with multiple cancers may continue fact that this subgroup is treated and then re-treated for to live for years. To our knowledge, little insight exists a new cancer over time. Thus, a history of more than in the literature on the attitudes of health care profes- 1 early-stage cancer should not negatively impact the sionals toward patients with a history of multiple ma- management of subsequent early-stage cancers. In ad- lignancies, and repeated cancer diagnoses followed by dition, although comparisons across data sets are peril- a cumulative history of cancer treatment could conceiv- ous and, although our sample size is too small to al- ably lead to a biased approach to repeat cancer therapy low for formal adjusted comparisons across data sets, over time. The findings from this study underscore the when we examined patients who were first diagnosed need to continue to try to apply curative therapy to pa- with breast cancer and then developed multiple malig- tients with a history of multiple cancers, regardless of nances thereafter, we did not observe major differenc- whether these cancers are primary, secondary, or ter- es in survival based on benchmark data.9 For example, tiary. These findings also suggest the need to further the latter shows that 88% of patients with breast can- investigate the prevalence and risk factors for the de- cer reached the 5-year mark,9 whereas a survival curve velopment of multiple cancers. from patients with breast cancer in our study shows that more than 90% did the same (data not shown). References Risk factors for early death remain relevant to pa- 1. American Cancer Society. Cancer Facts and Figures 2016. tients with multiple cancers. Older age, smoking, and http://www.cancer.org/acs/groups/content/@research/documents/ document/acspc-047079.pdf. Accessed September 2, 2016. stage IV disease are all associated with poor survival 2. Zeng C, Wen W, Morgans AK, et al. Disparities by race, age, and in patients with multiple cancers. These prognostic fac- sex in the improvement of survival for major cancers: results from the Na-

October 2016, Vol. 23, No. 4 Cancer Control 459 tional Cancer Institute Surveillance, Epidemiology, and End Results (SEER) program in the United States, 1990 to 2010. JAMA Oncol. 2015;1(1):88-96. 3. Utada M, Ohno Y, Hori M, et al. Incidence of multiple primary cancers and interval between first and second primary cancers. Cancer Sci. 2014;105(7): 890-896. 4. Xu LL, Gu KS. Clinical retrospective analysis of cases with multiple primary malignant neoplasms. Genet Mol Res. 2014;13(4):9271-9284. 5. Van den Broek AJ, Schmidt MK, van ’t Veer LJ, et al. Worse breast cancer prognosis of BRCA1/BRCA2 mutation carriers: what’s the evidence? A systematic review with meta-analysis. PLoS One. 2015;10(3):e0120189. 6. Gryfe R, Kim H, Hsieh ET, et al. Tumor microsatellite instability and clinical outcome in young patients with colorectal cancer. N Engl J Med. 2000;342(2):69-77. 7. Travis LB. The epidemiology of second primary cancers. Cancer Epidemiol Biomark Prev. 2006;15(11):2020-2026. 8. Travis LB, Fosså SD, Schonfeld SJ, et al. Second cancers among 40,576 testicular cancer patients: focus on long-term survivors. J Natl Cancer Inst. 2005;97(18):1354-1365. 9. US Cancer Statistics Working Group. United States Cancer Statistics: 1999–2013 Incidence and Mortality Web-Based Report. Atlanta, GA: US Department of Health and Human Services; Centers for Disease Control and Prevention; National Cancer Institute; 2016. https://nccd.cdc.gov/uscs /Survival/Relative_Survival_Tables.pdf. Accessed September 2, 2016. 10. World Health Organization. The top 10 causes of death. http://www. who.int/mediacentre/factsheets/fs310/en/index2.html. Accessed Septem- ber 13, 2016. 11. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2016. CA Cancer J Clin. 2016;66(1):7-30.

460 Cancer Control October 2016, Vol. 23, No. 4 Ten Best Readings Relating to Health Disparities in Oncology

Sharma P, Ashouri K, Zargar-Shoshtari K, et al. pared with the US population and under-represent Racial and economic disparities in the treat- Asians and Hispanics. All tumor types from whites ment of penile squamous cell carcinoma: results contained enough samples to detect a 10% mutational from the National Cancer Database. Urol Oncol. frequency; however, group-specific mutations with 2016;34(3):122.e9-15. a 10% frequency were detectable for black patients Researchers evaluated overall survival (OS) dif- with breast cancer alone. It is probable but not well ferences in economics and sociodemographics and understood that ethnic diversity may be related to the economic differences among 5,412 study patients pathogenesis of cancer and may have an impact on with penile squamous cell carcinoma. Survival did the generalizability of these study findings. not change over the study period, but the authors found that black patients had worse median OS and Dutil J, Golubeva VA, Pacheco-Torres AL, et al. The presented with a more advanced stage of disease. spectrum of BRCA1 and BRCA2 alleles in Latin Those who had median incomes of at least $63,000 America and the Caribbean: a clinical perspective. and carried private insurance had better median OS Breast Cancer Res Treat. 2015;154(3):441-453. and presented with a lower stage of disease. Thus, A review of the medical literature reporting on the researchers determined that racial and economic the prevalence and/or spectrum of BRCA1 and BRCA2 differences exist in the OS rates of patients with penile variants was conducted. A total of 4,835 individuals squamous cell carcinoma. from Latin America, the Caribbean, and Hispanics from the United States were included, of which 167 Langford AT, Resnicow K, Dimond EP, et al. Racial/ unique pathogenic variants were reported. In un- ethnic differences in clinical trial enrollment, re- selected breast cancer cases, the prevalence ranged fusal rates, ineligibility, and reasons for decline from 1.2% to 27.1%. Some countries presented with among patients at sites in the National Cancer few recurrent pathogenic variants, whereas others Institute’s Community Cancer Centers Program. were characterized by diverse, nonrecurrent variants. Cancer. 2014;120(6):877-884. The proportion of BRCA pathogenic variants shared Racial and ethnic differences were studied among between US Hispanics and Latin American popula- patients who had enrolled in, refused to enroll, were tions was estimated to be 10.4%. Within Latin America ineligible to enroll, or wanted to participate in a clini- and the Caribbean, 8.2% of the BRCA variants were cal trial enrollment. Age 65 years or older, male sex, present in more than 1 country. Countries with a and non-Hispanic black race were significantly associ- high prevalence of BRCA pathogenic variants may ated with more medical or physical conditions. Those benefit from more aggressive testing strategies, and who were aged 65 years or older were significantly testing of recurrent variant panels might present a more likely to have low clinical trial enrollment; by cost-effective solution for improving genetic testing contrast, being male was significantly associated with in some countries. low refusal rates. The authors concluded that better management of physical or medical conditions before Weitzel JN, Clague J, Martir-Negron A, et al. Preva- and during treatment might increase the pool of pa- lence and type of BRCA mutations in Hispanics tients eligible to enroll in clinical trials. undergoing genetic cancer risk assessment in the southwestern United States: a report from the Spratt DE, Chan T, Waldron L, et al. Racial/ethnic Clinical Cancer Genetics Community Research disparities in genomic sequencing. JAMA Oncol. Network. J Clin Oncol. 2013;31(2):210-216. 2016;2(8):1070-1074. Researchers sought to determine the prevalence A retrospective review of individual patient data and type of BRCA1 and BRCA2 mutations among His- was conducted to determine the racial distribution panics in the southwestern United States and their among samples sequenced within The Cancer Ge- potential impact on genetic cancer risk assessment. nome Atlas (TCGA) and the number of samples still Deleterious BRCA mutations were detected in 25% of needed to detect select mutational frequencies in ra- 746 samples. A total of 11% of 189 samples were large cial minorities. Of the 5,729 samples analyzed, 77% rearrangement mutations, of which 62% were BRCA1 were white, 12% were black, 3% were Asian, 3% were exon 9 to exon 12 deletions. Nine recurrent mutations Hispanic, and less than 0.5% were Native Americans. accounted for 53% of the total. Among these, BRCA1 These percentages, thus, over-represent whites com- exon 9 to exon 12 deletions appear be a Mexican

October 2016, Vol. 23, No. 4 Cancer Control 461 founder mutation representing up to 12% of all BRCA1 women and for lung and prostate cancers in men. mutations in select cohorts. The authors conclude that The racial gap in death rates has widened for breast the high frequency of large rearrangement mutations cancer in women and remained stable for colorectal warrants screening in every case. cancer in men. In men, incidence rates from 2003 to 2012 decreased for all cancers combined as well as for Meyskens FL Jr, Mukhtar H, Rock CL, et al. Cancer the top 3 cancer sites. In women, overall rates during prevention: obstacles, challenges and the road the corresponding time period remained unchanged, ahead. J Natl Cancer Inst. 2015;108(2):pii:djv309. reflecting increasing trends in breast cancer combined The genetic heterogeneity and complexity of ad- with decreasing trends in lung and colorectal cancer vanced cancers support rationale for interrupting the rates. The 5-year relative survival rate is lower for carcinogenic process early and target prevention to blacks than whites for most cancers at each stage of reduce the burden of cancer; however, the focus of diagnosis. The extent to which these disparities reflect cancer prevention should be on persons at high risk unequal access to health care vs other factors remains and on primary localized disease, for which screening an active area of research. and detection should also play a role. The timing and doses of chemopreventive interventions also affect Ward E, Jemal A, Cokkinides V, et al. Cancer dispari- response. The intervention may be ineffective if the ties by race/ethnicity and socioeconomic status. target population is very high risk or presents with CA Cancer J Clin. 2004;54(2):78-93. preneoplastic lesions and irreversible cellular changes. The authors sought to identify disparities in can- The field must begin to focus on targeted organ-site cer incidence, mortality, and survival rates in rela- prevention approaches in patients at high risk. The tion to race and ethnicity. For all cancer sites com- authors also suggest that comparative effectiveness re- bined, residents of poorer counties had 13% higher search designs and the value of information obtained death rates from cancer in men and 3% higher rates from large-scale prevention studies are necessary so in women when compared with individuals living in that preventive interventions become a routine part more affluent counties. Among both sexes, the 5-year of cancer management. survival rate for all cancers combined was 10% lower among persons who live in poorer rather than more Carlson RW, Scavone JL, Koh WJ, et al. NCCN frame- affluent census tracts. When the census tract poverty work for resource stratification: a framework for rate was accounted for, African American, American providing and improving global quality oncology Indian, Alaskan Native, and Asian/Pacific Islander men care. J Natl Compr Canc Netw. 2016;14(8):961-969. and African American, American Indian, and Alaskan This article describes the rationale for resource- Native women had lower 5-year survival rates than stratified guidelines and the methodology for devel- non-Hispanic whites. Detailed analyses of select can- oping the National Comprehensive Cancer Network cers showed large variations in rate of cancer survival (NCCN) Framework. Disparities in available resources by race and ethnicity. for cancer care are enormous, which is one reason why the NCCN developed a framework for stratifying Quinn GP, Sanchez JA, Sutton SK, et al. Cancer and its clinical practice for cancer. NCCN also hopes to lesbian, gay, bisexual, transgender/transsexual, help health care systems with varying levels of avail- and queer/questioning (LGBTQ) populations. able resources provide optimal care for patients with CA Cancer J Clin. 2015;65(5):384-400. cancer. The framework is modified from a method The authors review the literature on 7 cancer developed by the Breast Health Global Initiative. The sites that may disproportionately affect lesbian, gay, NCCN Framework for Resource Stratification identifies bisexual, transgender/trans-sexual, and queer/ques- 4 resource environments and presents the recommen- tioning populations. For each cancer site, descrip- dations in a graphic format that maintains the context tive statistics, primary prevention, secondary preven- of the NCCN guidelines. tion and preclinical disease, tertiary prevention and late-stage disease, and clinical implications are dis- DeSantis CE, Siegel RL, Sauer AG, et al. Cancer cussed. The authors also provide an overview of psy- statistics for African Americans, 2016: progress chosocial factors related to cancer survivorship as well and opportunities in reducing racial disparities. as strategies for improving access to care. CA Cancer J Clin. 2016;66(4):290-308. Approximately 189,910 new cases of cancer and 69,410 cancer-related deaths will occur among blacks in 2016. Although blacks continue to have higher rates of cancer-related mortality than whites, the disparity has narrowed for all cancers combined in men and

462 Cancer Control October 2016, Vol. 23, No. 4 Peer Reviewers, 2016

Peer review is the indispensable and critical element in the evaluation of all manuscripts being considered for publication in our journal. The oncology professionals who took part in our peer review process this past year are listed below, and we thank them for their invaluable efforts on behalf of Cancer Control. — The Editor

External Reviewers Deneve, Jeremiah L., MD Lichtman, Stuart, MD University of Tennessee Health Memorial Sloan Kettering Abate, Getahun, MD, PhD Science Center Saint Louis University Luque, John, MD Dickson, Paxton, MD Medical University of South Carolina Alberg, Anthony J., PhD, MPH University of Tennessee Health Medical University of South Carolina Sciences Center Mahtani, Reshma L., MD University of Miami Health System Alcaraz, Kassandra, PhD, MPH Drucker, Mitchell D., MD American Cancer Society University of South Florida Maluccio, Mary A., MD, MPH Indiana University School of Medicine Ahmad, Nazeel, MD Eisbruch, Avraham, MD James A. Haley Veterans Administration University of Michigan Margo, Curtis, MD Medical Center University of South Florida Friedman, Scott, MD Allott, Emma H., PhD Mount Sinai Hospital Margolies, Laurie, MD University of North Carolina at Chapel Hill Mount Sinai School of Medicine Glas, Martin, MD Alrabaa, Sally, MBBS Robert Janker Clinic Montero, Jose, MD University of South Florida University of Bonn Medical Center University of South Florida Apata, Jummai, MBBS, MPH Glezerman, Ilya G., MD Moore, Charles E., MD Morgan State University Memorial Sloan Kettering Emory University School of Medicine Cancer Center Bao, Renyue, MD Muro-Cacho, Carlos, MD University of Houston Goovaerts, Pierre, PhD James A. Haley Veterans Hospital BioMedware Biermann, Martin, MD Nevidjon, Brenda, MSN, RN University of Bergen, Norway Hara, Wendy, MD Oncology Nursing Society Stanford University Bowen, Deborah J., PhD Nguyen, Jennifer, PhD, MPH University of Washington Harbour, J. William, MD University of Florida Bascom Palmer Eye Institute Briere, Elizabeth C., MD, MPH Oehler, Richard, MD Centers for Disease Control and Hassanein, Ashraf, MD, PhD James A. Haley Veterans Prevention Florida Dermatologic Surgery Administration Hospital & Aesthetics Institute Brierley, James, D., MB Florida Pathology Pentz, Rebecca D., PhD Princess Margaret Cancer Centre Emory School of Medicine Hernandez, Brenda Y., PhD, MPH Bull, Joan M.C., MD University of Hawaii Phuphanich, Surasak, MD University of Texas Health Science Cen- Cedars-Sinai Medical Center ter at Houston Jatoi, Aminah, MD Mayo Clinic Razis, Evangelia, MD, PhD Chevez-Barrios, Patricia, MD Hygeia Hospital Houston Methodist Jones, Lovell A., PhD Prairie View A&M University College Rigakos George, MD, MSc Chowdary, Sajeel, MD of Nursing Hygeia Hospital Boca Raton Regional Hospital Kamen, Charles, PhD Rowe, Christopher, MD Climent, Consuelo, MD University of Rochester Medical Center University of Newcastle University of Puerto Rico Kanji, Jamil, MD Rusch, Valerie W., MD Colon-Otero, Gerardo, MD University of Alberta Memorial Sloan Kettering Mayo Clinic Cancer Center

Lavand’homme, Patricia, MD, PhD Dagan, Roi, MD Université Catholique de Louvain Saeed, Maythem, MD University of Florida University of California at San Francisco LeRuhn, Emilie, MD DeGregorio, Richard, MD University Hospital and Oscar Sheikhattari, Payam, MD, MPH Surgical Pathology Laboratories Lambret Center Morgan State University

October 2016, Vol. 23, No. 4 Cancer Control 463 Peer Reviewers, 2016

Simmons, Melissa A., MD Northwestern University

Sioutos, Panayiotis, MD Hygeia Hospital

Sniffin, Jason, DO Florida Hospital Orlando Infectious Disease Consultants

Tang, Haidong, PhD University of Texas Southwestern Medical Center

Thompson, Haley, PhD Barbara Ann Karmanos Cancer Institute Wayne State University

Tyson, Mark D., MD Mayo Clinic

Vander Poorten, Vincent, MD Leuven Cancer Institute

Wayman, Laura L., MD Vanderbilt University

Yom, Sue S., MD, PhD University of California at San Francisco

Zhang, Paul J., MD University of Pennsylvania

Internal Reviewers (H. Lee Moffitt Cancer Center & Research Institute) Ahmad, Nazeel, MD Arrington, John, MD Balducci, Lodovico, MD Caracciolo, Jamie, MD Caudell, Jimmy J., MD, PhD Chon, Hye Sook, MD Clark, Terry J., MScA Coppola, Domenico, MD Dessureault, Sophie, MD Dhillon, Jasreman, MD Dossett, Lesly, MD Green, B. Lee, PhD Greene, John N., MD Gwede, Clement K., PhD Harrison, Louis B., MD Hudson, Janella, PhD Kumar, Nagi B., PhD Laber, Damian A., MD Lancet, Jeffrey, MD Makanji, Rikesh, MD Mullinax, John E., MD Nanjappa, Sowmya, MD Niell, Bethany, MD, PhD Otto, Kristen, MD Pasikhova, Yanina, PharmD Pow-Sang, Julio M., MD Quinn, Gwendolyn P., PhD So, Wonhee, PharmD Sutton, Steve K., PhD Trotti, Andy, MD Wang, Xia, MD Zager, Jonathan S., MD

464 Cancer Control October 2016, Vol. 23, No. 4 Index for 2016, Volume 23

61 Special Report: Genetic Investigation of Vol. 23, No. 1: Regional Therapeutics Uterine Carcinosarcoma: Case Report and Cohort Analysis 3 Art, Science, and Personalized Cancer Care Timothy N. Hembree, DO, PhD, Jamie K. Teer, PhD, Through Regional Therapeutics Ardeshir Hakam, MD, and Alberto A. Chiappori, MD Kristen Otto, MD, and Jonathan S. Zager, MD 67 Special Report: Advances in the Endoscopic 6 Stereotactic Body Radiotherapy for Diagnosis of Barrett Esophagus Recurrent Unresectable Head and Neck Ashley H. Davis-Yadley, MD, Kevin G. Neill, MD, Cancers Mokenge P. Malafa, MD, and Luis R. Peña, MD Tobin Strom, MD, Christian Wishka, and Jimmy J. Caudell, MD, PhD 78 Special Report: Association of microRNA 21 With Biological Features and Prognosis of 12 Developments in Intralesional Therapy for Neuroblastoma Metastatic Melanoma Yaodong Zhou, MD, and Bo Sheng, MD Sarah Sloot, MD, Omar M. Rashid, MD, JD, Amod A. Sarnaik, MD, and Jonathan S. Zager, MD

21 Stereotactic Body Radiotherapy in Vol. 23, No. 2: Ocular Oncology the Management of Oligometastatic Disease 90 Are Ocular and Ocular Adnexal Cancers Kamran A. Ahmed, MD, and Javier F. Torres-Roca, MD Overdiagnosed? Historical Perspectives on Diagnosis 30 Local Ablation for Solid Tumor Liver Curtis E. Margo, MD Metastases: Techniques and Treatment Efficacy 93 Assessing Prognosis in Uveal Melanoma Joyce Wong, MD, and Amanda Cooper, MD Zélia M. Corrêa, MD, PhD

36 Hyperthermic Intraperitoneal Chemotherapy 99 Therapeutic Options for Retinoblastoma and Cytoreductive Surgery in the Management Pia R. Mendoza, MD, and Hans E. Grossniklaus, MD of Peritoneal Carcinomatosis Rahul Rajeev, MBBS, and Kiran K. Turaga, MD 110 Primary Vitreoretinal Lymphoma: Management of Isolated Ocular Disease 47 Disparities Report: Racial Disparity in Time Matthew T. Witmer, MD Between First Diagnosis and Initial Treatment of Prostate Cancer 117 Management of Primary Acquired Melanosis, Ballington L. Kinlock, PhD, Roland J. Thorpe Jr, PhD, Nevus, and Conjunctival Melanoma Daniel L. Howard, PhD, Janice V. Bowie, PhD, Louie E. Andrew Kao, MD, Armin Afshar, MD, Michele Bloomer, MD, Ross, PhD, David O. Fakunle, and Thomas A. LaVeist, PhD and Bertil Damato, MD, PhD

52 Disparities Report: Exploring the Perceptions 126 Sebaceous Carcinoma of the Eyelid of Anal Cancer Screening and Behaviors Carlos Prieto-Granada, MD, and Paul Rodriguez-Waitkus, Among Gay and Bisexual Men Infected MD, PhD With HIV Alexis M. Koskan, PhD, Natalie LeBlanc, MPH, and 133 Role of Vismodegib in the Management of Isabella Rosa-Cunha, MD Advanced Periocular Basal Cell Carcinoma Kyle F. Cox, MD, and Curtis E. Margo, MD 59 Case Report: Hyperuricemia in 2 Patients Receiving Palbociclib for Breast Cancer 140 Extranodal Marginal Zone B-cell Lymphoma David J. Bromberg, MD, Mauricio Valenzuela, MD, of the Ocular Adnexa Sowmya Nanjappa, MD, and Smitha Pabbathi, MD Jean Guffey Johnson, MD, Lauren A. Terpak, MS, Curtis E. Margo, MD, and Reza Setoodeh, MD

October 2016, Vol. 23, No. 4 Cancer Control 465 Index for 2016, Volume 23

150 Ophthalmic Complications Related to 213 Clinical Benefits of Proton Beam Therapy for Chemotherapy in Medically Complex Patients Tumors of the Skull Base Lynn E. Harman, MD Kamran A. Ahmed, MD, Stephanie K. Demetriou, Mark McDonald, MD, and Peter A.S. Johnstone, MD 157 Disparities Report: Disparities Among Minority Women With Breast Cancer Living in 220 Management of BCC and SCC of the Impoverished Areas of California Head and Neck Sundus Haji-Jama, Kevin M. Gorey, PhD, Isaac N. Tobin J. Strom, MD, Jimmy J. Caudell, MD, PhD, and Luginaah, PhD, Guangyong Zou, PhD, Caroline Hamm, MD, Louis B. Harrison, MD and Eric J. Holowaty, MD 228 Treatment of Head and Neck 163 Infectious Disease Report: Bordetella pertussis Paragangliomas Infection in Patients With Cancer Kenneth Hu, MD, and Mark S. Persky, MD Abraham Yacoub, MD, Sowmya Nanjappa, MD, Tyler Janz, and John N. Greene, MD 242 Multidisciplinary Management of Salivary Gland Cancers 167 Pharmacy Report: Megestrol Acetate-Induced Matthew J. Mifsud, MD, Jon N. Burton, MD, Adrenal Insufficiency Andy M. Trotti, MD, and Tapan A. Padhya, MD Sowmya Nanjappa, MD, Christy Thai, PharmD, Sweta Shah, MD, and Matthew Snyder, PharmD 249 Pathology Report: Practical Issues for Retroperitoneal Sarcoma 170 Pathology Report: Presacral Noncommuni- Vicky Pham, MS, Evita Henderson-Jackson, MD, cating Enteric Duplication Cyst Matthew P. Doepker, MD, Jamie T. Caracciolo, MD, Shabnam Seydafkan, MD, David Shibata, MD, Ricardo J. Gonzalez, MD, Mihaela Druta, MD, Yi Ding, MD, Julian Sanchez, MD, Nam D. Tran, MD, Marino Leon, MD, and Marilyn M. Bui, MD, PhD and Domenico Coppola, MD 265 Infectious Diseases Report: Primary 175 Symposium Report: Second Annual Moffitt Gangrenous Cutaneous Mold Anatomic Pathology Symposium Infections in a Patient With Cancer and Neutropenia Abraham Yacoub, MD, Kiran K. Soni, Lysenia Mojica, MD, Jane Mai, MD, Jamie Morano, MD, C. Wayne Cruse, MD, Vol. 23, No. 3: Head and Neck Cancers Ramon L. Sandin, MD, Sowmya Nanjappa, MBBS, MD, Chandrashekar Bohra, MBBS, Ganesh Gajanan, MBBS, 192 Maximizing Cures and Preserving Function and John N. Greene, MD in Head and Neck Cancers Louis B. Harrison, MD, and Jimmy J. Caudell, MD, PhD 272 Case Series: Diffuse Alveolar Hemorrhage in Acute Myeloid Leukemia 197 Management of Oropharyngeal Cancer in the Sowmya Nanjappa, MBBS, MD, Daniel K. Jeong, MD, HPV Era Manjunath Muddaraju, MD, Katherine Jeong, MD, Arash O. Naghavi, MD, Tobin J. Strom, MD, Kamran A. Eboné D. Hill, MD, and John N. Greene, MD Ahmed, MD, Michelle I. Echevarria, MD, Yazan A. Abuodeh, MD, Puja S. Venkat, MD, Jessica M. Frakes, MD, 278 Case Series: Marijuana Smoking in Patients Louis B. Harrison, MD, Andy M. Trotti, MD, and Jimmy J. With Leukemia Caudell, MD, PhD Sara Khwaja, MD, Abraham Yacoub, MD, Asima Cheema, MD, Nancy Rihana, MD, Robin Russo, MD, 208 Definitive Radiotherapy for Squamous Cell Ana Paula Velez, MD, Sowmya Nanjappa, MBBS, MD, Carcinoma of the Glottic Larynx Ramon L. Sandin, MD, Chandrashekar Bohra, MBBS, William M. Mendenhall, MD, Roi Dagan, MD, Curtis M. Ganesh Gajanan, MBBS, and John N. Greene, MD Bryant, MD, Robert J. Amdur, MD, and Anthony A. Mancuso, MD

466 Cancer Control October 2016, Vol. 23, No. 4 Index for 2016, Volume 23

284 Special Report: Platelet-to-Lymphocyte Ratio 338 Racial and Ethnic Differences in the and Use of NSAIDs During the Perioperative Epidemiology and Genomics of Lung Cancer Period as Prognostic Indicators in Patients Matthew B. Schabath, PhD, W. Douglas Cress, PhD, With NSCLC Undergoing Surgery and Teresita Muñoz-Antonia, PhD Brenda M. Lee, MD, Andrea Rodríguez, MD, Gabriel Mena, MD, Vijaya Gottumukkala, MB, MD, Reza J. 347 Black Heterogeneity in Cancer Mortality: Mehran, MD, David C. Rice, MB, Lei Feng, MS, US-Blacks, Haitians, and Jamaicans Jun Yu, MS, and Juan P. Cata, MD Paulo S. Pinheiro, MD, PhD, Karen E. Callahan, MPH, Camille Ragin, PhD, MPH, Robert W. Hage, MD, PhD, DLO, 295 Special Report: Resection of the First Rib MBA, Tara Hylton, MPH, and Erin N. Kobetz, PhD, MPH With Preservation of the T1 Nerve Root in Pancoast Tumors of the Lung 359 Genomic Disparities in Breast Cancer Andreas K. Filis, MD, Lary A. Robinson, MD, and Among Latinas Frank D. Vrionis, MD, PhD Filipa Lynce, MD, Kristi D. Graves, PhD, Lina Jandorf, MA, Charité Ricker, MS, Eida Castro, PsyD, Laura Moreno, 302 Special Report: Feasibility, Preliminary Bianca Augusto, Laura Fejerman, PhD, and Susan T. Efficacy, and Lessons Learned From a Vadaparampil, PhD Garden-Based Lifestyle Intervention for Cancer Survivors 373 Clinical Considerations of Risk, Incidence, and Colleen K. Spees, PhD, RDN, Emily B. Hill, Elizabeth M. Outcomes of Breast Cancer in Sexual Minorities Grainger, PhD, RDN, Jackie L. Buell, PhD, RDN, Anne E. Mattingly, MD, John V. Kiluk, MD, and Susan E. White, PhD, Matthew D. Kleinhenz, PhD, and M. Catherine Lee, MD Steven K. Clinton, MD, PhD 383 Construction and Validation of a Multi- 311 Special Report: Hypofractionated Stereotactic Institutional Tissue Microarray of Invasive Radiotherapy for Auditory Canal or Middle Ductal Carcinoma From Racially and Ear Cancer Ethnically Diverse Populations Taro Murai, MD, Shin-Etsu Kamata, MD, Kengo Sato, MD, Edward Seijo, MS, Diana Lima, MPH, Egiebade Iriabho, MS, Kouki Miura, MD, Mitsuhiro Inoue, Naoki Yokota, MD, Jonas Almeida, PhD, Jesus Monico, MPH, MS, Seiji Ohta, MD, Michio Iwabuchi, MD, Hiromitsu Iwata, MD, Margarita Echeverri, PhD, Sylvia Gutierrez, MD, and Yuta Shibamoto, MD Idhaliz Flores, PhD, Ji-Hyun Lee, PhD, Kate Fisher, MA, William E. Grizzle, MD, PhD, Gabriel L. Sica, MD, PhD, Charles Butler, Chindo Hicks, PhD, Cathy D. Meade, PhD, RN, Stephen Olufemi Sodeke, Krzysztof Moroz, MD, Vol. 23, No. 4 Supplement: Lung Cancer Domenico Coppola, MD, and Teresita Muñoz-Antonia, PhD

2 Best Practices in Treatment Selection for 390 Determinants of Treatment Delays Among Patients With Advanced NSCLC 2 Underserved Hispanics With Lung and Head Mark A. Socinski, MD, and Nathan Pennell, MD, PhD and Neck Cancers Evelinn A. Borrayo, PhD, Katie L. Scott, PhD, Ava R. Drennen, PhD, Tiare MacDonald, PhD, and Jennifer Nguyen, MS Vol. 23, No. 4: Health Disparities in Oncology 401 Geographical Factors Associated With 327 Barriers to Clinical Trial Enrollment in Racial Health Disparities in Prostate Cancer and Ethnic Minority Patients With Cancer Scott M. Gilbert, MD, Julio M. Pow-Sang, MD, and Lauren M. Hamel, PhD, Louis A. Penner, PhD, Hong Xiao, PhD Terrance L. Albrecht, PhD, Elisabeth Heath, MD, Clement K. Gwede, PhD, RN, and Susan Eggly, PhD 409 Disparities in Penile Cancer Pranav Sharma, MD, Kamran Zargar-Shoshtari, MD, Curtis A. Pettaway, MD, Matthew B. Schabath, PhD, Anna R. Giuliano, PhD, and Philippe E. Spiess, MD

October 2016, Vol. 23, No. 4 Cancer Control 467 Index for 2016, Volume 23

415 Chemoprevention in African American Men With Prostate Cancer Nagi B. Kumar, PhD, RD, Julio M. Pow-Sang, MD, Philippe E. Spiess, MD, Jong Y. Park, PhD, Ganna Chornokur, PhD, Andrew R. Leone, MD, and Catherine M. Phelan, PhD, MD

424 Disparities in Adolescents and Young Adults With Cancer Leidy L. Isenalumhe, MD, Olivia Fridgen, MPH, Lynda K. Beaupin, MD, Gwendolyn P. Quinn, PhD, and Damon R. Reed, MD

434 Tobacco-Related Health Disparities Across the Cancer Care Continuum Vani Nath Simmons, PhD, Bárbara Piñeiro, PhD, Monica Webb Hooper, PhD, Jhanelle E. Gray, MD, and Thomas H. Brandon, PhD

442 Case Series: Human Metapneumovirus Infection in Immunocompromised Patients Sharmeen Samuel, MD, Sowmya Nanjappa, MBBS, MD, Christopher D. Cooper, MD, and John N. Greene, MD

446 Special Report: Systematic Review and Case Series Report of Acinar Cell Carcinoma of the Pancreas Evan S. Glazer, MD, PhD, Kevin G. Neill, MD, Jessica M. Frakes, MD, Domenico Coppola, MD, Pamela J. Hodul, MD, Sara E. Hoffe, MD, Jose M. Pimiento, MD, Gregory M. Springett, MD, PhD, and Mokenge P. Malafa, MD, PhD

455 Special Report: A Single-Institution Study of Demographics and Outcomes of Adult Patients With Multiple Cancers Thanh P. Ho, MD, Nathan R. Foster, MS, and Aminah Jatoi, MD

468 Cancer Control October 2016, Vol. 23, No. 4 Current Perspectives in Oncology Nursing

Save the Date! | Call for Abstracts February 16–17, 2017 Moffitt Cancer Center, Tampa, FL

16th Annual Nursing Conference

Special Event The 22nd Annual Rhinehart Reception & Lecture | February 15, 2017

Moffitt Cancer Center Vincent A. Stabile Research Building 12902 Magnolia Drive, Tampa, FL 33612

For additional conference information please contact ( Gina Woodward at 813-745-1741 or [email protected] (