The Role of KRAS Rs61764370 in Invasive Epithelial Ovarian Cancer: Implications for Clinical Testing

Published OnlineFirst March 8, 2011; DOI: 10.1158/1078-0432.CCR-10-3405

Clinical Cancer Imaging, Diagnosis, Prognosis Research

The Role of KRAS rs61764370 in Invasive Epithelial Ovarian Cancer: Implications for Clinical Testing

Paul D. P. Pharoah1, Rachel T. Palmieri3, Susan J. Ramus6, Simon A. Gayther6, Irene L. Andrulis9, Hoda Anton-Culver12, Natalia Antonenkova13, Antonis C. Antoniou2, David Goldgar for the BCFR Investigators14, Mary S. Beattie15, Matthias W. Beckmann16, Michael J. Birrer18, Natalia Bogdanova20,21, Kelly L. Bolton22, Wendy Brewster23, Angela Brooks-Wilson24, Robert Brown26, Ralf Butzow28,29, Trinidad Caldes30, Maria Adelaide Caligo32, Ian Campbell33,34, Jenny Chang-Claude35, Y. Ann Chen36, Linda S. Cook38, Fergus J. Couch39, Daniel W. Cramer19, Julie M. Cunningham40, Evelyn Despierre43, Jennifer A. Doherty44, Thilo Dork€ 20, Matthias Durst€ 45, Diana M. Eccles46, Arif B. Ekici17, Douglas Easton for the EMBRACE Investigators47, Peter A. Fasching7, Anna de Fazio48,49, David A. Fenstermacher36, James M. Flanagan26, Brooke L. Fridley41, Eitan Friedman50, Bo Gao48,49, Olga Sinilnikova for the GEMO Study Collaborators51, Aleksandra Gentry- Maharaj26, Andrew K. Godwin52, Ellen L. Goode42, Marc T. Goodman53, Jenny Gross8, Thomas V. O. Hansen54, Paul Harnett49, Matti Rookus for the HEBON Investigators59, Tuomas Heikkinen28, Rebecca Hein35, Claus Høgdall55, Estrid Høgdall56,57, Edwin S. Iversen5, Anna Jakubowska60, Sharon E. Johnatty61, Beth Y. Karlan8, Noah D. Kauff62, Stanley B. Kaye65, Georgia Chenevix-Trench for the kConFab Investigators66 and the Consortium of Investigators of Modifiers of BRCA1/261, Linda E. Kelemen67, Lambertus A. Kiemeney68, Susanne Kruger€ Kjaer55,56, Diether Lambrechts70, James P. LaPolla71, Conxi Lazaro 72, Nhu D. Le25, Arto Leminen28, Karin Leunen43, Douglas A. Levine63,YiLu61, Lene Lundvall55, Stuart Macgregor61, Tamara Marees68, Leon F. Massuger69, John R. McLaughlin10, Usha Menon27, Marco Montagna73, Kirsten B. Moysich74, Steven A. Narod11, Katherine L. Nathanson75, Lotte Nedergaard58, Roberta B. Ness76, Heli Nevanlinna28, Stefan Nickels35, Ana Osorio31, Jim Paul77, Celeste Leigh Pearce6, Catherine M. Phelan36, Malcolm C. Pike6,64, Paolo Radice78,79, Mary Anne Rossing44, Joellen M. Schildkraut3, Thomas A. Sellers36, Christian F. Singer80, Honglin Song1, Daniel O. Stram6, Rebecca Sutphen37, Annika Lindblom for the SWE-BRCA Investigators81, Kathryn L. Terry19, Ya-Yu Tsai36, Anne M. van Altena69, Ignace Vergote43, Robert A. Vierkant41, Allison F. Vitonis19, Christine Walsh8, Shan Wang-Gohrke82, Barbara Wappenschmidt83, Anna H. Wu6, Argyrios Ziogas12, Andrew Berchuck4 and Harvey A. Risch84 for the Ovarian Cancer Association Consortium

Abstract Purpose: An assay for the single-nucleotide polymorphism (SNP), rs61764370, has recently been commercially marketed as a clinical test to aid ovarian cancer risk evaluation in women with family histories of the disease. rs67164370 is in a 30-UTR miRNA binding site of the KRAS oncogene and is a candidate for epithelial ovarian cancer (EOC) susceptibility. However, only one published article, analyzing fewer than 1,000 subjects in total, has examined this association.

Authors' Affiliations: 1Department of Oncology; 2Centre for Cancer Genetics, National Cancer Institute, NIH, Bethesda, Maryland; 23Depart- Genetic Epidemiology, Department of Public Health and Primary Care, ment of Obstetrics and Gynecology, School of Medicine, University of University of Cambridge, Cambridge, United Kingdom; 3Departments of North Carolina at Chapel Hill, North Carolina; 24Genome Sciences Centre; Community and Family Medicine, and 4Obstetrics and Gynecology, Duke 25Cancer Control Research, British Columbia Cancer Agency, Vancou- University Medical Center; 5Department of Statistical Science, Duke ver, British Columbia, Canada; 26Epigenetics Unit, Department of Surgery University, Durham, North Carolina; 6Department of Preventive Medicine, and Cancer, Imperial College London; 27Department of Gynaecological Keck School of Medicine and the USC Norris Comprehensive Cancer Oncology, University College London, EGA Institute for Women's Health, Center, University of Southern California; 7Division of Hematology and London, United Kingdom; 28Department of Obstetrics and Gynecology, Oncology, Department of Medicine, David Geffen School of Medicine, Helsinki University Central Hospital; 29Department of Pathology, University University of California at Los Angeles; 8Women's Cancer Research of Helsinki, Helsinki, Finland; 30Molecular Oncology Laboratory, Hospital Institute at the Samual Oschin Comprehensive Cancer Institute, Clínico San Carlos; 31Human Genetics Group, Human Cancer Genetics Cedars-Sinai Medical Center, Los Angeles, California; 9Ontario Cancer Programme, Spanish National Cancer Centre, Madrid, Spain; 32Section of Genetics Network, Cancer Care Ontario and Samuel Lunenfeld Research Genetic Oncology, University Hospital of Pisa, Pisa, Italy; 33Centre for Cancer Institute, Mount Sinai Hospital; 10Cancer Care Ontario; 11Women's Col- Genomics and Predictive Medicine, Peter MacCallum Cancer Centre, Mel- lege Research Institute, University of Toronto, Toronto, Ontario, Canada; bourne; 34Department of Pathology, University of Melbourne, Parkville, 12Department of Epidemiology, School of Medicine, University of Cali- Victoria, Australia; 35Division of Cancer Epidemiology, German Cancer fornia, Irvine, California; 13Byelorussian Institute for Oncology and Med- Research Center (DKFZ), Heidelberg, Germany; 36H. Lee Moffitt Cancer ical Radiology Aleksandrov N.N., Minsk, Belarus; 14Breast Cancer Family Center and Research Institute; 37Department of Pediatrics, University of Registry, Epidemiology and Genetics Research Program, DCCPS, South Florida College of Medicine, Tampa, Florida; 38Division of Epidemiol- National Cancer Institute, Rockville, Maryland; 15Cancer Risk Program, ogy and Biostatistics, Department of Internal Medicine, University of New Departments of Medicine, Epidemiology, and Biostatistics, University Mexico, Albuquerque, New Mexico; 39Department of Laboratory Medicine of California at San Francisco, San Francisco, California; 16Department of and Pathology, Mayo Clinic; 40Genomics Shared Resource, Department of Gynecology and Obstetrics, University Hospital Erlangen; 17Institute of Laboratory Medicine and Pathology; 41Department of Health Sciences Human Genetics, Friedrich-Alexander University Erlangen-Nuremberg, Research Divisions of Biostatistics and 42Epidemiology, Mayo Clinic College Erlangen, Bavaria, Germany; 18Department of Medicine, Massachusetts of Medicine, Rochester, Minnesota; 43Division of Gynaecologic Oncology, General Hospital; 19Obstetrics and Gynecology Epidemiology Center, Department of Obstetrics and Gynaecology, University Hospitals Leuven, Brigham and Women's Hospital, Boston, Massachusetts; 20Clinics of University of Leuven, Leuven, Belgium; 44Program in Epidemiology, Fred Obstetrics and Gynaecology and 21Radiation Oncology, Hannover Med- Hutchinson Cancer Research Center, Seattle, Washington; 45Clinic of ical School, Hannover, Germany; 22Division of Cancer Epidemiology and ObstetricsandGynaecology,FriedrichSchillerUniversity,Jena, Germany;

3742 Clin Cancer Res; 17(11) June 1, 2011

KRAS rs61764370 in Epithelial Ovarian Cancer

Experimental Design: Risk association was evaluated in 8,669 cases of invasive EOC and 10,012 controls from 19 studies participating in the Ovarian Cancer Association Consortium, and in 683 cases and 2,044 controls carrying BRCA1 mutations from studies in the Consortium of Investigators of Modifiers of BRCA1/2. Prognosis association was also examined in a subset of five studies with progression-free survival (PFS) data and 18 studies with all-cause mortality data. Results: No evidence of association was observed between genotype and risk of unselected EOC (OR ¼ 1.02, 95% CI: 0.95–1.10), serous EOC (OR ¼ 1.08, 95% CI: 0.98–1.18), familial EOC (OR ¼ 1.09, 95% CI: 0.78–1.54), or among women carrying deleterious mutations in BRCA1 (OR ¼ 1.09, 95% CI: 0.88–1.36). There was little evidence for association with survival time among unselected cases (HR ¼ 1.10, 95% CI: 0.99–1.22), among serous cases (HR ¼ 1.12, 95% CI ¼ 0.99–1.28), or with PFS in 540 cases treated with carboplatin and paclitaxel (HR ¼ 1.18, 95% CI: 0.93–1.52). Conclusions: These data exclude the possibility of an association between rs61764370 and a clinically significant risk of ovarian cancer or of familial ovarian cancer. Use of this SNP for ovarian cancer clinical risk prediction, therefore, seems unwarranted. Clin Cancer Res; 17(11); 3742–50. 2011 AACR.

Introduction component of disease risk (2–4). Many candidate gene studies have identified possible common ovarian cancer Epithelial ovarian cancer (EOC) is the fifth-most com- susceptibility alleles, but most are likely to represent false- mon cancer in women. It generally presents as advanced positive associations as none have been reported at the disease with poor prognosis. Family and twin studies have levels of statistical significance required, when testing suggested that inherited genetic variation plays an appreci- hypotheses with low prior probabilities of association (1). able part in determining individual risk. However, until In July 2010, a single-nucleotide polymorphism (SNP), 0 recently, knowledge of genetic susceptibility was limited to rs61764370, located in the 3 -UTR of the KRAS oncogene, rare, highly penetrant alleles in a handful of genes includ- was reported to be associated with risk of unselected EOC ing BRCA1, BRCA2 and the mismatch repair genes (1). In (5). The variant was also reported to be associated with the past 2 years, genome-wide association studies have a stronger risk in women carrying BRCA1 mutations, identified common susceptibility alleles at 4 loci at highly in women not carrying BRCA1 or BRCA2 mutations stringent levels of statistical significance (P < 10 8), but but with a family history of the disease, as well as associated these alleles have small effects on disease risk (per-allele with poorer progression-free survival (PFS: ref. 5). This SNP OR < 1.3) and explain a small fraction of the genetic was thought to be a strong candidate for cancer risk as it

46Wessex Clinical Genetics Service, Princess Anne Hospital, Southamp- tistics and HTA, and 69Gynecology, Radboud University Nijmegen Medical ton, United Kingdom; 47Centre for Cancer Genetic Epidemiology, Depart- Centre, Nijmegen, The Netherlands; 70Vesalius Research Center (VIB) and ment of Public Health and Primary Care, University of Cambridge, University of Leuven, Belgium; 71Bayfront Medical Center Obstetrics and Strangeways Research Laboratory, Worts Causeway, Cambridge, United Gynecology Residency Program and Women's Cancer Associates, St. Kingdom; 48Department of Gynaecological Oncology, Westmead Hospi- Petersburg, Florida; 72Hereditary Cancer Program, Instituto Catalan de tal; 49Westmead Institute for Cancer Research, University of Sydney at Oncología, Barcelona, Spain; 73Immunology and Molecular Oncology Westmead Millennium Institute, Westmead Hospital, Westmead, New Unit, Istituto Oncologico Veneto IOV–IRCCS, Padua, Italy; 74Department South Wales, Australia; 50Oncogenetics Unit, Sheba Medical Center, of Cancer Prevention & Control, Roswell Park Cancer Institute, Buffalo, Tel Hashomer, and Sackler School of Medicine, Tel Aviv University, Tel New York; 75Department of Medicine, Medical Genetics and Abramson Aviv, Israel; 51Cancer Genetics Network "Groupe Gen etique et Cancer", Cancer Center, University of Pennsylvania School of Medicine, Philadel- Fed eration Nationale des Centres de Lutte Contre le Cancer, Lyon, France; phia, Pennsylvania; 76University of Texas School of Public Health, Hous- 52Department of Pathology and Laboratory Medicine, University of Kansas ton, Texas; 77Cancer Research UK Clinical Trials Unit, University of Medical Center, Kansas City, Kansas; 53Cancer Research Center of Glasgow, Glasgow, Scotland; 78Unit of Molecular Bases of Genetic Risk Hawaii, University of Hawaii, Honolulu, Hawaii; 54Genomic Medicine, and Genetic Testing, Department of Preventive and Predictive Medicine, Department of Clinical Biochemistry, Rigshospitalet, Copenhagen Uni- Fondazione IRCCS Istituto Nazionale Tumori (INT); 79IFOM, Fondazione versity Hospital; 55Gynecologic Clinic, The Juliane Marie Centre, Rigshos- Istituto FIRC di Oncologia Molecolare, Milan, Italy; 80Division of Special pitalet, University of Copenhagen; 56Department of Viruses, Hormones Gynecology, Medical University of Vienna, Vienna, Austria; 81Karolinska and Cancer, Institute of Cancer Epidemiology, Danish Cancer Society; University Hospital, Stockholm, Sweden; 82Department of Obstetrics and 57Department of Pathology, Molecular Unit, Herlev Hospital; 58Department Gynecology, University of Ulm, Ulm, Germany; 83Center of Familial Breast of Pathology, Rigshospitalet, University of Copenhagen, Copenhagen, and Ovarian Cancer, Department of Obstetrics and Gynaecology and Denmark; 59Department of Epidemiology, The Netherlands Cancer Insti- Center for Integrated Oncology (CIO), University of Cologne, Cologne, tute, Amsterdam, The Netherlands; 60International Hereditary Cancer Germany; 84Department of Epidemiology and Public Health, Yale Uni- Center, Department of Genetics and Pathology, Pomeranian Medical versity School of Medicine, New Haven, Connecticut University, Szczecin, Poland; 61The Queensland Institute of Medical 62 Corresponding Author: Research, Brisbane, Queensland, Australia; Clinical Genetics Service; Harvey A. Risch, Department of Epidemiology Departments of 63Surgery, and 64Epidemiology and Biostatistics, Memor- and Public Health, Yale University School of Public Health, School of ial Sloan-Kettering Cancer Center, New York, New York; 65Section of Medicine, 60 College Street, New Haven, CT 06520-8034. Phone: 203- Medicine, Institute of Cancer Research, Sutton, United Kingdom; 66Peter 785-2848; Fax: 203-785-4497; E-mail: [email protected] 67 MacCallum Cancer Centre, East Melbourne, Victoria, Australia; Depart- doi: 10.1158/1078-0432.CCR-10-3405 ment of Population Health Research, Alberta Health Services-Cancer Care, Calgary, Alberta, Canada; 68Departments of Epidemiology, Biosta- 2011 American Association for Cancer Research.

www.aacrjournals.org Clin Cancer Res; 17(11) June 1, 2011 3743

Pharoah et al.

Translational Relevance

An assay for a single-nucleotide polymorphism (SNP) in a 30-UTR miRNA binding site of the KRAS gene has recently been commercially marketed as a clinical test to aid epithelial ovarian cancer (EOC) risk

assessment in women with family histories of the method disease. The justification for use of this assay was based on one published article which analyzed fewer than 1,000 subjects in total, including only 67 EOC cases 350 Fluidigm 195259 Fluidigm Fluidigm carrying BRCA1 mutations or with family histories of 173137 Fluidigm 698 Illumina Hap317 213 TaqMan Fluidigm EOC. The present article found no association between this SNP and ovarian cancer risk among 8,669 cases of unselected invasive EOC and 10,012 controls, or in 683 cases and 2,044 controls carrying BRCA1 mutations.TheresultssuggestthatevaluationofthisSNPis Number of subjectsControls Genotyping Cases not clinically useful for risk prediction in sporadic or familial ovarian cancer.

lies in a miRNA binding site, and associations between miRNA mutations or misexpression and risk of some human cancers have been seen. These observations sug- Population based 434 Hospital basedHospital based 151 426 Study type gested that miRNAs can function as tumor suppressors Hospital basedPopulation based 253 142 Population based 265 or oncogenes (6). An assay to determine genotype at rs61764370 has subsequently been marketed as a commercial test to determine risk in women with a family history of ovarian cancer (http://www.miradx.com). How- ever, as with other candidate gene studies, the reported association was not at a level of statistical significance that is regarded as definitive for common susceptibility alleles (7), nor was the magnitude of risk sufficient for this SNP Study population to be acceptable as a useful clinical marker of ovarian Belgium Washington StateGermany Population based 721 b b cancer risk. The present work therefore sought to achieve a the following: (i) replicate the association in a robust (Continued on the following page) DOV HOC Finland BEL Study abbrev. GER manner in multiple study populations, genotyped to a high standard with stringent quality assurance procedures; (ii) assess the association between genotype at this locus and ovarian cancer risk in women with family histories or who carry deleterious mutations in BRCA1; and (iii) examine the hypothesis that the SNP is associated with differences in postdiagnosis PFS or all-cause mortality. Materials and Methods

Study populations Minsk Ovarian Cancer Study HMO Belarus Jena Ovarian Cancer Study HJO Germany – Nineteen ovarian cancer case–control sets and 1 case – series participating in the Ovarian Cancer Association

Consortium (OCAC), and 1 additional case series, con- Evaluation Study Diseases of the Ovary and their Hannover Hannover Helsinki Ovarian Cancer Study Belgium Ovarian Cancer Study Study name tributed data to the analyses (Table 1). Three of the case– Brigham Women's Hospital StudyGerman Ovarian BWH Cancer Study Boston control sets were each comprised of a case series matched to controls from the same geographic region: PVM, UK2, and Description of participating studies UK-GWAS (genome-wide association studies). Survival

time analysis was based on data from 18 case series, control sets including the additional publicly available data for 359 – BWH DOV HMO HJO HOC BEL Case Table 1. Analysis set ovarian cancer cases from The Cancer Genome Atlas GER (http://cancergenome.nih.gov/) that had information on

3744 Clin Cancer Res; 17(11) June 1, 2011 Clinical Cancer Research

Downloaded from Analysis Study name Study Study Study type Number of subjects Genotyping set abbrev. population method Controls Cases

HOP Hormones and Ovarian Cancer HOPb Pittsburgh Population based 368 365 TaqMan Prediction Study MAY Mayo Clinic Ovarian Cancer Study MAYa Upper Midwest Hospital based 520 358 Illumina 610 Quad Published OnlineFirstMarch8,2011;DOI:10.1158/1078-0432.CCR-10-3405

clincancerres.aacrjournals.org NCO North Carolina Ovarian Cancer Study NCOb North Carolina Population based 655 494 Illumina 610 Quad NTH Nijmegen Ovarian Cancer Study NTH Netherlands Population based 327 296 Fluidigm OVA Ovarian Cancer Study OVA Alberta and British Population based 416 494 Fluidigm Columbia, Canada PVM Pelvic Mass Study PVDb Copenhagen, Denmark Hospital based 201 Fluidigm Malignant Ovarian Cancer Study MAL Copenhagen, Denmark Population based 215 Fluidigm TBO Tampa Bay Ovarian Cancer Study TBOb Tampa Population based 168 227 Illumina 610 Quad TOR Familial Ovarian Tumour Study TOR Ontario, Canada Population based 556 734 Illumina 610 Quad Cancer Research. UCI UC Irvine Ovarian Cancer Study UCIb Southern California Population based 372 192 Fluidigm b on September 25, 2021.©2011AmericanAssociation for SEARCH SEA UK Population based 670 Fluidigm Southampton Ovarian Cancer Study SOCb UK Population based 389 Fluidigm UK2 Scottish Randomized Trial in Ovarian SROa UK Hospital based 866 Fluidigm Cancer United Kingdom Ovarian Cancer UKOb UK Population based 655 Fluidigm Population Study SEARCH SEAb UK Population based 1,089 Illumina 610 Quad United Kingdom Ovarian Cancer UKOb UK Population based 500 Illumina 610 Quad KRAS Population Study UK-GWAS Cancer Research UK Familial Ovarian UCRb UK Familial cancer register 32 Illumina 610 Quad Cancer Register Cancer Ovarian Epithelial in rs61764370 lnCne e;1(1 ue1 2011 1, June 17(11) Res; Cancer Clin Royal Marsden Hospital Study RMH UK Hospital based 147 Illumina 610 Quad UK 1958 Birth cohort 58BC UK Cohort 1,438 Illumina Hap550 UK Colorectal control NSCR UK Population based 917 Illumina Hap550 USC Los Angeles County USC Los Angeles Population based 343 260 TaqMan Case–Control Studies of Ovarian Cancer Case-only studies LAX Case series LAXa Los Angeles Hospital based 263 Fluidigm TCGA The Cancer Genome Atlas TCGAa USA Hospital based 359 Illumina 1M Duo

aCase data included in analysis of all-cause mortality and PFS. bCase data included in analysis of all-cause mortality. 3745 Published OnlineFirst March 8, 2011; DOI: 10.1158/1078-0432.CCR-10-3405

Pharoah et al.

all-cause mortality. The analysis of PFS was based on data Statistical analyses from 5 case series. Finally, data from 683 cases and 2,044 Analyses were restricted to white non-Hispanic women controls enrolled in a stage I project of the Consortium of based on self-reported ethnic origin for all of the studies, Investigators of Modifiers of BRCA1/2 (see ref. 8 for details with the exception of the TCGA, MAY, NCO, TOR, TBO, of studies participating in CIMBA) were used to examine and UK-GWAS controls. For these studies, genome-wide risk among women carrying deleterious BRCA1 mutations. genotype data were used to estimate intercontinental Each study was approved by a governing research ethics ancestry and women of less than 90% European ancestry committee and all study subjects provided written were excluded (see methods in ref. 3 for details). Cases informed consent. Clinical and questionnaire data with borderline (low malignant potential) EOC were included tumor behavior, histology, stage and grade, age also excluded, as were 22 cases from TCGA that had been at diagnosis (or at comparable date for controls), family provided to TCGA as part of the MAY case–control history of ovarian cancer, and ethnicity/race. study. Departure of genotype frequencies from those Survival time data were available for cases from 18 expected under Hardy–Weinberg equilibrium was asses- studies (BEL, DOV, UCR, GER, HOP, LAX, MAY, NCO, sed by using a c2 test for each study that was directly PVD, RMH, SEA, SOC, SRO, TBO, TCGA, UCI, UKO, and genotyped. The association between SNP and disease USC) and clinical information on chemotherapy, residual risk was evaluated by using unconditional logistic regres- disease after surgery, and time to progression was collected sion in which number of copies of the minor (infrequent) in 5 studies (BEL, LAX, MAY, SRO, and TCGA). All of the allele was treated as a continuous variable. This provides women included in the analysis of PFS had at least 4 cycles an estimate of the per-allele OR and 95% CI. Models of carboplatin and paclitaxel as part of primary treatment. adjusted for age categories (<40, 40–49, 50–59, 60–69, PFS was defined as the time interval between the date of and 70 years) were also considered. Each case–control histologic diagnosis and the first confirmed sign of disease set was analyzed individually and the pooled result was recurrence or progression (9). obtained by combining the log odds-ratios by using standard inverse variance-weighted meta-analytic meth- Genotyping ods. Analysis of the BRCA1 mutation carrier cohort was Genotyping of 13 case–control sets was done in a single carried out by using a time-to-event analysis framework laboratory by using a 50-nuclease TaqMan allelic discrimi- that models the association between genotype and ovar- nation assay (Applied Biosystems) as part of a 96-SNP ian cancer risk as a HR. Because mutation carriers were Fluidigm multiplex (10 studies) or–with the same batch not sampled randomly with respect to their disease status, of TaqMan reagents–using the 7900HT Sequence Detection standard methods of survival analysis may lead to biased Software (Applied Biosystems; 3 studies; Table 1). Details estimates of associations. Therefore, analyses were car- of quality control criteria of OCAC have been described ried out by modeling the retrospective likelihood of previously (10); they include genotyping of a common set observed genotype conditional on disease phenotype of 95 DNAs (90 CEPH trios and 5 duplicate samples) and (seerefs.8,12fordetails). comparison with the genotypes for the same samples as Associations between genotype and PFS and all-cause reported by HapMap. However, rs61764370 was not mortality were evaluated by using proportional hazards genotyped in the HapMap project. On the basis of regression. Because the EOC cases were recruited at variable sequence data for 57 individuals of European origin from times after diagnosis, regression analysis of all-cause mor- the 1000 Genomes project (http://www.1000genomes. tality allowed for left truncation, with time at risk starting org), a HapMap SNP, rs17388148, was found to be on date of diagnosis and time under observation beginning strongly correlated (r2 ¼ 0.97) with rs61764370. The at the time of study entry. This method generates an concordance between the CEPH trio genotype data for unbiased estimate of the HR, provided that the propor- rs61764370 and the HapMap genotypes for rs17388148 tional hazards assumption is reasonably correct (13). was 100%. Therefore, data on rs17388148 were obtained Cause-specific mortality was not available for most studies, from 3 GWAS, in which cases and controls had been so the analysis of all-cause mortality was right censored at 5 genotyped by using Illumina genome-wide SNP arrays years after diagnosis to minimize the proportion of deaths (refs. 1, 2, 4, 8; Table 1). Neither rs61764370 nor from causes other than ovarian cancer. The analysis of PFS rs17388148 were included on the Illumina arrays used was adjusted for stage and residual disease, and survival in these GWAS studies, but imputed genotypes were time ended at time of progression or was censored at time available for rs17388148. These genotypes were provided of last follow-up. as the estimated number of rare alleles carried (0 to 2 on a continuous scale). The accuracy of the imputation as Results calculated by the program MACH of Li and Abecasis (11) for the North American studies (BWH, MAY, Details of the 19 case–control sets used in our analyses NCO, TBO, and TOR) was r2 ¼ 0.977. This high accuracy are given in Table 1. Genotype data from these sets were of the imputation was evidently because of the presence available for 8,669 cases and 10,012 controls (Table 2). All of a nearby SNP (rs12305513, 17 kb away) in high LD (r2 studies passed the OCAC criteria for genotyping quality. ¼ 1 in HapMap) with rs17388148. Genotype frequencies were close to those expected under

3746 Clin Cancer Res; 17(11) June 1, 2011 Clinical Cancer Research

KRAS rs61764370 in Epithelial Ovarian Cancer

Table 2. Genotype frequency numbers of controls and cases by study

Study abbreviation Control genotypesa Case genotypesa

0 1 2 Total Allele freq.b 0 1 2 Total Allele freq.b

BEL 202 49 2 253 0.105 142 28 3 173 0.098 DOV 610 104 7 721 0.082 570 123 5 698 0.095 GER 218 46 1 265 0.091 184 28 1 213 0.070 HJO 354 66 6 426 0.092 217 42 0 259 0.081 HMO 130 19 2 151 0.076 174 20 1 195 0.056 HOC 390 43 1 434 0.052 321 28 1 350 0.043 HOP 297 67 4 368 0.102 308 54 3 365 0.082 LAX 223 39 1 263 0.078 NTH 275 47 5 327 0.087 249 46 1 296 0.081 OVA 332 83 1 416 0.102 417 74 3 494 0.081 PVM 181 33 1 215 0.081 166 34 1 201 0.090 UCI 318 52 2 372 0.075 150 41 1 192 0.112 UK2 1,087 231 7 1,325 0.092 1,039 201 15 1,255 0.092 USC 281 59 3 343 0.095 214 45 1 260 0.090 BWHc 118 23 1 142 0.088 115 21 1 137 0.084 MAYc 438 78 4 520 0.083 283 71 4 358 0.110 NCOc 537 112 6 655 0.095 398 91 5 494 0.102 TBOc 141 26 1 168 0.083 181 43 3 227 0.108 TCGAc 286 69 4 359 0.107 TORc 458 93 5 556 0.093 587 139 8 734 0.106 UK-GWASc 1,923 410 22 2,355 0.096 1,446 306 16 1,768 0.096 Total 8,282 1,649 81 10,012 7,969 1,241 81 9,291

aColumns labeled 0, 1, and 2 refer to common homozygote, heterozygote, and infrequent homozygote, respectively. b Estimated frequency of minor (infrequent) allele: (f2 þ ½ f1)/(f2 þ f1 þ f0). cEstimated genotype frequencies based on imputed data.

Hardy–Weinberg equilibrium in both cases and controls were carriers of BRCA1 mutations (HR ¼ 1.09, 95% CI: for the 13 directly genotyped studies. No evidence was 0.88–1.36, P ¼ 0.40). found for association between rs61764370 and invasive Survival time data were available for 6,002 cases of the EOC in univariate analysis (OR ¼ 1.02, 95% CI: 0.95–1.10, 6,826 total from 18 case series, including 13,696 person- P ¼ 0.44), with minimal heterogeneity of risk between years at risk and 2,044 deaths. Little evidence of association studies (P ¼ 0.28). Study-specific ORs are shown in Figure was observed between rs61764370 genotype and all-cause 1A. When studies with directly genotyped data and with mortality within 5 years of diagnosis (HR ¼ 1.10, 95% CI: imputed data were analyzed separately, the overall OR in 0.99–1.22, P ¼ 0.08), with no evidence of heterogeneity of the genotyped studies was 0.96 (95% CI: 0.87–1.06, P ¼ the HR between studies (P ¼ 0.89). Results of analyses 0.42) compared with 1.08 (95% CI: 0.97–1.20, P ¼ 0.15) restricted to serous subtype (HR ¼ 1.12, 95% CI: 0.99– in the imputed data studies. Adjusting for age at diagnosis/ 1.28, P ¼ 0.08) or adjusted for tumor stage and grade (HR interview made little difference to the results (data ¼ 1.06, 95% CI: 0.96–1.18, P ¼ 0.27) were similar. There not shown). No differences in risk were observed when was also little evidence for association between genotype analyses were restricted to cases who provided blood sam- and PFS in 540 high-grade serous cases known to have been ples within 18 months of diagnosis (n ¼ 6,550, OR ¼ 1.02, treated with carboplatin and paclitaxel (HR ¼ 1.18, 95% 95% CI: 0.93–1.10, P ¼ 0.72), cases with serous tumors CI: 0.93–1.52, P ¼ 0.16). (Fig. 1B, n ¼ 4,706, OR ¼ 1.08, 95% CI: 0.98–1.18, P ¼ 0.11), or cases reporting a family history of ovarian Discussion cancer in a first-degree relative (from 6 studies; Fig. 1C, n ¼ 249, OR ¼ 1.09, 95% CI: 0.78–1.54, P ¼ 0.62). Tests for This study provides no evidence to support the pre- heterogeneity of risk between studies in all analyses were not viously reported associations between rs61764370 and risk statistically significant. A similar result was seen for risk of of EOC. The relative risk given by Ratner and colleagues for EOC by genotype among 683 cases and 2,044 controls who their replication data set of unselected cases was 1.70 (95%

www.aacrjournals.org Clin Cancer Res; 17(11) June 1, 2011 3747

Pharoah et al.

CI: 1.11–2.63, P ¼ 0.016; ref. 5). The CIs in the present A study do not overlap with these. The power of the present Study OR (95% CI) study to detect an allele conferring a relative risk of 1.3 under either a dominant or log-additive genetic model with HMO 0.74 (0.41–1.33) 4 BWH 0.96 (0.54–1.70) a type 1 error rate of 10 was greater than 99%, strongly PVM 1.11 (0.68–1.83) suggesting that the association observed by Ratner and GER 0.75 (0.46–1.22) colleagues was a chance finding or possibly because of TBO 1.27 (0.79–2.06) subtle genotyping errors. This observation is not surprising. HOC 0.82 (0.51–1.31) Associations with modest P values that are declared as BEL 0.93 (0.59–1.46) positive are very likely to be false positives, when the prior UCI 1.57 (1.03–2.41) probability of association is low (14). In genetic association USC 0.95 (0.64–1.41) NTH 0.93 (0.62–1.37) studies, even if evidence exists that a variant has functional HJO 0.88 (0.59–1.30) effects, the prior probability of association at a relative risk HOP 0.79 (0.56–1.13) of 1.5 is unlikely to be more than 1:100 and, given that only OVA 0.76 (0.55–1.06) a handful of loci conferring relative risks of more than 1.5 MAY 1.42 (1.02–1.96) have been found for any cancer, the prior probability is NCO 1.10 (0.83–1.45) likely to be much less. On the basis of the methodology of TOR 1.17 (0.90–1.52) DOV 1.18 (0.91–1.53) Wacholder and colleagues (14), assuming a prior of 1:100, UK2 0.99 (0.82–1.20) the probability that the association reported by Ratner and UK–GWAS 0.99 (0.85–1.16) colleagues was a false positive is 86%. Under more likely, Overall 1.02 (0.95–1.10) smaller prior probabilities, the false positive probability will approach 100%. For familial ovarian cancer, the power

0.5 1 2 4 of the present study to detect a relative risk of 2.0 was over B 95% at a type 1 error rate of 0.05, again suggesting that it is unlikely that the present analyses have missed a true GER 0.52 (0.25–1.06) association of this magnitude with familial ovarian cancer. HMO 1.06 (0.53–2.13) Statistical power to detect an allele with weaker effects is HJO 1.08 (0.58–2.02) HOC 1.01 (0.54–1.88) more limited. For example, if the true relative risk were 1.15 BWH 0.92 (0.52–1.65) under a codominant genetic model–similar to the relative NTH 0.93 (0.53–1.62) risk conferred by an allele with the biggest risk detected in a PVM 1.04 (0.60–1.80) recent well-powered lung-cancer GWAS (15)–power at a BEL 0.91 (0.55–1.53) type 1 error rate of 0.001 would be reduced to 70%. A very TBO 1.56 (0.94–2.60) modest association between the rs61764370 variant and a UCI 1.62 (0.98–2.68) slightly increased risk of ovarian cancer cannot, therefore, USC 1.24 (0.79–1.95) HOP 0.77 (0.50–1.18) be excluded. Neither has the present study shown much OVA 0.81 (0.55–1.19) evidence for association between genotype at this locus and MAY 1.44 (0.99–2.09) other related phenotypes: EOC risk in carriers of deleter- NCO 1.08 (0.77–1.50) ious BRCA1 mutations or in women with family histories TOR 1.29 (0.95–1.75) of EOC, all-cause mortality after diagnosis of EOC, and PFS DOV 1.30 (0.97–1.74) in EOC cases treated with carboplatin and paclitaxel. UK2 0.96 (0.73–1.28) In summary, the possibility that the minor (infrequent) UK–GWAS 1.01 (0.83–1.22) Overall 1.08 (0.98–1.18) allele of rs61764340 is associated with an appreciable risk of EOC is excluded. Furthermore, it is debatable whether a single-risk allele, even one conferring a relative risk as high 0.5 1 2 4 as 2.0, has clinical utility, particularly in a disease with C lifetime risk as low as it is in invasive EOC (1 in 70), and even among women with a family history of the disease UCI 1.26 (0.29–5.53) MAY 0.73 (0.17–3.20) (16). The marketing of a commercial assay for rs61764370 NCO 1.05 (0.36–3.05) for clinical use among unselected cases, or women with HOP 1.52 (0.57–4.05) family histories of ovarian cancer or who are carriers of DOV 1.46 (0.74–2.86) BRCA1 mutations, in the absence of convincing evidence UK–GWAS 0.87 (0.52–1.46) for association, is not warranted. Overall 1.09 (0.78–1.54)

0.5 1 2 4 OR Figure 1. Funnel plots of study-specific ORs for association between rs61764370 and ovarian cancer risk. A, all cases; B, serous cases; and C, cases with a family history of ovarian cancer in a first-degree relative.

3748 Clin Cancer Res; 17(11) June 1, 2011 Clinical Cancer Research

KRAS rs61764370 in Epithelial Ovarian Cancer

Disclosure of Potential Conflicts of Interest the support of Translational Research (BEL); NIH grants 5R01-CA054419 and 5P50-CA105009 (BWH); NIH grants 5R01-CA112523 and 5R01-CA087538 (DOV); the Ligue National Contre le Cancer, the Association "Le cancer du sein, parlons- The content of this article does not necessarily reflect the views or policies en!" and research grant AICR-07-0454 from the Association for International of the National Cancer Institute or any of the collaborating centers in the Cancer Research (GEMO); the German Federal Ministry of Education and BCFR, nor does mention of trade names, commercial products, or organiza- Research, the Programme of Clinical Biomedical Research grant 01 GB 9401, tions imply endorsement by the US Government or the BCFR. the state of Baden-Wurttemberg€ through the University of Ulm Medical Faculty grant P.685, and the German Cancer Research Center (GER); the Helsinki Acknowledgments University Central Hospital Research Fund, Academy of Finland and the Finnish Cancer Society (HOC); NIH grants 5R01-CA095023, 5R01-CA126841, and 5R03- CA121881, US Army MRMC grant DAMD 17-02-1-0669, and grants from the We thank all the individuals who took part in this study and all the Ovarian Cancer Research Foundation and the Roswell Park Alliance Foundation researchers, clinicians, and administrative staff who have enabled and carried (HOP); grant 628333 from Cancer Australia and grants from the National Breast out the many studies contributing to this work. In particular, we thank: E. Cancer Foundation, the National Health and Medical Research Council John, J.L. Hopper, S.S. Buys, M.B. Daly, M.B. Terry, and M. Southey (BCFR); S. (NHMRC), the Queensland Cancer Fund, the Cancer Councils of New South Manoukian, B. Peissel, and M.A. Pierotti of the Fondazione IRCCS Istituto Wales, Victoria, Tasmania and South Australia, and the Cancer Foundation of Nazionale dei Tumori, Milan, Italy; P. Peterlongo of the Fondazione Istituto Western Australia (kConFab); American Cancer Society Early Detection Pro- FIRC di Oncologia Molecolare, Milan, Italy; M. Barile and B. Bonanni of the fessorship grant SIOP-06-258-01-CCE and the L & S Milken Foundation (LAX); Istituto Europeo di Oncologia, Milan, Italy; L. Bernard of the Consortium for Mermaid 1, The Danish Cancer Society and the National Cancer Institute 5R01- Genomic Technologies, Milan, Italy; and A. Viel of the Centro di Riferimento CA061107 (MAL); National Cancer Institute 5R01-CA122443 (MAY); National Oncologico, Aviano, Italy (CONSIT); all of the GEMO study collaborating Cancer Institute 5R01-CA076016 (NCO); an Investment Grant of the Radboud groups for their contributions (GEMO); A.U. Eilber and T. Koehler (GER); L. University Nijmegen Medical Centre (NTH); grants from the Canadian Institutes € Gacucova (HMO); P. Schurmann, F. Kramer, W. Zheng, T-W. Park-Simon, P. of Health Research (CIHR) and the Workers’ Compensation Board of British Hillemanns, K. Beer-Grondke, D. Schmidt, and I. Runnebaum (HJO); A. Columbia (OVA); Cancer Research UK (SEA); Experimental Cancer Research Teule, G. Capella, I. Blanco, J. Brunet, L. Feliubadalo, and M. Salinas (ICO); E. Centres and Cancer Research UK A6689 (SRO); NIH grant 5R01-CA106414, Niedermeyr and H. Thorne (kConFab); I. Tomlinson and R. Houlston American Cancer Society grant CRTG-00-196-01-CCE and an Advanced Cancer (NSCR); the Nijmegen Biomedical Study for sharing questionnaire data Detection Center Grant, Department of Defense grant DAMD 17-98-1-8659, and and DNA from controls and the gynecologists in the catchment area of the funding from the Celma Mastry Ovarian Cancer Foundation (TBO); Canadian Comprehensive Cancer Center East for recruiting patients (NTH); P. Har- Institutes for Health Research, the National Cancer Institute of Canada, the rington and the SEARCH team (SEA); Ian Jacobs, Eva Wozniak, Andy Ryan, Canadian Cancer Society and the NIH grants 5R01-CA063682 and 5R01- Jeremy Ford, and Nayala Balogun (UKO). We thank the University of Cam- CA063678 (TOR); National Cancer Institute grants 5R01-CA058860 and 5K07- bridge (CamGrid) for providing computational resources. This study in part CA092044 and the Lon V Smith Foundation LVS-39420 (UCI); Cancer Research makes use of data generated by the Wellcome Trust Case-Control Consor- UK, the Eve Appeal, the OAK Foundation and the Department of Health's NIHR tium. A full list of the investigators who contributed to the generation of those Biomedical Research Centre funding scheme (UKO); the California Cancer data is available at http://www.wtccc.org.uk. The results published here are Research Program grants 00-01389V-20170 and 2110200, NIH grants 5P30- also in part based on data generated by The Cancer Genome Atlas Pilot Project CA014089, 5P01-CA017054, 5R01-CA061132, 5R01-CA063464, N01-PC67010, established by the National Cancer Institute and National Human Genome and 5R03-CA113148, and California Department of Health Services sub-contract Research Institute. Information about TCGA and the investigators and 050-E8709 (USC); Wellcome Trust award 076113 (WTCCC). Additional support institutions constituting the TCGA research network can be found at for the TOR, NCO, MAY, TBO, and NCI studies was provided by NIH grant 5R01- http://cancergenome.nih.gov/. We thank the patients who donated tissues CA114343. The Consortium of Investigators of Modifiers of BRCA1/2 GWAS and for research use, the tissue source sites, and the TCGA Research Network. its participating studies were funded by: the Breast Cancer Research Founda- Samples used in the BCFR from the NCCC, FCCC, and HCI were processed tion, Susan G. Komen for the Cure, the Ovarian Cancer Research Fund, the and distributed by the Coriell Institute for Medical Research. Eileen Stein Jacoby Fund, NIH grants 5R01-CA128978 and 1R01-CA140323, Cancer Research UK, the Neve Foundation, Project Hope for Ovarian Cancer Research and Education, the Australian National Breast Cancer Foundation, the Grant Support National Health and Medical Research Council of Australia. The costs of publication of this article were defrayed in part by the pay- The Ovarian Cancer Association Consortium is supported by a grant from ment of page charges. This article must therefore be hereby marked advertise- the Ovarian Cancer Research Fund. Funding for the individual participating ment in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. studies was provided by: cooperative agreements with the Breast Cancer Family Registry including Cancer Care Ontario, and US NIH grants 5U01-CA069467, 5U01-CA069398, 5U01-CA069631, 5U01-CA069446, 5U01-CA069417, 5U01- Received December 23, 2010; accepted January 19, 2011; published CA069638, and N02-PC45022 (BCFR); the National Cancer Plan–Action 29 for OnlineFirst March 8, 2011.

References 1. Gayther SA, Pharoah PD. The inherited genetics of ovarian and 8. Antoniou AC, Wang X, Fredericksen ZS, McGuffog L, Tarrell R, endometrial cancer. Curr Opin Genet Dev 2010;20:231–8. Sinilnikova OM, et al. A locus on 19p13 modifies risk of breast cancer 2. Song H, Ramus SJ, Tyrer J, Bolton KL, Gentry-Maharaj A, Wozniak in BRCA1 mutation carriers and is associated with hormone receptor- E, et al. A genome-wide association study identifies a new ovarian negative breast cancer in the general population. Nat Genet cancer susceptibility locus on 9p22.2. Nat Genet 2009;41: 2010;42:885–92. 996–1000. 9. Johnatty SE, Beesley J, Paul J, Fereday S, Spurdle AB, Webb PM, 3. Bolton KL, Tyrer J, Song H, Ramus SJ, Notaridou M, Jones C, et al. et al. ABCB1 (MDR 1) polymorphisms and progression-free survival Common variants at 19p13 are associated with susceptibility to among women with ovarian cancer following paclitaxel/carboplatin ovarian cancer. Nat Genet 2010;42:880–4. chemotherapy. Clin Cancer Res 2008;14:5594–601. 4. Goode EL, Chenevix-Trench G, Song H, Ramus SJ, Notaridou M, 10. Pearce CL, Near AM, Van Den Berg DJ, Ramus SJ, Gentry-Maharaj A, Lawrenson K, et al. A genome-wide association study identifies Menon U, et al. Validating genetic risk associations for ovarian cancer susceptibility loci for ovarian cancer at 2q31 and 8q24. Nat Genet through the international Ovarian Cancer Association Consortium. Br 2010;42:874–9. J Cancer 2009;100:412–20. 5. Ratner E, Lu L, Boeke M, Barnett R, Nallur S, Chin LJ, et al. A KRAS- 11. Scott LJ, Mohlke KL, Bonnycastle LL, Willer CJ, Li Y, Duren WL, et al. variant in ovarian cancer acts as a genetic marker of cancer risk. A genome-wide association study of type 2 diabetes in Finns detects Cancer Res 2010;70:6509–15. multiple susceptibility variants. Science 2007;316:1341–5. 6. Esquela-Kerscher A, Slack FJ. Oncomirs–microRNAs with a role in 12. Antoniou AC, Sinilnikova OM, Simard J, Leone M, Dumont M, Neu- cancer. Nat Rev Cancer 2006;6:259–69. hausen SL, et al. RAD51 135G!C modifies breast cancer risk among 7. Thomas DC, Clayton DG. Betting odds and genetic associations. J BRCA2 mutation carriers: results from a combined analysis of 19 Natl Cancer Inst 2004;96:421–3. studies. Am J Hum Genet 2007;81:1186–200.

www.aacrjournals.org Clin Cancer Res; 17(11) June 1, 2011 3749

Pharoah et al.

13. Azzato EM, Greenberg D, Shah M, Blows F, Driver KE, Caporaso NE, 15. Truong T, Hung RJ, Amos CI, Wu X, Bickeboller€ H, Rosenberger A, et al. Prevalent cases in observational studies of cancer survival: do et al. Replication of lung cancer susceptibility loci at chromosomes they bias hazard ratio estimates? Br J Cancer 2009;100:1806–11. 15q25, 5p15, and 6p21: a pooled analysis from the International Lung 14. Wacholder S, Chanock S, Garcia-Closas M, El Ghormli L, Rothman N. Cancer Consortium. J Natl Cancer Inst 2010;102:959–71. Assessing the probability that a positive report is false: an approach for 16. Vineis P, Schulte P, McMichael AJ. Misconceptions about the use of molecular epidemiology studies. J Natl Cancer Inst 2004;96:434–42. genetic tests in populations. Lancet 2001;357:709–12.

3750 Clin Cancer Res; 17(11) June 1, 2011 Clinical Cancer Research

The Role of KRAS rs61764370 in Invasive Epithelial Ovarian Cancer: Implications for Clinical Testing

Paul D. P. Pharoah, Rachel T. Palmieri, Susan J. Ramus, et al.

Clin Cancer Res 2011;17:3742-3750. Published OnlineFirst March 8, 2011.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-10-3405

Cited articles This article cites 16 articles, 3 of which you can access for free at: http://clincancerres.aacrjournals.org/content/17/11/3742.full#ref-list-1

Citing articles This article has been cited by 6 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/17/11/3742.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/17/11/3742. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.