AMPI2019-25Print-1UP 1..135

SYMPTOMS AND SURVIVORSHIP

11500 Oral Abstract Session, Mon, 8:00 AM-11:00 AM

A phase II RCT of high-dose vitamin D supplementation and exercise for cancer treatment- induced bone loss in breast cancer patients on aromatase inhibitors.

Luke Joseph Peppone, Jennifer E Reschke, Michelle Christine Janelsins, Julia Ellen Inglis, Karen Michelle Mustian, Eva Culakova, Amber Kleckner, Charles Stewart Kamen, Po-Ju Lin, Richard Francis Dunne, Ian Kleckner, Alissa Huston, Michelle Shayne, Gary R. Morrow; University of Rochester Medical Center, Rochester, NY; University of Rochester James P. Wilmot Cancer Institute, Strong Memorial Hospital, Rochester, NY; Univ of Rochester, Rochester, NY

Background: Cancer-treatment-induced bone loss (CTIBL) is a side effect of aromatase inhibitors (AIs) and can result in osteoporotic fractures. Vitamin D (VITD) protects against postmenopausal bone loss but it is unclear if the recommended daily allowance (RDA: 600 IU/day) of VITD is sufficient to prevent CTIBL. This phase II RCT aimed to assess the feasibility, safety, and preliminary efficacy of high-dose VITD (with and without exercise) on bone mineral density (BMD) compared to the RDA. Methods: Non-metastatic breast cancer patients starting AIs with low VITD (,32 ng/ml) were randomized 1:1:1 into 3 arms: 1) placebo 2) high-dose VITD (50,000 IU/week) or 3) high-dose VITD + Exercise for Cancer Patients (EXCAP): a home-based, personalized walking and resistance band training program for 24 weeks. All subjects received the RDA of VITD 600 IU/day. Serum VITD and calcium levels were assessed at baseline, weeks 6, 12, 18, and 24. BMD was assessed at the hip via DXA at baseline and week 24. Results: Of the 116 subjects randomized (mean age = 60; 94% white; mean baseline VITD = 24.6 ng/mL), 90 provided fully evaluable data. Compliance ($ 80% of instructed doses) exceeded 95% in all 3 arms with no between-group difference. ANCOVA showed significant differences between groups on final VITD levels (high-dose = 63.6 vs high-dose + EXCAP = 60.3 vs placebo = 32.0 ng/mL; p,0.001) without severe calcium toxicities, as indicated by final calcium level (high-dose = 9.4 vs high-dose + EXCAP = 9.5 vs placebo = 9.4 ng/mL; p = 0.78). The placebo group lost a significant amount of hip BMD (21.7%; p , 0.01) while hip BMD was maintained in the high-dose (20.1%; p = 0.77) and high-dose + EXCAP (20.2%; p = 0.74) resulting in significant between-group differences for high-dose + EXCAP vs placebo (p = 0.04) and high-dose vs placebo (p = 0.05). Conclusions: This is one of the first studies to show our novel high-dose VITD intervention, with and without exercise, significantly reduced hip BMD loss in breast cancer patients on AIs. Moreover, high-dose VITD supplementation is safe and feasible in this population. A phase III RCT is needed to confirm these findings. Funding: K07CA168911. Clinical trial information: NCT01419730.

11501 Oral Abstract Session, Mon, 8:00 AM-11:00 AM

A randomized trial comparing four-weekly versus 12-weekly administration of bone- targeted agents (denosumab, zoledronate, or pamidronate) in patients with bone metastases from either breast or castration-resistant prostate cancer.

Mark J. Clemons, Michael Ong, Carol Stober, D. Scott Ernst, Christopher M. Booth, Christina M. Canil, Mihaela Mates, Andrew George Robinson, Phillip S. Blanchette, Anil Abraham Joy, John Frederick Hilton, Olexiy Aseyev, Gregory Russell Pond, Brian Hutton, Ahwon Jeong, Lisa Vandermeer, Dean Fergusson; Division of Medical Oncology, Department of Medicine, The Ottawa Hospital and University of Ottawa, Ottawa, ON, Canada; The Ottawa Hospital Cancer Centre,Ottawa,ON,Canada;Cancer Research Group, The Ottawa Hospital Research Institute and the University of Ottawa,Ottawa,ON,Canada;Divisionof Medical Oncology, Department of Oncology, London Regional Cancer Program, London Health Sciences Centre and University of Western Ontario,, London, ON, Canada; Cancer Centre of Southeastern Ontario, Kingston, ON, Canada; Department of Oncology, University of Alberta, Cross Cancer Institute, Edmonton, AB, Canada; Regional Cancer Centre, Thunder Bay Regional Health Sciences Centre, Northern Ontario School of Medicine, Thunder Bay, ON, Canada; McMaster University, Department of Oncology, Hamilton, ON, Canada; Department of Epidemiology and Community Medicine, The Ottawa Hospital Research Institute and the University of Ottawa, Ottawa, ON, Canada; Clinical Epidemiology Program, The Ottawa Hospital Research Institute and University of Ottawa, Ottawa, ON, Canada

Background: Defining the optimal dosing interval of commonly used bone-targeted agents (BTAs), such as denosumab and bisphosphonates, for patients with bone metastases remains an important clinical question. We performed a pragmatic randomised trial comparing the non-inferiority of 12- versus 4- weekly BTAs in patients with bone metastases from breast and prostate cancer. Methods: Patients with bone metastases, who were either BTA-na¨ıve, or already receiving, denosumab, pamidronate or zoledronate were eligible. They were randomised to receive their chosen BTA every 12- or 4-weeks for one year. The primary endpoint was Health related quality of life (HRQL) (EORTC-QLQ-C30 Functional Domain - Physical Subdomain). Secondary endpoints included: pain (EORTC-QLQ-BM22 - pain domain), Global Health Status (EORTC-QLQ-C30), symptomatic skeletal events (SSE) rates and time to SSEs. Adverse events and toxicity profiles were also compared. Results: Of 263 patients (60.8% breast and 39.2% prostate), 130 (49.4%) were randomised to 12-weekly and 133 (50.6%) to 4-weekly therapy. 138 (52.5%) were bone-agent na¨ıve. The BTAs included; denosumab (n=148, 56.3%), zoledronate (n=63, 24.0%) and pamidronate (n=52, 19.8%). Study-reported outcomes showed no significant difference in; HRQL-physical domain (median [range]: 0 [-86, 40] vs. 0 [-66, 53.3]), pain (median [range]: 0 [-66, 72] vs. 0 [-100, 88]), Global Health Status (median [range]: 0 [-100, 66.7] vs. 0 [-83, 33.3]), SSE rates (N [%]: 24 [18.5%] vs. 22 [16.5%]), 1-year SSE-free rate (median, range; 73.2% [63.6, 80.7] vs. 77.9% [69.1, 84.4]) between the 12- and 4-weekly arms, respectively. Subgroup analyses for BTA na¨ıve and pre-treated patients, and for patients receiving denosumab, zoledronate and pamidronate, showed no significant difference between the 12- and 4-weekly arms. There was no significant difference in reported rates of renal impairment (2.3% vs. 3.0%), symptomatic hypocalcae- mia (1.5% vs. 1.5%) or osteonecrosis of the jaw (0.8% vs. 0.8%). Conclusion: The findings of this trial are consistent with those previously reported for de-escalating zoledronate. This trial also included patients receiving de-escalated denosumab and pamidronate. While the results of the Swiss REDUSE trial are awaited, the data presented would suggest that de-escalation of all commonly used BTAs is a reasonable treatment option. Clinical trial information: NCT02721433.

11502 Oral Abstract Session, Mon, 8:00 AM-11:00 AM

Osteonecrosis of the jaw in patients with cancer receiving zoledronic acid for bone metastases: SWOG S0702, NCT00874211.

Catherine H. Van Poznak, Joseph M. Unger, Amy K. Darke, Carol Moinpour, Robert A. Bagramian, Mark M. Schubert, Lisa Kathryn Hansen, Justin D. Floyd, Shaker R. Dakhil, Danika L. Lew, James Lloyd Wade, Michael Jordan Fisch, Norah Lynn Henry, Dawn L. Hershman, Julie Gralow; University of Michigan Comprehensive Cancer Center, Ann Arbor, MI; SWOG Statistics and Data Management Center, Fred Hutchinson Cancer Center, Seattle, WA; Fred Hutchinson Cancer Research Center, Seattle, WA; School of Dentistry, University of Michigan, Ann Arbor, MI; University of Washington, School of Dentistry, Seattle, WA; Legacy Good Samaritan Hospital, Portland, OR; Heartland NCORP/Cancer Care Specialists of Illinois, Swansea, IL; Wichita NCORP/Cancer Center of Kansas, Wichita, KS; SWOG Statistics and Data Management Center; Fred Hutchinson Cancer Center, Seattle, WA; Heartland NCORP/Cancer Care Specialists of Central Illinois, Decatur, IL; AIM Specialty Health, Chicago, IL; Huntsman Cancer Institute, University of Utah, Salt Lake City, UT; Columbia University Medical Center, New York, NY; University of Washington, Seattle Cancer Care Alliance, Seattle, WA

Background: Osteonecrosis of the jaw (ONJ) may occur in cancer patients (pts) with metastatic bone disease (MBD) treated with bone modifying agents. No large prospective studies have precisely determined the incidence of ONJ. A better understanding of the true incidence and predictors of ONJ is needed. Methods: SWOG S0702 was a prospective observational study that assessed the cumulative incidence (CI) of ONJ at 3 years in pts with MBD from any malignancy receiving zoledronic acid (Zol). Participants must have had either limited or no prior exposure to bone modifying agents and a clinical care plan that included use of Zol within 30 days of registration. Cancer treatments, bone modifying agents (including Zol), and dental care were administered as clinically indicated and were not directed by S0702. Baseline and every 6 m followup dental exams were recommended. Report forms (medical, dental and pt reported outcomes) were submitted every 6 m but if ONJ was diagnosed, follow up interval became every 3 m. Protocol defined ONJ required exposed bone in the maxillofacial region present 8 weeks or more in a pt who was receiving or had been exposed to a bisphosphonate, and had not had radiation therapy to the craniofacial region. Results: The study enrolled 3,491 evaluable pts (breast 1,120; myeloma 580; prostate 702, lung 666, other 423) between 2009-2013. About 2/3 of pts had a baseline dental exam. Overall, 87 pts had confirmed ONJ. The cumulative incidence of ONJ was 0.8% at year 1 (95% CI: 0.5%-1.1%), 2.0% at year 2 (95% CI: 1.5%-2.5%), and 2.8% at year 3 (95% CI: 2.3- 3.5%). Rates of 3-year confirmed ONJ were highest in myeloma pts (4.3%; 95% CI, 2.8%-6.4%). Pts with planned Zol dosing intervals of every 3-4 weeks (n = 3,032, 87.2%) were much more likely to experience ONJ than pts with planned dosing intervals of 5 weeks or greater (n = 447, 12.8%; 3.2% vs 0.7%; HR = 4.80, 95% CI, 1.52-15.18, p = .008). Fewer total number of teeth, the presence of dentures and any oral surgery at baseline were all associated with a higher rate of ONJ. Conclusions: About 1 in 40 patients receiving Zol for MBD developed ONJ. S0702 provides information to guide stratification of risk for developing ONJ in pts with MBD receiving Zol. Cancer type, oral health and frequency of Zol dosing affect risk of ONJ. Clinical trial information: NCT00874211.

11503 Oral Abstract Session, Mon, 8:00 AM-11:00 AM

A randomized, double-blind, placebo-controlled phase III trial evaluating olanzapine 5 mg combined with standard antiemetic therapy for the prevention of chemotherapy-induced nausea and vomiting in patients receiving cisplatin-based chemotherapy: J-FORCE Study.

Hironobu Hashimoto, Masakazu Abe, Masahiko Nakao, Hideaki Mizutani, Yasuhiko Sakata, Yukiyoshi Fujita, Tomoyasu Nishimura, Katsuya Hirano, Hideaki Okada, Naoki Inui, Yukio Sakata, Hirotoshi Iihara, Sadamoto Zenda, Yosuke Uchitomi, Takuhiro Yamaguchi, Yukari Hoshina, Takako Yanai, Satoru Iwasa, Noboru Yamamoto, Yuichiro Ohe; Department of Pharmacy, National Cancer Center Hospital, Tokyo, Chuo-Ku, Japan; Department of Gynecologic Oncology,Shizuoka Cancer Center, Shizuoka, Japan; Department of Pharmacy, Osaka City General Hospital, Osaka, Japan; Saitama Cancer Center, Saitama, Japan; Hiroshima City Hiroshima Citizens Hospital, Hirosh- ima, Japan; Gunma Prefectural Cancer Center, Gunma, Japan; Wakayama Medical University Hospital, Wakayama Prefecture, Japan; Department of Respiratory Medicine, Hyogo Prefectural Amagasaki General Medical Center, Amagasaki, Japan; Kobe Minimally Invasive Cancer Center, Hyogo, Japan; Department of Clinical Pharmacology and Therapeutics, Hamamatsu University School of Medicine, Hamamatsu, Japan; Hakodate Municipal Hospital, Hokkaido, Japan; Gifu University Hospital, Gifu, Japan; Japan Supportive, Palliative and Psychosocial Oncology Group, Tokyo, Japan; Tohoku University Graduate School of Medicine, Sendai, Japan; National Cancer Center Hospital, Tokyo, Japan; De- partment of Pharmacy, National Cancer Center Hospital, Tokyo, Tokyo, Japan; Department of Experimental Therapeutics, National Cancer Center Hospital, Tokyo, Japan; Department of Thoracic Oncology, National Cancer Center Hospital, Tokyo, Japan

Background: Olanzapine (OLZ) 10 mg added to standard antiemetic therapy including aprepitant (APR), palonosetron (PALO), and dexamethasone (DEX) has been recommended for the prevention of chemotherapy-induced nausea and vomiting (CINV) caused by highly emetogenic chemotherapy (HEC). Guidelines suggest that a dose of 5 mg should be taken into consideration in patients at risk of sedation. OLZ 5 mg showed an equivalent activity and favorable toxicity to somnolence in several phase II studies. We conducted a randomized, double-blind, placebo-controlled phase III trial to evaluate OLZ 5 mg in addition to standard antiemetic therapy for the prevention of CINV in patients receiving cisplatin- based chemotherapy. Methods: Patients receiving cisplatin ($ 50 mg/m2) were randomly assigned to either OLZ 5 mg or placebo on days 1–4, combined with APR, PALO and DEX. The primary endpoint was complete response (CR), defined as no vomiting and no rescue medications in the delayed phase (24–120 h). A total of 690 patients were required to detect a 10% increase in CR from 65% in the placebo to 75% in the OLZ, with a one-sided alpha of 2.5% and a power of 80%. Results: A total of 710 patients were enrolled (OLZ 356 and placebo 354). CR in the delayed phase was 79.1% (95% CI: 74.9–83.3) in the OLZ 5 mg and 65.8% (95% CI: 60.9–70.8) in the placebo (p , 0.001). Other efficacy results are summarized in Table. The most common treatment-related adverse events was somnolence (43.1% for OLZ vs. 33.0% for placebo). Conclusions: OLZ 5 mg combined with APR, PALO and DEX can be considered a new standard antiemetic therapy in patients receiving cisplatin-based chemotherapy. Clinical trial information: UMIN000024676.

OLZ 5 mg (n = 354) placebo (n = 351) % 95%CI % 95%CI CR Acute 94.9 92.6, 97.2 88.6 85.3, 91.9 Delayed 79.1 74.9, 83.3 65.8 60.9, 70.8 Overall 78.0 73.7, 82.3 63.5 58.5, 68.6 TC* Acute 85.9 82.3, 89.5 80.6 76.5, 84.8 Delayed 60.2 55.1, 65.3 50.1 44.9, 55.4 Overall 58.8 53.6, 63.9 48.1 42.9, 53.4 *TC (total control): no vomiting, nausea and rescue medications

11504 Oral Abstract Session, Mon, 8:00 AM-11:00 AM

Olanzapine (OLN) versus aprepitant (APR) in patients receiving high-emetogenic chemotherapy: Final results of randomized phase II trial.

Alexey Rumyantsev, Elena Glazkova, Rukhshona Nasyrova, Ekaterina Ignatova, Lia Chitia, Anna Popova, Dzhennet Chekini, Yaroslava Kochetkova, Sergei Kit, Kheda Elsnukaeva, Sophia Menshikova, Olga Sekhina, Ilya Pokataev, Alexandra Tyulyandina, Marina Stenina, Mona A. Frolova, Anatoly Bulanov, Mikhail Fedyanin, Alexey Tryakin, Sergei Tjulandin; Federal State Budgetary Institution "N.N. Blokhin National Medical Research Center of Oncology" pf the Ministry of Health of the Russian Federation (N.N. Blokhin NMRCO), Moscow, Russian Federation; National Medical Research Cancer CenterFederal State Budgetary Institution N.N. Blokhin National Medical Research Center of Oncology pf the Ministry of Health of the Russian Federation (N.N. Blokhin NMRCO), Moscow, Russian Federation; Federal State Budgetary Institution «N.N. Blokhin National Medical Research Center of Oncology» pf the Ministry of Health of the Russian Federation (N.N. Blokhin NMRCO), Moscow, Russian Federation; Federal State Budgetary Institution N.N. Blokhin National Medical Research Center of Oncology pf the Ministry of Health of the Russian Federation (N.N. Blokhin NMRCO), Moscow, Russian Federation; Federal State Budgetary Institution «N.N. Blokhin National Medical Research Center of Oncology» pf the Ministry of Health of the Russian Federation (N.N. Blokhin NMRCO), Moscow, Russia

Background: Management of chemotherapy-induced nausea and vomiting (CINV) remains challenging. OLN might provide several benefits over APR which is current standard of care – particularly in terms of nausea control and cost effectiveness. However, sedation associated with recommended doses of olanzapine precludes its wide use in oncology practice. Methods: This was randomized phase II single center study aimed to compare OLN and APR in CINV prophylaxis. Key inclusion criteria were: chemo- and radio-therapy na¨ıve patients, planned administration of high-emetogenic chemotherapy (cisplatin, carboplatin AUC$4, doxorubicin etc). Patients were randomized 1:1 ratio in the following arms: olanzapine 5 QD day 0-4 or aprepitant 125 mg day 1, 80 mg day 2,3. All patients received ondansetron 16 mg day 1 and dexamethasone 8 mg day 1-3. Primary endpoint was complete nausea control (no nausea and no rescue medication) 0-120 hours after chemotherapy. Complete response (no emesis and no rescue medication) was a key secondary end point. Nausea was assessed using MASCC Antiemesis Tool. Sample size: 94 patients to increase nausea control rate from 40 to 70% (a = 0.05; b = 0.80; 10% of estimated data loss). Results: We included in the analysis 93 patients who could be evaluated. The groups were well balanced, median age was 49 years, vast majority of patients (95.6%) were females. The proportion of patients with complete nausea control in OLN and APR groups was 44.2% and 24.0% respectively (RR 2.5; 95% CI 1.04-6.08; p = 0.039). Complete response was achieved in 74.4% and 54.0% patients respectively (RR 2.48; 95% CI 1.026-5.99; p = 0.041). No differences in rates of undesired sedations were detected. Conclusions: Our data suggests superiority of OLN regimen in terms of nausea control. This regimen deserves further investigation. Clinical trial information: NCT03478605.

11505 Oral Abstract Session, Mon, 8:00 AM-11:00 AM

Randomized controlled trial (RCT) of a patient navigation (PN) intervention to increase early access to supportive care (SC) for patients with metastatic cancer in a resource- limited setting.

Enrique Soto Perez De Celis, Yanin Chavarri Guerra, Wendy Alicia Ramos-Lopez, Alfredo Covarrubias- Gómez, Africa Navarro-Lara, Paulina Quiroz, Sofia Sanchez, Natasha Alcocer, Mirza Alcalde Castro, Jose Carlos Aguilar Velazco, Alexandra Bukowski, Juan Alberto Chavarri Maldonado, Sergio Contreras- Gardu~no, Lindsay Krush, Itoro E. Inoyo, Andrea Medina-Campos, Marıá Luisa Moreno-Garcıá, Viridiana Perez-Montessoro, ´ Paul E. Goss; Instituto Nacional de Ciencias Medicas y Nutricion Salvador Zubiran, Mexico City, Mexico; Instituto Nacional de Ciencias Medicas y Nutricion Salvador Zubiran, Mexico City, DF, Mexico; Global Cancer Institute, Boston, MA; Universidad Autonoma ´ de Ciudad Juarez, ´ Ciudad Juarez, ´ CI, Mexico; Massachusetts General Hospital Cancer Center and Harvard Medical School, Boston, MA

Background: Early integration of SC to the treatment of advanced cancer can improve outcomes, but this may be challenging in developing countries due to a lack of resources and knowledge. In this RCT, we examined whether PN could improve early access to SC among Mexican patients with metastatic solid tumors as recommended by ASCO guidelines. Methods: Adult patients with newly-diagnosed metastatic cancer were randomly assigned to PN or standard oncologic care. At baseline, a navigator assessed the patients’ SC needs (depression, anxiety, fatigue, pain, caregiver burden) using validated questionnaires administered with an electronic tablet. For those in the PN arm, a personalized SC plan was created and implemented by a multidisciplinary team (palliative care, physical therapy, geriatrics, psychology, psychiatry). The primary outcome was access to SC, defined as receipt of SC interventions in the first 3 months (mo) after diagnosis. Secondary outcomes included advanced directive (AD) completion (for patients with expected survival #6 mo in accordance to Mexican law), changes in SC needs, and changes in quality of life (assessed using FACT-G). Results: 133 patients (median age 60, range 23-93; 52% male) were randomized (66 PN, 67 control). 61% of patients had gastrointestinal tumors. 94% of patients in the PN arm completed baseline assessments and received recommendations from the navigator. At 3 mo, 37 patients died or were lost to follow-up (16 PN, 21 control; p = 0.45), and 96 completed assessments. SC interventions were provided to 73% of patients in the PN arm and 24% of controls (p , 0.01). In the PN arm, 48% of 29 eligible patients completed AD, compared to 0% of eligible controls (p , 0.01). At 3 mo, patients in the PN arm were significantly less likely to report moderate/severe pain than controls (10 vs 33%, p = 0.006). There were no significant differences in other symptoms or in FACT-G scores (76 vs 76.3, p = 0.46) between PN and control arms at 3 mo. Conclusions: PN can lead to significant improvements in early access to SC, AD completion, and pain control among patients with metastatic cancer treated in a resource-limited setting. Clinical trial information: NCT03293849.

11506 Oral Abstract Session, Mon, 8:00 AM-11:00 AM

Safety of pregnancy following breast cancer (BC) in patients (pts) carrying a BRCA mutation (mBRCA): Results of an international cohort study.

Matteo Lambertini, Lieveke Ameye, Anne-Sophie Hamy, Anna Zingarello, Philip Daniel Poorvu, Estela Carrasco, Albert Grinshpun, Sileny Han, Christine Rousset-Jablonski, Alberta Ferrari, Shani Paluch-Shimon, Laura Cortesi, Claire Senechal, Gianmaria Miolo, Katarzyna Pogoda, Jose Alejandro Perez-Fidalgo, Laura De Marchis, Lucia Del Mastro, Fedro Peccatori, Hatem A. Azim; Policlinico San Martino Hospital - University of Genova, Genova, Italy; Data Centre, Institut Jules Bordet - Universite ´ Libre de Bruxelles (ULB), Brussels, Belgium; Institut Curie, Paris, Gabon; Institut Gustave Roussy, Villejuif, France; Dana-Farber Cancer Institute, Boston, MA; Vall Hebron Institute of Oncology, Barcelona, Spain; Hadassah-Hebrew University Medical Center, Jer- usalem, Israel; BGOG & Department of Gynaecology and Obstetrics, University Hospitals Leuven, Leuven Cancer Institute, Leuven, Belgium; Centre Leon-B ´ erard, ´ Lyon, France; Fondazione IRCCS Policlinico San Matteo, University of Pavia, Pavia, Italy; Sheba Medical Center, Tel HaShomer, Ramat Gan, Israel; University Hospital of Modena, Modena, Italy; Bergonie Institute, Bordeaux, France; Dipartimento di Oncologia Medica, Centro di Riferimento Oncologico di Aviano (CRO), IRCCS, Aviano, Italy; Maria Sklodowska-Curie Institute - Oncology Center, Warsaw, Poland; Hospital Clı´nico Uni- versitario de Valencia, INCLIVA, Centro de Investigacion ´ Biomedica ´ en Red de Oncologı´a, CIBERONC- ISCIII, GEICAM Spanish Breast Cancer Group, Valencia, Spain; Policlinico Umberto 1, Rome, Italy; Ospedale Policlinico San Martino-Oncologia Medica, Genova, Italy; European Institute of Oncology IRCCS, Milan, Italy; American University of Beirut (AUB), Beirut, Lebanon

Background: Very limited data are available on the safety of pregnancy and reproductive outcomes in mBRCA pts with prior BC history. We report the results of the largest study to date addressing these questions. Methods: This international, multicenter, hospital-based, retrospective cohort study included consecutive pts with invasive early BC (stage I-III) diagnosed between Jan-2000 and Dec-2012 at the age of #40 years and carrying a deleterious germline mBRCA. Primary endpoints were pregnancy rate and disease-free survival (DFS); overall survival (OS) and pregnancy outcomes were secondary endpoints. To account for guarantee-time bias, we performed two survival analyses: 1) Case-control approach matching pregnant and non-pregnant (1:3) pts for classic prognostic factors (each non-pregnant control had a disease-free interval $ than the time elapsing between BC diagnosis and date of pregnancy of the matched pregnant case); 2) Extended Cox model with time-varying covariates including all pts. Results: 1,252 mBRCA BC pts (811 mBRCA1, 430 mBRCA2, 11 mBRCA1&2)wereincludedfrom30centers worldwide, of whom 195 pts had a pregnancy (pregnancy rate = 16% [95% CI 14-18]) after a median 4.5 years (range 3.1-6.7 years) following BC diagnosis. Pregnant pts were younger and had more ER-negative tumors (all p , 0.01). 16 (8.2%) and 20 (10.3%) pts had an induced and spontaneous abortion, respectively. Among the 150 (76.9%) pts who conceived (n = 170 babies), pregnancy complications and congenital anomalies were described in 13 (11.6%) and 2 (1.8%) cases, respectively. Median follow-up was 8.3 years (range 8.1-8.7 years). In the case-control analysis, pregnant pts had better DFS (HR 0.71; 95% CI 0.51-0.99; p = 0.045), with no difference in OS (HR 0.86; 95% CI 0.44-1.67; p = 0.65). Subgroup analysis suggested that the superior outcome was restricted to mBRCA1 pregnant pts (p- interaction , 0.01). Similar results were obtained in the second supportive analysis. Conclusions: Pregnancy following BC is safe in mBRCA pts, particularly mBRCA1, with no detrimental impact on maternal prognosis or fetal outcomes. These findings are of paramount importance for fertility counseling in young mBRCA BC pts.

11507 Oral Abstract Session, Mon, 8:00 AM-11:00 AM

Effects of exercise on cancer-related fatigue and muscular strength in patients with breast cancer.

Po-Ju Lin, Kah Poh Loh, Julia Ellen Inglis, Richard Francis Dunne, Ian Kleckner, Chunkit Fung, Nikesha Gilmore, Gilberto Lopez, Charles E. Heckler, Eva Culakova, Luke Joseph Peppone, Michelle Christine Janelsins, Charles Stewart Kamen, Karen Michelle Mustian; University of Rochester Medical Center, Rochester, NY; University of Rochester James P. Wilmot Cancer Institute, Strong Memorial Hospital, Rochester, NY

Background: Cancer-related fatigue (CRF) is a persistent daily lack of energy commonly experienced by breast cancer (BC) patients. Due to CRF, BC patients have difficulties carrying out daily activities, become less active and consequently reduce muscular strength. Exercise can improve muscular strength and increase energy level; therefore it may alleviate CRF. This phase II RCT assessed the effects of exercise on CRF and muscular strength in BC patients. Methods: Ninety BC patients (55.569.6 years, 79% white, 48% and 46% under radiation or hormone therapy) were randomized into two arms: a 6-week Exercise for Cancer Patients (EXCAP) program or standard care (Control). EXCAP is a home-based, personalized, progressive exercise program combining aerobic walking and resistance band training. The Brief Fatigue Inventory was used to assess CRF and CRF interference with daily activities and a 7-10 repetition maximum chest press and leg extension strength test was used to assess upper- and lower-body strength at pre- and post-intervention. T-tests and ANCOVA with pre-intervention as the covariate were used to analyze within- and between-group changes, respectively. Results: Participants in the EXCAP group decreased CRF (-0.960.3, p = 0.01) and CRF interference with daily activities (-1.160.3, p , 0.01) from pre- to post-intervention while participants in the Control group did not. The mean improvement (from pre- to post-intervention) in CRF and CRF interference of daily activities for the EXCAP group were significantly higher than the change in the Control group (both p , 0.01). Participants in the EXCAP group increased upper- (3.961.4, p , 0.01) and lower-body strength (6.461.3, p , 0.01) from pre- to post-intervention, while participants in the Control group did not. The mean increase (from pre- to post-intervention) in lower-body strength for the EXCAP group was significantly higher than the change in the Control group (p = 0.01). Conclusions: Exercise combining aerobic walking and resistance band training reduces CRF and CRF interference with daily activities and improves muscular strength in BC patients. Results from this study provide further evidence of the benefits of exercise for supportive cancer care. Clinical trial information: NCT00851812.

11508 Oral Abstract Session, Mon, 8:00 AM-11:00 AM

Association of baseline cardiovascular risk factors and health care utilization and costs in elderly breast cancer patients enrolled in SWOG clinical trials.

Dawn L. Hershman, Cathee Till, Jason Dennis Wright, Melissa Kate Accordino, Riha Vaidya, William E. Barlow, Scott Ramsey, Joseph M. Unger; Columbia University College of Physicians and Surgeons, New York, NY; Fred Huchinson Cancer Research Center, Seattle, WA; Columbia University College of Physicians and Surgeons and New York Presbyterian Hospital, New York, NY; New York- Presbyterian Hospital, New York, NY; Fred Hutchinson Cancer Research Center, Seattle, WA; Cancer Research and Biostatistics, Seattle, WA

Background: Cardiovascular-disease risk factors (CVD-RFs) increase the risk of cardiac events in women undergoing chemotherapy. Less is known about the impact of CVD-RFs on healthcare utilization and costs. Methods: We examined breast cancer patients treated uniformly on SWOG clinical trials from 1999- 2011. We identified baseline diabetes, hypertension, hypercholesterolemia, and coronary artery disease (CAD) by linking trial records to Medicare claims; obesity was identified using clinical records. The outcomes were emergency room visits (ER), hospitalizations and costs. Multivariable logistic and linear regression were used. Results: Among the 708 patients included in the analysis, 160 (22.6%) experienced 234 separate hospitalizations, and 193 (27.3%) experienced 311 separate ER visits. Diabetes, hypertension, hypercholesterolemia, and CAD were all associated with increased risk of hospitalizations and ER visit. Hypertension had the strongest association, with more than a threefold risk of hospitalization for those with hypertension compared to those without (OR [95% CI], 3.16 [1.85- 5.40], p,0.001). For those with $3 CVD-RFs, the risk of hospitalization was greater compared to 0 or 1 CVD-RFs (OR [95% CI], 2.74 [1.71-4.38], p,0.001). Similar results were seen for ER visits. In the first 12 months after trial registration, patients with diabetes ($38,324 vs $30,923, 23.9% increase, p=0.05), hypercholesterolemia ($34,168 vs $30,661, 11.4% increase, p=0.02), and CAD ($37,781 vs $31,698, 19.2% increase, p=0.04) had statistically significantly higher total healthcare costs. Addi- tionally, those with 2 significant CVD-RFs ($35,353 vs. $28,899, 22.3% increase, p=.005) had higher total healthcare costs. Conclusions: Our study demonstrates that the presence of both CVD-RFs and ER visits and hospitalizations are frequent among elderly BC patients. The risk of ER visits and hospitalizations is higher among patients with CVD-RFs, and increases with the number of RFs. Better management of CVD-RFs and more aggressive symptom management may be required to reduce both physical and financial toxicities to elderly patients undergoing BC therapy.

11509 Clinical Science Symposium, Mon, 3:00 PM-4:30 PM

Predictive factors of grade 3-5 toxicity in older patients with cancer treated with chemotherapy: A prospective multicenter study.

Jaime Feliu Batlle, Laura Basterretxea, Maria Dolores Torregrosa, Elisenda Llabres, Beatriz Losada Vila, Beatriz Jimenez-Munarriz, Maite Antonio Rebollo, Ana B. Custodio, Regina Girones, Patricia Cruz, Mar Munoz Sanchez, Alvaro Pinto, Gemma Soler, Teresa Soria Comes, Hellis Telleria, MJ Molina- Garrido; Medical Oncology Department, La Paz University Hospital, Madrid, Spain; Hospital Donostia, Donostia, Spain; Hospital Lluis Alcanyis, Valencia, Spain; H. U. Insular, Las Palmas, Spain; H. Fuenlabrada, Madrid, Spain; H. Clara Campal Madrid, Madrid, Spain; Oncogeriatrics Unit, Catalan Institute of Oncology, Hospitalet, Barcelona, Spain; Hospital Universitario La Paz, Madrid, Spain; Hospital UniversitarioLaF´e,Val`encia, Spain; Hospital Virgen de La Luz, Cuenca, Spain; Institut Catalad ` ’Oncologia, Barcelona, Spain; Hospital Universitario Doctor Peset, Valencia, Spain; ESI/OSI Donostialdea, San Sebastian, ´ Spain; Medical Oncology Department, Hospital Virgen de la Luz in Cuenca, Cuenca, Spain

Background: Older patients have increased risk of toxicity from chemotherapy. The purpose of this study was to analyse predictive factors for developing grade 3-5 toxicity in older patients treated with chemotherapy. Methods: This prospective multicenter study included 500 cancer patients $ 70 years between Feb 2014 and Jun 2018. A prechemotherapy assessment including sociodemographics, tumor/ treatment variables, laboratory test results, and geriatric assessment variables (function, comorbidity, cognition, psychological state, social activity/support, and nutritional status) was performed. Logistic regression was used to examine the association between these factors and the development of grade 3-5 toxicity. Results: Mean age of the patients was 77 years (70-92), ECOG PS 0/1/2: 25%/63%/12%. 223 (45%) had a primary dose reduction.167 (33%) patients developed grade 3-5 toxicity (28% grade 3, 5% grade 4, 1% grade 5). Univariate analysis found a higher risk of grade 3-5 toxicity in patients with creatinine clearance # 60 mL/min, IADL #7, VES13 $ 6, and the administration of standard chemotherapy doses. In multivariable analysis, only the chemotherapy dose (odds ratio [OR] 1.179; 95% confidence interval [CI] 1.215–2.655) and creatinine clearance (odds ratio [OR] 0.989; 95% confidence interval [CI] 0.981–0.997) were independently associated with toxicity. Conclusions: Renal function and chemotherapy dose were significant predictors of grade 3-5 toxicity among older patients treated with chemotherapy.

11510 Clinical Science Symposium, Mon, 3:00 PM-4:30 PM

Validating the Cancer and Aging Research Group (CARG) toxicity prediction tool in older men receiving chemotherapy for metastatic castration-resistant prostate cancer (mCRPC) and extending it to androgen receptor targeted agents.

Shabbir M.H. Alibhai, Henriette Breunis, Richard William Gregg, Aaron Richard Hansen, Padraig Richard Warde, Narhari Timilshina, George Tomlinson, Anthony M. Joshua, Neil Eric Fleshner, Urban Emmenegger; University Health Network, Toronto, ON, Canada; Cancer Centre of Southeastern Ontario, Kingston, ON, Canada; Princess Margaret Cancer Centre, Toronto, ON, Canada; Department of Radiation Oncology, University of Toronto; Radiation Medicine Program, Princess Margaret Cancer Centre, University Health Network, Toronto, ON, Canada; University of Toronto, Toronto, ON, Canada; Princess Margaret Cancer Centre, University Health Network, University of Toronto, Toronto, ON, Canada; De- partment of Surgery, Division of Urology, University Health Network, Princess Margaret Cancer Centre, Toronto, ON, Canada; Odette Cancer Centre, Sunnybrook Health Sciences Centre, University of Toronto, Toronto, ON, Canada

Background: Multiple treatment options are available for mCRPC. Being able to predict toxicity risk for different treatments in older adults can aid treatment decision-making and supportive care. The CARG tool is a promising toxicity risk prediction tool for chemotherapy, but has not been specifically validated in the mCRPC setting for either chemotherapy or androgen receptor targeted agents. We prospectively evaluated the ability of the CARG tool to predict risk of clinically relevant grade 2 and grade 3+ toxicity of treatment with chemotherapy (CHEMO) and abiraterone or enzalutamide (A/E) in older adults with mCRPC. Methods: Men age 65+ were enrolled in this prospective observational study at 3 academic centres, Princess Margaret Cancer Centre, Sunnybrook Health Sciences Centre, and Kingston Health Sciences Centre in Ontario, Canada. All grade 2 and grade 3+ toxicities were documented during cycle 1 of CHEMO or in the first 3 months of A/E via structured interviews and chart review. Lab abnormalities were documented only if resulting in emergency room visits, requiring treatment, or affecting subsequent oncologic treatment. Toxicity was rated using the Common Terminology Criteria for Adverse Events version 4. Logistic regression was performed to identify predictors of toxicity. Results: 64 men starting CHEMO (primarily docetaxel 60-75 mg/m^2, mean age 73y) and 59 men starting A/E (mean age 76y) were included. Clinically important grade 2 toxicities occurred in 86% and 70% of CHEMO and A/E patients, respectively. Grade 3+ toxicities occurred in 48% and 25% of CHEMO and A/E patients, respectively. The CARG tool was predictive of grade 3+ toxicities with CHEMO, which occurred in 22%, 53%, and 71% of low, moderate, and high risk groups (p = 0.017). However, the CARG tool was not predictive of grade 2 toxicities with CHEMO, or grade 2 or 3+ toxicities with A/E (Table). Conclusions: We provide external validation of the CARG tool in predicting grade 3+ toxicity in older men with mCRPC undergoing CHEMO. Grade 2 toxicities are very common with both CHEMO and A/E, and grade 3+ toxicity occurs in 1 in 4 older men on A/E. Additional efforts to identify men at higher risk of toxicity from various mCRPC treatments are warranted.

Toxicity CARG LowRisk CARG Moderate Risk CARG HighRisk p-value CHEMO Grade 2, n(%) 17 (94%) 26 (81%) 12 (86%) 0.43 Grade 3, n(%) 4 (22%) 17 (53%) 10 (71%) 0.017 A/E Grade 2, n(%) 8 (57%) 27 (73%) 5 (83%) 0.41 Grade 3, n(%) 3 (21%) 11 (30%) 0 0.28

11511 Clinical Science Symposium, Mon, 3:00 PM-4:30 PM

Outcomes that matter to patients: The Geriatric Oncology Surgical Assessment and Functional rEcovery after Surgery (GOSAFE) study—Analysis of 471 patients.

Isacco Montroni, Giampaolo Ugolini, Antonino Spinelli, Giorgio Ercolani, Michael T. Jacklitsh, Siri Rostoft, Ponnandai Sadasivan Somasundar, Barbara Van Leeuwen, Nicola de Liguori Carino, Nicole M. Saur, Giovanni Ferrari, Federico Ghignone, Giacomo Sermonesi, Francesca Rita Di Candido, Flavia Foca, Chiara Zingaretti, Bernadette Vertogen, Riccardo Audisio, SIOG Surgical Task Force/ESSO GOSAFE Study Group; Colorectal Surgery Unit, Ospedale per gli Infermi di Faenza, AUSL Romagna, Faenza, Italy; Division of Colon and Rectal Surgery, Humanitas Clinical and Research Center, Department of Biomedical Sciences, Humanitas University, Rozzano, Milan, Italy; U.O. Chirurgia Generale e Terapie Oncologiche Avanzate Ospedale "GB.Morgagni-L.Pierantoni", AUSL Romagna, Forli, Italy; Brigham and Women’s Hospital, Harvard Medical School, Boston, MA; University of Oslo, Oslo, Norway; Roger Williams Medical Center/Boston University School of Medicine, Providence, RI; Department of Surgical Oncology, University Medical Center Groningen, University of Groningen, Groningen, Netherlands; HPB Unit, Manchester Royal Infirmary, Central Manchester University Hospitals, Manchester, United Kingdom; University of Pennsylvania Perelman School of Medicine, Pennsylvania Hospital, Philadelphia, PA; ASST Grande Ospedale Metropolitano Niguarda, Milano, Italy; Colorectal Surgery, General Surgery, Ospedale per gli Infermi, Faenza, Italy; Colorectal Surgery General Surgery Ospedale per gli Infermi, Faenza, Italy; Division of Colon and Rectal Surgery, Humanitas Clinical and Research Center, Rozzano, Milan, Italy; IRCCS-IRST (Istituto Scientifico Romagnolo per lo Studio e la Cura dei Tumori), Meldola, Italy; Unita ` di Biostatistica e Sperimentazioni Cliniche, IRCCS/IRST Meldola, Ravenna, Italy; Istituto Scientifico Romagnolo per lo Studio e la Cura dei Tumori IRCCS, Meldola, Italy; University of Gothenburg - Institute of Clinical Sciences, Gothen- burg, Sweden

Background: Older cancer patients value functional outcomes as much as survival but surgical studies lack functional recovery (FR) data. The international, multicenter GOSAFE study (ClinicalTrials.gov NCT03299270) aims to evaluate patients’ quality of life (QoL)and FR after cancer surgery and to assess predictors of FR. Methods: GOSAFE prospectively collects functional and clinical data before and after major elective cancer surgery on senior adults ($70 years). Surgical outcomes are recorded (30 days, 90 days, and 180 dayspost-operatively) with QoL(EQ-5D-3L) and FR (Activities of Daily Living (ADL) + Timed Up and Go (TUG) + MiniCog), 28centers are prospectively enrolling patients; accrual ends February 2019. Results: 643 patients underwent major cancer surgery with curative(94%) or palliative (6%) intent (February 2017-September 2018). Median age was 78(range 70-94); 51.6% males, ASA III-IV 52%. Patients dependent (ADL , 5) were 8%. Frailty was detected by G8 . 14 in 32% and fTRST$2 in 36% of patients. 639 (99%) lived at home, 32% lived alone, and 88% were able to go out. Major comorbidities (CCI . 6) were detected in 36% and 22% had cognitive impairment according to MiniCog (5% self-reported). 26% had . 3 kg weight loss, 30% were hospitalized in the last 90 days, 45% had $3 medications (6% none). For 471 patients, a 90-day comprehensive evaluation was available. Postoperative morbidity was 42% (30 day) and 63.3% (90 day), but Clavien- Dindo III-IV complications were only 11.2% and 17.6%. 90-day mortality was 7.4% (5% 30-day). QoL improved 90 days after surgery (mean EQ-5D index from 0.76 to 0.80). Patients with EQ-5D VAS score . 60 raised from 73.9% at baseline to 82.8% at 90 days. 29% had complete FR (ADL score . 4, MiniCog . 2, TUG , 20). Decreased functional capacity was seen in 23.4% of patients alive at 90- days. Conclusions: GOSAFE is the largest prospective study on older cancer patients undergoing major surgery. Interim analysis reports decreased functional capacity in a quarter of patients. The study will allow clinicians to associate clinical outcomes with individual factors of the preoperative assessment and create a user-friendly tool to predict outcomes that matter to patients.

11512 Clinical Science Symposium, Mon, 3:00 PM-4:30 PM

Reduced 90-day postoperative mortality through geriatric comanagement after cancer surgery.

Armin Shahrokni, Amy Tin, Saman Sarraf, Koshy Alexander, Soo Jung Kim, Heidi Yulico, Sincere McMillan, Robert J. Downey, Andrew Vickers, Beatriz Korc-Grodzicki; Memorial Sloan Kettering Cancer Center, New York, NY

Background: We explored the association between geriatric comanagement and 90-day postoperative mortality of cancer patients aged 75 or older. Methods: A retrospective review of a prospectively maintained database was performed on patients over 75 years old who underwent elective surgery with hospital length of stay of $1 day at Memorial Sloan Kettering Cancer Center from 2015-2018. Geriatric comanagement group (GCG) patients had geriatric preoperative evaluation and inpatient geriatric comanagement. Patients in the surgical management group (SMG) did not have geriatric preoperative evaluation or postoperative geriatric comanagement. We utilized a multivariable logistic regression model with 90-day mortality as the outcome, geriatric co-management as the predictor, and adjusted for age, gender, American Society of Anesthesiology score, Memorial Sloan Kettering Frailty Index, preoperative albumin level, operation time, and estimated blood loss. The same logistic model was used to assess the association between adverse surgical events within 30-days (any major complication, readmission, or urgent care center visit) and geriatric comanagement. Results: Of 1,855 patients (median age 80), 1,009 patients (54%) were co-managed by geriatricians. GCG patients were slightly older, less likely to be male, had longer operation time, and stayed in the hospital longer. Adjusted rates of 90-day mortality was lower in GCG vs. SMG (4.3% and 9.2%, respectively; 95% CI around difference -7.3%, -2.5%; p-value , 0.0001). We did not find evidence of a difference in adverse surgical events between groups (OR 0.96, p- value = 0.8). A greater proportion of GCG patients received inpatient physical therapy (80% vs. 64%) and occupational therapy (37% vs. 25%) compared to SMG patients. Conclusions: Our study shows that geriatric comanagement is associated with reduced 90-day postoperative mortality in cancer patients aged $75. A randomized trial study is needed to confirm this finding.

11513 Clinical Science Symposium, Sat, 8:00 AM-9:30 AM

Randomized trial of a hospice video decision aid for patients with advanced cancer and their caregivers.

Areej El-Jawahri, Olivia Vanbenschoten, Alyssa L. Fenech, Amanda L. Jankowski, Netana Markovitz, Lara Traeger, Joseph Greer, Angelo E. Volandes, Jennifer S. Temel; Massachusetts General Hospital, Boston, MA

Background: Although hospice provides high-quality end-of-life (EOL) care for patients with advanced cancer and their family caregivers, the service remains underutilized in part due to lack of adequate information provided to patients and families about hospice care. Methods: We conducted a single-site randomized clinical trial of a hospice video decision aid versus a verbal description in 150 hospitalized patients with advanced cancer and their caregivers. Patients without an available caregiver were still eligible to participate. Intervention participants (75 patients; 18 caregivers) received a verbal description about hospice plus a six-minute video depicting hospice care. Control participants (75 patients; 26 caregivers) received only the verbal description. The primary endpoint was patient preference for hospice care immediately after the intervention, adjusting for baseline preferences. Secondary outcomes included patient and caregiver knowledge and perceptions of hospice (Hospice Perception and Knowledge Questionnaire). Results: Between 2/2017 and 1/2019, we enrolled 55.7% (150/269) of potentially eligible patients and 44 caregivers. Post-intervention, patients assigned to the video group were more likely to prefer hospice care at the EOL (86.7% vs. 82.7%, OR = 2.85, P = 0.08), but this was not statistically significant. Patients in the video group reported greater knowledge about hospice (B = 0.50, P = 0.024) and were less likely to endorse that hospice care is only about death (6.7% vs. 21.6%, OR = 0.28, P = 0.035). Post-intervention, caregivers assigned to the video were more likely to prefer hospice care for their loved ones (94.4% vs. 65.4%, P = 0.031), reported greater knowledge about hospice (B = 1.94, P , 0.001), and were less likely to endorse that hospice care is only about death (0.0% vs. 23.1%, P = 0.066). Conclusions: Patients with advanced cancer and their caregivers who viewed a hospice video decision support tool were more informed about hospice care and reported more favorable perceptions of hospice. Future work should examine the impact of the video on hospice utilization and length-of-stay among patients with advanced cancer. Clinical trial information: NCT03040102.

11514 Clinical Science Symposium, Sat, 8:00 AM-9:30 AM

A randomized controlled trial of a novel artificial intelligence-based smartphone application to optimize the management of cancer-related pain.

Mihir Kamdar, Amanda Jayne Centi, Stephen Agboola, Nils Fischer, Simone Rinaldi, Jacob J. Strand, Lara Traeger, Jennifer S. Temel, Joseph Greer, Areej El-Jawahri, Vicki Jackson, Joseph Kvedar, Kamal Jethwani; Massachusetts General Hospital, Boston, MA; Connected Health Innovation, Partners Healthcare, Boston, MA; Mayo Clinic, Rochester, MN; Partners HealthCare, Boston, MA

Background: Cancer pain is a significant problem that impairs patient quality of life and increases healthcare utilization. ePAL is a smartphone application that utilizes patient-reported outcomes (PROs) and artificial intelligence (AI) to optimize cancer pain management. This randomized controlled trial examined the impact of ePAL on cancer pain severity, attitudes toward cancer pain, and healthcare utilization. Methods: Patients with pain from metastatic solid tumors (n = 112) undergoing treatment in a palliative care clinic were randomized to either a control group (n = 56) that received usual care or an intervention group (n = 56) that received ePAL in addition to usual care for 8 weeks. Measures of pain severity (Brief Pain Inventory), attitudes towards cancer treatment (Barriers Questionnaire II) and anxiety (General Anxiety Disorder-7) were assessed. We used repeated measures mixed modeling to assess change in outcome measures over time. We also conducted a chart review to identify pain-related hospital admissions and emergency department (ED) visits and compared risk between study groups. Results: Pain severity (BPI) and negative attitudes toward cancer treatment (BQ-II) decreased significantly for those assigned to ePAL compared to controls (ß = -0.09, p = 0.034 and ß = -0.037, p = 0.042, respectively). Patients assigned to ePAL reported higher anxiety scores compared to controls (ß = 0.21, p = 0.015). Patients assigned to ePAL had significantly fewer pain-related hospital admissions (n = 4 vs. n = 20, per patient risk ratio 0.31, p = 0.018) and fewer pain-related admissions through the ED (n = 2 vs. n = 14, per patient risk ratio 0.18, p = 0.008) compared to control group. Conclusions: To our knowledge, this is the first mobile app to utilize patient reported outcomes and artificial intelligence to significantly decrease pain scores and pain-related hospitalizations in patients with cancer-related pain. Future directions include examining the efficacy of ePAL in settings with limited access to palliative care.

11515 Clinical Science Symposium, Sat, 8:00 AM-9:30 AM

Prediction of treatment (tx)-induced fatigue in breast cancer (BC) patients (pts) using machine learning on genome-wide association (GWAS) data in the prospective CANTO cohort.

Sangkyu Lee, Joseph O. Deasy, Jung Hun Oh, Antonio Di Meglio, Sandrine Boyault, Marina Rousseau- Tsangaris, Celine Besse, Emilie Thomas, Anne Boland-Auge, ´ Paul H. Cottu, Olivier Tredan, Christelle Levy, Anne-Laure Martin, Sibille Everhard, Patricia A. Ganz, Ann H. Partridge, Stefan Michiels, Jean-Francois Deleuze, Fabrice Andre, Ines Maria Vaz Duarte Luis; Memorial Sloan Kettering Cancer Center, New York, NY; Memorial Sloan-Kettering Cancer Center, New York, NY; Institut Gustave Roussy, Villejuif, France; Centre Leon ´ Berard, Lyon, France; Centre Leon-B´ ´ erard, Lyon, France; CNRGH, Evry, France; Fondation Synergie Lyon Cancer, Lyon, France; Institut Curie, Paris, France; Departement ´ d’Oncologie Medicale, ´ Centre L´eon B´erard, Lyon, France; Centre François Baclesse, Department of Medical Oncology, Caen, France; Unicancer, Paris, France; University of California, Los Angeles, Los Angeles, CA; Dana-Farber Cancer Institute, Boston, MA

Background: Many BC survivors report fatigue. The relevant genomic correlates of fatigue after BC are not well understood. We applied a previously validated machine learning methodology (Oh 2017) to GWAS data to identify biological correlates of fatigue induced after tx. Methods: We analyzed 3825 BC pts with GWAS data (Illumina InfiniumExome24 v 1.1) from the CANTO study (NCT01993498). The outcome of this study was post-tx fatigue 1 year after the end of primary chemotherapy/radiotherapy/surgery using the EORTC C30 fatigue subscale (overall fatigue) and the EORTC FA 12 fatigue domains (physical/emotional/ cognitive). For each domain, we limited the study group to those with zero baseline fatigue and defined severe fatigue change as score increase above the third quartile. We tested univariate correlations between severe fatigue in each domain and 496539 SNPs as well as relevant clinical variables. The machine learning prediction model based on preconditioning random forest regression (PRFR) (Oh et al., 2017), was then built using the SNPs with ancestry adjusted univariate p-value , 0.001 and clinical variables with Bonferroni adjusted p-value , 0.05. The model was validated in a holdout subset of the cohort. Gene set enrichment analysis (GSEA) was performed using MetaCore to identify key biological correlates relevant to tx-induced fatigue. Results: Distinct results were found by fatigue domain (table). GSEA showed that the cognitive fatigue model SNPs included biomarkers for cognitive disorders (p = 1.6 x 10-12) and glutamatergic synaptic transmission (p = 1.6 x 10-8). Conclusions: A SNP based model had differential performance by fatigue domain, with a potential genetic role on risk and biology for tx induced cognitive fatigue. Further research to explore biomarkers of tx induced fatigue are needed.

PRFR performance SNPs AUC (p) Samples with Fatigue (train/holdout, Event rate Significant clinical variables p < 0.001 SNP + domain N) (%) (N) (N) SNP only clinical Overall 377/161 13 0 309 0.42 NA (0.89) Physical 283/121 19 0 277 0.44 NA (0.78) Emotional 515/220 21 3 257 0.42 0.42 (0.96) (0.96) Cognitive 820/351 17 6 299 0.59 0.61 (0.003) (0.01)

Funding Source: French National Agency for Research (ANR), Susan Komen for Cure (CCR17483507) to IVZ

11516 Clinical Science Symposium, Sat, 8:00 AM-9:30 AM

Using machine learning to predict mortality in older patients with cancer: Decision tree and random forest analyses from the ELCAPA and ONCODAGE prospective cohorts.

Etienne Audureau, Pierre-Louis Soubeyran, Claudia Martinez-Tapia, Carine A. Bellera, Sylvie Bastuji- Garin, Pascaline Boudou-Rouquette, Muriel Rainfray, Anne Chahwakilian, Thomas Grellety, Olivier Hanon, Simone Mathoulin-Pelissier, ´ Elena Paillaud, Florence Canoui-Poitrine; Public Health Department, Hospital Henri Mondor, AP-HP, Creteil, ´ France, EA7376 CEpiA (Clinical Epidemiology and Ageing Unit), UPEC, Creteil, ´ France; Institut Bergonie, ´ Bordeaux, France; EA 7376 CEpiA (Clinical Epidemiology and Ageing Unit), Creteil, ´ France; Institut Bergonie, ´ Comprehensive Cancer Centre, Bordeaux, France; Department of Medical Oncology, ARIANE, Cochin Hospital, Paris Descartes University, AP-HP, CARPEM, Paris, France; Department of Clinical Gerontology, Bordeaux University Hospital Xavier Arnozan, Pessac, France; Broca Hospital, AP-HP, Paris Descartes University, Sorbonne Paris Cite, ´ Paris, France; Geriatric Department, Broca Hospital, APHP, Paris, France; INSERM CIC 14.01 Bordeaux, Clinical Epidemiology Unit, Bordeaux, France; Geriatric Department, Georges Pompidou European Hospital, Paris, France

Background: Accurate prognosis is crucial to decision making in oncology, but remains challenging in older patients due to the heterogeneity of this population and the lack of ability of current models to capture complex interactions between oncological and geriatric predictors. We aimed to develop new predictive algorithms based on machine learning to refine individualized prognosis in older patients with cancer. Methods: Data were collected from 3409 patients $70 years referred to geriatric oncology clinics for completion of a geriatric assessment (GA), including 2012 and 1397 patients from the ELCAPA (training set) and ONCODAGE (validation set) French prospective cohorts, respectively. Candidate predictors included baseline oncological and geriatric parameters, G-8 score and routine biological data (CRP/albumin ratio). Prognostic models for 12-months mortality were built using Cox regression model, single decision tree (DT) and random survival forest (RSF). Models performance was compared based on externally validated Harrell’s C-indexes. Results: During the 1-year study period, 875 (43%) and 219 (16%) patients died in the training and validation sets, respectively (mean age: 8166/7865, women 47% / 70%, metastasis 50% / 34%). Cox model identified 9 independent predictors of mortality: tumor site/metastatic status, anticancer treatment, weight loss . 3kg, drugs . 5, renal failure, increased CRP/Albumin, ECOG-PS$2, ADL#5 and altered TGUG. DT identified more complex combinations between features, yielding 16 patient groups with highly differentiated survival, notably depending on the G-8 ( , 10 vs. $10 as the root node). RFS had the highest C-index (0.86 [RFS], 0.82 [Cox], 0.81 [DT]), identifying the G-8, CRP/albumin and tumor site/metastasis as the most important features. Conclusions: While Cox modeling confirmed known independent prognostic factors, DT revealed more complex interactions between them and random forest achieved superior prognostic performance by better capturing patient’s complexity. The latter model has been implemented into an interactive web interface for easy and direct use in clinical practice. Clinical trial information: NCT02884375.

11517 Poster Discussion Session; Displayed in Poster Session (Board #209), Mon, 1:15 PM-4:15 PM, Discussed in Poster Discussion Session, Mon, 4:30 PM-6:00 PM

Impact of cancer on physical and mental activities of daily living in young adult (YA) survivors.

Tyler Garrett Ketterl, Mary S. McCabe, Donald L. Rosenstein, Linda A. Jacobs, Steven C Palmer, Patricia A. Ganz, Jacqueline N. Casillas, Betsy C Risendal, Linda Overholser, Ann H. Partridge, Karen L Syrjala, Kevin Scott Baker; Fred Hutchinson Cancer Research Center, Seattle, WA; Memorial Sloan Kettering Cancer Center, New York, NY; UNC Lineberger Comprehensive Cancer Center, Chapel Hill, NC; Abramson Cancer Center, Philadelphia, PA; University of Pennsylvania, Philadelphia, PA; University of California, Los Angeles, Los Angeles, CA; University of Colorado Cancer Center, Aurora, CO; University of Colorado School of Medicine, Denver, CO; Dana-Farber Cancer Institute, Boston, MA; Fred Hutchinson Cancer Research Center & University of Washington, Department of Public Health, Seattle, WA

Background: Young Adult (YA) survivors face challenges unique from those survivors of childhood cancer or of middle and older age adults. The potential impact of cancer, its treatment or the lasting effects on daily mental and physical tasks are not fully understood. Methods: Eligibility included diagnosis of malignancy between ages 18-39, 1-5 years from diagnosis and $1 year from therapy completion. Participants were randomly selected from tumor registries of 7 academic institutions. Enrolled participants were asked to complete an online patient reported outcomes (PRO) survey related to the effects of cancer and its treatment on daily functionality, finances, and cancer-related distress. All enrolled subjects had diagnostic and treatment information abstracted by a standardized protocol and entered into a database. Results: Enrollment and the online PRO survey was completed by 872 survivors (43.5% of eligible survivors). Survivors were 72.3% female, 90% non-Hispanic or Latino and the most common cancer for males and females were testicular and breast cancer respectively. Cancer, its treatment or the lasting effects limited the kind or amount of instrumental activities of daily living (IADLs) in 649 (76.3%) of survivors. Exposure to chemotherapy was associated with a 3.2-fold increased odds of interference with IADLs (CI 2.08-4.83, p , 0.01) after controlling for diagnosis, age, gender, radiation and surgery exposure. This impairment lasted longer than 1 year in 268 (41.3%) of survivors. Cancer, its treatment or the lasting effects interfered with their ability to perform any mental tasks as part of their IADLs in 454 (53.3%) of survivors. Exposure to chemotherapy was associated with a 2.8-fold increased odds of interference with mental impairment of IADLs (CI 1.94-3.96, p , 0.01) after controlling for diagnosis, age, gender, radiation and surgery exposure. Exposure to radiation was also associated with an increased odds of interference with mental impairment of daily activities (CI 1.05-1.91, p , 0.01) after controlling for diagnosis, age, gender, chemotherapy and surgery exposure. Conclusions: In YA cancer survivors, cancer, its treatment or the lasting effects of treatment commonly impact the daily physical and mental IADLs and many survivors report ongoing limitations . 1 year from therapy completion.

11518 Poster Discussion Session; Displayed in Poster Session (Board #210), Mon, 1:15 PM-4:15 PM, Discussed in Poster Discussion Session, Mon, 4:30 PM-6:00 PM

Effects of a structured intervention program to improve physical activity (PA) of adolescents and young adult cancer survivors (AYAs): Final results of the randomized Motivate AYA–MAYA trial.

Jannike Lisa Salchow, Wiebke Jensen, Barbara Koch, Julia von Grundherr, Simon Elmers, Gabriele Escherich, Rudiger ¨ Reer, Carsten Bokemeyer, Julia Mann, Alexander Stein; University Cancer Center Hamburg, University Medical Center Hamburg-Eppendorf, Hamburg, Germany; University Medical Center Hamburg-Eppendorf, Hamburg, Germany; Institute of Human Movement Science, University of Hamburg, Hamburg, Germany; University Hospital Hamburg-Eppendorf, Hamburg, Germany; University Medical Center Hamburg-Eppendorf, Department of Oncology, Haematology, Stem Cell transplantation and Pneumology, Hamburg, Germany

Background: Major cardiovascular (CV) events are the most common late toxicities among AYAs. Although regular PA of vigorous intensity ($ 9 metabolic equivalent [MET]-hours/week) lowers the risk for CV events and mortality, no larger randomized controlled trials on interventions are available. Our aim was to assess whether a 12-week structured intervention increases the vigorous PA of AYAs. Methods: AYAs aged 15 to 39 years, after curative intent cancer treatment with at least one CV risk factor, were randomized to usual- care control group (CG) and to intervention group (IG). The CG received standard recommendations, and the IG participated at a semi-structured interview and phone consulting focusing on PA and behavioral change. At baseline, post-intervention (12 wks), and at follow-up (52 wks), participants completed the International Physical Activity Questionnaire (IPAQ) and quality of life assessment (EORTC QLQ-C30). Primary endpoint was the rate of AYAs with $ 9 MET-hours/week of vigorous activity (IPAQ) at 12 weeks. This single center trial was registered (DRKS00009453). Results: Among 115 screened AYA 89 eligible patients were randomized; 69 (77.5%) completed the intervention and the endpoint assessment; 36 (52.2%) were in the IG and 33 (47.8%) in the CG. Median age was 24.3 years (range, 18 to 39). CV risk factors were use of anthracyclines (94.2%), chest radiation (47.8%), or both (44.9%). At baseline 49.2% of all AYAs reported to perform vigorous PA with at least 9 MET-hours/week, although reporting was individually biased. Post-intervention this rate significantly increased in the IG from 45.7% to 69.7% (p = 0.007), whereas in the CG only a modest non-significant increase was noted (53.3% to 65.6%, p = 0.134). Notably, upon long-term follow up (52 wks) AYAs did not keep their increased vigorous PA, whereas improved moderate PA was achieved (MET score in IG p = 0,044). Also, both groups reduced significantly the time they spent sitting from 6.5 (SD, 2.9) to 5.4 (SD, 2.7) hours/day (p = 0.001). Conclusions: Intensified PA counseling improves short term vigorous PA and long term moderate PA of AYAs and, should thus be part of survivorship programs. Further studies with AYAs will be required to establish reliable PA screening methods and to confirm the results in larger cohorts. Clinical trial information: DRKS00009453.

11519 Poster Discussion Session; Displayed in Poster Session (Board #211), Mon, 1:15 PM-4:15 PM, Discussed in Poster Discussion Session, Mon, 4:30 PM-6:00 PM

Predictors of suicide risk in adolescent and young adults (AYA) with cancer.

Jeremy Howard Lewin, Kate Thompson, Donovan Moncur, Sam Mancuso; Department of Cancer Medicine, Peter MacCallum Cancer Centre, Melbourne, Australia; ONTrac at Peter Mac Victorian Adolescent & Young Adult Cancer Service, Melbourne, Australia

Background: Whilst AYA accounts for a small proportion of annual cancer cases, the complexity of this subgroup places them at a disproportionately high risk of psychological distress and poor mental health. This study aimed to identify patient and disease characteristics associated with suicide in AYA patients. Methods: A retrospective analysis of AYA (aged 15-39) between the years 1973 to 2015 from the Surveillance, Epidemiology, and End Result Research database were included. A training and test set were used to develop a model predicting suicide. The SMOTE sampling algorithm was used for the training set due to severe class imbalances. A random forest model was trained with 10 clinical features (cancer subtype, age, sex, race, marital status, diagnostic year, number of in situ/malignant tumours, chemotherapy, surgery, and radiation therapy) with recursive feature elimination (RFE) via 10-fold cross validation. Results: There were 139,394 AYA included with 974 (0.7%) having a documented suicide or self/inflicted injury yielding an incidence of 1499 suicides / 100,000 person years. The standardized mortality ratio (SMR) for AYA was 34.4 (95% CI: 32-37)(Aged 15-24: 47.4 (39-57); Aged 25-39: 32.3 (29-35)). Suicide rates increased over time (1973-1980: SMR = 18.0 (15-22); 2001-2015: SMR = 127.0 (102-155)). Univariate analyses observed high rates of suicide in: females (43.8 (38-51)); single/ unknown relationship status (47.1 (42-52)); “other” race (85.2 (52-132)) and by cancer type: leukaemia (77.1 (42-129)), soft tissue sarcoma (73.7 (57-94)), unspecified malignancy (71.0 (9-256)) and brain (63.6 (43-91)). We randomly assigned AYAs to a training set (n = 97,576) and test set (n = 41,818), stratified by cause of death. All 10 clinical features were retained with RFE. The prediction model achieved an AUC of 0.59, accuracy of 0.78, sensitivity of 0.40 and specificity of 0.78. Conclusions: Risk of suicide in AYA is high compared to normative data, further heightened in females, single/unknown relationship and certain tumour subtypes. The prediction model performed poorly, potentially related to the exclusion of mental health variables, which are not collected by the SEER dataset. Our findings suggest that dedicated psychosocial supports and targeted mental health assessments are critical components of care for AYA.

11520 Poster Discussion Session; Displayed in Poster Session (Board #212), Mon, 1:15 PM-4:15 PM, Discussed in Poster Discussion Session, Mon, 4:30 PM-6:00 PM

Candidate SNPs enhance prediction of cognitive impairment after blood or marrow transplantation (BMT) for hematologic malignancy (HM).

Noha Sharafeldin, Joshua Richman, Alysia Bosworth, Yanjun Chen, Purnima Singh, Sunita K. Patel, Xuexia Wang, Emily Morse, Molly Mather, Can-Lan Sun, Liton Francisco, Stephen J. Forman, F. Lennie Wong, Smita Bhatia; Univ Alabama at Birmingham, Birmingham, AL; University of Alabama at Birmingham, Birmingham, AL; City of Hope, Duarte, CA; City of Hope Cancer Ctr, Duarte, CA; University of North Texas, Denton, TX; City of Hope National Medical Center, Duarte, CA

Background: We tested the hypothesis that candidate genetic variants are associated with cognitive impairment in BMT recipients for HM, and that inclusion of genetic variants improves the performance of a risk prediction model that includes only clinical and demographic characteristics. Methods: We used standardized tests to assess cognitive function in 277 adult BMT recipients at City of Hope (COH). Global Deficit Score (GDS) $0.50 was used as indicator of cognitive impairment. Generalized estimating equation models and logic regression were used to identify single-SNP and gene-level associations with cognitive impairment post-BMT. Three risk prediction models were developed in the COH cohort using elastic net regression: Base Model (sociodemographics); Clinical Model (Base Model + clinical characteristics, therapeutic exposures and baseline cognitive reserve); Combined Model (Clinical + Genetic Model). The Genetic Model included significant SNPs in blood brain barrier, telomere homeostasis and DNA repair identified from single- and gene-level analyses. Models were validated in an independent cohort of long-term BMT survivors (BMTSS) with (n = 141) and without (n = 258) memory problems. Results: Training set (COH): The cohort included 58.5% males; 68.6% non-Hispanic whites; 46.6% allogeneic BMT recipients; median age at BMT: 51.6y. The mean area under the receiver operating characteristic curve (AUC) was: Base Model: 0.69 (95%CI: 0.63-0.75); Clinical Model: 0.77 (95%CI: 0.71-0.83); Combined Model: 0.89 (95%CI: 0.84-0.92). Test set (BMTSS): Median age at BMT was 45y; 53.5% were males; 88.4% non-Hispanic whites. Testing the models in BMTSS yielded mean AUC of 0.57 (95%CI: 0.49-0.63) in the Clinical Model and 0.72, (95%CI: 0.65-0.78) in the Combined Model. Conclusions: These findings provide evidence on the utility of a validated risk prediction tool that incorporates genetic factors that could identify BMT recipients at risk for cognitive impairment, providing opportunities for targeted interventions.

11521 Poster Discussion Session; Displayed in Poster Session (Board #213), Mon, 1:15 PM-4:15 PM, Discussed in Poster Discussion Session, Mon, 4:30 PM-6:00 PM

Cognitive rehabilitation program to improve cognition of cancer patients treated with chemotherapy: A randomized controlled multicenter trial.

M´elanie Dos Santos, Olivier Rigal, Isabelle Leger, ´ Idlir Licaj, Sarah Dauchy, Christelle Levy, Sabine Noal, Carine Segura, Corinne Delcambre, Djelila Allouache, Aurelie ´ Parzy, Marie Lange, Aurelie ´ Capel, Jean-Michel Grellard, Ben´ ´ edicte Clarisse, Johan Lefel, Florence Joly; Centre François Baclesse, Clinical Research Department, Caen, France; Centre Henri-Becquerel, Department of Medical Oncology, Rouen, France; Institut Gustave Roussy, Department of Psycho Oncology, Villejuif, France; Centre François Baclesse, Department of Medical Oncology, Caen, France

Background: Cognitive impairment induced by cancer chemotherapy (CT) has been identified as an important side-effect with negative impact on quality of life (QoL) without specific treatment. We evaluated the impact of computer-assisted cognitive rehabilitation (CR) on cognitive complaint, objective cognitive dysfunction and QoL among cancer patients treated with CT. Methods: We included cancer patients with cognitive complaint occurring during CT or within 5 years of the end of CT. Patients were randomly assigned in a 1:1:1 ratio to face-to-face CR with a neuropsychologist (group A), home cognitive exercises (group B) or phone follow-up (group C) with 9 sessions over 3 months. Cognition was assessed by the Functional Assessment of Cancer Therapy Cognitive Function (FACT-Cog) completed by a neuropsychological battery of test and QoL assessment by the FACT-General (FACT-G). The primary endpoint was the proportion of patients with 7-point improvement in the perceived cognitive impairments (PCI) of the FACT- Cog between baseline (T0) and the end of the program (T3). Results: 167 patients were enrolled, median age was 50 years [43-59] and 96% were women with mainly breast cancer. Compliance rate with completion of all sessions was 76, 61 and 75% respectively. Proportion of patients with 7-point PCI improvement were 73, 55 and 56% without reaching the statistically significant difference between group A and B (p = 0.07) and group A and C (p = 0.08). The mean difference in PCI score were 17, 10 and 10 (p = 0.03). Patients with CR improved their working memory with significant difference between group A and C (1.4 versus 0.3, p , 0.001) but not between group A and B (1.4 versus 1.1, p = 0.43). There was a significant impact of CR on the FACT-Cog subscale score of QoL (p = 0.01) in favor of the group A, but not on the different dimensions of the FACT-G. Patients in group A presented improvement in depression compared to group B and C: -6.5 versus -1.7 and -2.3 (p = 0.03). Conclusions: CR with a neuropsychologist improves cognitive complaint. Cognitive stimulation showed improvements in working memory. CR was associated with better QoL linked to cognitive disorders and lower levels of depression. Clinical trial information: NCT01788618.

11522 Poster Discussion Session; Displayed in Poster Session (Board #214), Mon, 1:15 PM-4:15 PM, Discussed in Poster Discussion Session, Mon, 4:30 PM-6:00 PM

Acupuncture versus cognitive behavioral therapy for cognitive impairment in cancer survivors with insomnia: Implications for personalized medicine.

Jun J. Mao, Kevin Liou, James Root, Qing Susan Li, Ting Bao, Sheila N. Garland, Tim Ahles; Memorial Sloan Kettering Cancer Center, New York, NY; Center for Clinical Epidemiology and Biostatistics, Perelman School of Medicine at the University of Pennsylvania, Philadelphia, PA; Memorial University of Newfoundland, St. John’s, NF, Canada

Background: Cognitive impairment is a prevalent condition among cancer survivors that lacks effective treatment and can be maintained and exacerbated by poor sleep. This study explored whether treating insomnia with acupuncture or Cognitive Behavioral Therapy for Insomnia (CBT-I) improves subjective and objective cognitive functions in cancer survivors. Methods: We analyzed cognitive outcomes from a pragmatic randomized trial comparing acupuncture versus CBT-I for cancer survivors with insomnia. Analysis was limited to those reporting cognitive impairment at baseline. Acupuncture and CBT-I were delivered over 8 weeks. Perceived cognitive ability was assessed using the Brown Attention-Deficit Disorder Scale (BADDS). Objective cognitive function was evaluated with the Buschke Selective Reminding Test (BSRT). All outcomes were evaluated at baseline, Week 8 (end of intervention), and Week 20 (12 weeks post-intervention). Results: Among 99 cancer survivors, mean age was 60.4 years, 56.6% were women, and 26.3% were non-white. The most common cancer types were breast (31.3%) and prostate (19.2%). Perceived cognitive ability improved in both acupuncture and CBT-I groups at weeks 8 and 20 relative to baseline (all P , 0.001). No significant between-group differences were noted in BADDS total score (p = 0.28), but the CBT-I group demonstrated a better BADDS attention subscale score than the acupuncture group at weeks 8 and 20 (p = 0.031). With regards to objective cognitive functions assessed by BSRT, acupuncture improved attention (p = 0.017), learning (p = 0.040), and memory (p = 0.0020) at Week 8, whereas CBT-I only improved attention at Week 20 (p = 0.0002); between-group differences were not statistically significant. Conclusions: Among cancer survivors with insomnia, both acupuncture and CBT-I improved cognitive impairment relative to baseline, but their relative effects differed: the CBT-I group showed slightly better subjective attention, whereas the acupuncture group may have improved objective memory. Further investigation of these two therapies may lead to effective and personalized interventions for cancer survivors. Clinical trial information: NCT02356575.

11523 Poster Discussion Session; Displayed in Poster Session (Board #215), Mon, 1:15 PM-4:15 PM, Discussed in Poster Discussion Session, Mon, 4:30 PM-6:00 PM

Improving person-centered communication of goals, proxy, and advance directives in older patients with advanced cancer: Secondary analysis from a University of Rochester NCI Community Oncology Research Program (NCORP) cluster randomized controlled trial (CRCT).

Marie Anne Flannery, Eva Culakova, Kah Poh Loh, Ronald M. Epstein, Charles Stewart Kamen, Spencer Obrecht, Nataliya Melnyk, Mary I. Whitehead, Jodi Geer, Jeffrey K. Giguere, Karen Michelle Mustian, Paul Duberstein, William Dale, Supriya Gupta Mohile; University of Rochester Medical Center, Rochester, NY; Rutgers Robert Wood Johnson Medcl School, East Brunswick, NJ; Breast Cancer Options, Sharon, CT; Park Nicollet, Minneapolis, MN; Greenville Health System, Seneca, SC; Rutgers University, New Brunswick, NJ; CIY, Chicago, IL

Background: Quality person-centered care relies on effective communication between the clinical team and the patient/caregiver eliciting goals and discussing wishes. In a PCORI- and NCI-funded CRCT, we found that providing community-based oncologists with geriatric assessment-guided recommendations led to more and higher quality discussions of age-related issues for older patients with advanced cancer. In this secondary analysis, we assessed whether specific recommendations to oncologists to discuss patient goals, proxy and advance directives resulted in increased communication about these topics. Methods: Patients aged 70+ with advanced solid tumors or lymphoma and at least one impaired geriatric domain (e.g., function, cognition) were enrolled (URCC 13070; PI: Mohile). Oncology practices were randomized to the intervention (oncologists received recommendations to elicit goals and discuss wishes) or usual care. The clinic visit after the oncologist received recommendations was recorded and transcribed; two blinded coders evaluated the transcripts for discussion of the specific topic areas recommended in the intervention. Between arm differences were compared using generalized linear models controlling for practice cluster. Results: From 2014-17, 528 patients (284 intervention) provided transcripts from 31 practices (mean age = 77, range 70-96 years; 49% female; mixed cancer diagnoses). Topics related to patient goals, proxy and advance directive wishes were more often discussed in the intervention arm (goals of care preferences: 9 vs 2%, p = .02, treatment goals: 35 vs 20%. p = .04, elicit caregiver input: 28 vs 3%. p , .01, assess values and goals: 25 vs 7%, p = .07, health care proxy: 40 vs 1%, p = .004, advance directive: 25 vs 1%, p = .002). Conclusions: In this community-based study of older adults providing recommendations to oncologists to discuss specific topics resulted in increased person-centered discussions with patients and caregivers about goals, proxy and advance directive wishes. However, the content areas were discussed in less than half of all visits. Clinical trial information: NCT02107443.

11524 Poster Discussion Session; Displayed in Poster Session (Board #216), Mon, 1:15 PM-4:15 PM, Discussed in Poster Discussion Session, Mon, 4:30 PM-6:00 PM

Prognostic understanding in hematologic malignancies: A multicenter longitudinal study.

Kah Poh Loh, Thomas William LeBlanc, Stephanie Lee, Anthony Back, Paul Duberstein, Supriya Gupta Mohile, Ronald M. Epstein, Areej El-Jawahri; University of Rochester Medical Center, Rochester, NY; Duke University Medical Center, Durham, NC; Fred Hutchinson Cancer Research Center, Seattle, WA; Seattle Cancer Care Alliance, Seattle, WA; Massachusetts General Hospital, Boston, MA

Background: Accurate prognostic understanding facilitates the receipt of goal-concordant medical care. Few studies have evaluated prognostic understanding in patients with hematologic malignancies. In this secondary analysis of a multicenter, longitudinal study of patients with hematologic malignancies referred for a second opinion hematology subspecialty consultation, we assessed changes in prognostic understanding after consultation, and predictors of post-consultation patient-oncologist prognostic discordance. Methods: Patients were recruited from 4 academic centers. Before and 1-7 days after consultation, patients were asked about their perceived chance of cure (options , 10%, 10-19%, and up to 90-100% in 10% increments, and “do not wish to answer”). Oncologists were asked the same question after consultation. Discordance was defined as a difference in response by 2 levels in the patient- oncologist dyads. We used multivariate analysis to assess the demographic and clinical predictors of patient-oncologist discordance. Results: We included 216 patients (median age 55 years, range 22-79) and 46 oncologists (47, 30-70). Overall, $On multivariate analysis, discordance before consultation [Odds Ratio (OR) 6.05, 95% Confidence Interval (CI) 2.96-12.36) and , college education (vs. post- graduate education; OR 2.34, 95% CI 1.09-5.14) were associated with discordance after consultation. Other patient demographics, comorbidity, cancer type, psychological distress, social support, decision- making preference, and coping strategies were not associated with discordance. Conclusions: Patient- oncologist concordance in prognostic understanding improved after subspecialty consultation, but over half of patients’ views of their prognosis remained discordant with those of their oncologists. Interventions to improve patient-oncologist communication about prognosis are needed, especially in patients with lower education level.

11525 Poster Discussion Session; Displayed in Poster Session (Board #217), Mon, 1:15 PM-4:15 PM, Discussed in Poster Discussion Session, Mon, 4:30 PM-6:00 PM

Comparing nonverbal synchrony in racially concordant and racially discordant oncology interactions.

Lauren M. Hamel, Robert Moulder, Susan Eggly, Terrance Lynn Albrecht, Steven Boker, David W. Dougherty, Louis Penner; Karmanos Cancer Center, Wayne State University, Detroit, MI; University of Virginia, Charlottesville, VA; Wayne State University, Detroit, MI; Dana-Farber Cancer Institute, Boston, MA

Background: Communication in racially discordant (Black patient, non-Black physician) oncology interactions, which constitute about 80% of Black patients’ interactions, is generally poorer than in racially concordant interactions, and likely contributes to treatment disparities. However, the nonverbal behaviors that contribute to this problem are largely unknown. We examined nonverbal synchrony, or the nonconscious coordination of movement, which can reflect relationship quality. We hypothesized that racially discordant interactions will have lower levels of nonverbal synchrony. Methods: Data include video recordings of 68 Black patients and 163 White patients discussing treatment with their non-Black oncologists. Recordings were submitted to motion detection software to measure nonverbal synchrony. This software measures global synchrony (all correlated motion), peak synchrony (all positively correlated motion), who is leading the interaction (similar to who is leading in ballroom dancing), and how much synchrony occurs based on who is leading the interaction. Using multi-level models, we investigated whether nonverbal synchrony differed in racially concordant and racially discordant dyads. Results: Findings showed greater levels of global synchrony (p , .05) and greater peak synchrony (p , .05) in racially discordant interactions compared to racially concordant interactions. Global synchrony was the same in racially concordant interactions regardless of who was leading, but greater global synchrony occurred in racially discordant interactions when the patient was leading (p , .05). Conclusions: This is the first study to use a dynamic jointly determined measure of behavior to assess oncology interactions. Contrary to our hypothesis, nonverbal synchrony was greater in racially discordant interactions than in racially concordant interactions. It appears patients are driving more of the synchrony in racially discordant interactions. This may suggest that patients in racially discordant interactions adapt to their physicians to bridge racial differences. Findings could contribute to physician training to enhance coordination and outcomes in oncology interactions.

11526 Poster Discussion Session; Displayed in Poster Session (Board #218), Mon, 1:15 PM-4:15 PM, Discussed in Poster Discussion Session, Mon, 4:30 PM-6:00 PM

Accelerated sarcopenia and outcomes in older adults with cancer: The Health ABC Study.

Grant Richard Williams, Yanjun Chen, Kelly Kenzik, Andrew Michael McDonald, Shlomit S. Shachar, Heidi D. Klepin, Stephen Kritchevsky, Smita Bhatia; University of Alabama at Birmingham, Birming- ham, AL; Department of Radiation Oncology, University of Alabama at Birmingham, Birmingham, AL; The University of North Carolina at Chapel Hill, Chapel Hill, NC; Comprehensive Cancer Center, Wake Forest Baptist Health, Winston Salem, NC; Wake Forest University School of Medicine, Winston-Salem, NC

Background: Progressive loss of muscle mass and strength (sarcopenia) is a well-known phenomenon of aging; however, little is known about the contribution of a cancer diagnosis to sarcopenia and its subsequent impact on disability. Using a prospective cohort of older adults from pre- to post-cancer diagnosis and a similarly-followed non-cancer cohort, we examined the trajectory of sarcopenia measures and their association with overall survival (OS) and major disability among those with cancer. Methods: The Health, Aging, and Body Composition (Health ABC) Study is a prospective longitudinal study where 3,075 community-dwelling older adults (70-79y) underwent 6 annual assessments of body composition and were followed for development of sentinel events (cancer, disability, death). Appendicular lean mass (ALM [kg]) was a sum of DXA-based lean tissue of all extremities. Hand grip strength (HGS [kg]) was averaged from 2 trials per hand. Gait speed (GS) was evaluated over a 20m course. We used linear mixed effect models to compare the change in ALM, HGS, and GS between individuals who subsequently developed cancer and those who did not, adjusting for age, race, gender, enrollment site. Among patients with cancer, we used multivariable cox regression for time from cancer diagnosis to mortality and major disability (cane/walker, inability to walk 0.25 mile/climb 10 steps, assistance with activities of daily living) treating sarcopenia measures as time- varying covariates. Results: Mean age at enrollment was 75y; 52% female; 42% black; 515 new cancers (prostate: 23%, colorectal: 15%, lung: 13%, breast: 11%). Compared with non-cancer controls, we found significantly steeper declines in HGS (p= 0.03) and GS (p, 0.001), and a trend in ALM (p= 0.07) prior to cancer diagnosis; and a significantly steeper decline in ALM (p, 0.001), but no difference in HGS (p= 0.6) or GS (p= 0.4) after cancer diagnosis. Slow GS was associated with a 44% increase in mortality (p= 0.02) and a 70% increase in disability (p= 0.02), but not ALM or HGS. Conclusions: Accelerated loss in sarcopenia measures both prior to and after a cancer diagnosis, and association with disability and mortality in older adults with cancer, present opportunities for targeted interventions.

11527 Poster Discussion Session; Displayed in Poster Session (Board #219), Mon, 1:15 PM-4:15 PM, Discussed in Poster Discussion Session, Mon, 4:30 PM-6:00 PM

Age disparities among cancer clinical trial participants: The role of industry sponsorship.

Ethan B. Ludmir, Walker Mainwaring, Austin B. Miller, Timothy Lin, Amit Jethanandani, Andres F. Espinoza, Clifton David Fuller; The University of Texas MD Anderson Cancer Center, Houston, TX; Baylor College of Medicine, Houston, TX; The University of Texas Health Science Center McGovern Medical School, Houston, TX; The University of Tennessee Health Science Center College of Medicine, Memphis, TN

Background: Randomized controlled trials (RCTs) in oncology, which often establish the standard of care for cancer patients, do not necessarily enroll trial participants representative of the broader patient population. We sought to characterize age disparities for cancer patients enrolled on RCTs, and asked whether certain trial-related factors (such as industry sponsorship) predispose trials to larger age disparities between trial enrollees and the general population. Methods: All phase 3 RCTs in clinical oncology with results available were identified through ClinicalTrials.gov. Only randomized multi-arm trials assessing a therapeutic intervention for cancer patients were included. The scope of trials was limited to breast, colorectal, lung, and prostate disease sites. Trial participant median ages were compared to national SEER data for population median ages by disease site. Results: Three-hundred and two trials met inclusion criteria. For all trials, the trial median age was an average 6.23 years younger than the population median age (95% CI: -5.55 to -6.91 years, p , 0.001). Trials with industry sponsorship had significantly younger trial patient populations compared with non-industry-sponsored trials (mean difference from population -6.57 vs -4.48 years, p = 0.02). Younger patients were enrolled on trials evaluating targeted agents (p = 0.04), superiority-design trials (p = 0.02), and trials utilizing a surrogate primary endpoint (p = 0.03). By disease site, lung cancer trials enrolled the youngest patients relative to the population median (-8.98 years), followed by breast (-7.76 years), colorectal (-6.96 years), and prostate (+2.66 years, p , 0.001). Conclusions: Industry-funded clinical trials are associated with larger age disparities among trial participants; we believe this represents a novel finding both in clinical oncology and academic medicine more broadly. Underrepresentation of older patients on RCTs has major ramifications for the generalizability of results as well as equity of access to care; future efforts to address trial participant age disparities may focus on those trials at greatest risk for disparities based on the identified risk factors, including industry sponsorship.

11528 Poster Discussion Session; Displayed in Poster Session (Board #220), Mon, 1:15 PM-4:15 PM, Discussed in Poster Discussion Session, Mon, 4:30 PM-6:00 PM

Cumulative burden of new-onset chronic health conditions (CHCs) among older cancer survivors.

Kelly Kenzik, Grant R. Williams, Nickhill Bhakta, Leslie L. Robison, Smita Bhatia; University of Alabama at Birmingham, Birmingham, AL; St. Jude Children’s Research Hospital, Memphis, TN

Background: In the US there are an estimated 11 million survivors of cancer diagnosed at $65y of age. Description of the morbidity in these survivors has been limited to single complications or to prevalence of comorbidities. The cumulative burden of CHCs remains unstudied, and is critically needed to inform healthcare delivery in this burgeoning population. Methods: Using SEER-Medicare, we identified 300,082 patients with breast (34%), prostate (33%), colorectal (16%), non-small cell lung (NSCLC 10%) or non-Hodgkin lymphoma (NHL 7%) diagnosed between 2000 and 2011 at age . 65y (mean age at diagnosis: 75y; 47% males, 88% non-Hispanic whites). An age-, race-, and sex-similar non- cancer cohort (n = 97,842) was assembled. New-onset non-malignant health conditions (n = 109) were consolidated into 10 organ-specific CHCs. Inpatient CHC visits were used to describe severe CHCs. The cumulative incidence (CI) and cumulative burden (CB) of CHCs was described up to 10y from cancer diagnosis and by attained age – up to six months prior to death or until 12/31/2013. Subsequent malignant neoplasms (SMNs) were described 10y from primary cancer diagnosis. Results: The 10y CI of any CHC and severe CHC was 98% (95%CI 98-99%) and 73% (72-73%) in cancer patients and 92% (91-92%) and 55% (54-55%) in non-cancer controls (hazard ratio [HR10y]: 1.65, 95%CI 1.64-1.66). Cardiovascular conditions were the largest contributor to severe non-malignant CHCs (10y CI: 49%- 69%). Prostate cancer survivors had the highest 10y CI for SMNs (19.4%). The CI for severe CHCs was 44% by age 80y and 85% by age 90y, compared to 34% and 54% in controls (p , 0.001). The 10y CB of CHCs was highest among NSCLC (42 CHCs/survivor) and NHL (41 CHCs/survivor) survivors; in comparison, the 10y CB was 31 CHCs/individual in controls. Colorectal cancer survivors had greatest overall burden at age 80y (27 CHCs/survivor) and 90y (36 CHCs/survivor), compared to 13 and 16 CHCs/individual in controls. Conclusions: The cumulative burden of new-onset multimorbidity among older cancer survivors is substantially greater than that of their non-cancer counterparts, providing quantifiable evidence that survivor-adapted healthcare management policies and risk-based interventions are needed.

11529 Poster Session (Board #221), Mon, 1:15 PM-4:15 PM

Early mortality after resection of locally advanced rectal cancer in elderly United States patients.

Helmneh M. Sineshaw, K Robin Yabroff, Vassiliki Liana Tsikitis, Ahmedin Jemal, Timur Mitin; American Cancer Society, Atlanta, GA; Oregon Health and Science University, Portland, OR

Background: Early mortality after resection of locally advanced rectal cancer in patients age 75 and older has not been studied in the United States. This information could inform clinical decision-making for patients who achieve complete clinical response after neoadjuvant therapy and consider watchful waiting versus surgical resection. Methods: Using the National Cancer Data Base, we identified patients age 75 years and older who underwent surgery for clinical stage II or III rectal cancer between 2004- 2015. We performed multivariable logistic regression analyses to assess associations between patient and facility characteristics and 30-day, 90-day, and 6-month mortality. Results: Among 11,326 patients, 94% underwent resection and the remaining 6% underwent local excision. Overall early mortality rates after surgery were 4%, 7.6% and 11% for 30-day, 90-day and six-month, respectably. Six-month mortality varied by age subgroup (8% in 75-79 years old to 17.7% in 85 years and older), and comorbidity score (9.5% for comorbidity score = 0 to 18.5% for comorbidity score $ 2). Between 2004 and 2015, six-month mortality declined significantly from 11.9% in 2004-2007 to 9.8% in 2012-2015 (P trend = 0.0029), with the decline larger among patients age 85 years and older (from 19.4% in 2004-2007 to 15.3% in 2012-2015, P trend = 0.0377). In the multivariable analysis, older age, higher comorbidity score, and lower facility case volume were significantly associated with higher odds of six-month mortality. Patients treated at National Cancer Institute (NCI) designated centers had 32% lower odds of six-month mortality compared with those treated at non-NCI designated teaching/ research centers. Conclusions: Post-operative six-month mortality among patients age 75 years and older with locally advanced rectal cancer in the US declined steadily over the past decade. Older age, high comorbidity score, and low facility case volume were associated with higher six-month mortality after surgery. Additional efforts are needed to guide elderly patients and their physicians in discussing treatment options for locally advanced rectal cancer.

11530 Poster Session (Board #222), Mon, 1:15 PM-4:15 PM

Is there a shift in use of subacute rehabilitation (SAR) instead of hospice referral since immunotherapy became available?

Jonathan Yeh, Louise Knight, Joyce M. Kane, Danielle Doberman, Arjun Gupta, Thomas J. Smith; Johns Hopkins Program in Palliative Care, Baltimore, MD; Director of Patient and Family Services, The Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins, Baltimore, MD; The Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins University, Baltimore, MD; University of Texas Southwestern Medical Center, Dallas, TX; Johns Hopkins University School of Medicine, Baltimore, MD

Background: Immunotherapy has rapidly become mainstream treatment. Since the first drug approval in 2011, we have noted a decline in referrals from inpatient oncology to hospice, and an increase in referrals to sub-acute rehabilitation (SAR) facilities, possibly with the aim of “getting strong enough” for immunotherapy and other promising drugs. This study explores outcomes after discharge to SAR, including rates of cancer-directed therapy after SAR, overall survival, and hospice utilization. Methods: Electronic chart review of patients discharged from oncology units to SAR facilities from 2009-2017. Demographics, admission statistics, and post-discharge outcomes were gathered from discharge summaries and targeted chart searches. Results: SAR referrals increased from 28 in 2012 to 82 in 2016. Age 66, males 52%, solid tumors 58%. 358 patients were referred to SAR 413 times. 174 patients (49%) returned to the oncology clinic prior to re-admission or death, and only 117 (33%) ever received further cancer-directed treatment (chemotherapy, radiation, or immunotherapy). 219 of 358 (61%) died within 6 months. Only 3 individuals who were not on immunotherapy at time of admission went on to receive immunotherapy after discharge to SAR. Among all discharges, 28% led to readmissions within 30 days. 74 patients (21%) were deceased within 30 days, of whom only 31% were referred to hospice. Palliative care involvement resulted in more frequent do not resuscitate (DNR) code status (33 v 22%), documented goals of care (GOC) discussions (81 v 23%), and electronic advance directives (42 v 28%).(All p,0.05). Conclusions: A growing number of oncology inpatients are being discharged to SAR, but two-thirds do not receive further cancer therapy at any point, including a substantial fraction that are re-admitted or deceased within 1 month. Many patients lose the opportunity to use hospice for optimal end of life care, as few SAR facilities offer this. These data can help guide decision-making and discharge planning that aligns with patients’ goals of care. More clinical data are needed to predict who is most likely to benefit from SAR and proceed to further cancer therapy.

11531 Poster Session (Board #223), Mon, 1:15 PM-4:15 PM

End-of-life care and immune checkpoint inhibitors.

Hazel O’Sullivan, Dearbhaile Collins, Deirdre O’Mahony, Derek Power, Richard Bambury, Seamus O’Reilly; Cork University Hospital, Cork, Ireland

Background: In the era of cytotoxic chemotherapy, aggressive cancer treatment and hospitalization at the end of life (EOL) has been associated with a worse quality of death. Meanwhile, in the era of immunotherapy (IO), little is known of the impact of these novel agents on EOL care. The aim of this study was to evaluate the EOL care of metastatic cancer patients treated with immune checkpoint inhibitors. Methods: We conducted a retrospective analysis of patients prescribedPD1/L1 or CTLA-4 antibodies in Cork University Hospital (CUH) and Mercy University Hospital (MUH) between January 2013 to December 2018. Patients treated on a clinical trial were excluded. Results: We identified 224 patients treated with immune checkpoint inhibitors (outside of a clinical trial) in CUH and MUH over the described 6 year period. 108 of these patients were deceased, 102 electronic files were available for analysis. Of the 102 patients, 57 had metastatic melanoma, 33 non small cell lung cancer, 8 renal cell carcinoma, 4 had other advanced malignancies. 43% were female and 57% were male. 6% of patients had an ECOG performance status (PS) of 0 at diagnosis, 80% PS of 1 and 10% PS of 2. Median age at death was 62 years. 47 patients were treated with pembrolizumab, 26 nivolumab, 25 ipilimumab, 2 nivolumab/ipilimumab and 2 received atezolizumab. 29 patients received IO as first line treatment, 50 as second line, 17 as third line and 6 as fourth line. Median number of IO cycles received was 4 (range 1 - 41). Progression of disease (62%) and declining performance status (14%) were the most common reasons for discontinuation of IO treatment. 16 of the 102 patients received a further line of systemic therapy. Median time from last dose of IO to death was 57 days. 20 patients (20%) died within 30 days of last dose of IO. Of these 20 patients, the median number of cycles of IO received was 2 (range 1-7), 8 of these 20 patients received one cycle of IO only. 39 patients (38%) attended the ED in the last month of life. 47 (46%) patients had at least one hospital admission in the last month of life, the median hospital length of stay was 6 days (range 1-30) and 22 patients died in hospital. 94% of patients were referred to palliative care, the median time from palliative care referral to death was 64 days (range 1- 1010), 62% of patients died in hospice. Conclusions: Patients with advanced cancer treated with immunotherapy have high rates of hospital admissions and ED attendances despite early palliative care involvement. 20% of patients died within 30 days of IO. More research is needed to help physicians identify patients who are least likely to benefit from IO so as not to treat futile cases.

11532 Poster Session (Board #224), Mon, 1:15 PM-4:15 PM

Goals of care designation associated with improved survival and indicators of quality end- of-life care in pancreatic cancer (PC) patients (pts) undergoing palliative chemotherapy.

Matthew Anaka, Jennifer L. Spratlin, Winson Y. Cheung, Sunita Ghosh, Minji Lee; University of Alberta, Edmonton, AB, Canada; Alberta Health Services, Edmonton, AB, Canada; BC Cancer Agency, Vancouver, BC, Canada

Background: Discussion of goals of care (GoC) is a key part of quality care for pts with palliative cancer. Numerous studies have shown that documentation of GoC in this population remains low. Here we describe changes in GoC documentation and other indicators of quality end-of-life care in PC pts undergoing palliative chemotherapy during a health-system wide initiative to improve advanced care planning (ACP). Methods: This is a retrospective cohort analysis of 106 pts with locally advanced or metastatic PC treated with palliative chemotherapy from 2012-2015 in Northern Alberta (Canada). In 2014, an initiative was launched to provide pts with hard copies of their GoC designation intended to be available at all health-system interactions. Data were obtained from outpatient medical oncology (MO) and palliative care (PAL) notes and the provincial cancer registry. Survival analysis used a multivariate Cox-regression. All other tests were Chi-squared. Results: 50% (53/106) of pts had a documented GoC discussion, with 45% (48/106) receiving a specific GoC designation. In 2012, 31% (6/19) of pts had a GoC designation, which increased to 61% (20/33) in 2015. Of 84 individual GoC discussions documented, 34% (29/84) were by MO, 62% (52/84) were by PAL, and at least 8% (7/84) referenced prior discussions with a family physician or discussion while admitted to hospital. Pts with a GoC designation had increased median overall survival (287 vs 216 days; HR = 0.62; p = 0.041), and were less likely to receive chemotherapy in the last two weeks of life (p = 0.016). Conclusions: Rates of GoC discussions for PC pts undergoing palliative chemotherapy increased during a health-system wide ACP initiative. Having a GoC designation was associated with greater overall survival and indicators of higher quality end-of-life care.

11533 Poster Session (Board #225), Mon, 1:15 PM-4:15 PM

Immune checkpoint inhibitor use near the end of life.

Chad Glisch, Travis Snyders, Stephanie Gilbertson-White, Yuya Hagiwara, Mohammed M. Milhem, Laurel Lyckholm; University of Iowa, Iowa City, IA; University of Iowa College of Nursing, Iowa City, IA

Background: Studies of chemotherapy near the end of life reveal increased costs, adverse effects and minimal clinical benefit. Immune Checkpoint Inhibitor (ICI) use near the end of life has not been described. We studied factors related to ICI use near the end of life. Methods: We conducted a single- institution retrospective chart review of patients who received ICI and died between August 2014 and December 2018. End of life ICI (EOL-ICI) was defined as treatment within the last 30 days of life and comparisons were made to patients who received treatment . 30 days from end of life (non EOL-ICI). Results: 441 patients were reviewed. Mean age was 64 and 182 (41%) were female. 294 (67%) received ICI within the last 90 days of life and 117 (27%) within the last 30 days of life. At time of last ICI, 145 (33%) had brain metastases and 416 (94%) were stage 4. The most common cancers were melanoma (124, 28%), NSCLC (118, 27%) and urothelial carcinoma (34, 8%). The EOL-ICI group had a higher proportion of patients with ECOG $3 at time of last treatment (22% vs. 7%, p = , .001), higher rate of death in hospital (32% vs. 18%, p = 0.003) and lower hospice enrollment (52% vs. 76%, p=, .001). The EOL-ICI group received fewer ICI doses (mean 5.1 vs 6.7, p = 0.016). A higher proportion of patients in the EOL-ICI group were just beginning treatment and received # 3 doses (60% vs 40%, p , .001). There was no difference in mean age or presence of brain metastases between the groups. Even when the cutoff for EOL-ICI is extended from 30 to 90 days, there remain significant differences in ECOG, hospice enrollment and starting ICI but no difference in rate of dying in the hospital. Conclusions: One out of four patients studied received ICI within the last 30 days of life. EOL- ICI treatment is associated with higher rates of death in the hospital and lower hospice enrollment. Our results suggest that performance status is important when considering treatment with ICI. Use of ICI in the last 30 days of life had minimal clinical benefit and high cost.

11534 Poster Session (Board #226), Mon, 1:15 PM-4:15 PM

Unraveling the high cost of end-of-life care: An oncology care model experience.

J. Russell Hoverman, Cynthia B. Taniguchi, Jad Hayes, Kathryn Eagye, B. Brooke Mann, Marcus A. Neubauer; Texas Oncology, Dallas, TX; US Oncology Network/McKesson Specialty Health, The Woodlands, TX; McKesson Specialty Health, Westminster, CO; The US Oncology Network, McKesson Specialty Health, Seattle, WA

Background: Simply put, cancer care in the last 30 days of life (30DAY) is often aggressive and expensive without best meeting the wishes and needs of patients and caregivers. In this study, we evaluated total and 30DAY cost data for patients who died while enrolled from July 1, 2016 to June 30, 2017 in the Oncology Care Model (OCM), a Medicare (CMS) program initiated by the Center for Medicare and Medicaid Innovation. Methods: CMS provided claims data on 16 OCM-participating US Oncology Network practices. We measured 30DAY OCM episode expenditures (“EXP30”) for n = 5017 deceased patients, as well as patient demographics, clinical outcomes (hospital, ICU, and ER visits) and 30DAY treatments. The population was divided into two cohorts: those with $ 3 days of hospice care (“HC”) and those without (“NOHC”). EXP30 was compared between HC and NOHC at the univariate level using Wilcoxon Rank-sums; categorical variables were compared using Chi-squared tests. Multivariate regression was used to determine the effect of HC on EXP30 adjusted for demographic and disease factors. Results: HC had mean EXP30 reduction of $7,192 vs. NOHC [95%CI -$7,628, -$6,730] adjusted for demographics and disease. HC had lower rates of death in hospital (0.03% vs. 43.3%, p , .0001). Mean expense by days before death: NOHC values 0-30 days = $20,701; 31-60 days = $12,962; 61-90 days = $9,952 and HC values 0-30 days = $10,877; 31- 60 days = $10,376; 61-90 days = $9,064. Conclusions: All categories of care except HC are associated with significantly higher cost in the last 30 days of life. The dramatic and steep trajectory in the last 30 days suggests high expense but also loss of choices as to how best to live the end of one’s life. The results will be updated as new data become available.

Mean 30DAY Expense by Category (unadjusted). HC NOHC (n = 2930) (n = 2087) IV Chemotherapy $ 865 $ 2,128 Part D (Oral) Chemo 254 548 Supportive Care Rx 100 377 Hospitalization 4,255 12,694 ICU 2,466 10,744 SNF 281 801 Hospice 3,366 373 All 30DAY Expense: 10,682 20,530 1) p-value , .0001 for all

11535 Poster Session (Board #227), Mon, 1:15 PM-4:15 PM

Economics of using telehospice to supplement hospice care in rural areas.

Adam Lomenick, Sandy Kuhlman, Tami Gurley-Calvez, Joe Barnes, Hope Krebill, Ashley Spaulding, Gary C. Doolittle; University of Kansas Medical Center, Westwood, KS; Hospice Services and Palliative Care of Northwest Kansas, Inc., Phillipsburg, KS; University of Kansas Medical Center, Kansas City, KS; Midwest Cancer Alliance, University of Kansas Medical Center, Fairway, KS

Background: Hospice care in rural areas is often characterized by provider shortages and vast geographical service areas to cover, making access to quality end-of-life care challenging. Telehospice, the utilization of interactive video (ITV) technology to provide health services over a distance for hospice patients, has been proposed as a solution to address access issues. In 2017, the University of Kansas Medical Center (Kansas City, Kansas) partnered with Hospice Services, Inc. (HSI) (Phillipsburg, Kansas) to augment traditional, in-person hospice care with hospice care delivered via mobile tablets.This work examines the costs of Telehospice when compared with the costs of in-person hospice services. Methods: Detailed administrative data from July 1, 2018 to December 31, 2018 was analyzed to estimate the costs of service after Telehospice use was inculcated into routine practice. Results: Hospice Services, Inc., which averages a daily census of 34 patients, conducted 257 calls, averaging 28 hours a month. The average time for a Telehospice call was 21 and 18 minutes for nursing and medical director calls, respectively. Through various hospice functions, including administrative, patient, and non-patient related connec- tions, HSI saved over $115,000 in staff travel time and mileage reimbursement. Administratively, by hosting their weekly 14-member interdisciplinary meeting via ITV, HSI saved $29,869 of staff travel time and mileage reimbursement. Conclusions: Our estimates indicate substantial cost saving potential with the use of Telehospice services. Further research is needed to assess the effects of Telehospice utilization on the experiences and subsequent cost of hospice care. Additional cost data specifics will be presented in poster form.

In-person In-person Telehospice Telehospice Visits Avg. Cost / Visit Call Avg. Cost / Call TOTAL RN and LPN 2176 26 Social Worker 563 54 Medical Director 28 42 AVERAGE RN and LPN 363 $197 4 $21 Social Worker 94 $205 9 $12 Medical Director 5 $1,572 7 $57* *data from July–September 2018 1. Doolittle GC, Nelson EL, Spaulding A, et al. TeleHospice: A Community-Engaged Model for Utilizing Mobile Tablets to Enhance Rural Hospice Care. Am J Hosp & Pall Med, 2019, in press.

11536 Poster Session (Board #228), Mon, 1:15 PM-4:15 PM

Write to recover: The impact of group led creative writing on behavioral health outcomes in cancer patients.

Darya Nesterova, Junjia Zhu, Courtney Kramer, Monali K. Vasekar, Jolene Collins, Cristina I. Truica, Aditya Joshi, Michael Hayes, Erika Saunders, Joseph J. Drabick, Monika Joshi; Penn State Hershey College of Medicine, Hershey, PA; Penn State Health Milton S. Hershey Medical Center, Hershey, PA; Penn State College of Medicine, Hershey, PA; Penn State Cancer Institute, Hershey, PA; Penn State Hershey Medical Center, Hershey, PA

Background: The diagnosis of cancer can adversely affect mental wellbeing. In addition to treating cancer, the emotional wellbeing of patients must simultaneously be addressed. A previous pilot exploring the feasibility of creative writing workshop (CWW) in cancer patients showed apositive effect on patients’ mental health. Methods: To longitudinally evaluate the efficacy of CWW on mood, we conducted a phase II study with cancer patients (any stage, any cancer type); randomized 2:1 to CWW vs. active control (AC). Patients in the CWW arm attended at least 4, 1.5-hour bi-monthly CWW x 8 wks, whereas AC patients completed independent writing at home with the help of a book (bi-monthly x 8wks). We used validated tools, [Emotional Thermometer Scales (ETS), PHQ-9, GAD-7] to assess changes in overall mood, depression, and anxiety. Primary end point: a) ETS scores before and after intervention b) Changes in depression and anxiety based on PHQ-9 and GAD-7 scores. We present results from ETS scores. Descriptive statistics were generated for these quantitative scales measured in each group, pre and post intervention. Comparisons between groups (gp) were made using Wilcoxon Rank-sum tests. All tests were two sided and the statistical significance level used was 0.05. Results: Amongst evaluable patients, N of 50 (demographics in table below), twenty-six patients in the CWW gp attended at least one class and 19 attended at least 4 classes. Patients in CWW showed significant mood improvement vs. AC when comparing the final overall ETS (p=0.0063). Three of the five sub-scale ETS scores were significantly lower for the CWW vs. AC gp: anxiety (p=0.0027), depression (p=0.0009), and anger (p=0.0027). Conclusions: Group led CWW have a positive effect on mood. Our results suggest potential therapeutic benefit of this intervention on the emotional wellbeing of cancer patients. Larger studies are needed to evaluate the effect of CWW in cancer patients. Clinical trial information: NCT03536702.

Demographics. Variables Categories Overall (N=50) CWW (N=33) AC (N=17) Pval Age (years) Median 60.5 61.4 58.5 0.3056 Gender F 31 (62%) 23 (69.7%) 8 (47.1%) 0.1372 M 19 (38%) 10 (30.3%) 9 (52.9%) Status of Ca New Dx 18 (36.7%) 14 (43.8%) 4 (23.5%) 0.2213 Progression 10 (20.4%) 5 (15.6%) 5 (29.4%) Recurrence 1 (2%) 0 (0%) 1 (5.9%) Stable 20 (40.8%) 13 (40.6%) 7 (41.2%)

11537 Poster Session (Board #229), Mon, 1:15 PM-4:15 PM

Palliative Prognostic Index application at an inpatient palliative care service from a Brazilian tertiary Hospital: A prospective observational study.

Tiago Pugliese Branco, Alze Pereira dos Santos Tavares, Mariana Sarkis Braz, Mariana Ribeiro Monteiro, Ana Beatriz Kinupe Abrahao, Thiago A. F. de Menezes, Carolina Paparelli, Carolina Sassaki Kishimoto, Helena M. F. Medeiros, Karen Ebina, Luciana R. B. Cabral, Maria Clara Fernandes, Silvia Cortizo, Fabiola Borges, Simone Nagashima, Pedro Nazareth Aguiar; Hospital Paulistano, S~ao Paulo, Brazil; Americas Oncologia, S~ao Paulo, Brazil

Background: Palliative Care Index (PPI) has been proposed to improve the accuracy of survival prediction for advanced cancer patients. The aim of this study is to investigate the feasibility and real-world prognosis survival of oncology inpatients from a Brazilian tertiary hospital using PPI. Methods: Hospitalized advanced cancer patients who have been referred to the Palliative Care Team were enrolled from May 2011 to December 2018. The PPI was collected within 24 hours of the referral by the palliative care physician. Primary endpoint was median overall survival (OS), estimated with the use of the Kaplan–Meier method, in three groups: PPI , 4.0; 4.0 # PPI . 6.0 and PPI $ 6.0. Secondary endpoints were OS rate at 3-week for patients with PPI $ 6.0, and the most accurate PPI value to predict 6 and 3-week survival, calculated by ROC curve. Results: Total of 1.381 patients were included in this cohort with a median age of 68-year-old, and 51.3% of females. The most frequent primary cancer sites were lung/chest (17,2%), colorectal (14,3%), breast (11,2%), and biliopancreatic (10,9%). Among 454 patients with PPI , 4.0, median OS was 44 days (95% CI: 35,5-52,4); 20 days (95% CI: 15,4-24,5) for 260 patients with 4.0# PPI , 6.0 and 8 days (95% CI: 7-8,9) between 655 patients with PPI $ 6. Differences in OS among the groups adjusted for primary site, age and gender were significant (p , 0,001). OS rate at 3 weeks for PPI$ 6.0 was 28.1% (OR 5,39 p , 0.001). PPI value of , 5,5 best predicted 6-week OS (79% sensibility, 55% specificity, AUC 0,714) and the PPI value of $ 5,5 predicted 3-week OS (67% sensibility, 73% specificity, AUC 0,753). Conclusions: PPI is feasible and suitable for routine clinical practice to predict survival among Brazilian patients with advanced cancer. In our study, PPI 5.5 seems to be the most accurate value to predict survival within 3 weeks.

11538 Poster Session (Board #230), Mon, 1:15 PM-4:15 PM

Comparing outcomes data from dedicated oncology-subspecialist hospitalists to those from a traditional care model on an inpatient oncology service.

Kristina Fanucci, Andrew Chung Yang, Steven E. Reinert, Rabin Niroula; Lifespan Health System, Providence, RI; Brown University Hematology and Oncology-Rhode Island Hospital, Providence, RI; Lifespan Information Services, Providence, RI; Roger Williams Med Center, Providence, RI

Background: Changes in the complexity and delivery of health care and efforts to contain rising costs gave rise to the hospitalist model of care delivery in the 1990s. More recently some hospitals have begun using the hospitalist model on inpatient oncology floors. Oncology patients are a unique population who require high rates of hospitalization for problems arising from their progressive disease burden and side effects of cancer-directed therapy and have high rates of 30 day readmissions (30DR), an important quality metric. The majority of studies in this area have shown similar outcomes comparing care given by internal medicine hospitalists working full time on an inpatient oncology service compared to multiple oncologists working in rotating shifts on the same inpatient oncology service (Traditional). A new care model using dedicated hematology/oncology subspecialist hospitalists (H-O) to care for patients admitted to the oncology service has been implemented at our academic medical center. Methods: We conducted a retrospective chart review to identify patients with a cancer diagnosis admitted to the oncology service over a six year period. 7/1/2012-6/30/2015 marked the Traditional care model and 7/1/2015-7/1/2018 the H-O model. We compared 30DR, discharge to hospice, length of stay (LOS), and inpatient mortality betweenthesetwogroups.Results: We identified a total of 3778 patients admitted to the oncology service over this six year period—1932 patients admitted to the Traditional service and 1846 patients admitted to the H-O service. There was a significant difference in 30DR between the Traditional v. H-O service (36.7% v. 29.5%, x2= 21.1, p , 0.00001) and discharge to hospice (7.6% v. 12.9%, x2=29.0,p, 0.0000001). There was no significant difference in LOS between the Traditional and H-O services (6.40 days v. 6.03 days, p = 0.122) or in-hospital mortality (2.6% v. 1.9%, x2=2.16,p=0.14). Conclusions: The shift to dedicated hospitalist-oncologists caring for patients on the inpatient oncology service significantly decreased 30DR at our institution. There were also significantly more patients transferred to hospice care. We found no significant difference between the two groups in LOS or inpatient mortality. This care model may help reduce the cost of caring for oncology patients by decreasing 30DR.

11539 Poster Session (Board #231), Mon, 1:15 PM-4:15 PM

Variation in the intensity of end-of-life care, hospice enrollment, and the cost of care by cancer site.

Shitanshu Uppal, Versha Pleasant, Sarah Bell, R. Kevin Reynolds, Laurel Rice, Ryan J. Spencer; University of Michigan, Ann Arbor, MI; Univ of Virginia Health Science Ctr, Charlottesville, VA; University of Wisconsin Hospital and Clinics, Madison, WI

Background: To compare end-of-life (EOL) care intensity across multiple cancer sites and its impact on the cost of care during the last 30-days of life. Methods: Cross-sectional retrospective study using Surveillance, Epidemiology, and End Results Program (SEER)-Medicare linked database from 2008-2013. Utilization of the following in the EOL period (last 30 days of life prior to death) was examined: 2 or more visits to the Emergency Department (ED), hospitalizations, receipt of any life-extending procedures, admission to intensive care unit, any hospice use, hospice used #3 days, death in hospital, or receipt of chemotherapy (last 14 days). We tallied the claims made during the EOL period for each patient. Median expenditures during the EOL period were tabulated by cancer site and stratified by receipt of hospice care. Results: EOL care utilization varied widely between cancer sites. The rates of any hospice utilization were the highest for breast cancer (48.2%) compared to ovarian (17.4%) (p , 0.001). Chemotherapy during the last two weeks of life was the highest for ovarian cancer (1.7%) and the lowest for colorectal cancer (0.5%). The median cost of care during the EOL period for patients who received hospice care was $7,547.84 (Interquartile range [IQR] $5,422-$17,293) compared to $17,179 (IQR $10,323 - $30,824) for those never received hospice care (p , .0001). There was no significant variation in median cost by cancer site for those who received hospice services (range $7,317 - $9,816). However, the cost of care varied for those who did not receive hospice care. Median costs for this subgroup was the lowest for pancreatic cancer ($14,635) and the highest for colorectal cancer ($24,324) (p , 0.001). Conclusions: Despite increasing acceptance of palliative care services and hospice in cancer care, there continues to be intensive and costly care administered during the EOL period. These costs vary considerably across cancer sites when patients do not receive hospice care, but costs are relatively similar when patient are enrolled in hospice Further research to investigate the causes of this variation should be undertaken to design interventions aimed at improving timely hospice enrollment at the EOL and reducing costly care in the EOL period.

11540 Poster Session (Board #232), Mon, 1:15 PM-4:15 PM

Prognostic factors among older adults with advanced non-small cell lung cancer (NSCLC): A multisite cohort study.

Melisa L. Wong, Christine Miaskowski, Alexander K. Smith, W. John Boscardin, Harvey Jay Cohen, Arti Hurria, Vivian Lam, Carling Jade Ursem, Collin Michael Blakely, Matthew A. Gubens, Thierry Marie Jahan, Caroline Elizabeth McCoach, Gregory Maners Allen, Julia Rotow, Niharika Dixit, Vivek Musinipally, Katey R. Webber, Louise Christie Walter; University of California San Francisco, San Francisco, CA; University of California, San Francisco and San Francisco Veterans Affairs Medical Center, San Francisco, CA; Duke University Medical Center, Durham, NC; City of Hope National Medical Center, Duarte, CA; University of California, San Francisco and Zuckerberg San Francisco General, San Francisco, CA; University of California, San Francisco and Berkeley, San Francisco, CA

Background: Prognosis in NSCLC is vital for clinical decision making. With the aging US population and rapidly changing treatment landscape, we aimed to identify prognostic factors among older adults with advanced NSCLC receiving chemo-, immuno-, and/or targeted therapy. Methods: We conducted a prospective cohort study of adults age $65 with advanced NSCLC starting a new non-curative systemic treatment (chemo-, immuno-, and/or targeted therapy) at a Comprehensive Cancer Center, Veterans Affairs Medical Center, and safety-net hospital. Prior to treatment initiation, patients completed a geriatric assessment including cognition, function, comorbidities, mood, social support, and quality of life. Cox proportional hazards models were performed to identify prognostic factors for overall survival (OS). Results: In a sample of 51 patients, median age was 73 (range 65-94). The majority of patients had stage IVB (59%) or IVA (39%) NSCLC. Current treatment included immunotherapy (37%), targeted therapy (29%), chemoimmunotherapy (18%), and chemo (16%). Most patients had received prior NSCLC treatment (80%): chemo (51%), targeted therapy (35%), immunotherapy (22%), radiation (RT; 47%), and/or surgery (19%). At enrollment, 73% had an abnormal Montreal Cognitive Assessment score , 26 (MoCA; median score 23) and 35% had an abnormal Timed Up and Go time $13.5 secs (TUG; median time 12.7 secs). Median OS was 12.5 months. In univariable analyses, stage IVB disease (HR 6.99, 95% CI 1.55-31.5), prior RT (HR 2.98, 95% CI 1.08-8.21), worse MoCA score (HR 1.15 per 1 point change, 95% CI 1.03-1.28), and longer TUG time (HR 1.13 per 1 sec change, 95% CI 1.05- 1.23) were associated with worse OS. Of note, age, current NSCLC treatment, line of therapy, and Karnofsky Performance Status were not associated with OS. In multivariable analysis, MoCA score was the only statistically significant prognostic factor (HR 1.15, 95% CI 1.01-1.30). Conclusions: We found that abnormal pretreatment cognition is very common and an important prognostic factor among older adults with advanced NSCLC. Pretreatment screening for cognitive impairment should be considered to inform prognostication, decision making, and treatment planning.

11541 Poster Session (Board #233), Mon, 1:15 PM-4:15 PM

ELAN-ONCOVAL (Elderly Head and Neck Cancer-Oncology Evaluation) study: Evaluation of the G8 screening tool and the ELAN geriatric evaluation (EGE) for elderly patients (pts) with head and neck squamous cell carcinomas (HNSCC).

Cécile Mertens, Herve ´ Le Caer, Cecile ´ Ortholan, Emmanuel Blot, Caroline Even, Hubert Rousselot, Frederic Peyrade, Christian Sire, Didier Cupissol, Yoann Pointreau, Philippe Debourdeau, Frederic Rolland, Jerome Fayette, Olivier Capitain, Xu Shan Sun, Moufida Debbah, Rabia Boulahssass, Dominique Schwob, Anne Auperin, Joel Guigay; Institut Bergonie, Bordeaux, France; Centre Hospitalier de Saint Brieuc, Saint- Brieuc, France; Princess Grace Hospital, Monaco, Monaco; Centre d’Oncologie Saint-Yves, Vannes, France; Gustave Roussy, Villejuif, France; Institut de Cancerologie ´ de Lorraine-Alexis Vautrin, Vandoeuvre- Les-Nancy, France; Centre Antoine-Lacassagne, Nice, France; Groupe Hospitalier Bretagne Sud-Radio- therapie-Oncologie, ´ Lorient, France; Institut du Cancer de Montpellier, Montpellier, France; Centre Jean Bernard-Clinique Victor Hugo, Le Mans, France; Institut Sainte-Catherine, Avignon, France; Institut de Cancerologie ´ de l’Ouest, Department of Medical Oncology, St Herblain, France; Centre Leon ´ Bérard, Medical Oncology, Lyon, France; Institut de Cancerologie de l’Ouest, Site Paul Papin, Angers, France; Centre Hospitalier de Belfort-Montbeliard, ´ Montbéliard, France; Centre Hospitalier Universitaire de Nice, Hopitalˆ de Cimiez, Nice, France; Centre Antoine Lacassagne, FHU OncoAge, UniversiteC ´ oteˆ d’Azur, Nice, France

Background: About 47% and 10% of HNSCC occur in pts aged $65 or $80 respectively. This population being heterogeneous, balancing efficacy with toxicity is challenging. As the G8 screening tool includes 3 nutritional items, we need to evaluate its usefulness in HNSCC pts and to establish an adapted tool for HNSCC pts. Methods: ONCOVAL is the first step of the ELAN program for pts $70 with HNSCC not amenable to surgery. Pts were assessed for geriatric frailties before inclusion in appropriate trials (ELAN- RT, FIT, UNFIT) depending on curative or palliative setting and fit or unfit status. The aim was to evaluate the G8 tool and EGE usable by oncologists to classify pts as fit or unfit. The EGE was elaborated by the French GERICO group for HNSCC pts, evaluating functional status, comorbidity, cognition/mood, social status. To estimate sensitivity and specificity of G8 and EGE, we used comprehensive geriatric assessment (CGA) performed by geriatricians as gold standard. Results: From 06/2013 to 08/2018, 633 pts were included in 44 centers. Data are available for 613 pts. Median age 79 years (70-97) with 45% $ 80. 75% males. G8 score was #14, i.e. at risk of geriatric frailty, in 82% pts. Assessment by EGE was done by physicians (64%), nurses (11%), ELAN staff (21%) in 20 minutes of median time. After EGE, oncologists classified 71% pts as unfit. 49% of pts were assessed by CGA. The G8 score with cutoff # 14 has sensitivity 91% and specificity 30%. Lowering cut-off to 12 increased specificity to 56% but decreased sensitivity to 75%. Sensitivity of EGE was 95% and specificity 60%. Multidisciplinary interventions were more frequently implemented in pts assessed by geriatricians than in others (71% vs 47%), even after adjusting for frailty. 58% of pts included in ONCOVAL were further included in ELAN therapeutic trials. Assessment by geriatricians did not modify the rate of inclusion in trials. Conclusions: The G8 screening tool is not appropriate for HNSCC pts. The EGE was feasible and had better sensitivity and specificity. Oncologists and geriatricians must continue such collaboration to propose tailored treatments. Clinical trial information: NCT03614936.

11542 Poster Session (Board #234), Mon, 1:15 PM-4:15 PM

Fitness scale-4 item (FS4) to predict six-month postoperative mortality of cancer patients aged 75.

Armin Shahrokni, Saman Sarraf, Robert J. Downey, Koshy Alexander, Beatriz Korc-Grodzicki; Memorial Sloan Kettering Cancer Center, New York, NY

Background: We have been prospectively collecting Geriatric Assessment (GA) data on patients aged 75+ undergoing oncologic surgery at a single institution. We aimed to to determine if a fitness scale based on a limited number of easily captured datafields can be developed, validated, and associated with 6 month postoperative mortality. Methods: From patients referred to Memorial Sloan Kettering (MSK) Geriatrics Service for preoperative GA since 2015, we have selected patients who were aged 75+, had elective surgery, with hospital length of stay (LOS) of $1 day, and had at least 6-month follow up. The Fitness Scale-4 item (FS4) is composed of 3 variables from the GA; Karnofsky Performance Scale (KPS), ability to walk outside, ability to perform housekeeping, and additional variable of preoperative albumin level. The FS4 total score ranges from 0-8 (frail = 0-2, vulnerable = 3-5, and fit = 6+). We validated this scale against prevalence of geriatric syndromes in each fitness category. We performed multivariable regression (MVR) analysis to assess the odds of six-month postoperative mortality among vulnerable and frail patients vs. fit patients. Results: Out of 1270 patients (median age 79, IQR 77-83), 792 (61.4%), 327 (25.7%) and 151 (11.9%) were fit, vulnerable, and frail, respectively. All of the geriatric syndromes were more prevalent among frail and vulnerable patients than fit patients. Six-month mortality rate was 4.7%, 12.5%, and 21.2% for fit, vulnerable, and frail patients, respectively. In the MVR analysis, and after adjustment for age, gender, ASA PS, LOS, operation time, KPS, and preoperative albumin, the odds of six-month mortality among frail (OR = 4.58, p = 0.004) and vulnerable patients (OR = 2.63, p = 0.004) vs. fit patients were much higher. Conclusions: Frail and vulnerable patients based on FS4 are at significantly higher risk for six-month postoperative mortality vs. fit patients. Future studies should assess the external validation of FS4, and its impact on surgical decision making, and perioperative care of older cancer patients.

11543 Poster Session (Board #235), Mon, 1:15 PM-4:15 PM

Does distress screening lead to mental health care in older adult surgical patients?

Kelly Marie Trevino, Christian J. Nelson, Rebecca Saracino, Beatriz Korc-Grodzicki, Saman Sarraf, Armin Shahrokni; Memorial Sloan Kettering Cancer Center, New York, NY

Background: Surgery is a notable stressor for older adults with cancer who are often medically complex and frail. The American College of Surgeons Commission on Cancer requires distress screening in accredited cancer care settings. The degree to which distress screening leads to mental health use is unclear. This study examined rates and predictors of post-surgical mental health care in older adults referred for a preoperative evaluation. Methods: Patients aged 75 years or older (n = 1,008) referred to the Geriatrics Service at a comprehensive cancer center were enrolled. Patients underwent elective surgery with a length of stay of 3 days or longer and were followed for at least 30 days after surgery. A comprehensive geriatric assessment (CGA) was administered as part of routine care. Surgical characteristics and post-surgical encounters with social work, psychology, and psychiatry were abstracted from the electronic medical record. Bivariate relationships between demographic and surgical characteristics and the CGA and postoperative receipt of mental health services were examined. Characteristics with significant (p , .01) bivariate relationships were entered into a multivariable regression predicting post-operative mental health service use. Results: One-quarter of the total sample (n = 246, 24.4%) received post-operative mental health services. In multivariable analyses, high distress (Distress Thermometer score$4; p = .01), poor social support (p = .01), iADL dependence (p = .04), and longer length of stay (p , .001) were associated with receipt of mental health services after controlling for significant sociodemographic and surgical characteristics and CGA variables. Of patients with high distress, poor social support, or iADL dependence, only one-third (29-33%) received mental health care. Conclusions: Distressed older adults and those with low levels of support pre-operatively were more likely to receive mental health services after surgery, controlling for sociodemographic and surgical characteristics. Yet, only one-third of these patients received mental health care. These findings suggest that barriers to translating distress screening into provision of mental health services remain.

11544 Poster Session (Board #236), Mon, 1:15 PM-4:15 PM

Associations of uncertainty with psychological status and quality of life (QoL) among 527 older patients with advanced cancer.

Dilip Sankar Babu, Kah Poh Loh, Javier Bautista, Huiwen Xu, Eva Culakova, Beverly E. Canin, Alison Katherine Conlin, James Bearden, Jeffrey L. Berenberg, Yingzi Zhang, Megan Wells, Ronald M. Epstein, William Dale, Paul Duberstein, Supriya Gupta Mohile, Mohamedtaki Abdulaziz Tejani; University of Rochester Medical Center, Rochester, NY; University of Rochester, Rhinebeck, NY; Earle A. Chiles Research Institute at the Robert W. Franz Cancer Center, Portland, OR; Southeast Clinical Oncology Research Consortium (SCOR), Winston-Salem, NC; Hawaii MUNCORP, Honolulu, HI; University of Rochester, Roches- ter, NY; City of Hope National Medical Center, Duarte, CA; Rutgers University, New Brunswick, NJ

Background: Older patients with advanced cancer face considerable uncertainty related to their disease and treatment. The aim of our study was to evaluate the associations of uncertainty with psychological status and QoL. Methods: This is a secondary analysis of baseline data from a national geriatric assessment (GA) cluster randomized trial (URCC 13070; PI: Mohile). Patients aged $70 years with $1GAdomain impairment (e.g., function, cognition) and advanced cancer who were considering or receiving any line of cancer treatment were enrolled (n=541). Uncertainty was measured using the modified 9-item Mishel Uncertainty in Illness (MUIS), where respondents with higher scores perceive more uncertainty (range 9- 45). QoL and psychological measures consisted of Functional Assessment of Cancer Therapy-General (FACT-G), emotional wellbeing (EWB; FACT-G subscale), distress (distress thermometer), anxiety (Gen- eralized Anxiety Disorder-7), and depression (Geriatric Depression Scale-15). Multiple linear regressions were used to evaluate the associations of MUIS scores with each measure, adjusted for demographics, cancer type, and number of impaired GA domains. Results: Mean age was 77 years (SD 5, range 70-96); 26% had gastrointestinal cancer and 26% had lung cancer. Mean number of GA domain impairments was 4 (SD 1, range 1-7). Mean MUIS score was 20 (SD 5, range 9-37). On multivariate analyses, higher MUIS score was associated with lower QoL (b=-1.08, SE=0.11) and EWB (b=-0.29, SE=0.03), as well as higher distress (b=0.12, SE=0.02), anxiety (b=0.11, SE=0.04), and depression (b=0.09, SE=0.03; all P,0.01). Conclusions: Distress associated with uncertainty was common in a vulnerable population of frail older patients with advanced cancer and $1 GA domain impairment. A higher degree of uncertainty was associated with poorer psychological health and QoL. Our results underscore the important role that uncertainty plays in older patients’ psychological status. Previous tested uncertainty management interventions (mainly including information and coping strategies) could be revised, tailored and tested to meet the unique needs of older patients with cancer. Clinical trial information: NCT02107443.

11545 Poster Session (Board #237), Mon, 1:15 PM-4:15 PM

Predictive value of an electronic frailty index (FI) in U.S. Veterans with newly diagnosed non-small cell lung cancer (NSCLC).

Ayesha Rizwan Sheikh, David Cheng, Nathanael Fillmore, Nhan Do, Mary T. Brophy, David P. Tuck; Boston Medical Center/ Boston University School of Medicine, Boston, MA; VA Boston Healthcare System, Boston, MA; VA Boston Healthcare System and Dana-Farber Cancer Institute, Boston, MA; VA Boston Healthcare System and Boston University School of Medicine, Boston, MA; VA Boston Healthcare System and Boston University Medical Center, Boston, MA

Background: Frailty is a well-established predictor of survival, but assessing it in a real-world healthcare setting is often time-consuming and impractical. We sought to determine if a quantitative measure of frailty, computed automatically from electronic health records using the cumulative deficit model, also predicts overall survival in patients with NSCLC. Methods: We conducted a retrospective cohort study of Veterans identified to have NSCLC in the VA Precision Oncology Data Repository. Frailty was assessed using a previously validated FI, the Frailty Index for Veterans Affairs (VA-FI), consisting of 31 deficits. We categorized non-frail patients into non-frail (0-0.1) and pre-frail (0.1-0.2) groups, and frail patients into mild (0.2-0.3), moderate (0.3-0.4) and severe ( . 0.4) groups. Associations of frailty and survival were assessed using Kaplan-Meier estimates and Cox models. The added value of frailty status for predictive performance was quantified using the continuous net-reclassification index (NRI). Results: The cohort included 359 patients with NSCLC. The mean age was 66 years, and 96.1% were male, including 76.3% Caucasian and 16.2% African-American. The majority had an advanced stage cancer (15.9% and 44.3% with stages III and IV respectively) with minority falling into the early stage category (15.3% and 6.1% with stages I and II respectively). A substantial proportion of patients were identified to be frail (30%). The frailty categories differentiated survival profiles of patients with 2-year survival ranging from 0.41 for non-frail patients to 0.08 for the severely frail patients (log-rank p , 0.001). Frail patients exhibited significantly shorter survival than non-frail group even after adjusting for age, gender, and stage in a multivariate Cox model (hazard ratio 1.7, p , 0.001). Inclusion of frailty status in the Cox model significantly improved 2-year risk estimates of survival (NRI 0.35; 95% CI 0.27–0.47). Conclusions: Frailty as determined by VA-FI is a significant predictor of survival independent of stage and demographic factors among NSCLC patients in this VA cohort. VA-FI is an automated and a practically feasible tool to better estimate life expectancy and help individualize patient care.

11546 Poster Session (Board #238), Mon, 1:15 PM-4:15 PM

Older adults’ preferred and perceived roles in decision making about palliative chemotherapy: Their decision priorities, and information preferences.

Erin Moth, Belinda Emma Kiely, Andrew James Martin, Vasi Naganathan, Stephen A. Della-Fiorentina, Florian Honeyball, Rob Zielinski, Christopher B. Steer, Hiren A. Mandaliya, Abi Ragunathan, Prunella Blinman; Concord Cancer Centre, Concord Repatriation General Hospital, Concord, NSW, Australia; NHMRC Clinical Trials Centre, University of Sydney, Sydney, Australia; Centre for Education and Research on Ageing, Concord Clinical School, University of Sydney, Sydney, NSW, Australia; Southern Highlands Cancer Centre, Bowral, NSW, Australia; Dubbo Health Service, Dubbo, NSW, Australia; Central West Cancer Centre, Orange, NSW, Australia; Border Medical Oncology, Albury Wodonga Regional Cancer Centre, Albury, NSW, Australia; Macarthur Cancer Therapy Centre, Camp- belltown, NSW, Australia

Background: We sought older adults’ preferred and perceived roles in decision-making about palliative chemotherapy; their decision-making priorities; and information received and desired. Methods: Patients aged $65 years with incurable cancer who had discussed palliative chemotherapy with an oncologist and made a decision about whether or not to receive palliative chemotherapy were invited to complete a written questionnaire. Preferred and perceived decision-making roles were assessed by the Control Preferences Scale (CPS). Associations with preferred decision-making role were examined using Wilcoxon rank sum tests. Factors important in making a decision about chemotherapy, and receipt of and desire for information were described. Results: The 179 respondents had a median age of 74 years (range 65 to 92 years). Most were male (114, 64%) and had chosen to receive chemotherapy (148, 83%). Half (92, 52%) were vulnerable by the Vulnerable Elders Survey-13 (score $3). Preferred decision-making roles (n = 173) were active in 39%, collaborative in 27%, and passive in 35%. Perceived decision-making roles (n = 172) were active in 42%, collaborative in 22%, and passive in 36%, and matched the preferred role for 63% of patients. Preference for an active role was associated with being single/widowed (p = 0.004, OR 1.49) and declining chemotherapy (p = 0.02, OR 2). Factors ranked most important when making a decision about chemotherapy (n = 159) were “doing everything possible” (30%), “my doctor’s recommendation” (26%), “my quality of life” (20%), and “living longer” (15%). A minority expected chemotherapy to cure their cancer (14%). Most had discussed expectations of cure (70%), side effects (88%) and benefits (82%) of chemotherapy, though fewer had received quantitative prognostic information (49%) than desired this (67%). Conclusions: Older adults showed varied preferences for involvement in decision-making about palliative chemotherapy, and most played the role that they preferred. To facilitate shared decision-making, oncologists should seek patients’ decision- making preferences, priorities and information needs when discussing palliative chemotherapy.

11547 Poster Session (Board #239), Mon, 1:15 PM-4:15 PM

Cancer-specific mortality and competing causes of death in older adults: A prospective, multicenter cohort study (ELCAPA-19).

Deborah Assouan, Elena Paillaud, Philippe Caillet, Emmanuelle Kempf, Helene Vincent, Etienne Brain, Godelieve Rochette de Lempdes, Cyril Touboul, Claudia Martinez-Tapia, Mylene Allain, Sylvie Bastuji- Garin, Olivier Hanon, Marie Laurent, Florence Canoui-Poitrine; EA 7376 CEpiA (Clinical Epidemiology and Ageing Unit), Creteil, France; Geriatric Department, Georges Pompidou European Hospital, Paris, France; Gustave-Roussy Institute, Oncology Department, Villejuif, France; Curie Institute, France, Saint Cloud, France; Curie Institute, Saint Cloud, France; Gynecological Department, Henri-Mondor Hospital, Creteil, France; EA 7376 CEpiA (Clinical Epidemiology and Ageing Unit), Creteil, ´ France; Clinical Research Unit, Henri-Mondor Hospital, Creteil, ´ France; Geriatric Department, Broca Hospital, APHP, Paris, France

Background: Among older adults with cancer, comorbidities compete with cancer as the cause of death. The objectives were to quantify the proportion and rate of cancer-specific death in older patients with cancer, and to analyze the associations between geriatric factors and cancer death. Methods: Between January 2007 and December 2014, older patients with cancer were prospectively included by the ELCAPA cohort study’s eight investigating centers. Competing risk methods were used to estimate 6- month and 3-year cancer mortality rates and to probe associations between geriatric factors and cancer death. Results: A total of 1678 patients were included (mean 6 standard deviation age: 81.3 6 5.8; women: 49%). The most common cancers were colorectal (19%), breast (17%) and urinary (15%) cancers and 49% had metastasis. After a median follow-up period of 34 months, a total of 948 deaths were observed. Of the 282 deaths in non-metastatic patients, 203 (72%; 95% confidence interval (CI): [66%-77%]) were attributable to cancer. This proportion was 92% (89–94; N = 448/498) for metastatic patients. The 6-month and 3-year cancer mortality rates was respectively 12% (9–15) and 34% (29-38) for non-metastatic tumors and 45% (41–49) and 83% (80–87) for metastatic stage tumors. At 6 months, the geriatric factors independently associated with cancer death were a dependency in activities of daily living (ADL) score # 5 (adjusted subhazard ratio: 2.11 (95%CI: [1.68–2.64]), mobility impairment (Timed Get Up and Go (TGUG) test time . 20 s (1.40 [1.05–1.87]) or inability to perform the TGUG (2.41 [1.67–3.48])) and comorbidities (total Cumulative Index Rating Scale-Geriatric score $13) (1.59 [1.23–2.06]). At 3 years, the independently associated factors were ADL # 5 (1.60, [1.34–1.91]), TGUG . 20 s (1.28, [1.04–1.59]) or inability to perform TGUG (2.02 [1.47–2.79]), and cognitive impairment (1.23 [1.01–1.50]). Conclusions: Most older adults with cancer die from this disease and not from other comorbidities. However, geriatric parameters (dependency, impaired mobility, comorbidities, and cognitive impairment) are independently associated with cancer death. These geriatric impairments should be taken into account when assessing the cancer patient’s prognosis in clinical practice. Clinical trial information: NCT02884375.

11548 Poster Session (Board #240), Mon, 1:15 PM-4:15 PM

Guided geriatric interventions (GI) in older adults with cancer: What, how, and for whom? The French PACA EST Cohort Experience.

Rabia Boulahssass, Sebastien Gonfrier, Jean-Marc Ferrero, Noemie Champigny, Cyrielle Rambaud, Remy Largillier, Damien Chauviere, Emmanuel Barranger, Jean Marc Bereder, Emmanuel Benizri, Delphine Borchiellini, Matthieu Durand, Patrick Baque, Jerome Delotte, Jerome Barriere, Jean-Michel Hannoun-Levi, Joel Guigay, Thierry Piche, Eric FRANCOIS, Olivier Guerin; Centre Hospitalier Universitaire de Nice, Hopitalˆ de Cimiez, Nice, France; UCOG PACA EST CHU DE NICE, Nice, France; Department of Medical Oncology, Centre Antoine Lacassagne, Nice, France; CHU de Nice, Nice, France; GINECO-Centre Azureen ´ de Cancerologie, ´ Mougins, France; Centre Antoine Lacassagne, Nice, France; University Hospital Archet, Nice, France; Hopital Archet, Nice, France; Centre Antoine Lacassagne, UniversiteC ´ oteˆ d’Azu, Nice, France; CHU Nice, Nice, France; Clinique Saint Jean, Cagnes Sur Mer, France; Radiation Oncology, Antoine Lacassagne Cancer Center, University of Cote d’Azur, Nice, France; Department of Medical Oncology, Antoine Lacassagne Comprehensive Cancer Centre, FHU OncoAge, UniversiteC ´ oteˆ d’Azur, Nice, France; Department of Medical Oncology, Centre Antoine-Lacassagne, Nice, France; Department of Geriatric Medicine, Cimiez Hospital, FHU Oncoage, University Coteˆ d’Azur, Nice, France

Background: Some previous studies in geriatric oncology have described the GI and their adherence. Today’s challenge is to screen patients needing specific GI and repeated Comprehensive Geriatric Assessments (CGA). We recently analyzed a phenotype of patients requiring more GI (Boulahssass et al, Cancers 2019). The main purpose of the present study is to compare types of GI implemented, according to patient frailty levels, in order to better understand the necessary care plan. Methods: Between April 2012 and May 2018, 3530 consecutive patients with solid tumors were enrolled in this multicentric, prospective cohort. 3140 patients (mean age:82y) were finally included and a CGA was performed at Baseline. Twelve GI were standardized, individualized or based on experience if no guidelines were available. Within 1 month, geriatricians including patients in the cohort received standardized training. Logistic regression was performed to compare types of GI in the 3 groups using the Balducci Score (B1/B2/B3). Results: 8819 GI were implemented for the 3140 patients. On average, fit patients had 1.5 GI (n = 146), vulnerable patients 2.4 GI (n = 1568) and frail patients 3.3 GI (n = 1426). We observed no significant differences between the 3 groups concerning specific pain management (Fit vs B2: p = 0.19; Fit vs B3: p = 0.57) and psychological care (Fit vs B2:p = 0.03; Fit vs B3:p = 0.24). In vulnerable and frail patients, we recorded more significant GI for nutritional care, delirium prevention, comorbidity management, nursing interventions, social worker interventions, treatment modifications for optimization or iatrogenic disorders and physiotherapy, with the highest Odds Ratio for nursing interventions (Fit vs B2 OR:2.9 p = 0.011; Fit vs B3 OR:9 P , 0.001) and physiotherapy (Fit vs B2:OR 4.3 p , 0.001; Fit vs B3:OR:9 p , 0.001). B3 patients had significantly more GI on care pathways modifications (OR:3.1, p = 0.002) and caregiver care (OR:2, p = 0.049). Conclusions: Fit patients also needed GI. We observed differences in types of GI between the groups. However, the aims and levels also seemed to differ and need further studies to analyze their impact.

11549 Poster Session (Board #241), Mon, 1:15 PM-4:15 PM

Pilot randomized trial of a transdisciplinary geriatric intervention for older adults with cancer.

Paul Kay, Areej El-Jawahri, Charn-Xin Fuh, Brandon Temel, Sophia Landay, Daniel Lage, Esteban Franco-Garcia, Erin Scott, Erin Stevens, Terrence O’Malley, Supriya Gupta Mohile, William Dale, Lara Traeger, Vicki Jackson, Joseph Greer, Jennifer S. Temel, Ryan David Nipp; Massachusetts General Hospital Cancer Center, Boston, MA; Massachusetts General Hospital, Boston, MA; University of Rochester Medical Center, Rochester, NY; CIY, Chicago, IL

Background: Oncologists often struggle with managing the unique care needs of older adults with cancer. We sought to determine the feasibility of delivering a transdisciplinary geriatric intervention designed to address the geriatric (physical function & comorbidity) and palliative care (symptoms & prognostic understanding) needs of older adults with cancer. Methods: We randomly assigned patients age $65 with newly diagnosed incurable gastrointestinal (GI) or lung cancer to receive a transdisciplinary geriatric intervention or usual care. Intervention patients received two visits with a geriatrician who was trained to address patients’ palliative care needs in addition to conducting a geriatric assessment. We defined the intervention as feasible if . 70% of patients enrolled in the study and . 75% completed study visits and surveys. At baseline and week 12, we assessed patients’ quality of life (QOL, Functional Assessment of Cancer Therapy General), symptoms (Edmonton Symptom Assessment System), and communication confidence (Perceived Efficacy in Patient Physician Interactions). As this was a pilot study, we calculated mean change scores in outcomes and estimated intervention effect sizes (ES). Results: From 2/2017-6/ 2018, we randomized 62 patients (55.9% enrollment rate [most common reason for refusal was feeling too ill]; median age = 72.3 [range 65.2-91.8]; 45.2% female; cancer types: 56.5% GI, 43.5% lung). Among intervention patients, 82.1% attended the first visit and 76.2% attended both. Overall, 77.8% completed all study surveys. Compared to usual care, intervention patients had less decrement in QOL scores (-0.77 vs -3.84, ES = .21), greater reduction in the number of moderate/severe symptoms (-0.69 vs +1.04, ES = .58), and more improvement in communication confidence (+1.06 vs -0.80, ES = .38). Conclusions: In this trial of older adults with advanced cancer, more than half enrolled in the study and over 75% of those who enrolled completed all study visits and surveys. Our effect size estimates suggest that a transdisciplinary intervention targeting patients’ geriatric and palliative care needs may be a promising approach to improve patients’ QOL, symptom burden, and communication confidence. Clinical trial information: NCT02868112.

11550 Poster Session (Board #242), Mon, 1:15 PM-4:15 PM

Predicting severe toxicity of targeted therapies in elderly patients with cancer (PreToxE): A multicenter, prospective, and retrospective study.

Coriolan Lebreton, Coralie Cantarel, Emilie Toulza, Romain Desgrippes, Laurence Bozec Lemoal, Esma Saada, Agnes Ducoulombier, Magalie Pascale Tardy, Elena Paillaud, Caroline Lalet, Simone Mathoulin-Pelissier, ´ Antoine Italiano; Institut Bergonie, Bordeaux, France; CH Saint Malo, Saint Malo, France; Institut Curie-Hopitalˆ Rene ´ Huguenin, Saint-Cloud, France; Centre Antoine Lacassagne, Nice, France; Antoine Lacassagne Center, Nice, France; Geriatric Department, Georges Pompidou European Hospital, Paris, France; Institut Bergonie, ´ Bordeaux, France; INSERM CIC 14.01 Bordeaux, Clinical Epidemiology Unit, Bordeaux, France

Background: It is crucial that targeted therapies are also studied in senior patients to establish predictive factors of severe toxicity. Methods: The PRETOXE study includes 3 multicentric independent cohorts of patients $70 years old with advanced solid tumor (2 retrospectives and one prospective) and treated with a tyrosine/threonine kinase inhibitor (TKI) as per drug label. Data on clinical and biological characteristics of the patient, disease and treatment were centrally collected at the beginning of the treatment. Primary endpoint is severe toxicity defined as treatment-related death, persistant or significant disability/ incapacity, hospitalization or discontinuation of treatment for more than three weeks. Predictive factors of severe toxicity were first identified in a training retrospective cohort by multivariate analysis. Two independent cohorts ( retrospective and prospective) will be used for external validation. Results: 371 patients entered the study (training retrospective cohort n = 171, 46.1 %; validation retrospective cohort: n = 160, 43.1%, validation prospective cohort: n = 40, 10.8%). Median age was 74.0 (range 70.0-88.0) in the training retrospective cohort. 73 patients (42.7%) were male. The most frequent solid tumors were lung 64 (37.4%), breast 50 (29.2%), sarcomas 27 (15.8%), colon 10 (5.8%) and kidney 8 (4.7%). The five most frequent prescribed TKIs were everolimus 51 (29.8%), erlotinib 43 (25.1%), pazopanib 18 (10.5%), gefitinib 17 (9.9%) and regorafenib 14 (8.2%). The prescribed dose was lower than that recommended in the drug label in 32.1% of cases. 46 (26.9%) patients had a severe toxicity as per protocol definition in retrospective cohort: hospitalization 5 (10.9%), discontinuation of treatment more than 3 weeks 4 (8.7%), definitive discontinuation of treatment 38 (82.6%), persistent or significant disability/incapacity 14 (30.4%). On multivariate analysis, female gender, $3 concomitant medications and anti-angiogenic activity of TKI were independent predictive factors of severe toxicity. External validation on the other two independent cohorts (retrospective and prospective) will be presented at the meeting. Conclusions: One in four cancer patient $ 70 years old and treated with a TKI has severe toxicity impacting treatment outcome. The role of geriatric intervention to prevent such toxicities should be considered particularly in female patients, patients with $3 concomitant medications or when the TKI targets the VEGF receptors family. Clinical trial information: NCT02751827.

11551 Poster Session (Board #243), Mon, 1:15 PM-4:15 PM

Prognostic value of routine biomarkers in older patients with cancer: Pooled analysis of three prospective cohorts.

Elena Paillaud, Pierre Soubeyran, Nadia Oubaya, Etienne Brain, Marianne Fonck, Nicoleta Reinald, Damien Heitz, Mylene Allain, Sonia Zebachi, Hubert Rousselot, Marie Laurent, Jer ´ omeˆ Dauba, Philippe Caillet, Sylvie Bastuji-Garin, Muriel Rainfray, Michael Bringuier, Thomas Grellety, Carine A. Bellera, Simone Mathoulin-Pelissier, ´ Florence Canoui-Poitrine; Hopital Europeen Geoges Pompidou, Paris, France; Institut Bergonier, Bordeaux, France; Hopital Henri Mondor, Creteil, ´ France; Institut Curie, Paris & Saint-Cloud, France; Institut Bergonie, ´ CLCC Bordeaux, Bordeaux, France; Henri Mondor’s Hospital, Creteil, ´ France; CHU Strasbourg, Strasbourg, France; Clinical Research Unit, Henri-Mondor Hospital, Creteil, ´ France; Universite ´ Paris Est Creteil, Clinical Epidemiology And Ageing, Creteil, ´ France; Institut de Cancerologie ´ de Lorraine-Alexis Vautrin, Vandoeuvre-Les-Nancy, France; EA 7376 CEpiA (Clinical Epidemiology and Ageing Unit), Creteil, ´ France; CH Mont de Marsan, Mont- De-Marsan, France; Hopital Europeen Georges Pompidou, Paris, France; Department of Clinical Gerontology, Bordeaux University Hospital Xavier Arnozan, Pessac, France; Institut Curie Saint- Cloud, Saint-Cloud, France; Institut Bergonie, ´ Bordeaux, France; Institut Bergonie, ´ Comprehensive Cancer Centre, Bordeaux, France; INSERM CIC 14.01 Bordeaux, Clinical Epidemiology Unit, Bor- deaux, France; Assistance Publique–Hopitauxˆ de Paris, Hopital Mondor-Chenevier, Universite Paris Est, Creteil, ´ France

Background: To assess prognostic value of routine biomarkers in older patients with cancer. Methods: A pooled analysis of three prospective multicentre cohorts, ELCAPA, PHRC Aquitaine and ONCODAGE was conducted. Patients aged 70 years or older, with cancer were included. Biomarkers collected were plasmatic C-reactive protein, albumin and a combined score: Glasgow Prognostic Score (GPS). The GPS comprised three categories (0: CRP#10 mg/L, albumin$35 g/L; 1: CRP#10 mg/L and albumin , 35 g/L, or CRP . 10 mg/L and albumin$35 g/L; 2: CRP . 10 mg/L and albumin , 35 g/L).The primary endpoint was overall survival at 12 months. Multivariable Cox models were used, adjusting for age, sex, localisation, metastatic status, performance status, frailty screening index, the G8. Discriminative properties were assessed using Harrell C index and NRI (Net Reclassification Improvement). Results: Overall 1800 patients were analyzed (ELCAPA: N = 543, PHRC Aquitaine: N = 253, ONCODAGE: N = 1004; mean age: 78.565.5 years; 61.7% of men; 37% metastatic; most frequent localisations: breast (34.9%) and colon- rectum (17.7%); 70.7% of patients screened at risk of frailty with G8). Overall survival was 71.1%. GPS was independently associated with death (among normal G8: GPS 1: Hazard Ratio (HR) = 4.48; 95% Confidence Interval (95% CI) = [2.03; 9.89], GPS 2: 11.64 [4.54; 29.81], among abnormal G8: GPS 1: 2.45 [1.79; 3.34], GPS 2: 3.97 [2.93; 5.37]. The addition of GPS to the clinical model (Harell C: 0.82 [0.80; 0.83]) improved discrimination (Harell C: 0.84 [0.82; 0.85], NRI: 11% [5; 19]). Conclusions: GPS could be used in older patients with cancer to help decision-making and prognosis assessment.

11552 Poster Session (Board #244), Mon, 1:15 PM-4:15 PM

Role of the oncological-multidimensional prognostic index in older patients with metastatic colorectal cancer treated in a real-world setting.

Letizia Procaccio, Antonella Brunello, Pasquale Fiduccia, Annunziata Lettiero, Giuseppina Tierno, Grazia Pusole, Francesca Bergamo, Marta Schirripa, Chiara Manai, Vittorina Zagonel, Sara Lonardi; Department of Clinical and Experimental Oncology, Medical Oncology 1 Unit, Istituto Oncologico Veneto IOV-IRCCS, Padua, Italy; Clinical Trials and Biostatistics Unit, Istituto Oncologico Veneto IOV - IRCCS, Padua, Italy

Background: About 50% of diagnoses of colorectal cancer (CRC) occur in patients (pts) older than 70 years. Though a comprehensive geriatric assessment (CGA) is recommended for proper management of older cancer pts, there is still no consensus on the best form of geriatric assessment. We investigated possible prognostic factors in elderly metastatic (m)CRC pts in a real-world setting, focusing on the role of the oncological-multidimensional prognostic index (onco-MPI). Methods: Pts aged $ 70 years with mCRC referred to the Medical Oncology 1 Unit from May 2010 to May 2017 were assessed by a multidisciplinary team and received a basal CGA. Onco-MPI was calculated by a validated algorithm as a weighted linear combination of the CGA domains, as previously described. The following 3 different prognostic groups were identified: low (scores 0.0-0.46), medium (scores 0.47-0.63) and high risk (scores 0.64-1.0). Results: A total of 206 mCRC pts were included, 123 males. Mean age was 76.1 years (69.2-90.8). ECOG PS was , 2 in 90% and mini-mental state examination was $ 24 in 85% of pts. Primary tumor was located in rectum, left and right side in 18%, 42% and 40% of pts, respectively. RAS and BRAF mutations were detected in 44% and 9% of pts, respectively. According to onco-MPI score, 32%, 39% and 28% of cases were low, medium and high risk, respectively. According to CGA as per Balducci’s criteria, 56% of pts were classified as fit, 31% vulnerable and 13% frail. Median overall survival (OS) was 26 months (95% CI 19.7-32.4). The following factors were significantly associated with OS: ECOG PS (0-1 vs . 1, 31% vs 15%, p = 0.004); onco-MPI score (low vs medium vs high risk, 29% vs 38% vs 19%, p = 0.005), treatment (monotherapy vs doublet vs triplet, 20% vs 31% vs 30%, p = 0.01). No significant difference in OS was observed in CGA-based groups (p = 0.15). In high onco-MPI score, doublet-regimen correlated with higher OS compared to monotherapy (79% vs 51%, p = 0,03). Conclusions: Onco-MPI emerged as a significant prognosticator in mCRC elderly pts. It may be useful in daily clinical practice for driving decision-making in this age group. Thanks to its marked standard- ization it may be also applied in clinical trials.

11553 Poster Session (Board #245), Mon, 1:15 PM-4:15 PM

Treatment completion and toxicity of trastuzumab or trastuzumab + lapatinib in older patients (pts): BIG 2-06; NCCTG N063D (Alliance).

Noam Falbel Ponde, Dominique Agbor-Tarh, Lissandra Dal Lago, Larissa A. Korde, Florentine Hilbers, Christian Jackisch, Olena Werner, Richard D. Gelber, Aminah Jatoi, Amylou C. Dueck, Alvaro Moreno- Aspitia, Christos Sotiriou, Evandro de Azambuja, Martine Piccart; AC Camargo Cancer Center, S~ao Paulo, Brazil; Frontier Science Scotland (FSS), Kincraig, United Kingdom; UFSM, Santa Maria, Brazil; Head, Breast Cancer Therapeutics, Clinical Investigations Branch, National Cancer Institute, Bethesda, MD; Breast International Group, Brussels, Belgium; Sana Klinikum, Offenbach, Germany; Novartis, Basel-City, Switzerland; Dana-Farber Cancer Institute, Harvard Medical School, Harvard TH Chan School of Public Health, and Frontier Science and Technology Research Foundation, Boston, MA; Mayo Clinic, Rochester, MN; Mayo Clinic, Scottsdale, AZ; Mayo Clinic, Jacksonville, FL; Institut Jules Bordet, Brussels, Belgium

Background: Little is known about the toxicity of trastuzumab (T) or of trastuzumab + lapatinib (T + L), approved in the advanced setting, in older pts. We have performed a sub-analysis of the Adjuvant Lapatinib and/or Trastuzumab Treatment Optimisation (ALTTO) trial focused on treatment completion and toxicity of T and T+L in older pts (aged $65 years (yr)). Methods: The ALTTO trial (NCT00490139, NCCTG N063D) randomised 8381 pts with early HER2+ BC into 4 arms and we included the T and T+L arms in our analysis. Eligible pts for our study were those having received at least one dose of assigned treatment. Treatment completion was evaluated through the rate of temporary treatment interruptions (TTI), permanent treatment discontinuations (PTD) and lapatinib dose reductions (LDR). Toxicity was evaluated via a selected set of adverse events of interest (AEIs). Risk factors for TTI, PTD, LDR and AEIs were assessed, including comorbidities and polypharmacy at baseline (defined as use 5 or more co- medications) and AEIs during treatment. Results: A total of 430 pts$65-year-old were identified for this sub-analysis, out of a total of 4190 pts with a median age of 68 yrs (range 65-80). Older pts were more likely to have comorbidities (70% vs 38%). Treatment completion was worse among older pts in the T+L arm but not in the T arm (Table). AEIs were more common in the T+L arm in all patients, with older patients having higher AEI rates (78.04% in older vs 63.38% in younger), particularly diarrhea (60.75% vs 38.0%). Identified risk factors (multivariate) for worse treatment completion in the T and T+ L arms included concomitant use of chemotherapy and the occurrence of grade 3 adverse events, among others. Conclusions: T + L has worse treatment completion and is more toxic in older patients, while T was well tolerated. Identifiable risk factors at baseline and during the course of treatment could be used to aid in regimen selection and management for both T and T + L in their respective indications. Support: UG1CA189823, Novartis;https://acknowledgments.alliancefound.org.

TT+L >=65 < 65 years years < 65 years >= 65 years Events Events Multivariate* OR (95% Events Events Multivariate* OR (95% (% of N) (% of N) CI) (% of N) (% of N) CI) PTD 300 44 (20.37) 1.33 625 (33.26) 115 2.08 (15.95) (0.93 to 1.91) (53.74) (1.55 to 2.79) P= 0.122 P,0.001 TTI 573 75 (34.72) 1.09 1176 162 1.72 (30.46) (0.80 to 1.48) (62.59) (75.70) (1.23 to 2.42) p= 0.602 p=0.002

*Multivariate analysis includes comorbidity, polypharmacy, type and timing of chemotherapy.

11554 Poster Session (Board #246), Mon, 1:15 PM-4:15 PM

Functional impairment on admission and associated symptom burden and health outcomes among hospitalized patients with advanced cancer.

Daniel E Lage, Areej El-Jawahri, Charn-Xin Fuh, Richard Newcomb, Vicki Jackson, Joseph Greer, Jennifer S. Temel, Ryan David Nipp; Massachusetts General Hospital, Boston, MA

Background: Hospitalized patients with cancer often have impaired function, as measured by activities of daily living (ADLs), related to age, comorbidities, and both cancer and treatment-related morbidity. However, the relationship between functional impairment and patients’ symptom burden and clinical outcomes has not been well described. Methods: We prospectively enrolled patients with advanced cancer with unplanned hospitalizations at an academic medical center. Upon admission, nurses assessed patients’ ADLs (mobility, feeding, bathing, dressing, and grooming). We used the Edmonton Symptom Assessment Scale (ESAS) and Patient Health Questionnaire-4 to assess physical and psychological symptoms, comparing symptom burden between patients with and without ADL impairment. We used regression models adjusted for age, sex, education, Charlson comorbidity index, months since advanced cancer diagnosis, and cancer type to assess the relationship between any ADL impairment on admission and hospital length of stay, the composite outcome of death or readmission within 90 days of discharge, and survival. Results: Among 932 patients, 40.2% had at least one ADL impairment. Patients with ADL impairment were older (Mean = 67.2 vs 60.8 years, p , 0.001), had higher Charlson comorbidity index (Mean = 1.1 vs 0.7, p , 0.001), and higher physical symptom burden (ESAS Physical Mean = 35.2 vs 30.9, p , 0.001). Those with ADL impairment were more likely to have moderate to severe constipation (46.7% vs. 36.0%, p , 0.01), pain (74.9% vs. 63.1%, p , 0.01), drowsiness (76.6% vs. 68.3%, p , 0.01), as well as symptoms of depression (38.3% vs. 23.6%, p , 0.01) and anxiety (35.9% vs. 22.4%, p , 0.01). In adjusted models, ADL impairment was associated with longer hospital length of stay (B = 1.30, p , 0.01), higher odds of death or readmission within 90 days (odds ratio = 2.26, p , 0.01), and higher mortality (hazard ratio = 1.73, p , 0.01). Conclusions: Hospitalized patients with advanced cancer who have functional impairment experience a significantly higher symptom burden and worse health outcomes compared to those without functional impairment. These findings highlight the need to assess and address functional impairment among this population to enhance their quality of life and care.

11555 Poster Session (Board #247), Mon, 1:15 PM-4:15 PM

Obesity paradox in older cancer patients for middle and long-term mortality: A prospective multicenter cohort study of 2,071 patients.

Claudia Martinez-Tapia, Thomas Diot, Nadia Oubaya, Elena Paillaud, Johanne Poisson, Mathilde Gisselbrecht, Philippe Caillet, Aurelie ´ Baudin, Fred´ ´ eric Pamoukdjian, Sylvie Bastuji- Garin, Marie Laurent, Florence Canoui-Poitrine; EA 7376 CEpiA (Clinical Epidemiology and Ageing Unit), Creteil, ´ France; UPEC, Creteil, ´ France; Hopital Henri Mondor, Creteil, ´ France; Hopital Europeen Geoges Pompidou, Paris, France; Hopitalˆ Europeen ´ Georges-Pompidou, Paris, France; Unitede ´ Recherche Clinique, Hopitalˆ Henri-Mondor, Creteil, ´ France; Clinical Epidemiology and Aging (CEPIA)-EA7376, UPEC. APHP, Hopitalˆ Avicenne, H.U.P.S.S.D., Bobigny, France; Sante Publique Methodologie Hopital Henri Mondor APHP, Creteil, ´ France

Background: Overweight and obesity are associated with numerous adverse health outcomes. However, among older adults, substantial literature suggests an improved survival among overweight and obese patients. This phenomenon, referred to as the “obesity paradox” remains controversial. In the context of cancer, the association between overweight/obesity and mortality is complex due to the concomitant weight loss and cachexia. We aim to assess the impact of high Body Mass Index (BMI) on mortality in a large population of older cancer patients. Methods: We studied patients aged $70 from the ELCAPA prospective open-cohort (2007-2016; 10 geriatric oncology clinics, Great Paris urban area). Endpoints were 12- and 48-months mortality. A variable combining BMI at cancer diagnosis and weight loss (in the 6 months preceding the diagnosis) was created. BMI categories considered: underweight, (BMI , 22.4kg/m²), normal weight (BMI 22.5-24.9), overweight (BMI 25-29.9), and obese (BMI $30); weight loss (WL) categories: , 5%, 5- , 10%, $10%. Univariate and multivariate Cox proportional- hazards analysis were conducted in males and females. Results: A total of 2071 patients were included (mean age, 81; female, 48%; metastases, 49%; main localizations: digestive (37%), urinary (26%), breast (16%); underweight (30%), normal weight (23%), overweight (33%), obese (14%)). After adjustment for age, smoking, inpatient status, cancer site, metastasis, performance status and comorbidities, overweight women with , 5% WL were at lower risk of 12- and 48-months mortality compared to normal weight women with , 5% WL (aHR = 0.56; p = 0.043 and 0.65; p = 0.031, respectively). Obese women with , 5% WL had a lower mortality risk only at 48-months (aHR = 0.63; p = 0.036). The obesity paradox was stronger in overweight metastatic women. Overweight or obese women with 5- , 10% or $10% WL had no reduced risk of mortality compared to normal weight women. Overweight and obese men had no reduced risk of mortality irrespective of weight loss. Conclusions: By taking into account initial weight loss, we did not found evidence for obesity paradox in older patients with cancer except in the subgroup of women with minimal weight loss. Clinical trial information: NCT02884375.

11556 Poster Session (Board #248), Mon, 1:15 PM-4:15 PM

Identifying patient-reported anxiety and depression in older adults with cancer.

Reena Jayani, Can-Lan Sun, Kemeberly Charles, Enrique Soto Perez De Celis, Leana Chien, Elsa Roberts, Jeanine Moreno, William Dale, Supriya Gupta Mohile, Mina S. Sedrak, Marianna Koczywas, Vincent Chung, Marwan Fakih, Joseph Chao, Mihaela C. Cristea, Sumanta K. Pal, Vani Katheria, Arti Hurria, Daneng Li; H. Lee Moffitt Cancer Center and Research Institute, Tampa, FL; City of Hope, Duarte, CA; Instituto Nacional de Ciencias Medicas y Nutricion Salvador Zubiran, Mexico City, Mexico; CIY, Chicago, IL; University of Rochester Medical Center, Rochester, NY; City of Hope National Medical Center, Duarte, CA; Department of Medical Oncology and Therapeutics Research, City of Hope, Duarte, CA; The Judy and Bernard Briskin Center for Clinical Research, City of Hope, Duarte, CA; City of Hope Comprehensive Cancer Center, Duarte, CA; Memorial Sloan Kettering Cancer Center, New York, NY

Background: Anxiety and depression are associated with decreased quality of life, treatment adherence, and survival in patients with cancer. Mental Health Inventory (MHI-17) is a validated screening tool for psychological well-being, but cut points for older adults with cancer are unknown. The goal of this study is to identify cut points on MHI-17 Anxiety (MHI-A) and Depression (MHI-D) subscales which correlate with patient-reported anxiety and depression in older adults with cancer. Methods: This is a secondary analysis of baseline data from a randomized controlled trial in adults aged 65+ with solid tumors starting chemotherapy. At baseline, patients completed MHI-17. MHI-A and MHI-D were calculated (range 0-100; higher scores represent better mental health). Self-reported anxiety was obtained from single-item Linear Analog Scale Assessment (0-5 = low, 6-10 = high). Self-reported depression was obtained from Yale Depression Screen, “Do you often feel sad or depressed?” The association of MHI-A and MHI-D with the patient-reported outcomes was analyzed using logistic regression. Youden’s index was used to determine the optimal cut points for MHI-A and MHI-D for identifying patients with high anxiety and depression. Results: 458 patients (median age 71 (range 65-91), 57% female, 55% non-Hispanic white) were included in this analysis. The most common cancer types were: GI (31%), breast (19%), GU (18%), and pulmonary (16%); 75% had stage IV cancer. Twenty-four percent (N = 110) reported high anxiety and 21% (N = 97) depression. Median scores for MHI-A and MHI-D were 75 (range 0-100) and 80 (range 0-100). The optimal cut point for high anxiety on MHI-A was 65; this had an accuracy of 76.1%, a sensitivity of 71.8%, and a specificity of 77.5%. The optimal cut point for depression on MHI-D was 70; this had an accuracy of 80.1%, a sensitivity of 80.4%, and a specificity of 79.8%. Conclusions: The current study identified optimal cut points for MHI-Anxiety and MHI-Depression subscales to identify older adults with cancer starting chemotherapy with self-reported anxiety and depression. In the absence of patient-reported anxiety and depression, these cut points could be used to identify older patients with cancer at risk for poor mental health.

11557 Poster Session (Board #249), Mon, 1:15 PM-4:15 PM

Association of obesity with breast cancer outcome in relation to cancer subtypes.

Ana Elisa Lohmann, Sara V. Soldera, Isabel Pimentel, Domen Ribnikar, Marguerite Ennis, Eitan Amir, Pamela Jean Goodwin; Lunenfeld-Tanenbaum Research Institute, Mount Sinai Hospital, Toronto, ON, Canada; Princess Margaret Cancer Centre, Toronto, ON, Canada; Mount Sinai Hospital- University of Toronto, Toronto, ON, Canada; Applied Statistician, Markham, ON, Canada; Lunenfeld-Tanenbaum Research Institute, Sinai Health System, Toronto, ON, Canada

Background: Obesity at breast cancer (BC) diagnosis is associated with poor outcome, although the magnitude of effect in different BC subtypes is uncertain. Here we report on the association of obesity at BC diagnosis with disease-free (DFS) and overall survival (OS) in the following subtypes: (i) hormone receptor (ER/PgR) +ve, HER2-ve, (ii) HER2+ve, any ER/PgR and (iii) triple negative (TN). Methods: We searched MEDLINE, EMBASE and COCHRANE databases to December 31, 2018 and meeting pre- sentations (past 5 years) using predefined search terms. Study eligibility, data abstraction were performed independently by two authors; those reporting hazard ratios (HR) for obesity and DFS/OS in BC subtypes were included. Using Review Manager pooled HRs were computed and weighted using generic inverse variance in fixed and random effects models (results were similar, random effects are presented). Results: Of 10,702 titles, 26 studies (108,793 patients) were included. Pooled HR for DFS for obese vs non-obese were (i) ER/PgR+ve HER2-ve 1.21 (95% Confidence interval, CI; 1.12-1.31, p , 0.00001), (ii) HER2+ve, any ER/PgR 1.16 (95%CI, 1.06-1.26; p = 0.0006) and (iii) TN, 1.13 (95%CI; 1.05-1.22 p = 0.002). Pooled HRs for OS were (i) ER/PgR+ve, HER2-ve 1.45 (95%CI; 1.30-1.62 p , 0.00001), (ii) HER2+ve any ER/PgR 1.21 (95%CI; 1.10-1.34 p = 0.0001) and (iii) TN 1.13 (95%CI, 1.04-1.23, p = 0.003).PooledHR for OS (but not DFS) were somewhat higher in observational vs interventional studies in (i) ER/PgR+ve, HER2-ve 1.57 vs 1.36, HER2+ve any ER/PgR (ii) 1.37 vs 1.09 but not (iii) TN 1.12 vs 1.22 (p = 0.21, 0.03 and 0.48, respectively). Conclusions: Obesity was associated with a worse outcome in all BC subtypes. Higher HR for OS in observational studies in (i) ER/PgR+ve, HER2- and (ii) HER2+ve any ER/PgR BC may reflect selection of healthier patients for intervention trials.

11558 Poster Session (Board #250), Mon, 1:15 PM-4:15 PM

Physiological and psychosocial effects of a highly structured exercise program on breast cancer survivorship.

Judy A. Tjoe, Linda B. Piacentine, Jun Yin, Paula Papanek, Alexander V. Ng; TORQUE, Aurora Health Care/ECOG/ACRIN, Milwaukee, WI; Marquette University College of Nursing, Milwaukee, WI; Aurora Research Institute, Milwaukee, WI; Marquette University, Milwaukee, WI; Marquette University College of Health Sciences, Milwaukee, WI

Background: Exercise after breast cancer treatment improves cancer-related outcomes, although the mechanism of action is unclear. Engagement in healthy active lifestyles after cancer treatment may also impact overall survival. We aimed to determine effectiveness of a highly structured, clinically overseen, goal oriented, group triathlon training program on improving physiological and psychosocial outcomes in female breast cancer survivors (BCS). Methods: After stage appropriate local and systemic breast cancer treatment, 53 female BCS were recruited to participate in this study. The 14-week group triathlon training program was individually adjusted for treatment side effects. 28 similar BCS who did not participate in the training served as controls. Pre- and post-exercise training measures included: Functional endurance (Timed 6MWT), quality of life (QOL, FACT-B), cancer-related fatigue (CRF, FACIT-F), exercise self-efficacy (ESE) via questionnaires, estradiol, and inflammatory biomarkers (C-reactive protein, TNFr1, leptin and adiponectin). Results: Complete data were obtained from 41 [mean age 51 (7) yr, mean BMI 29.5 (6.2)] triathlon finishers and 16 [mean age 56 (10) yr, mean BMI 31.5 (8.0)] controls. 6MWT improved (26 m) more in the triathlon group (p , 0.05). FACT-B, FACIT-F, and ESE all improved compared to controls (all p , 0.05). Body mass and BMI decreased in the training group compared to controls (p = 0.01, p = 0.04 respectively). Arm circumference decreased in the trained group but increased in controls (p , 0.05). Estradiol and leptin positively correlated with initial body weight in both groups but did not change after training. Adiponectin significantly decreased in the triathlon group (p = 0.01), perhaps due to selection bias of controls. No significant changes were seen in other serological markers. Conclusions: Ahighly structured, clinically overseen, moderate intensity exercise program can improve endurance, QOL, CRF, body mass, and possibly improve survival after breast cancer treatment. Furthermore, improvements in ESE may provide tools for patients to continue adherence to regular exercise guidelines that may lower obesity and its related comorbidities, including arm lymphedema.

11559 Poster Session (Board #251), Mon, 1:15 PM-4:15 PM

Tobacco retail availability and tobacco cessation in lung and head and neck (HN) cancer survivors.

Lawson Eng, Jie Su, Steven Habbous, Katrina Hueniken, M. Catherine Brown, Deborah Saunders, John R. de Almeida, Andrew Hope, Andrea Bezjak, Frances A. Shepherd, Natasha B. Leighl, Andrew Pierre, Peter Selby, David Paul Goldstein, Wei Xu, Meredith Elana Giuliani, William K. Evans, Geoffrey Liu, Michael Chaiton; Division of Medical Oncology and Hematology, Princess Margaret Cancer Centre, Toronto, ON, Canada; Department of Biostatistics, Princess Margaret Cancer Centre, Toronto, ON, Canada; Princess Margaret Cancer Centre, Toronto, ON, Canada; Northeast Cancer Centre of Health Sciences North, Sudbury, ON, Canada; Department of Otolaryngology-Head & Neck Surgery/Surgical Oncology, Princess Margaret Cancer Centre, Toronto, ON, Canada; Department of Radiation Oncology, Princess Margaret Cancer Centre, Toronto, ON, Canada; National Cancer Institute of Canada Clinical Trials Group, Kingston, Toronto, ON, Canada; Cancer Clinical Research Unit, Princess Margaret Cancer Centre, Toronto, ON, Canada; Department of Thoracic Surgery, Toronto General Hospital, University Health Network, Toronto, ON, Canada; Centre for Addiction and Mental Health, Toronto, ON, Canada; McMaster University, Hamilton, ON, Canada

Background: Continued smoking after a cancer diagnosis is associated with poorer outcomes. Tobacco retail availability is negatively associated with cessation in non-cancer patients (pts), but has not been explored in cancer survivors. We evaluated the impact of tobacco retail availability on cessation in lung and HN cancer pts. Methods: Lung and HN cancer pts (Princess Margaret Cancer Centre, Toronto) completed questionnaires evaluating changes in tobacco use with a median of 26 months apart. Validated tobacco retail location data were obtained from Ministry of Health and pt home addresses were geocoded using ArcGIS 10.6.1, which calculated walking time/distance to nearest vendor, and vendor density within 250 meters (m) and 500m from pts. Multivariable logistic regression and Cox proportional hazard models evaluated the impact of vendor availability on cessation and time to quitting after diagnosis respectively, adjusting for significant clinico-demographic and tobacco covariates. Results: 242/721 lung and 149/445 HN pts smoked at diagnosis; subsequent overall quit rates were 66% and 49% respectively. Mean distance and walking time to a vendor was 1 km (range 0-13) and 11 min (range 0-156). On average, there was one vendor (range 0-19) within 250m and four vendors (range 0-40) within 500m from pts; 37% and 61% of pts lived within 250m and 500m from at least one vendor respectively. Greater distance (aOR 1.18 per 1000m [95% CI 1.00-1.38] p = 0.05) and increased walking time (aOR 1.01 per minute [1.00-1.02] p = 0.05) were associated with quitting at one year. Living within 250m (aOR 0.52 [0.32-0.84] p = 0.008) or 500m (aOR 0.57 [0.35-0.92] p = 0.02) to at least one vendor reduced quitting at one year. Living near more vendors within 500m had an increasing dose effect on reducing cessation rates at one year (aOR 0.96 per vendor [0.93-1.00] p = 0.05). Living within 500m to a vendor reduced chance of quitting at any time (aHR 0.66 [0.48-0.91] p = 0.01). HN and lung subgroups revealed similar associations. Conclusions: Close access to tobacco retail outlets is associated with reduced cessation rates for lung and HN cancer survivors. Reducing density of tobacco vendors is a cessation strategy that can positively impact cancer pt outcomes.

11560 Poster Session (Board #252), Mon, 1:15 PM-4:15 PM

Analysis of financial sustainability of survivorship clinics led by advanced practice providers.

Maria Alma Rodriguez, Guadalupe R. Palos, Katherine Ramsey Gilmore, Paula A. Lewis-Patterson, Patricia Chapman, Weiqi Bi; The University of Texas MD Anderson Cancer Center, Houston, TX; University of Texas MD Anderson Cancer Center, Houston, TX

Background: Disease specific Survivorship Care Clinics (SCs) have been established within a comprehensive cancer center. Clinics are staffed by Advanced Practice Providers (APPs), Physician Assistants and Advanced Practice Registered Nurses, with experience in the management of each disease type. To determine the sustainability of this model of survivorship care, we analyzed the professional fees’ revenue generated by APPs’ billings for 6 clinics and then compared the APPs’ salaries across all clinics. Methods: A retrospective analysis was conducted of 6 survivorship clinic’s patient volumes and clinic days supported by APPs from 9/1/16-4/30/17. The full FTE salary of the APPs, including benefits were prorated to the time dedicated to each of the SCs. Institutional financial data was used to align professional fees to actual reimbursements received. Salary recovery percentage was calculated as the ratio of reimbursement received to prorated FTE salary. Results: Table shows variation in APPs’ salary commensurate to FTE proportion. Results also indicate there was an average of 99% professional fee recovery. Clinics with an FTE proportion . 0.5 had recovery higher than the anticipated prorated salary, suggesting there is a threshold to maximize efficacy and sustainability. Conclusions: APPs professional fees for care provided to cancer survivors are reimbursable, across disease types or payers, and proportionally supports their salaries. Our findings suggest delivery models based on APPs to manage care of long-term survivors can be self-supporting.

Comparison of APP’s Professional Fees Across 6 Survivorship Clinics. APP salary % of salary re- prorated to covered through Clinic FTE (includ- Professional Professional professional Clinic FTE ing fringe) Billing Reimbursement reimbursement Genitourinary 0.9 $86,982.00 $224,385.00 $92,861.35 106.76% Gynecology 0.2 $19,329.33 $40,544.00 $17,361.58 89.82% Head & Neck 0.4 $38,658.67 $90,582.00 $37,037.54 95.81% Lymphoma 0.2 $19,329.33 $39,835.00 $17,658.98 91.36% Melanoma 0.3 $28,994.00 $53,497.00 $27,428.26 94.60% Thyroid 0.6 $57,988.00 $140,946.00 $67,748.53 116.83%

11561 Poster Session (Board #253), Mon, 1:15 PM-4:15 PM

Impact of health behavior change on health utility (HU) and financial toxicity in head and neck cancer (HNC) survivors.

Lawson Eng, Katrina Hueniken, M. Catherine Brown, Andrew Hope, Meredith Elana Giuliani, Peter Selby, Kelvin K. Chan, Nicole Mittmann, Wei Xu, David Paul Goldstein, Geoffrey Liu, John R. de Almeida; Division of Medical Oncology and Hematology, Princess Margaret Cancer Centre, Toronto, ON, Canada; Princess Margaret Cancer Centre, Toronto, ON, Canada; Department of Radiation Oncology, Princess Margaret Cancer Centre, Toronto, ON, Canada; Centre for Addiction and Mental Health, Toronto, ON, Canada; Sunnybrook Health Sciences Centre, Odette Cancer Centre, University of Toronto, Toronto, ON, Canada; HOPE Research Centre, Sunnybrook Hospital, Toronto, ON, Canada; Department of Otolaryngology-Head & Neck Surgery/Surgical Oncology, Princess Margaret Cancer Centre, Toronto, ON, Canada

Background: Health behavior changes including tobacco cessation and increasing physical activity (PA) are important aspects of cancer survivorship. Understanding how these behaviours impact on HU and financial toxicity will help when evaluating survivorship programs. We evaluated the impact of tobacco cessation and PA on HU, function and financial toxicity among HNC patients (pts). Methods: HNC pts from Princess Margaret Cancer Centre completed questionnaires at baseline (diagnosis) and 12 months between 2014-2018 evaluating tobacco use, PA with the Godin questionnaire, cancer related monthly out of pocket costs (OOPC), HU using HU Index Mark 3, function using Lawton Brody Scale (LBS) and lost annual income. Multivariable linear regression analyses evaluated the impact of health behaviour change on OOPC, HU, LBS and lost income. Results: Among 296 pts, mean age 61, 76% male; 29% smoked at diagnosis, 60% quit 1 year after; 26% met PA guidelines at diagnosis, 52% continued to meet guidelines at 1 year. 19% of those not meeting PA guidelines at diagnosis, met them at 1 year. Among all, mean HU [SEM] was 0.84 [0.01] (baseline), 0.80 [0.01] (12 months); mean monthly OOPC [SEM] were $171 [27] (12 months); mean annual lost individual income was $25897 [2945]. Among smokers at diagnosis, those continuing to smoke at 1 year lost a mean of $21272 (95% CI [$2783-39761] P= 0.03) more in individual annual income compared to pts who quit, adjusted for baseline income and education. Current smokers who quit smoking at 1 year had an adjusted mean increase in HU of 0.15 ([0.00-0.30] P=0.05) greater than pts continuing to smoke. Pts who continued meeting PA guidelines at 1 year had an adjusted mean increase in HU scores of 0.11 ([0.02-0.20], P= 0.02) compared to those reducing PA levels after diagnosis. Changes in PA and tobacco were not associated with change in function or OOPC; improving to meet PA guidelines after diagnosis was not associated with HU or lost income (P. 0.05). Conclusions: Quitting smoking and maintaining PA levels after diagnosis were associated with improvements in HU scores; quitting smoking reduced lost income. Cancer survivors should be made aware of the potential economic impact of behaviour change.

11562 Poster Session (Board #254), Mon, 1:15 PM-4:15 PM

Impact of overweight, obesity, and post-treatment weight changes on occupational reintegration of breast cancer (BC) survivors.

Antonio Di Meglio, Gwenn Menvielle, Agnes Dumas, Arnauld S. Gbenou, Thomas Bovagnet, Elise Martin, Arlindo R. Ferreira, Laurence Vanlemmens, Olivier Arsene, Mahmoud Ibrahim, Johanna Wassermann, Anne-Laure Martin, Jerome Lemonnier, Lucia Del Mastro, Lee Jones, Ann H. Partridge, Jennifer A. Ligibel, Fabrice Andre, Stefan Michiels, Ines Maria Vaz Duarte Luis; Institut Gustave Roussy, Villejuif, France; Institut Pierre Louis d’Epidemiologie ´ et de Sante ´ Publique, Paris, France; Centre Oscar Lambret, Lille, France; Centre Hospitalier de Blois, Blois, France; Centre Hospitalier Regional ´ d’Orleans, ´ Orleans, ´ France; Pitie-Salp ´ etriˆ ere ` Hospital, Sorbonne University, Cancer University Institute, Paris, France; Unicancer, Paris, France; Department of Medical Oncology, IRCCS AOU San Martino–IST, National Cancer Institute, Genova, Italy; Memorial Sloan Kettering Cancer Center, New York, NY; Dana-Farber Cancer Institute, Boston, MA

Background: Overweight and obesity are strongly linked to poorer BC-specific outcomes, quality of life and financial burden in cancer care. Weight loss interventions have the potential to improve such outcomes. Fewer data exist on whether excess weight and post-diagnosis weight changes impact the ability of BC survivors to return to work (RTW). Methods: CANTO (NCT01993498) is a multicenter prospective longitudinal study of 12000 patients (pts) with stage I-III BC that characterizes long-term toxicities of BC treatment. Of 5801 pts enrolled from 2012-2014 (last data lock), we identified 1874 pts who were professionally active at BC diagnosis, $5 years (yrs) younger than minimum legal retirement age (62 yrs) and with updated work status 2 yrs after diagnosis. Logistic regression models evaluated the impact of body mass index (BMI) at diagnosis and of weight changes over 2 yrs after diagnosis on odds of non-RTW, adjusting for age, education, income, BC treatment and recreational physical activity (PA). Results: 37% pts were overweight or obese at diagnosis (BMI $25 kg/m2): 34% of them gained $5% and 16% lost $5% weight after diagnosis. Rates of non-RTW at 2 yrs were significantly higher in overweight or obese vs under or normal weight pts (27% vs 18%, p#.001; adjusted odds ratio 1.37, 95% Confidence Interval [CI] 1.04-1.80, p = .017). Overweight and obese pts who did not RTW experienced higher increments in weight (mean [95% CI]: +3.6% [+2.3, +4.9] vs +1.5% [+0.8, +2.2]) and reported more modest changes in PA (mean [95% CI]: +1.0 [-1.4, +3.5] vs +2.1 [+0.8, +3.3] MET-h/week) vs those who did RTW. Weight changes independently impacted odds of non-RTW in overweight and obese pts (p for interaction weight change*BMI #.001): a 5% weight gain was associated with 17% increase in adjusted odds of non-RTW (95% CI 2-35%, p = .024), whereas a loss $5% with 60% reduced odds of non-RTW vs weight gain (95% CI 18-82%, p = .013). Conclusions: Excess weight and weight changes are significantly associated with occupational reintegration after BC in overweight and obese pts. Randomized studies testing dedicated weight control interventions should also measure outcomes of social rehabilitation in this large subset of survivors. Clinical trial information: NCT01993498.

11563 Poster Session (Board #255), Mon, 1:15 PM-4:15 PM

Neurocognitive outcomes in adult survivors of neuroblastoma: A report from the Childhood Cancer Survivor Study.

Caroline Hesko, Wei Liu, Deokumar Srivastava, Tara M. Brinkman, Lisa Diller, Todd M. Gibson, Kevin C. Oeffinger, Wendy Leisenring, Rebecca M. Howell, Gregory T. Armstrong, Kevin R. Krull, Tara O. Henderson; University of Chicago Comer Children’s Hospital, Chicago, IL; St. Jude Children’s Research Hospital, Memphis, TN; Dana-Farber Cancer Institute and Harvard Medical School, Boston, MA; Duke University, Durham, NC; Fred Hutchinson Cancer Research Center, Seattle, WA; Department of Radiation Physics, The University of Texas MD Anderson Cancer Center, Houston, TX; The University of Chicago, Chicago, IL

Background: Long-term survivors of neuroblastoma may be at risk for neurocognitive impairment due to young age at diagnosis and intensive multimodal therapies. Methods: 837 survivors of neuroblastoma (57% female; median [range] age 25 [17-58] years, age at diagnosis 1 [0-21] years) and 728 siblings (56% female; age 32[16-43] years) self-reported neurocognitive problems using a neurocognitive questionnaire. Impairment was defined as scores $90th percentile of siblings in emotional regulation (ER), organization, task efficiency (TE), and memory. Multivariable log-binomial models evaluated associations with treatment exposures, era and chronic conditions (Grade 2-4 CTCAE v5) adjusting for sex, age, and race. Analyses were stratified by age at diagnosis (#1and. 1 year) as proxy for risk group. Results: Rates of impairment were 19.7% (ER), 25.3% organization, 21.9% TE and 19.4% for memory. Survivors had 50% higher risk of impaired TE (#1 year relative risk [RR] 1.48, 95% confidence interval [CI] 1.08-2.03; . 1 year: RR 1.58, CI 1.22-2.06) and ER (#1 year RR 1.51, CI 1.07-2.12; . 1 year RR 1.44, CI 1.06-1.95) versussiblings. Among survivors #1 year at diagnosis, treatment with platinum (RR 1.74, CI 1.01-2.97), hearing loss (RR 1.95, CI 1.26-3.00), cardiovascular (RR 1.83, CI 1.15-2.89) and neurologic (RR 2.00, CI 1.32-3.03) conditions were associated with higher risk of impaired TE. Female sex (RR = 1.54, CI, 1.02-2.33), cardiovascular (RR 1.71, CI 1.08-2.70) and respiratory (RR 1.99, CI 1.14-3.49) conditions were associated with higher risk of impaired ER. Among survivors . 1 year at diagnosis those treated in 1970-79 vs. 1990-99 had 80% higher risk of impaired ER (RR 1.77, CI 1.02-3.06). Hearing loss (RR 1.56 (1.09-2.24), respiratory (RR 2.35, CI 1.60-3.45) and cardiovascular (RR 1.74, CI 1.12-2.69) conditions were associated with higher risk of impaired TE. Conclusions: Adult survivors of neuroblastoma are at-risk for neurocognitive impairment. Differences associated with age at diagnosis, chronic disease and treatment exposures may inform risk-stratified inventions to improve neurocognitive outcomes. Reduced risk in later eras may reflect improved supportive care and knowledge of late effects.

11564 Poster Session (Board #256), Mon, 1:15 PM-4:15 PM

Return to work after breast cancer: Comprehensive longitudinal analyses of its determinants.

Agnes Dumas, Ines Maria Vaz Duarte Luis, Thomas Bovagnet, Antonio Di Meglio, Mayssam El- Mouhebb, Sandrine Pinto, Cecile Charles, Sarah Dauchy, Charles Coutant, Paul H. Cottu, Anne Lesur, Florence Lerebours, Olivier Tredan, Laurence Vanlemmens, Christelle Levy, Jerome Lemonnier, Christel Mesleard, Patrick Arveux, Fabrice Andre, Gwenn Menvielle; Institut Gustave Roussy, Villejuif, France; Dana-Farber Cancer Institute, Boston, MA; Institut Pierre Louis d’Epidemiologie ´ et de Sante ´ Publique, Paris, France; Gustave Roussy, Villejuif, France; INSERM, Paris, France; Institut Gustave Roussy, Department of Psycho Oncology, Villejuif, France; Centre Georges-François Leclerc, Dijon, France; Institut Curie, Paris, France; Institut de Cancerologie ´ de Lorraine, Vandœuvre-Les- Nancy, France; Departement ´ d’Oncologie Medicale, ´ Centre Leon ´ B´erard, Lyon, France; Centre Oscar Lambret, Lille, France; Centre François Baclesse, Department of Medical Oncology, Caen, France; Unicancer, Paris, France; R&D Unicancer, Paris, France; Biostatistics and Quality of Life Unit, Centre Georges François Leclerc and EA 4184, Dijon, France

Background: The interplay between breast cancer (BC) late effects, psychosocial and work-related factors in return to work (RTW) is not well understood. Previous reports were retrospective and did not combine all these features. Methods: We used data of a French prospective cohort study (CANTO, NCT01993498) of stage I-III BC patients (pts) including detailed clinical data of 1,874 pts working at diagnosis (dx) and $5 years younger than legal retirement age. The outcome was non-RTW 2 years after dx. Multivariable regressions were conducted to identify correlates of non-RTW. First, we examined the independent effect of treatments, toxicities (Common Toxicity Criteria Adverse Events), and patient reported outcomes (EORTC BR23 and FA12; Hospital Anxiety and Depression Scale) collected shortly after end of primary treatment. Then, in a restricted sample of 1,003 pts with working conditions (WC) information available, we fitted models to account for detailed pre-dx WC including type of contract, working hours, strenuous postures, supportive environment, degree of autonomy and perception of work. All models were adjusted for age, stage, marital status, socioeconomic status and comorbidities. Results: Two years after dx, 21% of pts did not work. Adjusted odds of non-RTW were increased among pts treated with combinations of chemotherapy (CT) and trastuzumab (TR) (e.g. OR of CT-TR = 2.20 [95% CI 1.24-3.88] and OR of CT-TR-hormonotherapy (HT) = 1.72 [1.13-2.63] vs. treated only with CT-HT), who had severe arm morbidity (OR = 1.73 [1.27-2.36] vs. no), severe emotional fatigue (OR = 1.55 [1.03-2.32] vs. no), anxiety (OR = 1.51 [1.02-2.23] vs. no), or depression (OR = 2.23 [1.27- 3.94] vs no). In addition, we also found that the odds of non-RTW were increased among pts who had shift working hours (OR = 2.23 [1.32-3.76] vs. no), who did not work in a supportive environment before dx (OR = 2.24 [1.44-3.50] vs. supportive) and who perceived their job as boring (OR = 3.57 [1.71-7.46] vs. not boring). Conclusions: More than 1/5 of pts did not RTW 2 years after dx, with treatment (trastuzumab), clinical, psychological and work-related factors being associated with job reintegration. Multidisciplinary strategies are needed to support BC survivors.

11565 Poster Session (Board #257), Mon, 1:15 PM-4:15 PM

Polypharmacy risk among five-year cancer survivors.

Christine D. Hsu, Jennifer Leigh Lund, Hazel Nichols; University of North Carolina at Chapel Hill, Chapel Hill, NC; Department of Epidemiology, University of North Carolina at Chapel Hill, Chapel Hill, NC; University of North Carolina, Chapel Hill, NC

Background: Cancer survivors are increasing in the United States. Survivors often face cancer sequelae and side effects from treatment, which can arise during treatment or even months or years later. Survivors may therefore experience greater medication burden than the general population, increasing concerns for polypharmacy and risk of drug interactions and non-adherence. The objective of this study is to characterize prevalence of polypharmacy by cancer history in a nationally-representative sample of U.S. adults. Methods: Using National Health and Nutrition Examination Survey (NHANES) data from 2003-2014, we identified 32,330 individuals 20 years or older, 1,672 of whom had been diagnosed with cancer (excluding non-melanoma skin) at least five years before the survey. The association between cancer history and polypharmacy (5+ medications) was examined using Poisson regression models to calculate multivariable risk ratio (RRs) and 95% confidence intervals (CIs), adjusted for education, ethnicity, marital status, and age at survey. Results: Among five-year cancer survivors, 35% had a diagnosis within 5-9 years of the survey year, 26% within 10-14 years, 13% within 15-19 years, and 26% had a diagnosis 20 or more years before the survey year. Breast cancer was the most common type of cancer (23%), followed by cervical cancer (18%), prostate cancer (12%), and colon cancer (7%). Prevalence of polypharmacy was higher in cancer survivors (33.6%; 95% CI: 30.7-36.6%) than in those with no cancer history (13.0%, 95% CI: 12.4-13.7%) (RR: 1.32, 95% CI: 1.20, 1.45). Polypharmacy prevalence differed by age group (4.83% in 20-39 year olds; 40.37% in 40-64 year olds; 54.8% in 65 years and older), and within each age group, risk of polypharmacy was higher in cancer survivors than in those with no cancer history, with the most pronounced difference in cancer survivors 20-39 years old (RR: 2.93, 95% CI: 1.69-5.09), followed by 40-64 year olds (RR: 1.52, 95% CI 1.27-1.83) and those 65 years and older (RR: 1.17, 95% CI: 1.07-1.29). Conclusions: Cancer survivors are more likely to experience polypharmacy burden than those with no cancer history. Findings from this study can increase awareness about the unique challenges cancer survivors face and encourage medication reconciliation services.

11566 Poster Session (Board #258), Mon, 1:15 PM-4:15 PM

High temporal variability of clinical side effects with and without adjuvant chemotherapy in 4,684 early breast cancer patients in the CANTO trial.

Paul H. Cottu, Marie-Laure Tanguy, Enora Laas, Olivier Tredan, Sylvie Giacchetti, Sophie Guillermet, Jean-Marc Ferrero, Olivier Rigal, Carole Tarpin, Mario Campone, Christelle Levy, Patrick Soulie, Laurence Vanlemmens, Marion Fournie, ´ Christelle Jouannaud, Sibille Everhard, Jerome Lemonnier, Patrick Arveux, Ines Maria Vaz Duarte Luis, Fabrice Andre; Institut Curie, Paris, France; Institute Curie, Paris, France; Departement ´ d’Oncologie Medicale, ´ Centre Leon ´ B´erard, Lyon, France; Hopital Saint Louis, Breast Disease Center, Paris, France; Centre Eugene ` Marquis, Rennes, France; Department of Medical Oncology, Centre Antoine Lacassagne, Nice, France; Centre Henri-Becquerel, Department of Medical Oncology, Rouen, France; Institut Paoli Calmettes, Marseille, France; Institut de Cancerologie ´ de l’Ouest, Rene ´ Gauducheau, St Herblain, France; Centre François Baclesse, Department of Medical Oncology, Caen, France; Institut de Cancerologie ´ de l’Ouest, Angers, France; Centre Oscar Lambret, Lille, France; Institut Bergonie, ´ Bordeaux, France; Institut Jean Godinot, Reims, France; Unicancer, Paris, France; Biostatistics and Quality of Life Unit, Centre Georges François Leclerc and EA 4184, Dijon, France; Institut Gustave Roussy, Villejuif, France

Background: CANTO (CANcer TOxicities - NCT01993498) is a French multicenter prospective longitudinal study dedicated to the quantification and characterization of side effects after treatment for patients with stage I-III breast cancer, and to the development of predictors of toxicities. Methods: The CANTO study has included . 12,000 patients, assessed at diagnosis, and 3-6 months (T0), 15 months (T12), and year 3 and 5 after treatment completion. In the current report, we focus on key toxicities at T0 and T12 according to chemotherapy (CT) administration. For each side effect, 4 populations were defined according to its occurrence at T0 and/or T12 (no/no, no/yes, yes/no, yes/yes). Results: We analyzed 4684 patients with T0 and T12 consolidated data. Median age at diagnosis was 57y (22-89). Patients (pts) had HR+/HER2-, HER2+ or triple negative tumors in 78.9%, 12.4% and 8.7% of cases, respectively. Overall, 2516 pts (53.7%) received CT. Most CT pts (81%) received a sequential anthracyclines– taxanes schedule. As an example, a high proportion of patients presented neurological symptoms including cognitive symptoms, sensory or motor neuropathy, paresthesia, headache, etc (all grades) at either T0 or T12. Overall, CT was strongly associated with neurotoxicity at all times (OR = 2.27, p , 0.0001). However, proportions of patients with neurological side effects changed between T0 and T12. The table shows the proportions of the 4 categories of pts in the CT and no CT groups. Furthermore, at T12, neurological symptoms remained more frequent in the CT group, whether pts had symptoms at T0 (CT vs no CT, 81% vs 77%, p = 0.007) or not (CT vs no CT, 41% vs 36%, p = 0.03). Similar temporal trends were observed (with specific percentages for each considered side effect) for detailed neurological toxicities, pain and joint/bone toxicity (stratified on endocrine therapy), gastro-intestinal, pulmonary and cardiac toxicities. Conclusions: Overall, symptoms burden is extremely high at T0 and T12 after treatment, and much higher in pts receiving CT. A high temporal variability was observed in all subsets, including a clinically meaningful delayed onset of e.g. neurological side effects. Clinical trial information: NCT01993498. neurological side effects at T0 then at T12 CT (%) no CT (%) No No 17.7 32.8 No Yes 12.3 18.3 Yes No 13.4 11.5 Yes Yes 56.7 37.4

11567 Poster Session (Board #259), Mon, 1:15 PM-4:15 PM

Female fertility preservation (FP) at pediatric cancer centers: A report from the Children’s Oncology Group (COG).

Jennifer Levine, Gwendolyn P. Quinn, James Klotsky, Joanne Frankel Kelvin, Brooke Cherven, Elyse Bryson, Sameeya Ahmed-Winston, Natasha Frederick, Julienne Brackett, Lillian Meacham, David Robert Freyer, Christopher Dvorak, Eric Jessen Chow; Weill Cornell Medicine, New York, NY; NYU Langone Health, New York, NY; Childrens Healthcare of Atlanta, Atlanta, GA; Memorial Sloan Kettering Cancer Center, New York, NY; Children’s Healthcare of Atlanta, Atlanta, GA; Children’s National, Washington, DC; Connecticut Children’s Medical Center, Hartford, CT; Baylor College of Medicine - Texas Children’s Cancer and Hematology Centers, Houston, TX; Emory University, Atlanta, GA; Children’s Hospital Los Angeles, Los Angeles, CA; University of California San Francisco, San Francisco, CA; Fred Hutchinson Cancer Research Center, Seattle, WA

Background: Preserved fertility after cancer is a priority for female survivors and their families. Embryo/ oocyte cryopreservation are standard of care (SOC) for post-pubertal females. Experimental ovarian tissue cryopreservation (OTC) is the only current option for pre-pubertal girls. We surveyed COG sites about their FP infrastructure and practices. Methods: A REDcap survey was emailed to one individual previously identified as knowledgeable about FP or the Principal Investigator at each COG site. Site specific factors associated with outcomes were determined using logistic regression. All study procedures were IRB-approved. Results: Responses were received from 144 of 220 sites (65%). Discussions about fertility at diagnosis were reported as routinely held with all females “at risk” of infertility, all post pubertal females, and all females at 113 (78%), 94 (65%), and 65 (45%) of sites respectively. Embryo/oocyte cryopreservation was offered at 95 (70%) institutions and independently associated with large (.120 new patients/year) sites (OR 6.0 95%CI 1.6-22.8) and presence of a FP navigator/team (OR 4.7 95%CI 1.7-13.5). OTC was offered at 64 (48%) sites: 34 (25%) by referral to another institution, 18 (13%) under an IRB protocol, and 12 (9%) as a clinical service. OTC accessibility was associated with large sites (OR 3.2 95% CI 1.1-8.9) and a FP navigator/team (OR 3.2 95%CI 1.4-7.0). A total of 102/133 (77%) sites use gonadotropin releasing hormone analogues (GnRHa) for any indication; 90 (68%) for menstrual suppression, 75 (56%) with the goal of ovarian suppression for fertility preservation, and 27 (20%) for contraception. Conclusions: Variation in FP services exists across COG. Discussion of infertility risk is not universal. The availability of OTC at treating institutions is limited. The presence of an FP navigator/team is a modifiable factor associated with greater likelihood of accessing SOC and experimental options. Despite conflicting evidence and lack of endorsement from professional societies, GnRHa’s are commonly used for FP. These survey results suggest FP services remain inadequate but highlight opportunities for improvement and areas of needed research.

11568 Poster Session (Board #260), Mon, 1:15 PM-4:15 PM

Effect of race/ethnicity on long-term cytopenias and major infections in adolescent young adult breast cancer survivors.

Candice Sauder, Melanie Goldfarb, Alicia Gingrich, Qian Li, Theodore Wun, Theresa Keegan; Com- prehensive Cancer Center, University of California Davis, Sacramento, CA; John Wayne Cancer Institute, Santa Monica, CA; Center for Oncology Hematology Outcomes Research and Training (COHORT), UC Davis Comprehensive Cancer Center, Sacramento, CA

Background: Many Adolescent and Young Adult (AYA) Breast Cancer (BC) patients receive chemotherapy as part of their initial treatment. Long-term bone marrow suppression is a potential complication, but no studies have evaluated the impact of race/ethnicity on the development in AYA BC survivors. Methods: Female patients ages 15-39 diagnosed with BC during 1996-2012 and surviving $ 2 years were obtained from the California Cancer Registry and linked to statewide hospitalization data. We estimated the cumulative incidence of developing anemia, leukopenia, or major infection/sepsis ($ 2 years after diagnosis), accounting for death as a competing risk, and examined the impact of race/ethnicity using multivariable Cox proportional hazards regression. Results: Of 14,729 patients, 48.8% were non- Hispanic white, 8.3% non-Hispanic black, 25.5% Hispanic, and 16.5% Asian/Pacific Islander. At diagnosis, 95.5% had local or regional disease (27.7% stage I, 49.4% stage II), and were mostly treated with surgery (96.2%) and chemotherapy (74.3%). The 10-year cumulative incidence of anemia (16.8% vs 11.7%), leukopenia (4.6% vs 2.1%), and major infection/sepsis (13.2% vs 7.9%) was greater following initial treatment with chemotherapy (p , 0.0001 for all vs no chemotherapy). In multivariable analyses controlling for sociodemographic factors, baseline comorbidities, treatment and stage, Blacks had the highest risk (vs. non-Hispanic whites) of medical late effects, including anemia [HR: 1.62, CI 1.41-1.86], leukopenia (HR: 1.53, CI 1.17-2.00), and major infection/sepsis (HR: 1.36, CI 1.16-1.60). Similarly, Hispanic and Asian/Pacific Islanders had a higher risk of developing anemia (HR: 1.16, CI 1.04-1.29; HR: 1.17, CI 1.03-1.33) and trended toward developing more leukopenia (HR: 1.24, CI 1.00-1.54; HR: 1.25, CI 0.98-1.61). Conclusions: AYAs of Black, Hispanic, and Asian/Pacific Islander race/ethnicity are at an increased risk of anemia and leukopenia after chemotherapy compared with non- Hispanic Whites. With improvements in prognostic testing resulting in potential decreased chemotherapy usage, there may be a decrease in long-term late effects for these young cancer survivors.

11569 Poster Session (Board #261), Mon, 1:15 PM-4:15 PM

The CAROLE (CArdiac Related Oncologic Late Effects) Study: A Phase II, single-arm feasibility trial.

Lindsay Puckett, Shahryar Saba, Sonia Henry, Stacey Rosen, Elise Rooney, Alison Laxer, Philip Gilbo, Karalyn Pappas, Katherine Eacobacci, Amitha Kapyur, Justin Robeny, Samaria Filosa, Samantha Musial, Anisha Chaudhry, Rahul Chaudhry, Martin Lesser, Adam Riegel, Jessica Charlton, Sariah Ramoutarpersaud, Lucille Lee; Medical College of Wisconsin, Milwaukee, WI; Northwell Health and Zucker School of Medicine at Hofstra/Northwell, New Hyde Park, NY

Background: Long-term breast cancer survivors are at risk of late effects from cardiotoxic (tox) radiation (RT) and chemotherapy (chemo). However, there is a paucity of data to recommend appropriate cardiac screening for those ~10 years (yrs) after diagnosis (dx). This phase II, single arm study assessed multimodality screening in a heterogeneous cohort of long-term survivors. Methods: Our prospective, single center pilot study enrolled 201 (of 200 planned) eligible patients (pts): age 18-65 at dx, with treatment (tx) for breast cancer (any stage/tx) in 2004-2011 ($6 yrs prior), with no cardiac disease (dz) at dx. The primary endpoint was presence of any cardiac dz (preclinical or clinical) on electrocardiogram (EKG), echocardiogram (echo), or coronary artery calcium CT (CAC CT); secondary endpoint included clinical (clin) dz alone. Subgroups assessed were: no tox chemo or RT (no tox), tox RT (left-sided) alone, tox chemo (anthracycline) alone, and both tox chemo and tox RT (tox chemoRT). Results: After enrollment (6/2017-7/2018), 200 pts had sufficient data for analysis. Median age was 50 (29-65) at dx, 63 (37-77) at imaging, and interval (dx to imaging) was 11.5 (6.7-14.5) yrs. There were 44% no tox, 31.5% tox RT, 16% tox chemo, and 8.5% tox chemoRT pts. Among all pts, 77.6% displayed evidence of any cardiac dz, 51.5% had clin dz. Per modality, rates of any and clin dz were: 27.1%/10% EKG, 50.0%/25.3% echo, and 50.8%/45.8% CAC CT. Among subgroups, these rates were: 73.9%/53.4% no tox, 82.5%/58% tox RT, 75%/38.7% tox chemo, and 82.4%/35.3% tox chemoRT (NS on x2 test, p = .58/p = .15). Multivariable logistic regression with tox RT, tox chemo, and age as covariates, showed no significant tox RT (NS) or tox chemo effect (NS), however, increasing age was significantly associated with higher incidence of cardiac dz (p , 0.01; OR = 1.12 per yr, 95% CI: 1.07-1.18). Conclusions: Among pts with a median of . 10 yrs post breast cancer tx, multimodality CAROLE screening was feasible and showed high rates of cardiac dz (any/clin = 77.6%/51.5%) among all subgroups, including those without tox tx. This study provides evidence for long-term cardiac screening in a diverse group of breast cancer survivors and provides a novel pathway for evaluation. Clinical trial information: NCT03235427.

11570 Poster Session (Board #262), Mon, 1:15 PM-4:15 PM

Temporal trends among survivors of rhabdomyosarcoma: A report from the Childhood Cancer Survivor Study (CCSS).

Pooja Hingorani, Mingjuan Wang, Deokumar Srivastava, Paul C. Nathan, Kevin C. Oeffinger, Yutaka Yasui, Wendy M. Leisenring, Rebecca M. Howell, Emily L. Mueller, Todd M. Gibson, Suzanne L. Wolden, Kari L. Bjornard, William H. Meyer, Carola A. S. Arndt, Leslie L. Robison, Daniel M. Green, Gregory T. Armstrong, Kirsten K. Ness; Phoenix Children’s Hospital, Phoenix, AZ; St. Jude Children’s Research Hospital, Memphis, TN; The Hospital for Sick Children, Toronto, ON, Canada; Duke University, Durham, NC; Fred Hutchinson Cancer Research Center, Seattle, WA; Department of Radiation Physics, The University of Texas MD Anderson Cancer Center, Houston, TX; Indiana University, Indianapolis, IN; Memorial Sloan Kettering Cancer Center, New York, NY; University of Oklahoma Health Science Center, Oklahoma City, OK; Mayo Clinic, Rochester, MN

Background: Intergroup Rhabdomyosarcoma Study Group (IRSG) protocols included treatment modifications, which may have ameliorated late health outcomes for rhabdomyosarcoma (RMS) survivors treated in more recent era. Methods: We evaluated chronic health conditions (CHCs) and late mortality ( . 5 years from diagnosis) among survivors treated 1970-1990 (IRSG I-III) and 1991-1999 (IRSG IV), and associations with specific treatments to identify treatment-related factors for adverse outcomes. Asso- ciations between treatments and CHCs and mortality were evaluated using Fine and Gray’s proportional hazards method accounting for competing risks. Results: 856 survivors treated 1970-90 (median diagnosis age 5.4 years [0- 20]) and 306 treated 1991-99 (median diagnosis age 5.5 years [0-20]) were included. Significant exposure differences between eras included higher percentage (53% vs. 17%, p , 0.01) receiving $ 20gm/m2 cumulative alkylators in 1991-99, but more receiving platinums (13% vs 5%, p , 0.01) and abdomen/ pelvis radiation (29% vs. 23%, p = 0.04) in 1970-90. 20-year cumulative incidence for any (40% vs. 28%, p , 0.01), $2(16%vs.7%,p, 0.01), and endocrine (8% vs. 2.5%, p , 0.01) grade 3-5 CHCs was higher in 1970-90 compared to 1991-99. The hazard ratio (HR) for any (HR 0.7, 95% Confidence Interval [CI] 0.55-0.9), $2 (HR 0.38, 95% CI 0.22-0.66) and endocrine (HR 0.25, 95% CI 0.09-0.67) grade 3-5 CHC was lower for 1991-99 survivors than 1970-90. The effect of era (1991-99 vs 1970-90, HR 0.73; 95% CI 0.59-0.91) on CHC was not attenuated when treatment variables were added to multivariable model. Exposures with increased risk of grade 3-5 CHC included platinums (hearing, HR 2, 95% CI 1.07-3.8), anthracycline $250mg/m2 (cardiovascular, HR 2.7, 95% CI 1.2-6) and abdomen/ pelvis radiation (second malignant neoplasms, HR 2.1, 95% CI 1.1-4, gastrointestinal, HR 7.4, 95% CI 3.5-16 and endocrine, HR 2.5, 95% CI 1.4-4.4). Gonadal dysfunction was the most common endocrine CHC. There was no difference in all cause or cause-specific mortality between the two cohorts. Conclusions: RMS survivors from the IRSG IV era are at reduced risk for late onset chronic health conditions compared to previous era.

11571 Poster Session (Board #263), Mon, 1:15 PM-4:15 PM

Cardiovascular disease risk in survivors of 20 adult cancers.

Helen Strongman, Sarah Gadd, Anthony Matthews, Kathryn Mansfield, Susannah Jane Stanway, Alexander R. Lyon, Isabel dos-Santos-Silva, Liam Smeeth, Krishnan Bhaskaran; London School of Hygiene and Tropical Medicine, London, United Kingdom; University of Leeds, Leeds, United Kingdom; Royal Marsden Hospital, London, United Kingdom; Imperial College London, London, United Kingdom

Background: There are concerns about long-term cardiovascular disease (CVD) risk in cancer survivors, but few studies have quantified the risks for a wide range of cancers and specific CVD outcomes. Methods: Using UK electronic health records, we identified cohorts of adults alive one year after a cancer diagnosis at 20 different sites. Risks of a range of CVD outcomes were compared to age, sex and general practice matched cancer free controls using Cox regression; crude and adjusted models were compared to investigate the role of shared cancer/CVD risk factors (e.g. smoking and diabetes). Results: 126 120 cancer survivors and 603 144 controls were followed over a median (IQR) 4.6 (2.5-8.1) and 5.6 (3.2-9.2) years. Crude and adjusted hazard ratios (HRs) were similar. In adjusted models, there was strong evidence (p,0.01) of increased risk of CVDs among cancer survivors compared with controls: venous thrombo- embolism (VTE, 18 cancers), heart failure/cardiomyopathy (7 cancers), arrhythmia (4 cancers), and stroke (3 cancers). In stratified analyses HRs were higher in younger people and continued beyond 5 years post diagnosis. Conclusions: We found increased long term CVD risk among survivors of several cancers compared to the general population, which varied by cancer site and specific CVD outcome.

Adjusted HRs (95% CIs) for the 10 most common cancers. Coronary artery Heart failure/ Cancer site disease Stroke Arrhythmia cardiomyopathy VTE Female breast 0.8 (0.7-0.9) 1.1 (0.9- 1.0 (0.9- 1.2 (1.1-1.3) 2.1 (1.9- 1.2) 1.1) 2.4) Prostate 1.0 (0.9-1.1) 1.0 (0.9- 1.0 (1.0- 1.0 (0.9-1.1) 1.8 (1.6- 1.1) 1.1) 2.0) Colorectal 0.8 (0.7-0.9) 1.0 (0.9- 1.0 (0.9- 0.9 (0.8-1.0) 2.6 (2.3- 1.2) 1.1) 3.0) Bladder 1.3 (1.1-1.5) 1.2 (1.0- 1.1 (1.0- 0.9 (0.8-1.1) 2.0 (1.6- 1.3) 1.2) 2.3) Malignant skin 1.1 (0.9-1.3) 1.2 (1.0- 1.3 (1.1- 1.0 (0.8-1.2) 1.8 (1.4- melanoma 1.6) 1.4) 2.3) Lung 1.1 (0.8-1.4) 1.6 (1.2- 1.5 (1.3- 1.4 (1.1-1.8) 4.7 (3.6- 2.1) 1.8) 6.2) Non Hodgkin 1.2 (1.0-1.5) 1.4 (1.1- 1.4 (1.2- 1.9 (1.5-2.3) 2.2 (1.7- lymphoma 1.9) 1.6) 2.9) Leukemia 0.9 (0.7-1.2) 0.9 (0.7- 1.1 (0.9- 1.3 (1.0-1.6) 1.9 (1.5- 1.2) 1.3) 2.6) Uterus 0.9 (0.6-1.1) 1.0 (0.7- 1.1 (0.9- 0.9 (0.6-1.2) 2.3 (1.7- 1.4) 1.3) 3.0) Ovary 1.1 (0.7-1.7) 1.2 (0.8- 1.0 (0.8- 2.0 (1.4-3.0) 5.6 (4.1- 1.8) 1.3) 7.6)

11572 Poster Session (Board #264), Mon, 1:15 PM-4:15 PM

Total body irradiation (TBI) and risk of breast subsequent malignant neoplasm (SMN) after blood or marrow transplantation (BMT): A report from the BMT survivor study (BMTSS).

Andrew Michael McDonald, Yanjun Chen, Jessica Wu, Lindsey Hageman, Liton Francisco, Michelle Kung, Emily Ness, Kevin Battles, Wendy Landier, Donna E. Salzman, Daniel Jordan Weisdorf, Stephen J. Forman, Mukta Arora, Saro Armenian, Smita Bhatia; Department of Radiation Oncology, University of Alabama at Birmingham, Birmingham, AL; University of Alabama at Birmingham, Birmingham, AL; City of Hope, Monrovia, CA; Univ of Alabama Birmingham, Birmingham, AL; University of Minnesota, Minneapolis, MN; City of Hope, Duarte, CA; City of Hope Comprehensive Cancer Center, Duarte, CA

Background: The association between TBI and breast SMN in BMT recipients is not clearly defined. We address this limitation to develop evidence for screening guidelines for those at risk. Methods: Study participants were drawn from BMTSS – a multi-site retrospective cohort of patients who had received BMT between 1974 and 2014 and survived $2y post-BMT. Participants completed the BMTSS survey that included sociodemographics and health conditions. Breast SMN was confirmed by pathology report review. Clinical characteristics, pre-BMT and BMT exposures were abstracted from medical records. Using multivariable Cox proportional hazards models, freedom from breast SMN was measured from birth and TBI was treated as a time-varying covariate in analyses stratified by type of BMT (allogeneic; autologous). Results: 1546 female participants (allogeneic 808; autologous 738) received BMT at a mean age of 43.1617.6y and were followed for 9.467.7y; primary diagnoses: HL/NHL (28%), AML/ MDS (27%), PCD (19%), other (28%). TBI was used in 671 patients (allogeneic 57%; autologous 30%). Patients with pre-BMT chest radiation (n = 45) were excluded from the analysis. A total of 38 cases of breast SMN were identified (allogeneic 19; autologous 19). Allogeneic BMT: exposure to TBI (HR = 3.4; 95%CI 1.1-10.9, p = 0.04) and age at BMT , 40y (HR = 4.0; 95%CI 1.3-12.8, p = 0.02) were associated with increased risk of breast SMN. Risk of breast SMN was higher and breast SMN developed earlier among those exposed to TBI before age 40y vs. those exposed after 40y (8.7% vs. 0% by attained age 50; 10.6% vs. 5.8% by attained age 60, p = 0.003). Autologous BMT: Age , 40y was associated with a 4.8-fold higher risk (95%CI 1.1-20.1, p = 0.03) of breast SMN. TBI was associated with a 2-fold higher risk (p = 0.2). The cumulative incidence of breast SMN among those exposed to TBI at age , 40y vs. age $40y was 5.1% vs. 2.4% by attained age 60 (p = 0.07). Conclusions: TBI is associated with breast SMN among allogeneic BMT recipients. The risk is especially high and occurs at a younger age, among those exposed at age , 40y. These findings provide evidence for initiating screening at a young age after exposure to TBI.

11573 Poster Session (Board #265), Mon, 1:15 PM-4:15 PM

Second solid (SMN) and hematologic malignant neoplasms (HMN) among 24,900 United States testicular cancer survivors (TCS) after chemotherapy (CHEM), radiotherapy (RT), or surgery only (SURG).

Mohammad Issam Abu Zaid, Michael T. Milano, Paul C Dinh, Sophie Fossa,˚ Hongmei Yang, Darren R. Feldman, Lois B. Travis, Chunkit Fung; Indiana University School of Medicine, Indianapolis, IN; University of Rochester Medical Center, Rochester, NY; National Advisory Unit on Late Effects After Cancer Treatment, Oslo, Norway; Memorial Sloan Kettering Cancer Center, New York, NY

Background: No large, population-based U.S. study has comprehensively examined SMN and HMN risks after TC, taking into account initial therapy and focusing on recent decades. Methods: Standardized incidence ratios (SIR) for SMN and HMN stratified by site and time since TC diagnosis were calculated for 24,900 TCS (median age, TC diagnosis: 33 y) reported to population-based cancer registries in the NCI SEER program (1973-2014). TCS were initially given CHEM (n=6,340), RT (n=9,058), or SURG (n=8,995), with each group accruing 80,700, 156,735, and 128,039 person-years (PY) of follow-up, respectively. Results: During 372,709 PY of follow-up, 1,625 TCS developed SMN and 228 developed HMN, including 107 lymphomas, 92 leukemias, and 29 plasma cell dyscrasias. Among all TCS, overall risk of SMN was increased by 1.06-fold (95% CI 1.01-1.1). Risks of SMN were increased after RT (SIR 1.1, 95% CI 1.06-1.2) and CHEM (SIR 1.3, 95% CI 1.1-1.4); but not after SURG (SIR 0.8). After CHEM, significant excesses of SMN of the pancreas (SIR 2.2), soft tissue (SIR 4.0), kidney (SIR 1.7), thyroid (SIR 3.3) occurred; after RT, significantly elevated risks for SMN of the stomach (SIR 1.7), rectum/ sigmoid (SIR 1.4), pancreas (SIR 2.7), soft tissue (SIR 2.2), bladder (SIR 1.5), and thyroid (SIR 2.0) were observed. The 30 year cumulative incidence of SMN after SURG, CHEM, and RT was 8.9% (95% CI 7.8- 9.9), 10.1% (95% CI 8.8-11.5) and 17.0% (95% CI 15.8-18.2), respectively. Significantly increased risks of leukemia followed CHEM (SIR 2.7) and SURG (SIR 1.8) and were driven by increased risks for acute myeloid leukemia, with significant excesses restricted to 1-10 y and 1-5 y after TC diagnosis, respectively; nonsignificant 2-fold excesses occurred 1-10 y after RT. Risks for lymphoma and plasma cell dyscrasias were not elevated (1.02 and 1.27, respectively). Conclusions: In the largest population- based study of U.S. TCS to date, we report significant 6% excesses of SMN and 2-fold increased risks of leukemias. Efforts to minimize CHEM exposure and decrease doses/field size of RT in TC should continue. TCS should be educated about cancer prevention and screening.

11574 Poster Session (Board #266), Mon, 1:15 PM-4:15 PM

Body weight changes in young breast cancer survivors and associated predictors.

Tal Sella, Zhenying Tan-Wasielewski, Shoshana M. Rosenberg, Philip Daniel Poorvu, Kathryn Jean Ruddy, Shari I. Gelber, Rulla M. Tamimi, Lidia Schapira, Steven E. Come, Jeffrey M. Peppercorn, Virginia F. Borges, Ann H. Partridge, Jennifer A. Ligibel; Dana Farber Cancer Institute, Boston, MA; Department of Data Sciences, Dana-Farber Cancer Institute, Boston, MA; Dana-Farber Cancer Institute, Boston, MA; Mayo Clinic, Department of Oncology, Rochester, MN; Brigham and Women’s Hospital, Boston, MA; Stanford Cancer Center, Palo Alto, CA; Beth Israel Deaconess Medical Center, Boston, MA; Massachusetts General Hospital, Boston, MA; University of Colorado Comprehensive Cancer Center, Aurora, CO

Background: Weight gain after cancer diagnosis is common in cancer survivors and has been linked to increased treatment toxicity, poor quality of life, and increased risk of second cancers and overall mortality. Young breast cancer (BC) survivors may be especially susceptible to weight changes given the impact of treatments such as chemotherapy and hormonal therapy on menopausal status. Methods: We identified women with Stage 0-III breast cancer diagnosed at #40 years (y) between 2006-2016 from a multi-center prospective cohort study. Clinical data including self-reported pre-diagnosis and follow-up weights were obtained using baseline and follow-up patient surveys. Participants missing baseline weight, pregnant at diagnosis/within 1y of diagnosis or with BC recurrence within 1y were excluded; those pregnant or with BC recurrence between 1-3y from diagnosis were excluded from the 3y analysis. Menopausal status at baseline and treatment-related amenorrhea (TRA) in follow-up were defined by self-reported last menstrual period. Factors associated with weight gain (.5%) were evaluated using univariate two-sided Fisher’s exact test. Results: At baseline, 1y and 3y post diagnosis, 956, 899 and 687 women were eligible for analysis respectively. Median age at diagnosis was 37y (17 - 40), 65% received endocrine therapy and 74% chemotherapy. Premenopausal status was verified in 94% at baseline. Mean BMI at baseline was 24.4 (SD 5.3) kg/m2; 20% (187/956) were overweight and 12% (116/956) obese. At 1y and 3y, mean BMI increased modestly to 24.7 (SD 5.6) and 24.9 (SD 5.2), respectively with weight gain (.5%) observed in 18% (164/899) and 13% (87/687) respectively. 37% (300/804) and 32% (196/615) of eligible premenopausal subjects experienced TRA at 1y and 3y, respectively. Receipt of chemotherapy, receipt of endocrine therapy and TRA were not associated with weight gain at any timepoint. Conclusions: In this large prospective cohort of young BC survivors, mean BMI increased only modestly over time. Self-reported weight gain was not associated with treatment and not exacerbated by TRA. Further analysis to understand the effects of physical activity and other predictors of weight gain in this population are ongoing.

11575 Poster Session (Board #267), Mon, 1:15 PM-4:15 PM

Prospective trial in early-stage breast cancer (EBC) patients (pts) submitted to nutrition evidence-based educational intervention: Early results of adherence to dietary guidelines (ADG) and body weight change (BWC).

Luisa Carbognin, Ilaria Trestini, Isabella Sperduti, Clelia Bonaiuto, Valentina Zambonin, Elena Fiorio, Daniela Tregnago, Veronica Parolin, Sara Pilotto, Giovanni Scambia, Giampaolo Tortora, Michele Milella, Emilio Bria; Fondazione Policlinico Universitario Agostino Gemelli IRCCS, Rome, Italy; Azienda Ospeda- liera Universitaria Integrata, Verona, Italy; Regina Elena National Cancer Institute IRCCS, Biostatistics, Rome, Italy; University of Verona, Azienda Ospedaliera Universitaria Integrata, Verona, Italy; Fondazione Policlinico Universitario A Gemelli IRCCS, Universita ` Cattolica del Sacro Cuore, Rome, Italy; Fondazione Policlinico Universitario Gemelli IRCCS, Rome, Italy; Fondazione Policlinico Universitario Agostino Gemelli IRCCS, Universita ` Cattolica del Sacro Cuore, Rome, Italy

Background: Weight gain and overweight have been related to an increased risk of recurrence and mortality in patients affected by EBC. However, the adherence to nutritional intervention is not entirely explored. The aims of this prospective study were to evaluate the ADG and the effectiveness of nutritional intervention in terms of BWC in pts with EBC undergoing treatment. Methods: Entry criteria: EBC pts candidate to neoadjuvant/adjuvant therapy. At study entry, pts received a nutrition evidence- based tailored educational intervention by a skilled dietitian. Anthropometric and dietary assessments were performed. ADG was estimated through the validated Med-Diet 14-item questionnaire. Health- Related Quality of Life was analyzed with the EORTC QLQ-C30. Descriptive statistics was adopted. Associations between variables and groups according to nutritional variables were analysed (Chi-square test). Results: From February 2016 to December 2018, 204 pts were enrolled (median age 49 years): 27.5%/72.5% neoadjuvant/adjuvant treatment. At baseline, 2.5% of pts were underweight, 41.7% were normal weight, 33.3% were overweight and 22.5% were obese. Moreover, 47.5% of pts gained $5% of their usual weight and 25.5% performed physical activity. Significant nutritional impact symptoms were reported: dyspepsia (51.5%), constipation (62.3%) and dysgeusia (34.8%). The majority of pts presented dietary patterns high in fat (median fat intake: 35.2% from total energy intake) and low in dietary fiber (median dietary fiber intake: 17.2 g/day). A significant correlation between baseline BMI and tension was observed (p , 0.0001) as well as BMI and worry, irritability and depression (p , 0.0001, p , 0.0001 and p = 0.008, respectively). Six months after the intervention, the median ADG was high (median Med-Diet score was 12). A high adherence to nutrition guidelines (defines as a Med-Diet score $10, 112 patients) significantly correlated with a weight loss $5% from the baseline weight (p= 0.005). Furthermore, the weight loss $5% was correlated with a lower rate of depression (p=0.05). Conclusions: A tailored nutritional intervention for EBC pts undergoing adjuvant or neoadjuvant therapy has the potential to improve their ADG, in order to achieve a weight loss.

11576 Poster Session (Board #268), Mon, 1:15 PM-4:15 PM

Cardiorespiratory fitness and cardiovascular mortality after prolonged androgen deprivation therapy for prostate cancer.

Jingyi Gong, David Payne, Jesse Pittard Caron, Camden Bay, Bradley Alexander McGregor, Jon Hainer, Ann H. Partridge, Tomas G. Neilan, Marcelo Di Carli, Anju Nohria, John Groarke; Brigham and Women’s Hospital Heart and Vascular Center, Boston, MA; Brigham and Women’s Hospital Heart and Vascular Center, Boston, MA; Brigham and Women’s Hospital Center for Clinical Investigation, Boston, MA; Dana-Farber Cancer Institute, Boston, MA; Brigham and Women’s Hospital, Boston, MA; Massachu- setts General Hospital, Boston, MA; Brigham and Women’s Hospital Noninvasive Cardiovascular Imaging Program, Department of Medicine (Cardiovascular Division) and Department of Radiology, Boston, MA

Background: Androgen deprivation therapy (ADT) plays a pivotal role in management of prostate cancer (PC), with prolonged ADT favored over short-term use in the definitive treatment of high risk PC with radiation. Objectives: To compare cardiorespiratory fitness (CRF) and cardiovascular (CV) mortality among patients with PC with and without ADT exposure, and to explore how duration of ADT exposure influences CRF and CV mortality risk. Methods: This is a retrospective study of patients referred for exercise treadmill testing (ETT) after a diagnosis of PC. PC risk classification was based on Gleason score (GS) at diagnosis: high risk GS $ 8, intermediate risk GS= 7, and low risk GS # 6. CRF was categorized according to metabolic equivalents (METs): METs $ 8 defined as good CRF and METs , 8 as reduced CRF. ADT exposure was grouped as short-term (# 6 months) versus prolonged (. 6 months). Results: 616 patients underwent an ETT a median of 4.8 years (interquartile range: 2.0-7.9) after diagnosis of PC. 150 patients (24.3%) received ADT prior to ETT; 51 with short-term versus 99 with prolonged exposure. 524 (85.1%) patients had $ 2 CV risk factors, and 28 CV deaths occurred over 4.2 (interquartile range: 2.3- 7.1) years following the ETT. Reduced CRF was more frequent among ADT-exposed versus ADT-naive patients (48.7 versus 32.6%, p, 0.001). Prolonged ADT was associated with reduced CRF (odds ratio (OR): 2.71; 95% confidence interval (CI): 1.31-5.61; p=0.007) and increased CV mortality (hazard ratio (HR): 3.87; 95% CI: 1.16-12.96; p=0.03) in adjusted analyses. In contrast, short-term ADT exposure was not independently associated with either reduced CRF (OR 1.71; 95% CI: 1.00-2.94); p=0.05) or CV mortality (HR: 1.60; 95% CI: 0.51-5.01; p=0.42). Conclusions: Among patients with PC and high baseline CV risk, . 6 months ADT exposure but not less was associated with reduced CRF and increased CV mortality. Reduced CRF may in part mediate increased CV mortality risk. Exercise interventions concurrent with prolonged ADT warrants investigation to potentially offset risk.

11577 Poster Session (Board #269), Mon, 1:15 PM-4:15 PM

Combination of olanzapine and aprepitant in the prevention of chemotherapy-induced nausea and vomiting (CINV) in breast cancer patients.

Mukesh Shanthilal, Sathya M, Akshay Jk; Mysore Medical College and Research Institute, Mysore, India

Background: Olanzapine and Aprepitant have been shown to be a safe and effective agent for the prevention of CINV. This study aims to compare Olanzapine, Aprepitant and their combination in the prevention of CINV. Methods: Prospective randomized controlled study in breast cancer patients receiving doxorubicin 60mg/m2 and cyclophosphamide 600mg/m2 chemotherapy. Female patient; age, $ 18; chemotherapy na¨ıve; were included. Patients with seizure disorder, brain metastasis were excluded. Olanzapine group received Tablet Olanzapine 10 mg on day 1 to 3. Aprepitant group received Tablet. Aprepitant 125mg on day 1, 80mg on days 2-3. The combination group received Aprepitant 125mg on day 1, 80mg on days 2-3 and Olanzapine 10mg on day 1. All groups received Palnosteron 0.25mg and Dexamethasone 8mg on day 1. The primary end point of the study was complete response (CR) for nausea that is no nausea in the acute, delayed and overall periods. Secondary endpoint was CR for vomiting and no use of rescue drugs in all periods. Beginning with the first day of chemotherapy and daily through day 5, patients were asked to record daily episodes of nausea using a visual analogue scale from 0 to 10, with 0 indicating no nausea and 10 indicating a maximal level of nausea. They were asked to record daily episodes of vomiting (number and time) and the utilization of rescue therapy. Results: A total of 141 patients were evaluated and consented for the study. The median age was 47 years; range 29-80; CR for nausea in the acute period (within 24 hours) was 83%, 63.8% and 78.7% (p=0.078); for the delayed period (days 2-5) 59.6%, 55.3% and 63.8% (p=0.702); for the overall period (0-120 hours) 57.4%, 53.2% and 59.6% (p=0.817) for the Olanzapine, Aprepitant and combination arm respectively. CR for vomiting in the acute period was 91.5%, 91.5% and 97.9.7% (p=0.344); for the delayed period 74.5%, 85.1% and 97.9% (p=0.005); for the overall period 70.2%, 85.1% and 97.9% (p=0.001) for the Olanzapine, Aprepitant and combination arm respectively. There were no Grade 3/4 toxicities. Conclusions: The combination strategy shows trend towards better prevention of CINV. Clinical trial information: CTRI/2017/12/010864.

11578 Poster Session (Board #270), Mon, 1:15 PM-4:15 PM

Single-fraction stereotactic versus standard conventional multifraction radiation for predominantly non-spine bone metastases: A randomized phase II trial.

Quynh Nguyen, Edward Chow, Stephen G. Chun, Zhongxing X. Liao, Rensi Fnu, James William Welsh, Stephen Hahn, Clifton David Fuller, Bryan Moon, Justin E. Bird, Robert L. Satcher, Patrick P. Lin, Melenda Jeter, Michael O’Reilly, Valerae O. Lewis; University of Texas MD Anderson Cancer Center, Houston, TX; Odette Cancer Centre, Sunnybrook Health Sciences Centre, University of Toronto, Toronto, ON, Canada; The University of Texas Southwestern Medical Center, Dallas, TX; M.D. Anderson Cancer Center, Houston, TX; The University of Texas MD Anderson Cancer Center, Houston, TX; Department of Orthopedic Oncology, University of Texas MD Anderson Cancer Center, Houston, TX; Department of Orthopedic Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX; Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX

Background: There lacks a consensus as to the optimal radiotherapy dose and fractionation schedule for treating bone metastases. We assessed the relative efficacy of single high-dose stereotactic radiation therapy (SBRT) versus standard multifraction radiation therapy (MFRT) for alleviation of pain in patients with mostly non-spine bone metastases. Methods: This prospective, randomized, single-institution phase II non-inferiority trial enrolled patients with radiologically confirmed painful bone metastases from September 2014 through June 2018. Patients were randomly assigned in a 1:1 ratio to receive either single-fraction SBRT (12 Gy for $4-cm lesions or 16 Gy for , 4-cm lesions) versus MFRT to 30 Gy in 10 fractions. Results: The primary endpoint was pain response, defined by international consensus criteria as a combination of pain score and analgesic use (daily morphine-equivalent dose [MED]). Failure of pain response was defined as worsening pain score ($2 points on a 0-to-10 scale), an increase in morphine- equivalent opioid dose of $50%; re-irradiation; or pathologic fracture. Among evaluable patients who received treatment per protocol, the single-fraction SBRT group had more pain responders (CR+PR) at 2 weeks (62% vs. 36% MFRT, P= 0.01), at 1 month (62% vs 36%, P= 0.01), at 3 months (72% vs. 49% MFRT, P= 0.03), and at 9 months (77% vs. 12% MFRT, P= 0.03). No differences were found in treatment-related toxicity or quality of life scores after SBRT versus MFRT; local control rates at 1 year were higher in patients receiving single-fraction SBRT. Conclusions: Delivering high-dose single-fraction SBRT is an effective, convenient treatment option for patients with painful bone metastases. Among evaluable patients, SBRT led to higher rates of pain response (CR+PR) than did MFRT and thus should be considered for patients expected to have relatively long survival. Clinical trial information: Clinical- Trials.gov ID NCT02163226.

11579 Poster Session (Board #271), Mon, 1:15 PM-4:15 PM

“If you don’t ask, you won’t know”: Do patient-reported outcome (PRO) instruments capture the symptom experience of patients treated with immune checkpoint inhibitors (ICIs)?

Brooke Peterson Gabster, Evan Thomas Hall, Surbhi Singhal, James Dickerson, Lidia Schapira; Stanford Hospital & Clinics, Stanford, CA; Stanford University School of Medicine, Stanford, CA; Stanford Hospital & Clinics, stanford, CA; Stanford Cancer Center, Palo Alto, CA

Background: PROs are increasingly used as a key tool in patient-focused treatment decisions. However, many cancer PROs were designed to capture the patient experience of those treated with chemother- apeutic agents. Less is known about the utility of PRO instruments in assessing symptoms from ICIs. Methods: We systematically searched the literature to identify peer-reviewed publications that reported PROs for patients receiving ICIs. We excluded case reports/series, narrative reviews, and publications without original data. We then selected the studies that compared ICIs to cytotoxic chemotherapy. Clinician-documented adverse events (AEs) occurring in at least 10% of patients in a study arm were extracted and examined for concordance with symptoms included in the PROs administered to study patients. Results: Of 1,450 identified studies, eight met criteria for inclusion. Seven assessed PROs with the European Organization for the Research and Treatment of Cancer Core Quality of Life Questionnaire (EORTC QLQ-C30), six used the EuroQOL EQ-5D, and one used the Lung Cancer Symptom Scale (LCSS). Across the studies, fatigue, nausea, vomiting, appetite loss, diarrhea, constipation, pruritus, rash, and pyrexia were among the most common clinician-documented AEs. Of these AEs, six are directly correlated with questions on the PRO instruments, and three (pruritus, rash, and pyrexia) are not. AEs with corresponding PRO questions were more common in chemotherapy patients. Pruritus, rash, and pyrexia – AEs without corresponding PRO questions – were more common in patients treated with ICIs (see Table). Conclusions: Existing PRO instruments do not specifically query important symptoms associated with ICIs, underscoring the need to revise the instruments to more appropriately reflect the toxicity profiles of novel agents.

# Studies Reporting ICI arms: Average % patients Chemotherapy arms: Average % patients Symptom Data with AE with AE Fatigue* 7 17% 27% Nausea* 7 10% 32% Vomiting* 6 3% 16% Appetite 7 7% 16% loss* Diarrhea* 7 11% 15% Constipation* 6 4% 12% Pruritus 6 14% 2% Rash 5 10% 5% Pyrexia 38%6%

* AE directly corresponds to a PRO instrument question

11580 Poster Session (Board #272), Mon, 1:15 PM-4:15 PM

Patient level factors associated with chronic opioid use in cancer patients.

Colleen Ann Cuthbert, Yuan Xu, Devon J. Boyne, Shiying Kong, Brenda R. Hemmelgarn, Winson Y. Cheung; University of Calgary, Calgary, AB, Canada; Alberta Health Services, Calgary, AB, Canada

Background: Opioid prescribing in oncology is increasingly scrutinized given public health concerns about chronic opioid use, misuse, and harms. We aimed to evaluate patient reported pain scores, mental health indicators, prior opioid use, and number of opioid prescribers as potential risk factors for chronic opioid use in a large Canadian province. Methods: This was a population-based cohort study using administrative health data of patients in Alberta, Canada, diagnosed between Jan 2016 and Jan 2017, and completed a prospective comprehensive symptom survey within +/- 60 days of diagnosis. Patients were divided into two groups: chronic opioid use (COU) (defined as continuous prescriptions for opioids for at least 90 days post diagnosis) and non-chronic opioid use (NCOU). Logistic regression models were used to evaluate factors associated with COU. Results: We included 694 patients. Most had breast (20%), colorectal (13%), and lung (33%) cancers. There were no differences in mean age (65 years) or gender (50% female) between the groups. In total, 32% had moderate to high pain scores at diagnosis. Of the 14% with COU, 79% were opioid na¨ıve at diagnosis. Those in the COU group were more often diagnosed with advanced stage of disease (66% vs 40%), had lung cancer (47%), and were opioid tolerant at diagnosis (defined as . 90 days of continuous opioids within 1 year prior to their diagnosis) (21% vs 3%). In comparison, 64% of COU versus 27% of NCOU had moderate to severe pain scores at diagnosis (p , 0.001). COU had significantly higher anxiety and depression scores at diagnosis versus NCOU (p = 0.004). Among patients with COU, morphine equivalent daily doses increased from 27.3 (pre-diagnosis) to 65.1 (post-diagnosis). Irre- spective of treatment type or stage, those who had moderate to high pain scores, were opioid tolerant at diagnosis, or had multiple prescribers were at greater risk for COU (see Table). Conclusions: Specific patient groups were at increased risk of COU and should be the focus of adaptive prescribing approaches to ensure that opioid use is appropriate.

Variable Adjusted OR (95%CI) P value Pain score Mild pain referent Moderate pain 1.7 (1.1-3.6) 0.03 Severe pain 2.6 (1.1-6.1) 0.03 Opioid use pre-diagnosis Naive referent Tolerant 3.3 (1.3-8.5) 0.01 Number of prescribers (continuous) 2.3 (1.8-2.9) , .001

11581 Poster Session (Board #273), Mon, 1:15 PM-4:15 PM

Comparison of the effectiveness of oral morphine versus oral tramadol on early pain control in opioid-naive patients with moderate cancer pain.

Ramila Shilpakar, Bishnu D. Paudel, Aarati Shah, Soniya Dulal; NAMS, Bir Hospital, Kathmandu, Nepal; National Academy of Medical Sciences, Kathmandu, Nepal; National Academy of Medical Sciences (NAMS), Bir hospital, Kathmandu, Nepal

Background: Adequate pain control, an essential part of cancer care remains a challenge in resource limited countries like Nepal. Early use of morphine for Moderate Cancer Pain (MCP) and not sequential World Health Organization (WHO) analgesic ladder seems reasonable in the setting of limited access to health care. The purpose of this study was to compare efficacy of oral Morphine (MOR) with oral Tramadol (TRM) in control of pain as well as physical wellbeing in patients (pts) with MCP using Edmonton Symptom Assessment Scale (ESAS). Methods: An IRB approved randomized phase II trial was performed in opioid-naive pts with MCP as defined by pain score in Numerical rating score (NRS)of 4-6.Patients were randomized to receive MOR syrup 5 mg 4 hourly or TRM 50 mg four times a day. Titration of dose was done in both groups for 3 days as per standard recommendation for MOR or till maximum recommended daily dose for TRM. MOR was changed to prolonged release form at Day 4.The primary endpoint was number of early responders, defined as pts with at least 20% reduction in pain intensity on NRS on Day 3. Secondary outcome was number of patients with highly meaningful pain reduction, defined as decrease in pain intensity on NRS by $ 5 and improvement in physical well-being with ESAS at Day 7. Results: 68 pts consented and were randomized, 34 in each arm. The primary endpoint occurred in 94.1% pts in MOR and 55.9% in TRM (p ,0.001). Number of patients with highly meaningful pain reduction was significantly higher in MOR than in TRM (76.5% vs. 32.35%; p,0.001). Improvement in general physical wellbeing as assessed by ESAS was better in morphine group. No difference in adverse effects was noted between the treatment arms. Conclusions: In this study Morphine was superior to Tramadol in the control of pain with statistically significant difference in the primary and secondary endpoints. So, early use of morphine skipping the WHO sequential analgesic ladder for moderate cancer pain seems a higher value option in resource scarce country with limited access to healthcare.

11582 Poster Session (Board #274), Mon, 1:15 PM-4:15 PM

Naldemedine for opioid-induced constipation in patients receiving palliative care: A real- world registry study (Phase-R OIC Study).

Masaki Shimizu, Takaomi Kessoku, Hiroto Ishiki, Tetsuya Matsuura, Yusuke Hiratsuka, Yoshinobu Matsuda, Takaaki Hasegawa, Kengo Imai, Isseki Maeda, Shunsuke Oyamada, Eriko Satomi; Department of Palliative Medicine,National Cancer Center Hospital, Tokyo, Japan; Department of palliative care center, Yokohama City University Hospital, Yokohama, Japan; Department of Palliative Medicine, Tohoku University School of Medicine, Sendai, Japan; National Hospital Organization Kinki-Chuo Chest Medical Center, Sakai, Japan; Division of Psycho-oncology and Palliative Care, Nagoya City University Hospital, Nagoya, Japan; Department of Palliative Medicine,Seirei Mikatahara General Hospital, Hamamatsu, Japan; Gratia Hospital, Minoh, Osaka, Japan; Department of Biostatistics, JORTC Data Center, Tokyo, Japan

Background: Although naldemedine, a new peripherally-acting mu-opioid receptor antagonist, was approved for the management of opioid-induced constipation (OIC), its safety and effectiveness in real world clinical practice is unknown. Methods: We conducted a real world registry study in 14 hospital palliative care teams and inpatient palliative care units in Japan between April and December 2018. Consecutive cancer patients who received naldemedine for OIC were enrolled in a 7-day observational study. All treatment and assessment procedures were performed according to the accepted clinical practice. The primary outcome of the study was the proportion of patients who experienced spontaneous bowel movement within 24 hours after the initial administration of naldemedine. Adverse events, which indicated a possible or stronger causal relationship with naldemedine treatment, were reported according to the Common Terminology Criteria for Adverse Events (CTCAE) ver 4.0. Results: Overall, 204 patients were enrolled in the study. The mean age of the patients was 63614 years, and 103 (50.5%) were male. The most common primary cancer site was lung (23.5%), followed by gastrointestinal (13.7%), and urological organs (9.3%). The proportion of patients undergoing active cancer treatment was 59.9%. Oxycodone was the most frequently used regular opioid (n = 115, 56.4%), and the median oral morphine-equivalent daily dose of opioids was 30 mg (interquartile range: 20-60 mg). Magnesium oxide (64.2%) and Senna (17.2%) were used as concomitant laxatives. All patients received 0.2 mg of naldemedine orally once daily. Most patients (90.2%) completed the 7-day observation. In 146 patients, spontaneous bowel movement was observed within 24 hour after the first administration of naldemedine (71.6%, 95% confidence interval 65.4-77.8%). Nearly two-thirds of the patients experienced increased frequency of spontaneous bowel movement in the week after naldemedine administration. The most prevalent adverse events were diarrhea (CTCAE grade 1-2, 35 cases; grade 3, 1 case) and abdominal pain (CTCAE grade 1-2, 10 cases; grade 3, 1 case). No serious adverse events including gastrointestinal perforation were reported. Conclusions: Naldemedine for opioid-induced constipation is safe and effective in the real world oncology and palliative care settings. Clinical trial information: UMIN000031381.

11583 Poster Session (Board #275), Mon, 1:15 PM-4:15 PM

Effect of patient education on palliative care knowledge and acceptability of outpatient palliative care service among gynecologic oncology patients.

Ashley Graul, Ashley Ford Haggerty, Carolyn Stickley, Pallavi Kumar, Knashawn Morales, Hillary Bogner, Robert Burger, Morgan Mark, Emily Ko; University of Pennsylvania, Philadelphia, PA; The University of Pennsylvania, Philadelphia, PA; University of Pennsylvania, Philadephia, PA

Background: This was a randomized control trial to estimate the effect of an interventional video on improving palliative care knowledge and acceptability of outpatient services in gynecologic oncology patients. Methods: Women receiving active treatment for gynecologic malignancy (persistent or progressive disease despite primary treatment) were recruited at an academic tertiary care center from 2/2018 to 1/2019 and randomized to: palliative care educational video or non-directive cancer center informational video (control). The primary outcome was desire for referral to palliative care. Function and knowledge were assessed using the Functional Assessment of Cancer Therapy (FACT-G) and the Palliative Care Knowledge Scales. Data analyses were performed using t-tests, Wilcoxon rank sum or Fisher’sexact tests with significance level of a=0.05. Results: 111 women were enrolled. Demographic characteristics were equally distributed between groups (mean age 63.4 vs 65.4 years; 78% vs 82% Caucasian, 58% vs 68% stage III, 71% vs 64% ovarian cancer, 65% vs 72% platinum-sensitive). There was no statistical difference in knowledge scores or in desire for referral to palliative care (29% vs 27%; p=0.79). Secondary analysis showed a statistically significant increase in utilization of palliative care services compared to historic institutional data (8.8% to 29.7%; p=,0.001). Further, those that desired referral had significantly worse FACT-G scores at time of referral choice (table). Conclusions: Use of a palliative care educational video did not increase knowledge or acceptability of palliative care services within this RCT. However, the rate of patients seeking palliative care referral tripled compared to historic rates. Further studies should investigate whether discussion regarding palliative care services alone may increase desire for referral, and if use of Fact-G scores may identify patients in greatest need of services.

Chose Referral Did Not Choose Referral n=31 n=80 Function and knowledge at time of referral choice (Median, IQR) (Median, IQR) p-value Physical Welling Being Sub-score 24.0 (18.0, 25.0) 19.0 (14.0, 23.0) 0.007 Social Well Being Sub-score 25.8 (23.9, 28.0) 24.0 (18.0, 26.0) 0.005 Emotional Well Being Sub-score 19.0 (15.5, 22.0) 17.0 (11.0, 20.0) 0.008 Functional Well Being Sub-score 23.0 (17.0, 26.0) 17.0 (14.0, 21.0) 0.002 FACT-G Total Score 89.0 (74.8, 97.0) 74.3 (63.0, 86.0) ,0.001 PaCKs Knowledge Score 12.0 (12.0, 13.0) 12.0 (12.0, 13.0) 0.98

11584 Poster Session (Board #276), Mon, 1:15 PM-4:15 PM

Cognitive function in chronic lymphocytic leukemia (CLL): Examining effects of disease, treatment, and inflammation.

AnnaLynn Williams, Edwin van Wijngaarden, Christopher Seplaki, Charles E. Heckler, Miriam Weber, Paul M. Barr, Clive Zent, Michelle Christine Janelsins; University of Rochester Medical Center, Rochester, NY; University of Rochester School of Medicine and Dentistry, Rochester, NY

Background: Cancer-related cognitive impairment (CRCI) is an important clinical problem that may occur through pro-inflammatory pathways, cancer-related pathology, or cancer treatment. However, their relative contribution to CRCI is unknown. We used CLL, an indolent cancer of the immune system, as a novel model to study differential roles of disease, treatment, and inflammation in CRCI. Methods: We assessed cognitive function and serum pro-inflammatory marker levels in 150 CLL patients (100 treatment na¨ıve and 50 chemotherapy treated) including 84 patients with high risk of CLL progression (deletion 17p, deletion 11q, unmutated IGHV, Zap70+, or CD38+). Objective neuropsychological tests assessed global cognition, processing speed, attention, memory, and executive function (NIH Toolbox Cognition, Hopkins Verbal Learning Test Revised (HVLT), Trail Making Test (TMTA/B)). Linear regression models examined cognitive outcomes in relation to risk, treatment, and pro-inflammatory markers while adjusting for sociodemographic covariates. Results: High risk patients recalled almost 2 fewer words on a memory task (HVLT immediate recall b = -1.8, 95%CI -3.3, -0.3) and took 15 seconds longer on an executive function task (TMTB b = 15.4, 95%CI 3.1, 27.6) than low risk patients, independent of treatment status. Treated patients reported significantly greater cognitive difficulties than treatment naive patients (FACT-Cog b = -5.2, 95%CI -8.7, -1.6) but did not perform worse on objective cognitive measures. Higher CRP was significantly associated with more self-reported cognitive difficulties, but higher levels of interleukin-6 and interferon-g were significantly associated with better performance on objective measures of global cognition, attention, and executive function. Conclusions: High risk patients experienced impairments in executive function and memory suggesting that disease biology contributes to CRCI independent of treatment. Increasing levels of pro-inflammatory cytokines were associated with better cognitive performance indicating that the role of inflammation with respect to CRCI in CLL may be complex, possibly due to tumor-related immune deficiencies.

11585 Poster Session (Board #277), Mon, 1:15 PM-4:15 PM

A community oncology palliative program: Early results for cost and quality measures within OCM program claims data.

Adil Jamal Akhtar, Jeffrey Margolis, Karma Maxwell, Andrew A. Muskovitz, Richard Philip Zekman, Karna Sheth, Samer Ballouz, Yusuf Qamruzzaman, Mohammed Ibrahim, Tammy Scott-Barney, Lexi Stortz, Maria Fabbiano; Michigan Health Professionals, Sterling Heights, MI; Michigan Health Professionals, Farmington Hills, MI; Michigan Health Professionals/Premiere Hospice, Sterling Heights, MI; Michigan Health Professionals, Royal Oak, MI; Integra Connect, West Palm Beach, FL; Michigan Health professionals, Royal Oak, MI; Michigan Health Professionals, Troy, MI; Premier Hospice, Troy, MI

Background: Oncology Care Model (OCM) is an initiative of the Centers for Medicare and Medicaid Innovation which aims to provide higher quality and more coordinated oncology care while lowering the cost. Oncology Division of Michigan Health Professionals (MHP) participates in OCM. Palliative and End of Life care was identified as one of the quality improvement areas. A community oncology Palliative care (PC) program was launched in October 2017. Methods: The multidisciplinary PC team was led by Board certified palliative care and hospice physicians. Patients appropriate for PC referral were identified by participating medical oncologists. Patients were contacted by the PC team. If the patients agreed a Nurse Practitioner (NP) would assess and follow the patients at home. Care was coordinated by the NP’sin communication with the palliative care team and the primary medical oncologists. Last 30-day (limited by the OCM episode or patient death) OCM program claims data was analyzed by IntegraConnect. Results: From October 2017 to October 2018 a total of 273 patients were referred to the PC program. Fifty-eight patients were identified as having OCM episodes, of these 36 patients had claims data through June 30, 2018. Twenty patients accepted and were engaged with PC,16 patients declined or were unable to reach for PC and formed the comparison group. Even when drug and office costs were excluded, PC engaged patients spent 17% less versus the comparison group (93k vs 112k) in last 30-day claims data. PC engaged patients had a lower acute care facility costs which accounted for 50% (46k) of reimbursement, compared with 95% (105k) for the comparison group. Fourteen OCM patients referred to Palliative program died within episode. 80% (8/10) of engaged patients met quality measure for OCM-3, at least 3 days in hospice vs. 0% (0/4) of patients who declined palliative care, before episode-death. Conclusions: Palliative engaged OCM patients experienced more care at their homes at a lower cost. Palliative program improved practice performance in OCM-3 quality measure. MHP Palliative program is reaching patients in OCM episodes but the numbers are still small.

11586 Poster Session (Board #278), Mon, 1:15 PM-4:15 PM

Preemptive versus reactive topical clobetasol for regorafenib-induced hand-foot reactions: Results from the ReDOS trial.

Aminah Jatoi, Fang-Shu Ou, Daniel H. Ahn, Patrick McKay Boland, Kristen Keon Ciombor, Nisha Lassi Jacobs, Boris Pasche, James M. Cleary, Jeannine S. McCune, Katrina Pedersen, Afsaneh Barzi, E. Gabriela Chiorean, Erica N. Heying, Heinz-Josef Lenz, Jeff A. Sloan, Mario E. Lacouture, Axel Grothey, Tanios S. Bekaii-Saab; Mayo Clinic, Rochester, MN; Ohio State University Arthur G. James Cancer Hospital and Richard J. Solove Research Institute, Columbus, OH; Roswell Park Cancer Institute, Buffalo, NY; Vanderbilt University Medical Center, Nashville, TN; Minnesota Oncology Hematology, Minneapolis, MN; Wake Forest Baptist Health, Winston-Salem, NC; Dana-Farber Cancer Institute, Boston, MA; University of Washington and Fred Hutchinson Cancer Research Center, Seattle, WA; Mayo School of Graduate Medical Education, Rochester, MN; USC Keck School of Medicine Norris Comprehensive Cancer Center, Los Angeles, CA; Mayo Clinic/Alliance Statistics and Data Center, Rochester, MN; University of Southern California, Los Angeles, CA; Memorial Sloan Kettering Cancer Center, New York, NY; West Cancer Center, Memphis, TN; Mayo Clinic, Phoenix, AZ

Background: Hand-foot skin reaction (HFSR) is the most common regorafenib adverse event. However, no intervention has demonstrated efficacy in preventing or palliating this adverse event, short of modulating the regorafenib dose. Methods: The Regorafenib Dose Optimization Study (ReDOS) was a randomized trial (https://clinicaltrials.gov/ct2/show/NCT02368886); here we describe an apriorisecondary analysis of ReDOS with the goal of assessing whether clobetasol 0.05% cream (a potent corticosteroid) applied to the palms and soles twice per day is more effective when prescribed pre-emptively (before the development of HFSR) versus reactively (after the development of HFSR). Patients were assessed during the 1st 2 cycles of regorafenib. Results: Among 116 evaluable patients, 61 received pre-emptive clobetasol, and 55 reactive clobetasol. Baseline demographics were comparable between groups. During the 1st regorafenib cycle, 46% and 52% of patients developed HFSR with pre-emptive and reactive clobetasol, respectively (p = 0.52). However, during the 2nd cycle, 56% (of 47 total patients) and 80% (of 41 total patients), developed this toxicity with pre-emptive and reactive clobetasol, respectively (p = 0.02). Of note, during the 2nd cycle, rates of grade 1,2, and 3 HFSR were 40%, 11%, and 4% (pre-emptive) and 45%, 25%, and 10% (reactive) (p = 0.07). A sensitivity analysis to examine rates of no HFSR over all 2 regorafenib cycles showed that no HFSR occurred in 33% (of 61 total patients) with pre-emptive clobetasol versus 15% (of 55 total patients) with reactive clobetasol (p = 0.02). Patient-reported outcomes showed HFSR compromised activities of daily living, including getting dressed, preparing meals, walking, and even driving, with seemingly worse descriptive outcomes in patients who received reactive therapy. No adverse events from clobetasol were reported. Conclusions: Compared to reactive therapy, pre-emptive topical clobetasol might lessen regorafenib-induced HFSR. Clinical trial information: NCT02368886.

11587 Poster Session (Board #279), Mon, 1:15 PM-4:15 PM

Correlates of distress for cancer patients: Results from multi-institution use of holistic patient-reported screening tool.

Christine B. Weldon, James I. Gerhart, Frank J. Penedo, Paramjeet Khosla, Betty Roggenkamp, Mary Pasquinelli, Joanna Martin, Teresa Lillis, Shelly S. Lo, Lawrence Eric Feldman, Catherine Deamant, Rosa Berardi, Harry Miranda, Carol Newsom, Anne Bowman, Julia Rachel Trosman; Northwestern University Feinberg School of Medicine, Chicago, IL; Central Michigan University, Mount Pleasant, ME; Mount Sinai Hospital, Chicago, IL; Center for Business Models in Healthcare, Chicago, IL; University of Illinois Hospital and Health Sciences System, Chicago, IL; Jesse Brown VA Medical Center, Chicago, IL; Rush University Medical Center, Chicago, IL; Loyola University Medical Center, Maywood, IL; University of Illinois at Chicago Cancer Center, Chicago, IL; Cook County Hospital and Health System, Chicago, IL; Coleman Foundation, Chicago, IL; John H. Stroger, Jr. Hospital of Cook County, Chicago, IL; Mercy Hospital and Medical Center, Chicago, IL; UnityPoint Health Methodist, Peoria, IL

Background: The Commission on Cancer (CoC) Standard 3.2 requires distress screening and indicated action for cancer patients. NCCN and ASCO supportive care and age-related guidelines include patient reported concerns beyond distress. This study compares PHQ4 scores to other patient reported concerns. Methods: The Coleman Supportive Oncology Collaborative aggregated “best of” screening tools to assess patient reported needs and concerns aligned with CoC, NCCN and ASCO guidance. This supportive care screening tool was implemented at 8 sites from July 2015 thru July 2018. Analysis used chi squared test. Results: Most patients, 86% (10,635/12,295), reported one plus concerns and/or above threshold scores on PHQ4, PROMIS Pain, Fatigue or Physical Function. A chi squared comparison of patients with at least mild distress on PHQ4 to patients with no distress resulted in p values , .0001 for every screening category. Conclusions: Patients with a PHQ4 distress score of mild, moderate or severe also reported statistically significant levels of practical, family, physical, nutrition and treatment concerns. These patients also scored threshold levels for PROMIS Pain, Fatigue, and Physical Function. Screening only for distress without screening for other patient concerns may direct patients to services that do not address or focus on the underlying cause of the distress.

Endorsed concern Mild+ Moderate+ Severe+ or threshold score Standard PHQ4 vs PHQ4 vs PHQ4 vs Screening Item (n = 12,295) Rate Average Deviation lower lower lower Patient Health 1.89 2.88 Questionnaire 4 (PHQ4) Mild +, score of 3.0+ 3564 29% Moderate +, 6.0+ 1430 12% Severe +, 9.0+ 516 4% Practical Concerns 3168 26% 0.42 0.86 , 0.0001 , 0.0001 , 0.0001 Family Concerns 3115 25% 0.42 0.85 , 0.0001 , 0.0001 , 0.0001 Physical Concerns 7573 62% 1.98 2.42 , 0.0001 , 0.0001 , 0.0001 2+ Physical 5475 45% , 0.0001 , 0.0001 , 0.0001 Concerns Nutrition Concerns 3961 32% 0.88 0.85 , 0.0001 , 0.0001 , 0.0001 PROMIS Pain, 5+ 4284 35% 5.37 3.45 , 0.0001 , 0.0001 , 0.0001 PROMIS Fatigue, 3683 30% 8.85 4.60 , 0.0001 , 0.0001 , 0.0001 10+ PROMIS Physical 1800 15% 19.52 6.35 , 0.0001 , 0.0001 , 0.0001 Function, 15- Treatment Concerns 5509 45% 1.19 1.51 , 0.0001 , 0.0001 , 0.0001

11588 Poster Session (Board #280), Mon, 1:15 PM-4:15 PM

Sex differences in adverse event reporting in SWOG chemotherapy, biologic/immunotherapy, and targeted agent cancer clinical trials.

Joseph M. Unger, Riha Vaidya, Kathy S. Albain, Michael Leo LeBlanc, Lori M. Minasian, Carolyn Gotay, Norah Lynn Henry, Michael Jordan Fisch, Scott David Ramsey, Charles David Blanke, Gary H. Lyman, Dawn L. Hershman; Fred Hutchinson Cancer Research Center, Seattle, WA; Loyola University Chicago Stritch School of Medicine, Cardinal Bernardin Cancer Center, Maywood, IL; Southwest Oncology Group Statistical Center, Seattle, WA; National Cancer Institute, Rockville, MD; University of British Columbia, School of Population and Public Health, Vancouver, BC, Canada; University of Utah, Salt Lake City, UT; AIM Specialty Health, Chicago, IL; Oregon Health and Science University, Portland, OR; Columbia University Medical Center, New York, NY

Background: Women have more adverse events (AEs) from chemotherapy than men, but few studies have explored sex differences in biologic/immunotherapies (BIs) or targeted therapies. We examined subjective (symptomatic) and objective AEs by sex across different treatments. Methods: We analyzed drug-related severe (grade 3) or worse AEs by sex in SWOG phase II and III clinical trials conducted between 1980- 2018, excluding sex-specific cancers. AE codes and grade were categorized using the Common Termi- nology Criteria for Adverse Events (CTCAE). Subjective or symptomatic toxicities were defined as those aligned with the NCI’s new Patient-Reported Outcome (PRO) CTCAE; lab-based or physician-determined AEs were designated as objective. Multivariable logistic regression was used, adjusting for age, race, and disease prognosis. Thirteen symptomatic and 19 objective AE categories were examined. Results: In total, 36,397 patients (women, 13,907 [38.2%]; men, 22,490 [61.8%]) experiencing 522,835 AEs on 297 trials with 385 treatment arms were analyzed. Overall, 29.1% (n = 10.860) had severe or worse toxicity. Women experienced an increased risk of severe symptomatic AEs for BIs (OR = 1.53, 95% CI: 1.32-1.78, p , .0001), chemotherapy (OR = 1.31, 95% CI: 1.24-1.39, p , .0001), and targeted therapies (OR = 1.23, 95% CI: 1.06-1.43, p = .008). Women also had an increased risk of severe objective AEs for BIs (OR = 1.53, 95% CI: 1.32-1.78, p , .0001), chemotherapy (OR = 1.35, 95% CI: 1.28-1.43, p , .0001), but not targeted therapies (OR = 1.08, 95% CI: 0.94-1.25, p = .28). Across all treatments, sex differences were greater for hematologic (OR = 1.29, 95% CI: 1.24-1.35, p , .0001) v. non-hematologic (OR = 1.13, 95% CI: 1.08-1.18, p , .0001) objective AEs. Conclusions: The greater severity of both symptomatic and objective – especially hematologic – AEs in women across multiple treatment paradigms indicates broad- based sex-differences exist. This could be due to AE reporting, pharmacogenomics of drug metabolism and disposition, total dose received, and/or adherence to therapy. Particularly large sex differences were observed for patients receiving BIs, suggesting studying AEs from these agents is a priority.

11589 Poster Session (Board #281), Mon, 1:15 PM-4:15 PM

Risk factors for opioid abuse/dependence in hospitalized cancer patients in the United States.

Veli Bakalov, Amy Tang, Amulya Yellala, Laila Babar, Rupin Shah, Santhosh K. Sadashiv, Robert B. Kaplan, John Lister, Elizabeth Cuevas, Dulabh K. Monga; Allegheny Health Network, Department of Internal Medicine, Pittsburgh, PA; Biostatistics Department, Henry Ford Health System, Detroit, MI; Department of Medicine, Allegheny Health Network, Pittsburgh, PA; The Esophageal and Lung Institute, Allegheny Health Network, Pittsburgh, PA; Allegheny General Hospital, Pittsburgh, PA; Division of Hematology and Cellular Therapy, Allegheny Health Network, Pittsburgh, PA; Western Pennsylvania Cancer Inst, Pittsburgh, PA; Allegheny Health Network, Pittsburgh, PA; Medical Oncol- ogy, Allegheny Health Network, Pittsburgh, PA

Background: Opioid medications are the mainstay for treating cancer pain. Goal of this study was to identify risk factors for opioid abuse/dependence in patients hospitalized with cancer, explore whether risk of opioid abuse/dependence varies by cancer type and to assess whether opioid abuse/dependence in cancer patients effects the outcomes of hospitalization. Methods: The Nationwide Inpatient Sample for the years of 2011-2015 was queried for the analysis. We used ICD-9-CM codes of solid tumors as a primary diagnosis for hospitalization, and opioid abuse/dependence as a secondary diagnosis of the hospitalization. We performed univariate and multivariate logistic regression analyses to examine the association between risk factors and opioid abuse/dependence. Data were analyzed using SAS v9.4 (SAS Institute, Cary, NC). Results: Total of 524,624 patients were included in our cohort. Rate of opioid abuse/dependence was highest in patients with liver cancer (1.77%). Opioid abuse/dependence was less associated with age (.65 years old: OR 0.29, 95% CI 0.21-0.39). Patients with Medicaid insurance associated with increased risk of opioid abuse/dependence comparing to other insurances (OR 5.29, 95% CI 4.78-5.86). Strongest association with opioid abuse/dependence were in patients with liver cancer (OR 6.07, 95% CI 5.11-7.20) followed by head and neck cancer (OR 3.20, 95% CI 2.67-3.84). Substance abuse (OR 9.9, 95% CI 9.04- 10.84), mental disease (OR-2.87, 95% CI 2.64-3.13) and nutrition deficiency (OR-2.09, 95% CI 1.90- 2.31) were highly associated with opioid abuse dependence. Inhospital mortality rate, total cost of hospitalization, and length of stay were significantlyhigherinpatientswithopioidabuse/dependence (Table). Conclusions: We identified risk factors for opioid abuse/dependence in hospitalized patients with cancer and demonstrated that risk of opioid abuse varies by cancer type, and opioid abuse/dependence affects the outcomes of hospitalization. Findings of our study can be used for development of the screening tools with higher sensitivity and specificity for predicting the risk of opioid abuse/dependence in cancer patients.

Outcomes of hospitalization. Without opioid abuse/dependency Opioid cases (n=524509) (n=2531) p-value Inhospital mortality rate (%) 4.67 (n=24343) 5.53 (n=140) 0.046 Cost of hospitalization ($) 44,429 (26-537-74,561) 51,187(28,513- ,0.001 92,976) Length of stay (d) 3.3 (1.4-6.6) 5.5 (2.8-9.6) ,0.001

11590 Poster Session (Board #282), Mon, 1:15 PM-4:15 PM

The effect of structured exercise during chemotherapy on chemotherapy-induced peripheral neuropathy (CIPN): A role for interoceptive brain circuitry.

Ian Kleckner, Jennifer S. Gewandter, Charles E. Heckler, Susan Staples, Ann Colasurdo, Po-Ju Lin, Michelle Shayne, Alissa Huston, Allison Magnuson, Mohamedtaki Abdulaziz Tejani, Richard Francis Dunne, Chunkit Fung, Brea Lipe, Frank Charles Passero, Nimish Mohile, Gary R. Morrow, Michelle Christine Janelsins, Karen Michelle Mustian; University of Rochester Medical Center, Rochester, NY; Univ of Rochester, Rochester, NY; University of Rochester James P. Wilmot Cancer Institute, Strong Memorial Hospital, Rochester, NY; University of Rochester, Rochester, NY; Tufts Medcl Ctr, Boston, MA

Background: Over half of patients receiving “neurotoxic” taxane, platinum, or bortezomib chemotherapy experience CIPN—a dose-limiting toxicity involving numbness and pain in the extremities. There are no FDA-approved drugs for CIPN, but exercise may help. This randomized pilot study explored whether structured exercise during chemotherapy ameliorates CIPN symptoms and whether improvements involve changes in the brain’s sensory processing (interoceptive) circuitry. Methods: Nineteen patients scheduled to receive taxane, platinum, or bortezomib were randomized to exercise (home-based, low-moderate intensity, walking and resistance training; EXCAP) or nutrition education (control) for 12 weeks starting at their first infusion. At 0, 6, and 12 weeks, we assessed CIPN symptoms using the CIPN-20 questionnaire (sensory scale, ranges 9-36, higher is worse) and a finger tactile sensitivity task. We assessed resting functional connectivity between the insula and thalamus via fMRI at 0 and 12 weeks. We used linear regression to model each outcome, tested for an effect of exercise, and controlled for baseline value and age because controls were older. Given the pilot nature of this study we present effect sizes, not p-values. Results: The 19 patients were 65611 years old, 52% women, with cancer: 42% breast, 32% gastrointestinal, 16% myeloma, and 10% genitourinary. Exercise mitigated CIPN symptoms per the CIPN-20 sensory scale. At 6 weeks, exercisers increased from 11.0 to 12.5 whereas controls increased from 11.0 to 15.5 (+1.5 vs. +4.5; ES = 0.81). At 12 weeks, exercisers increased from 11.0 to 14.8 whereas controls increased from 11.0 to 16.2 (+3.8 vs. +5.2, ES = 0.46). The finger-touching test corroborated results at 6 and 12 weeks (ES = 1.03 and 0.07). Exercisers showed better (reduced) insula- thalamus connectivity vs. controls (ES = 0.41). Reductions in connectivity were correlated with smaller increases in CIPN symptoms (r = 0.74). Conclusions: Exercise during neurotoxic chemotherapy mitigated CIPN symptoms, perhaps via improvements in interoceptive brain circuitry. Future work should test for replication with a larger sample. Clinical trial information: NCT03021174.

11591 Poster Session (Board #283), Mon, 1:15 PM-4:15 PM

Clinical impact of neutropenia and febrile neutropenia in mCRC pts treated with FOLFOXIRI/ bevacizumab (bev): A pooled analysis of TRIBE and TRIBE2 studies.

Daniele Rossini, Andrea Sbrana, Francesca Bergamo, Chiara Manai, Daniele Santini, Michele Ghidini, Carlotta Antoniotti, Roberto Moretto, Federica Marmorino, Federica Urbano, Monica Ronzoni, Silvia Noventa, Giovanni Randon, Chiara Carlomagno, Tiziana Pia Latiano, Stefano Sergio Cordio, Cristina Granetto, Chiara Cremolini, Alfredo Falcone, Andrea Antonuzzo; Department of Translational Research and New Technologies in Medicine and Surgery, Unit of Medical Oncology 2, Azienda Ospedaliera Universitaria Pisana, Pisa, Italy; Department of Clinical and Experimental Oncology, Medical Oncology 1 Unit, Istituto Oncologico Veneto IOV-IRCCS, Padua, Italy; Department of Medical Oncology, Campus Bio-Medico - University of Rome, Rome, Italy; Oncology Unit, ASST Cremona, Cremona, Italy; Department of Radiological Science, Oncology and Patology, Policlinico Umberto I, "Sapienza" University of Rome, Rome, Italy; Department of Oncology, Istituto Scientifico San Raffaele IRCSS, Milan, Italy; Medical Oncology Unit, Casa di Cura Poliambulanza, Brescia, Italy; Medical Oncology Department, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy; Department of Clinical Medicine and Surgery, University Federico II, Naples, Italy; Oncology Unit, Fondazione IRCCS Casa Sollievo della Sofferenza, San Giovanni Rotondo (FG), Italy; Oncologia Medica, Azienda Ospe- daliera ARNAS Garibaldi, Catania, Italy; Oncologia Medica, Azienda Sanitaria Ospedaliera Santa Croce e Carle, Cuneo, Italy; Unit of Medical Oncology 1, Azienda Ospedaliera Universitaria Pisana, Pisa, Italy

Background: FOLFOXIRI/bev is a valid option as first-line therapy for unresectable mCRC. TRIBE and TRIBE2 trials reported better activity and efficacy of the triplet/bev when compared with doublets/bev at the price of a higher incidence of chemo-related toxicities, including neutropenia (N). Here we aim at providing a detailed description of this adverse event, including the occurrence of febrile neutropenia (FN) and the use of granulocyte-colony stimulating factors (G-CSFs), in order to estimate the clinical relevance of N during FOLFOXIRI/bev. Methods: Safety data of 1175 pts enrolled in the TRIBE and TRIBE2 studies were reviewed. The incidence of N, the incidence and severity of FN, and the use of G-CSF in the triplet/bev and in the doublets/bev arms were compared using the Chi-square or the Fisher exact test as appropriate. Results: Out of 1175 pts included in the final analysis, 586 (49.8%) were treated with FOLFOXIRI/bev. Five pts (0.8%) in the doublets/bev arms and 29 (4.9%) in the triplet/bev arms received a primary prophylaxis with G-CSF. Among other pts, 118 (20.2%) in the doublets/bev arms and 276 (49.9%) in the triplet/bev arms experienced $ G3 N (p , 0.001). FN occurred in 25 (4.3%) and 41 (7.4%) cases respectively (p=0.041). Out of 78 FN episodes, 4 (13.3%) out of 30 in the doublets/bev arms and 13 out of 48 (27.1%) in the triplet/bev arms were associated with a poor MASCC score (,21) (p=0.17). G-CSF was used in 1069 (10.8%) cycles, 270 (5.3%) in doublets/bev and 799 (16.6%) in triplet/bev arms. In both arms, the majority of N and FN episodes were observed in the first two months (318 $ G3 N episodes out of 675 (47.1%), and 54 FN episodes out of 78 (69.2%)). Conclusions: FOLFOXIRI/bev was associated with a higher risk of N and FN than doublets/bev. However, the risk of FN was lower than 10%, thus not requiring a systematic use of primary G-CSF prophylaxis. The majority of FN episodes was associated with a good MASCC score, thus having a limited clinical impact. The vast majority of FN episodes occurred in the first two months of treatment, suggesting a closer monitoring of this adverse event during the first courses of therapy.

11592 Poster Session (Board #284), Mon, 1:15 PM-4:15 PM

Supportive care medications (SCMs) and pharmacogenomics (PGx) relevance in 6,985 cancer patients (pts) undergoing distress screening.

Jai Narendra Patel, Danielle Boselli, Elizabeth Jandrisevits, Issam Hamadeh, Ahmed Salem, Patrick Leland Meadors; Levine Cancer Institute, Charlotte, NC; Levine Cancer Institute, Atrium Health, Charlotte, NC

Background: SCMs are prescribed based on symptom burden, but response is variable, possibly due to PGx. We investigated the association between symptom burden, SCM prescribing, and frequency of SCMs with PGx evidence. Methods: Cancer pts $ 18 years old and completing electronic distress screening within 90 days of intake between 1/1/2017-12/31/2017 were included. Anxiety was measured using Generalized Anxiety Disorder 2-item (0-6) and depression using Patient Health Questionnaire-2 (0-6). Fatigue, nausea, neuropathy, pain and sleep were measured on a 0-10 scale. SCM prescribing within 90 days of intake was documented. Logistic regression compared symptom scores and SCM prescribing. Receiver Operating Characteristics analysis estimated sensitivity/specificity. Optimal symptom thresh- olds were selected according to Youden’s J statistic. SCMs with PGx evidence level A or B (according to Clinical Pharmacogenetics Implementation Consortium) were summarized. Results: Of 6985 pts, 65% were female, 75% Caucasian, 20% African American and median age was 60. 49% reported $ 1severe symptom, which correlated with SCM prescribing (p , 0.001). 3208 (46%) were prescribed SCM(s), mainly for pain (69%) or nausea (46%). Of these, 2759 (86%) received $ 1 SCM with PGx evidence and 2695 (84%) received a SCM metabolized by CYP2D6 - hydrocodone (47%), ondansetron (41%), and oxycodone (28%). Based on reported CYP2D6 allele frequencies conferring altered metabolism (~20%), 539 of the 2695 pts may have altered drug response. Threshold scores for each symptom are summarized in the table. Fatigue and nausea were not associated with SCM prescribing. Conclusions: Symptom burden is high in cancer pts and correlates with SCM prescribing. Many SCMs have PGx evidence, suggesting preemptive testing, particularly for CYP2D6, may have broad applicability in this population.

Threshold Sensitivity Specificity % of SCMs with PGx Symptom (N) (%) (%) OR [95% CI]a evidence Pain $ 7 (1380) 26 83 1.74 [1.54, 97 1.97] Anxiety $ 2 (2705) 59 63 2.41 [2.00, 0 2.90] Depression $ 2 (2180) 53 70 2.63 [2.17, 63 3.20] Sleep $ 4 (3276) 74 53 3.28 [1.99, 0 difficulty 5.40] Neuropathy $ 1 (2611) 52 63 1.90 [1.48, 10 2.43] ap , 0.001

11594 Poster Session (Board #286), Mon, 1:15 PM-4:15 PM

Safety of intravenous (IV) NEPA and oral NEPA for prevention of CINV in patients (pts) with breast cancer (BC) receiving anthracycline/cyclophosphamide (AC) chemotherapy (CT).

Lee S. Schwartzberg, Daniel Voisin, Giada Rizzi, Kamal Patel, Matti S. Aapro; West Cancer Center, Memphis, TN; Helsinn Healthcare SA, Lugano, Switzerland; CARTI Cancer Center, Little Rock, AR; Clinique de Genolier, Genolier, Switzerland

Background: NEPA, a combination antiemetic agent [NK1 receptor antagonist (RA) netupitant (oral) or fosnetupitant (IV) + 5-HT3RA palonosetron] offers 5-day CINV prevention with a single-dose. Unlike other IV NK1RAs, the fosnetupitant solution does not require a surfactant, emulsifier, or solubility enhancer and contains no allergenic excipients. In a Phase 3 study in pts receiving cisplatin-based CT, there were no infusion site or anaphylactic reactions related to IV NEPA. In contrast, hypersensitivity reactions and anaphylaxis have been reported with IV aprepitant, fosaprepitant and rolapitant, with the highest rate (35%) for fosaprepitant in the AC setting. This study (NCT03403712) evaluates the safety of IV NEPA in BC pts receiving repeat cycles of AC CT. Methods: This was a Phase 3b, double-blind study in females with BC na¨ıve to highly/moderately emetogenic CT. Pts were randomized 1:1 to receive a single 30-min infusion of IV NEPA or a single oral NEPA capsule on Day 1, prior to AC. Oral dexamethasone was also given to all patients before CT. The primary objective was a safety evaluation of IV NEPA based primarily on treatment-emergent adverse events (TEAEs). No formal between groups statistical comparison was planned. Results: 402 pts were treated with IV NEPA or oral NEPA and included in the safety population. The AE profiles were similar for the two groups; cycle 1 results are reported (Table). Comparable complete response (no emesis, no rescue) rates were seen during the cycle 1 overall phase (0-120h) (73.0% IV NEPA, 77.2% oral NEPA). Conclusions: There were no infusion-site AEs related to IV NEPA and no anaphylaxis reported for either formulation. Consistent with the pivotal study, IV NEPA is safe and effective in pts receiving AC. As a simplified single-dose formulation, IV NEPA may be better tolerated than other NK1 RAs. Clinical trial information: NCT03403712.

IV NEPA Oral NEPA n (%) pts with at least one (N = 200) (N = 202) TEAE 121 (60.5) 122 (60.4) Severe TEAEs 11 (5.5) 10 (5.0) Serious TEAEs 2 (1.0) 1 (0.5) Treatment-related (TR) TEAE 13 (6.5) 12 (5.9) TR infusion-site TEAE 00 TR TEAE leading to discontinuation or death 00 Most common (‡2%) TR TEAEs Headache 5 (2.5) 3 (1.5) Dizziness 4 (2.0) 1 (0.5)

11595 Poster Session (Board #287), Mon, 1:15 PM-4:15 PM

Physician concordance with update to ASCO guidelines for antiemetic use with carboplatin AUC ‡ 4.

Rudolph M. Navari, Kathryn Jean Ruddy, Thomas William LeBlanc, Rebecca Anne Clark-Snow, Rita S. Wickham, Gary Binder, Tammy Bratton Coberly, Ravi C. Potluri, Luke M. Schmerold, Eric Roeland; University of Alabama at Birmingham, Birmingham, AL; Mayo Clinic, Department of Oncology, Rochester, MN; Duke University Medical Center, Durham, NC; Oncology Consultant, KS; Rush University College of Nursing, Chicago, IL; Helsinn Therapeutics, Inc., Iselin, NJ; SmartAnalyst Inc., New York, NY; Massachusetts General Hospital, Boston, MA

Background: In 2017, NCCN (2/2017) and ASCO (8/2017) each amended antiemetic guidelines to classify carboplatin AUC $4 as highly emetogenic chemotherapy (HEC), recommending upfront triple prophylaxis with an NK1 receptor antagonist (RA) + 5HT3 RA + dexamethasone. Physician concordance with the new recommendations, and the consequences for avoidable post-chemotherapy acute care, merit study. Methods: In a large electronic health record database (IBM Explorys), we identified carboplatin courses of therapy ($14-day cycles as a proxy for AUC $4) and courses with $7-day cycles of other HEC and non-HEC therapy from 4Q 2012 through August 2018. Guideline concordance, defined as triple prophylaxis at HEC initiation, was evaluated. We also assessed 30-day post-chemotherapy acute care (inpatient or emergency department) associated with nausea or vomiting (NV) or eight other toxicities deemed avoidable in the US Centers for Medicare & Medicaid’s new oncology outcome measure OP-35. Results: 11,554 carboplatin courses were identified. Before the guideline change, rates of upfront triple prophylaxis grew from 14% in 2013 to 16% in mid-2017. Rates then rose to 26% by 1Q 2018 before dropping to 21% by 3Q 2018; quarterly rates averaged 20% (range 15%-26%) following the guideline change. In 31% of carboplatin courses we noted 30-day acute care use, of which 75% involved $1ofthe ten OP-35 toxicities. NV (with or without acute care use) was reported in 24% of courses, and 27% of total OP-35 acute care events involved NV. Rates for NV, and for OP-35-related and NV-related acute care after carboplatin, were similar to rates after other HEC chemotherapy, and higher than rates after other non- HEC IV chemotherapy or oral HEC/MEC agents. Conclusions: Use of upfront triple antiemetic prophylaxis has increased only marginally for carboplatin AUC $ 4 since its 2017 re-classification as HEC in national guidelines, perhaps due to low awareness of the change. Patients receiving carboplatin had similar rates of NV and related 30-day acute care events as seen for other HEC, confirming that the new HEC definition fits clinical experience. More triple prophylaxis use is needed to reduce NV and NV-related avoidable acute care seen with carboplatin AUC $ 4.

11596 Poster Session (Board #288), Mon, 1:15 PM-4:15 PM

Survival and safety of monosialotetrahexosylganglioside in GI cancer patients with oxaliplatin-induced peripheral neurotoxicity-result from TJMUCH-GI-001 trial.

Zhou Likun, Dingzhi Huang, Rui Liu, Hongli Li, Tao Ning, Le Zhang, Shaohua Ge, Ming Bai, Xia Wang, Rubing Han, Yuchong Yang, XinYi Wang, Xinyun Chen, Zhiying Gao, Kaijun Niu, Laizhi Luo, Xi Wu, Ting Deng, Yuanquan Yang, Yi Ba; Tianjin Medical University Cancer Institute and Hospital, National Clinical Research Center for Cancer Tianjin’s Clinical Research Center for Cancer, Key Laboratory of Cancer Prevention and Therapy Tianjin Medical University, Tianjin, China; Tianjin Medical University Cancer Institute and Hospital, Tianjin, China; Tianjin Medical University Cancer Institute and Hospital, National Clinical Research Center for Cancer, Key Laboratory of Cancer Prevention and Therapy, Tianjin, China; Tianjin Cancer Hospital, Tianjin, China; Tianjin Medical University Cancer Institute and Hospital, National Clinical Research Center for Cancer, Key Laboratory of Cancer Prevention and Therapy, Tianjing, China; Tianjin Medical University Tianjin Cancer Hospital, Tianjin, China; Qiqihar Medical University Research Institute of Medicine and Pharmacy, Qiqihar, China; Tianjin Medical University, Tianjin, China; Capital Medical University, Tianjin, China; Cancer Hospital Chinese Academy of Medical Sciences, Beijing, China; Roswell Park Comprehensive Cancer Center, Buffalo, NY; Department of Medical Oncology, Tianjin Cancer Hospital, Tianjin, China

Background: TJMUCH-GI-001 Trial was a randomized, double-blind, placebo-controlled phase III trial to study the efficacy of Monosialotetrahexosylganglioside (GM1) for oxaliplatin-induced peripheral neurotoxicity(OIPN)inGIcancerpatients. Majority patients (. 80%) in both arms continued receiving oxaliplatin on the trial. The results showed GM1 effectively reduced OIPN in GI cancer patients. Here we report the survival and safety results of this trial. Methods: Patients were randomized in a 1:1 ratio to receive GM1 or placebo. Patients with OIPN . = G2 by CTCAE 4.03 persisting during or after oxaliplatin-based chemotherapy were eligible. The patients who remained on oxaliplatin after enrollment, received concurrent placebo or GM1 x 7 days with each chemotherapy cycle. The patients who stopped taking oxaliplatin, were treated with placebo or GM1 x 14 days every 3 weeks. GM1 was dosed at 60mg daily for every 3-week or 40mg daily for every 2-week schedule. Trial was continued until modified EORTC QLQ-CIPN20 ( MCIPN） increased by 30% or stayed unchanged after two more treatments beyond completion of oxaliplatin. Survival data for the treatment arms were compared using a log-rank test and Chi-square tests were used for safety analysis. Results: From May 2015 to Dec 2017, 73 patients were enrolled in GM1 and 72 in placebo arm. The median follow-up was 16.6 months (0.8-43.1 months) as of Dec.2018. Four patients lost to follow up. There was no deleterious impact of GM1 on survival. As a matter of fact, receiving GM1 was associated with a trend toward improved PFS and OS (HR=0.74,95%CI, 0.469 - 1.156 for PFS and HR=0.76, 95% CI0.469 - 1.156 for OS). The most frequent Grade 3 or 4 adverse events included neutropenia (8 patients in GM1 group VS. 4 in placebo group) and hypoleukemia (4 patients in GM1 group VS. 1 in placebo group). Other 3 or 4 adverse events (all less than 3 patients) included anorexia, hypercalcemia, nausea, vomiting, proteinuria, hyperbilirubinemia, hypokalemia, hypertension and appendicitis. All the 3 or 4 adverse events were related to chemotherapy, not to GM1. Conclusion: In this placebo-controlled phase III trial, GM1 showed acceptable toxicity with trends favorable PFS and OS in GI cancer patients. Clinical trial information: NCT02486198.

11597 Poster Session (Board #289), Mon, 1:15 PM-4:15 PM

Validation of a new prognostic body composition parameter in cancer patients.

Paolo Pedrazzoli, Riccardo Caccialanza, Nicole Stobaus, ¨ Annalisa Turri, Catherine Klersy, Marilisa Caraccia, Silvia Brugnatelli, Simona Secondino, Daniela Cicognini, Ilaria Imarisio, Andrea Riccardo Filippi, Teresa Monaco, Atonello Giannoni, Kristina Norman, Emanuele Cereda; Fondazione IRCCS Policlinico San Matteo, Pavia, Italy; Charite ´ University Hospital, Berlin, Germany; Civil Hospital, Carrara, Italy

Background: In cancer patients protein-calorie imbalances are responsible for decreased lean body mass and, in turn, for worse clinical outcome. We evaluated the prognostic value of a new body composition parameter (creatinine height index [CHI]) obtained from bioimpedance vectorial analysis-derived body cell mass and its association with nutritional and functional status. Methods: Data from Italian and German cancer patients based on information from previous prospective cohort studies were used. Multivariable models (adjusted for age, gender, hydration status, performance status, and disease stage) were built in both cohorts to assess the association between body composition outcome parameters (low fat-free mass [FFM], ,15 [females] and ,17 [males] kg/m2; low standardized phase angle [SPA], ,-1.65; low CHI, ,510 [females] and ,660 [males] mg/24h/m) and 1-year all-cause mortality, low body mass index (BMI; ,20 [,70 years] and ,22 [./=70 years] kg/m2), clinically significant weight loss (WL; ./=10% in 6 months) and low handgrip strength (HG; ,20 [females] and ,30 [males] kg). Results: Overall, 1084 cancer patients were included (Italians, N=454; Germans, N=630). Low CHI was independently associated with mortality in both Italian and German cohorts (Table). Low FFMI and low SPA did not predict survival in the German cohort. In patients with low CHI, worse nutritional and functional status were observed in both study populations. Performance of models addressing the study endpoints showed substantial consistency with both cohorts, particularly of those including low CHI. Conclusions: We validated a new prognostic body composition parameter, which is easier to interpret than standard nutritional parameters and may be useful for identifying cancer patients at nutritional risk, requiring early nutritional support.

Risk estimates and performance of multivariable models addressing the association between body composition parameters and 1-year survival. Predictor Cohort HR (95%CI) P-value Harrell’sc Low FFMI Italian 2.08 (1.41-3.07) ,0.001 0.718 (TPI, 0.696) German 1.16 (0.79-1.72) 0.448 0.688 (TPI, 0.659) Low SPA Italian 3.07 (1.72-5.47) ,0.001 0.724 (TPI, 0.691) German 1.33 (0.95-1.86) 0.098 0.693 (TPI, 0.660) Low CHI Italian 1.84 (1.18-2.86) 0.007 0.716 (TPI, 0.689) German 1.52 (1.17-2.07) 0.008 0.693 (TPI, 0.664) TPI, tertiles of predictor index.

11598 Poster Session (Board #290), Mon, 1:15 PM-4:15 PM

Effect of whey protein isolate supplementation on body composition, muscle strength, and treatment tolerance in malnourished advanced cancer patients undergoing chemotherapy.

Paolo Pedrazzoli, Emanuele Cereda, Silvia Cappello, Luca Arcaini, Valeria Borioli, Elisa Ferraris, Silvia Chiellino, Giulia Stella, Marco Benazzo, Alessandra Ferrari, Catherine Klersy, Riccardo Caccialanza; Fondazione IRCCS Policlinico San Matteo, Pavia, Italy

Background: Malnutrition is frequent in cancer pts, particularly in advanced disease, which requires appropriate multidisciplinary interventions. We evaluated the benefit of whey protein isolate (WPI) supplementation in addition to nutritional counseling in malnourished cancer pts undergoing chemotherapy (CT). Methods: In a single-center, randomized, pragmatic, parallel-group controlled trial (ClinicalTrials.gov: NCT02065726), 166 malnourished advanced cancer pts undergoing CT were randomly assigned to receive nutritional counseling with (N=82) or without (N=84) WPI supplementation (20 grams/daily) for 3 months. Primary endpoint was the change in phase angle (PhA). Secondary endpoints included changes in standardized PhA (SPA), fat-free mass index (FFMI), body weight, muscle strength, quality of life and CT toxicity. Results: In pts with the primary endpoint assessed (modified intention-to-treat population), counseling plus WPI (N=66) resulted in improved PhA compared to nutritional counseling alone (N=69): mean difference, 0.48° [95%CI, 0.05 to 0.90] (P=0.027). Imputation of missing outcomes yielded consistent findings. WPI supplementation resulted also in improved SPA (P=0.021), FFMI (P=0.041), body weight (P=0.023), muscle strength (P,0.001) and in reduced risk of CT toxicity, particularly of severe (grade $3) events (Table). Conclusions: In malnourished advanced cancer patients undergoing CT and receiving nutritional counseling, 3-month supplementation with WPI resulted in improved body composition, muscle strength, body weight and reduced CT toxicity. Further trials in newly diagnosed specific cancer types are warranted. Clinical trial information: NCT02065726.

Treatment toxicity according to CTCAE 4.03 in the randomized population. Counseling (N=84) n Counseling + WPI (N=82) n Risk difference % P- Endpoints [%] [%] (95%CI) value Any 83 [98.8] 73 [89.0] -9.8 (-16.9, -2.6) 0.009 Hematological 17 [20.2] 10 [12.2] -8.0 (-19.3, 3.2) 0.21 Gastrointestinal 44 [52.4] 38 [46.3] -6.0 (-21.2, 9.2) 0.44 Others 64 [76.2] 57 [69.5] -6.7 (-20.2, 6.8) 0.38 Multiple 41 [48.8] 23 [28.0] -20.8 (-35.2, -6.3) 0.007 Grade 3-4 44 [52.4] 18 [22.0] -30.4 (-44.4, -16.5) 0.001 Hematological 12 [14.3] 4 [4.9] -9.4 (-18.4, -0.4) 0.06 Gastrointestinal 21 [25.0] 10 [12.2] -12.8 (-24.7, -0.9) 0.046 Others 11 [13.1] 4 [4.9] -8.2 (-16.9, 0.5) 0.10 CT suspension 10 [11.9] 5 [6.1] -5.8 (-14.5, 2.9) 0.28

11599 Poster Session (Board #291), Mon, 1:15 PM-4:15 PM

Interim analyses of the efficacy of a collaborative care intervention for patients diagnosed with comorbid cancer and depression.

Jennifer Lynne Steel, Jessica Miceli, Carol Lynn Hecht, Allan Tsung, Wallis Marsh, Michael Antoni, Yoram Vodovotz, Shaymal Peddada, Michael Spring, Heather Jackson, Yanet Vanegas, Donna Olejniczak, Qi Chen, David A. Geller; Univ of Pittsburgh, Pittsburgh, PA; University of Pittsburgh, Pittsburgh, PA; Comprehensive Liver Cancer Center, University of Pittsburgh Medical Center, Pitts- burgh, PA; Sylvester Comprehensive Cancer Center, University of Miami, Miller School of Medicine, Miami, FL; Unviersity of Pittsburgh, Pittsburgh, PA; University of Pittsburgh Medical Center Liver Cancer Center, Pittsburgh, PA

Background: There is an urgent need for evidence-based and scalable interventions to reduce depression, pain, and fatigue and improve quality of life in patients diagnosed with cancer. The aims of this study were to share the interim analyses of testing the efficacy of a stepped collaborative care intervention for patients diagnosed with cancers affecting the hepatobiliary and pancreatic system. Methods: Patients were screened for clinical levels of depression, pain, or fatigue and were enrolled in the study if they screened positive for depression, pain, and/or fatigue. After completing a baseline battery of instruments, patients were randomized to the stepped collaborative care intervention or the screening and referral arm. Post-treatment data was collected at 6 months and 12 months to assess efficacy and maintenance of change in depressive symptoms. Results: A total of 100 patients have completed the post-treatment assessment. Interim data analyses revealed that the mean age of patients was 64.0 years (SD = 10.3) and the majority of patient were male (51%), Caucasian (89%), diagnosed with liver cancer (47%) and stage III and IV (60%). Patients randomized to the stepped collaborative care intervention reported significant reductions in depressive symptoms (F(1,92) = 6.22, p = 0.014) and improvements in quality of life (F(1,92) = 7.36, p = 0.008) with moderate effect sizes (Cohen’s d = 0.547 and 0.652, respectively) at 6- months. The mean change in depressive symptoms from randomization to 6-month post- treatment was -4.3 (SD = 9.7) for the patients randomized to the collaborative care intervention and +0.71 (SD = 9.4) for the patients randomized to the screening and referral arm of the study. The mean change in quality of life from randomization to 6-month post-treatment was +4.5 (SD = 16.2) for the patients randomized to the collaborative care intervention and -4.4 (SD = 15.2) for the patients randomized to the screening and referral arm of the study. Conclusions: This promising evidence-based, scalable intervention to treat comorbid cancer and depression was shown to be effective in reducing depressive symptoms and improving quality of life in patients with cancer. Clinical trial information: NCT02939755.

11600 Poster Session (Board #292), Mon, 1:15 PM-4:15 PM

Effect of high flow oxygen on exertional dyspnea in cancer patients: A double-blind randomized clinical trial.

David Hui, Liliana Larsson, Sajan Thomas, Carol Harrison, Jimin Wu, Donald Mahler, Kenneth R. Hess, Juan Lopez-Mattei, Kara Thompson, Daniel Richard Gomez, Melenda Jeter, Steven H. Lin, Anne S. Tsao, George Eapen, Karen Basen-Engquist, Eduardo Bruera; University of Texas MD Anderson Cancer Center, Houston, TX; MD Anderson Cancer Center, Houston, TX; The University of Texas MD Anderson Cancer Center, Houston, TX; Valley Regional Hospital, Claremont, NH; Department of Cardiology, The University of Texas MD Anderson Cancer Center, Houston, TX; Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX

Background: High flow oxygen therapy is effective for hypoxemic respiratory failure. However, its effect on dyspnea in non-hypoxemic patients is unknown. In this 2x2 factorial, double-blind randomized clinical trial, we assessed the effect of flow rate (high vs. low) and gas (oxygen vs. air) on exertional dyspnea in cancer patients. Methods: Non-hypoxemic patients with cancer completed two structured cycle ergometer exercise tests with Low Flow Air [LFAir] at 2 L/min. They were then randomized to receive High Flow Oxygen [HFOx] with up to 60 L/min, High Flow Air [HFAir], Low Flow Oxygen [LFOx] or LFAir during a constant work rate exercise test at 80% maximum. Dyspnea intensity was assessed with the modified 0-10 Borg scale. The primary outcome was difference in the slope of dyspnea intensity vs. time during the third test. Secondary outcomes included difference in exercise time, vital signs, and adverse events. We estimated that 10 patients per arm will provide 86% power to detect a 1-standard deviation main effect and 86% power to detect a 2-SD interaction effect with an alpha of 5%. A linear mixed effects model was used to assess the impact of flow rate and gas on study outcomes. Results: 45 patients were randomized and 44 completed the study (10, 11, 12, 11 patients on HFOx, HFAir, LFOx, LFAir, respectively). The mean age was 63 (range 47-77); 18 (41%) were female; 34 (44%) had lung cancer; and 20 (46%) had metastatic disease. In mixed effects model, the association between the change in dyspnea intensity over time with flow rate differed significantly between oxygen and air (P = 0.04). Specifically, HFOx (slope difference -0.20, P , 0.001) and LFOx (-0.14, P = 0.01) were significantly better than LFAir, but not HFAir (+0.09, P = 0.09). Exercise time also significantly increased with HFOx (difference +2.5 min, P = 0.009) compared to LFAir, but not HFAir (+0.63 min, P = 0.48) or LFOx (+0.39 min, P = 0.65). HFOx was well tolerated without significant adverse effects. Conclusions: The combination of high flow rate and oxygen improved dyspnea and exercise duration during constant work exercise test in non-hypoxemic cancer patients. Larger trials are needed to confirm the benefits of HFOx during exercises. Clinical trial information: NCT02357134.

11601 Poster Session (Board #293), Mon, 1:15 PM-4:15 PM

Growth of palliative care (PC) in United States cancer centers: A national survey.

David Hui, Allison De La Rosa, Joseph Chen, Marvin Delgado-Guay, Yvonne Heung, Diane D. Liu, Kenneth R. Hess, Eduardo Bruera; University of Texas MD Anderson Cancer Center, Houston, TX; MD Anderson Cancer Center, Houston, TX; Loma Linda University Medical Center, Loma Linda, CA; Department of Biostatistics, The University of Texas MD Anderson Cancer Center, Houston, TX; The University of Texas MD Anderson Cancer Center, Houston, TX

Background: Outpatient PC facilitates timely referral and improved outcomes for cancer patients. We examined the change in outpatient PC services at US cancer centers over the past decade. Methods: Between April and August 2018, we surveyed all 62 National Cancer Institute designated cancer centers (NCI-CCs) and a random sample of 61 out of 1306 non-NCI-CCs. Two surveys previously used in a national study (Hui et al. JAMA 2010) were sent to each institution: a 22-question executive survey inquired about PC infrastructure and attitudes toward PC and an 82-question PC program leader survey inquired about the PC structures, processes and outcomes in detail. Generalized linear mixed model and logistic regression were used to examine the change in availability of outpatient PC services between 2009 and 2018 among NCI-CCs and non-NCI-CCs, respectively. Results: Among NCI-CCs, 40/ 62 (65%) executives and 52/61 (85%) program leaders responded. Among non-NCI-CCs, 41/61 (67%) executives and 27/39 (69%) program leaders responded. For NCI-CCs, we observed a significant increase in outpatient PC clinic between 2009 and 2018 (59% v. 95%; OR 13.1, 95% CI 2.6-66.8; P = 0.004) but no significant change in inpatient consultation team (92% v. 90%), PC unit (26% v. 40), nor institute-run hospice (31% v. 18%). For non-NCI-CCs, there was a significant increase in outpatient PC clinics (22% v. 42%; OR 2.51, 95% CI 1.01, 6.26; P = 0.05) and decrease in institute-run hospice (42% v. 22%; P = 0.05) over the past decade but no significant change in inpatient consultation team (56% v. 68%) and PC unit (20% v. 17%). The median (IQR) duration from outpatient referral to death increased from 90 (84, 120) days to 180 (131, 220) days for NCI-CCs and from 41 (28, 54) days to 84 (48, 120) days for non-NCI-CCs, respectively. We also observed significant growth in staffing, service hours, number of referrals, fellowship programs and rotations for oncology fellows, although research activity remains low. Conclusions: Cancer centers reported significant growth in outpatient PC clinics and overall PC infrastructure since 2009. However, major gaps in structures and processes exist, such as the lack of outpatient clinics at non-NCI-CCs, absence of PC units and limited research.

11602 Poster Session (Board #294), Mon, 1:15 PM-4:15 PM

The development of a nomogram to determine the risk for inappropriate opioid use in cancer patients.

Sriram J. Yennu, Rony Dev, Tonya Edwards, Joseph Anthony Arthur, Zhanni Lu, Elif Erdogan, Jimi S. Malik, Syed M. Naqvi, Jimin Wu, Diane D. Liu, Janet L. Williams, David Hui, Suresh Reddy, Eduardo Bruera; Department of Palliative Care, Rehabilitation, and Integrative Med- icine, Division of Cancer Medicine, The University of Texas MD Anderson Cancer Center, Houston, TX; The University of Texas MD Anderson Cancer Center, Houston, TX; University of Texas MD Anderson Cancer Center, Houston, TX; Unit 1414, Houston, TX; UT MD Anderson Cancer Center, Houston, TX; Department of Biostatistics, The University of Texas MD Anderson Cancer Center, Houston, TX

Background: Non-Medical opioid use is a growing crisis. Cancer patients at risk of harmful use of prescribed opioids are frequently underdiagnosed. The aim was to develop a nomogram to predict the probability of occurrence of Inappropriate opioid use that is, presence of SOAPP $ 7) among patients receiving outpatient supportive care consultation at a comprehensive cancer center. Methods: 3588 consecutive cancer patients referred to a supportive care clinic from March 1, 2016 to July 15, 2018 were reviewed. Patients were eligible if they had diagnosis of cancer, and were on opioids for pain for at least a week. All patients were assessed using Edmonton Symptom Assessment Scale with spiritual pain and financial distress (ESAS-FS), MEDD (morphine equivalent daily dose), SOAPP-14 (validated questionnaire for assessment of risk of inappropriate opioid use, and CAGE-AID (screening questionnaire for alcoholism/substance use disorder). Patients at with SOAPP+ were defined by SOAPP score $7. A nomogram was devised based on the risk factors determined in the multivariate logistic regression model and it can be used to estimate the probability of inappropriate opioid use. Results: Median age was 62yrs. Median ESAS pain item score on consultation was 5, Median ECOG was 2.20.4% were SOAPP+ and 10.1% were CAGE-AID+. SOAPP+ was significantly associated with gender, race, marital status, smoking status, depression, anxiety, financial distress, MEDD and CAGE score. The C-index is 0.8(CI 0.78, 0.82). A nomogram was developed. For example, for a male Hispanic patient, who is married, never smoked, with the following ESAS scores: (depression = 3, anxiety = 3, financial distress = 8), CAGE score of 0, and MEDD of 20, the total score is 9+9+0+0+6+10+26+0+1 = 61. In the nomogram a score of 58 indicates the probability of inappropriate opioid use being 0.1 and a score of 88 indicates the probability of 0.2. Based on the patient’s total score of 61, the probability of his aberrant behavior is between 10% to 20% (close to 10%). Conclusions: A nomogram can predict the risk of inappropriate opioid use in cancer patients.

11603 Poster Session (Board #295), Mon, 1:15 PM-4:15 PM

Predictors and trajectories of fatigue in ovarian and uterine cancer.

Hanneke Poort, Belle Hadewijch de Rooij, Shicheng Weng, Nicole Ezendam, Lonneke V van de Poll- Franse, Alexi A. Wright; Dana-Farber Cancer Institute, Boston, MA; Netherlands Comprehensive Cancer Organisation (IKNL), Eindhoven, Netherlands; Comprehensive Cancer Organisation Netherlands, Eindhoven, Netherlands

Background: Fatigue is one of the most common and distressing symptoms reported by patients with gynecological cancers, but few studies have empirically examined whether it resolves without intervention. The aims of this study were to identify: 1) clinically-distinct subgroups of patients with fatigue over time and 2) medical and psychological predictors of clinically-significant fatigue one-year post-diagnosis. Methods: Secondary analysis of a prospective cohort study. Symptoms of fatigue, depression, and anxiety were assessed at diagnosis, 6-months, and 12-months with the 10-item Fatigue Assessment Scale (FAS), and the Hospital Anxiety and Depression Scale (HADS), respectively. Group-based trajectory modeling was used to classify patients by their fatigue scores over time, and logistic regression models were fit to examine associations between clinically-significant fatigue and demographic, clinical, and psychosocial characteristics. Patients with recurrent or primary progressive cancers were excluded from the primary analysis. Results: Among 312 participants with newly diagnosed ovarian (n = 112) or endometrial (n = 200) cancers, the median age was 66 years (IQR = 60-72 years), 36% had ovarian cancer, and 79% had early stage disease. At baseline, 49% reported clinically significant fatigue and one year later, 42% had persistent fatigue. During the year after diagnosis, there were three distinct trajectories of fatigue that persisted: (1) severe fatigue (15%), (2) moderate fatigue (45%), and (3) no fatigue (41%). Patients with $2 comorbid conditions (odds ratio [OR] 2.52, 95% confidence interval [CI] = 1.21-5.27, P = 0.01), clinically significant fatigue at baseline (OR 5.47, 95% CI = 2.71-11.03, P , .0001), and those reporting depressive symptoms at baseline (OR 3.45, 95% CI = 1.13-10.55, P = 0.03) were more likely to report clinically-significant fatigue at 12 months. Conclusions: Half of women with gynecologic cancers have clinically-significant fatigue at diagnosis and 42% of survivors have persistent fatigue one year later, suggesting spontaneous regression of symptoms is rare. Importantly, depressive symptoms contribute to persistent fatigue and are modifiable with psychological interventions. Future studies should test scalable psychological interventions to address depressive symptoms, reduce fatigue, and improve quality of life in women with gynecologic cancers.

11604 Poster Session (Board #296), Mon, 1:15 PM-4:15 PM

Phase 2 trial of Symptom screening with Targeted Early Palliative care (STEP) for patients with advanced cancer.

Camilla Zimmermann, Breffni Hannon, Monika K. Krzyzanowska, Madeline Li, Gary Rodin, Ashley Pope, Nadia Swami, Mohana Giruparajah, Doris Howell, Amit M. Oza, David Warr, Jennifer J. Knox, Natasha B. Leighl, Srikala S. Sridhar, Rebecca M. Prince, Stephanie Lheureux, Aaron Richard Hansen, Christopher M. Booth, Deborah Jane Dudgeon, Lisa W Le, STEP Trial Group; Princess Margaret Cancer Centre, University of Toronto, Toronto, ON, Canada; Princess Margaret Hospital, Toronto, Ontario, Canada; Princess Margaret Cancer Centre, Toronto, ON, Canada; Palliative Care and Psychosocial Oncology, Princess Margaret Cancer Centre, University of Toronto, Toronto, ON, Canada; Princess Margaret Hospital, University Health Network, Toronto, ON, Canada; University Health Network, Toronto, ON, Canada; Princess Margaret Hospital, Toronto, ON, Canada; University Health Network, Princess Margaret Cancer Centre, Toronto, ON, Canada; Kingston Health Sciences Centre, Kingston, ON, Canada; Queen’s University, Kingston, ON, Canada

Background: Routine early palliative care (EPC) by specialized teams improves quality of life for patients with advanced cancer, but may not be scalable. To plan for a larger randomized controlled trial, we conducted a phase 2 trial of STEP, a novel intervention of targeted EPC based on symptom screening. Methods: Participants with advanced cancer, ECOG 0-2, and clinical prognosis $6 months were recruited from lung, gastrointestinal, genitourinary, breast and gynecology outpatient clinics. Symp- toms were screened at every outpatient visit using the Edmonton Symptom Assessment System-revised (ESAS-r). Moderate to severe symptom scores ($4/10 for pain, nausea, dyspnea, depression, anxiety; $7/10 for fatigue, appetite, drowsiness, well-being) triggered an e-mail to a nurse, who called the patient, offering an EPC visit. Participants completed outcome measures at baseline, 2, 4 and 6 months (primary endpoint). Trial feasibility criteria were: i) $100 patients accrued in 12 months; ii) $70% complete screening for $70% of visits; iii) $60% of those for whom a call is triggered meet at least once with the EPC team; iv) $60% complete measures at each endpoint. Results: From 11/2016- 1/2018, 116 patients were enrolled; 89/116 (77%) completed screening for $70% of visits and 59% (69/116) received a call triggered by symptoms. Of those receiving a call, 62% (43/69) received EPC; 3 further patients were referred by their oncologist. Measure completion was 79% (81/116) at 2, 61% (71/116) at 4, and 57% (66/116) at 6 months. By trial end (6 months), patients who received a call and accepted EPC involvement had better symptom control (ESAS-r-CS mean change in those receiving vs. deferring EPC: -0.07616.9 vs 11.8613.7, p = 0.02) and less deterioration in mood (PHQ-9: -0.463.4 vs 2.662.3, p = 0.003). There was no difference between those receiving versus deferring EPC in quality of life (FACIT-Sp: -4.7613.6 vs -3.2615.4, p = 0.75; QUAL-E: -2.668.4 vs -1.5611.0, p = 0.62), or satisfaction with care (FAMCARE-P-16: -3.267.9 vs -3.566.1, p = 0.79). Conclusions: STEP is feasible in patients with advanced cancer. More than half of patients have moderate to severe symptoms, and acceptance of the triggered EPC visit should be encouraged.

11605 Poster Session (Board #297), Mon, 1:15 PM-4:15 PM

Effect on pain outcomes in patients treated for painful bone metastases with biopolar RFA: The OPuS One study.

Sandeep Bagla, Jason Levy, Thomas Hopkins, OPuS One Study Investigators; Vascular Institute of Virginia, Woodbridge, VA; Northside Hospital, Atlanta, GA; Duke University Health System, Durham, NC

Background: To report the preliminary outcomes from the prospective, nonrandomized, multicenter study of the OsteoCool Radiofrequency Ablation system used for the palliative treatment of patients with painful bone metastases (OPuS One Study). Methods: Subjects were prospectively enrolled (NCT03249584) from October 2017 through August 2018. Inclusion criteria were: 2 or less painful target sites involving the thoracolumbar spine, pelvis and/or sacrum, with a worst pain $4/10 measured by the Brief Pain Inventory (BPI) within the previous 24 hours. RFA was performed with the OsteoCool device (Medtronic) under imaging guidance and followed by cementoplasty at operator discretion. Subjects were evaluated prior to RFA, at discharge, 3 days, 1 week, and 1-, 3-, 6-, and 12-months. Demographics and disease characteristics were collected at baseline and outcomes were assessed using validated measures. Device-, procedure- and/or therapy-related adverse events (AEs) were collected. Results: At the time of the analysis, 37 subjects at 8 US centers underwent RFA. The mean age was 64 years (range 30-89) and 68% were female. Mean baseline BPI worst pain was 8.6. The most common histology was adenocarcinoma; breast (32%) and lung (24%) were reported as the most common primary cancer. Thirty-one subjects (84%) were treated for lesions involving the thoracolumbar spine, while 6 subjects (16%) were treated for lesions located the pelvis and/or sacrum. All RFA procedures were technically successful. Of the 14 subjects presenting at the 3-month visit, 93% did not undergo radiation therapy at the targeted site(s) between baseline and 3 months. Seventy- four percent (25/34) of subjects at 1 week and 93% (13/14) of subjects at 3 months reported a clinically significant change from baseline in worst pain of $ 2 points. One AE, drug hypersensitivity, resolved without intervention. All eleven deaths reported were attributed to the subjects underlying malignancy and not related to RFA. Conclusions: The preliminary results of the OPuS One study show rapid pain improvement at 1 week and sustained long-term relief through 3 months in patients with metastatic bone disease. Clinical trial information: NCT03249584.

11606 Poster Session (Board #298), Mon, 1:15 PM-4:15 PM

Progress and trends in the utilization and adoption of palliative care in patients with terminal gynecologic cancer in the United States: A study of 4,264 women.

Anthony Milki, Amandeep Kaur Mann, Daniel Stuart Kapp, John K. Chan; Stanford University, School of Medicine, Stanford, CA; Palo Alto Medical Foundation Research Institute, Palo Alto, CA; California Pacific Medical Center Research Institute, San Francisco, CA

Background: To determine the factors associated with the utilization of palliative care (PC) in patients with metastatic gynecologic cancer who died while hospitalized. Methods: Data were extracted from the National Inpatient Sample (NIS) database for patients with cervical, uterine, and ovarian cancers from 2005 to 2011. Chi-squared and logistic regression models were employed for statistical analyses. Results: Of 4264 women with gynecologic cancer, 983 (23.0%) utilized PC (median age: 66 years). 2633 (61.8%), 1034 (24.3%), and 597 (14.0%) patients had ovarian, uterine, and cervical cancer, respectively. The majority were white (57.9%), and the remainder were black (12.6%), Hispanic (7.8%), and Asian (3.1%). 24.9, 23.9, 23.6, and 25.2% were low, low-middle, middle-high, and high with respect to median income. Medicare, Medicaid, and private insurance was found in 46.7, 10.8, and 37.6% of patients. 36.9, 21.0, 18.0, and 24.1% of patients were treated in hospitals in the South, West, Midwest, and Northeast. 11.3, 23.9, and 64.2% of patients were treated in small, medium, and large hospitals, designated as teaching (53.1%) and nonteaching (46.3%) institutions. The utilization of PC increased from 5.2% in 2005 to 30.4% in 2011. Older age (OR: 1.34; 95% CI: 1.10-1.64; P = 0.01), high SES (OR: 1.37; 95% CI: 1.09-1.72; P = 0.01), privately insured (OR: 1.81; 95% CI: 1.46-2.24; P , 0.001), treatment at Western (OR: 1.94; 95% CI: 1.56-2.41; P , 0.001) and Midwestern hospitals (OR: 1.43; 95% CI: 1.15-1.77; P = 0.001), and depression (OR: 1.34; 95% CI: 1.00-1.80; P = 0.05) were associated with higher PC use. However, race, cancer type, hospital teaching status, and hospital location were not associated with PC use. Conclusions: The use of inpatient palliative care has increased significantly over the study period. The lower utilization of palliative care for terminal illness associated with younger, lower socioeconomic status, Southern and smaller volume hospitals warrants further attention.

11607 Poster Session (Board #299), Mon, 1:15 PM-4:15 PM

A phase II, double-blinded, randomized study using a proprietary amino acid mixture as diarrhea prevention in patients undergoing autologous stem cell transplantation (AST) for multiple myeloma (MM) and non-Hodgkin lymphoma (NHL).

Laura Luque, Robert Soiffer, Brett Glotzbecker, Haesook Kim, Zachariah Michael DeFilipp, Dorothy Mary Kate Keefe; Entrinsic Health Solutions, Inc., Norwood, MA; Center for Stem Cell Transplantation, Division of Hematologic Malignancy, Department of Medical Oncology, Dana- Farber Cancer Institute, Harvard Medical School, Boston, MA; Dana-Farber Cancer Institute, Boston, MA; Dana Farber Cancer Institute, Boston, MA; Winship Cancer Institute, Atlanta, GA; Royal Adelaide Hospital, Adelaide, Australia

Background: Melphalan, remains the mainstay of conditioning for autologous hematopoietic stem cell transplantation (HSCT) in multiple myeloma (MM) and non-Hodgkin lymphoma (NHL) patients. Gas- trointestinal symptoms represent the most significant non-hematologic toxicities following high-dose melphalan conditioning, with approximately 40% of patients experiencing CTCAE grade 2 or higher diarrhea following conditioning regimens containing melphalan. Enterade is a proprietary blend of electrolytes and five amino-acids that can facilitate retention of the absorbing capacity of the small intestine by rebuilding the villi and reduce antigenic translocation by tightening the mucosal barrier. In a study of irradiated mice, enterade improved survival and improved body weight following irradiation. The goal was to investigate the effectiveness of enterade to reduce GI toxicities after high-dose melphalan chemotherapy. Methods: The trial was designed as a Phase 2, multi-center, double-blinded, 2-arm randomized study. 114 MM or NHL patients were enrolled between October 2016 and October 2017. Patients received either two 8oz bottles/day of enterade or placebo starting on the day of admission through Day +14. GI toxicity was scored by the CT-CAE 4.0 system from admission through Day + 14. Compliance was arbitrarily set at consumption of 2 bottles daily for 11+ days. Results: Of the 114 enrolled patients, 99 (61 MM, 38 NHL) attempted to consume at least 1 dose of enterade/placebo. Compliance overall was much lower than anticipated; with no MM patients achieving compliance compared to 34.2% in the NHL group. Compliance in NHL patients was 31.6% in the enterade group versus 36.8% in the placebo group. Analysis of primary endpoint in NHL patients showed a 16% incidence of diarrhea $ grade 2 in enterade compliant patients versus 86% in the placebo group (p= 0.02). Conclusions: Eleven days of two 8oz bottles of liquid is a difficult task during ASCT, especially for MM with nausea, altered taste and poor appetite. For those NHL patients, compliant per protocol (who consumed $11 days), enterade significantly reduced diarrhea. The use of enterade to prevent diarrhea following high-dose chemotherapy should be explored further in populations capable of reasonable oral intake. Clinical trial information: NCT02919670.

11608 Poster Session (Board #300), Mon, 1:15 PM-4:15 PM

Safety, tolerability, and PK of topical calcitriol formulation for treatment of chemotherapy- induced alopecia (CIA) in patients receiving taxane-based regimen: Final results.

Mario E. Lacouture, Rangaprasad Sarangarajan, Joaquin Juan Jimenez, Michael A. Kiebish, Niven R. Narain, Brian Berman, Shari Beth Goldfarb; Memorial Sloan Kettering Cancer Center, New York, NY; BERG, LLC, Framingham, MA; University of Miami Miller School of Medicine, Miami, FL

Background: Chemotherapy-induced alopecia (CIA) following taxane-based treatment regimen is due to direct toxicity of these agents on rapidly dividing cells within hair follicles. Currently no oral or topical agents have been approved for prevention of CIA. In murine studies, topical calcitriol reduced CIA, likely due to arrest of cell cycle in healthy hair follicles, and reduction in the sensitivity of follicular epithelium to chemotherapy. Methods: A 3+3 dose-escalation strategy was used in this Phase 1 study with 3-6 patients at six dose levels (5/10/20/40/60/80 mg/mL) were assessed for safety and tolerability of BPM31543. Patients with a diagnosis of breast cancer, gynecologic cancer or sarcomas receiving a taxane-based chemotherapy regimen applied 1mL of the formulation to the scalp BID, $ 5 days prior to starting chemotherapy for at least 3 months or until treatment completed. Safety and efficacy assessments included AE monitoring, PK analysis, blinded photographic assessments and patient self-assessment. Results: 22/23 (95.7%) female patients receiving treatment and included in the safety population experienced at least one TEAE. The most frequently experienced TEAEs were alopecia (14 pts; 60.9%), fatigue (11 pts; 47.8%), nausea (9 pts; 39.1%), peripheral sensory neuropathy (7 pts; 30.4%), and maculopapular rash and vitamin D increased each in six patients (26.1%). Of these, elevated vitamin D and rash were possibly or probably related to treatment. Fatigue, nausea, and neuropathy were likely due to chemotherapy. In 18 patients included in the post-dose versus pre-treatment comparison, there was no dose-dependency on systemic levels of calcitriol. Hair loss , 50% from baseline was observed in 8 patients at week 7 that was maintained at week 15 in 2 patients. Conclusions: Study results showed BPM31543 applied topically twice daily to the scalp in patients receiving taxane-based chemotherapy was safe and well-tolerated. No DLT was observed up to 80 mg/mL dose and no MTD level was reached. There was a signal of potential efficacy detected at each dose level. A seamless Phase 2/3 trial strategy for clinical development is planned. Clinical trial information: NCT01588522.

11609 Poster Session (Board #301), Mon, 1:15 PM-4:15 PM

Lower costs of care, improved discharge disposition, and improved survival of advanced cancer patients (ACP) receiving early inpatient palliative care (PC) compared to standard oncologic care (SOC).

Christopher Daugherty, Julie Johnson, Stacie K. Levine, Kristen Wroblewski, Bradford Lane, William Dale, Monica Malec, Fay J. Hlubocky; University of Chicago, Chicago, IL; University of Chicago Medicine, Chicago, IL; CIY, Chicago, IL; University of Chicago Pritzker School of Medicine, Chicago, IL

Background: Outpatient PC improves ACP symptom burdens, end-of-life care transitions, and mortality thereby enhancing quality of life. Yet, the financial implications, discharge disposition, and survival benefits of early, inpatient PC compared to SOC remains less understood. Methods: Retrospective cohort analysis of ACP receiving either PC or SOC between Jan 2015-Dec 2015 (N = 810). ACP cohorts were compared for demographics, costs, disposition, and survival. Financial costs collected included: fixed (overhead expenditures, facility maintenance, hospital property); variable (patient care supplies, diagnostic/therapeutic supplies, medications); operating (fixed, variable, breaking-even costs); direct (labor, materials, commissions, piece-rate wages, manufacturing supplies); indirect (production- supervision salaries, quality control, insurance, depreciation). Univariate and multivariate analyses were completed. Results: 468wereadmittedtoPCand342toSOC.ComparedwithSOC,PCweremorelikely to be: younger (61.1613.2 v. 62.5613.0, p = 0.02); African American (48% v. 36%, p = 0.0045); female (50% v. 40%, p = 0.005); and have shorter length of stay (5.764.9 v. 6.266.5, p = 0.01). PC had significantly less 30-day readmissions (16% v 23%, p = 0.03) and lower costs: direct ($9,478 v. $10,416, p = 0.01); indirect ($9,538 v. $10,999, p = 0.002); fixed ($10,308 v. $12,076, p = 0.001); variable ($8,709 v. $ 9,339, p = 0.02); operating ($19,017 v. $21,416, p = 0.003).Compared with SOC, ACP receiving PC were more likely to be discharged to: home (55% v.45%, p = 0.01); healthcare facilities (e.g. skilled nursing, inpatient rehabilitation) (36.1% v. 20%, p = 0.04); and hospice (home and inpatient) (7.7% v 5.8%, p = 0.02). PC had overall greater median survival from the time of discharge (106.8699.95 v. 73.8661.93, p = 0.03) compared to SOC. Conclusions: Early PC results in less financial strain, greater cost savings, and improved outcomes for younger and underserved inpatient ACP. Our results provide additional evidence for policies supporting that ACP access to routine PC must become a healthcare priority.

11610 Poster Session (Board #302), Mon, 1:15 PM-4:15 PM

Learning from best scalp cooling practices in a registry: Differences in results from n>7000 patients with solid tumors.

Corina van den Hurk, Martin W Dercksen, J. W. R. Nortier, Wim Breed; Netherlands Comprehensive Cancer Organisation, Utrecht, Netherlands; Maxima ´ Medical Center, Eindhoven, Netherlands; Leiden Univ Med Ctr, The Hague, Netherlands; Catharina Hospital, Eindhoven, Netherlands

Background: Hair loss is a frequently occurring and stigmatizing side effect of chemotherapy. Worldwide scalp cooling is being introduced to prevent chemotherapy-induced alopecia (CIA). In the Netherlands scalp cooling is implemented in many hospitals since 2005. FDA clearance has been approved for scalp cooling among breast cancer patients in the USA in 2016. Recently approval has been expended for solid tumors. Methods: In a prospective, longitudinal registry data have been collected between 2006 and 2017. Patients could be included if they received chemotherapy that induced severe alopecia, regardless type of solid tumor, stage of disease, age, gender or receiving adjuvant or palliative treatment. Patients were eligible for evaluation of hair loss after they received at least 2 cycles of chemotherapy or if they ceased scalp cooling because of severe hair loss after the first cycle. Failure was defined as feeling the need to use a wig or head cover. Data will be presented using descriptive statistics and multivariate regression analysis to explore determinants of scalp cooling efficacy per largest groups of chemotherapy regimen, and evaluate variety between hospitals. Results: Preliminary results show data of 7378 patients from 68 hospital locations of whom 75% had breast cancer and 8% prostate cancer. Overall 57% of patients did not feel the need to wear a wig or head cover. Variation was observed between hospitals in scalp cooling procedures: e.g. wetting the hair (0-100% of patient) and cooling- and infusion times that varied per type of chemotherapy. Also results varied between hospitals per type and dose of chemotherapy (n . 10 patients), e.g. minimal 37% and maximal 86% success rates between hospitals for paclitaxel-carboplatin (n = 498), 17-54% success for irinotecan monotherapy (n = 275), 25-79% success for 5FU/epirubicin/cyclophosphamide-docetaxel (n = 843) and 79-94% for docetaxel monotherapy (n = 824). Also variation in satisfaction with information about scalp cooling and nursing expertise was observed. Results from the regression analyses will be presented at the conference. Conclusions: Scalp cooling efficacy varied enormously between hospitals. A registry is a useful tool to identify best practices and to provide guidance to further improve results. An international registry has been set up to also collect data on CIA among scalp cooled and non-scalp cooled patients in the USA, Australia and the UK.

11611 Poster Session (Board #303), Mon, 1:15 PM-4:15 PM

Therapy discontinuation processes in a gynecologic oncology population.

Catherine H. Watson, Allison Puechl, Tracy Truong, Stephanie Lim, Laura Jean Havrilesky, Brittany Anne Davidson; Duke University, Durham, NC; Division of Gynecologic Oncology, Duke Cancer Institute, Duke University Medical Center, Durham, NC; Duke University School of Medicine, Durham, NC

Background: The decision to discontinue anti-cancer therapies in oncology patients is a complex one and may occur in several unique ways. We sought to understand the processes of therapy discontinuation in a gynecologic oncology population and to discern possible changes in the distribution of these processes after implementation of a palliative care (PC) quality improvement project. Methods: Women with incurable gynecologic malignancies seen in the outpatient setting at an academic center were identified with the recommendation for ‘goals of care’ discussion within 3 visits following identification. Processes of discontinuing chemotherapy for this population were assessed and categorized into 1 of 4 categories: definitive outpatient decision, definitive inpatient decision, interruption by hospitalization, treatment holiday, and “no decision.” Retrospective chart review identified a similar cohort of women prior to the implementation of our PC intervention. Univariate analyses were conducted to determine associations between characteristics and binary outcome of definitive outpatient decision versus all other decision processes. Results: 90/102 (88%) pre-intervention subjects and 83/157 (53%) post-intervention subjects had died at time of analysis. Of the total deceased cohort, 59/173 (34%) made a definitive decision to stop therapy in the outpatient setting. After implementation of the PC initiative, there was a trend towards fewer women identified as having made “no decision” (18.1% vs 30%). Those who made a definitive outpatient decision were less likely to die within 30 days of hospitalization than those who did not (OR 0.16 [95% CI 0.07,0.390], p , 0.0001). Conclusions: Discontinuation of therapy is a nuanced concept in gynecologic oncology patients that can be stratified into several processes. While our data demonstrates a possible increase in active decision-making with the PC initiative, a majority of patients near the end of life still did not make definitive therapy cessation decisions. This reveals an urgent need for the development of initiatives to enhance patient engagement and shared decision making for women near the end of life.

11613 Poster Session (Board #305), Mon, 1:15 PM-4:15 PM

Communicating Oncologic Prognosis with Empathy (COPE) Pilot Study.

Boone W. Goodgame, Farya Phillips, Wenhui Vivian Zhang, Barbara L. Jones; Dell Medical School, Austin, TX; The University of Texas at Austin, Austin, TX; UT Austin, Austin, TX

Background: Multiple studies show that patients with advanced cancer often believe the goal of treatment is cure. Our interprofessional team designed the Communicating Oncologic Prognosis with Empathy (COPE) guide to enhance communication for patients, families, and the healthcare team. This unique, single-page tool allows treating physicians to estimate prognosis for patients, including likelihood of living 6 mo, 1y, 2y, and 5y, with or without treatment. Methods: Patients with incurable cancer were enrolled at one oncology clinic. Using a pre/post-test design, participants completed the validated prognosis and treatment perception survey (PTPQ) then met with their oncologist. The COPE guide was used for the discussion and the patient received a copy of the completed COPE. The COPE guide was placed in the patient chart and the healthcare team (social work, spiritual care etc.) used it to provide patient support and explore the patient’s response to the information. Post-test was completed 4-8 weeks later. Qualitative exit interviews were used to assess patient reported impact on communication with medical team, caregivers and family, and quality of life. Results: 43 patients were enrolled and 25 completed all study measures. All patients had advanced cancer with 96% of patients (24) extremely unlikely to be cured ( , 1% chance). Oncologist’s treatment goals were better quality of life and longer life. All participants completing the study found the communication guide helpful, scoring 9-10 out of 10. 95% of patients enrolled wanted to hear details about their prognosis. Pre-intervention, 28% (7) of patients reported their primary goal of treatment as cure (n = 7; 28%) or to extend life as long as possible (n = 7; 28%). Pre-intervention, patient’s perceptions of their oncologist’s goals of treatment were cure (n = 11; 44%) or extend life (n = 8; 32%). At post-test, 8% of participants (2) inaccurately reported goal of treatment as cure (p , 0.05). Conclusions: These findings highlight major gaps in patients’ understanding of their prognosis. COPE is a promising mechanism to enhance patient-centered communication about prognosis and treatment for patients with incurable cancer.

11614 Poster Session (Board #306), Mon, 1:15 PM-4:15 PM

Engaging care partners in breast cancer care and communication.

Jennifer Aufill, Jennifer L. Wolff, Diane Echavarria, JaAlah-Ai Heughan, Kimberley T. Lee, Elissa Thorner Bantug, Howard Levy, Sydney M. Dy, Antonio C. Wolff; Johns Hopkins University Bloomberg School of Public Health, Baltimore, MD; Johns Hopkins Sidney Kimmel Comprehensive Cancer Center, Baltimore, MD; Johns Hopkins University School of Medicine, Baltimore, MD

Background: Family is often overlooked in cancer care. Little is known about the patient preferences for involving family in communication, whether preferences may be elicited and supported at the point of care, and impact on care quality. Methods: We conducted a two-group pilot randomized controlled trial (NCT03283553) of patients on active treatment for breast cancer and the “care partner” who accompanies them to routine visits (n = 132 dyads). Intervention dyads (n = 69) completed a self- administered checklist to clarify the care partner role, establish a shared visit agenda, and facilitate access to the electronic health record (MyChart) patient portal. Control dyads (n = 63) received usual care. Intervention acceptability and short-term effects were assessed from post-visit surveys and MyChart utilization at 6 weeks. Results: At baseline, most patients (89.4%) but few care partners (1.5%) were registered for MyChart. Most patients (79.4%) wanted their care partner to have access to their records and 39.4% of care partners reported accessing it in the past year using the patient’s account login/password. In completing the checklist, intervention patients and care partners identified an active communication role for the care partner and similar issues for the visit agenda: topics most frequently selected were treatment goals/expectations (75.4% & 66.7%, respectively), symptoms/side effects (73.9% & 62.3%) and chances of cancer recurrence/spread (49.3% & 44.9%). More than 90% of intervention participants reported that completing the checklist was easy, useful, and recommended it to others. At 6 weeks, intervention (vs control) care partners were more likely to be registered for MyChart (75.4 % vs 1.6%; p , 0.001), to have logged in (43.5% vs 0%; p , 0.001) and viewed clinical visit notes (30.4% vs 0%; p , 0.001), but no more likely to have exchanged direct messages with the clinical team (1.5% vs 0%; p = 0.175). No difference in patient MyChart registration, use, or messaging was found at 6 weeks, but intervention patients more often viewed clinical visit notes (50.7% vs 9.5%; p , 0.001). Conclusions: A self-administered patient-family communication intervention affected online practices of patients and cancer care partners. Follow-up continues. Clinical trial information: NCT03283553.

11615 Poster Session (Board #307), Mon, 1:15 PM-4:15 PM

Current practice for screening and management of financial distress at NCCN member institutions.

Nandita Khera, Jessica Sugalski, Diana Krause, Richard Butterfield, Nan Zhang, Warren Smedley, F. Marc Stewart, Joan M. Griffin, Yousuf Zafar, Stephanie Lee; Mayo Clinic, Phoenix, AZ; National Comprehensive Cancer Network, Fort Washington, PA; National Comprehensive Ca, Plymouth Meeting, PA; Mayo Clinic, Scottsdale, AZ; Patient Care Connect - Powered By UAB Health System, Birmingham, AL; Fred Hutchinson/UW/Children’s Hospital, Seattle, WA; Mayo Clinic, Rochester, MN; Duke University Medical Center, Durham, NC; Fred Hutchinson Cancer Research Center, Seattle, WA

Background: Deficiencies and barriers exist to delivering comprehensive and affordable cancer care. Understanding the variation in organizational commitment, existing programs, and expected outcomes for screening and management of financial distress is needed. Methods: Representatives from 17 of 27 NCCN Member Institutions (63%) completed an online survey in November 2018 conducted by the NCCN Best Practices Committee. Centers were classified based on number of unique patients seen per year, as large ( . 10,000) (76%), or small ( , 10,000) (34%). The survey focused on institutions’ screening and management practices for patient financial distress, perceived barriers in implementation, and leadership attitudes. Results: Routine screening for financial distress was reported by 77% of centers, and most used social worker assessments (94%). 56% screened patients throughout the cancer journey. Help with drug costs, meal or gas vouchers and payment plans were offered by 100% of centers. Formal pre-authorization programs and assistance with claims and denials was offered by 81%. Charity care for medical costs was provided by 100% of the large centers compared to only 33% of small centers (p = 0.03). Median number of social workers (24 vs. 3; p = 0.01) and pharmacy representatives (6 vs. 2; p = 0.02) was also different between large and small centers. 76% evaluated the impact of financial advocacy services through number of patients assisted (85%), bad debt and charity write-offs (85%) or patient satisfaction surveys (54%). 6% and 12% reported overall effectiveness of institutional practice for screening and management of financial distress as poor/ very poor respectively. Inadequate staffing and real time resources (69%), limited institutional budget (50%), lack of reimbursement (50%), and clinical time constraints (50%) were reported as potential barriers in provision of these services. 94% agreed about stronger integration of financial advocacy services into oncology practice and 84% felt that success of these services should be a quality metric. 31% of large centers vs. 100% of small centers plan to increase staffing in this area in the next 5 years. Conclusions: Majority of NCCN Member Institutions report screening and management programs for financial distress, though the actual practices and range of services vary widely. Information from this study can help centers benchmark their performance relative to similar cancer programs and identify best practices in this area.

11616 Poster Session (Board #308), Mon, 1:15 PM-4:15 PM

How far do laypersons believe in the cure of cancer? Results of the EDIFICE6 survey.

Thibault De La Motte Rouge, Sebastien Couraud, Morgan Roupret, Chantal Touboul, Christine Lhomel, Jerome Viguier, Francois Eisinger, Laurent Greillier, Jean F. Morere; Centre Eugene-Marquis, ` Rennes, France; Acute Respiratory Medicine and Thoracic Oncology Department, & CIRCAN Program Co- ordinator, Cancer Institute of Hospices Civils de Lyon, Lyon Sud Hospital, Pierre Benite, France; Urology University Paris 6, Paris, France; Kantar Health, Paris, France; Roche, Boulogne-Billancourt, France; Hopital Bretonneau, Tours, France; Institut Paoli-Calmettes, Marseille, France; Assistance Publique–Hopitauxˆ de Marseille, Aix Marseille University, Marseille, France; Paul Brousse Hospital, Villejuif, France

Background: Advances in therapy and early detection, by population-based screening in particular, have significantly increased the number of patients cured from cancer. However, cancer patients’ chances of being cured varies strongly from one cancer to another. We studied the understanding of the concept of cure in the lay population and the factors associated with believing that cancer can be cured. Methods: The French nationwide observational survey, EDIFICE 6, was conducted online (June 26-July 28, 2017) on a core sample of 12 046 individuals (age, 18-69y). Representativeness was ensured by quota sampling on age, sex, profession, and stratification by geographical area/type of urban district. This analysis focused on understanding of the meaning of cure in breast (BC), cervical (CC), colorectal (CRC), lung (LC) and bladder (BLC) cancer for individuals with no history of cancer. Results: The majority of respondents believed that cure exists (BC 95%; CC 91%; CRC 89%; BLC 87%, LC 71%). Some agreed with the definition that cure is the disappearance of the disease (BC 42%; CC 38%; CRC 35%; BLC 33%, LC 25%), while others preferred the definition that cure is several years without disease (BC, CC 53%; CRC, BLS 54%; LC 46%). More men than women (P , 0.05) believed that cure exists for CRC, BLC and LC. Socially non-vulnerable individuals were more likely to believe in cure than their vulnerable counterparts (P , 0.05), as were individuals aged 50-69y (P , 0.05) versus those of 18-50y, and for all cancer types except LC. In multivariate analysis, the variable “clinical research enables progress” was correlated with believing that cure exists (BC, OR = 2.93; CC, OR = 1.86; CRC, OR = 2.22; LC, OR = 1.57, BLC, OR = 2.06), as was “progress is rapid” (BC, OR = 1.61; CC, OR = 1.66; CRC, OR = 1.7; LC, OR = 1.84; BLC OR = 1.68), and also social non-vulnerability. However, the variables “prevention”, respectively screening/treatments, “are important for cancer control” had a low impact on the belief in cure (OR~1). Conclusions: The lay population is relatively optimistic about the cure for cancer. Confidence in the existence of cure relies on medical progress. However, factors related to individual behavior, e.g., prevention and screening, did not affect the perception of cure.

11617 Poster Session (Board #309), Mon, 1:15 PM-4:15 PM

Advanced cancer patients’ changes in terminal illness understanding (TIU) and their psychological well-being.

Login S. George, Megan Johnson Shen, Paul K. Maciejewski, Andrew S. Epstein, Holly Gwen Prigerson; Memorial Sloan Kettering, New York, NY; Weill Cornell Medical College, New York, NY; Memorial Sloan Kettering Cancer Center, New York, NY

Background: Although accurate TIU is necessary for informed treatment decision-making, clinicians worry that patients’ recognition of the terminal nature of their illness may lower psychological well- being. This study examines if such recognition is associated with lowered psychological well-being, that persists over time. Methods: Data came from 87 advanced cancer patients, with a life expectancy of less than 6 months. Patients were assessed pre and post an oncology visit to discuss cancer restaging scan results, and again one month later (follow-up). TIU was assessed at pre and post as the sum of four indicator variables — understanding of terminal nature of illness, curability, stage, and life-expectancy — and a TIU change score was computed (post minus pre). Psychological well-being (psychological symptoms subscale, McGill questionnaire) was assessed at pre, post, and follow-up, and two change scores were computed (post minus pre; follow-up minus post). Results: Changes toward more accurate TIU was associated with a corresponding decline in psychological well-being (r=-0.33, p , .01), but thereafter was associated with subsequent improvements in psychological well-being (r = .40, p , .001). This pattern persisted even after adjustment for relevant demographic factors, prognostic discussion, scan results, and physical well-being change. TIU change scores ranged from positive to negative, with some participants showing improvements in TIU (n = 19), some showing decrements in TIU (n = 14), and others showing stable TIU (n = 54). Among patients with improved TIU, psychological well-being initially decreased, but subsequently recovered [7.03 (2.23) to 6.30 (1.80), to 7.63 (2.08)]; the stable TIU group showed relatively unchanged well-being [7.34 (2.37) to 7.45 (2.32), to 7.36 (2.66)], and the less accurate TIU group showed an initial improvement followed by a subsequent decline [6.30 (2.62) to 7.36 (2.04), to 5.63 (3.40)]. Conclusions: Improved TIU may be associated with initial decrements in psychological well-being, followed by patients rebounding to baseline levels. Concerns about psychological harm may not need to be a deterrent to having prognostic discussions with patients.

11618 Poster Session (Board #310), Mon, 1:15 PM-4:15 PM

After cancer, what’s more important? The survivorship care plan or the survivorship care visit?

Harish Saiganesh, Christine M. Duffy, Camille Higel-Mcgovern, Mary Lorraine Lopresti, Mary Anne Fenton, Rochelle Strenger, Susan F. Korber, Tara Szymanski, Don S. Dizon; Lifespan Cancer Institute, Providence, RI; Lifespan, Providence, RI; Rhode Island Hospital, Providence, RI; Rhode Island Hospital Comprehensive Cancer Center, Providence, RI; Comprehensive Cancer Center of Rhode Island Hospital, Providence, RI; Massachusetts General Hospital, Boston, MA

Background: The Commission on Cancer has made the provision of survivorship care plans (SCPs) a prerequisite for accreditation of cancer programs. However, whether all patients desire or derive benefits from an SCP is not established. This study sought to characterize the provision of survivorship care at the Lifespan Cancer Institute (LCI) to (1) determine clinical and distress characteristics at first visit to the LCI among patients treated with curative intent for a solid tumor diagnosis; (2) characterize which factors are associated with receipt of an SCP with or without a survivorship care visit (SCV); and (3) determine referral patterns and outcomes among patients with an SCP with or without an SCV. Methods: We have retrospectively reviewed 650 patients at Lifespan Cancer Institute so far, all of whom were initially seen between the years 2014-2017. As part of routine practice, all new patients are screened with the NCCN Distress Thermometer (DT). Data related to demographics, treatment variables, and presenting distress score were gathered using our electronic medical records. Categorical data were analyzed using Fisher’s Exact Test or Chi-Square. Multinomial logistic regression was performed for multivariate analysis. All analyses were performed in STATA 15.0. Results: Themedianagewas63(range29to90),74.8% were female, and 39.5% were married or partnered. The major cancers represented included breast (52.2%), lung (16.6%), and prostate (6%) cancers. Severe distress (DT$4) was reported in 49.5% of patients at their first visit with top concerns being worry (54.3%), nervousness (46.6%) and fatigue (37.3%). An SCP was documented in 461 patients (71.1%) and of these, 267 (57.9%) were seen in an SCV as well. On multivariate analysis, gender, marital status, and stage were significantly associated with the receipt of an SCP. However, only disease site was significantly associated with being seen in an SCV. Compared to those with an SCP but not seen in an SCV, those who attended an SCV were significantly more likely (p , 0.005) to be referred for post-treatment evaluation, including psychological counseling (15.7 vs 6.7%), physical therapy (56.9 vs 22.7%), and nutrition (26.6 vs 8.2%). The majority of patients followed up with referrals, though it was significantly higher among those with an SCV (95.4 vs 87%, p = 0.015). Conclusions: Survivorship care varies by cancer type, with highest rates of visits for patients with gynecologic and GI cancers. Referral to other resources is significantly higher among those seen in visits, and those who are referred will follow up.

11619 Poster Session (Board #311), Mon, 1:15 PM-4:15 PM

Lung cancer stigma: A ten-year look at patient and oncologist attitudes about lung cancer.

Jennifer C. King, Eleni Rapsomaniki, Maureen Rigney; Lung Cancer Alliance, Washington, DC; AstraZeneca, Gaithersburg, MD

Background: The presence of lung cancer stigma is well documented and has been shown to impact the care and treatment of lung cancer survivors (Tod et al. 2008; Carter-Harris et al. 2014). In the past decade, there has been considerable research progress in lung cancer but it is unknown if the level of stigma has changed and how that affects patient care. Methods: 205 oncologists who treat lung cancer, 208 patients with lung cancer, and 1001 members of the general public were surveyed with the same survey instrument from a 2008 survey (Weiss et al. 2014) plus 5-15 additional questions at the end. The survey was carried out with identical methodology by phone and online between June 6 and July 26, 2018. Statistical analysis was performed comparing 2008 and 2018 datasets using paired t-tests if normally distributed or Mann-Whitney U tests for continuous data and Chi-squared or Fisher’s exact test for categorical data. Results: In 2018, significantly more oncologists feel they have adequate treatment options for metastatic lung cancer (67% vs 36%, p , 0.001) and the majority of patients report being satisfied with their medical care (87%) and treatment options (71%). Despite these advances, there was a non-significant increase in oncologists indicating both that there is a stigma associated with lung cancer (68% in 2018 vs 60% in 2008) and that patients blame themselves (67% vs 57%). Significantly more patients felt that there was a stigma associated with having lung cancer (70% vs. 54%, p , 0.001). In addition, 57% of oncologists indicated that patients with different cancers are thought about, approached, or handled differently, similar to 2008. Lung cancer patients were most frequently cited as treated differently. In 2018, 40% of patients agreed with the statement “patients with lung cancer are treated differently by doctors and nurses” compared to 26% a decade ago (p = .01). Both groups felt the most common way patients were treated differently was “received less sympathy from medical staff.” Conclusions: After a decade of research progress in lung cancer, stigma surrounding the disease remains a critical problem even in a healthcare setting. Patients are perceiving stigma at higher levels and oncologists are not reporting any improvement. This work underscores the need to address stigma with proactive multilevel approaches including the need for medical providers to practice empathic communication.

11620 Poster Session (Board #312), Mon, 1:15 PM-4:15 PM

Differences among Asian/Asian American, and Caucasian breast and gynecologic cancer patient-reported survivorship needs, symptoms, and illness mindsets (N=220).

Lidia Schapira, Elisa Hofmeister, Allison W. Kurian, Sean Zion, Hanyang Shen, Tara Torres, Jonathan S. Berek, Oxana Palesh; Stanford Cancer Center, Palo Alto, CA; Stanford University, Stanford, CA; Stanford School of Medicine, Stanford, CA; Stanford University, Palo Alto, CA; Stanford Women’s Cancer Center, Stanford, CA; Department of Psychiatry and Behavioral Sciences, Stanford University, Stanford, CA

Background: Cancer experiences are mediated by host and disease factors and affected by social and cultural determinants of health. We sought to characterize perceived supportive care needs and domains of psychosocial functioning among a diverse group of women attending routine appointments for treatment or follow-up of breast and gynecologic cancers. Methods: From July 2018 until January 2019, all patients seen at the Stanford Women’s Cancers Program (serving women treated for breast and gynecologic cancers) were approached to participate. 220 cancer patients (78% breast and 22% gynecologic), mean age 54 (SD = 12) completed an online survey on their supportive care needs assessment (The Short-form Supportive Care Needs Survey Questionnaire) which contains physical, social, information, sexual and psychological scales, anxiety and depression (Patient Health Questionnaire-4) and mindsets about illness and the body mindset (Brief Illness Mindset Inventory). Only 14% of patients refused to take part in the survey. Results: T-test showed that Asian/Asian- American cancer patients (n = 57, 26%) reported fewer needs related to coping with physical symptoms, side effects of treatment and performing usual tasks and activities compared to Caucasian patients (n = 137, 62%), p = .038. Moreover, Asian/Asian-American cancer patients reported fewer depression symptoms (p = .049). Chi-square analyses showed that Asian/Asian-American patients reported less use of psychotherapy over their lifetimes (24.6% vs. 54.0%, p , .001), but significantly higher utilization of Chinese Medicine for cancer symptom management (24.6% vs. 5.8%; p , . 001) compared to the Caucasian patients. T-tests showed that Asian/Asian American patients reported significantly greater agreement with the mindsets that ‘cancer is a catastrophe’ (p = .006) and that their ‘body is an adversary’ (p = .047). Conclusions: In this population of breast and gynecologic patients and survivors we found differences in cancer mindsets, coping and use of mental health and complementary therapies between Asian and Asian-American and Caucasian women. Further analysis to understand the cultural differences are ongoing.

11621 Poster Session (Board #313), Mon, 1:15 PM-4:15 PM

Relationship between perceptions of treatment goals and psychological distress in patients with advanced cancer.

Amanda L. Jankowski, Deborah Anne Forst, Joseph Greer, Ryan David Nipp, Lauren Waldman, Jennifer S. Temel, Areej El-Jawahri; Massachusetts General Hospital, Boston, MA

Background: Several studies have demonstrated discordance between how patients perceive their goal of treatment versus how they perceive their oncologist’s goal. Studies evaluating the extent and risk factors of this discordance are lacking. Methods: We conducted a cross-sectional study of 559 patients with incurable lung, gastrointestinal, breast, and brain cancers. We used the Perception of Treatment and Prognosis Questionnaire to assess patients’ perceptions of both their treatment goal and their oncologist’s goal and categorized responses: 1) patients who reported that both their goal and their oncologist’s goal was concordant (either to cure or not to cure); and 2) patients who reported discordant perceptions of their goal versus their oncologist’s goal. We assessed patients’ psychological distress using the Hospital-Anxiety-and-Depression-Scale and used linear regression to assess the relationship between patients’ perceptions of their treatment goal and psychological outcomes. Results: 61.7% of patients reported that both their goal and their oncologist’s goal was non-curative; 19.3% reported that both their goal and their oncologist’s goal was to cure their cancer; and 19.0% reported discordance between their goal and their perception of the oncologist’s goal. Older age (OR = 0.98, P = 0.01), non- Hispanic ethnicity (OR = 0.31, P = 0.049), and higher education (OR = 0.62, P = 0.042) were associated with lower likelihood of reporting discordant goals. Patients with discordant perceptions of their goal and their oncologist’s goal reported higher anxiety (B = 1.56, P = 0.003) compared to those who reported that both their goal and their oncologist’s goal was curative. Patients who reported both their goal and the oncologist’s goal was non-curative had higher depression symptoms (B = 1.06, P = 0.013) compared to those who reported that both their goal and the oncologists’ goal was curative. Conclusions: One-fifth of patients with advanced cancer report discrepancies between their perceptions of their own and their oncologists’ treatment goal which is associated with psychological distress. Tools are needed to identify patients at risk of cognitive dissonance about their prognosis.

11622 Poster Session (Board #314), Mon, 1:15 PM-4:15 PM

The individualized Goals of Care Discussion Guide: A simple tool to empower patients with metastatic breast cancer.

Jeffrey M. Peppercorn, Yvonne Y. Lei, Nora Horick, Katharine M. Quain, Don S. Dizon, Rachel Jimenez, Jennifer Adrienne Shin, Karen Sepucha, Giselle Katiria Perez; Massachusetts General Hospital, Boston, MA

Background: Individualized treatment planning is a critical part of quality cancer care, but how best to achieve this for patients with metastatic breast cancer (MBC) is unclear. We evaluated the feasibility, acceptability and impact of using a simple and scalable “Individualized Goals of Care Discussion Guide” (IGCDG) to facilitate patient-provider communication at the time of treatment decisions. Methods: We developed the IGCDG based on structured interviews with MBC patients and input from experts in cancer care, decision sciences, psychology and palliative care. We then conducted a single arm feasibility trial among patients with newly diagnosed or progressive MBC. Prior to clinic, patients received the IGCDG, an 8-page MBC informational brochure and 1-page questionnaire regarding treatment preferences, personal goals and priorities for care planning. The completed questionnaire was provided to the oncology team at the patient’s visit. Pre and post assessment included the Distress Thermometer (DT), Patient Satisfaction with Cancer Care Scale and the Control Preferences Scale. Feasibility was defined as: 1) accrual of . 50%, 2) attrition rate , 32%, and 3) , 50% of patients experiencing increased distress following the intervention. Results: Among 60 eligible patients, 42 participated (70% accrual), 40 completed all surveys (2% attrition), and only 7 (18%) reported increased distress. Mean age was 57 (range 31 – 79), 85% were white, 7% black, 5% Hispanic, 66% were college graduates, and 40% reported high baseline distress (DT . 4). Patient priorities for discussion included cancer directed therapy (70%), symptom management (70%), and prognosis/planning ahead (60%). At 2-month follow-up, 53% reported decreased distress compared to baseline. Satisfaction with cancer care was high at baseline and follow-up. Most patients preferred shared decision making (77%), and 79% reported decision roles concordant with preferences. Overall, 72% of participants found the IGCDG helpful, 93% found the questionnaire easy to complete, and 44% felt it improved communication with their doctor (49% unsure). Conclusions: Administration of the Individualized Goals of Care Discussion Guide is feasible and provides patients with MBC an opportunity to define their goals of care and priorities for discussion in clinic. Clinical trial information: NCT03375827.

11623 Poster Session (Board #315), Mon, 1:15 PM-4:15 PM

Depression and anxiety symptoms in bereaved caregivers of patients with advanced cancer.

Olivia Vanbenschoten, Joseph Greer, Vicki Jackson, Mihir Kamdar, Emily R. Gallagher, Maya V. Anand, Jennifer S. Temel, Areej El-Jawahri; Massachusetts General Hospital, Boston, MA

Background: Caregivers of patients with advanced cancer experience substantial caregiving burden and psychological distress during the illness course. However, data on depression and anxiety symptoms in bereaved caregivers and factors associated with their psychological distress are lacking. Methods: We conducted a secondary analysis of 168 caregivers enrolled in a randomized trial of early palliative care integrated with oncology care versus oncology care alone for patients newly-diagnosed with incurable lung and non-colorectal gastrointestinal cancers and their caregivers who completed bereavement assessments at 3 months after their loved one’s death. We used the Hospital Anxiety and Depression Scale (HADS) to assess patients’ and caregivers’ depression and anxiety symptoms at baseline within 8 weeks of diagnosis, and at 3-4 months after the patient’s death (for caregivers). We asked caregivers to rate patient’s physical and psychological distress in the last week of life on a 10-point scale. We used linear regression adjusting for randomization and cancer type to explore associations between patient and caregiver factors and bereaved caregivers’ depression and anxiety. Results: 30.4% (51/168) and 43.4% (73/168) of bereaved caregivers reported clinically significant depression and anxiety symptoms, respectively. Younger patient age (B = -0.06, P = 0.041), higher patient baseline anxiety (B = 0.28, P = 0.002), and caregiver rating of worse physical (B = 0.28, P = 0.035) and psychological (B = 0.41, P , 0.001) distress experienced by the patient at the end of life (EOL) were associated with worse depression symptoms in bereaved caregivers. Only caregiver factors, including age (B = -0.07, P = 0.004), female sex (B = 1.60, P = 0.024), and rating of worse psychological distress experienced by the patient at the EOL (B = 0.42, P , 0.001) were associated with worse bereaved caregivers’ anxiety symptoms. Conclusions: Bereaved caregivers of patients with advanced cancer experience substantial psychological distress which is associated with their perceptions of their loved one’s distress at the EOL. Interventions to optimize EOL care for patients and reduce bereaved caregivers’ psychological distress are needed.

11624 Poster Session (Board #316), Mon, 1:15 PM-4:15 PM

Effectiveness of psychological distress reduction with cognitive behavioral therapy for oncological patients: A one-year follow-up study.

Alessandro Rossi, Maria Marconi, Stefania Mannarini, India Minelli, Monica Anderboni, Chiara Rossini, Giuseppe Di Lucca, Claudio Verusio; Department of Philosophy, Sociology, Education, and Applied Psychology, Section of Applied Psychology, University of Padova, Padova, Italy; Department of Medical Oncology, ASST Valle Olona, Presidio Ospedaliero di Saronno, Saronno, Italy; Department of Medi- caDepartment of Philosophy, Sociology, Education, and Applied Psychology, Section of Applied Psychology, University of Padova, Padova, Italy; Department of Medical Oncology, Presidio Ospedaliero di Saronno, ASST Valle Olona, Saronno, Italy; Ospedale Civile Di Saronno, Lissone, Italy; Department of Oncology, ASST Valle Olona, Busto Arsizio, Italy

Background: Distress has a negative impact on medical treatment (Di Matteo, Lepper & Croghan, 2006) and it is considered one of the most important indexes of psychological suffering in oncological patients (NCCN, 2015). Thus, the purpose of this study was to determine the long term effectiveness of brief Cognitive Behavioral Therapy for patients with cancer (CBT-C) compared with a control group (CG) of oncological patients without any psychotherapy intervention - at one year after a chemotherapy treatment. Methods: Participants (n = 80; mean age = 63.3, SD = 13.4; 54 female) enrolled at the Oncology Day Hospital at the “Presidio Ospedaliero” of Saronno, ASST Valle Olona, Italy who undertook (CBT-C: n =40) or non-undertook (CG: n = 40) a psychotherapy intervention. Individual psychotherapy sessions strictly followed the IPOS guidelines (Watson & Kissane, 2017). Participants were tested with the Psychological distress Inventory (PDI) at the baseline (T1; Cronbach a =.88)attheendofthechemotherapytreatment (T2; Cronbach a = .87), at the end of the psychotherapy intervention (T3; Cronbach a = .88), 6-month follow-up (T4; Cronbach a = .85), and 1-year follow-up (T5; Cronbach a =.84).Results: Multilevel growth curve modeling – controlling for age, number of sessions, type and localization of tumor – showed a sharper reduction of distress for CBT-C participants that continue after posttreatment until 1-year follow-up (p , .001); whereas for CG participants it reduced more gradually from pretreatment to 1-year follow-up (p , .001). The results revealed a significant difference between the linear slopes for each treatment condition (p , .001). The overall Hedges’ g comparing the two groups for distress reduction between pretreatment and 1-year follow-up was 2.14 (p , .001) in favor of CBT-C. Conclusions: Given that psychological distress occurs frequently among oncological patients this study is into an important area of study. Results suggest that CBT-C is statistically and clinically effective in treating psychological distress 1 year after the chemotherapy treatment. These findings revealed a kind of long-term effectiveness psychological intervention able both to reduce psychological suffering and improve a better quality of life in oncological settings.

11625 Poster Session (Board #317), Mon, 1:15 PM-4:15 PM

Evaluating the effects of a structured exercise intervention on physical self-worth in men with prostate cancer: Addressing an unmet need.

Richard Francis Dunne, Charles E. Heckler, Julia Ellen Inglis, Po-Ju Lin, Chunkit Fung, Luke Joseph Peppone, Gilberto Lopez, Eva Culakova, Ian Kleckner, Michelle Christine Janelsins, Aminah Jatoi, Supriya Gupta Mohile, Karen Michelle Mustian; University of Rochester James P. Wilmot Cancer Institute, Strong Memorial Hospital, Rochester, NY; University of Rochester Medical Center, Rochester, NY; Mayo Clinic, Rochester, MN

Background: Improving body image and self-esteem are top ASCO priorities in the survivorship care of men with prostate cancer (PCa). Body image and global self-esteem, influenced by physical self-worth, are negatively affected by PCa treatment. We investigate whether exercise can improve physical self- worth in men treated for PCa and if improving self-worth is associated with changes in quality of life (QoL) and mental health. Methods: We performed a secondary analysis of a phase II randomized controlled trial comparing the effects of Exercise for Cancer Patients(EXCAP), a structured, 6-week, home-based exercise intervention, to usual care (UC) in men with non-metastatic PCa receiving radiation or Androgen Deprivation Therapy (ADT). The Physical Self-Perception Profile (PSPP), a valid 30-item questionnaire where higher scores indicate greater physical self-worth, was assessed at pre- and post-intervention. Changes between arms were compared using ANCOVA. Spearman correlations were calculated for pre/post-intervention change scores for PSPP and QoL, depression, and anxiety as measured by the Functional Assessment of Cancer Therapy (FACT), Center of Epidemiologic Studies Depression (CES-D) scale, and State-Trait Anxiety Inventory (STAI), respectively. Results: Fifty-eight men were randomized; average age was 67.1 years. Physical self-worth at baseline moderated the effect of the intervention. Compared to UC, EXCAP improved physical self-worth in those with baseline PSPP scores above the median (p , 0.04). Exercisers with baseline PSPP scores in the top quartile demonstrated a more significant improvement over UC (p , 0.01). Improvements in physical self-worth were associated with improved QoL (r = 0.29, p = 0.04), depression (r = -0.28, p = 0.04) and anxiety (r = -0.30, p = 0.03). Conclusions: Exercise significantly improves physical self-worth in men with PCa on radiation or ADT, and greater physical self-worth is associated with improved QoL, depression and anxiety. Those with higher baseline physical self-worth derived the most benefit from exercise. Exercise should be prescribed to boost self-esteem and body image in men receiving radiation or ADT for PCa. Clinical trial information: NCT00815672.

11626 Poster Session (Board #318), Mon, 1:15 PM-4:15 PM

The importance of altruism to biomarker development for pancreatic cancer.

Danny HJ Heo, Corinne DeCicco, Tori Singer, Vijetha Vemulapalli, Christopher Yee, Michael A. Kiebish, Eric M. Grund, Viatcheslav R Akmaev, Niven R. Narain, Kevin F. Kennedy, A. James Moser; Beth Israel Deaconess Medical Center, Boston, MA; BERG, LLC, Framingham, MA; Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA

Background: Longitudinal donation of tissue and health information by patients with pancreatic ductal adenocarcinoma (PDAC) is critical to validating algorithms for precision treatment. Unlike therapeutic trials offering potential survival benefit, factors motivating long-term participation in biomarker studies are poorly understood. We hypothesized that long-term participation (PART) depends on identifiable factors motivating study volunteers in non-therapeutic biomarker studies. Methods: A prospective, single-institution biomarker study was screened to identify participants with, or at elevated risk for, PDAC and periampullary cancer (Jul 2015-Dec 2018). Study consent precluded individual receipt of value or research benefit. Detailed biospecimen and clinical data (n = 294) were analyzed using multivariable modeling and Bayesian AI (bAIcis) to characterize motivators of PART. Results: Of 294 participants, 185 had PDAC (63%); 89 had premalignant lesions (30%) or tumors (20; 7%) mimicking PDAC. Mean age was 68 years, 54% male, 26% with prior cancers, and 18% family PDAC history. Treatment was not indicated (37% no cancer), potentially curative (27%), non-curative (35%); or refused (1.4%). Bayesian AI showed PDAC diagnosis to be the leading predictor of PART. Using multivariable modeling and adjusting for the competing risk of death, none of the following domains predicted PART: demographics, perceived personal/family cancer risk, health insurance type, median income, housing price, travel time/distance or charitable expense, or primary institution providing clinical treatment. The following motivators were significant predictors of PART: worsening symptoms (McGill Brisbane . 9; HR 0.54; p = 0.045), non-curative treatment (HR 0.55; p = 0.042); and shorter AJCC predicted survival (HR 0.77, p = 0.042). Conclusions: Volunteers with symptomatic PDAC and shorter predicted survival who received non-curative treatment were more motivated to participate in biomarker discovery. Volunteers were undeterred by apparent self-interests such as travel time, insurance type, income, charitable expense, and personal life expectancy. We attribute these findings to altruistic behaviors among patients most likely to die of PDAC.

11627 Poster Session (Board #319), Mon, 1:15 PM-4:15 PM

The psychosocial and financial burden on caregivers of young-onset colorectal cancer patients.

Kim Lynn Newcomer, N2Y Board, Ronit Yarden; Colorectal Cancer Alliance, Washington, DC

Background: The Colorectal Cancer Alliance launched a survey for caregivers of young-onset colorectal cancer (CRC) patients and survivors. A caregiver of young-onset CRC is an unpaid or paid member of a person’s social network who helps them with activities of daily living. Methods: Caregivers who participated in our survey (n = 427) are diverse in age, gender, and race and ethnicity, however, the majority of respondents were between the ages of 30-39. Results: The caregivers reported they faced many challenges including insufficient psychosocial (66%) and financial support (43%) to effectively care for their patients, especially as many (67%) of the caregivers are parents who juggled work and raising young children in addition to caring for the patient. Caregivers indicated a lack of sufficient resources and assistance with transportation to treatment, doctors visits, child care,and household maintenance. Overall, about one in three caregivers reported they were missing three days or more or more of work each month to care for their loved one. The majority of caregivers (59%) reported that their loved one experienced changes in their ability to perform expected social tasks, including those of a spouse, child rearer, friend, or worker. Some caregivers also mentioned the loss of sexuality, depression, pain, despair, lack of sleep, sadness, and stress, along with a loss of faith and hope, which may cause additional strain on their relationships. A caregiver stated, "It changed my role as a daughter, to daughter and caregiver. Friends became distant, relationships have been lost." Conclusions: Our survey results indicate that caregivers need additional resources to manage everyday tasks. With additional tools and support, a caregiver may be able to devote a little more time to self-care, which may alleviate some of the psychosocial burdens of caregiving, especially, when caring for a young person with particular obligations to work and family. Finally, caregivers noted a limited amount of information on prevention and surveillance are available to family members. The Colorectal Cancer Alliance uses these survey results to understand the self-reported medical, psychosocial, and quality of life experiences of this often overlooked group.

11628 Poster Session (Board #320), Mon, 1:15 PM-4:15 PM

Till death do us part: Existential loneliness (EL), psychosocial distress, and survival of advanced cancer patients (ACPs), and their spousal caregivers (SCs).

Fay J. Hlubocky, Tamara Sher, David Cella, Mark J. Ratain, Christopher Daugherty; University of Chicago Medicine, Chicago, IL; Northwestern University Family Institute, Chicago, IL; Robert H. Lurie Comprehensive Cancer Center, Northwestern University, Chicago, IL; University of Chicago, Chicago, IL

Background: Terminally-ill ACP often experience existential suffering as they near the end-of-life. However, the specific impact of EL on ACP and SC psychosocial wellness and survival has yet to be identified. Methods: A prospective ACP cohort enrolled in phase I trials was assessed at baseline (T1) and at one month (T2) utilizing psychosocial measures: state/trait anxiety (STAI-S/T), depression (CES- D), quality of life/QOL (FACIT-Pal), and global health (SF-36). Semi-structured interviews evaluated the ACP-SC EL experience: cognitive awareness (e.g. death, terminal prognosis); and, emotions (physical isolation; emptiness; abandonment). Results: To date, 160 participants (80 Phase I ACP and 80 SC) have been separately interviewed at T1 and T2. Total population demographics include: median age 62 (28-78y); 50% male; 100% married; 89% Ca; 68% . HS educ; 58% GI dx; ACP median survival 8.3 months (.51-18.9) 55% income , $65,000 yr. At T1, 77% of ACP acknowledged a cognitive awareness of death, 72% reported physical isolation; 54% felt emotional emptiness. For SC at T1, 65% recognized a cognitive awareness of death; 89% felt isolated; and 82% sensed emotional abandonment. Over time, rates of EL remained consistent for ACP-SC with the exception of increased self- reported isolation at 79% and 92% respectively at T2. At T2, ACP with death cognitions had higher STAI-S (32610 v. 30611, p = 0.03) and CES-D scores (13612 v. 11610, p = 0.02). SC with self- reported abandonment had higher STAI-S anxiety (39617 v. 35613, p = 0.03) at T2. Regression analyses revealed ACP with EL death cognitions had poorer FACIT-Pal QOL over time. Moreover, SC with self-reported physical isolation at T2 was negatively associated with SF-36 scores. ACP with EL had shorter overall survival compared to ACP without EL (4.1 v. 6.9 months, p = 0.02). ACP and SC qualitative inquiry re EL exposed unique themes: difficulty articulating EL experience; acceptance of death; finding meaning within crisis. Conclusions: EL is negatively associated with QOL for ACP participating in Phase I trials. Supportive couples-based, dyadic psychological interventions at the end- of-life to assist with coping are warranted.

TPS11629 Poster Session (Board #321a), Mon, 1:15 PM-4:15 PM

Improve: A community-based physical activity intervention to improve functional and health outcomes in older breast cancer survivors: Rationale, design, and methods.

Cynthia Owusu, Nora Nock, Paul Hergenroeder, Kris Austin, Beth Bennett, Stephen Cerne, Halle C. F. Moore, Jean Petkac, Mark D. Schluchter, KATHRYN H. SCHMITZ, Lindsay Atkins, Oghenerukeme Asagba, Leonard Wimbley, Nathan A. Berger; Case Western Reserve University School of Medicine, Cleveland, OH; Metrohealth Medical Center, Cleveland, OH; The Gathering Place, Beachwood, OH; Cleveland Clinic, Cleveland, OH; University Hospitals of Cleveland, Cleveland, OH; Case Western Reserve University, Cleveland, OH; University of Pennsylvania, Philadelphia, PA; Case Western Reserve Univ, Cleveland, OH

Background: African-Americans (AA) and lower socioeconomic status (SES) older breast cancer survivors (BCS) are more likely to experience poor functional and health outcomes. Promotion of healthy behaviors, such as physical activity (PA), is critical to addressing these health disparities. Here we describe the rationale, design and methods of a randomized controlled trial testing the effectiveness of a physical activity intervention among older BCS from diverse racial and SES backgrounds. Methods: The IMPROVE Study is a community-based randomized-controlled trial designed to recruit 320 BCS, 80 in each of four strata defined by race (AA vs. Non-Hispanic Whites [NHW]) and SES (low vs. high). Participants are aged $ 65 years, AA or NHW and are within five years from treatment completion for stage I-III breast cancer. Participants are recruited utilizing the Ohio Cancer Incidence and Surveillance System database or directly from three area hospitals in Cleveland, Ohio and randomized to one of two arms: a 52-week moderate intensity aerobic and resistance group training intervention (n = 160) versus attention-control (support group sessions), (n = 160). The first 20 weeks of the PA intervention includes 3x per week 60- minute supervised sessions. The last 32 weeks of the PA intervention are unsupervised. Each of the 60- minute supervised PA sessions include 30 minutes of moderate intensity aerobic activity at 50%-70% of HRmax (maximum heart rate) and 30 minutes of resistance training based on 1-RM (repetition maximum) for chest and leg press. The attention-control group attend a once per week 60-minute support group session for the first 20 weeks and have unsupervised group sessions during the last 32 weeks. Exit interviews are being conducted at 52 weeks. The primary outcome is change in Short Physical Performance Battery (SPPB) Scores at 20 weeks. Secondary outcomes include change in SPPB scores at 52 weeks, and change in body composition and biomarkers of breast cancer prognosis at 20 and 52 weeks. One hundred and seventy-four participants have been enrolled as of 02/07/2019. Discussion: This study includes three underserved populations, (older BCS in general, older AA BCS and older low SES BCS) in one study. Results may contribute to a better understanding of factors associated with recruitment, sustained participation and acceptability, and will inform physical activity programs that will optimally improve the functional and health outcomes for older women during breast cancer survivorship. Clinical trial information: CNCT02763228.

TPS11630 Poster Session (Board #321b), Mon, 1:15 PM-4:15 PM

Implementation of complex perioperative intervention in older patients with cancer (IMPROVED program).

Elena Paillaud, Frederique Peschaud, Tristan Cudennec, Philippe Caillet, Mathilde Gisselbrecht, Laure De Decker, Nicola De Angelis, Pierre Cattan, Benoit Plaud, Christophe Tournigand, Thomas Aparicio, Sandrine Touzet, David Moszkowicz, Marie Laurent, Florence Canoui-Poitrine, Etienne Brain; Hopital Europeen Geoges Pompidou, Paris, France; Hospital Ambroise Pare, AP-HP, University Versailles-SQY, Boulogne-Billancourt, France; Hopitalˆ Ambroise Pare, ´ APHP, Boulogne- Billancourt, France; Hopitalˆ Europeen ´ Georges-Pompidou, Paris, France; EA 1156-12, Nantes University Hospital, Nantes, France; Hopital Henri Mondor, Creteil, ´ France; Hopital Saint Louis, Paris, France; Hopitaux Universitaires Henri Mondor, AP-HP, Creteil, ´ France; Department of Gastro- enterology, Saint Louis Hospital, Paris, France; CHU Lyon, Lyon, France; Hopital Amboise Pare, ´ Boulogne Billancourt, France; EA 7376 CEpiA (Clinical Epidemiology and Ageing Unit), Creteil, ´ France; Institut Curie, Paris & Saint-Cloud, France

Background: Nearly 50% of patients are older than 70 years at diagnosis of digestive cancer. Surgical resection is the first line strategy of treatment. Despite improvement in surgical techniques and development of rehabilitation programs, the rate of postoperative complications remains high. Peri-operative involvement of geriatricians may improve care management older cancer patients. Methods: During a 6 months run-up period (emerging project), we structured a multi-professional network (digestive surgeons, anesthetists, geriatricians, digestive oncologists, epidemiologists), we elaborated a innovative peri-operative geriatric intervention (Improved program) in digestive surgery setting based on evidence- based data. We build a dedicated evaluation plan by determinate the best design for assessing geriatric intervention in this complex context and choose the more appropriate endpoints. Results: We will include 554 patients aged 75 or more with resectable digestive cancer in a stepped wedge cluster randomized trial. The intervention is based on 1/ a preoperative geriatric assessment, focusing on frailty parameters and developing a coordinated program of tailored geriatric interventions 2/ a postoperative shared care with an integrated care model where both surgeon and geriatrician share responsibility for the patient management in surgical ward. This geriatric postoperative management will be focus on prevention and correction of complications, early mobilization, optimal nutritional support. The main endpoint Is is Grade II or higher post-surgical complications rate according Clavien-Dindo classification within 30 days after the surgical procedure. Conclusion: We expected to demonstrate a benefit of a peri-operative shared management model to decrease the risk of post-surgical complications In older patients with digestive cancer.

TPS11631 Poster Session (Board #322a), Mon, 1:15 PM-4:15 PM eMouvoir: Randomized study estimating the impact of a personalized and remote support centered on physical activity (PA) for patients (pts) after breast cancer (BC).

Laurence Vanlemmens, Amelie Anota, Emilie Bogart, Virginie Nerich, Didier Cauchois, Alain Dewitte, Etienne Dormeuil, Eric Lartigau, Flora Le Gall, Herve ´ Mocaer, Aurelie Thenot, Marc-Antoine Brochard, Marie-Cecile Le Deley; Centre Oscar Lambret, Lille, France; Methodology and Quality of Life Unit, Department of Oncology, INSERM UMR 1098, University Hospital of Besancon; French National Platform Quality of Life and Cancer, Besançon, France; Institut Regional ´ Federatif de Cancerologie, Besançon, France; Stimulab, Paris, France; Oscar Lambret Comprehensive Cancer Center, Lille, France

Background: BC pts can reduce their health-related quality of life (HRQoL) encompassing physical, psychological and social components due to cancer and treatments (trts). Despite the strong evidence of the beneficial effects of EPA for BC pts on HRQoL, the cancer pts’ EPA levels most often decline after diagnosis and trts due to physical and psychological components, accessibility to exercise programs, time constraints. Connected watches can now contribute to pts’ deeper commitment to their program by providing reliable trends of their EPA metrics. Notwithstanding this recent progress, cancer pts keep facing challenges to maintain a regular EPA. We aim at evaluating a holistic intervention including a physical activity educator, coaching pts remotely for 4 months (mo), to improve HRQoL after BC. Methods: eMouvoir is a multicenter randomized, controlled phase 3 trial started 4-6 mo after the end of trt among non-metastatic BC pts. It assesses the benefit of a personalized remote EPA coaching, including a connected watch, access to a digital platform, personalized objectives, $2 messages per week, weekly lessons, for a 4-mo duration, compared to the standard supportive approach (recommendations for EPA made during visits with the oncologist) in terms of HRQoL. The randomization is balanced 1:1 and controlled for HRQoL at inclusion, trts, exercise practice, age, access to the Internet and center. Main eligibility criteria are: age $18; adjuvant or neoadjuvant trt for non-metastatic BC; medical certificate for sports practice. Patients without a connected device are eligible. HRQoL is evaluated using the SF-36 questionnaire, at 4, 8 and 12 mo. Both components (physical and mental summary) at 12 mo are used as co-primary endpoints. The study includes a health-economics evaluation, using EQ-5D and EQ VAS, to estimate the incremental cost-utility ratio. Based on the following assumptions: expected mean difference for each SF-36 component at 12 mo =3 points, standard deviation=15, power=90%, 2-sided alpha=2.5%, 10% drop-outs, 1242 evaluable pts are required, leading to 1380 pts. The trial is funded by the French ministry of health and should open enrollment soon.

TPS11632 Poster Session (Board #322b), Mon, 1:15 PM-4:15 PM

Physical activity platform to improve bone health in cancer survivors.

Cathy Skinner; Thrivors Inc., Saint Paul, MN

Background: Cancer treatment-induced bone loss and the subsequent risk of fractures in both men and women is a significant burden on national health care. US cancer survivors will reach 18 million by 2022 and by 2025 and the projected burden of osteoporosis is $25.3B. Cancer patients treated with medications that lower hormone levels face an increased risk of fractures. Clinical studies report that moderate-intensity resistance exercises prevent a decline in bone health in female cancer survivors, and that strength training may reduce complications such as fatigue, muscle wasting, and bone loss. Unfortunately, only a small percentage of female cancer survivors engage in and adhere to a strength training exercise routine. Several factors account for this low level of engagement such as poor awareness of the benefits of exercise, lack of motivation, and the lack of affordable and easily accessible exercise programs designed for cancer survivors. Internet-based self-managed programs of exercises and guidance can potentially address these problems, and can be scaled for broader dissemination. Methods: The Thrivors platform seeks to shift the current paradigm of poor adherence to exercise by providing an affordable and accessible means of improving outcomes for cancer survivors. Thrivors plus Bone Health (Thrivors+BH), will be a first-of-its-kind offering of clinically validated bone health exercises that can be customized to a cancer survivor’s pain and energy levels, with several novel interactive feedback features. The product has supportive modules for connecting with a community of other users, mindfulness training, nutritional advice and educational information. The Thrivors platform is scalable, device-agnostic, and allows analysis of de-identified usage data. The two versions of the platform will be deployed in a 20-week, two-arm randomized controlled trial of 50 breast cancer survivors. Adherence to exercise sessions (primary outcome) and platform usage will be measured by platform software, while user engagement (secondary outcomes) will be assessed by surveys. Differ- ences between primary and secondary outcomes for the control and intervention groups will be evaluated. Usage statistics, motivational responses and user feedback will be used to further refine platform features and content. Clinical trial information: NCT03651037.

TPS11633 Poster Session (Board #323a), Mon, 1:15 PM-4:15 PM

A randomized trial of a healthy lifestyle intervention versus usual care on chemotherapy and endocrine therapy adherence rate in women with breast cancer: The Lifestyle Exercise and Nutrition Early After Diagnosis (LEANER) Study.

Tara B. Sanft, Maura Harrigan, Brenda Cartmel, Leah M. Ferrucci, Karen Basen-Engquist, Dawn L. Hershman, Jennifer A. Ligibel, Marian Neuhouser, Anees B. Chagpar, Beth A. Jones, M. Tish Knobf, Andrea Silber, Fangyong Li, Melinda L. Irwin; Yale School of Medicine, New Haven, CT; Yale School of Public Health, New Haven, CT; Yale University School of Public Health, New Haven, CT; The University of Texas MD Anderson Cancer Center, Houston, TX; Columbia University Medical Center, New York, NY; Dana-Farber Cancer Institute, Boston, MA; Fred Hutchison Cancer Center, Seattle, WA; Yale Cancer Center, New Haven, CT; Yale, New Haven, CT; Yale Univ School Nursing, Orange, CT; Yale Center for Analytical Sciences, New Haven, CT

Background: The World Cancer Research Fund and the American Cancer Society provide diet and exercise guidelines for cancer survivors. Many women with breast cancer do not follow these guidelines. Adoption of recommended lifestyle behaviors soon after diagnosis may prevent adverse changes in body composition, breast cancer biomarkers, and may improve adherence to treatment thereby improving breast cancer prognosis. The Lifestyle, Exercise, and Nutrition Early after Diagnosis (LEANER) study is testing the impact of a healthy lifestyle intervention on chemotherapy completion and endocrine therapy adherence. Secondary endpoints include changes in inflammatory and metabolic biomarkers, body composition, and patient reported outcomes. Methods: Eligible participants are women with stage I-III breast cancer undergoing chemotherapy. 250 participants are being recruited and randomized 1:1 to a yearlong, 16 session, healthy diet and exercise counseling intervention or usual care group. The intervention is delivered in person and/or via telephone by registered dietitians with training in exercise science. Materials include workbooks, videos, cookbooks, fit bits, and home-based exercise equipment. Intervention is focused on graduated goal setting to meet the recommended diet and exercise guidelines for cancer survivors. The primary endpoint, chemotherapy completion rate, is gathered from the Electronic Medical Record and the average Relative Dose Intensity for the originally planned regimen is calculated based on standard formulas. Assessments are completed at baseline, post chemotherapy, 1- and 2- year time points (to assess adherence to endocrine therapy). Body composition is measured using dual energy X-ray absorptiometry, blood samples and patient reported outcomes are collected. At time of submission, 39 women have been randomized. Discussion: If successful, this study has the potential to make healthy lifestyle interventions initiated shortly after diagnosis a standard component of breast cancer treatment. Clinical trial information: NCT03314688.

TPS11634 Poster Session (Board #323b), Mon, 1:15 PM-4:15 PM

Understanding and predicting fatigue, cardiovascular (CV) decline, and events after breast cancer treatment (UPBEAT): A prospective cardio-oncology study.

Kerryn Reding, Ralph D’Agostino, Glenn Jay Lesser, Heidi D. Klepin, Lynne I. Wagner, Dalane Kitzman, Peter H. Brubaker, Shannon Mihalko, Jennifer Jordan, Nancy E. Avis, Karen Marie Winkfield, Bonnie Ky, Susan Faye Dent, Teresa Crotts, Tonya Calhoun, Kimberly Lane, W. Gregory Hundley; Fred Hutchinson Cancer Research Center & University of Washington, School of Nursing, Seattle, WA; Wake Forest Baptist Health, Winston-Salem, NC; Wake Forest Baptist Medical Center, Winston-Salem, NC; Comprehensive Cancer Center, Wake Forest Baptist Health, Winston Salem, NC; Northwestern University Feinberg School of Medicine, Chicago, IL; Wake Forest School of Medicine, Winston-Salem, NC; Wake Forest University, Winston-Salem, NC; Virgnia Commonwealth University, Richmond, VA; Massachusetts General Hospital, Boston, MA; Hospital of the University of Pennsylvania, Philadelphia, PA; The Ottawa Hospital Cancer Centre, Ottawa, ON, Canada; Virginia Commonwealth University, Richmond, VA; Wake Forest School of Medicine, Winston Salem, NC

Background: Modern treatment for breast cancer (BC) has led to improved survival; however, this improvement can be offset by an increase in cancer therapy-related morbidity and mortality. Over one-third of early stage BC patients treated with cancer therapy experience CV injury, left ventricular (LV) dysfunction, exercise intolerance, or fatigue. CV disease is a leading cause of mortality in BC survivors. There is limited information on the time course and long-term CV health of BC survivors. UPBEAT, a multicenter study, will prospectively evaluate CV risk factors and outcomes in early stage BC patients, treated with modern cancer therapies. This will facilitate evaluation of primary CV prevention strategies in this patient population. Methods: This is a prospective cohort study of 840 patients with early stage (I-III) BC treated with chemotherapy +/- radiation and 160 controls. Baseline and serial longitudinal measures will examine the influence of cancer treatment on CV function, exercise capacity and fatigue, and future development of CV events. The comprehensive assessment of factors includes ascertainment of cardiac biomarkers, CV risk factors, comorbidities, functional status (e.g., disability measures, Expanded SPPB), neurocognitive tests, behavioral risk factors, socio-demographics, and quality of life at baseline, 3-, 12-, and 24-mos. Outcomes measured at the same time points, include a deep phenotyping of CV dysfunction (via cardiac MRI assessing LV end diastolic volume, LV end systolic volume, LV ejection fraction, myocardial strain, strain rate, left atrial volumes and mass, and aortic stiffness), exercise intolerance (submaximal as 6-minute walk test and maximal as VO2peak via cardiopulmonary exercise test), fatigue (via FACT-F). Eligibility criteria are: age .18 years; ECOG 0-2, able to walk without symptoms; and for BC patients, treatment with chemotherapy. 143 participants are accrued and currently enrolling through ECOG and NCORP sites. Participants will be followed for 9 years with active surveillance of CV events, i.e., heart failure, myocardial infarction, stroke, all-cause and CV death. Clinical trial information: NCT02791581.

TPS11635 Poster Session (Board #324a), Mon, 1:15 PM-4:15 PM

Patient controlled analgesia (PCA) versus non-PCA intravenous hydromorphone titration for severe cancer pain: A randomized, controlled, multicenter, phase III trial, HMORCT09-1.

Rongbo Lin, Jinfeng Zhu, Shuitu Feng, Sunzhi Lin, Jianqian Fu, Yongzhi Yao, Lixia Hong, Min Xin, Zhangshu Chen, Zhenyu Cai, Chunxiang Li, Shen Zhao, Qingyi Wu, Shangwang Yang, Cheng Huang; Gastrointestinal Medical Oncology, Fujian Cancer Hospital., Fuzhou, China; Quanzhou First Hospital, Quanzhou, China; Xiamen Haicang Hospital, Xiamen, China; Mindong Hospital of Ningde City, Ningde, China; Xiamen Fifth Hospital, Xiamen, China; Putian People Hospital, Putian, China; Hanjiang Hospital of Putian City, Putian, China; Fuzhou Armed Police Hospital, Fuzhou, China; The First Affiliated Hospital of Xiamen University, Xiamen, China; The Second Affiliated Hospital of Fujian Medical University, Quanzhou, China; Department of Gastrointestinal Medical Oncology, Fujian Cancer Hospital, Fujian Medical University Cancer Hospital, Fuzhou, China; Guangqian Hospital of Quanzhou City, Quanzhou, China; Rehabilitation Hospital Affiliated to Fujian University of Traditional Chinese Medicine, Fuzhou, China; Fujian Cancer Hospital, The Affiliated Cancer Hospital of Fujian Medical University, Fuzhou, China

Background: The opioid dose for an individual with cancer pain to provide adequate relief of pain with an acceptable degree of side effects is variable. Opioid titration is a process to obtain the tailored dose. Conventional titration is administered by a clinician or nurse. PCA is that patients control cancer pain by self-administration of intravenous opioids using programmable pump. The aim of our study is to evaluate the efficacy of PCA titration versus conventional titration intravenously for severe cancer pain (10-point numerical rating scale, NRS $ 7). Injectable Hydromorphone was selected as pharmaceu- tical analgesics, which works as well as morphine and oxycodone and had similar side effects. Methods: This is currently enrolling patients (n=230) with severe cancer pain during previous 24 hours. Patients are randomized 1:1 and stratified by opioid intolerance or opioid tolerance into PCA or non-PCA titration. PCA titration using programmable pump: bolus hydromorphone at 0.5mg (for opioid intolerance) or hydromorphone dose equivalent to 10% to 20% of the total opioid taken in the previous 24 hours with a lockout time 15 minutes (for opioid tolerance) was administered by the patients educated. No basal infusion was set in the pump. Non-PCA titration administered by a nurse or clinician: Initial hydromorphone doses were same with PCA titration. Reassess pain at 15 minutes. Increased dose of hydromorphone by 50%-100% if pain unchanged or increased, or repeat same dose if pain decreased to NRS 4-6, or continue at current effective dose as needed over initial 24 hours. The primary endpoint is the time needed to successful titration was defined the time from the first dose of hydromorphone after randomization to achieve satisfied pain control. The satisfied pain control was defined NRS pain score # 3 at rest in at least 2 consecutive assessment (15 minutes interval). The time needed to successful titration was extended to achieve satisfied pain control again if NRS pain score $ 7 after satisfied pain control within 24 hours. The failure of successful titration was defined that satisfied pain control does not achieve within 24 hours. Secondary endpoints include the percentage of patients titrated successfully, the mean NRS pain score of 24 hours, the total dose of hydromorphone titrated, and adverse events. Clinical trial information: NCT03375515.

TPS11636 Poster Session (Board #324b), Mon, 1:15 PM-4:15 PM

Impact of a regular exercise program on amount of exercise and QOL metrics in patients on immune therapy.

Nicole Brenna Quenelle, Kathryn Blount Bollin; Univ of Connecticut, West Hartford, CT; Scripps Green, La Jolla, CA

Background: Studies show physical activity has a positive impact on fatigue and quality of life both during cancer treatment with chemotherapy and radiation and post-treatment (1, 2). There may also be a survival benefit to increasing physical activity both during and after treatment (3). To date there is no published research on the role of exercise in ameliorating the fatigue patients can experience during treatment with immune therapy. Our study proposes to use the existing framework of the LIVESTRONG at the YMCA program to objectively measure improvement in activity level and objective quality of life measurements. Methods: Randomized controlled prospective study evaluating patient participation in LIVESTRONG at YMCA program during active cancer treatment to assess change in minutes per week of self-reported physical activity over 12 weeks. Assessments will be done based on attendance of 12 week program, activity log, functional assessments of physical activity pre- and post- program (6 min walk test, % change in weight, % change in max weight lifted and flexibility), and questionnaires evaluating fatigue (PROMIS 13a FACIT-F), pain (PROMIS pain intensity scale, ASCQ-Me short form), quality of life (FACT-G), Godin Leisure Time Activity Questionnaire, and inflammatory markers (ESR, CRP). Data will be analyzed on an intention-to-treat analysis. A sample size of 100 participants per group will achieve 80% power to detect a 60 minute difference with a standard deviation of 150 minutes 1 and with a significance level (alpha) of 0.050 using a two-sided two-sample t-test. Enrollment is targeted at 108 participants per arm to allow for 8% attrition, 216 total. Secondary endpoints will be assessed at a baseline functional assessment session for all participants and a follow up session after 12 weeks, including administration of questionnaires at both sessions. For physical activity measurements and survey completions, percent change in baseline and completion measurements will be calculated for each patient, then comparison using a chi-square test will be done to determine statistical significance.(Tomlinson et al. Effect of exercise on cancer- related fatigue: a meta-analysis. Am J Phys Med Rehabil. 2014;93:675-686; Irwin et al. Effect of the LIVESTRONG at the YMCA Exercise Program on Physical Activity, Fitness, Quality of Life, and Fatigue in Cancer Survivors. 2016 (published online October 28, 2016); Li T et al. The dose–response effect of physical activity on cancer mortality: findings from 71 prospective cohort studies. Br J Sports Med. 2016; 50:339-345).