AMPI2019-11Print-1UP 1..95

GENITOURINARY (NONPROSTATE) CANCER

4500 Oral Abstract Session, Mon, 8:00 AM-11:00 AM

Pembrolizumab (pembro) plus axitinib (axi) versus sunitinib as first-line therapy for metastatic renal cell carcinoma (mRCC): Outcomes in the combined IMDC intermediate/poor risk and sarcomatoid subgroups of the phase 3 KEYNOTE-426 study.

Brian I. Rini, Elizabeth R. Plimack, Viktor Stus, Rustem Gafanov, Robert Hawkins, Dmitry Nosov, Frederic Pouliot, Denis Soulieres, Bohuslav Melichar, Ihor Vynnychenko, Sergio Jobim Azevedo, Delphine Borchiellini, Raymond S. McDermott, Jens Bedke, Satoshi Tamada, Shuyan Wan, Rodolfo F. Perini, Mei Chen, Michael B. Atkins, Thomas Powles; Cleveland Clinic Taussig Cancer Institute, Cleveland, OH; Fox Chase Cancer Center, Philadelphia, PA; Dnipropetrovsk Medical Academy of Ministry of Health of Ukraine, Dnipro, Ukraine; Russian Scientific Center of Roentgenoradiology, Moscow, Russian Federation; The Christie NHS Foundation Trust, Manchester, United Kingdom; Central Clinical Hospital with Outpatient Clinic, Moscow, Russian Federation; CHU de Quebec ´ and Universite ´ Laval, Quebec City, QC, Canada; Centre Hospitalier de l’Universite ´ de Montreal, ´ Montreal, QC, Canada; Palacky ´ University Medical School and Teaching Hospital, Olomouc, Czech Republic; Sumy State University, Sumy Regional Oncology Center, Sumy, Ukraine; Hospital de Clı´nicas de Porto Alegre, Porto Alegre, Brazil; Centre Antoine Lacassagne, UniversiteC ´ oteˆ d’Azu, Nice, France; Adelaide and Meath Hospital and University College Dublin, Dublin, Ireland; Department of Urology, Eberhard- Karls University Tubingen, ¨ Tubingen, ¨ Germany; Osaka City University Hospital, Osaka, Japan; Merck & Co., Inc., Kenilworth, NJ; Georgetown Lombardi Comprehensive Cancer Center, Washington, DC; Barts Health and the Royal Free NHS Trusts, Barts Cancer Institute, Queen Mary University of London, London, United Kingdom

Background: In KEYNOTE-426, pembro + axi significantly improved OS (HR 0.53, P , .0001), PFS (HR 0.69, P = .0001), and ORR (59.3% vs 35.7%, P , .0001) vs sunitinib and had manageable toxicity as first-line therapy for mRCC (NCT02853331). The pembro + axi benefit was observed across all IMDC risk groups and regardless of PD-L1 expression. We present data for the combined intermediate/poor risk group and for patients (pts) with sarcomatoid features. Methods: 861 eligible pts with clear-cell mRCC, no prior systemic therapy for mRCC, and KPS $70 were randomized 1:1 to pembro 200 mg IV Q3W for a maximum of 35 cycles plus axi 5 mg orally BID (N = 432) or sunitinib 50 mg orally QD (4-wk on/2-wk off) (N = 429). Primary endpoints were OS and PFS (RECIST v1.1 by blinded, independent central review [BICR]). ORR (RECIST v1.1 by BICR) was the key secondary endpoint. The intermediate/poor risk group was prespecified; the sarcomatoid group was exploratory. HRs and their 95% CIs were calculated with a Cox proportional hazards model. None of the analyses were multiplicity-controlled. Results: 592 (68.8%) of all randomized pts were of IMDC intermediate/poor risk — 294 in the pembro + axi arm, 298 in the sunitinib arm. Pembro + axi improved OS (HR 0.52, 95% CI 0.37-0.74; 12-mo rate 87.3% vs 71.3%), PFS (HR 0.67, 95% CI 0.53-0.85; median 12.6 vs 8.2 mo), and ORR (55.8% [95% CI 49.9-61.5] vs 29.5% [24.4-35.1]) in pts with intermediate/poor risk; CR rates were 4.8% (95% CI 2.6-7.9) vs 0.7% (0.1-2.4). Of the 578 pts with known status, 105 (18.2%) had sarcomatoid features — 51inthepembro+ axi arm, 54 in the sunitinib arm. Pembro + axi improved OS (HR 0.58, 95% CI 0.21-1.59; 12-mo rate 83.4% vs 79.5%), PFS (HR 0.54, 95% CI 0.29-1.00; mediannotreachedvs8.4mo),andORR(58.8% [95% CI 44.2-72.4] vs 31.5% [19.5-45.6]) in pts with sarcomatoid features; CR rates were 11.8% (95% CI 4.4-23.9) vs 0% (0.0-6.6). Conclusions: Pembro + axi provides benefit in the combined population of pts with IMDC intermediate or poor risk and in pts whose tumors had sarcomatoid features. The observed benefits were consistent with those seen in the total population. Clinical trial information: NCT02853331.

4501 Oral Abstract Session, Mon, 8:00 AM-11:00 AM

A pilot randomized study evaluating nivolumab (nivo) or nivo + bevacizumab (bev) or nivo + ipilimumab (ipi) in patients with metastatic renal cell carcinoma (MRCC) eligible for cytoreductive nephrectomy, metastasectomy or post-treatment biopsy (Bx).

Jianjun Gao, Jose A. Karam, Nizar M. Tannir, Matthew T Campbell, Rebecca Slack Tidwell, Kamran Ahrar, Priya Rao, Chaan S. Ng, Eric Jonasch, Surena F. Matin, Amado J. Zurita, Amishi Yogesh Shah, Yu Shen, Jorge M. Blando, Luis M. Vence, Sreyashi Basu, Hao Zhao, James Patrick Allison, Christopher G. Wood, Padmanee Sharma; The University of Texas MD Anderson Cancer Center, Houston, TX; University of Texas MD Anderson Cancer Center, Houston, TX; Department of Immunology, The University of Texas MD Anderson Cancer Center, Houston, TX

Background: Cytoreductive surgery (CS) including cytoreductive nephrectomy and metastasectomy provides clinical benefits to patients with MRCC. However, CS has not been evaluated in the setting of immune checkpoint (IC) therapy. We performed a pre-surgical/biopsy trial to evaluate biological and clinical activity of nivo +/- (bev or ipi) in patients with MRCC. Methods: In this open-label, randomized trial (NCT02210117), patients with MRCC w/o prior IC therapy and anti-VEGF therapy were randomized 2:3:2 to receive nivo (3mg/kg q2wks x3), nivo + bev (10mg/kg q2wks x3) or nivo + ipi (1mg/kg q3wks x2), followed by CS or Bx, and then nivo maintenance therapy up to 2 yrs. Clinical response per RECIST criteria was assessed at $12 wks. Pre- and post-treatment blood and tumors were obtained for correlative studies. Results: All of 105 patients have been accrued and 104 are evaluable for response. For all patients, best overall response (BOR = complete response [CR] + partial response [PR]) including surgery effect was 55% nivo, 44% nivo + bev, 43% nivo + ipi. Median PFS (95% confidence interval) was 14.5 months (5.5, not reached [NR]) nivo, 7.6 (4.8, 8.9) months nivo + bev, 7.5 (2.0, 12.4) months nivo + ipi. Overall survival (OS) at one year was 86% nivo, 73% nivo + bev, 83% nivo + ipi. Grade 3 or higher toxicities related to therapy were 38% for nivo, 42% for nivo + bev (including 18% hypertension), and 47% for nivo + ipi. For patients with CS, BOR including surgery effect was: 86% nivo, 88% nivo + bev, and 69% nivo + ipi. Median PFS was 17.3 months (6.1, NR) nivo, 7.6 (5.7, 10.3) months nivo + bev, 8.9 (2.9, 23.0) months nivo + ipi. OS at one year was 100% nivo, 94% nivo + bev, 92% nivo + ipi. Median OS has not yet reached with a median follow-up of 24.6 months. Immune and gene profiling analyses demonstrate: 1) tumor infiltrating CD8 T cells correlate with clinical responses to nivo or nivo + bev, but not to nivo + ipi; 2) tumor IFN pathway gene expression correlates with responses; and 3) PD-L1 status, tumor mutation or mutation burden, neoantigens did not correlate with response. Conclusions: ICtherapyplusCSissafeandbeneficial to patients with MRCC and therefore, warrants testing, along with a few correlative biomarkers, in a larger phase 3 trial. Clinical trial information: NCT 02210117.

4502 Oral Abstract Session, Mon, 8:00 AM-11:00 AM

Randomized, double-blind phase III study of pazopanib versus placebo in patients with metastatic renal cell carcinoma who have no evidence of disease following metastasectomy: A trial of the ECOG-ACRIN cancer research group (E2810).

Leonard Joseph Appleman, Maneka Puligandla, Sumanta K. Pal, Wayne Harris, Neeraj Agarwal, Brian Addis Costello, Christopher W. Ryan, Michael Pins, Jill Kolesar, Daniel A. Vaena, Rahul Atul Parikh, Mehmood Hashmi, Janice P. Dutcher, Robert S. DiPaola, Naomi B. Haas, Michael Anthony Carducci; UPMC Hillman Cancer Center, Pittsburgh, PA; Dana-Farber Cancer Institute, Boston, MA; City of Hope National Medical Center, Duarte, CA; Emory University School of Medicine, Department of Hematology and Medical Oncology, Winship Cancer Institute of Emory University, Atlanta, GA; University of Utah Huntsman Cancer Institute, Salt Lake City, UT; Mayo Clinic, Rochester, MN; Oregon Health & Science University, Knight Cancer Institute, Portland, OR; University of Illinois College of Medicine, Chicago, IL; University of Wisconsin Carbone Cancer Center, Madison, WI; University of Iowa Hospitals and Clinics, Holden Comprehensive Cancer Center, Iowa City, IA; University of Kansas Cancer Center, Westwood, KS; University of Kansas, Kansas City, KS; Cancer Research Foundation of NY, Bronx, NY; University of Kentucky, Lexington, KY; Penn Medicine Abramson Cancer Center, Philadelphia, PA; Sidney Kimmel Cancer Center At Johns Hopkins, Baltimore, MD

Background: Patients with no evidence of disease (NED) after metastasectomy for metastatic renal cell carcinoma (mRCC) are at high risk of recurrence, but no systemic therapy has been shown to benefit this population. Pazopanib is an inhibitor of VEGFR and other kinases that improves progression-free survival in patients with measurable RCC metastatic disease. We performed a randomized, double-blind, placebo- controlled multicenter study to test the hypothesis that pazopanib would improve disease-free survival in patients with mRCC rendered NED after metastasectomy Methods: Patients with NED following metastasectomy were randomized 1:1 to receive pazopanib starting at 800 mg daily vs. placebo for 52 weeks. Patients were stratified by 1 vs. . 1 site of resected disease, and by disease-free interval # vs. . 1year. Clinical assessment for toxicity and patient-reported outcomes were performed every 4 weeks, and restaging scans every 12 weeks. The study was designed to observe a 42% improvement in disease-free survival (DFS) from 25% to 45% at 3 years. Results: From August 2012 to July 2017, 129 patients were enrolled. The study was unblinded after 83 DFS events had been observed (92% information). The median follow-up from randomization was 30 months (range 0.4 – 66.5 months). The study did not meet the primary endpoint: hazard ratio (95% CI) for DFS was 0.85 (0.55, 1.31) p= 0.47 in favor of pazopanib. At the time of unblinding, 22/129 (17%) of subjects had died. The HR for overall survival (OS) was 2.65 (1.02, 6.9) in favor of placebo (p= 0.05). Patient-reported outcomes and laboratory correlates will be reported separately. Conclusions: 52 weeks of pazopanib did not improve DFS compared to blinded placebo in patients with mRCC who were NED after metastasectomy. There was a trend toward worse overall survival with pazopanib. Clinical trial information: NCT01575548.

4503 Oral Abstract Session, Mon, 8:00 AM-11:00 AM

CALGB 90601 (Alliance): Randomized, double-blind, placebo-controlled phase III trial comparing gemcitabine and cisplatin with bevacizumab or placebo in patients with metastatic urothelial carcinoma.

Jonathan E. Rosenberg, Karla V. Ballman, Susan Halabi, Colleen Watt, Olwen Mary Hahn, Preston D. Steen, Robert Dreicer, Thomas W. Flaig, Walter Michael Stadler, Christopher Sweeney, Amir Mortazavi, Michael J. Morris; Memorial Sloan Kettering Cancer Center, New York, NY; Weill Cornell Medicine, New York, NY; Duke University Medical Center, Durham, NC; University of Chicago, Chicago, IL; University of Chicago Medical Center, Chicago, IL; Roger Maris Cancer Center, Fargo, ND; Cleveland Clinic, Cleveland, OH; Division of Medical Oncology, School of Medicine, University of Colorado, Aurora, CO; Lank Center for Genitourinary Oncology, Dana-Farber Cancer Institute, Boston, MA; Arthur G. James Cancer Hospital, Ohio State University Wexner Medical Center, Columbus, OH

Background: The combination of gemcitabine (G) and cisplatin (C) is a standard therapy for metastatic urothelial carcinoma (mUC). Based on data that angiogenesis plays a role in UC growth and progression, a randomized placebo-controlled trial was performed. Methods: Patients mUC, no prior chemotherapy for metastatic disease and .12 months from prior (neo)adjuvant chemotherapy and ECOG PS 0-1 were randomized 1:1 to G 1000 mg/m2 IV days 1 and 8 and C IV 70 mg/m2 day 1 with bevacizumab (GCB) 15 mg/kg IV or placebo (GCP) day 1 every 21 days. Randomization was stratified by the presence of visceral metastases and prior chemotherapy. The primary endpoint was overall survival (OS) defined as the time from randomization to death or last follow-up (FU). Secondary endpoints included progression-free survival (PFS), objective response rate (ORR), and $ grade 3 toxicity. With 445 deaths, the log-rank test had an 87% power to detect a hazard ratio (HR) of 0.74 with a 2-sided a=0.05. The primary analysis was based on the stratified log-rank test adjusting on stratification factors. Alliance Data Safety and Monitoring Board approved the final OS analysis be performed at 420 events due to lower than expected event rates. Results: 506 patients were randomly assigned (252 GCB, 254 GCP) stratified by the presence of visceral disease and prior chemotherapy for UC. The median FU for patients still alive was 46.2 months. Median OS was 14.5 months for patients treated with GCB and 14.3 months for patients treated with GCP with a HR of 0.87 (95%CI 0.72-1.06; 2-sided Wald p=0.17). The HR for PFS was 0.77 (95%CI 0.63-0.93) in favor of GCB (p=0.0074). Grade 3 or greater adverse event rate was 83.5% with GCB compared to 80.7% with GCP. Conclusions: The addition of bevacizumab to GC chemotherapy did not result in improved OS (primary endpoint) in patients with mUC but there was a PFS improvement. The observed median OS of about 14 months is consistent with prior phase III trials of cisplatin-based chemotherapy. Support: U10CA180821, U10CA180882, U10CA180820, U10CA180853, U10CA180888, Genentech https://acknowledgments.alliancefound.org. Clinical trial information: NCT00942331.

4504 Oral Abstract Session, Mon, 8:00 AM-11:00 AM

Randomized double-blind phase II study of maintenance pembrolizumab versus placebo after first-line chemotherapy in patients (pts) with metastatic urothelial cancer (mUC): HCRN GU14-182.

Matt D. Galsky, Sumanta K. Pal, Amir Mortazavi, Matthew I. Milowsky, Saby George, Sumati Gupta, Mark T. Fleming, Long H. Dang, Daniel M. Geynisman, Radhika Walling, Robert S. Alter, Erwin L. Robin, Jue Wang, Shilpa Gupta, David D. Chism, Joel Picus, George Philips, David I. Quinn, Noah M. Hahn, Menggang Yu; Department of Medicine, Icahn School of Medicine at Mount Sinai, Tisch Cancer Institute, New York, NY; City of Hope National Medical Center, Duarte, CA; Arthur G. James Cancer Hospital, Ohio State University Wexner Medical Center, Columbus, OH; University of North Carolina at Chapel Hill School of Medicine, Chapel Hill, NC; Roswell Park Cancer Institute, Buffalo, NY; Huntsman Cancer Institute-University of Utah Health Care, Salt Lake City, UT; Virginia Oncology Associates, US Oncology Research, Norfolk, VA; Ochsner Medical Center, Baton Rouge, LA; Fox Chase Cancer Center, Phila- delphia, PA; Community Cancer Center, Indianapolis, IN; John Theurer Cancer Center at Hackensack University Medical Center, Hackensack, NJ; Community Hospital, Munster, IN; University of Arizona Cancer Center at Dignity Health St. Joseph’s Hospital and Medical Center, Phoenix, AZ; Department of Medicine, Masonic Cancer Center, University of Minnesota, Minneapolis, MN; Vanderbilt University Medical Center, Nashville, TN; Washington University in St. Louis School of Medicine, St. Louis, MO; Georgetown Univ Hosp, Washington, DC; USC Norris Comprehensive Cancer Center, Los Angeles, CA; Johns Hopkins University School of Medicine, Baltimore, MD; University of Wisconsin Department of Biostatistics and Medical Informatics, Madison, WI

Background: Platinum-based chemotherapy for 1st-line treatment of pts with metastatic urothelial cancer (mUC) is typically administered for a fixed duration followed by observation until recurrence. PD-1 blockade with pembro improves survival of pts with mUC progressing despite platinum-based chemotherapy. We explored the potential benefit of earlier use of PD-1 blockade using a "switch maintenance" approach. Methods: Pts with mUC achieving at least stable disease after up to 8 cycles of 1st-line platinum- based chemotherapy were enrolled. Pts were randomized 1:1 to pembro 200 mg IV q3 weeks versus placebo for up to 24 months; pts progressing on placebo could cross over to pembro. Randomization was stratified based on pre-chemotherapy visceral metastases (Y/N) and response to 1st-line chemotherapy (CR/PR vs. SD). The primary objective was to determine the progression-free survival (PFS) as per irRECIST among pts treated with pembro versus placebo. Results: Between 12/2015 and 11/2018, 107 pts were randomized to placebo (n=52) versus pembro (n=55). The baseline pt characteristics are shown in the Table. Pts randomized to placebo and pembro received a median of 6 and 8 cycles, respectively. Excluding patients with baseline CRs, the objective response rate was 12% (5/42) on placebo and 22% (10/46) on pembro. Grade 3-4 treatment emergent adverse events occurred in 48% of pts on placebo and 56% on pembro. At a median follow-up of 14.7 months, 41 pts have died and 26/52 pts randomized to placebo have crossed over to pembro. PFS was significantly longer in patients randomized to pembro vs. placebo (Maximum Efficiency Robust Test p=0.036; log-rank p = 0.038). The 18-month restricted mean progression-free survival time was 5.6 months with placebo and 8.2 months with pembro (p=0.023). Conclusions: Switch maintenance pembro may “deepen” responses achieved with 1st-line chemotherapy. Switch maintenance pembro prolongs PFS in pts with mUC completing 1st-line platinum-based chemotherapy. Clinical trial information: NCT02500121.

Baseline Characteristics. Placebo (N=52) Pembro (N=55) Age, median (range) 65 (44-87) 68 (41-83) Pre-chemo metastatic disease Visceral metastases 32 (62%) 39 (71%) No visceral metastases 20 (38%) 16 (29%) Median # cycles of 1st line chemo 65 Response to 1st line chemotherapy CR/PR 36 (69%) 40 (73%) SD 16 (31%) 15 (27%) 1st line chemo Carboplatin-based 11 (21%) 16 (29%) Cisplatin-based 41 (79%) 39 (71%)

LBA4505 Oral Abstract Session, Mon, 8:00 AM-11:00 AM

EV-201: Results of enfortumab vedotin monotherapy for locally advanced or metastatic urothelial cancer previously treated with platinum and immune checkpoint inhibitors.

Daniel Peter Petrylak, Arjun Vasant Balar, Peter H. O’Donnell, Bradley Alexander McGregor, Elisabeth I. Heath, Evan Y. Yu, Matt D. Galsky, Noah M. Hahn, Elaina M Gartner, Juan Pinelli, Amal Melhem-Bertrandt, Jonathan E. Rosenberg; Yale School of Medicine, New Haven, CT; Perlmutter Cancer Center at NYU Langone Health, New York, NY; University of Chicago Comprehensive Cancer Center, Chicago, IL; Dana-Farber Cancer Institute, Boston, MA; Barbara Ann Karmanos Cancer Institute, Wayne State University, Detroit, MI; University of Washington, Seattle, WA; Department of Medicine, Icahn School of Medicine at Mount Sinai, Tisch Cancer Institute, New York, NY; Johns Hopkins University School of Medicine, Baltimore, MD; Seattle Genetics, Bothell, WA; Seattle Genetics, Inc., Bothell, WA; Astellas Pharma US, Inc, Northbrook, IL; Memorial Sloan Kettering Cancer Center, New York, NY

The full, final text of this abstract will be available at abstracts.asco.org at 7:30 a.m. ET on Monday, June 3. Onsite at the Meeting, this abstract will be printed in the Monday edition of ASCO Daily News.

4506 Oral Abstract Session, Mon, 8:00 AM-11:00 AM

SWOG S1314: A randomized phase II study of co-expression extrapolation (COXEN) with neoadjuvant chemotherapy for localized, muscle-invasive bladder cancer.

Thomas W. Flaig, Catherine M. Tangen, Siamak Daneshmand, Ajjai Shivaram Alva, Seth P. Lerner, M. Scott Lucia, David James McConkey, Dan Theodorescu, Amir Goldkorn, Matthew I. Milowsky, Richard Carlton Bangs, Gary R. MacVicar, Bruno R. Bastos, Daniel Gustafson, Melissa Plets, Ian Murchie Thompson; Division of Medical Oncology, School of Medicine, University of Colorado, Aurora, CO; Fred Hutchinson Cancer Research Center, Seattle, WA; Institute of Urology, Keck School of Medicine, University of Southern California, Los Angeles, CA; University of Michigan, Ann Arbor, MI; Baylor College of Medicine, Houston, TX; University of Colorado Anschutz Medical Campus, Aurora, CO; Johns Hopkins School of Medicine, Baltimore, MD; Samuel Oschin Comprehensive Cancer Institute at Cedars-Sinai, Los Angeles, CA; Division of Medical Oncology, Department of Medicine, Keck School of Medicine and Norris Comprehensive Cancer Center, University of Southern California, Los Angeles, CA; University of North Carolina at Chapel Hill School of Medicine, Chapel Hill, NC; SWOG, San Antonio, TX; Illinois CancerCare PC, Peoria, IL; Cleveland Clinic Florida, Weston, FL; Colorado State University, Fort Collins, CO; CHRISTUS Medical Center Hospital, San Antonio, TX

Background: Both dose-dense Methotrexate-Vinblastine-Adriamycin/doxorubicin-Cisplatin (ddMVAC) and Gemcitabine-Cisplatin (GC) are accepted neoadjuvant regimens for muscle-invasive bladder cancer (BC). We investigated COXEN, a gene expression model, as a predictive biomarker. Methods: Eligibility included Stage cT2-T4a N0 M0, urothelial BC (mixed histology allowed), $ 5mmofviabletumor, cisplatin eligible, with plan for cystectomy. 237 patients were randomized between ddMVAC, given every 14 days for 4 cycles, and GC, given every 21 days for 4 cycles. The primary objective was to assess whether the pre-specified dichotomous treatment-specific COXEN gene expression profile is prognostic of pT0 rate or # pT1 at surgery, and to assess whether COXEN score is a predictive factor between regimens and response. Logistic regression was used to model response, adjusting for stratification factors. Results: 167 patients were included; the ddMVAC/GC arms had a median age of 65/64, PS = 0 in 80%/75%, Male proportion of 88%/79% and T2 stage of 87%/92%. All had at least 3 cycles of chemo and surgery/progression within 100 days of last chemo. There were favorable COXEN ddMVAC scores in 32% and GC score in 26%. The pT0 rates for ddMVAC and GC were 32% and 35%; the rates of # pT1 were 55% and 49%, respectively. Conclusion: The COXEN scores were not significantly prognostic for response in their individual arms; The COXEN GC score was significant predictor for downstaging in pooled arms. There was no evidence of an interaction between COXEN score and regimen in predicting response. The prospective data and samples from this study will allow for further development of COXEN and other predictive biomarkers. Clinical trial information: NCT02177695.

Logistic Regression of COXEN Score Factor N Odds Ratio ** 95% CI ** p-value** GC score* for pT0 in GC arm 82 2.63 0.82, 8.36 0.10 For downstaging 82 1.79 0.60, 5.34 0.30 MVAC score* for pT0 in MVAC arm 85 1.12 0.42, 2.95 0.82 For downstaging 85 0.92 0.37, 2.27 0.86 GC score* for downstaging in both arms 167 2.33 1.11, 4.89 0.02 MVAC score* for downstaging in both arms 167 0.90 0.46, 1.75 0.76 * favorable based on prespecified algorithm and dichotomous cut point ** adjusted for two strat factors – clinical stage at baseline (T2 vs T3, T4a), PS (0 vs 1)

4507 Oral Abstract Session, Mon, 8:00 AM-11:00 AM

Randomized trial of adjuvant chemotherapy versus adjuvant radiation therapy for locally advanced bladder cancer after radical cystectomy.

Mohamed S. Zaghloul, John Paul Christodouleas, Tarek Zaghloul, Andrew Smith, Ahmed Abdalla, Hany William, Wei-Ting Hwang, Brian Christopher Baumann; Children’s Cancer Hospital, Cairo, Egypt; University of Pennsylvania, Department of Radiation Oncology, Philadelphia, PA; National Cancer Institute, Cairo, Egypt; University of Pennsylvania, Department of Biostatistics and Epidemiology, Philadelphia, PA; Ahmed Maher Teaching Hospital, Cairo, Egypt; Washington University in St. Louis, Department of Radiation Oncology, St. Louis, MO

Background: Some chemotherapy-na¨ıve patients with locally advanced bladder cancer (LABC) after radical cystectomy (RC) are sufficiently de-conditioned that they are not candidates for adjuvant chemotherapy or decline it, even though such treatment may be warranted. There is no clear alternative adjuvant therapy for these patients, who are usually observed. In this study, we compare post-op radiotherapy (PORT) vs. adjuvant chemotherapy in a randomized clinical trial. We hypothesized that PORT can achieve comparable disease-free survival (DFS). Methods: A randomized phase III trial was opened to compare PORT vs. sequential chemo+PORT after RC for LABC & accrued from 2002–2008 at the NCI in Cairo. In 2007, a third arm comparing adjuvant chemo was added. Herein, we report the results of PORT vs. adjuvant chemo. Patients #70 y/o with $1 of the following factors ($pT3b/T4a, grade 3, or positive nodes) with negative margins after RC + pelvic node dissection were eligible. Routine follow-up & pelvic CT q6 months were performed. PORT included 3D conformal pelvic RT (45Gy/1.5Gy BID). Chemo included gemcitabine/cisplatin x 4. Post-hoc non-inferiority exploratory analysis was performed. Results: The PORT arm accrued 78; the chemo arm accrued 45. 51% had urothelial carcinoma; 49% had squamous cell carcinoma/other. The two arms were well-balanced except for gender (p = 0.06). Two-year outcomes & overall adjusted hazard ratios (HR) for PORT vs. chemo alone were 54% vs. 47% (HR 0.65(95%CI 0.35-1.19, p = 0.16) for DFS; 92% vs. 69% (HR 0.28(95%CI 0.10-0.82), p = 0.02 for LRFS; 75% vs. 79% (HR 2.39(95%CI 0.94-6.09), p = 0.07) for DMFS; 61% vs. 60% (HR 0.94(95%CI 0.52-1.69), p = 0.83) for OS. Late grade $3 GI toxicity was observed in 6 PORT patients (8%) & 1 chemo patient (2%). Based on our data, there is a greater than 90% probability that the true difference in 2 yr DFS is less than 10%, the pre-specified non-inferiority margin. Conclusions: This randomized study demonstrates superior local control with PORT vs. adjuvant chemo with no significant differences in DFS, DMFS or OS. Results suggest that PORT could be an option for patients with LABC after RC who are medically unfit for adjuvant chemo or who decline it. Clinical trial information: NCT01734798.

4508 Oral Abstract Session, Mon, 8:00 AM-11:00 AM

Cytoreductive nephrectomy (CN) in metastatic renal cancer (mRCC): Update on Carmena trial with focus on intermediate IMDC-risk population.

Arnaud Mejean, Simon Thezenas, Christine Chevreau, Karim Bensalah, Lionnel Geoffrois, Antoine Thiery- Vuillemin, Luc Cormier, Herve Lang, Laurent Guy, Gwenaelle Gravis, Frederic Rolland, Claude Linassier, Marc-Olivier Timsit, Laurence Albiges, Stephane Oudard, Thierry Lebret, jean-Marc Treluyer, Sandra Colas, Bernard Escudier, Alain Ravaud; Department of Urology, Hopitalˆ Européen Georges- Pompidou - Paris Descartes University, Paris, France; Institut du Cancer de Montpellier (ICM), Univ Montpellier, Montpellier, France; IUCT-Oncopoleˆ Institut Claudius Regaud, Toulouse, France; University Hospital Pontchaillou Service d’urologie CHU Rennes, Université de Rennes, Rennes, France; Centre Alexis Vautrin, Vandoeuvre-Lès-Nancy, France; CU-PH Medical Oncology, Besançon, France; Centre Hospitalier Universitaire de Dijon, Dijon, France; Department of Urology, CHU Strasbourg, Strasbourg University, Strasbourg, France; Hopital Gabriel Montpied CHU, Clermont-Ferrand, France; Medical Oncology, Institut Paoli-Calmettes, Marseille, France; Institut de Cancerologie ´ de l’Ouest, Department of Medical Oncology, St Herblain, France; CHU Tours, Tours Cedex 9, France; Hopital Europeen Georges Pompidou, Paris, France; Medical Oncology, Gustave Roussy, Université Paris-Saclay, Villejuif, France; Hopitalˆ Foch, Suresnes, France; URC/CIC Paris Descartes Necker Cochin, Paris, France; Gustave Roussy, Villejuif, France; CHU Hopitaux de Bordeaux-Hopitalˆ Saint-André, Bordeaux, France

Background: Carmena was a randomized phase III trial, testing the benefit of CN followed by sunitinib (arm A) vs sunitinib alone (arm B), with stratification by MSKCC risk groups in 450 mRCC patients. Based on this trial, CN is not anymore recommended in mRCC (NEJM, Mejean et al, 2018). However there are questions about which patients could still benefit from CN, especially in intermediate risk group. In the present study, we investigated different subgroups from the Carmena trial to answer these questions. Methods: Carmena trial was initially stratified according to MSKCC risk groups. For the purpose of this analysis, we reclassified the patients based on IMDC risk groups. We also analyzed patients with one metastatic site vs more than one, as well as patients with secondary nephrectomy in arm B. Overall survival (OS) was the primary endpoint. Results: With a updated median FU of 61.5 months (mo), the median OS by ITT analysis was 15.6 vs 19.8 mo in arm A and B respectively stratified on MSKCC (HR 0.933 ; 95% CI [0.76- 1.15]) / stratified on IMDC (HR 0.957 ; 95% CI [0.78- 1.18]). Using IMDC risk group factors, 58.6% patients were intermediate and 41.4 % were poor risk. When looking at intermediate risk group only, 48.1% had only one risk factor (interval between diagnosis and treatment , 1y), with a median OS of 30.5 and 25.2 mo in arm A and B respectively (HR 1.24 [0.81 – 1.90]). By contrast, 51.9 % had two risk factors (mostly low hemoglobin, high corrected calcium or neutrophils), with a median OS of 16.6 and 31.2 mo in arm A and B respectively (HR 0.61 [0.41 – 0.91] p = 0.015). Regarding number of metastatic sites, 33% had only one metastatic site. Median OS was 23.6 and 22.7 mo in arm A and B respectively (HR 1.08 [0.75 – 1.57]. Finally, 40 patients had a secondary nephrectomy in arm B, with median OS of 48.5 mo [CI 95%: 27.9-64.4] vs 15.7 mo [CI 95%: 13.3-20.5] in patients who never had surgery. Conclusions: With longer FU, Carmena trial confirms that CN is not superior to sunitinib alone in ITT population, both with MSKCC and IMDC risk groups. However for patients with only one IMDC risk factor, CN might be beneficial. Number of metastatic site is not helpful to define good candidates for surgery. Finally, patients with secondary nephrectomy have very long OS, supporting this strategy. Clinical trial information: NCT00930033.

4509 Poster Discussion Session; Displayed in Poster Session (Board #335), Mon, 1:15 PM-4:15 PM, Discussed in Poster Discussion Session, Mon, 4:30 PM-6:00 PM

A phase II study of RC48-ADC in HER2-positive patients with locally advanced or metastatic urothelial carcinoma.

Xinan Sheng, Ai-Ping Zhou, Xin Yao, Yanxia Shi, Hong Luo, Benkang Shi, Jiyan Liu, Guohua Yu, Zhisong He, Changlu Hu, Weiqing Han, Jianmin Fang, Jun Guo; Peking University Cancer Hospital, Beijing, China; National Cancer Center/Cancer Hospital,Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China; Tianjin Cancer Hospital, Tianjin, China; Sun Yat-Sen university Cancer Center, Guangzhou, China; Chongqing Cancer Hospital, Chongqing, China; Qilu Hospital of Shandong University, Jinan, China; West China Hospital, Sichuan University, Chengdu, China; Weifang People9s Hospital, Weifang, China; Peking University First Hospital, Beijing, China; Anhui Provincial Cancer Hospital, Hefei, China; Hunan Cancer Hospital, Changsha, China; School of Life Science and Technology, Tongji University, Shanghai, China; Peking University Cancer Hospital and Institute, Beijing Shi, China

Background: Urothelial carcinoma (UC) is the third largest cancer type with HER2 positive cancer. RC48- ADC is a novel humanized anti-HER2 antibody-drug conjugate (ADC). This study was to evaluate the activity of RC48-ADC in HER2-positive patients with locally advanced or metastatic urothelial carcinoma (mUC). Methods: This study is an open-label, multicenter, single-arm, non-randomized phase II study. Eligibility criteria include: histologically confirmed UC, HER2-positive (IHC 2+ or 3+), ECOG PS 0-1, treated with $1 prior systemic treatment. The patients received RC48-ADC treatment alone (2 mg/kg IV infusion, q2w) until disease progression, unacceptable toxicity, withdrawal, or study termination. The primary endpoint was objective response rate (ORR). Progress-free survival (PFS), overall survival (OS), and safety was also assessed. Results: Patient enrollment for this study was completed in November 2018. A total of 43 patients were enrolled, with a median age of 64 years old. At baseline, most patients (37/43) had visceral metastasis. Fourteen (32.6%) patients had received $ 2 lines treatment and 8 (18.6%) patients had prior immune checkpoint inhibitor (CPI) therapy in second line treatment. The objective response rate was 60.5% (95% CI: 44.4%, 75.0%) and the DCR was 90.7% (39/43). As of Jan 23, 2019, the median PFS for the overall study population was not yet reached, and the median PFS was 7.8 months (95% CI: 4.9, 10.7) for the 9 patients who started RC48-ADC prior to Jun 30, 2018. The ORR was 70.6% (12/17) in patients with HER2 FISH+ or IHC3+. The ORR was 64.9% (24/37) in patients with visceral metastasis and was 70.0% (14/20) in liver metastasis patients especially. The ORR was 64.3% in patients post to $ 2 lines treatment and 75.0% in patients post to immunotherapy. Common treatment-related AEs were leukopenia (51.2%), hypoesthesia (41.9%), alopecia (41.9%), neutropenia (37.2%), fatigue (34.9%), ALT increase (32.6%), and AST increase (32.6%); Most were Grade 1 or 2. Conclusions: RC48-ADC has demonstrated a clinically meaningful ORR of 60.5% in pretreated HER-2 positive mUC patients including those who underwent failure to the immunotherapy. Clinical trial information: NCT03507166.

4510 Poster Discussion Session; Displayed in Poster Session (Board #336), Mon, 1:15 PM-4:15 PM, Discussed in Poster Discussion Session, Mon, 4:30 PM-6:00 PM

Infigratinib in upper tract urothelial carcinoma vs urothelial carcinoma of the bladder and association with comprehensive genomic profiling/cell-free DNA results.

Nazli Dizman, Jonathan E. Rosenberg, Jean H. Hoffman-Censits, David I. Quinn, Daniel Peter Petrylak, Matt D. Galsky, Ulka N. Vaishampayan, Ugo De Giorgi, Sumati Gupta, Howard A. Burris, Harris S. Soifer, Gary Li, Carl L. Dambkowski, Susan Moran, Yining Ye, Siamak Daneshmand, Dean F. Bajorin, Sumanta K. Pal; City of Hope Comprehensive Cancer Center, Duarte, CA; Memorial Sloan Kettering Cancer Center, New York, NY; Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins, Baltimore, MD; USC Norris Comprehensive Cancer Center, Los Angeles, CA; Yale Cancer Center, Smilow Cancer Hospital, New Haven, CT; Department of Medicine, Icahn School of Medicine at Mount Sinai, Tisch Cancer Institute, New York, NY; Wayne State University, Detroit, MI; Istituto Scientifico Romagnolo per lo Studio e la Cura dei Tumori, Meldola, Italy; Huntsman Cancer Institute-University of Utah Health Care, Salt Lake City, UT; Sarah Cannon Research Institute, Nashville, TN; QED Therapeutics Inc, San Francisco, CA; Keck School of Medicine of USC, Los Angeles, CA; City of Hope National Medical Center, Duarte, CA

Background: Infigratinib (BGJ398) is a potent and selective FGFR1–3 inhibitor with significant activity in patients (pts) with advanced or metastatic urothelial carcinoma (mUC) bearing FGFR3 alterations [Pal et al 2018]. Given the distinct biologic characteristics of upper tract UC (UTUC) and urothelial carcinoma of the bladder (UCB), we sought to determine if infigratinib had varying activity in these settings. Methods: Eligible pts had mUC with activating FGFR3 mutations/fusions and prior platinum-based chemotherapy, unless contraindicated. Pts received infigratinib 125 mg orally daily (3 wks on/1 wk off). Overall response rate (ORR: CR+PR) and disease control rate (DCR; CR+PR+SD) were characterized in UCB and UTUC pts. Comprehensive genomic profiling was performed on FFPE tissues in a CLIA-certified lab (Foundation Medicine; Cambridge, MA). Blood was collected for cell-free (cf)DNA analysis using a 600-gene panel on an Illumina HiSeq 2500 sequencer. Results: 67 pts were enrolled; the majority (70.1%) had received $2 prior antineoplastic therapies. ORR was 25.4% and DCR was 64.2%. In the 8 pts with UTUC, 1 CR and 3 PRs were observed (ORR 50%); the remainder had a best response of SD (DCR 100%). UTUC pts were predominantly 2nd line (62.5%), with only 2 (25%) showing response to previous treatment. In pts with UCB, 13 PRs were observed (ORR 22%), and 22 pts had a best response of SD (DCR 59.3%). Notable differences in genomic alterations between cohorts included a higher frequency of FGFR3-TACC3 fusions (12.5% vs 5.8%) and FGFR3 R248C mutations (50% vs 11.5%), and a lower frequency of FGFR3 S249C mutations (25% vs 59.6%) in UTUC vs UCB. Consistent with previous reports [Sfakianos et al 2016], UTUC pts had a differential frequency of alterations in HRAS, CDKN2B and ARID1A.Sufficient cfDNA yield was obtained in UTUC and UCB pts and a comprehensive comparison of these data will be presented. Conclusions: Differences in cumulative genomic profile were observed between UCB and UTUC in this FGFR3-restricted experience, underscoring the distinct biology of these diseases. Results with infigratinib in UTUC support a planned phase III adjuvant study predominantly in this population. Clinical trial information: NCT01004224.

4511 Poster Discussion Session; Displayed in Poster Session (Board #337), Mon, 1:15 PM-4:15 PM, Discussed in Poster Discussion Session, Mon, 4:30 PM-6:00 PM

FIERCE-22: Clinical activity of vofatamab (V) a FGFR3 selective inhibitor in combination with pembrolizumab (P) in WT metastatic urothelial carcinoma, preliminary analysis.

Arlene O. Siefker-Radtke, Graeme Currie, Esteban Abella, Daniel A. Vaena, Arash Rezazadeh Kalebasty, Giuseppe Curigliano, Krzysztof Tupikowski, Zoran Gojko Andric, Iwona Lugowska, William Kevin Kelly; The University of Texas MD Anderson Cancer Center, Houston, TX; Rainier Therapeutics Inc, San Leandro, CA; Rainier Therapeutics, Inc, San Leandro, CA; University of Iowa Hospitals and Clinics, Holden Comprehensive Cancer Center, Iowa City, IA; Norton Cancer Institute, Louisville, KY; University of Milano, European Institute of Oncology, Division of Early Drug Development, Milan, Italy; Dolnoslaskie Centrum Onkologii, Dolnoslaskie, Poland; Clinical Hospital Center, Bezanijska Kosa, Belgrade, Serbia; Centrum Onkologii - Instytut im. Marii Sklodowskiej - Curie, Warsaw, Poland; Sidney Kimmel Cancer Center at Thomas Jefferson University, Philadelphia, PA

Background: Patients (pts) with mUC who have failed platinum-based chemotherapy have a poor prognosis. Reported response rates to immune checkpoint inhibitors (ICI) are approximately 20%. 20% of pts with mUC harbor FGFR3 mutations or fusions (M/F), which may result in lower sensitivity to ICI. V (B-701) is a fully human monoclonal antibody against FGFR3 that blocks activation of both the wildtype and genetically activated receptor. This is a Phase 1b/2 study designed to evaluate V monotherapy window followed in combination of V with P(VP) (NCT03123055). Methods: This trial enrolled mUC pts with failure to $ 1 prior line of chemotherapy or recurrence # 12 months of (neo) adjuvant chemotherapy, measurable disease and ECOG ,2. Treatment consisted of v at 25 mg/kg alone for 2 week monotherapy window followed by combination with P 200 mg q3w.during the V window paired tumor biopsy were obtained. Efficacy was assessed by investigators (RECIST 1.1). Primary objectives were safety and activity [ORR]). Results: 35 pts have received treatment (Ph1b:8, WT:20, M/F+: 7). WT patients were unselected for PD-1 status, predominately male (55%) white (95%), all had received at least 1 line of prior chemo and 60% had Bellmunt scores of . 1. The safety profile is consistent with previously reported data for P. TEAE occurring in .20% of patients were nausea, anemia, diarrhea and fatigue. Six WT patients (30%) had responses (4 confirmed responses, 1 unconfirmed), and an additional patient with an iRECIST response. Responses occurred at a median of 3.5 months. At 4+ months of follow up, 13(65%) remain on treatment @ a median of 8 cycles (range: 1-13). At 5+ months the median PFS has not been reached. Conclusions: VP combination therapy is well tolerated with an encouraging ORR and prolonged PFS in the WT cohort; greater than one would anticipate from P alone based upon historical data. We will be reporting 9+ month preliminary PFS/OS and updated OOR/DOR data. Clinical trial information: NCT03123055.

4512 Poster Discussion Session; Displayed in Poster Session (Board #338), Mon, 1:15 PM-4:15 PM, Discussed in Poster Discussion Session, Mon, 4:30 PM-6:00 PM

Atezolizumab (atezo) + bevacizumab (bev) versus sunitinib (sun) in pts with untreated metastatic renal cell carcinoma (mRCC) and sarcomatoid (sarc) histology: IMmotion151 subgroup analysis.

Brian I. Rini, Robert J. Motzer, Thomas Powles, David F. McDermott, Bernard Escudier, Frede Donskov, Robert E. Hawkins, Sergio Bracarda, Jens Bedke, Ugo De Giorgi, Camillo Porta, Alain Ravaud, Francis Parnis, Enrique Grande, Wei Zhang, Mahrukh A. Huseni, Susheela Carroll, Roxana Ioana Sufan, Christina Schiff, Michael B. Atkins; Cleveland Clinic Taussig Cancer Institute, Cleveland, OH; Memorial Sloan Kettering Cancer Center, New York, NY; Barts Cancer Institute, Queen Mary University of London, Royal Free NHS Trust, London, United Kingdom; Beth Israel Deaconess Medical Center, Boston, MA; Gustave Roussy, Villejuif, France; Department of Oncology, Aarhus University Hospital, Aarhus, Denmark; The Christie NHS Foundation Trust, Manchester, United Kingdom; Azienda Ospedaliera S.Maria, Terni, Italy; Department of Urology, University of Tubingen, ¨ Tubingen, Germany; Istituto Scientifico Romagnolo per lo Studio e la Cura dei Tumori (IRST) IRCCS, Meldola, Italy; IRCCS San Matteo University Hospital Foundation, Pavia, Italy; CHU Hopitaux de Bordeaux-Hopitalˆ Saint-Andre, ´ Bordeaux, France; Ashford Cancer Centre Research, Kurralta Park, SA, Australia; MD Anderson Cancer Center Madrid, Madrid, Spain; Genentech, Inc., San Francisco, CA; Genentech, Inc., South San Francisco, CA; Genetech Inc, South San Francisco, CA; Georgetown Lombardi Comprehensive Cancer Center, Washington, DC

Background: In the Phase 3 IMmotion151 trial, atezo + bev showed improved PFS vs sun in untreated mRCC pts expressing PD-L1. Here we report results of a prespecified subgroup analysis in pts whose tumors have sarc histology, an independent predictor of poor survival. Methods: Pts were randomized to receive atezo 1200 mg IV q3w + bev 15 mg/kg IV q3w or sun 50 mg po qd for 4w on, 2w off. Coprimary endpoints were reported previously (Motzer ASCO GU 2018). Secondary endpoints included INV-PFS and OS in sarc pts and are shown here with INV-ORR, safety, PRO and biomarker data. Results: 142 randomized pts (16%) from IMmotion151 had tumors with any component of sarc histology; mPFS was 8.3 vs 5.3 mo with atezo + bev vs sun and mOS was NR vs 15.0 mo, respectively (see Table for PD-L1+). ORR was 49% vs 14% and CR rate was 10% vs 3% in the atezo + bev vs sun arms. Grade 3-4 AEs occurred in 27 pts (40%) with atezo + bev and 34 (49%) with sun. Using the MDASI scale, sarc pts reported longer median time to deterioration (TTD) of symptom interference with daily activities with atezo + bev vs sun (11.3 vs 4.9 mo). Prevalence of AngiogenesisHigh gene expression (GE) signature subset was lower (34% vs 65%) and T-effectorHigh GE subset was higher (54% vs 40%) in sarc vs non-sarc tumors. PD-L1+ disease was more common in sarc vs non-sarc tumors (63% vs 39%). Conclusions: mRCC pts with sarc histology had longer OS and PFS and a higher ORR/CR rate when treated with atezo + bev vs sun, regardless of PD-L1 status. Biomarker data support a biological correlate for the increased responsiveness to atezo + bev in sarc pts. Clinical trial information: NCT02420821.

All Sarc PD-L1+ Sarc Atezo + Bev Sun Atezo + Bev Sun n=68 n = 74 Stratified HR n=36 n = 50 Stratified HR mPFS, mo 8.3 5.3 0.52 8.6 5.6 0.45 (5.4, 12.9) (3.3, 6.7) (0.34, 0.79) (3.9, 15.3) (3.3, 6.7) (0.26, 0.77) mOS, mo NR 15.0 0.56 NR 15.0 0.53 (18.3, NR) (8.7, NR) (0.32, 0.96) (17.1, NR) (8.4, NR) (0.27, 1.06) 12-mo OS, % 69 60 — 71 61 — (58, 81) (48, 71) (56, 86) (47, 75) ORR, % 49 14 — 56 12 — (36, 61) (7, 23) (38, 72) (5, 24) CR, % 10 3 — 14 4 — TTD MDASI, mo 11.3 4.9 0.61 —— — (6.3, 17.5) (3.5, 7.6) (0.33, 1.11) m, median NR, not reached Parentheses denote 95% CI Stratification factors: MSKCC risk score, liver mets, PD-L1 status

4513 Poster Discussion Session; Displayed in Poster Session (Board #339), Mon, 1:15 PM-4:15 PM, Discussed in Poster Discussion Session, Mon, 4:30 PM-6:00 PM

CheckMate 214 post-hoc analyses of nivolumab plus ipilimumab or sunitinib in IMDC intermediate/poor-risk patients with previously untreated advanced renal cell carcinoma with sarcomatoid features.

David F. McDermott, Toni K. Choueiri, Robert J. Motzer, Osvaldo Rudy Aren, Saby George, Thomas Powles, Frede Donskov, Michael Roger Harrison, Jeronimo Rafael Rafael Rodriguez Cid, Yuko Ishii, M. Brent McHenry, Sabeen Fatima Mekan, Brian I. Rini, Nizar M. Tannir; Beth Israel Deaconess Medical Center, Dana-Farber/Harvard Cancer Center, Boston, MA; Dana-Farber Cancer Institute, Brigham and Women’s Hospital, and Harvard Medical School, Boston, MA; Memorial Sloan Kettering Cancer Center, New York, NY; Centro de Investigacion ´ Clı´nica Bradford Hill, Santiago, Chile; Roswell Park Cancer Institute, Buffalo, NY; Barts Cancer Institute, Queen Mary University of London, Royal Free NHS Trust, London, United Kingdom; Aarhus University Hospital, Aarhus, Denmark; Duke Cancer Institute, Durham, NC; Centro Oncologico, ´ Hospital Medica Sur, Mexico City, Mexico; Bristol- Myers Squibb, Princeton, NJ; Monter Cancer Ctr, New York, NY; Cleveland Clinic Taussig Cancer Institute, Cleveland, OH; University of Texas MD Anderson Cancer Center, Houston, TX

Background: Patients (pts) with advanced renal cell carcinoma with sarcomatoid features (sRCC) have poor prognosis and suboptimal outcomes with anti-VEGF targeted therapy. Nivolumab plus ipilimumab (N+I) demonstrated superior objective response rate (ORR) and overall survival (OS) vs sunitinib (S) in previously untreated pts with International Metastatic RCC Database Consortium (IMDC) intermediate/poor (I/P)-risk, clear-cell, advanced RCC in the phase 3 CheckMate 214 trial. Methods: We performed a post-hoc exploratory analysis of N+I vs S in CheckMate 214 sRCC pts. The presence of sarcomatoid features was assessed by keyword search for “sarcomatoid” in pts with available local pathology reports accompanying pretreatment tumor samples. Results: 842 (77%) of 1096 intention-to-treat pts had local pathology reports available, including 112 randomized pts with I/P-risk sRCC (N+I, n = 60; S, n = 52). Baseline characteristics of sRCC pts were balanced between arms. Notably, 47% vs 53% of I/P-risk sRCC pts in the N+I and S arms had tumor PD-L1 expression $1% at baseline, which was higher than in all I/P-risk pts (N+I, 26% vs S, 29%). In descriptive analyses performed at a minimum follow-up of 30 months, confirmed ORR and complete response rate per investigator (RECIST v1.1), OS, and progression-free survival (PFS) per investigator were improved with N+I vs S in I/P-risk pts with sRCC (Table). No new safety signals were seen in sRCC pts. Conclusions: In this post-hoc descriptive subgroup analysis of CheckMate 214, N+I demonstrated promising efficacy and prolonged survival vs S, with consistent safety, in previously untreated, I/P-risk, advanced clear-cell RCC with sarcomatoid features. Prospective studies of N+I that include pts with sRCC are ongoing. Clinical trial information: NCT02231749.

Randomized I/P-risk Randomized I/P-risk sRCC pts sRCC pts P value/HR (95% CI) N+I (N = 60) S (N = 52) ORR, % (95% CI) 56.7 (43.2–69.4) 19.2 (9.6–32.5) P, 0.001 Complete re- 18.3 0 – sponse, % Median PFS, mo 8.4 (5.2–24.0) 4.9 (4.0–7.0) HR (95% CI), 0.61 (0.38‒0.97); P, 0.03 Median OS, mo 31.2 (23.0‒NE) 13.6 (7.7‒20.9) HR (95% CI), 0.55 (0.33‒0.90); P, 0.0155 NE, not estimable

4514 Poster Discussion Session; Displayed in Poster Session (Board #340), Mon, 1:15 PM-4:15 PM, Discussed in Poster Discussion Session, Mon, 4:30 PM-6:00 PM

Efficacy of immune checkpoint inhibitors (ICI) and genomic characterization of sarcomatoid and/or rhabdoid (S/R) metastatic renal cell carcinoma (mRCC).

Ziad Bakouny, Natalie Vokes, Xin Gao, Amin Nassar, Sarah Abou Alaiwi, Ronan Flippot, Gabrielle Bouchard, John A. Steinharter, Pier Nuzzo, Wenting Pan, Abdallah Flaifel, Gwo-Shu Mary Lee, David A. Braun, Xiao X. Wei, Sabina Signoretti, Bradley Alexander McGregor, Lauren Christine Harshman, Eliezer Mendel Van Allen, Toni K. Choueiri; Dana-Farber Cancer Institute, Boston, MA; Brigham and Women’s Hospital, Boston, MA; Dana–Farber Cancer Institute, Boston, MA; Dana-Farber Cancer Institute, Brigham and Women’s Hospital, and Harvard Medical School, Boston, MA

Background: S/R mRCC are poorly characterized rapidly progressing tumors associated with poor prognosis. Although conventional therapies are less effective for these tumors, emerging data suggests that ICIs may be especially effective. Our aim was to characterize the genomic alterations (GA) in S/R mRCC tumors and evaluate their response to ICIs. Methods: We retrospectively compared the activity of first-line ICIs to non-ICI-based therapies for S/R mRCC patients (pts) treated at DFCI and analyzed sequencing data from an NGS panel (275-447 genes) on a subset of these patients (matched by histology to non-S/R mRCC). For S/R mRCC pts treated with ICI vs non-ICI therapies, overall survival (OS) and time to treatment failure (TTF) were compared by Cox regression and objective response rate (ORR) by logistic regression. GA frequencies were compared by Fisher’s test and tumor mutational burden (TMB) by Mann Whitney U between S/R and non-S/R mRCC. Results were considered statistically significant if p , 0.05 or q , 0.10. Results: 125 S/R mRCC pts were included (88 S, 23 R, 14 S&R) among which 103 were clear cell and 48 had sequencing data. GA in BAP1 were significantly more frequent in S/R vs non-S/R (25% vs 4.3%; q = 0.096) while other GA had similar frequencies and TMB (median [IQR]) was similar (7.2 [5.2- 8.4] vs 6.8 [5.3-9.1] mut/Mb; p = 0.98). Median follow-up was 35.4 (95% CI = 24.9 – 46.0) months (m). On multivariable analysis, S/R mRCC pts treated with ICI had significantly better clinical outcomes (Table). Conclusions: Pts with S/R mRCC have a higher frequency of BAP1 GA and better outcomes on ICIs compared to non-ICI-based therapies. Future studies should determine the molecular mechanisms underlying the improved response to ICIs in S/R mRCC.

ICI Non-ICI Univariable Multivariable* (N = 39) (N = 86; 90.7% TKI-based) (OR/HR) (adjusted OR/HR) ORR 16/37 (43.2%) 20/72 (27.8%) 1.98 (0.86 – 4.54) 3.02 (1.15 – 7.98) (N = 109) Median TTF 4.8 m (1.6 – 8.0) 4.9 m (3.6 – 6.2) 0.69 (0.45 – 1.05) 0.65 (0.41 – 1.01) (111/125 events) 24m OS rate 48.8% (30.6 – 67.0) 29.7% (19.7 – 39.7) 0.58 (0.34 – 0.99) 0.52 (0.30 – 0.90) (89/125 events)

* Adjusted for IMDC risk group and histology (clear cell vs non-clear cell)

4515 Poster Discussion Session; Displayed in Poster Session (Board #341), Mon, 1:15 PM-4:15 PM, Discussed in Poster Discussion Session, Mon, 4:30 PM-6:00 PM

Patient-reported outcomes (PROs) in IMmotion150: Atezolizumab (atezo) alone or with bevacizumab (bev) versus sunitinib (sun) in first-line metastatic renal cell carcinoma (mRCC).

Sumanta K. Pal, David F. McDermott, Michael B. Atkins, Bernard Escudier, Brian I. Rini, Robert J. Motzer, Lawrence Fong, Richard Wayne Joseph, Stephane Oudard, Alain Ravaud, Sergio Bracarda, Cristina Suarez Rodriguez, Elaine Tat Lam, Toni K. Choueiri, Beiying Ding, Caroleen Quach, Kenji Hashimoto, Christina Schiff, Elisabeth Piault, Thomas Powles; City of Hope National Medical Center, Duarte, CA; Beth Israel Deaconess Medical Center, Boston, MA; Georgetown Lombardi Comprehensive Cancer Center, Washington, DC; Gustave Roussy, Villejuif, France; Cleveland Clinic Taussig Cancer Institute, Cleveland, OH; Memorial Sloan Kettering Cancer Center, New York, NY; University of California San Francisco, San Francisco, CA; Mayo Clinic Hospital, Jacksonville, FL; Paris Descartes University, Paris, France; CHU Hopitaux de Bordeaux-Hopitalˆ Saint-Andre, ´ Bordeaux, France; Ospedale San Donato, Arezzo, Italy; Vall d’Hebron Institute of Oncology, Universitat Autonoma ` de Barcelona, Barcelona, Spain; University of Colorado Cancer Center, Denver, CO; Dana-Farber Cancer Institute, Brigham and Women’s Hospital, and Harvard Medical School, Boston, MA; Gen- entech, Inc., South San Francisco, CA; Roche Products Ltd., Welwyn Garden City, United Kingdom; Barts Cancer Institute, Queen Mary University of London, Royal Free NHS Trust, London, United Kingdom

Background: The phase 2 IMmotion150 study showed improved PFS with atezo + bev vs sun in patients (pts) whose tumors were PD-L1+ and corroborated the activity of atezo monotherapy in previously untreated mRCC. This study offers an opportunity to evaluate PROs with immunotherapy and VEGF- directed therapy alone and in combination. Methods: Pts randomized to atezo 1200 mg IV q3w alone or with bev15mg/kgIVq3w,orsun50mgpoqdfor4wkon/2wkoffcompletedtheMDAndersonSymptom Inventory (MDASI) and Brief Fatigue Inventory (BFI) q3w of each 6-wk cycle until progression (RECIST 1.1). Time to deterioration (TTD; first $ 2-point score increase over baseline [BL]) and change from BL (effect size [ES] $ 0.2 suggests a clinically important difference vs sun) in MDASI symptom severity and interference and BFI fatigue severity and interference scores are reported for all-comers. Results: Completion rates were . 90% at BL and $ 80% at most visits across arms. BL PRO scores indicating mild symptoms and interference were similar across arms. Delayed TTD of symptom severity and interference with daily life was seen with atezo vs sun and atezo + bev vs sun and was longest with atezo alone (Table). Pts reported milder symptoms and less interference during the first 6 cycles with atezo vs sun: ES mean (range) was 0.36 (0.07-0.68) for core symptom severity and 0.36 (0.04-0.83) for symptom interference. The 5 worst pt-reported symptoms during tx (dry mouth, fatigue, rash, drowsiness, lack of appetite) were all in the sun arm; all 16 symptoms measured were milder with atezo vs sun. Conclusions: PROs suggest that atezo alone or with bev maintained daily function with minimal symptom interference vs sun. Clinical activity, safety and PRO data for atezo support its investigation in adjuvant tx of high-risk RCC pts (IMmotion010; NCT03024996). Clinical trial information: NCT01984242.

TTD: HR (95% CI) vs Sun. Atezo Atezo + Bev MDASI core symptom severity 0.39 (0.22, 0.71) 0.74 (0.45, 1.20) MDASI RCC symptom severity 0.22 (0.12, 0.41) 0.60 (0.38, 0.94) MDASI symptom interference 0.36 (0.22, 0.58) 0.70 (0.47, 1.04) BFI fatigue severity 0.48 (0.33, 0.70) 0.75 (0.53, 1.06) BFI fatigue interference 0.38 (0.24, 0.60) 0.67 (0.46, 0.99)

4516 Poster Discussion Session; Displayed in Poster Session (Board #342), Mon, 1:15 PM-4:15 PM, Discussed in Poster Discussion Session, Mon, 4:30 PM-6:00 PM

Active surveillance in metastatic renal cell carcinoma (mRCC): Results from the Canadian Kidney Cancer information system (CKCis).

Igal Kushnir, Naveen S. Basappa, Sunita Ghosh, Aly-Khan A. Lalani, Denis Soulieres, Georg A. Bjarnason, Lori Wood, David Dawe, Christian K. Kollmannsberger, Daniel Yick Chin Heng, Anil Kapoor, Aaron Richard Hansen, Frederic Pouliot, M. Neil Reaume; Ottawa Hospital Cancer Centre, Ottawa, ON, Canada; Cross Cancer Institute, University of Alberta, Edmonton, AB, Canada; Cross Cancer Institute/University of Alberta, Edmonton, AB, Canada; Juravinski Cancer Centre, McMaster University, Hamilton, ON, Canada; Centre Hospitalier de l’Universite ´ de Montreal, ´ Montreal, QC, Canada; Sunnybrook Odette Cancer Centre, Toronto, ON, Canada; Dalhousie University, Halifax, NS, Canada; CancerCare Manitoba, University of Manitoba, Winnipeg, MB, Canada; BC Cancer–Vancouver Cancer Centre, Vancouver, BC, Canada; Tom Baker Cancer Centre, Calgary, AB, Canada; McMaster Institute of Urology, St Joseph’s Healthcare, Hamilton, ON, Canada; Princess Margaret Cancer Centre, Toronto, ON, Canada; Dept. of Surgery, Urology Division, Laval University, Quebec, QC, Canada; The Ottawa Hospital Cancer Centre, Ottawa, ON, Canada

Background: Active surveillance (AS) is a commonly used strategy in patients (pts) with low tumor burden or slow growing disease. However, few studies have assessed AS for mRCC compared to immediate treatment. We aimed to assess the outcomes and safety of AS in comparison to immediate systemic treatment for mRCC pts. Methods: Using CKCis, mRCC pts diagnosed between January 1, 2011 and December 31, 2016 were identified. AS strategy was defined as: (1) start of systemic therapy $6 months after diagnosis of mRCC; or (2) never receiving systemic therapy for mRCC with an overall survival (OS) $1yr(OS$ 1 yr a surrogate to exclude pts not started on treatment due to poor prognosis). Pts starting systemic treatment , 6 months after diagnosis of mRCC were defined as receiving immediate systemic treatment. OS and time until 1st line treatment failure (TTF) between the two cohorts were compared. Results: A total of 863 pts met criteria for AS (cohort A). Of these, 370 started treatment $ 6 months after their initial diagnosis (cohort A1) and 493 never received systemic treatment and were alive for $1 year (cohort A2). 848 pts received immediate systemic treatment (cohort B). Median age for pts in cohort A and B was 65.1 (19.0-91.5) vs. 62.2 yrs (23.1-87.1) (p , 0.0001). Sex distribution was not statistically different. Pts in cohort A had fewer sites of metastatic disease vs. cohort B ( , 0.0001) and 23% of pts in cohort A had metastasectomy vs. 5% in cohort B (P = , 0.0001). Five-year OS probability was significantly greater for cohort A than for cohort B (70.2% vs. 32.1%; P , 0.0001). After adjusting for IMDC risk criteria and age, both OS (HR 0.46, 0.38-0.56, P , 0.0001) and TTF (HR 0.79, 0.69-0.92, P = 0.0021) were greater in cohort A1 vs. B. For cohort A1 the median time on AS was 14.2 m (range 6 – 71). Conclusions: Based on the largest analysis of AS in mRCC to date, our data suggest that a subset of pts may be safely observed without immediate initiation of systemic therapy. Prospective validation is required in the contemporary immunotherapy era.

4517 Poster Discussion Session; Displayed in Poster Session (Board #343), Mon, 1:15 PM-4:15 PM, Discussed in Poster Discussion Session, Mon, 4:30 PM-6:00 PM

Safety and efficacy of nivolumab plus ipilimumab (NIVO+IPI) in patients with advanced renal cell carcinoma (aRCC) with brain metastases: Interim analysis of CheckMate 920.

Hamid Emamekhoo, Mark Olsen, Bradley Curtis Carthon, Alexandra Drakaki, Ivor John Percent, Ana M. Molina, Daniel C. Cho, Johanna C. Bendell, Lucio N. Gordan, Arash Rezazadeh Kalebasty, Daniel J. George, Thomas E. Hutson, Richard J. Lee, Tina C. Young, Jennifer Johansen, Scott S. Tykodi; University of Wisconsin School of Medicine and Public Health, Madison, WI; Oklahoma Cancer Specialists and Research Institute, Tulsa, OK; Emory University Winship Cancer Institute, Atlanta, GA; University of California, Los Angeles, Los Angeles, CA; Florida Cancer Specialists South/Sarah Cannon Research Institute, Port Charlotte, FL; Weill Cornell Medicine, New York, NY; New York University Langone Hospitals, New York, NY; Sarah Cannon Research Institute/Tennessee Oncology, Nashville, TN; Florida Cancer Specialists North/Sarah Cannon Research Institute, Gainesville, FL; Norton Cancer Institute, Louisville, KY; Duke University Medical Center, Durham, NC; Texas A&M College of Medicine, Bryan, TX; Bristol-Myers Squibb, Princeton, NJ; University of Washington and Fred Hutchinson Cancer Research Center, Seattle, WA

Background: Previous clinical trials of patients (pts) with aRCC, including CheckMate 214, have mostly excluded pts with brain metastases. However, antitumor activity in pts with brain metastases has been observed in pts with melanoma treated with NIVO 1 mg/kg + IPI 3mg/kg and pts with non-small cell lung cancer treated with NIVO 240 mg + IPI 1mg/kg. CheckMate 920 is an ongoing, phase 3b/4 clinical trial of NIVO + IPI treatment in pts with aRCC with a high unmet medical need. Here, we present the safety and efficacy interim results for the cohort of pts with brain metastases. Methods: Pts with previously untreated aRCC of any histology, with asymptomatic brain metastases (not on corticosteroids or receiving radiation), and Karnofsky performance status $70% were assigned to treatment with NIVO 3 mg/kg + IPI 1 mg/kg every 3 weeks for 4 doses, followed by NIVO 480 mg every 4 weeks. Pts were treated until disease progression, unacceptable toxicity, or for a maximum of 2 years. The primary endpoint was the incidence of high-grade immune-mediated adverse events (IMAEs). Key secondary endpoints included progression-free survival (PFS) and objective response rate (ORR) by RECIST v1.1 per investigator. Exploratory endpoints included additional safety analyses and overall survival (OS). Results: Overall, 28 patients were enrolled in the brain metastases cohort. With a minimum follow-up of 6.47 months, grade 3-4 IMAEs within 100 days of last dose were reported in 6 cases. The grade 3-4 IMAEs observed in $ 1 patient were diarrhea, colitis, diabetic ketoacidosis, immune-mediated hepatitis, hypophysitis, and rash of any type (n = 1 each). No treatment-related grade 5 IMAEs were reported. ORR by RECIST v1.1 per investigator in all treated subjects was 28.6% (95% CI 13.2–48.7). Median PFS in all treated subjects was 9.0 months (95% CI 2.9–not estimable [NE]). Median OS has not been reached (95% CI 13.1–NE). Conclusions: In pts with aRCC and brain metastases who are often excluded from clinical trials, NIVO + IPI treatment showed a safety profile consistent with previous reports of this dosing regimen, with encouraging antitumor activity. Clinical trial information: NCT02982954.

4518 Poster Discussion Session; Displayed in Poster Session (Board #344), Mon, 1:15 PM-4:15 PM, Discussed in Poster Discussion Session, Mon, 4:30 PM-6:00 PM

Phase II study of nivolumab and ipilimumab for advanced bladder cancer of variant histologies (BCVH).

Bradley Alexander McGregor, Matthew T Campbell, Wanling Xie, Arlene O. Siefker-Radtke, Amishi Yogesh Shah, Aradhana M. Venkatesan, Kerry L. Kilbridge, Guru Sonpavde, Glenn Bubley, Rana R. McKay, Toni K. Choueiri; Dana-Farber Cancer Institute, Boston, MA; The University of Texas MD Anderson Cancer Center, Houston, TX; UT MD Anderson Cancer Center, Houston, TX; Lank Center for Genitourinary Malignancy, Dana-Farber Cancer Institute, Boston, MA; Beth Israel Deaconess Medical Center, Boston, MA; Dana-Farber Cancer Institute, Brigham and Women’s Hospital, and Harvard Medical School, Boston, MA

Background: Patients with BCVH have poor outcomes and data regarding the management of this heterogeneous group of patients is limited. Nivolumab and ipilimumab has demonstrated safety and efficacy in urothelial carcinoma and other malignancies. In this multicenter, single arm, multi-cohort phase II trial we evaluate the efficacy of nivolumab and ipilimumab in patients with BCVH and other advanced rare genitourinary cancers (NCT 03333616). Herein, we report the preliminary results of the fully accrued BCVH cohort. Methods: Eligible patients had metastatic BCVH, ECOG performance status of 0-2 and were either untreated or had received any number of lines of prior therapy excluding prior immunotherapy. Patients underwent a baseline biopsy and blood collection for correlative studies and received treatment with nivolumab 3 mg/kg and ipilimumab 1 mg/kg intravenously every 3 weeks for 4 cycles with continued maintenance of nivolumab 480 mg IV every 4 weeks. The primary endpoint was overall response rate (ORR) by RECIST 1.1. Results: 19 BCVH patients were enrolled at 4 institutions between 4/2018 and 1/2019: squamous cell (n = 6), small cell (n = 3), adenocarcinoma (n = 3), urachal (n = 5), plasmacytoid (n = 1), and spindle cell (n = 1). 13 (68%) patients had received prior systemic therapy including platinum-based chemotherapy in 92% patients. Median number of cycles of ipilimumab plus nivolumab received was 3 (range 1-8) and median follow-up was 3.6 (0.3-8.8) months. 13 patients had undergone at least one scan; ORR was 31% (4/13, 80%CI: 14-52%), with partial responses seen in small cell carcinoma (n = 2), urachal (n = 1) and a complete response in 1 patient with plasmacytoid carcinoma. 3 patients (16%) developed treatment-related grade 3 toxicities with 1 (5%) grade 4 toxicity. Conclusions: Nivolumab and ipilmumab resulted in objective responses in a subset of patients with BCVH with manageable toxicities. Updated clinical and correlative data will be presented. This combination may warrant further investigation in patients with BCVH, which has substantial unmet needs. Clinical trial information: NCT 03333616.

4519 Poster Discussion Session; Displayed in Poster Session (Board #345), Mon, 1:15 PM-4:15 PM, Discussed in Poster Discussion Session, Mon, 4:30 PM-6:00 PM

Clinical outcomes according to PD-L1 status and age in the prospective international SAUL study of atezolizumab (atezo) for locally advanced or metastatic urothelial carcinoma (UC) or non-UC of the urinary tract.

Cora N. Sternberg, Axel Stuart Merseburger, Ernest Choy, Daniel E. Castellano, Fernando Lopez-Rios, Nicholas James, Giuseppe Luigi Banna, Ugo De Giorgi, Cristina Masini, Aristotelis Bamias, Xavier Garcia del Muro, Thomas Powles, Ignacio Duran, Craig Gedye, Marija Gamulin, Friedemann Zengerling, Lajos Geczi, Sabine de Ducla, Simon Fear, Yohann Loriot; Weill Cornell Medicine, New York, NY; University Hospital Schleswig-Holstein, Campus Lubeck, ¨ Lubeck, ¨ Germany; CREATE Centre, Section of Rheumatology, Division of Infection and Immunity, Cardiff University School of Medicine, Cardiff, United Kingdom; Medical Oncology Service, Hospital Universitario 12 de Octubre, Madrid, Spain; Hospital Universitario HM Sanchinarro, Madrid, Spain; Institute of Cancer and Genomic Services, University of Birmingham, and Cancer Centre, Queen Elizabeth Hospital, Birmingham, United Kingdom; Cannizzaro Hospital, Catania, Italy; IstitutoScientificoRomagnoloperloStudioelaCuradeiTumori(IRST)IRCCS,Meldola,Italy; Medical Oncology Unit, AUSL-IRCCS Reggio Emilia, Reggio Emilia, Italy; National and Kapodis- trian University of Athens, Alexandra Hospital, Athens, Greece; Institut Catala d’Oncologia, IDIBELL, Barcelona, Spain; Barts Cancer Institute, Experimental Cancer Medicine Centre, Queen Mary University of London, St Bartholomew’s Hospital, London, United Kingdom; Hospital Uni- versitario Virgen del Rocio, Seville (current affiliation: Hospital Universitario Marques de Valde- cilla, Santander, Spain; Calvary Mater Newcastle, Waratah, Australia; University Hospital Centre ‘Zagreb’, Zagreb, Croatia; Department of Urology, University Hospital Ulm, Ulm, Germany; National Institute of Oncology, Budapest, Hungary; F. Hoffmann-La Roche Ltd, Basel, Switzerland; Department of Cancer Medicine and INSERM U981, Universite ´ Paris-Sud, Universite ´ Paris-Saclay, Gustave Roussy, Villejuif, France

Background: Atezo, a monoclonal antibody targeting PD-L1, is an approved therapy for locally advanced/ metastatic UC based on IMvigor210 and IMvigor211 phase II and III trials. The single-arm SAUL study (NCT02928406) with a broader patient (pt) population demonstrated median overall survival (OS) of 8.7 months and a safety profile consistent with previous atezo trials. Methods: Pts with locally advanced/ metastatic UC or non-UC of the urinary tract received atezo 1200 mg every 3 weeks until disease progression or unacceptable toxicity. Populations excluded from IMvigor211 (renal impairment, ECOG PS 2, treated asymptomatic CNS metastases, stable controlled autoimmune disease, concomitant steroids, HIV positive, non-UC) were eligible. The primary endpoint was safety; OS and overall response rate (ORR) were secondary endpoints. Predefined subgroup analyses included outcomes according to PD-L1 status (VENTANA SP142) and age in the overall population (and the IMvigor211-like subgroup for PD-L1). Results: Between Nov 2016 and Mar 2018, 1004 pts were enrolled; 997 received atezo. Efficacy is summarized below. Incidences of grade $3 treatment-related adverse events were similar irrespective of PD-L1 status (overall IC 0/1 vs 2/3: 11% vs 16%; IMvigor211-like IC 0/1 vs 2/3: 11% vs 15%) or age ($65 y: 13%; $75 y: 12%; $80 y: 10%). Conclusions: OS and ORR appear more favorable in IC 2/3 vs IC 0/1 subgroups (overall and in the IMvigor211-like population). Atezo was effective and well tolerated across subgroups including elderly pts. Clinical trial information: NCT02928406.

All pts, PD-L1 IMvigor211-like, PD-L1 All pts, age IC 0/1 IC 2/3 IC 0/1 IC 2/3 ‡65 y ‡75 y ‡80 y Endpoint (n=666) (n=268) (n=427) (n=176) (n=624)a (n=227)a (n=78)a Deaths, n (%) 388 (58) 132 (49) 235 (55) 82 (47) 335 (54) 128 (56) 44 (56) Median OS, 7.9 11.6 9.0 14.5 8.5 8.3 8.3 months (6.8–9.1) (8.8–18.8) (7.8–10.4) (9.5–18.8) (7.5–10.9) (7.3–10.9) (5.4–11.2) (95% CI) 6-month OS rate, 57 (53–61) 67 (61–72) 61 (56–66) 72 (65–78) 60 (56–64) 61 (54–67) 59 (47–69) % (95% CI) ORR, % (95% CI) 10 (8–13) 21 (16–26) 10 (7–13) 23 (17–30) 14 (12–17) 13 (9–18) 8 (3–16) aSubgroups not mutually exclusive

4521 Poster Session (Board #347), Mon, 1:15 PM-4:15 PM

Correlation of methylthioadenosine phosphorylase (MTAP) loss with response to anti-folate therapy in urothelial bladder carcinoma (UBC).

Omar Alhalabi, Jianfeng Chen, Matthew T Campbell, Rebecca Slack Tidwell, Guangchun Han, Wei-Lien Wang, Jian H. Song, Sumankalai Ramachandran, Lidia Lopez, Anh Hoang, Arlene O. Siefker-Radtke, Mark Anton Titus, Charles Guo, Gary Gallick, Pavlos Msaouel, Eleni Efstathiou, Christopher Logothetis, Thai Huu Ho, Linghua Wang, Jianjun Gao; 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 M.D. Anderson Cancer Center, Houston, TX; Department of Pathology, The University of Texas MD Anderson Cancer Center, Houston, TX; MDACC, Houston, TX; Department of Genitourinary Medical Oncology, Division of Cancer Medicine, Houston, TX; Department of Genitourinary Medical Oncology, Division of Cancer Medicine, The University of Texas MD Anderson Cancer Center, Houston, TX; Mayo Clinic Arizona, Scottsdale, AZ; U.T. M.D. Anderson Cancer Center, Houston, TX

Background: The MTAP gene encodes an essential enzyme for the salvage pathway of nucleotide synthesis and is frequently deleted in UBC. Anti-folate agents such as pemetrexed can effectively inhibit the de novo pathway of nucleotide synthesis and as a result, create a synthetic lethality in MTAP deficient UBC. We hypothesize that MTAP gene loss correlates with enhanced response to pemetrexed in UBC. Methods: We investigated MTAP gene deletion rates in the TCGA database and determined MTAP protein loss rates by immunohistochemistry (IHC) using a UBC tissue microarray (TMA) from 151 patients (pts). We then performed in vitro and in vivo studies using MTAP proficient and MTAP deficient bladder cancer cell lines. At the clinical level, we performed a retrospective analysis based on MTAP status of pts treated with pemetrexed as 2nd line at our institution between 2014 and 2018. We are now enrolling pts in a single- arm, open-label, phase II clinical trial (NCT02693717) with pemetrexed in pts with MTAP deficient UBC. Results: Per our TCGA and TMA IHC analyses, MTAP deficiency rate was 25.9% and 27.8%, respectively. MTAP deficient UBC cell lines were at least 40 times more sensitive to pemetrexed than MTAP proficient lines. Knockdown of the MTAP gene increased apoptosis rate by pemetrexed from approximately 20% to 60%. Additionally, pemetrexed significantly inhibited the growth of MTAP deficient or knockdown xenograft tumors but not MTAP proficient tumors. Retrospective analysis of 12 pts using RECIST criteria indicated that all 4 MTAP deficient UBC pts responded to pemetrexed whereas only 1 of 8 (12.5%) MTAP proficient UBC pts responded. Of the 6 pts enrolled on the clinical trial, 3 (50%) had complete or partial response, 1 had stable disease, 1 was not evaluable and 1 had disease progression. Combined analysis of the entire experience demonstrates a higher response rate in MTAP deficient UBC (70%) as compared to MTAP proficient UBC (12.5%). Conclusions: Our preclinical and clinical data demonstrate that MTAP loss in UBC leads to a state of synthetic lethality when treated with pemetrexed and should be further investigated as a novel biomarker to predict response to anti-folate agents.

4522 Poster Session (Board #348), Mon, 1:15 PM-4:15 PM

Interim analysis of ibrutinib plus paclitaxel for patients with advanced urothelial carcinoma previously treated with platinum-based chemotherapy.

Daniel E. Castellano, Rafael Morales-Barrera, Ignacio Duran, Bhumsuk Keam, Ik Joo Chung, Hendrik-Tobias Arkenau, Ulka N. Vaishampayan, Mark Tuthill, Pablo Gajate, Sang Joon Shin, S. Paul Dang, Chia-Hsin Ju, Jennifer Lin, George W. Cole, Danelle Frances James, Oscar Reig; Medical Oncology Service, Hospital Universitario 12 de Octubre, Madrid, Spain; Vall d’Hebron Institute, Barcelona, Spain; Department of Medical Oncology, Hospital Universitario Marques ´ de Valdecilla, Santander, Spain; Department of Internal Medicine, Hemato-Oncology, Seoul National University Hospital, Seoul, South Korea; Chonnam National University Hwasun Hospital, Hwasun, South Korea; Sarah Cannon Research Institute, Cancer Institute, University College London, London, United Kingdom; Wayne State University, Detroit, MI; Oncology Department, Churchill Hospital, Oxford, United Kingdom; Hospital Ramon ´ y Cajal, Madrid, Spain; Division of Medical Oncology, Department of Internal Medicine, Yonsei University College of Medicine, Seoul, South Korea; Clearview Cancer Institute, Huntsville, AL; Pharmacyclics LLC, an AbbVie Company, Sunnyvale, CA; Pharmacyclics LLC, Sunnyvale, CA; Clinic and Provincial University Hospital, Barcelona, Spain

Background: Patients (pts) with advanced urothelial carcinoma (aUC) who progressed after platinum- based chemotherapy (PBCT) have a median overall survival (OS) of ~6 mo on CT and 8-10 mo on novel immune checkpoint inhibitors (ICIs). Treatment (tx) with ibrutinib (ibr), a first-in-class, once-daily inhibitor of Bruton’s tyrosine kinase (BTK) with immunomodulatory properties, plus paclitaxel (pac) may improve outcomes for pts with aUC. Methods: Pts with aUC who had 1-2 prior tx ($1 PBCT and/or ICIs) formed one cohort of this phase 1b/2 study. Recommended phase 2 dose per phase 1b was ibr 840 mg/d + weekly pac 80 mg/m2 in 21-d cycles; starting ibr dose of 560 mg/d. nanoString gene expression assay (NanoString Technologies) was used for preliminary assessment of ibr-targeted kinases in baseline tumor biopsies. Results: This interim analysis included the first 29 pts treated (ibr 840/560 mg/d, n = 25/4; median age 67 y). 52% had 2 prior regimens (prior PD-L1 inhibitors, 100%). Median time on study/tx was 7.2/2.3 mo (max tx: 11.5 mo). Unconfirmed overall response rate (ORR) was 41% (complete/partial response: 10%/31%). Disease control rate was 62%. ORR was 71% for 7 pts with only lymph node metastases; ORR was 36% for 14 pts with 1 prior tx and 47% for 15 pts with 2 prior tx. Median duration of response was 4.2 mo (90% CI: 1.9-7.1). Median progression-free survival was 3.6 mo (90% CI: 1.6-5.4). Median OS was 14.7 mo (90% CI: 7.7- 15.9). Follow-up is ongoing for durability/survival with 6 pts alive without progression, including 4 still on ibr. 76% of pts had grade $3 adverse events (AEs); 4 discontinued ibr + pac due to AEs and 1 had major hemorrhage. Pac exposure was not impacted by ibr; pac geometric mean Cmax (2066 ng/mL) and AUC0-‘ (3813 ng$h/mL) were consistent with historical data. Preliminary tumor biopsy data showed a nonsignificant trend of higher BTK/ITK/BMX expression in tx responders; analyses are ongoing. Conclusions: Ibr + pac shows promising results for ORR/OS in pts with aUC previously treated with PBCT and ICIs. No unexpected safety signals were seen; ibr + pac seemed tolerable. Follow-up of these and additional pts will support the activity and safety of ibr + pac in this study. Clinical trial information: NCT02599324.

4523 Poster Session (Board #349), Mon, 1:15 PM-4:15 PM

Circulating cell-free DNA (cfDNA) levels and fragmentation pattern can distinguish nonmuscle invasive (NMI) from muscle-invasive (MI) and metastatic (met) bladder cancer (BC).

Jaleh Fallah, Shinjini Ganguly, Hong Li, Wei (Auston) Wei, Aysegul Balyimez, Thirunavukkarasu Sitalaximi, C. Marcela Diaz-Montero, Patricia A. Rayman, Marcelo Lamenza, Priscilla Dann, Donna Company, Rahul D. Tendulkar, Mohamed Abazeed, Jorge A. Garcia, Moshe Chaim Ornstein, Brian I. Rini, Byron H Lee, Petros Grivas, Omar Y. Mian; Cleveland Clinic Foundation, Cleveland, OH; Cleveland Clinic, Cleveland, OH; Cleveland Clinic Lerner Research Institute, Cleveland, OH; Cleveland Clinic Taussig Cancer Insitute, Cleveland, OH; Cleveland Clinic Taussig Cancer Institute, Cleveland, OH; University of Wash- ington, School of Medicine, Seattle, WA; Cleveland Clinic, Dept. of Radiation Oncology, Dept. of Trans- lational Hematology Oncology Research, Cleveland, OH

Background: Occult MI and met BC may be under-staged. Circulating cfDNA may be a dynamic, low-cost and minimally invasive biomarker. We evaluated correlations between total circulating cfDNA and presence of MIBC and met BC. We hypothesized that the relative abundance of circulating low molecular weight cfDNA would correlate with BC stage. Methods: Peripheral blood from pts with BC was collected in Streck BCT tubes and processed to obtain cf nucleic acid extracts. Total cfDNA quantity (ng/ml) was assessed by fluorimetry. cfDNA fragment size was measured by Bioanalyzer DNA analysis. Wilcoxon rank sum test and Fisher’s Exact test were used to compare cfDNA quantity and fragmentation pattern among pts with NMIBC, MIBC, met BC. Results: Blood was obtained from 58 pts with BC (20% women, 34% never smokers, median age 71 (29-89). There was no significant difference in cfDNA between MIBC and met BC, however, it was significantly lower in pts with NMIBC vs MIBC and met BC (table). The concentration of low molecular weight fragments (LMW-frags) (100 - 400) base pairs and the ratio of LMW-Frag to cfDNA were significantly different between pts with NMIBC and pts with MIBC or met BC (table). Using median values as the cutoff, there was a significantly higher proportion of pts with cfDNA . 7 ng/ml and LMW-frags . 1.6 ng/mL, in MIBC & met BC vs NMIBC (p , 0.001). The % of pts with LMW- frags to cfDNA . 30%, was significantly different among NMIBC, MIBC and met BC groups: 16%, 53%, 78%, respectively (p , 0.001). Conclusions: This exploratory study suggests that cfDNA levels may correlate with BC stage. Measuring the relative abundance of LMW-frags with the expected size of cf DNA can enhance the specificity of cfDNA analysis for distinction between MIBC and met BC. Further studies are needed to confirm findings and define the optimal cut-points for optimal BC staging.

NMIBC MIBC met BC P-value (N 19) (N 15) (N 24) (NMIBC vs MIBC & met BC) Median cfDNA (ng/ml), 0.7 9.7 9.3 , 0.001 (IQR) (0.4,1.2) (4.0,13.3) (5.9,15.3) Median LMW-frags (ng/ml), 0.0 3.3 3.8 , 0.001 (IQR) (0.0,0.0) (0.0,8.4) (1.4,31.5) % LMW-Frags to cfDNA, 0.0 34.4 50.3 , 0.001 (IQR) (0.0,0.0) (0.0,60.8) (31.2,79.3)

4524 Poster Session (Board #350), Mon, 1:15 PM-4:15 PM

Nivolumab monotherapy in patients with advanced platinum-resistant urothelial carcinoma: Efficacy and safety update from CheckMate 275.

Arlene O. Siefker-Radtke, Ari David Baron, Andrea Necchi, Elizabeth R. Plimack, Sumanta K. Pal, Jens Bedke, Yousef Zakharia, Marc-Oliver Grimm, Sergio Bracarda, Margitta Retz, Chikara Ohyama, Gary Grossfeld, Sandra Collette, Padmanee Sharma, Matt D. Galsky; The University of Texas MD Anderson Cancer Center, Houston, TX; California Pacific Medical Center, San Francisco, CA; Fonda- zione IRCCS Istituto Nazionale dei Tumori, Milan, Italy; Fox Chase Cancer Center, Philadelphia, PA; City of Hope National Medical Center, Duarte, CA; University Tubingen, ¨ Tubingen, ¨ Germany; University of Iowa, Iowa City, IA; Jena University Hospital, Jena, Germany; Azienda Ospedaliera S. Maria, Terni, Italy; Rechts der Isar Medical Center, Technical University of Munich, Munich, Germany; Hirosaki University, Hirosaki Aomori, Japan; Bristol-Myers Squibb, Princeton, NJ; Department of Medicine, Icahn School of Medicine at Mount Sinai, Tisch Cancer Institute, New York, NY

Background: In the open-label, single-arm, phase 2 CheckMate 275 trial, objective response rate (ORR) for patients (pts) with metastatic urothelial carcinoma (mUC) with nivolumab (NIVO) was 20.4% with minimum follow-up of 21.3 mo. Here, we report updated efficacy and safety data with minimum follow-up of 33.7 mo. Methods: Pts with platinum-resistant locally advanced or metastatic urothelial carcinoma received NIVO 3 mg/kg until disease progression or unacceptable toxicity. The primary endpoint was ORR by blinded independent review committee (BIRC) by RECIST v1.1 (including duration of response [DOR]). Secondary endpoints included progression-free survival (PFS) by BIRC, overall survival (OS), and ORR per investigator. Efficacy was evaluated in all treated pts and by tumor PD-L1 expression. Safety and PFS by investigator were exploratory endpoints. Results: ORR by BIRC was 20.7% (95% CI 16.1–26.1) including 18 (7%) complete responses (CR; with 1 additional CR since the last report; Table). ORR per investigator was similar (24.8%). Median DOR by BIRC was 20.3 mo (95% CI 11.5–31.3). Of 56 pts with best overall response (BOR) of CR or partial response (PR), 59% had a DOR $12 mo. Median PFS (mPFS) was 1.9 mo per BIRC (95% CI 1.9–2.3; Table) and 2.0 mo per investigator (95% CI 1.9–2.5). Median OS (mOS) was 8.6 mo (95% CI 6.1–11.3; Table). 12, 24, and 36-mo OS rates were 40%, 30%, and 22%. While efficacy was numerically higher in pts with tumor PD-L1 expression $1%, efficacy was observed in all pts (Table). Any-grade treatment-related adverse events occurred in 69% of pts (grade 3–4, 25%), mostly (59%) within the first 3 mo of initiating therapy. Conclusions: With long-term follow-up from CheckMate 275, NIVO continues to provide durable antitumor activity in pts with mUC. No new safety signals were noted. Clinical trial information: NCT02387996.

All treated pts PD-L1 < 1% PD-L1 ‡1% N = 270 N = 146 N = 124 ORR (95% CI) 20.7 (16.1–26.1) 16.4 (10.8–23.5) 25.8 (18.4–34.4) BOR by BIRC, n (%) CR 18 (7) 6 (4) 12 (10) PR 38 (14) 18 (12) 20 (16) Stable disease 56 (21) 24 (16) 32 (26) Progressive disease 111 (41) 71 (49) 40 (32) mPFS by BIRC (95% CI), mo 1.9 (1.9–2.3) 1.9 (1.7–2.0) 3.5 (1.9–3.7) mOS (95% CI), mo 8.6 (6.1–11.3) 6.0 (4.4–8.1) 11.9 (9.1–19.1)

4525 Poster Session (Board #351), Mon, 1:15 PM-4:15 PM

Outcomes of patients (pts) with metastatic urothelial cancer (mUC) and poor performance status (PS) receiving anti-PD(L)1 agents.

Ali Raza Khaki, Leonidas Nikolaos Diamantopoulos, Ang Li, Michael Edward Devitt, Alexandra Drakaki, Evan Shreck, Monika Joshi, Pedro Isaacsson Velho, Lucia Alonso, Ariel Ann Nelson, Sandy Liu, Marcus W. Moses, Pedro C. Barata, Christopher J. Hoimes, Matt D. Galsky, Guru Sonpavde, Evan Y. Yu, Veena Shankaran, Gary H. Lyman, Petros Grivas; University of Washington, Seattle, WA; Fred Hutch Cancer Research Center, Seatle, WA; University of Virginia, Charlottesville, VA; University of California, Los Angeles, Los Angeles, CA; Montefiore Medical Center, Bronx, NY; Penn State Health Milton S. Hershey Medical Center, Hershey, PA; Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins, Baltimore, MD; Hospital National MARQUES VALDECILLA, Santander, Spain; Medical College of Wisconsin, Cleveland, OH; Department of Hematology and Oncology, David Geffen School of Medicine at University of California, Los Angeles, CA; Tulane University, New Orleans, LA; Cleveland Clinic, Cleveland, OH; University Hospitals Cleveland Medical Center/Case Western Reserve University, Cleveland, OH; Department of Medicine, Icahn School of Medicine at Mount Sinai, Tisch Cancer Institute, New York, NY; Dana-Farber Cancer Institute, Boston, MA; Fred Hutchinson Cancer Research Center, Seattle, WA; University of Washington, School of Medicine, Seattle, WA

Background: Anti-PD(L)1 immune checkpoint inhibitors (ICI) prolong overall survival (OS) after platinum chemotherapy in mUC. However, clinical outcomes in pts with poor PS at time of ICI initiation are unknown. We hypothesized that ICI initiation in pts with ECOG PS 2-3 would be associated with worse outcomes vs. pts with ECOG PS , 2, and impact death location. Methods: A retrospective cohort study in 8 institutions identified pts with mUC who received ICI. Demographic, clinicopathologic, treatment (tx) patterns, tx response, and outcomes were collected. Primary endpoint: overall response rate (ORR). Secondary endpoints: median (m) OS in pts receiving ICI as 1st and 2nd line (1L, 2L); odds of dying in hospital (vs elsewhere) for pts receiving ICI (vs no tx) within 30 days of death; and estimated drug cost for pts with ICI within 30 days of death based on average wholesale price. Unadjusted logistic regression was used to assess association between ORR and ECOG PS (2-3 vs , 2) and wald test was used to compare mOS between ECOG PS (2-3 vs , 2). Results: 194 consecutive pts (30% women, 41% never smokers, median age at diagnosis 69) treated with ICI for mUC were identified. Median number of total tx lines was 2; all pts received $1 ICI line (6 pts received 2 ICI lines); 97, 79, 17 and 7 pts received ICI in 1L, 2L, 3L and 4L, respectively; 26% pts with ICI in 1L and 2L had ECOG PS 2-3. ORR and mOS are shown in table. Among 106 pts who died, 96 had available death location; of those, 8% received ICI within 30 days of death. Starting ICI within 30 days of death (vs no tx) was associated with higher odds of hospital death (OR 6.05, 95%CI 1.3-27.6). Estimated average ICI cost/pt within 30 days of death was $1400.58. Conclusions: Pts with ECOG PS 2-3 at time of ICI initiation had similar ORR vs ECOG PS , 2 but worse mOS. ICI initiation within 30 days from death was associated with higher likelihood of hospital death. ICI may not circumvent the negative prognostic role of poor PS, so biomarker-based pt selection is critical. Limitations include lack of adjustment for selection bias and other confounders at time of ICI initiation; data validation is ongoing.

ECOG PS 1L ORR 1L mOS (mo) 2L ORR 2L mOS (mo) < 2 28% 14 23% 16 2-3 27% 6 35% 5 p value NS 0.002 NS , 0.001

4526 Poster Session (Board #352), Mon, 1:15 PM-4:15 PM

Treatment (tx) characteristics of patients (pts) with locally advanced or metastatic urothelial cancer (mUC) receiving checkpoint inhibitor (CPI) monotherapy in a US clinical practice.

Alicia K. Morgans, Osama E. Rahma, Shivani K. Mhatre, Ching-Yi Chuo, Jessica Davies, Jorge Martinalbo, Nicole N. Davarpanah, Russell Pachynski; Northwestern University, Chicago, IL; Dana-Farber Cancer Institute, Boston, MA; Genentech, Inc., South San Francisco, CA; Roche Products, Ltd., Welwyn Garden City, United Kingdom; F. Hoffmann-La Roche, Ltd, Basel, Switzerland; Division of Oncology, Washington University Medical School, St. Louis, MO

Background: Approval of anti–PD-L1/anti–PD-1 CPI agents has changed the mUC tx landscape, but real- world (RW) tx patterns are not well described. Here, we describe pt characteristics, time on tx (TOT), tx- cycle distribution, relative dose intensity (RDI) and subsequent tx for pts receiving atezolizumab (atezo), nivolumab (nivo) or pembrolizumab (pembro) monotherapy. Methods: Pts diagnosed with mUC who completed atezo, nivo or pembro in the first-line (1L) or prior-platinum second-line and beyond (2L+) settings by April 30, 2018, were identified from the US-based Flatiron Health electronic health record– derived database. TOT was defined as time from first to last CPI administration + 1 cycle, tx cycles as number of CPI doses received during TOT and RDI as ratio of actual to planned dose per week to reflect any dose interruption. Results: RW data from pts receiving atezo, nivo and pembro were analyzed (Table). Up to 38% of pts had ECOG PS . 1. Median TOT ranged from 2.1-2.8 mo, with overlapping 95% CIs; mean TOT ranged from 2.7-4.1 mo. Over 50% of pts had # 4 tx cycles. 21%-38% of pts did not have RDI within 95%-105% of the labeled dose. Most common subsequent txs were platinum-based chemotherapy combinations with gemcitabine or taxanes (post–1L CPI) and taxane monotherapy or other CPI monotherapy/combinations (post–2L+ CPI). Conclusions: Here, we present the largest analysis of RW CPI use in mUC to date. Overall, this unadjusted descriptive analysis showed relative comparability of pt and tx characteristics and TOT across CPI-treated groups. Insights into RW tx allow for an understanding of how clinical trial data translate to broader pt populations, including those with ECOG PS . 1, and may be useful for practitioners.

Atezo, 1La Atezo, 2L+ Nivo, 2L+ Pembro, 1La Pembro, 2L+ n 79 174 55 34 27 ‡ 65 years, % 86 84 69 88 78 ECOG PS > 1, %b 38 19 6 33 20 Median TOT, mo (95% CI) 2.1 (0.7, 4.2) 2.8 (1.4, 4.9) 2.0 (1.2, 3.6) 2.8 (1.4, 3.8) 2.7 (1.4, 3.5) Mean TOT, mo (SD) 2.9 (2.9) 4.0 (3.8) 2.7 (2.1) 2.9 (1.9) 2.9 (1.8) > 4 tx cycles, % 35 43 49 41 30 95%-105% RDI, % 76 79 62 68 70 a Unknown cisplatin eligibility and prior (neo)adjuvant platinum use. b % based on pts with known ECOG PS.

4527 Poster Session (Board #353), Mon, 1:15 PM-4:15 PM

Durability of complete response (CR) with atezolizumab (atezo) in locally advanced/ metastatic urothelial carcinoma (mUC).

Yohann Loriot, Arjun Vasant Balar, Robert Dreicer, Jean H. Hoffman-Censits, Jose Luis Perez-Gracia, Daniel Peter Petrylak, Michiel Simon Van Der Heijden, Xiaodong Shen, Qian (Cindy) Zhu, Beiying Ding, Constanze Kaiser, Jonathan E. Rosenberg; Institut de Cancerologie ´ Gustave Roussy, Villejuif, France; Perlmutter Cancer Center at NYU Langone Health, New York, NY; University of Virginia, Charlottesville, VA; Johns Hopkins Sidney Kimmel Comprehensive Cancer Center, Baltimore, MD; Department of Medical Oncology, Clinica Universidad de Navarra, Pamplona, Spain; Yale School of Medicine, New Haven, CT; Netherlands Cancer Institute, Amsterdam, Netherlands; Oncology, Genentech, Inc., South San Francisco, CA; Genentech, Inc., San Francisco, CA; Genentech, Inc., South San Francisco, CA; Genentech, South San Francisco, CA; Memorial Sloan Kettering Cancer Center, New York, NY

Background: Atezo (anti–PD-L1) has been shown to elicit CRs in a number of mUC patients (pts) in clinical trials. We sought to describe the kinetics, durability and outcomes associated with these CRs in Ph I (PCD) and II (IMvigor210) atezo studies, each with long-term follow-up. Methods: In PCD (pre-treated mUC) and IMvigor210 (Cohort 1, cisplatin-ineligible untreated mUC; Cohort 2, platinum-treated mUC), pts received atezo per protocol (Petrylak JAMA Oncol 2018; Balar Lancet 2017; Rosenberg Lancet 2016). This post hoc analysis descriptively assessed pt disposition, time to and duration of RECIST 1.1 response and overall survival in pts with CR. Results: CR rates were 13%, 8% and 7% in PCD, IMvigor210 Cohort 1 and Cohort 2, respectively. First response was PR in most pts with CR. Median CR duration was . 3 y in PCD, not estimable (NE) in IMvigor210 Cohort 1 and . 2 y in Cohort 2 (Table). At data cutoff, all but 2, 0 and 1 pts were alive, respectively; across studies, $ 40% of pts with CR were on treatment. CR pts had a first response (PR/CR) by a median of 3.5 cycles. Further pt characteristics and survival outcomes will be reported. Conclusions: Across Ph I/II atezo mUC studies, CRs appeared durable (median duration . 2 y) despite small pt numbers. Most pts with CR were alive, with responses ongoing after long-term follow-up (median follow-up . 30 mo). Clinical trial information: NCT01375842, NCT02951767, NCT02108652.

IMvigor210 Cohort 1 IMvigor210 Cohort 2 PCD (n = 12)a (n = 10)b (n = 22)c Median atezo duration (range), mo 33.6 (10.4- 21.5 (6.2-35.9) 32.6 (8.8-35.6) 44.4) Median time to first response (range), mod 1.4 (1.2-8.3) 2.2 (2.0-10.3) 2.0 (1.9-8.2) Median doses to first response (range), n 2.0 (2-12) 3.5 (3-15) 3.0 (3-12) Median time to CR (range), mo 8.1 (1.2-28.3) 6.2 (2.1-13.8) 4.2 (1.9-13.2) Median CR duration (95% CI), mo 37.5 (33.3, NE 28.4 (18.6, NE) 37.5) Pts with ongoing CR, n (%)e 8 (67) 8 (80) 12 (55) Pts with CR as first response, n (%) 2 (17) 3 (30) 8 (36) Treatment discontinuations due to PD, n (%) 2 (17) 0 3 (14) Treatment discontinuations due to non-PD, 2 (17) 6 (60) 3 (14) n (%)f

PD, progressive disease. Data cutoff (median follow-up, mo): a Dec 31, 2016 (40.8); b Jul 12, 2017 (31.7); c Jul 12, 2017 (33.1). d First response of CR or PR. e No PD/death. f Excluded 1 PCD pt who completed only 16 cycles per an earlier protocol.

4528 Poster Session (Board #354), Mon, 1:15 PM-4:15 PM

Association of cell-free DNA (cfDNA) levels with myeloid-derived suppressor cells (MDSC) levels in blood of patients (pts) with muscle invasive (MI) and metastatic (met) bladder cancer (BC).

Jaleh Fallah, Shinjini Ganguly, Patricia A. Rayman, Wei (Auston) Wei, Aysegul Balyimez, Thirunavukkarasu Sitalaximi, Marcelo Lamenza, Kevin L. Stephans, Priscilla Dann, Donna Company, Rahul D. Tendulkar, Mohamed Abazeed, Jorge A. Garcia, Brian I. Rini, Byron H Lee, Moshe Chaim Ornstein, Petros Grivas, Omar Y. Mian, C. Marcela Diaz-Montero; Cleveland Clinic Foundation, Cleveland, OH; Cleveland Clinic, Cleveland, OH; Cleveland Clinic Lerner Research Institute, Cleveland, OH; Cleveland Clinic Taussig Cancer Insitute, Cleveland, OH; Cleveland Clinic Taussig Cancer Institute, Cleveland, OH; University of Washington, School of Medicine, Seattle, WA; Cleveland Clinic, Dept. of Radiation Oncology, Dept. of Translational Hematology Oncology Research, Cleveland, OH

Background: cfDNA can be detected in healthy individuals but higher concentrations are present in pts with cancer. MDSC are immature immunosuppressive cells that can be mobilized from bone marrow by tumor-related factors. Higher blood MDSC levels have been associated with worse outcomes in pts with solid tumors including BC. We assessed correlations between cfDNA and MDSC levels in pts with MIBC and met BC. Methods: Peripheral blood from pts with MIBC and met BC was collected in Streck BCT tubes and processed to obtain cf nucleic acid extracts. Total cfDNA was determined by fluorimetry. Cell-free DNA fragment size was measured by Bioanalyzer DNA analysis; 100-400 bp fragments (mono- and di- nucleosomal fragments linked to granulocytic processing of apoptotic and necrotic tumor cells) were designated low molecular weight (LMW-frags). The % of MDSC (CD33+/HLADR-) and subtypes were measured. MDSC subtypes were defined as polymorphonuclear (PMN-MDSC: CD15+/CD14-), monocytic (M-MDSC: CD15-/CD14+) and uncommitted (UNC-MDSC: CD15-/CD14-). Spearman’s correlation test was used for analysis. Results: Blood from 37 pts (19% women, 40% never smokers) with MIBC or met BC was collected: 15 (41%) with MIBC and 22 (59%) with met BC at time of collection. There was a significantly positive correlation between total MDSC and cfDNA levels (r = 0.57, P = 0.0003). Among MDSC subtypes, there was a significantly positive correlation between PMN-MDSC and cfDNA levels (r = 0.61, P , 0.0001). The higher level of LMW-frags was significantly but moderately associated with higher total MDSC (r = 0.43, P 0.008) and PMN-MDSC (r = 0.41, P 0.01) levels. There was no significant correlation between cfDNA level and other MDSC subtypes. Conclusions: There was a positive correlation between total and PMN-MDSC with cfDNA levels in blood from pts with MIBC and met BC. That may suggest a putative role for MDSC in mediating cfDNA release into the circulation, consistent with prior reports of granulocyte-mediated ctDNA processing. Further studies need to identify mechanisms and implications of our findings and potential correlation with clinical outcomes.

4529 Poster Session (Board #355), Mon, 1:15 PM-4:15 PM

Molecular biomarker analysis and survival in patients (pts) with advanced urothelial cancer (UC) previously treated with chemotherapy.

Bernadett Szabados, Marlon Rebelatto, Craig Barker, Alvin Milner, Arthur Lewis, Michael Stokes, Magda Zajac, Thomas Powles, Norah J Shire; Barts Cancer Centre, Queen Mary University of London, London, United Kingdom; MedImmune, Gaithersburg, MD; AstraZeneca, Cambridge, United Kingdom; MedImmune, Cambridge, United Kingdom; Real-world Evidence, Evidera, Waltham, MA; Barts Cancer Institute, Queen Mary University of London, Royal Free NHS Trust, London, United Kingdom; AstraZeneca, Gaithersburg, MD

Background: The biomarkers PD-L1, FOXP3, and CD8 have been explored in pts with advanced UC who progressed after platinum-based chemotherapy (CTx). However, their relevance earlier in the disease process is less well understood. Methods: The Phase 2/3 LaMB study (NCT00949455) compared maintenance lapatinib vs placebo after first-line (1L) platinum-based CTx in pts with HER1/HER2- overexpressing stage IV advanced UC. Pre-CTx archival samples from this study were retrospectively analyzed and included both randomized and screen failure pts. PD-L1 expression was assessed (VENTANA SP263 Assay) and categorized as high ($25% of tumor cells [TC] and/or immune cells [IC]) or low/negative ( , 25% TC and IC). Overall survival (OS) and progression-free survival (PFS) were estimated via Kaplan-Meier method; results were stratified by PD-L1 expression. The exploratory biomarkers CD8 and FOXP3 were also analyzed. The prognostic significance of the biomarkers was explored by multivariable Cox proportional hazards models and a bootstrap method for model selection. Results: Of 446 pts (232 randomized; 214 screened), 243 (54.5%) were assessed for PD-L1 expression, with 61 (25.1%) PD-L1 high and 158 (65.0%) PD-L1 low/negative. In PD-L1 high and low/negative pts, respectively, median OS (95% CI) was 12.0 (9.4–19.7) vs 12.5 months (10.4–15.5); median PFS (95% CI) was 6.5 (3.5–8.8) vs 5.0 months (4.3–6.3). PD-L1 expression was not associated with OS or PFS in univariate analysis or in a multivariate model for OS (hazard ratio [HR] for PD-L1 high vs low/negative 1.4 [95% CI, 0.8–2.3]). In a multivariate model for PFS, PD-L1 expression improved accuracy of the model by 23% and was a significant variable (HR, 2.1 [95% CI, 1.2–3.5]). Results of analyses of CD8 and FOXP3 will also be reported. Conclusions: Overall, these data suggest a lack of association between PD-L1 expression and survival in pts receiving 1L platinum-based CTx. Mechanisms underlying the potential association of PD-L1 expression with PFS remain unclear. CD8 and FoxP3 exploratory analyses may help to elucidate these results. Clinical trial information: NCT00949455.

4530 Poster Session (Board #356), Mon, 1:15 PM-4:15 PM

Pembrolizumab (pembro) for patients (pts) with high-risk (HR) non–muscle invasive bladder cancer (NMIBC) unresponsive to Bacillus Calmette-Gu´erin (BCG): Updated follow-up from KEYNOTE-057.

Ronald De Wit, Girish S. Kulkarni, Edward M. Uchio, Laurence Eliot Miles Krieger, Joost L. Boormans, Mathieu Roumiguie, ´ Eric A. Singer, Dean F. Bajorin, Ashish M. Kamat, Petros Grivas, Ho Kyung Seo, Hiroyuki Nishiyama, Badrinath R. Konety, Kijoeng Nam, Ekta Kapadia, Tara L. Frenkl, Arjun Vasant Balar; Erasmus University Medical Center, Rotterdam, Netherlands; UHN Princess Margaret Cancer Center, University of Toronto, Toronto, ON, Canada; UC Irvine Health, Orange, CA; Royal North Shore Hospital, Northern Cancer Institute, St Leonards, NSW, Australia; Institut Universitaire du Cancer de Toulouse Oncopole CHU, Toulouse, France; Rutgers Cancer Institute of New Jersey, New Brunswick, NJ; Memorial Sloan Kettering Cancer Center, New York, NY; The University of Texas MD Anderson Cancer Center, Houston, TX; University of Washington, Seattle, WA; National Cancer Center, Goyang, South Korea; University of Tsukuba, Tsukuba, Japan; University of Minnesota, Minneapolis, MN; Merck & Co., Inc., Kenilworth, NJ; Perlmutter Cancer Center, NYU Langone Health, New York, NY

Background: Upregulation of the PD-1 pathway has been observed in BCG-unresponsive NMIBC, suggesting that pembro may be beneficial. Efficacy and safety of pembro in pts with HR, BCG- unresponsive NMIBC was evaluated in the single-arm phase 2 KEYNOTE-057 study (NCT02625961); updated results for pts with carcinoma in situ (CIS) with or without papillary tumors (cohort A) are reported. Methods: Pts with histologically confirmed HR, BCG-unresponsive CIS with or without papillary tumors who received adequate BCG therapy and were unable/refused to undergo radical cystectomy received pembro 200 mg Q3W for 24 mo or until recurrence, progression, or unacceptable toxicity. Pts who developed HR NMIBC or progressive disease during treatment were required to discontinue. Key end points were complete response rate (CRR), duration of response, and safety. Results: 102 pts (median age, 73 years; CIS alone, 63.7%; median number of prior BCG instillations, 12) had enrolled in cohort A as of enrollment cutoff. Median (range) duration of follow-up was 15.8 mo (4.6-28.2); 3-mo CRR was 40.2% (95% CI, 30.6-50.4) by central assessment. Among 41 pts who had CR at 3 mo, median CR duration was 12.7 mo (range, 0+ to 20.5+ mo); 75.4% had a CR duration $6 mo; 52.6% had a CR duration $12 mo (Kaplan-Meier method); 24 pts (58.5%) maintained CR at last follow-up, and 15 (36.6%) experienced recurrent NMIBC after CR; at the time of analysis, none progressed to muscle-invasive or metastatic disease. CRR was 44.6% for pts with CIS alone (n = 65), 41.7% for CIS with T1 tumors (n = 12), and 28.0% for CIS with high-grade Ta tumors (n = 25). Treatment-related adverse events (AEs) occurred in 66 (64.7%) pts; most frequent ($10%) were pruritus (10.8%), diarrhea (10.8%), and fatigue (9.8%). Grade 3/4 treatment-related AEs occurred in 13 (12.7%) pts. Immune-mediated AEs occurred in 19 (18.6%) pts. Conclusions: Pembro continued to show encouraging antitumor activity in pts with HR, BCG-unresponsive CIS with or without papillary tumors and a safety profile consistent with that of previous experience. Updated data using additional follow-up will be presented. Clinical trial information: NCT02625961.

4531 Poster Session (Board #357), Mon, 1:15 PM-4:15 PM

Impact of immune-related adverse events on survival in patients with metastastic urothelial carcinoma treated with immune-checkpoint inhibitors.

Rafael Morales-Barrera, Cristina Suarez Rodriguez, Macarena Gonzalez, Javier Ros, Maria Eugenia Semidey, Ester Serra Hernandez, Joaquin Mateo, Carlos Fernandez ´ S´aez, Fernando Lozano, Richard Mast, Sarai Roche, Angela Quintana, Sara Gutierrez ´ Fernandez, ´ Cesar Serrano, Claudia Valverde, Ines de Torres, Xavier Maldonado, Juan Morote, Joan Carles; Vall d’Hebron Institute, Barcelona, Spain; Vall d’Hebron Institute of Oncology, Universitat Autonoma ` de Barcelona, Barcelona, Spain; Vall d’Hebron Institute of Oncology, Vall d’ Hebron University Hospital, Universitat Autonoma ` de Barcelona, Barcelona, Spain; Vall D´Hebron University Hospital, Barcelona, Spain; Vall d´Hebron University Hospital, Barcelona, Spain; VHIO, barcelona, Spain; Vall d’Hebron Institute of Oncology, Barcelona, Spain; VHIO, Barcelona, Spain; Vall d’Hebron University Hospital, Barcelona, Spain; Pathology Department, Vall d’Hebron University Hospital, Barcelona, Spain; University Hospital Vall d’Hebron, Barcelona, Spain; Urology Department, Vall d’Hebron University Hospital, Barcelona, Spain

Background: Immune-checkpoints inhibitors (ICIs) represents the standard of care for platinum- pretreated advanced urothelial cancer patients (pts). By enhancing T-cell activation, a unique spectrum of inflammatory side effects has emerged, also known as immune-related adverse events (irAEs). Data regarding the association between irAEs and pts outcomes are conflicting. Here we conducted a retrospective analysis to investigate the association between irAEs profile and disease outcome in metastastic urothelial carcinoma (mUC) pts. Methods: Medical records from pts with mUC included in clinical trials between July 2013 and June 2018 and treated with ICIs were reviewed. Pts previously treated with platinum-based chemotherapy or cisplatin ineligible pts who had not been previously treated with chemotherapy were included. Clinical responses were assessed as complete response (CR), partial response (PR), stable disease (SD), and progressive disease (PD) according to RECIST v1.1. Adverse events were graded based CTCAE v4.03. Overall survival (OS) was calculated from the date of initiation of ICI to the date of death. X2 test was used to determine differences in rates. OS was estimated using Kaplan-Method and long rank test was used to assess differences between groups. All analyses were performed using SPSS v21. Results: From a total of 52 pts, 44 (84.6%) were treated with ICI monotherapy and 8 (15.3%) in combination (anti-CTLA4 or targeted therapy). Median age was 65 years, 42 pts (80.8%) were male, 44 patients (84.6%) had ECOG PS 0-1, 14 pts (26.9%) had liver metastasis. Overall irAEs were observed in 30 pts (57.7%) and 10 pts (19.2%) developed grade 3/4 irAES. Most common grade 3/4 irAEs were diarrhea (6.6%), rash (6.6%) and hepatitis (6.6%). Disease control rate (CR [26%]+PR[33%]+SD[20%]) was higher for patients with irAEs compared to those patients who did not developed irAEs (CR [13.6%]+PR[0%]+SD[22.7%], this difference was statically significant (P = 0.002). Median OS was 11.23 mo (CI 95%, 3.76-18.70) for the overall cohort, while median OS was 21.91 mo for those patients with irAEs compared to 6.47 mo in patients who did not developed irAEs (P = 0.004). Conclusions: In this analysis we found that the development of irAEs was a strong predictor of improved OS in mUC patients treated with ICI.

4532 Poster Session (Board #358), Mon, 1:15 PM-4:15 PM

Increasing use of neoadjuvant chemotherapy (NAC) in muscle-invasive bladder cancer (MIBC): Prognostic impact of non-standard of care (SOC) regimens.

Yaw A. Nyame, Sarah K Holt, Brian Winters, Sarah P. Psutka, Atreya Dash, George R Schade, Daniel W. Lin, Leonidas Nikolaos Diamantopoulos, Petros Grivas, Evan Y. Yu, John L. Gore, Jonathan L. Wright; Department of Urology, University of Washington Medical Center, Seattle, WA; University of Washington Medical Center, Seattle, WA; University of Washington, Seattle, WA; University of Washington, School of Medicine, Seattle, WA

Background: Cisplatin-based NAC can prolong overall survival (OS) in patients (pts) with MIBC. Utilization of NAC has increased to about 20% of pts with MIBC over the last decade. We evaluated NAC utilization with and without SOC cisplatin-based combination regimens and oncologic outcomes using registry data. Methods: This is a population-based analysis of linked SEER-Medicare data (2004-2011). We identified 4534 pts with MIBC (cT2-4N0-1) undergoing radical cystectomy (RC). Based on pharmacy records data, pts were stratified into 3 groups: SOC, non-SOC, and immediate cystectomy (IC). We used descriptive statistics to compare groups, and multivariate logistic regression to define factors associated with receiving SOC NAC. Competing risk bladder cancer-specific mortality (BCSM) incidence curves were generated and KM analysis was used to assess OS from time of RC. The impact of NAC on OS was evaluated with Cox regression analysis. Results: 694 (15.3%) pts received NAC, increasing from 11% in 2004 to 24.8% in 2011, with 345 (50%) receiving non-SOC, e.g. gemcitabine/carboplatin (49.3%), gemcitabine alone (21.2%), carboplatin alone (14.8%), cisplatin alone (8.4%), and methotrexate/vinblastine/ adriamycin/ carboplatin (0.8%). On logistic regression, increasing age (OR 0.91, 95%CI 0.88 – 0.94, p , 0.0001), Hispanic/Latin ethnicity (OR 0.49, 95%CI 0.22 – 1.10, p = 0.08), and $moderate renal dysfunction (OR 0.20, 95%CI 0.08 – 0.51, p , 0.001) were associated with lower odds of SOC NAC. Non-SOC NAC was associated with higher BCSM (competing risk) and lower OS (KM) vs. IC and SOC NAC. On multivariable analysis, non-SOC NAC was associated with higher risk of BCSM (HR 1.35, 95%CI 1.06 – 1.72, p = 0.01) and lower OS (HR 1.38, 95%CI 1.11 – 1.70, p = 0.003) vs. SOC NAC. Conclusions: About 50% of pts receiving NAC were not treated with SOC regimens. Non-SOC NAC was associated with higher bladder cancer death risk. This stresses the role of SOC NAC ideally in a multidisciplinary expert setting, as well as the need for timely RC and neoadjuvant clinical trials, including cisplatin-ineligible pts.

4533 Poster Session (Board #359), Mon, 1:15 PM-4:15 PM

Adenocarcinoma (ACB), urothelial carcinoma (UCB) and squamous cell carcinoma (SCCB) of the bladder: A Comprehensive Genomic Profiling (CGP) Study.

Joseph Jacob, Gennady Bratslavsky, Oleg Shapiro, Nick Liu, Elizabeth Kate Ferry, Alina Basnet, Julia Andrea Elvin, Jo-Anne Vergilio, Jonathan Keith Killian, Shakti Ramkissoon, Eric Allan Severson, Siraj Mahamed Ali, Jon Chung, Alexa Betzig Schrock, Prasanth Reddy, Kimberly McGregor, Brian Alexander, Vincent A. Miller, Andrea Necchi, Jeffrey S. Ross; SUNY Upstate Medical University, Syracuse, NY; Urologic Oncology Branch, National Cancer Institute at the National Institutes of Health, Bethesda, MD; SUNY Upstate University Hospital, Syracuse, NY; SUNY Upstate Medical University Hospital, Syracuse, NY; Foundation Medicine, Inc., Cambridge, MA; Foundation Medicine, Cambridge, MA; Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy

Background: We performed a CGP to compare the genomic alterations (GA) in ABC, UBC and SCCB. Methods: 143 cases of ACB, 2,142 cases of UCB and 83 cases of SCCB were subjected to CGP using a hybrid-capture based assay. Tumor mutational burden (TMB) was determined on 1.1 Mbp of sequenced DNA and microsatellite instability (MSI) was determined on 114 loci. PD-L1 expression was determined by IHC. Results: ABC patients were younger and more often female than UBC and SCCB (P , 0.0001). UBC and SCCB had a higher GA/tumor than ABC (P = 0.01). Un-targetable GA were similar in all 3 groups involving TP53 and KRAS. APC GA were more frequent in ABC whereas TERT, CDKN2A/B and DNA- repair genes (ARID1A and KDM6A) more frequently altered in UBC and SCCB. Targetable MTOR pathway GA (PIK3CA, TSC1, PTEN) were more frequent in UBC and SCCB as were targetable kinase alterations (FGFR3 and ERBB2). The UBC and SCCB had a significantly higher TMB than ABC (P , 0.0001) including mean TMB and TMB . 20 mut/Mb (P , 0.0001). CD274 (PD-L1) was amplified more frequently in SCCB than ACB or UBC (P , 0.0001). MSI high status was very uncommon in all tumor types. Conclusions: Deep sequencing reveals that ABC features a widely different genomic profile from UBC and SCCB. UBC has the highest frequencies of targetable kinase GA and high TMB. SCCB has the highest frequencies of IO efficacy predicting biomarkers including mean TMB and PD-L1 amp. Nonethe- less, ABC does feature potential kinase targets such as FGFR3 and ERBB2.

Top TMB TMB Age GA/ Un-targetable Top Targetable Mean ‡10 mut/ ‡20 mut/ PD-L1 IHC Tumor Type (M/F) MedianRange tumor GA GA MSI High TMB Mb Mb Positive

ACB 143 cases 57/86 58 (24-83) 5.4 TP53 79% PIK3CA 10% 2/106 (2%) 2.4 14 (10%) 4 (3%) 2/11 (18%) KRAS 30% ERBB2 8% SMAD4 14% PTEN 4% MYC 13% MET 4% CDKN2A EGFR 4% 13% TERT 12% CD274 0% ARID1A 10% APC 8% UCB 2,142 1597/ 67 (19-88) 7.7 TERT 68% PIK3CA 21% 11/1661 9.9 697 243 75/243 (31%) cases 545 TP53 56% FGFR3 19% (1%) (32%) (11%) CDKN2A ERBB2 16% 35% CDKN2B TSC1 9% 38% ARID1A 23% PTEN 5% KDM6A 23% CD274 1% KMT2D 22% KMT2D 22% APC 3% SCCB 45/38 62 (31-88) 8.2 CDKN2A PIK3CA 43% 1/69 10.4 26 13 7/18 70% 83 cases TP53 70% PTEN 8% (1%) (31%) (16%) (39%) TERT 62% FGFR3 7% CDKN2B ERBB2 6% 37% KMT2D 31% EGFR 6% APC 5% CD274 5%

4534 Poster Session (Board #360), Mon, 1:15 PM-4:15 PM

Gemcitabine plus carboplatin versus gemcitabine plus oxaliplatin in cisplatin unfit patients with advanced urothelial carcinoma: A randomized phase II study (COACH, KCSG GU10-16).

Jae-Lyun Lee, Bong-Seog Kim, Ho Yeong Lim, Hee Jun Kim, Inkeun Park, Yoon Ji Choi, Kyong Hwa Park, Kyung Hee Lee, Shinkyo Yoon, Bumsik Hong, Jun Hyuck Hong, Hanjong Ahn; Asan Medical Center and University of Ulsan College of Medicine, Seoul, South Korea; VHS Medical Center, Seoul, South Korea; Samsung Medical Center, Sungkyunkwan University, Seoul, South Korea; Chung- Ang University Hospital, Seoul, South Korea; Division of Hematology and Oncology, Department of Internal Medicine, Gachon University Gil Medical Center, Incheon, South Korea; Division of Oncology, Department of Internal Medicine, Korea University Anam Hospital, Korea University College of Medicine, Seoul, South Korea; Division of Oncology/Hematology, Department of Internal Medicine, Korea University Anam Hospital, Seoul, South Korea; Division of Oncology-Hematology, Department of Medicine, Yeungnam University College of Medicine, Daegu, South Korea; Department of Oncology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, South Korea; Department of Urology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, South Korea

Background: We investigated the activity and safety of first-line gemcitabine-oxaliplatin (GemOx) compared with gemcitabine-carboplatin (GCb) in cisplatin-ineligible patients with advanced UCCC. Methods: Treatment na¨ıve, cisplatin-ineligible patients with advanced UCCC were randomly assigned to GemOx [gemcitabine 1,000 mg/m2, oxliplatin 100 mg/m2 on day 1 (D1) every 2 weeks] or GCb (gemcitabine 1,000 mg/m2 on D1 and D8, carboplatin AUCC 4.5 on D1 every 3 weeks) stratified by ECOG performance status (PS) and visceral metastases. The primary endpoint was objective response rate (ORR), and secondary endpoints were progression-free survival (PFS) and overall survival (OS). Results: Between January 2011 and Mar 2017, 80 patients were enrolled; 39 patients and 40 patients were allocated to GCb and GemOx arm, respectively. Median age was 72 years (range 46-84) in GCb arm and 72.5 (55-85) in GemOx arm. ECOG PS was 2 in 41% (GCb arm) and 43% (GemOx arm), and median GFR was 45 ml/min (interquartile range 36-56) in GCb and 47 ml/min (37-56) in GemOx arm. ORR were 48.7% and 55.0% in GCb and GemOx, respectively. With a median follow-up duration of 37.8 months, median PFS and OS in GCb and GemOx arm were 5.5 months (95% CI, 4.8-6.2) vs. 4.4 months (95% CI, 2.7-6.1) and 9.1 months (95% CI, 5.2-13.0) vs. 11.0 months (95% CI, 6.9-15.0), respectively. Grade 3 leukopenia, neutropenia and fatigue were significantly more common in GCb arm (26% vs. 3%, p=0.003; 33% vs. 10%, p=0.014; 15% vs. 3%, p=0.057) while any grade neuropathy was more common in GemOx (8% vs. 60%, p,0.001). Conclusions: GemOx has showed comparable efficacy with GCb and favorable hematologic toxicity profile. GemOx may be used as a new option for UCCC patients who are not suitable for platinum-containing chemotherapy. Clinical trial information: NCT01487915.

4535 Poster Session (Board #361), Mon, 1:15 PM-4:15 PM

Squamous-cell carcinoma variant histology (SCC-VH) in muscle-invasive bladder cancer (MIBC): A comprehensive clinical, genomic, and therapeutic assessment from multiple datasets.

Marco Bandini, Filippo Pederzoli, Russell Madison, Alberto Briganti, Elizabeth R. Plimack, Jeffrey S. Ross, Guenter Niegisch, Evan Y. Yu, Aristotelis Bamias, Neeraj Agarwal, Srikala S. Sridhar, Jonathan E. Rosenberg, Joaquim Bellmunt, Matt D. Galsky, Andrea Gallina, Andrea Salonia, Francesco Montorsi, Siraj Mahamed Ali, Jon Chung, Andrea Necchi; Vita-Salute San Raffaele University, Milan, Italy; Universit`a Vita-Salute San Raffaele, Milan, Italy; Foundation Medicine, Inc., Cambridge, MA; Fox Chase Cancer Center, Philadelphia, PA; SUNY Upstate Medical University, Syracuse, NY; Heinrich-Heine-University, Medical Faculty, Department of Urology, Duesseldorf, Germany; University of Washington, Seattle, WA; Haematology- Oncology Unit, Department of Clinical Therapeutics, Alexandra Hospital, National and Kapodistrian University of Athens, Athens, Greece; University of Utah Huntsman Cancer Institute, Salt Lake City, UT; Princess Margaret Hospital, Toronto, ON, Canada; Memorial Sloan Kettering Cancer Center, New York, NY; IMIM-Hospital del Mar Medical Research Institute, Barcelona, Spain; Department of Medicine, Icahn School of Medicine at Mount Sinai, Tisch Cancer Institute, New York, NY; Vita Salute San Raffaele University and Urological Research Institute (URI), IRCCS San Raffaele Hospital, Milano, Italy; Vita-Salute San Raffaele University, Urological Research Institute, IRCCS San Raffaele Hospital, Milano, Italy; Universita Vita Salute San Raffaele, Milan, Italy; Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy

Background: Pure or predominant SCC-VH is not uncommon in MIBC. Nevertheless, very few data are available about the efficacy of neoadjuvant chemotherapy (NAC). Here, we examined the outcomes after NAC, explored novel therapeutic targets, and propose new results in these patients (pts) by integrating multiple datasets. Methods: Within RISC and San Raffaele databases (1990-2018), we identified 2858 MIBC pts with urothelial cancer (UC, N = 2229) or VH (N = 629) who received RC +/- NAC. Kaplan-Meier and Cox regression analyses compared cancer-specific survival (CSS) between SCC and UC with NAC stratification. Logistic regression models tested the odds of clinical-to-pathological downstaging (cT . pT). Foundation Medicine (FMI) dataset was queried for SCC-VH. 97 pts were assayed with hybrid-capture based comprehensive genomic profiling (CGP). Finally, we looked at the results from the PURE-01 study, that is now amended and enrolling pts with VH (NCT02736266). Results: Overall, 127 (4.4%) had predominant SCC-VH, 157 (5.5%) UC+SCC. Among the NAC-treated pts, SCC was the only VH (N = 44) significantly associated with worse CSS, (p , 0.001) and higher mortality (HR 2.10, p = 0.003) vs. UC. After NAC adjustment, SCC-VH showed lower rate of downstaging (3.7 vs 9.3%, OR 0.4, p = 0.028) vs. UC. Similar negative trends were confirmed in pN0 pts, where SCC exhibited worse CSS (p = 0.006) and higher mortality (HR 5.15, p = 0.002). In the FMI cohort, the median tumor mutational burden (TMB) of SCC was 6.25 mut/mb (vs 6.9 mut/mb of 1984 UC), 27% of pts having . 10 mut/mb and 14% . 20 mut/mb. Clinically relevant alterations occurred in PIK3CA (42%), CCND1 (15%), PTEN (9.3%), FGFR3 (9.3%), and ERBB2 (6.2%). In the PURE-01 study, 13/84 (15.5%) SCC-VH pts received pembrolizumab before RC. PD-L1 combined positive score was $10 in 11/13 pts; results yielded 4 pT0 (30.8%), 10 pT#1 (76.9%), and no deaths (median FUP: 10.4 mo). Conclusions: We present a comprehensive assessment of SCC-VH in MIBC. SCC represents the VH with the lowest activity of NAC. While CGP revealed multiple opportunities for targeted therapy, the efficacy of neoadjuvant pembrolizumab in SCC is encouraging.

4536 Poster Session (Board #362), Mon, 1:15 PM-4:15 PM

Defects in DNA repair genes and long-term survival in cisplatin-based neoadjuvant chemotherapy for muscle invasive bladder cancer (MIBC).

Benjamin Miron, Eric A. Ross, Fern Anari, John O’Neill, Jean H. Hoffman-Censits, Matthew R. Zibelman, Alexander Kutikov, Rosalia Viterbo, Richard E. Greenberg, David Chen, Costas D. Lallas, Edouard John Trabulsi, R. Katherine Alpaugh, Essel Dulaimi, Erica Golemis, Robert Uzzo, Elizabeth R. Plimack; Fox Chase Cancer Center, Philadelphia, PA; Sidney Kimmel Comprehensive Cancer CenteratJohnsHopkins,Baltimore, MD; Department of Urology, Sidney Kimmel Cancer Center, Thomas Jefferson University, Philadelphia, PA

Background: Although cisplatin-based neoadjuvant chemotherapy (NAC) has demonstrated an overall survival (OS) benefit in MIBC, only a subset of patients have pathologic complete response (pT0) at cystectomy. ATM, RB1 and FANCC mutations have shown correlation with pT0 to cisplatin-based NAC, as previously published. We now report updated OS and disease specific survival (DSS) from two phase II trials using these gene alterations as biomarkers. Methods: Patients with stage T2-T4 (N0 or N1) MIBC were enrolled in phase II trials of dose-dense NAC with MVAC (methotrexate, vinblastine, adriamycin, and cisplatin; NCT01031420) or GC (gemcitabine and cisplatin; NCT01611662). Patients were treated with NAC with plan for curative cystectomy. DNA from pretreatment tumor tissue was sequenced for coding exons of 287 cancer-related genes and analyzed for mutations. Survival in patients with one or more mutations in ATM, RB1,orFANCC genes was compared to those without mutations. Results: Of 58 pts treated, 38% (22/58 pts) had relevant mutations in the combined group of MVAC (13/34 pts) and GC (9/24 pts) trials. At a median follow-up of 56 months and minimum follow up of 16 months, patients with mutations had statistically significantly greater OS (p = 0.0043) and DSS (p = 0.0015). Median OS/DSS was not reached for patients with a mutation in any group. At 5 years post treatment, OS/DSS were greater in mutated vs non-mutated patients in all groups (see table). Conclusions: Long-term follow up reveals that previously reported improved responses to cisplatin- based NAC associated with mutations in ATM, RB1 and FANCC also confer a clinically meaningful and statistically significant survival benefit in these patients. These alterations may be useful as predictive biomarkers to allow clinicians to prioritize patients most likely to benefit from NAC prior to radical cystectomy.

5-yr DSS 5-yr OS (mutant) 5-yr OS (non-mut) 5-yr DSS (mutant) (non-mut) Combined 85% (60.4%, 46% (29.5%, 90% (64.8%, 49% (31.6%, 94.9%) 61.7%) 97.3%) 64.9%) MVAC 85% (51.2%, 52% (28.7%, 85% (67.1%, 52% (28.7%, 95.9%) 70.4%) 100%) 70.4%) GC 86% (33.4%, 40% (16.5%, 100% (100%, 47% (19.5%, 97.9%) 62.8%) 100%) 70.1%)

4537 Poster Session (Board #363), Mon, 1:15 PM-4:15 PM

Bladder Cancer Multidisciplinary Clinic (BCMC) model: Impact on imaging, pathology and treatment recommendations.

Brian Winters, Leonidas Nikolaos Diamantopoulos, Maria S. Tretiakova, Lawrence D. True, Jean H Lee, Jing Zeng, Jay Justin Liao, Heather H. Cheng, Michael Thomas Schweizer, Andrew Caleb Hsieh, Evan Y. Yu, George R Schade, John L. Gore, Daniel W. Lin, Manjiri Dighe, Funda Vakar-Lopez, Kenneth J. Russell, Petros Grivas, Robert B. Montgomery, Jonathan L. Wright; Department of Urology, University of Washington Medical Center, Seattle, WA; University of Washington, Seattle, WA; University of Washington Seattle Cancer Care Alliance, Seattle, WA; University of Washington/Fred Hutchinson Cancer Research Center, Seattle, WA; Fred Hutchinson Cancer Research Center, Seattle, WA; University of Washington Medical Center, Seattle, WA; University of Washington Medicine, Seattle, WA; Seattle Cancer Care Alliance, Seattle, WA; University of Washington, School of Medicine, Seattle, WA

Background: Despite guideline-based standard of care recommendations in BC and upper tract urothelial carcinoma (UTUC), treatment remains variable across US. Experts recommend focusing BC care in tertiary centers. We hypothesized that a BCMC model, with expert central pathology and radiology review, may result in changes in corresponding reports, and, thus, treatment recommendations. Methods: Our BCMC clinic format includes simultaneous consultation with Urologic, Medical and Radiation Oncology, with real time expert genitourinary pathology and radiology review. We retrospectively assessed the concordance between outside (pre-BCMC) imaging & pathology review and BCMC review. Differences between pre- and post- BCMC recommendations on management were also assessed; descriptive statistics were used. Results: We identified 233 BC/UTUC patients (pts) referred to BCMC. Complete radiographic and pathologic data were available for 209 pts. Median age at time of evaluation was 68 (27-93) and 85% were PS ECOG 0-1. After BCMC review of outside records, 112 (53.6%) imaging and/or pathology changes were noted, with 57 (27%) pts upstaged. Overall, imaging interpretation was changed in 25% of cases, and 20% of pts were upstaged. BCMC pathology review resulted in changes in 59 (28%) pts. Among those, 42 (71%) had histologic subtype addition or change, 9 (15%) had LVI/CIS status change, and 2 (3.4%) had low to high grade conversion. In terms of pathology staging, 7 (12%) were downstaged, and 5 (8.5%) upstaged. Further diagnostic work-up was recommended in 71/209 (34%) pts, resulting in upstaging in 11/71 (15.5%) of cases. Pre- and post- BCMC-recommended treatment modality differed in 55/209 (26%) pts, while a new treatment modality was added in 28/209 (13%) pts. These recommendations were followed 91.4% of the time (191/209 pts). Conclusions: BCMC initiation at our institution resulted in imaging and/or pathology diagnostic changes in almost half of cases, with approximately a quarter of pts being upstaged. Findings reveal the importance of expert radiology and pathology review in BC. Further study is needed to confirm the proposed benefits and impact of BCMC on treatment response and outcomes.

4538 Poster Session (Board #364), Mon, 1:15 PM-4:15 PM

Proximity to oil refineries and risk of bladder cancer: A population-based analysis.

Tamer Dafashy, Daniel Phillips, Mohamed Danny Ray-Zack, Preston Kerr, Yong Shan, Jacques G. Baillargeon, Yong-Fang Kuo, Hemalkumar B Mehta, Stephen Bentley Williams; The University of Texas Medical Branch at Galveston, Galveston, TX; The University of Texas Medical Branch, Galveston, TX; University of Texas Medical Branch at Galveston, Galveston, TX; University of Texas Medical Branch, Galveston, TX

Background: Exposure to aromatic amines is a risk factor for bladder cancer. Incidence rates according to proximity to oil refineries are largely unknown. We sought to determine proximity of oil refineries and bladder cancer incidence in the State of Texas which is home to the largest number of oil refineries in the United States. Methods: We used the Texas Cancer Registry database to identify patients diagnosed with bladder cancer from January 1, 2001 to December 31, 2014. The U.S. census data from 2010 was used to ascertain overall population size, age and sex distributions. Heat maps of the 28 active oil refineries in Texas were developed. Incidence of bladder cancer were compared according to proximity ( , 10 vs. $ 10 miles) to an oil refinery. Risk ratios were adjusted using a Poisson regression model. Results: A total of 45,517 incident bladder cancer cases were identified of which 5,501 cases were within 10 miles of an oil refinery. In adjusted analyses, bladder cancer risk was significantly greater among males vs. females (Relative Risk (RR) 3.41, 95% Confidence Interval (CI), 3.33-3.50), and greater among people living within 10 miles from an oil refinery than those living outside a 10-mile radius from an oil refinery (RR 1.19, 95% CI, 1.08-1.31). Conclusions: People living within 10 miles from oil refineries were at greater risk for bladder cancer. Further research into exposure to oil refineries and bladder cancer incidence is warranted.

4540 Poster Session (Board #366), Mon, 1:15 PM-4:15 PM

Towards the noninvasive identification of pathologic responders to neoadjuvant pembrolizumab in muscle-invasive urothelial bladder cancer (MIBC).

Giuseppina Calareso, Marco Bandini, Daniele Raggi, Elena Fare, ` Andrea Gallina, Renzo Colombo, Maurizio Colecchia, Roberta Luciano Luciano, Andrea Salonia, Filippo Pederzoli, Jeffrey S. Ross, Russell Madison, Siraj Mahamed Ali, Jon Chung, Francesco Montorsi, Alberto Briganti, Antonella Messina, Andrea Necchi; Radiology Unit, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy; Vita-Salute San Raffaele University, Milan, Italy; Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy; Fondazione IRCCS Istituto Nazionale Tumori, Milan, Italy; Vita Salute San Raffaele University and Urological Research Institute (URI), IRCCS San Raffaele Hospital, Milano, Italy; Division of Oncology/ Unit of Urology, Urological Research Institute, IRCCS Ospedale San Raffaele, Milan, Italy; Department of Pathology and Laboratory Medicine, Fondazione IRCCS Istituto Nazionale Tumori di Milano, Milan, Italy; IRCCS San Raffaele Hospital, Milan, Italy; Vita-Salute San Raffaele University, Urological Research Institute, IRCCS San Raffaele Hospital, Milano, Italy; Universita ` Vita-Salute San Raffaele, Milan, Italy; SUNY Upstate Medical University, Syracuse, NY; Foundation Medicine, Inc., Cambridge, MA; Universita Vita Salute San Raffaele, Milan, Italy

Background: The possibility to predict the pathologic complete response (pT0) or pT#1 after neoadjuvant immunotherapy may deeply impact the management of MIBC and orient clinical trials. The PURE01 study (Necchi, JCO 2018) evaluated preoperative pembrolizumab before radical cystectomy (RC) in T2- 3bN0MO patients (pts). The trial was amended to increase the sample size. Methods: Pts were assessed with bladder multiparametric magnetic resonance imaging (mpMRI: T2-weighted imaging, diffusion- weighted imaging, dynamic contrast enhancement [CE]) before and after treatment (3 cycles every 3 weeks) prior to RC. All mpMRI were made with bladder catheterization. We tested the following complete response (CR) definition: any measurable residual disease (T2W, morphology), but no residual restricted diffusion and no early CE signals. Exams were independently assessed by 2 senior radiologists and tested against the pT0/pT#1 endpoints. Logistic regression models analyzed mpMRI and biomarker data on comprehensive genomic profiling (CGP) with FoundationONE CDx assay and PD-L1 combined positive score (CPS). Results: From 02/17 to 11/18, 84 pts were assessed before and after treatment (168 total mpMRI) and had the pathological response available from RC. Baseline CR (TURB effect) was observed in 14 (16.7%) pts. The mean overall percentage of agreement (OPA) across the mpMRI data was 98.2%. The sensitivity of post-therapy (PT) CR for pT0 response was 0.80 (95%CI: 0.61-0.92); the specificity for pT#1 response was 0.94 (95%CI: 0.80-0.99). PT-CR was associated with both pT0/pT#1 endpoints (p , 0.001) after adjusting for clinical T-stage and histology. Baseline CPS$10 and TMB$20 mut/Mb were strongly associated with pT0 response (OR: 4.46 and 14.9). mpMRI CR post- pembrolizumab significantly improved the c-index of the CPS/TMB models: up to 0.81 and 0.87 for pT0/pT#1endpoints.Conclusions: In post-pembrolizumab MIBC, the proposed definition of radiological CR through mpMRI is endowed with limited inter-observer variability and high accuracy in predicting pT#1. CPS/TMB-selected pts who achieve CR with mpMRI can be reliably spared any surgical treatment. Clinical trial information: NCT02736266.

4541 Poster Session (Board #367), Mon, 1:15 PM-4:15 PM

Treatment sequencing of anti-PD-1/PD-L1 and carboplatin (carbo)-based chemotherapy (chemo) in cisplatin-ineligible patients (pts) with metastatic urothelial cancer (mUC).

Xiao X. Wei, Lillian Werner, Min Yuen Teo, Jonathan E. Rosenberg, Vadim S Koshkin, Petros Grivas, Bernadett Szabados, Laura Morrison, Lucia Carril, Daniel E. Castellano, Pedro Isaacsson Velho, Noah M. Hahn, Rana R. McKay, Daniele Raggi, Andrea Necchi, Ravindran Kanesvaran, Parissa Alerasool, Jacob Gaines, Joaquim Bellmunt, Guru Sonpavde; Dana-Farber Cancer Institute, Boston, MA; Memorial Sloan Kettering Cancer Center, New York, NY; University of California San Francisco, San Francisco, CA; University of Washington, School of Medicine, Seattle, WA; Barts Cancer Centre, Queen Mary University of London, London, United Kingdom; Barts Cancer Institute, Queen Mary University of London, London, London, United Kingdom; Hospital 12 de Octubre, Madrid, Spain; Medical Oncology Service, Hospital Universitario 12 de Octubre, Madrid, Spain; Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins, Baltimore, MD; Johns Hopkins University School of Medicine, Baltimore, MD; Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy; National Cancer Centre Singapore, Singapore, Singapore; Tisch Cancer Institute at Mount Sinai, New York, NY; IMIM-Hospital del Mar Medical Research Institute, Barcelona, Spain

Background: Anti-PD-1/PD-L1 agents and carbo-based chemo are therapy options in 1st-line (1L) setting for cisplatin-ineligible pts with mUC. However, optimal sequencing is unclear. Methods: We conducted a multicenter retrospective analysis of cisplatin-ineligible pts with mUC treated with 1L PD-1/PD-L1 monotherapy followed by carbo-based chemo (IO→Cb) or the reverse order (Cb→IO) without intervening systemic therapy. Perioperative cisplatin-based chemo was allowed if completed . 1 year from 1L mUC therapy initiation. To assess association between overall survival (OS) and therapy sequence, a multivariate analysis (MVA) was performed from initiation of 2L therapy, adjusted for treatment sequence, time interval between initiation of 1L and 2L therapies, Hb ( , 10 vs $10 g/dl), ECOG PS (0-1 vs 2-3), and metastatic site (LN/soft tissue only vs non-liver vs liver). Results: 146 pts (IO→Cb n = 43, Cb→IO n = 103) were evaluable with median age 72, 76% men, 78% ECOG PS 0-1, 17.8% with liver metastasis. Baseline factors were balanced except for higher proportion of men in IO→Cb group (91% vs 70%, p = 0.01). Median time interval between initiation of 1L and 2L therapy for IO→Cb and Cb→IO were 15.6mo (4.8-78.1) and 23.0mo (2.1-103.3), respectively. Response rates are summarized (Table). On MVA, treatment sequence was not associated with OS (HR 1.05, p = 0.85). Site of metastasis was the only factor significantly associated with OS (p = 0.002). Conclusions: In our retrospective analysis of cisplatin- ineligible pts with mUC regardless of PD-L1 expression, anti-PD-1/PD-L1 followed by carbo-based chemo or the reverse sequence appeared to confer comparable OS. The observed response rates and time interval between initiation of 1L and 2L therapy are likely contributed by pt selection, where all pts received 2L. Further investigation of the ‘PD-L1 high’ population is warranted, given higher response rates with anti- PD-1/PD-L1 vs ‘PD-L1 low’ population. Ongoing phase III trials will help inform optimal sequencing.

IO→Cb (N = 43) Cb→IO (N = 103) 1L 2L 1L 2L Best overall response PD-1/L1 Carbo Carbo PD-1/L1 Unknown 0 (0%) 8 (18.6%) 3 (2.9%) 15 (14.6%) PD 28 (65.1%) 8 (18.6%) 32 (31.1%) 53 (51.5%) SD 11 (25.6%) 8 (18.6%) 21 (20.4%) 13 (12.6%) PR 4 (9.3%) 19 (44.2%) 43 (41.7%) 19 (18.4%) CR 0 (0%) 0 (0%) 4 (3.9%) 3 (2.9%)

4542 Poster Session (Board #368), Mon, 1:15 PM-4:15 PM

FGFR-altered, advanced urothelial carcinoma (UC) and response to chemotherapy prior to receiving erdafitinib.

Andrea Necchi, Arlene O. Siefker-Radtke, Bob Zhong, Parthiv Jasvant Mahadevia, Ademi E. Santiago- Walker, Peter De Porre, Yohann Loriot; Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy; The University of Texas MD Anderson Cancer Center, Houston, TX; Janssen Research & Development, Spring House, PA; Janssen Research & Development, Raritan, NJ; Janssen Research & Development, Beerse, Belgium; Institut de Cancerologie ´ Gustave Roussy, Villejuif, France

Background: FGFR-altered, advanced UC has predominantly a luminal 1 subtype, which is associated with lower response rates to immunotherapy and possibly also to chemotherapy. Objective response rates (ORR) for first-line cisplatin-based regimens, such as gemcitabine-cisplatin (gem/cis) and methotrexate- vinblastine-doxorubicin-cisplatin (MVAC), historically range between 45-60% and for gemcitabine- carboplatin (gem/carbo) 35-45%. However, the ORR on chemotherapy for the ~20% of patients with FGFR-altered tumors is unknown. Methods: BLC2001 (NCT02365597) is an ongoing global open-label phase 2 study of the pan-FGFR inhibitor erdafitinib in patients with locally advanced or metastatic UC with specific FGFR2/3 gene alterations. Patients who had received first-line (1L) or second-line (2L) chemotherapy for advanced UC were identified. Investigator-reported ORR (complete + partial responses) and median time to progression (TTP) on these pretreatments were analyzed. Results: Of 210 patients treated with erdafitinib in BLC2001, 191 had received prior systemic therapy including 184 and 83 patients who had received 1L and 2L chemotherapy, respectively. ORR were 29.3% (54/184; 95% CI 22.8%, 35.9%) to 1L chemotherapy and 24.1% (20/83; 95% CI 14.9%, 33.3%) to 2L chemotherapy. 1L therapy consisted of gem/cis in 94 patients, gem/carbo in 59 patients, and MVAC in 22 patients, with ORR (95% CI) of 35.1% (25.5%, 44.8%), 25.4% (14.3%, 36.5%), and 22.7% (5.2%, 40.2%), respectively. In the 2L setting, of 46 patients who had received a regimen containing a taxane (paclitaxel or docetaxel) or vinflunine, 8 patients (17.4%; 95% CI 6.4%, 28.3%) achieved an objective response. Median TTP was 7.16 mo (95% CI 6.18, 7.49) after 1L chemotherapy (7.6 mo for gem/cis, 6.3 mo for gem/carbo, and 5.3 mo for MVAC) and 4.35 mo (95% CI 3.3, 5.5) after 2L chemotherapy (3.6 mo for taxane or vinflunine). Conclusions: In this post-hoc analysis, the overall ORR to prior 1L chemotherapy was lower, but within the range expected based on historical data. Further investigation into the response to chemotherapy in FGFR alteration positive patients is warranted and may be useful for the development of 1L trials of combination therapy. Clinical trial information: NCT02365597.

4543 Poster Session (Board #369), Mon, 1:15 PM-4:15 PM

Erdafitinib in high-risk patients (pts) with advanced urothelial carcinoma (UC).

Se Hoon Park, Yohann Loriot, Bob Zhong, Syed A. Hussain, Parthiv Jasvant Mahadevia, Peter De Porre, Arlene O. Siefker-Radtke; Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, South Korea; Institut de Cancerologie ´ Gustave Roussy, Villejuif, France; Janssen Research & Development, Spring House, PA; Plymouth University, Plymouth, United Kingdom; Janssen Research & Development, Raritan, NJ; Janssen Research & Development, Beerse, Belgium; The University of Texas MD Anderson Cancer Center, Houston, TX

Background: The pan-FGFR inhibitor erdafitinib exhibited a robust objective response rate (ORR) and tolerability among pts with FGFR2/3-altered advanced UC in the BLC2001 (NCT02365597) phase 2 study (Siefker-Radtke ASCO 2018 #4503). Here we report a post hoc subgroup analysis to explore efficacy among high-risk pts. Methods: The analysis included 99 BLC2001 pts who received the optimized dose regimen of 8 mg/d continuous (pharmacodynamically guided uptitrated to 9 mg/d per serum phosphate). Results for ORR, duration of response (DOR), progression-free survival (PFS), and overall survival (OS) were analyzed by select baseline variables, with high-risk defined as age .75 y, ECOG PS 2, hemoglobin ,10 g/ dL, visceral metastases, and 2 or 3 Bellmunt risk factors. Results: Efficacy results (Table) show investigator- assessed ORR .36% and median PFS .5 mo across all subgroups except ECOG PS 2. OS data are immature but generally follow the trend of PFS. With the exception of ECOG 2, there were no differences in G3/4 serious AE proportions by subgroup. Conclusions: The post hoc subgroup findings support that erdafitinib generally provides comparable efficacy in high-risk pts with FGFR-altered advanced UC as the overall population. Clinical trial information: NCT02365597.

ORR, %, n/N DOR, mo PFS, mo OS, mo (95% CI) (95% CI) (95% CI) (95% CI) Primary analysis of all patients 40.4, 40/99 5.59 8.08 13.80 (30.7, 50.7) (4.24, 7.23) (5.65, 8.77) (9.82, NE) Age 38.6, 32/83 5.59 7.39 12.02 <75 y (28.1, 49.9) (4.21, 7.00) (5.55, 8.51) (8.97, NE) ‡75 y 50.0, 8/16 13.37 14.75 14.03 (24.7, 75.3) (2.56, 13.37) (3.98, 14.75) (3.98, NE) ECOG 42.4, 39/92 5.98 8.08 14.03 0-1 (32.1, 53.1) (4.24, 7.39) (5.65, 11.07) (10.71, NE) 2 14.3, 1/7 2.79 4.17 5.13 (0.4, 57.9) (NE, NE) (NE, NE) (2.99, 8.02) Hb 53.3, 8/15 5.98 8.31 NE <10 g/dL (26.6, 78.7) (2.96, NE) (4.34, NE) (5.98, NE) ‡10 g/dL 38.1, 32/84 5.55 7.39 13.80 (27.7, 49.3) (4.21, 7.39) (5.62, 8.77) (9.82, NE) Visceral metastases 38.5, 30/78 5.98 8.25 13.80 Y (27.7, 50.2) (4.21, 13.37) (5.62, 14.75) (8.02, NE) N 47.6, 10/21 4.57 5.95 14.03 (25.7, 70.2) (2.96, NE) (4.34, NE) (10.71, NE) Bellmunt risk factor 41.6, 32/77 5.55 7.39 14.03 0or1 (30.4, 53.4) (4.21, 7.39) (5.62, 8.77) (10.71, NE) 2or3 36.4, 8/22 5.98 8.31 10.32 (17.2, 59.3) (2.79, NE) (4.17, NE) (5.13, NE)

4544 Poster Session (Board #370), Mon, 1:15 PM-4:15 PM

Prognostic value of sequential 18F-FDG + Na18F PET/CT (NaF+FDG PET) in metastatic genitourinary (GU) cancer patients (pts) treated with cabozantinib/nivolumab +/- ipilimumab (CaboNivoIpi).

Nicholas Peter Verdini, Ilhan Lim, Esther Mena, Liza Lindenberg, Amir Mortazavi, Sumanta K. Pal, Primo Lara, Katherine Graap, Marissa Mallek, Corrine Keen, Ryan Thompson, Christian Mayfield, Corinne Turrell, Jacqueline Cadena, Biren Saraiya, Ray L Lilly, Michael V. Knopp, Chadwick Wright, Seth M. Steinberg, Andrea B. Apolo; National Cancer Institute at the National Institutes of Health, Bethesda, MD; Korea Cancer Center Hospital, Seoul, South Korea; Molecular Imaging Program, Center for Cancer Research, National Cancer Institute at the National Institutes of Health, Bethesda, MD; National Cancer Institute, Bethesda, MD; Arthur G. James Cancer Hospital, Ohio State University Wexner Medical Center, Columbus, OH; City of Hope National Medical Center, Duarte, CA; University of California Davis Comprehensive Cancer Center, Sacramento, CA; Thoracic and Gastrointestinal Oncology Branch, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Bethesda, MD; Genitourinary Malignancies Brach, National Cancer Institute at the National Institutes of Health, Bethesda, MD; National Cancer Institute, National Institutes of Health, Bethesda, MD; Office of Clinical Research, UC Davis Comprehensive Cancer Center, Sacramento, CA; Genitourinary Malignancies Branch, National Cancer Institute, Bethesda, MD; The Cancer Institute of New Jersey, Lawernceville, NJ; The James Comprehensive Cancer Center, Columbus, OH; Department of Radiology, The Ohio State University Wexner Medical Center, Columbus, OH; Division of Molecular Imaging and Nuclear Medicine, Department of Radiology, The Ohio State University, Columbus, OH; Biostatistics and Data Management Section, NCI, NIH, Bethesda, MD; Genitourinary Malignancies Branch, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Bethesda, MD

Background: NaF+FDG PET imaging are used to assess soft tissue and bone metastases. The prognostic value of NaF+FDG PET in GU cancer pts was assessed as a secondary endpoint within a phase I trial of combination CaboNivoIpi. Methods: NaF+FDG PET scans were collected at baseline and cycle (C)3 day (D)1 for 50pts. Up to 50 lesions/patient were analyzed at baseline, up to 10 lesions/organ in case of extensive disease. Lesion number and whole-body metabolic tumor volume (wMTV) were recorded at baseline, C3D1 and percent change for FDG and NaF scans. Whole-body total lesion glycolysis (wTLG) and its percent change was obtained for FDG scans. Parameters were evaluated with respect to OS using Kaplan-Meier and log-rank, with quartiles, then refined to show strongest distinction, adjusting p-values to account for this exploration. Parameters with strongest OS association were also analyzed for associations with OS in urothelial carcinoma (UC) pts. Results: 50 pts, (UC (n = 20); others (renal cell, prostate, urachal/adenocarcinoma, germ cell, penile, bladder squamous cell (n = 2-7 each)). Median (m) overall survival (OS) was 23.9 months (mo) (95% CI: 13.7mo – NE) with 29.7mo m potential follow up and mOS of 24.7mo (95% CI: 13.7mo-NE) for UC. For FDG in all pts, wMTV: baseline # vs . 51.6, mOS (NR vs 10mo, p = .0006), C3D1 # vs . 85 mOS (25.9mo vs 5.1mo, p = .0001), percent change (0/increase vs decrease, mOS 14mo vs 25.9mo, p = .015); wTLG: baseline # vs . 178, mOS (NR vs 11.5mo, p = .011), C3D1 # vs . 300, mOS (25.9mo vs 8.3mo, p , .0001), percent change decrease vs increase, mOS (25.9mo vs 14.0mo, p = .016); and lesion number: baseline # vs . 13, mOS (25.9mo vs 9.9mo, p = .0090), C3D1 # vs . 13 mOS (25.9mo vs 9.9mo, p , .0001) significantly predicted OS. In UC pts, wMTV percent change (0/increase vs decrease, mOS 14mo vs. 25.9mo, p = .057), wTLG percent change decrease vs increase, mOS (NR vs 8.4mo, p = .0015), and lesion number C3D1 # vs . 13 mOS (25.9mo vs 2.8mo, p = .022) significantly predicted OS. NaF parameters failed to do so. Conclusions: FDG wMTV and wTLG at baseline, C3D1, and percent change and lesion number at baseline and C3D1, predicted OS in GU cancer pts on CaboNivoIpi. Clinical trial information: NCT02496208.

4545 Poster Session (Board #371), Mon, 1:15 PM-4:15 PM

Analysis of EGFR mutant urothelial carcinoma (UC) reveals distinct mutational landscape.

Russell Madison, Sumati Gupta, Jeffrey S. Ross, Sumanta K. Pal, Alexa Betzig Schrock, Vincent A. Miller, John Heymach, Andrea Necchi, Luke Juckett, Siraj Mahamed Ali, Jon Chung, Venkataprasanth P. Reddy; Foundation Medicine, Inc., Cambridge, MA; Huntsman Cancer Institute- University of Utah Health Care, Salt Lake City, UT; SUNY Upstate Medical University, Syracuse, NY; City of Hope National Medical Center, Duarte, CA; The University of Texas MD Anderson Cancer Center, Houston, TX; Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy; Foundation Medicine, Inc, Cambridge, MA; Shawnee Mission Cancer Center, Westwood, KS

Background: Genomic alterations (GA) of EGFR, are well recognized as druggable oncogenic drivers in NSCLC, but the druggable GA EGFR L858R and exon 19 deletion (ex19del), are rarely observed in genitourinary cancer. We reviewed the genomic landscape of advanced upper tract and bladder UC (UTUC and BUC) to assess the frequencies of druggable EGFR GA. Methods: Tissue specimens from patients with UTUC (407) and BUC (2,402) were assayed using hybrid capture-based comprehensive genomic profiling (CGP). Tumor mutational burden (TMB) was determined on up to 1.1 Mbp of sequenced DNA and microsatellite instability (MSI) was determined on 95 or 114 loci. Results: EGFR alterations (EGFRalt) were present in 17 UTUC and 93 BUC (4.2% and 3.9%). Age distribution between the two subgroups was similar, but UTUC was more prevalent in female patients (47% v 29%). BUC had a higher median TMB (5.2mut/mb v 7.8 mut/mb; p = 0.046) and the prevalence of MSI-H cases was not significantly different. TERT (55% v 71%) and TP53 (59% v 74%) were the most frequently mutated genes in EGFRalt UTUC and BUC. Within EGFRalt, amplifications were the most common alterations in both UTUC and BUC (13/17, 76%; 57/93, 61%). Amplifications were mutually exclusive from cases with EGFR short variants (SV) in BUC (0/34, 0%), and co-occurred with EGFR SV in four UTUC cases (4/8 50%). The majority of EGFR SV were EGFR exon 20 insertions (EGFRexon20), which made up a larger proportion of EGFRalt in UTUC than BUC (7/17, 41% v 13/93, 14%; p = 0.01). Compared to other EGFRalt, EGFRexon20 trended towards mutual exclusivity of GA in commonly altered UC genes: TP53 (UTUC EGFRexon20 v EGFRalt other: 0% v 100%, p = 5.1E-5; BUC EGFRexon20 v EGFRalt other: 0% v 86%, p = 2.2E-7), PIK3CA (14% v 10%; 0% v 19%), RAF1 (0% v 10%; 0% v 16%), or FGFR3 (0% v 10%; 0% v 6.3%) alterations. Only 2.2% (2/93) of BUC EGFRalt were L858R mutations and none were ex19del (0/93), while neither mutation was detected in UTUC. Conclusions: EGFRexon20 defines a subset of UC cases, and is the most common non-amplification GA seen in EGFR in UC, with some enrichment in UTUC. Consideration should be given to developing a trial for EGFR exon20 UC patients given the recent investigational work on inhibitors with activity against EGFRexon20, such as poziotinib and TAK-788.

4546 Poster Session (Board #372), Mon, 1:15 PM-4:15 PM

KEYNOTE-052: Phase 2 study evaluating first-line pembrolizumab (pembro) in cisplatin- ineligible advanced urothelial cancer (UC)— Updated response and survival results.

Peter H. O’Donnell, Arjun Vasant Balar, Jacqueline Vuky, Daniel E. Castellano, Joaquim Bellmunt, Thomas Powles, Dean F. Bajorin, Petros Grivas, Noah M. Hahn, Elizabeth R. Plimack, Mary J. Savage, Xiao Fang, James Luke Godwin, Tara L. Frenkl, Ronald De Wit; The University of Chicago, Chicago, IL; Perlmutter Cancer Center, NYU Langone Health, New York, NY; Oregon Health & Science University, Portland, OR; Hospital Universitario 12 de Octubre, Madrid, Spain; Dana-Farber Cancer Institute, Boston, MA; Barts Cancer Institute, Queen Mary University of London, London, United Kingdom; Memorial Sloan Kettering Cancer Center, New York, NY; University of Washington, Seattle, WA; The Sidney Kimmel Comprehensive Cancer Center, Johns Hopkins Medicine, Baltimore, MD; Fox Chase Cancer Center, Philadelphia, PA; Merck & Co., Inc., Kenilworth, NJ; Erasmus MC Cancer Institute, Rotterdam, Netherlands

Background: Initial results of the phase 2 KEYNOTE-052 (NCT02335424) study led to approval of pembro for cisplatin-ineligible patients (pts) with advanced UC. Updated results representing follow-up of over 2 y since last pt enrolled are presented. Methods: Pts had confirmed advanced UC, were cisplatin- ineligible (ECOG PS 2, CrCl $30 to ,60 mL/min, grade $2 neuropathy/hearing loss, NYHA Class III heart failure), and received no prior chemotherapy for metastatic disease. Pts received pembro 200 mg IV Q3W until progression, unacceptable toxicity, withdrawal, or 24 mo of therapy, whichever occurred first. Primary end point was confirmed ORR (RECIST v1.1, independent central review). Key secondary end points: duration of response (DOR), overall survival (OS), and safety. Data cutoff was September 26, 2018. Results: Among pts assessed (N = 370), median age was 74 y, 85% had visceral disease, and 30% were PD-L1 positive (combined positive score [CPS] $10). Median follow-up was 11.4 mo (range, 0.1- 41.2) for all pts and 29.3 mo (range 7-41.2) for responders. Confirmed ORR was 29% (95% CI, 24-34): complete response, 9% (n = 33); partial response, 20% (n = 73). Median DOR was 30.1 mo (95% CI, 18.1-not reached [NR]); 67% and 52% of pts had DOR $12 and $24 mo, respectively. Median OS was 11.3 mo (range 9.7-13.1); 12- and 24-mo OS rates were 47% and 31%, respectively. In pts with CPS ,10 (n = 251) and $10 (n = 110), respectively, confirmed ORR was 20% (95%CI, 16-26) and 47% (95% CI, 38-57). Median DOR for pts with CPS , 10 and $10 was 18.2 mo (95% CI, 9.7-NR) and NR (95% CI, 18.1-NR); DOR $24 mo was 45% and 57%, respectively. Median OS for pts with CPS , 10 and $10 was 9.7 mo (95% CI, 7.6-11.5) and 18.5 mo (95% CI, 12.2-28.5); 24-mo OS rates were 24% and 47% respectively. Treatment-related adverse events (AEs) occurred in 67% of pts. Most common were fatigue and pruritus (18% each); 21% were grade $3, including 1 death (myositis). Conclusions: With extended follow-up, pembro continued to elicit clinically meaningful, durable antitumor activity in cisplatin- ineligible pts with advanced UC and was more pronounced in those with PD-L1 expression CPS $10. Pembro safety profile was as expected. Clinical trial information: NCT02335424.

4547 Poster Session (Board #373), Mon, 1:15 PM-4:15 PM

Interim analysis of the fierce-21 phase 2 (P2) study of vofatamab (B-701), a selective inhibitor of FGFR3, as salvage therapy in metastatic urothelial carcinoma (mUC).

Begona Mellado, Daniel E. Castellano, S Pang, Yuksul Urun, Se Hoon Park, Ulka N. Vaishampayan, Sumanta K. Pal, Graeme Currie, Esteban Abella, Florian D. Vogl, Andrea Necchi; Hospital Clinic de Barcelona, Barcelona, Spain; Medical Oncology Service, Hospital Universitario 12 de Octubre, Madrid, Spain; Linkou Chang Gung Memorial Hospital, Taoyuan City Taiwan, Taoyuan, Taiwan; Ankara University, Ankara, Turkey; Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, South Korea; Wayne State University, Detroit, MI; City of Hope National Medical Center, Duarte, CA; Rainier Therapeutics Inc, San Leandro, CA; Rainier Therapeutics, Inc, San Leandro, CA; Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy

Background: Patients (pts) with mUC with FGFR3 mutations who have failed platinum-based chemotherapy have a poor prognosis. Their response to immune checkpoint inhibitors appears diminished 10% or less compared to WT pts. 20% of mUC pts harbor FGFR3 mutations or fusions (M/F). Vofatamab is a fully human monoclonal antibody against FGFR3 that blocks activation of the wildtype and genetically activated receptor. FIERCE-21 is a Phase 1b/2 study designed to evaluate vofatamab monotherapy (VFM) or in combination with docetaxel (VFD). Methods: The P2 expansion enrolled mUC pts with FGFR3 M/F+ tumor (identified with FoundationONE CDx™), who failed $ 1 prior line of chemotherapy (including prior taxane for pts receiving VFM) or recurred # 12 months of (neo)adjuvant chemotherapy. Pts had measurable disease and ECOG # 1. Treatment consisted of vofatamab at 25 mg/kg alone and in combination with docetaxel at 75 mg/m2 q3w. Efficacy was assessed by investigators (RECIST 1.1). Primary objectives were safety and objective response-rate (ORR). Results: In the P2, 21 pts each received VFM and VFD. 57% of VFD pts had received at least 2, and 71% of VFM at least 3 prior lines of therapy. Best response to prior therapy was PD for 67% of VFD and 38% of VFM. The safety profile is consistent with previously reported data. TEAEs occurring in . 20% of pts were decreased appetite, diarrhea, pyrexia, asthenia, anemia, dyspnea, and fatigue. Most common vofatamab-related TEAEs in . 10% of pts were asthenia, diarrhea, decreased appetite and rash; all were Grade 1 or 2. In VFM, only 1 pt had a grade 3 TEAE and no pt discontinued treatment due to an AE. There were no cases of hyperphosphatemia, ocular or nail toxicity; 1 pt reported grade 2 skin toxicity. For pts receiving VM, median age was 70 yrs, ECOG 1 = 67%, Hgb , 10 g/dL 5%, liver metastases 19%. Responses have been seen in 7 pts to date including those receiving both VFM and VFD. Conclusions: Vofatamab both alone and combined with D in a q3w schedule are well tolerated with a low frequency of grade 3 TEAEs. Both VFM and VFD have demonstrated efficacy in terms of ORR. PFS/OS and DOR data will be presented at 7+ months for VFD and 9+ months for VFM. Clinical trial information: NCT02401542.

4548 Poster Session (Board #374), Mon, 1:15 PM-4:15 PM

Immune correlates of CD73 expression in patients with urothelial carcinoma (UC).

Edwin Lin, Roberto Nussenzveig, Andrew W Hahn, Mark Yandell, Lauren Christine Harshman, Neeraj Agarwal, Abhishek Tripathi; University of Utah/Huntsman Cancer Institute, Salt Lake City, UT; University of Utah Hunstman Cancer Institute, Salt Lake City, UT; University of Utah, Salt Lake City, UT; Dana-Farber Cancer Institute, Boston, MA; University of Utah Huntsman Cancer Institute, Salt Lake City, UT; University of Oklahoma, Stephenson Cancer Center, Oklahoma City, OK

Background: Checkpoint inhibitors have improved outcomes in UC. However, response rates are low and additional mechanisms of immune evasion need to be ascertained. CD73 (encoded by NT5E) converts extracellular AMP to adenosine, which exerts an immunosuppressive effect in the tumor microenvironment by inhibiting infiltrating T and NK cells. Utilizing The Cancer Genome Atlas (TCGA) bladder cancer dataset, we evaluated correlations between NT5E expression and the immune milieu in UC. Methods: RNA-seq data from 411 primary UC tumor samples were obtained from the TCGA. Patients were split into low, intermediate, and high NT5E expression groups (# -1, -1 to 1 and $1 standard deviation from the overall mean). A tumor inflammation signature (TIS) reflecting an inflamed tumor phenotype was calculated based on the averaged expression of 18 previously validated genes (Ayers et al, 2017). NT5E expression was compared between tumors with high and low TIS scores and among the TCGA molecular subtypes. Abundance of infiltrating immune cell subsets was estimated based on expression of previously identified 782 immune metagenes and compared between NT5E expression groups (Charoentong et al, 2017). The Mann-Whitney U test assessed statistical significance, and the Bonferroni correction was used to control for false discovery rate. Results: NT5E expression was significantly higher in tumors with a high TIS score compared to those with low TIS score (P,0.0001) and correlated with expression of other immune checkpoints such as PD-L1, IDO and LAG-3 (each P,0.01). Patients with basal/squamous subtype had the highest NT5E expression compared to luminal or neuronal subtypes. High NT5E expression was associated with increased infiltrating NK cells, neutrophils, Tregs and decreased Type 2 T helper cells. Conclusions: High expression of NT5E in UC patients with an inflamed tumor phenotype was associated with an increase in infiltrating Tregs, and the basal/squamous subtype. Our findings highlight a potential role of CD73-adenosine pathway as a mechanism of immune evasion and a novel therapeutic target in UC. Further studies to assess the clinical impact of NT5E expression on outcomes in UC patients treated with immunotherapy are needed. AT and NA: equal contribution.

4549 Poster Session (Board #375), Mon, 1:15 PM-4:15 PM

An FDA analysis of the association between adverse events and outcome in patients with urothelial cancer receiving a programmed death protein 1 or programmed death ligand 1 (anti-PD-1/L1) antibody.

Chana Weinstock, Virginia Ellen Maher, Laura L. Fernandes, Shenghui Tang, Sundeep Agrawal, Michael Holman Brave, Yang-Min Ning, Harpreet Singh, Daniel L. Suzman, James Xu, Kirsten B. Goldberg, Rajeshwari Sridhara, Amna Ibrahim, Marc Robert Theoret, Julia A. Beaver, Richard Pazdur; U. S. Food and Drug Administration, Silver Spring, MD; U.S. Food and Drug Administration, Silver Spring, MD; US Food and Drug Administration, Silver Spring, MD; Food and Drug Administration, Silver Spring, MD

Background: To assess the relationship between tumor response rate, overall survival, and the development of related adverse events of special interest (AESIs) or related immune-mediated adverse events (imAEs) in patients with urothelial cancer treated with anti-PD-1/L1 antibodies. Methods: We examined seven trials that led to drug approval and which included 1747 patients with metastatic or locally advanced urothelial cancer treated with an anti-PD-1/L1 antibody. Five trials enrolled patients who had received prior platinum-based therapy and two enrolled patients who were cisplatin-ineligible. The datasets were searched for AESIs, related AESIs, imAEs, and related imAEs. The relationship to study drug was determined by the Investigator. Immune-mediated adverse events were defined as AESIs treated with topical or systemic corticosteroids. Results: In these exploratory analyses, a related AESI was reported in 64% of responding patients and in 34% of patients who did not respond to the anti-PD-1/L1 antibody while a related imAE occurred in 28% and 12% of patients who did and did not respond to study drug, respectively. In a responder analysis, an increase in overall survival was seen in patients with related AESIs compared to those with no related AESI [hazard ratio (HR) 0.42; 95% CI: 0.37, 0.49]. Fifty- seven percent of responding patients with a related AESI reported a related AESI prior to documentation of response. Conclusions: Patients who responded to treatment with an anti-PD-1/L1 antibody were more likely to report a related AESI or related imAE. This relationship did not appear to be due to the increased duration of exposure in responding patients. Systemic corticosteroid use did not appear to affect the duration of response.

4550 Poster Session (Board #376), Mon, 1:15 PM-4:15 PM

PD-L1/PD-1 expression as a predictor of response to BCG in patients with high-risk non– muscle invasive bladder cancer.

Mathieu Roumiguie, Eva Comperat, Yann Neuzillet, Francois-Xavier Nouhaud, Vivien Graffeille, Alexandra Masson-Lecomte, Pierre Colin, Francois Audenet, Nadine Houede, Stephane Larre, Evanguelos Xylinas, Jeanne PIana Thomassin, Juliette Cotte, Geraldine Pignot, Leonor Chaltiel, Morgan Roupret; Department of Urology Institut Universitaire du Cancer de Toulouse – Oncopole - Rangueil University, Toulouse, France; Sorbonne University, GRC n°5, ONCOTYPE-URO, AP-HP Service of Pathology, Tenon Hospital, Paris, France; Versailles-Saint-Quentin-en-Yvelines University, Urology Department, Foch Hospital, Suresnes, France; Urology Department, Rouen University Hos- pital, UNIROUEN, IRON Group, Rouen, France; Department of Urology, Rennes University Hospital, Rennes, France., Rennes, France; Paris-Diderot University Urology Department, Saint-Louis Hospital (AP-HP), Paris, France; Urology Department Hopitalˆ prive ´ de la Louviere, ` Lille, France; Universite ´ Paris Descartes, AP-HP, Urology Department, Georges Pompidou European Hospital, Paris, France; Mont- pellier University, Departement ´ d’oncologie Medicale, ´ CHU Caremeau,, Nıˆmes, France; Urology Department, CHU de Reims, Reims, France; Paris Descartes University, AP-HP, Urology Department Bichat-Claude Bernard Hospital, Paris, France; Department of Pathology, Institut Paoli-Calmettes,, Marseille, France; Sorbonne University, GRC n°5, ONCOTYPE-URO, AP-HP, Urology department, Hopitalˆ Pitie-Salp ´ etriˆ ere,, ` Paris, France; Department of Surgical Oncology, institut Paoli-Calmettes,, Marseille, France; Biostatistics Unit, Institut Claudius Regaud-Institut Universitaire du Cancer de Toulouse (IUCT)-Oncopole, Toulouse, France; Sorbonne Universite, ´ GRC n°5, ONCOTYPE-URO, AP-HP, Hopitalˆ Piti´e-Salpetri`ˆ ere, Paris, France

Background: Intravesical BCG instillation (IBI) is the gold standard adjuvant treatment after transurethral resection of the bladder in high risk non muscle invasive bladder cancer (HR-NMIBC). IBI induce a type of Th1 immune response requiring a recruitment of cytotoxic cells. This response is downregulated by the PD-1/PD-L1 checkpoint inhibitors through an inhibition of the action of CD8+ T cells against their target. Our purpose was to assess whether PD-1/PD-L1 expression was associated with IBI response in HR- NMIBC. Methods: Histologically confirmed HR-NMBC from 5 academic French institutions which underwent maintenance IBI were retrospectively included. The following data were collected: pathological stage, grade, concomitant carcinoma in situ, number of lesions and size. From a paraffin embedded samples of initial resection, a unique dedicated uropathologist quantified immunochemistry expression of CD3, CD8, PD-L1 (antibody SP263/ SP142, E1L3N, 28 8) in both tumour cells and tumoral microenvironment. Univariate and multivariate analyses were performed using Cox proportional hazards model. Results: Overall140 patients (median age 66.5 years, range:35-87; sex ratio male vs female:6:1) were included. The distribution of NMIBC tumour for stage and grade was: Ta 37.2% (n = 52), T1 62.8% (n = 88) and high grade 100% (n = 140) respectively. The median number of IBI which were delivered was 12 (range 7-36). The median length of follow-up was 54.24 mo (95% CI = 49.91- 58.68). Overall, 25 patients (17.9%) had a recurrence/or progression. The 72mo Disease free survival (DFS) rate was 81.11% (95% CI = 72.20-87.41). Using univariate analysis, we found that Age (HR = 1.07; [1.02-1.12] p = 0.005), CD3/CD8 ratio (HR per 10 units = 3.43 [1.62-7.23] p = 0.014) and PD-L1(HR per 10 units = 1.65 [1.15-2.38] p = 0.046) were associated with DFS. In multivariate analysis, Age (HR = 1.07 [1.02-1.13] p = 0.009), CD3/CD8 ratio (HR per 10 units = 1.96 [1.28-3.00] p = 0.02) and PD-L1 expression in tumour cells (HR per 10 units = 3.38 [1.61-7.11] p = 0.01) remained significantly associated with DFS. Conclusions: PD-L1 expression in tumour cells and the T cells population in tumour microenvironment were both predictive factors of BCG response in HR-NMIBC. These results build a scientific rationale for pharmacological intervention on a molecular target using immuno-oncology.

4551 Poster Session (Board #377), Mon, 1:15 PM-4:15 PM

A pilot presurgical study evaluating anti-PD-L1 durvalumab (durva) plus anti-CTLA-4 tremelimumab (treme) in patients (pts) with high-risk muscle-invasive bladder carcinoma (MIBC) who are ineligible for cisplatin-based neoadjuvant chemotherapy (NAC).

Jianjun Gao, Arlene O. Siefker-Radtke, Neema Navai, Matthew T Campbell, Rebecca Slack Tidwell, Charles Guo, Ashish M. Kamat, Surena F. Matin, John C. Araujo, Amishi Yogesh Shah, Pavlos Msaouel, Jorge M. Blando, Luis M. Vence, Fei Duan, Sreyashi Basu, Shalini Singh, Hao Zhao, James Patrick Allison, Colin P.N. Dinney, Padmanee Sharma; The University of Texas MD Anderson Cancer Center, Houston, TX; Department of Immunology, The University of Texas MD Anderson Cancer Center, Houston, TX

Background: In MIBC pts, especially in those with high risk features including lymphovascular invasion, hydronephrosis, T3b disease, or variant histology, cisplatin-based NAC followed by cystectomy improves overall survival as compared with cystectomy alone. However, it is estimated that over 50% of pts with MIBC are ineligible for cisplatin-containing therapy. Therefore, we propose this pre-surgical trial with durva + treme for this population of pts. Methods: This is a single-arm, pre-surgical clinical trial with durva + treme in pts with localized, high-risk MIBC (cT2-T4a) who are ineligible for cisplatin-based NAC due to decreased renal function, neuropathy, hearing loss, or heart failure; or refuse cisplatin-based NAC (NCT02812420). Each patient receives durva (1500 mg) plus treme (75 mg) on weeks 1 and 5. Pts then undergo surgery at week 9-11. Pre- and post-treatment blood and tumor samples are collected for correlative biological analyses. Results: Twenty eight of 35 pts have been enrolled on this trial. Twenty-one pts have completed cystectomy as of 11/16/19. Of these 21 pts, 9 (43%) had pathologically complete response (pCR) and two (10%) had pathologic T1N0 (pT1) disease (#pT1N0 rate = 52%). Fourteen of 21 (67%) had down-staging of disease. Of note, 10 of these 21 pts had 3-D mass (T3) on exam under anesthesia or clinical T4a disease; 5 of these 10 pts (50%) had pCR and one (10%) had pT1 disease (#pT1N0 rate = 60%). Only 5 of 28 (17%) pts developed grade 3 immune related toxicity including hepatitis and amylase/lipase elevation, and two (7%) resulted in surgery delay for . 30 days. Immune profiling with CyTOF analysis of baseline peripheral blood indicates that pts with pCR have significantly lower frequency of a Th2 subset as compared to pts with up-staging of disease. In addition, gene expression profiling analysis of baseline tumor tissues demonstrates a significantly less immunosuppressive microenvironment in pts with pCR as compared to pts with up-staging of disease. Conclusions: Our data indicate that durva plus treme is an effective and safe neoadjuvant therapy for pts with MIBC ineligible for cisplatin-based therapy. Therefore, neoadjuvant therapy with durva + treme and a number of potential biomarkers warrant testing in a larger phase 3 trial. Clinical trial information: NCT 02812420.

4552 Poster Session (Board #378), Mon, 1:15 PM-4:15 PM

5-factor prognostic model for survival of patients with metastatic urothelial carcinoma receiving three different post-platinum PD-L1 inhibitors.

Guru Sonpavde, Juliane Manitz, Chen Gao, Daniel Hennessy, Doris Makari, Guenter Niegisch, Jonathan E. Rosenberg, Dean F. Bajorin, Petros Grivas, Andrea B. Apolo, Robert Dreicer, Noah M. Hahn, Matt D. Galsky, Andrea Necchi, Sandy Srinivas, Thomas Powles, Ashok Kumar Gupta, Shaad Essa Abdullah, Gregory Russell Pond; Dana-Farber Cancer Institute, Boston, MA; EMD Serono, Inc., Billerica, MA; MedImmune, Gaithersburg, MD; AstraZeneca, Gaithersburg, MD; Heinrich-Heine- Universitat, ¨ D¨usseldorf, Germany; Memorial Sloan Kettering Cancer Center, New York, NY; University of Washington, Seattle, WA; National Cancer Institute, National Institutes of Health, Bethesda, MD; University of Virginia, Charlottesville, VA; Johns Hopkins University School of Medicine, Baltimore, MD; Department of Medicine, Icahn School of Medicine at Mount Sinai, Tisch Cancer Institute, New York, NY; Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy; Stanford University Medical Center, Palo Alto, CA; Royal Free Hospital, London, United Kingdom; McMaster University, Hamilton, ON, Canada

Background: A prognostic model for overall survival (OS) of metastatic urothelial carcinoma (mUC) was previously reported in the setting of post-platinum atezolizumab (Pond GR, GU ASCO 2018). This model was limited by employing only atezolizumab treated patients (pts), small size of the validation dataset and unclear applicability to other PD-1/L1 inhibitors. Hence, we constructed a robust prognostic model utilizing the combined atezolizumab cohort as the discovery dataset and used 2 separate validation datasets comprised of post-platinum avelumab or durvalumab treated pts. Methods: The discovery dataset consisted of pt level data from 2 phase I/II trials (IMvigor210 and PCD4989g) evaluating atezolizumab (n = 405). Pts enrolled on 2 separate phase I/II trials, EMR 100070-001 that evaluated post-platinum avelumab (n = 242) and CD1108 that evaluated durvalumab (n = 189) comprised the validation datasets. Cox regression analyses evaluated the association of candidate prognostic factors with OS. Factors were dichotomized and laboratory values were normalized by logarithmic transformation. Step- wise selection was employed to propose an optimal model using the discovery dataset. Discrimination and calibration were assessed in the avelumab and durvalumab datasets following the validation procedure by Royston and Altman (2013). Results: The 5 factors included in the optimal prognostic model in the discovery dataset were ECOG-PS (1 vs. 0; HR 1.80; 95% CI [1.36-2.36]), presence/absence of liver metastasis (HR 1.55; 95% CI [1.20-2.00]), number of platelets (HR 2.22; 95% CI [1.54-3.18]), neutrophil-lymphocyte ratio (NLR; HR 1.94; 95% CI [1.57-2.40]) and lactate dehydrogenase (LDH; HR 1.60; 95% CI [1.28-1.99]). There was robust discrimination of survival between low, intermediate and high-risk groups based on 0-1, 2-3 and 4 factors. The concordance of survival was 0.692 in the discovery and 0.671 and 0.775 in the avelumab and durvalumab validation datasets, respectively. Acceptable or good calibration of expected 1-year survival rate was observed. Conclusions: A 5-factor prognostic model is prognostic for survival across 3 different PD-L1 inhibitors (atezolizumab, avelumab, durvalumab) in this large study totaling 836 pts overall in the setting of post-platinum therapy for mUC. This model may assist in prognostic stratification and interpreting nonrandomized trials of post-platinum PD1/L1 inhibitors.

4553 Poster Session (Board #379), Mon, 1:15 PM-4:15 PM

Correlation of circulating tumor DNA (ctDNA), tissue-based genomic profiling and clinical efficacy in the biomarker directed Ph1b trial in metastatic bladder cancer (BISCAY).

Danielle Carroll, Rob McEwen, Iwanka Kozarewa, Elizabeth Harrington, Anne L’Hernault, Jayantha Ratnayake, Richard Mather, Darren Hodgson, J Carl Barrett, Jan G. C. E. Cosaert, Pierre Emmanuel Brachet, Petros Grivas, Simon Chowdhury, Thomas Powles, Donal Landers; Translational Science, Oncology, IMED Biotech Unit, AstraZeneca, Cambridge, United Kingdom; AstraZeneca, Cambridge, United Kingdom; Oncology Translational Science, IMED Biotech Unit, AstraZeneca, Cambridge, United Kingdom; Astrazeneca, Cambridge, United Kingdom; AstraZeneca, Royston, United Kingdom; AstraZeneca, Boston, MA; Translational Sciences, IMED Biotech Unit, AstraZeneca, Waltham, MA; AstraZeneca, Cambridge, NJ, United Kingdom; GINECO-Centre François Baclesse, Caen, France; University of Washington, School of Medicine, Seattle, WA; Guy’s, King’s and St. Thomas’ Hospitals, and Sarah Cannon Research Institute, London, United Kingdom; Barts Cancer Institute, Queen Mary University of London, Royal Free NHS Trust, London, United Kingdom; Astra- Zeneca, Macclesfield, United Kingdom

Background: BISCAY is a biomarker-directed Ph1b multi-arm platform study exploring the combination of targeted therapies with anti-PD-L1, Durvalumab, in advanced urothelial cancer. Methods: Next generation sequencing (NGS) of tumour tissue samples from . 380 patients(pts) was performed using the FoundationOne assay alongside IHC for PD-L1. ct DNA from pts enrolled in trial modules at treatment initiation was profiled using the Guardant Health OMNI platform assessing a panel of 500 genes. For a subset of pts, serial plasma samples were also analysed to monitor early signs of response vs. resistance and changes in ct DNA dynamics using a bespoke NGS panel of 10 genes. Results: To date 149 pts have been actively enrolled across 7 different biomarker selected and unselected treatment modules. Across all screened pts the most prevalent genomic alterations in tumour tissue were TERT promoter (65%), TP53 (59%), KMT2D 21%, KDM6A 21%, with the most common CNV CDKN2A/B loss (32 %). All enrolled pts tested had detectable ctDNA in plasma. Similar genomic alterations, both frequency and type, were detected in both plasma ctDNA and tumour tissue with high concordance for module specific biomarkers used for patient allocation (80% (8/10) for ATM, BRCA1 and 2). Alterations in putative biomarkers predictive of response to anti-PD-L1, such as HRR/MMR alterations and high bTMB levels ( . 20mut/Mb) were observed in22% and 40% patient plasma samples, respectively. Correlations between biomarkers across modules treatment efficacy have been explored. Conclusions: All pts with advanced bladder cancer enrolled on BISCAY who were plasma profiled had detectable ctDNA; frequencies of genomic alterations (in both tumour tissue and plasma) were comparable to prior published data sets. ctDNA may be an attractive alternative to tissue-based NGS, providing comprehensive dynamic snapshots of genomic landscapes at the start and during therapy, and warrants further prospective investigation in trials.

4554 Poster Session (Board #380), Mon, 1:15 PM-4:15 PM

Recombinant humanized anti-PD-1 monoclonal antibody toripalimab in patients with metastatic urothelial carcinoma: Preliminary results of an open-label phase II clinical study.

Xinan Sheng, Haige Chen, Xin Yao, Yi Hu, Xudong Yao, Ziling Liu, Fangjian Zhou, Yiran Huang, Jun Guo, Shanghai Junshi Biosciences Co., Ltd.; Peking University Cancer Hospital, Beijing, China; Ren Ji Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China; Tianjin Cancer Hospital, Tianjin, China; Department of Medical Oncology, Chinese PLA General Hospital, Beijing, China; Shanghai the Tenth People’s Hospital of Tongji University, Shanghai, China; Department of Tumor Center, First Hospital, Jilin University, Changchun, China; Sun Yat-sen University Cancer Center, Guangzhou, China; Renji Hospital Shanghai Jiaotong University School of Medicine, Shanghai, China; Key Laboratory of Carcinogenesis and Translational Research (Ministry of Education/Beijing), Department of Renal Cancer and Melanoma, Peking University Cancer Hospital & Institute, Collab- orative Innovation Center for Cancer Medicine, Beijing, China

Background: Patients with advanced metastatic urothelial carcinoma (UC) who experience disease progression after standard therapy have limited treatment options. Phase I studies of toripalimab in subjects with heavily pretreated metastatic UC have demonstrated an acceptable safety profile and promising clinical activity. Here we report the preliminary safety and efficacy result of toripalimab in a phase II clinical study in Chinese patients with refractory/metastatic urothelial carcinoma. (Clinical trial ID: NCT03113266). Methods: Metastatic UC Patients will receive toripalimab, also known as JS001, 3 mg/kg Q2W until disease progression or unacceptable toxicity. All patients with measurable disease will be assessed for clinical response every 8 weeks according to RECISTv1.1. Tumor PD-L1 expression and tumor mutational burden will be measured for correlation with clinical response. Results: From May 2017 to February 10, 2019, 79 patients were enrolled from 7 participating centers. The median age was 61 years with 57.5% male. By the cut-off date of Jan 20, 2019, common treatment related AEs were mostly grade 1 or 2, including anemia, hyperglycemia, ALT increased, AST increased and hypothyroidism. Among 65 evaluable patients, 2 complete responses, 18 partial responses, and 13 stable diseases were observed, for an objective response rate (ORR) of 30.8% and a disease control rate of 50.8%. 70% (14/20) responses were ongoing by the cut-off date. PD-L1 expression results were obtained from 56 subjects. PD-L1+ patients (n=16, 28.6%) had significant better ORR than PD-L1- patients (n=40), 62.5% versus 15.0% (p,0.01). Conclusions: Toripalimab has demonstrated encouraging clinical activity in chemo- refractory UC patients and a manageable safety profile. Toripalimab elicited a favorable 62.5% ORR in PD-L1 positive patients, while PD-L1 negative patients also achieved a 15% ORR, including one complete response. Patients will be continuously monitored for additional safety and efficacy readouts (DOR, PFS and OS). Clinical trial information: NCT03113266.

4555 Poster Session (Board #381), Mon, 1:15 PM-4:15 PM

Circulating tumor cell (CTC) enumeration in patients (pts) with metastatic genitourinary (mGU) tumors treated in a phase I study of cabozantinib and nivolumab (CaboNivo) +/- ipilimumab (CaboNivoIpi).

Andrea B. Apolo, Amir Mortazavi, Zishuo Ian Hu, Joseph Schonhoft, Lincy Chu, Amanda K. L. Anderson, Yipeng Wang, Ryan Dittamore, Sumanta K. Pal, Primo Lara, Mark N. Stein, Seth M. Steinberg, Christian Mayfield, Lisa M. Cordes, Marissa Mallek, Rene Costello, Carlos Diaz, Jane B Trepel, Donald P. Bottaro; Genitourinary Malignancies Branch, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Bethesda, MD; Arthur G. James Cancer Hospital, Ohio State University Wexner Medical Center, Columbus, OH; Stony Brook Hospital, Stony Brook, NY; Epic Sciences, Inc., San Diego, CA; Epic Sciences, San Diego, CA; City of Hope National Medical Center, Duarte, CA; University of California, Davis, Sacramento, CA; Columbia University Medical Center, New York, NY; Biostatistics and Data Management Section, NCI, NIH, Bethesda, MD; National Cancer Institute, National Institutes of Health, Bethesda, MD; National Institutes of Health, Bethesda, MD; National Cancer Institute, Bethesda, MD; Center for Cancer Research, NCI, NIH, Bethesda, MD; Center for Cancer Research, Division of Cancer Treatment and Diagnosis, Bethesda, MD

Background: CTCs may serve as biomarkers for clinical outcomes in GU tumor pts. We examined the association between baseline CTC enumeration, CTC heterogeneity, CTC morphologic subtypes, and progression-free-survival, overall survival and response to therapy with combination CaboNivo or CaboNivoIpi. Methods: 123 samples from 52 pts with mGU tumors treated with CaboNivo (38 pts) or CaboNivoIpi (14 pts) drawn at Baseline, Cycle (C) 2 Day (D) 1, and C3D1 were processed using the Epic Sciences platform. CTCs were defined as cytokeratin (CK)+, CD45-, distinct morphology, intact nucleus. PD-L1 expression was also assessed. Results: From 07/20/2016-09/01/2018, 52 pts [urothelial carcinoma (UC) N = 33; plasmacytoid N = 1; Clear cell renal cell carcinoma N = 4; bladder adenocarcinoma N = 8; bladder squamous cell carcinoma N = 2; bladder small cell N = 2; renal medullary N = 2] were treated. Median age was 61.5 years (range 20-82); 35 (67%) were male. N = 37 (71%) had visceral involvement, N = 15 (29%) with liver involvement, N = 11 (21%) with bone involvement. CTCs were found in the peripheral blood of 26/40 (65%) pts at baseline. 1/40 pts (bladder adenocarcinoma) had PDL1+ CTCs at Baseline. Median CTC/mL at Baseline, C2D1, C3D1, were not significantly different. CTC counts of . 2and. 4 at C2D1 were potentially associated with shorter OS (p = 0.071 and p = 0.045 without adjustment for multiple cutoffs for evaluation). PFS results exhibited similar trends. Unsupervised clustering of CTC images identified 5 main CTC subtypes, of which the presence of one was significantly associated with shorter OS (p = 0.0014, not adjusted for multiple testing). Additionally, we observed a trend towards patients with higher CTC heterogeneity at C2D1 having longer OS, when adjusting for the risk associated with CTC enumeration (p = 0.084). Conclusions: CTC were detected in pts with mGU tumors treated with CaboNivo and CaboNivoIpi. CTC values were somewhat but not statistically lower in responders vs. non-responders. On treatment lower CTCs and the absence of aggressive CTC subtypes were associated with better clinical outcomes. Ongoing analyses include single cell genomics, and analysis of T-cell populations. Clinical trial information: NCT02496208.

4556 Poster Session (Board #382), Mon, 1:15 PM-4:15 PM

Causes of patient ineligibility in clinical trials of metastatic urothelial cancer.

Gwynn Ison, Virginia Ellen Maher, Chana Weinstock, Sundeep Agrawal, Michael Holman Brave, Yang-Min Ning, Harpreet Singh, James Xu, Amna Ibrahim, Julia A. Beaver, Richard Pazdur; U.S. Food and Drug Administration, Silver Spring, MD; U. S. Food and Drug Administration, Silver Spring, MD; Food and Drug Administration, Silver Spring, MD; US Food and Drug Administration, Silver Spring, MD

Background: Registrational trials of PD-1/PD-L1 inhibitor therapy help inform clinicians and patients about the expected outcomes of patients receiving these drugs. However, the clinical trial population does not reflect all patients with the underlying disease. There have been ongoing efforts to expand eligibility criteria for clinical trial enrollment that are expected to mitigate some of these factors. We reviewed four registrational trials of PD-1/PD-L1 inhibitor therapy to better understand why patients were deemed ineligible for trial enrollment and whether expanded eligibility criteria may have addressed the reasons for ineligibility. Methods: We reviewed four trials that led to approval of PD-1/PD-L1 inhibitor therapy. These trials screened 1931 and enrolled 1253 patients with metastatic or locally advanced urothelial cancer who had previously received platinum-based therapy. We examined the datasets to determine patient demographics and reasons for trial ineligibility. Results: There were no differences in the demographics characteristics of patients who were eligible or ineligible for study treatment. However, when compared to the SEER database the screened population was younger, and minorities were under-represented. Causes of patient ineligibility included: 1) Lack of Tumor Tissue/PD-L1 Low Tumor Staining (23%), 2) Underlying Disease characteristics were not met (19%), 3) Laboratory Abnormalities (18%), 4) Excluded Co-morbid Conditions (16%), 5) Poor Performance Status (15%), 6) Refused Consent (10%), 7) Other (7%), and 8) Incorrect Prior Therapy (5%). Conclusions: Clinical trial accrual should be representative of all patients with the underlying disease. Changes in trial eligibility criteria such as less stringent requirements concerning tumor tissue (when possible), co-morbid conditions, laboratory abnormalities, and performance status may improve patient accrual. These, when added together, formed the majority of the reasons for clinical trial ineligibility.

4557 Poster Session (Board #383), Mon, 1:15 PM-4:15 PM

Impact of antibiotic use on clinical outcomes in patients with urothelial cancer receiving a programmed death protein 1 or programmed death ligand 1 (anti-PD-1/L1) antibody.

Chana Weinstock, Virginia Ellen Maher, Laura L Fernandes, Shenghui Tang, Sundeep Agrawal, Michael Holman Brave, Yang-Min Ning, Harpreet Singh, Daniel L. Suzman, James Xu, Kirsten B. Goldberg, Rajeshwari Sridhara, Amna Ibrahim, Marc Robert Theoret, Julia A. Beaver, Richard Pazdur; U. S. Food and Drug Administration, Silver Spring, MD; U.S. Food and Drug Administration, Silver Spring, MD; FDA, Silver Spring, MD; US Food and Drug Administration, Silver Spring, MD; U.S. Food and Drug Administration, Washington, DC; Food and Drug Administration, Silver Spring, MD

Background: Previous data has suggested that patients treated with anti-PD-1/L1 antibodies who receive antibiotics during their therapy might have dramatically decreased progression-free and overall survival 1,2. This has clinical implications for management of patients with suspected bacterial infection while on treatment with these agents. We assessed the relationship between antibiotic use and tumor response rate, progression-free survival, and overall survival in a large dataset of patients with urothelial cancer treated with anti-PD-1/L1 antibodies. Methods: We examined seven trials that led to drug approval and which included 1747 patients with metastatic or locally advanced urothelial cancer treated with an anti- PD-1/L1 antibody. Five trials enrolled patients who had received prior platinum-based therapy and two enrolled patients who were cisplatin-ineligible. Six were single arm trials and one was a randomized controlled trial whose control arm is not included in these analyses. Concomitant medication datasets were searched for systemic antibiotic used by each patient while on treatment. Results: Overall, 51% of patients (n=892) were exposed to antibiotics (ABX+) and 49% (n=855) were not exposed (ABX-). In these exploratory analyses, small numeric differences in OS, PFS, and ORR were seen in ABX+ vs. ABX- patients. Median OS was 9.23 vs. 9.86 months, median PFS was 105 vs 101 days, and ORR was 20% vs. 21% in ABX+ vs. ABX- patients, respectively. Conclusions: Patients who were treated with antibiotics while on therapy with an anti-PD-1/L1 antibody for urothelial cancer had similar outcomes to those who were not treated with antibiotics. Numeric differences in outcomes were not significant and did not duplicate previous analysis demonstrating a median OS that was doubled in ABX- patients1.Our exploratory analyses do not appear to demonstrate a clear need for practitioners to avoid antibiotic use in patients treated with PD-1/L1 agents for fear of significantly impacting clinical outcomes. References: 1) Tinsley et. al., ASCO annual meeting 2018, abstract 3010 2) Routy et. al., Science 05 Jan 2018: Vol. 359, Issue 6371.

Antibiotics = Yes Antibiotics = No N = 892 N = 855 Hazard Ratio Overall Survival 9.23(7.62,10.78) 9.86(8.74,10.97) 1.06(95%CI: 0.94,1.21) Median(95%CI) months Progression Free Survival 105(86,113) 101(84,122) 1.06(95%CI: 0.94,1.2) Median(95%CI) days Overall response rate, % 20(17,22) 21(18,24) N/A (confirmed, per independent review)

4558 Poster Session (Board #384), Mon, 1:15 PM-4:15 PM

RNAseq and DNA whole-exome sequence analysis reveal novel response signatures to IO treatment in muscle invasive bladder cancer (MIBC) patients.

Gregory Mayhew, Yoichiro Shibata, Joshua M Uronis, Michele C Hayward, Tracy L Rose, William Y. Kim, Charles M. Perou, Myla Lai-Goldman, Michael Vance Milburn; GeneCentric Therapeutics, Inc., Re- search Triangle Park, NC; University of North Carolina at Chapel Hill, Chapel Hill, NC; University of North Carolina Hospital, Chapel Hill, NC; UNC Lineberger Comprehensive Cancer Center, Chapel Hill, NC; The University of North Carolina at Chapel Hill, Chapel Hill, NC

Background: Objective: To examine in a cohort of anti-PD-(L)1 immune checkpoint inhibitors (ICP) treated urothelial cancer patients a strategy combining treatment outcomes with molecular alterations, pathways, and immune/tumor microenvironment features to determine potential responder and rapid- progression signatures. Methods: De-identified clinical history and treatment outcomes were collected on 109 MIBC patients treated with ICP agents. Archived FFPE samples from these patients were obtained and processed for mRNAseq, exome-seq, tumor mutation burden (TMB), microsatellite instability (MSI) and mutation panel testing. Comprehensive tumor/immune profiling is being analyzed in the context of ICP treatments and RECIST 1.1 outcomes. A 60 gene MIBC 4-typer expression subtyper and other response associated predictors are used to stratify and identify positive/negative ICP response indicators. Results: 109 patients were identified (median age 75, 64% male, 78% white, 17% black). 74% of patients had received prior platinum-based chemotherapy, and 12% had received 2 or more prior lines of therapy. At initiation of ICP, 28% of patients had hemoglobin , 10, 30% had liver metastases, and 59% had ECOG performance status . 0. Mutation analysis of the first 66 patients showed TP53 (n = 34, 52%), FGFR (n = 17, 26%), CDKN2A (n = 13, 20%) and RB1 (n = 12, 18%) as the top alterations. No patients (0/8) with known pathogenic mutations in FGFR3 (S249C and TACC3- fusion) responded to ICP. Of patients with T2 staging prior to ICP (37/66), overall survival was markedly shorter (2.7 years) in those possessing FGFR3 mutations (n = 6/37) compared to that for FGFR3 WT patients (5.7 years, n = 31/37; p = 0.045). Further analyses of molecular features relative to treatment outcomes are ongoing to characterize response signatures. Conclusions: Our preliminary cohort of patients with pathogenic FGFR3 alterations showed 0% favorable response to ICP. We are expanding on this observation with further comprehensive molecular analyses and retrospective treatments/ outcomes data. We anticipate identifying expression signatures that reflect ICP patient responder/non- responder signatures that may aid in future therapy decisions.

4559 Poster Session (Board #385), Mon, 1:15 PM-4:15 PM

A pilot safety study of gemcitabine and cisplatin (GC) with atezolizumab (A) as first-line therapy in patients (pts) with metastatic urothelial cancer (mUC).

Samuel Aaron Funt, Karan Jatwani, Michelle Makris, Ashley Marie Regazzi, Chung-Han Lee, Min Yuen Teo, Deaglan Joseph McHugh, Asia S. McCoy, Grace Hettich, Phillip Wong, Moshen Abu- Akeel, Jedd D. Wolchok, Taha Merghoub, Hikmat Al-Ahmadie, Irina Ostrovnaya, Joshua Chaim, Jeremy C. Durack, Gopa Iyer, Dean F. Bajorin, Jonathan E. Rosenberg; Memorial Sloan Kettering Cancer Center, New York, NY; Mount Sinai St Luke’s and West Hospital, New York, NY

Background: GC has a high overall response rate (ORR) but a high relapse rate in pts with untreated mUC. Inhibition of programmed death-ligand 1 (PD-L1) with A can lead to long-term survival, but single-agent ORR is modest. We report the outcomes of GC+A in a cohort pts with mUC. Methods: This study was designed to assess the safety of GC + A in 10 pts with untreated mUC prior to testing GC + A in a neoadjuvant study in pts with muscle-invasive disease. The primary endpoint was safety as assessed by a predefined dose limiting toxicity (DLT) rate during the first cycle in the first 6 pts. Total accrual goal was 10 pts to collect preliminary data on ORR and progression-free survival (PFS). RECIST 1.1 assessments were performed every 9 wks. Pts received 6 cycles of GC + A induction and then A maintenance every 3 wks. Results: No DLTs occurred during the first cycle in the first 6 pts. Grades 3-4 neutropenia and anemia occurred in 6/10 and 7/10 pts, respectively. Three pts required gemcitabine dose reductions for hematologic toxicity and 2 pts had febrile neutropenia. One pt discontinued cisplatin after 2 cycles for grade 3 hearing impairment but completed induction with gemcitabine and A. Only 1 pt discontinued study therapy due to treatment-related adverse events (AEs), including A-related grade 4 encephalopathy and grade 3 polyneuropathy. Three of 10 pts had visceral (liver or bone) metastases. Of the 10 pts, 1 pt is completing induction but meets initial criteria for partial response (PR), 8 pts had confirmed PR, and 1 pt had progressive disease (PD). Of 9 pts with confirmatory scans, the median PFS was 10.6 months (95% CI 6.7, N/A). Of 8 pts with confirmed PR, 5 eventually had PD, 1 has just completed induction, 1 remains without PD at 25 months, and 1 had consolidation surgery with a pathologic complete response and remains disease-free at 21 months. Conclusions: This 10 pt study met its primary safety endpoint. The neoadjuvant study is ongoing (NCT02989584). Although there were a substantial number of grade 3-4 toxicities, therapy was discontinued due to treatment-related AEs in only 1 pt. Immune correlative studies are ongoing. Clinical trial information: NCT02989584.

4560 Poster Session (Board #386), Mon, 1:15 PM-4:15 PM

Multi-omics profiling of upper-tract urothelial carcinomas.

Gabriel G. Malouf, Hui Yao, Roger Mouawad, Morgan Roupret, Jean-Emannuel Kurtz, Jean-Philippe Spano, INSTITUT CANCEROLOGIE Meyer KHAYAT, Eva Comperat, Xiaoping Su; Strasbourg University Hospital, Strasbourg, France; University of Texas MD Anderson Cancer Center, Houston, TX; Pitie-Salpetriere Hospital, Paris, France; Sorbonne Universite, ´ GRC n°5, ONCOTYPE-URO, AP-HP, Hopitalˆ Piti´e-Salpetri`ˆ ere, Paris, France; CHU Stasbourg, Strasbourg, France; Pitie-Salp ´ etri`ˆ ere Hospital, Sorbonne University, Cancer University Institute, Paris, France; Pitie-Salp ´ etri`ˆ ere Hospital, Paris, France; Sorbonne University, GRC n°5, ONCOTYPE-URO, AP-HP Service of Pathology, Tenon Hospital, Paris, France

Background: Upper-tract urothelial carcinomas (UTUC) may harbor similar genetic profile as compared to bladder cancer, although frequencies of mutated genes have been shown to differ between them. However, to the best of our knowledge, the epigenetic landscapes of UTUC and their association with genetic alterations and clinico-pathological tumor features remain unknown. Methods: We collected 40 UTUC samples (20 non-muscle invasive (NMI) and 20 muscle-invasive (MI) and carried out whole- exome sequencing (n = 30), DNA methylation using Infinium EPIC arrays (n = 35) and RNA sequencing (n = 20). Validation was performed on TCGA bladder cancer dataset. Results: We identified 3232 putative somatic mutations with an average of 2.1+/-2.6 mutations per megabase. Significantly mutated genes were FGFR3 (50%), KDM6A (27%), MLL2 (27%) and ARID1A/B (23%). No difference in term of genetic alterations were identified between MI- and NMI- UTUC. Unsupervised hierarchical clustering using most variable DNA methylation probes uncovered two robust DNA methylation epi- clusters. Epi-cluster C1 (n = 23; 65.7%) displayed markedly higher DNA methylation relative to Epi- cluster C2 (n = 12; 34.3%). Notably, all muscle-invasive samples were enriched in C1 (16/17, 94.1%); conversely, C2 was enriched with non-muscle-invasive samples (p = 0.0009). Overall, 14.209 probes were significantly hypermethylated in C1 as compared to C2 epi-cluster; Gene Set Enrichment Analysis (GSEA) demonstrated that those were enriched for PRC2 targets (p = 6x10-65). Integrative analysis with tumor genetic landscape showed that C1 epi-cluster was enriched for mutations in SWI/SNF complex as compared to C2 (p = 0.02). We then applied our epi-signature to bladder TCGA cohort and obtained two similar epi-clusters associated with patients overall survival (p = 0.035). Conclusions: Our study demonstrate for the first time that difference between MI- and NM- invasive UTUC might be related to epigenetic rather than genetic alterations. This might pave the way for testing epigenetic therapies in non-muscle invasive tumors with the aim to prevent recurrence and distant metastasis.

4561 Poster Session (Board #387), Mon, 1:15 PM-4:15 PM

Preliminary phase 2 clinical results of IL-15RaFc superagonist N-803 with BCG in BCG- unresponsive non-muscle invasive bladder cancer (NMIBC) patients.

Karim Chamie, John H. Lee, Amy Rock, Peter R. Rhode, Patrick Soon-Shiong; Institute of Urologic Oncology (IUO), Department of Urology, David Geffen School of Medicine at UCLA, Los Angeles, CA; Nantkwest, Culver City, CA; NantCell, Inc., Miramar, FL

Background: Patients with non-muscle-invasive bladder cancer (NMIBC) unresponsive to BCG therapy have limited treatment options. N-803 (also known as ALT-803) is an IL-15-based immunostimulatory protein complex (IL-15RaFc) that promotes proliferation and activation of natural killer (NK) cells and CD8+ T cells, but not regulatory T cells. Phase Ib data in BCG-na¨ıve patients with NMIBC demonstrate that intravesical administration of N-803 with BCG induced complete response in all patients, without recurrences for the study duration of 24 months. Methods: An open-label, single-arm multicenter Phase 2 study of intravesical BCG plus N-803 in patients with BCG-unresponsive high-grade NMIBC (NCT03022825) was opened. The study has two cohorts: Cohort A, patients with BCG-unresponsive carcinoma in situ (CIS) [with or without Ta or T1 disease] and Cohort B, patients with BCG-unresponsive high-grade Ta/T1 disease. All treated patients receive intravesical N-803 plus BCG, similar to a standard induction and maintenance treatment schedule. The primary endpoint for Cohort A is incidence of complete response (CR) of CIS at any time, and the primary endpoint for Cohort B is disease-free rate at 12 months. Results: To date, forty-six patients have enrolled in this phase 2 trial (Cohort A (CIS), n = 23, Cohort B (Papillary), n = 23). Of eleven evaluable patients in Cohort A, nine patients (82%) have a reported CR. In addition, seven out of nine (78%) patients in Cohort A demonstrated CR at their 6-month response assessment. Of thirteen evaluable patients in Cohort B, ten patients (77%) showed no evidence of recurrence at their 3-month response assessment; of these, none (0/8) evaluated past 3 months have had disease recurrence. Three serious adverse events (AEs) have been reported (E coli infection, anemia, and bacteremia), with no immune-related AEs. Conclusions: Nine out of eleven (82%) patients with BCG- unresponsive CIS of the bladder demonstrated a complete response. Ten out of thirteen patients with BCG- unresponsive papillary NMIBC show no evidence of disease at first assessment. Intravesical N-803 plus BCG was well-tolerated and no patients experienced immune-related AEs. Clinical trial information: NCT03022825.

4563 Poster Session (Board #389), Mon, 1:15 PM-4:15 PM

Biomarkers of outcomes in a randomized phase II trial of first-line paclitaxel, ifosfamide, and cisplatin (TIP) versus bleomycin, etoposide, and cisplatin (BEP) for intermediate- and poor-risk germ cell tumors (GCT).

Darren R. Feldman, James S Hu, Sujata Patil, Victor E. Reuter, Sandy Srinivas, Walter Michael Stadler, Brian Addis Costello, Matthew I. Milowsky, Leonard Joseph Appleman, Tanya B. Dorff, Maria Bromberg, Gabriella Joseph, Samuel Aaron Funt, Dean F. Bajorin, George J. Bosl, David I. Quinn, Robert J. Motzer; Memorial Sloan Kettering Cancer Center, New York, NY; University of Southern California Norris Comprehensive Cancer Center, Los Angeles, CA; Stanford University Medical Center, Palo Alto, CA; University of Chicago, Chicago, IL; Mayo Clinic, Rochester, MN; University of North Carolina at Chapel Hill School of Medicine, Chapel Hill, NC; University of Pittsburgh Medical Center, Pittsburgh, PA; City of Hope, Duarte, CA

Background: We previously reported no difference in favorable response rate (FRR) or PFS for TIP vs BEP. Here we present results of a pre-planned analysis of biomarkers of outcome. Methods: HCG and AFP were drawn on days 1 and 15 of each cycle and rates of decline classified as satisfactory [S] or unsatisfactory [US] by MSK (Motzer JCO 2007) and GETUG (Fizazi Lancet Oncol 2014) methods. IHC for ERCC1, RAD51, PARP1, HER-2, and p-AKT was performed on pre-treatment tumor samples. An H-score (0 – 300) was calculated for each stain (H = stain intensity [0 – 3] x % positive cells [0-100]). H-score and marker decline category were correlated with FRR (PR + CR) and PFS. Patients (pts) who received disease-stabilizing chemotherapy were excluded from marker analyses. Results: Of 91 pts, 80 did not receive disease-stabilizing treatment with 79 having sufficient marker values for analysis by the MSK method and 75 by GETUG. By MSK, 49 had S decline vs 30 US; by GETUG, 34 S vs 41 US. FRR and PFS were improved for pts with S vs US decline by both methods and remained significant by the MSK method when stratified by IGCCCG group (Table). IHC (n=77) quality was adequate in 71 to 73 pts (varied by stain) and was positive (H .0) for PARP in 68/73, ERCC1 in 54/71, RAD51 in 54/73, p-AKT in 5/72, and HER2 in 4/72. Only PARP1 was associated with outcome with worse PFS for the lowest expression tertile (H , 180; p=0.013). Conclusions: PARP1 expression and tumor marker decline rates, particularly by MSK method, were significantly associated with outcome to initial chemotherapy in int/ poor risk GCT. Future trials incorporating marker decline into treatment allocation and validating the prognostic effect of PARP1 expression are warranted. Clinical trial information: NCT01873326.

MSK Method GETUG Method SvsUS SvsUS Outcome by marker decline.* (n = 49 vs 30) (n = 34 vs 41) All patients FRR, % 94 vs 53 91 vs 66 (p,0.001) (p=0.01) 2y PFS, % 90 vs 50 88 vs 60 (p,0.001) (p=0.02) Intermediate-risk FRR, % 100 vs 75 100 vs 85 (p=0.008) (p=0.07) 2y PFS, % 100 vs 63 100 vs 76 (p,0.001) (p=0.02) Poor-risk FRR, % 85 vs 45 77 vs 57 (p=0.006) (p=0.22) 2y PFS, % 76 vs 45 69 vs 53 (p=0.02) (p=0.59)

*log-rank p value used for 2y PFS comparisons.

4564 Poster Session (Board #390), Mon, 1:15 PM-4:15 PM

A reduced pazopanib dose with food: Is it more patient-friendly and does it reduce drug costs?

Floor Lubberman, Hans Gelderblom, Paul Hamberg, Walter Vervenne, Sasja F. Mulder, Frank Jansman, Angela Colbers, Winette T.A. Van Der Graaf, David Burger, Saskia Luelmo, Dirk Jan A.R. Moes, Carla M.L.- Van Herpen, Nielka P. Van Erp; Radboud University Medical Center, Department of Pharmacy, Nijmegen, Netherlands; Leiden University Medical Center, Department of Medical Oncol- ogy, Leiden, Netherlands; Department of Internal Medicine, Franciscus Gasthuis & Vlietland, Rotter- dam, Netherlands; Deventer Ziekenhuis, Deventer, Netherlands; Radboud University Medical Center, Nijmegen, Netherlands; Deventer Ziekenhuis, Department of Clinical Pharmacy, Deventer, Nether- lands; Radboud University Medical Center, Department of Medical Oncology, Nijmegen, Netherlands; Department of Clinical Pharmacy & Toxicology, Leiden University Medical Center, Leiden, Netherlands

Background: Pazopanib has been licensed for advanced soft tissue sarcoma and metastatic renal cell carcinoma in a fixed oral daily dose of 800mg taken fasted. We hypothesized that ingesting pazopanib with food may improve patients’ comfort and reduce gastro-intestinal adverse events. Moreover, a food intervention, resulting in a better absorption, can lead to a lower dose, which could significantly reduce treatment costs. Methods: Part 1 of the study was performed to determine whether 600mg pazopanib taken with a continental breakfast was bioequivalent to 800mg pazopanib taken fasted. In part 2, differences in GI-toxicity and patient satisfaction were assessed by the cancer-therapy-satisfaction- questionnaire after both intake regimens. Finally, patient’s preference for either intake regimen was asked. Results: 16 patients were included in the bioequivalence study. The geometric mean ratio (fed/ fasted) of the area under the plasma concentration time curve was 1.10 (90% CI 1.00-1.19), maximum peak concentration was 1.12 (90% CI 1.02-1.22) and pazopanib trough concentration was 1.10 (90% CI 1.02-1.18). In part 2, 60 patients were included. No differences were seen in the occurrence of GI-toxicities under both intake regimens. Patients seem to be more positive about their feelings about side effects (72.3(95% CI 68.1-76.5) vs 68$2 (62.7-73.6); p=.092) and satisfaction with therapy scores were higher (84.7(95% CI 81.4-87.9) vs 81.9 (78.7-85.2); p= .059) when pazopanib was taken with food. 41 (68%) of the patients preferred the intake with continental breakfast. Conclusions: Intake of 600mg pazopanib with food results in bioequivalent exposure and was preferred over a standard pazopanib dose without food. Moreover, with this simple food intervention a large cost reduction can be realized in patients treated with pazopanib. Clinical trial information: NCT02138526.

4565 Poster Session (Board #391), Mon, 1:15 PM-4:15 PM

Comparative genomic profiling from tumor tissue and circulating tumor DNA (ctDNA) in 111 patients (Pts) with metastatic clear cell renal cell carcinoma.

Ritesh Kotecha, Erika Gedvilaite, Samuel J. Murray, Robert J. Motzer, Dana Tsui, Martin Henner Voss; Memorial Sloan Kettering Cancer Center, New York, NY

Background: Circulating tumor DNA (ctDNA) assessment is a non-invasive approach for genomic interrogation of solid tumors. As a novel tool, key benchmarks for applications in metastatic clear cell renal cell carcinoma (mccRCC) are yet to be determined. To understand the utility of ctDNA, we performed a large cohort analysis using a comparative genomics approach integrating matched primary tissue and ctDNA genomic data. Methods: Pts with prior tumor mutational profiles generated via next generation sequencing (NGS) from nephrectomy or metastatic specimens underwent single-time point plasma collection. Targeted NGS sequencing with MSK-IMPACT was performed on tumor and ctDNA with subsequent bi-directional cross genotyping using Waltz 2.0. All pts had matched germline comparison from peripheral blood; clinical data was extracted from medical record. Liberal (1-2 reads) and stringent ($3 reads) filters were applied, with a cut-off of , 30% allele frequency to remove germline mutations. Results: 111 mccRCC pts, of whom available IMDC-risk was favorable (35%), intermediate (60%), and poor-risk (5%) were included for analysis. The median time between tissue and ctDNA collection was 23 months (R: 1-177), and 96% of patients had undergone nephrectomy prior to ctDNA collection. In primary tissue sequencing, 64/111 (58%) from nephrectomy and 42/111 (42%) from metastatic sites, 569 unique alterations were identified across the whole cohort, with a median of 4 mutations/pt (R:1-23). RCC-specific alterations included VHL (88%), PBRM1 (48%), SETD2 (34%), KDM5C(17%),TP53(14%). Across the cohort, 176 alterations were identified in ctDNA. With cross genotyping, ctDNA alterations concordant with primary tumors were detected in 20% (22/111 pts, 28 unique alterations) using stringent criteria with a median of 1 mutation/pt (R:1-2). Using liberal criteria, concordance with primary tumors was 59% (66/111 pts, 142 unique alterations) with a median of 2mutations/pt(R:1-8).Conclusions: This large cohort study matching oncogenomics from tumor and ctDNA highlights complexities and challenges of applying liquid biopsy in biomarker development in mccRCC.

4566 Poster Session (Board #392), Mon, 1:15 PM-4:15 PM

Clinical and economic outcomes associated with sequential treatment in patients with advanced renal cell carcinoma (aRCC).

Meredith M. Regan, David F. McDermott, Michael B. Atkins, Apoorva Ambavane, Shuo Yang, Sumati Rao, M Dror Michaelson; 1 Beth Israel Deaconess Medical Center, Dana-Farber/Harvard Cancer Center, Boston, MA; Beth Israel Deaconess Medical Center, Dana-Farber/Harvard Cancer Center, Boston, MA; Georgetown Lombardi Comprehensive Cancer Center, Washington, DC; Evidera, Inc., Bethesda, MD; Bristol-Myers Squibb, Princeton, NJ; Massachusetts General Hospital Cancer Center, Boston, MA

Background: Immuno-oncology therapies (IOs) and tyrosine kinase inhibitors (TKIs) are recommended for the treatment of aRCC. As new drugs and combination regimens emerge, there is interest in gaining a deeper understanding of optimal treatment sequencing. We aimed to assess clinical and economic outcomes associated with treatment sequences for untreated aRCC patients with IMDC intermediate/poor risk. Methods: A discrete event simulation model was developed to estimate the total costs and survival (in life-years; LYs) over patients’ lifetimes when receiving sequential treatment with nivolumab + ipilimumab (N+I), sunitinib (SUN), pazopanib (PAZ), or cabozantinib (CAB) as first-line (1L) treatment, and nivolumab (NIVO), axitinib (AXI), PAZ, CAB, or lenvatinib + everolimus (LEN+EVE) as second line (2L). Efficacy inputs were derived from the CheckMate 214 trial and a network meta-analysis based on available literature. Safety and cost data were obtained from literature and publicly available sources. Results: N+I initiating sequences were estimated to provide longer survival in mean LYs and lower mean costs/LY versus sequences with 1L TKIs (table). The estimates of incremental cost-effectiveness ratio (ICER) for N+I initiating sequences with 2L TKI monotherapy were well below the willingness-to-pay threshold of $50,000. Using 2L LEN+EVE, compared with 2L monotherapies, provided an incremental survival gain but at costs/LY close to $100,000. Conclusions: Use of 1L N+I followed by TKI monotherapy is estimated to provide longer survival while being more cost-effective versus TKIs followed by IOs or sequences cycling TKIs, mainly driven by a longer time to 2L treatment and longer treatment-free survival with N+I. Clinical trials with head-to-head comparisons of treatment sequences would be necessary to validate the findings of the study.

1L 2L LYs* Total costs* Costs/LY* N+I CAB/AXI/PAZ 4.8–6.8 $264,722–$355,174 $48,832–$55,700 CAB NIVO/AXI/PAZ 4.6–5.1 $343,874–$359,143 $70,340–$77,566 SUN NIVO/CAB/AXI 3.1–3.7 $221,111–$277,802 $61,274–$74,794 PAZ NIVO/CAB/AXI 3.1–3.7 $221,597–$278,288 $61,405–$74,924 * Range for the sequences with different 2L options.

4567 Poster Session (Board #393), Mon, 1:15 PM-4:15 PM

Phase I/II study of axitinib (axi) and nivolumab (nivo) in patients with metastatic renal cell carcinoma (mRCC).

Matthew R. Zibelman, Daniel M. Geynisman, Ana M. Molina, Lois Malizzia, Karthik Devarajan, Diana Luong, Elizabeth Sullivan, Michael Anthony Carducci, Elizabeth R. Plimack; Fox Chase Cancer Center, Philadelphia, PA; Weill Cornell Medicine, New York, NY; Fox Chase Cancer Center, Temple Health, Philadelphia, PA; Sidney Kimmel Cancer Center At Johns Hopkins, Baltimore, MD

Background: Drug combinations targeting vascular endothelial growth factor (VEGF) and the programmed death one (PD-1) pathway have demonstrated encouraging efficacy in patients (pts) with mRCC. We are conducting a phase I/II trial combining the VEGF-targeting tyrosine kinase inhibitor (TKI) axi and the PD-1 inhibitor nivo in mRCC pts. Methods: The phase I portion of this investigator-initiated, multi-center trial enrolled pts with TKI-refractory mRCC (any line) and no prior anti-PD-1. It used a 3+3 design, with axi dosing starting at 3 mg orally BID, escalating to a target dose of5 mg BID. Nivo dosing was fixed at 240 mg IV Q2 weeks or 480 mg Q4 weeks. Eligible pts received axi alone for a one week induction, followed by combination dosing. There was a 4 week period to assess for dose limiting toxicities (DLTs). The primary endpoint for phase I was safety and establishment of a recommended phase II dose (RP2D) for axi. Safety and early efficacy data for the phase I cohort are presented. Results: Twelve pts received treatment during the phase I portion, with all evaluable for toxicity. Ten were 2nd line and 2 were 3rd line or beyond. One DLT was noted in the first cohort of three pts (grade 3 autoimmune disorder), thus that cohort was expanded. No further DLTs were reported at axi 3 mg BID, or at further expansion to 5 mg BID. Grade 3 adverse events (AEs) at least possibly related to one of the drugs included expected side effects HTN, hyperglycemia, and transaminitis. One patient received steroids for an immune related AE (grade 3 transaminitis) that resolved to grade 1 with steroids. No grade 4-5 AEs occurred. Axi at 5 mg BID was chosen as the RP2D. With 10 pts evaluable for efficacy, 4 pts had partial responses, 4 pts had stable disease, and 2 had progressive disease. Updated efficacy data will be presented. Conclusions: The phase I portion of this phase I/II trial has demonstrated expected safety and established 5 mg BID as the RP2D axi dose. This ongoing VEGF TKI/PD1 trial incorporating two drugs already approved for mRCC pts has shown encouraging signs of efficacy and has now expanded to include treatment na¨ıve pts in addition to TKI- refractory pts as part of phase II. Clinical trial information: NCT03172754.

4568 Poster Session (Board #394), Mon, 1:15 PM-4:15 PM

Association of human endogenous retrovirus (hERV) expression with clinical efficacy of PD-1 blockade in metastatic clear cell renal cell carcinoma (mccRCC).

Jean-Christophe Pignon, Opeyemi Jegede, Sachet A Shukla, David A. Braun, Christine Horak, Megan Wind-Rotolo, Yuko Ishii, Paul J. Catalano, Gordon J. Freeman, Rebecca B Jennings, Aguirre De Cubas, Kimryn Rathmell, Toni K. Choueiri, Michael B. Atkins, David F. McDermott, Sabina Signoretti; Brigham and Women’s Hospital, Boston, MA; Dana-Farber Cancer Institute, Boston, MA; Bristol-Myers Squibb, Princeton, NJ; Vanderbilt University, Nashville, TN; Vanderbilt University Medical Center, Nashville, TN; Dana-Farber Cancer Institute, Brigham and Women’s Hospital, and Harvard Medical School, Boston, MA; Georgetown Lombardi Comprehensive Cancer Center, Wash- ington, DC; Beth Israel Deaconess Medical Center, Boston, MA

Background: hERV levels positively correlate with tumor immune infiltrate and were recently shown to be associated with clinical benefit to PD-1/PD-L1 blockade in two small cohorts of patients (pts) with mccRCC (Smith C.C. et al and Panda A. et al; 2018). We tested whether hERV levels correlate with efficacy of nivolumab in a prospective phase II study of pts with mccRCC (Checkmate 010). Methods: Reverse transcribed RNA extracted from 99 FFPE pretreatment tumors were analyzed by RT-qPCR to assess levels of pan-ERVE4, pan-ERV3.2, hERV4700 GAG or ENV, and the reference genes 18S and HPRT1. Normalized hERV levels were transformed as categorical value (high or low) using population quartiles as cutoffs. For each cutoff, samples with non-quantifiable hERV levels for which the limit of quantification was above the tested cutoff could not be categorized and were excluded from analysis. Log rank test was used to test the association of hERV levels with PFS/irPFS (RECISTv1.1/irRECIST) at each cutoff using Holm-Bonferroni correction for Type I error control; adjusted P-values are reported. Fisher’s exact test was then used to explore the association with ORR/irORR (RECISTv1.1/irRECIST). Results: Among the hERV studied, only hERV4700 ENV was significantly associated with PFS/irPFS. At the 25th percentile cutoff, 45 pts had high levels of hERV4700 ENV and 24 pts had low levels of hERV4700 ENV. Median PFS and irPFS were significantly longer in the high-hERV4700 ENV group [7.0 (95% CI: 2.2 - 10.2) and 8.5 (95% CI: 4.2 - 14.1) months, respectively] versus the low- hERV4700 ENV group [2.6 (95% CI: 1.4 - 5.4) and 2.9 (95% CI: 1.4 - 5.7) months, respectively] (P = 0.010 for PFS and P = 0.028 for irPFS). At the same cutoff, ORR and irORR rates were significantly higher in the high-hERV4700 ENV group [35.6 (95% CI: 21.9 - 51.2) % for both ORR/irORR] versus the low-hERV4700 ENV group [12.5 (95% CI: 2.7 - 32.4) and 8.3 (95% CI: 1.0 - 27.0) %, respectively] (P = 0.036 for ORR and P = 0.012 for irORR). Conclusions: hERV4700 ENV levels may predict outcome on nivolumab in mccRCC. Validation of our results and correlation of hERV levels with immune markers in a controlled phase III trial (CheckMate 025) is ongoing.

4569 Poster Session (Board #395), Mon, 1:15 PM-4:15 PM

KEYNOTE-427 cohort B: First-line pembrolizumab (pembro) monotherapy for advanced non‒clear cell renal cell carcinoma (NCC-RCC).

Jae-Lyun Lee, Marek Ziobro, Rustem Gafanov, Vsevolod Borisovich Matveev, Cristina Suarez, Frede Donskov, Frederic Pouliot, Boris Yakovlevich Alekseev, Pawel J. Wiechno, Piotr Tomczak, Miguel Angel´ Climent, Sang Joon Shin, Rachel Kloss Silverman, Rodolfo F. Perini, Charles Schloss, David F. McDermott, Michael B. Atkins; Asan Medical Center and University of Ulsan College of Medicine, Seoul, South Korea; Centrum Onkologii-Instytut im. Marii Sklodowskiej, Cracow, Poland; Russian Scientific Center of Roentgenoradiology, Moscow, Russian Federation; N.N. Blokhin National Medical Research Center of Oncology, Moscow, Russian Federation; Vall d’Hebron University Hospital and Institute of Oncology, Universitat Autonoma de Barcelona, Barcelona, Spain; Aarhus University Hospital, Aarhus, Denmark; CHU de Quebec and Laval University, Quebec, QC, Canada; P. A. Herzen Moscow Oncology Research Institute, Ministry of Health of the Russian Federation, Moscow, Russian Federation; Maria Skłodowska-Curie Memorial Cancer Center, Warsaw, Poland; Clinical Hospital No. 1 of the Poznan University of Medical Sciences, Poznan, ´ Poland; Instituto Valenciano de Oncologı´a, Valencia, Spain; Yonsei University College of Medicine, Seoul, South Korea; Merck & Co., Inc., Kenilworth, NJ; Beth Israel Deaconess Medical Center, Boston, MA; Georgetown Lombardi Comprehensive Cancer Center, Washington, DC

Background: Efficacy of PD-1 inhibitors (or any therapy) in NCC-RCC has not been established. KEYNOTE- 427 (NCT02853344) is a single-arm, open-label, phase 2 study of pembro monotherapy in patients (pts) with advanced clear cell RCC (cohort A) and NCC-RCC (cohort B). Cohort B results are presented. Methods: 165 pts with histologically confirmed NCC-RCC, no prior systemic therapy, measurable disease (RECIST v1.1), and KPS $70% enrolled. Pts received pembro 200 mg IV Q3W for 35 cycles (~2 y) or until progressive disease (PD), unacceptable toxicity, or withdrawal. Primary end point: objective response rate (ORR) per RECIST v1.1 by blinded independent central review. Additional end points: duration of response (DOR), population description by sarcomatoid differentiation, histology and PD-L1 expression (combined positive score [CPS] $1 for PD-L1+). Results: Histology was confirmed by a central pathologist: papillary 72% (n = 118), chromophobe 13% (n = 21), unclassified 16% (n = 26); 62% were PD-L1+. At analysis, 49 pts had died and 3 had withdrawn. At median follow-up of 11.1 mo (range, 0.9-21.3), 56% of pts discontinued pembro due to PD or clinical progression. Overall ORR was 24.8% (95% CI, 18.5-32.2; 8 [4.8%] CR, 33 [20.0%] PR). Median DOR was not reached. For responding pts, 81.5% had a response $6 mo. 12-mo PFS and OS rates were 22.8% and 72.0%, respectively. ORR (95% CI) was 25.4% (17.9- 34.3) with papillary, 9.5% (1.2-30.4) with chromophobe, and 34.6% (17.2-55.7) with unclassified NCC- RCC; for responding pts, 82.1%, 50.0%, and 87.5% had a response $6 mo, respectively. Median DOR was not reached in any group. ORR (95% CI) was 44.7% (28.6-61.7) for pts with sarcomatoid differentiation (n = 38). ORR (95% CI) was 33.3% (24.3-43.4) and 10.3% (3.9-21.2) with CPS$1 and CPS , 1, respectively. Grade 3-5 treatment-related adverse events (TRAEs) occurred in 11% of pts. 6 pts died of AEs, 2 of TRAEs (pneumonia and cardiac arrest). Conclusions: Single-agent pembro showed encouraging antitumor activity in NCC-RCC, especially with papillary or unclassified histology. Safety profile of pembro was as expected. Updated data with additional follow-up will be presented. Clinical trial information: NCT02853344.

4570 Poster Session (Board #396), Mon, 1:15 PM-4:15 PM

First-line pembrolizumab (pembro) monotherapy in advanced clear cell renal cell carcinoma (ccRCC): Updated results for KEYNOTE-427 cohort A.

Scott S. Tykodi, Frede Donskov, Jae-Lyun Lee, Cezary Szczylik, Jahangeer Malik, Boris Yakovlevich Alekseev, James M.G. Larkin, Vsevolod Borisovich Matveev, Rustem Gafanov, Piotr Tomczak, Poul F. Geertsen, Pawel J. Wiechno, Sang Joon Shin, Frederic Pouliot, Teresa Alonso-Gordoa, Rachel Kloss Silverman, Rodolfo F. Perini, Charles Schloss, David F. McDermott, Michael B. Atkins; University of Washington and Fred Hutchinson Cancer Research Center, Seattle, WA; Aarhus University Hospital, Aarhus, Denmark; Asan Medical Center and University of Ulsan College of Medicine, Seoul, South Korea; Postgraduate Education Center-Oncology, Ecz Otwock, Poland; Edinburgh Cancer Centre, Western General Hospital, Edinburgh, United Kingdom; P. A. Herzen Moscow Oncology Research Institute, Ministry of Health of the Russian Federation, Moscow, Russian Federation; Institute of Cancer Research, London, United Kingdom; N.N. Blokhin Russian Cancer Research Center, Moscow, Russian Federation; Russian Scientific Center of Roentgenoradiology, Moscow, Russian Federation; Clinical Hospital No. 1 of the Poznan University of Medical Sciences, Poznan, ´ Poland; Herlev Hospital, University of Copenhagen, Herlev, Denmark; Maria Skłodowska-Curie Memorial Cancer Center, War- saw, Poland; Yonsei University College of Medicine, Seoul, South Korea; Universite ´ Laval, Quebec, QC, Canada; Hospital Universitario Ramon ´ y Cajal, Madrid, Spain; Merck & Co., Inc., Kenilworth, NJ; Beth Israel Deaconess Medical Center, Boston, MA; Georgetown Lombardi Comprehensive Cancer Center, Washington, DC

Background: KEYNOTE-427 (NCT02853344) is an open-label, single-arm, phase 2 study to evaluate efficacy and safety of first-line single-agent pembro, a programmed death 1 (PD-1) inhibitor, in patients (pts) with ccRCC (cohort A) and non–clear cell RCC (cohort B). Updated follow up from cohort A are presented. Methods: Pts with histologically confirmed ccRCC, measurable per RECIST v1.1, and no prior systemic therapy were eligible. Pts received pembro 200 mg IV Q3W for 2 y or until confirmed progressive disease, unacceptable toxicity, or pt decision to withdraw. Primary end point was objective response rate (ORR; per RECIST v1.1 blinded independent central review). Additional end points included duration of response (DOR), disease control rate (DCR), progression-free survival (PFS), overall survival (OS), and safety. Results: 110 pts enrolled; median (range) follow-up was 18.0 (2.5-22.7) mo. Median age (range) was 64 (29-87); 38.2%, 47.3%, and 14.5% had favorable, intermediate, and poor IMDC risk, respectively; 47.3% were PD-L1 positive. Confirmed ORR was 36.4% with 3 (2.7%) CRs and 37 (33.6%) PRs. Median DOR was not reached. Median PFS was 7.1 mo (95% CI, 5.6-11.0) and median OS was not reached. Results by IMDC category are outlined in the table. By PD-L1 status, confirmed ORR was 44.2% and 29.3% for positive and negative, respectively. By sarcomatoid differentiation (n=11), confirmed ORR was 63.6%. Treatment-related AEs occurred in 80.9%, with pruritus (28.2%) and fatigue (28.2%) most commonly reported. One pt died of treatment-related pneumonitis. Conclusions: With a median 18-months’ follow up, first-line pembro monotherapy continued to show antitumor activity in pts with ccRCC. Meaningful responses were observed in pts with intermediate/poor IMDC risk, PD-L1 positive and sarcomatoid differentiated tumors. Safety profile was comparable to previously reported. Clinical trial information: NCT02853344.

IMDC Total Favorable IMDC Intermediate/Poor N=110 N=42 N=68 ORR, % (95% CI) 36.4 (27.4-46.1) 31.0 (17.6-47.1) 39.7 (28.0-52.3) DCR, % (95% CI) 57.3 (47.5-66.7) 61.9 (45.6, 76.4) 54.4 (41.9-66.5) DOR, median (range), mo Not reached 13.7 (4.2-18.0) Not reached DOR ‡ 6 mo (responders), % 76.8 68.4 80.9

4571 Poster Session (Board #397), Mon, 1:15 PM-4:15 PM

Retrospective analysis of the safety and efficacy of immune checkpoint inhibitors (CPI) among patients (pts) with pre-existing autoimmune disorders (AD) and renal cell carcinoma (RCC) or urothelial carcinoma (UC).

Nieves Martinez Chanza, Wanling Xie, Majd Issa, Hannah Elizabeth Dzimitrowicz, Abhishek Tripathi, Benoit Beuselinck, Elaine Tat Lam, Yousef Zakharia, Rana R. McKay, Sumit Shah, Amir Mortazavi, Michael Roger Harrison, Toni K. Choueiri, Lauren Christine Harshman; Jules Bordet Institute, Brussels, Belgium; Dana-Farber Cancer Institute, Boston, MA; Ohio State University - James Cancer Hospital Solove Research Institute, Columbus, OH; Yale School of Medicine, New Haven, CT; University of Oklahoma, Stephenson Cancer Center, Oklahoma City, OK; University Hospitals Leuven, Leuven Cancer Institute, Leuven, Belgium; University of Colorado Cancer Center, Denver, CO; University of Iowa Holden Comprehensive Cancer Center, Iowa City, IA; University of California, San Diego, San Diego, CA; Stanford Univ, San Francisco, CA; Arthur G. James Cancer Hospital, Ohio State University Wexner Medical Center, Columbus, OH; Duke Cancer Institute, Duke University Medical Center, Durham, NC; Dana-Farber Cancer Institute, Brigham and Women’s Hospital, and Harvard Medical School, Boston, MA

Background: RCC and UC pts with clinically significant AD were generally excluded from CPI trials due to potential AD exacerbation. Thus, the safety and clinical activity of CPI in AD pts are not well characterized. Methods: We retrospectively collected data from RCC and UC pts with AD treated with CPI at 9 centers. Adverse events (AEs) were assessed using CTCAEv5 criteria. Objective response rate (ORR) was assessed by RECIST principles. Overall survival (OS) was estimated by Kaplan Meier. Results: Of 103 pts (57 RCC & 46 UC) with a broad spectrum of AD such as psoriasis (22%), thyroiditis (20%), rheumatoid arthritis (13%), polymyalgia rheumatica (8%), inflammatory bowel disease (6%), multiple sclerosis (3%) and lupus (3%), most received CPI as 1st or 2nd line (77% RCC, 93% UC) and anti-PD-1/ L1 monotherapy (65% RCC, 98% UC). At CPI start, 36 had clinically active AD (all grade 1-2) and 4(11%) requiring systemic immunosuppression. AD exacerbations occurred in 37% (n = 38); most frequent: arthritis (12% RCC, 24% UC), rash (11% RCC, 9% UC). New onset immune related (ir) AEs occurred in 36% (n = 37); most frequent: colitis (12% RCC, 4% UC), rash (11% RCC, 9% UC), hypothyroid (each 7%), nephritis (7% UC). Table details timing and management. Median followup was 12.5 (1-52) mos for RCC and 14.5 (1-53) mos for UC. Median time on CPI was 6 mos (1-36) RCC and 4 mos (0.5-40) UC. At data cutoff, 39 RCC & 36 UC had discontinued CPI; 16% for toxicity. ORR was 31% for RCC and 35% for UC. 1 yr-OS rate was 74% (95%CI 58-84) for RCC and 60% (95%CI 43-74) for UC. Conclusions: AD exacerbations or new irAEs occurred in 37% and 36% respectively and were generally manageable. Only a minority required CPI cessation due to toxicity.

AD EXACERBATIONS NEW irAEs RCC UC RCC UC N=57 N=46 N=57 N=46 N(%) 18 (32) 20 (43) 21 (37) 16 (35) 38 (37) 37 (36) Grade (n, %) 1/2 13(23) 14(30) 14(25) 9(20) 3/4 2(4) 4(9) 7(12) 5(11) Unkn 3(5) 2(4) - 2(4) Med Time CPI to AE, days (range) 69 (25-315) 32 (3-368) 48 (2-305) 120 (12-443) Systemic steroid (n, %) 5 (28) 12 (60) 11 (52) 9 (56) DMARDs (n, %) 1 (6) 2 (10) 0 (0) 0(0) CPI stopped for irAE (n, %) Temporary 3 (17) 5 (25) 8 (38) 6 (36) Permanent 2 (11) 4 (20) 3 (14) 3 (19)

4572 Poster Session (Board #398), Mon, 1:15 PM-4:15 PM

TIVO-3: Subgroup analysis of progression-free survival of tivozanib compared to sorafenib in subjects with refractory advanced renal cell carcinoma (RCC).

Camillo Porta, Elena Verzoni, Bernard Escudier, Sumanta K. Pal, Michael B. Atkins, Thomas E. Hutson, Michael N. Needle, David F. McDermott, Brian I. Rini; Department of Internal Medicine, University of Pavia and Division of Traslational Oncology, IRCCS Istituti Clinici Scientifici Maugeri, Pavia, Italy; Fondazione IRCCS Istituto Nazionale Tumori, Milan, Italy; Gustave Roussy, Villejuif, France; City of Hope National Medical Center, Duarte, CA; Georgetown Lombardi Comprehensive Cancer Center, Washington, DC; Baylor Sammons Cancer Center-Texas Oncology, Dallas, TX; Aveo Oncology, Cam- bridge, MA; Beth Israel Deaconess Medical Center, Boston, MA; Cleveland Clinic Taussig Cancer Institute, Cleveland, OH

Background: Tivozanib (T) is a biochemically potent and highly selective VEGF tyrosine kinase receptor inhibitor in clinical development in RCC. The TIVO-3 trial in 3rd and 4th line subjects with metastatic (m) RCC showed a median progression free survival (mPFS) of 5.6 months (mos) for T compared to 3.9 mos for sorafenib (S) (p = 0.017, HR = 0.73). Methods: Subjects with mRCC who failed 2 or 3 prior systemic regimens, one of which included a VEGFR TKI other than S or T, were stratified by IMDC risk category and type of prior therapy (two TKIs; TKI plus checkpoint; TKI + other) then randomized in a 1:1 ratio to T or S. The primary objective was to compare PFS by blinded independent radiological review. Pre-specified subgroups included prior treatment, IMDC prognostic group, and demographic characteristics. Results: Tivozanib demonstrated PFS benefit in all subgroups including men, women, patients over 65, and under 65. The hazard ratio was the same for patients enrolled in North America and the EU. Patients with ECOG performance status (PS) of 0 had a lower HR than patients with ECOG PS of 1. There was an increase in HR from IMDC favorable vs IMDC intermediate vs IMDC poor. PFS favored tivozanib in patients who had two prior lines of therapy, those treated with three prior lines, those with a prior checkpoint inhibitor, or with two prior VEGFR TKIs. Conclusions: Tivozanib improved PFS vs. sorafenib across several subgroups in TIVO-3. Patients with favorable and intermediate IMDC risk and ECOG PS 0 seemed to derive the most benefit. Patients treated with a prior checkpoint inhibitor or two VEGFR-TKIs had a longer PFS than patients treated on sorafenib. Clinical trial information: NCT02627963.

Subgroup HR Subgroup HR Male 0.64 IMDC Favorable 0.46 Female 0.72 IMDC Intermediate 0.69 Age £65 0.74 IMDC Poor 1.15 Age >65 0.59 Two prior VEGFR-TKIs 0.57 ECOG 0 0.54 Prior Checkpoint Ab 0.55 ECOG 1 0.87 3rd Line 0.69 North America 0.71 4th Line 0.64 EU 0.69 HR , 1 favors tivozanib

4573 Poster Session (Board #399), Mon, 1:15 PM-4:15 PM

Adjuvant axitinib dose modification in renal cell carcinoma (RCC): Analysis of the ATLAS study.

David I. Quinn, Tae Gyun Kwon, Masatoshi Eto, Dingwei Ye, Hideaki Miyake, Seong Il Seo, Seok-Soo Byun, Jae-Lyun Lee, Viraj A. Master, Chi-Fai Ng, Rolf Gerhard Linke, Brad Rosbrook, Mahgull Nazar Thakur, Enrique Grande, Marine Gross-Goupil, Chao-Hsiang Chang; USC Norris Comprehensive Cancer Center, Los Angeles, CA; Kyungpook National University Medical Center, Daegu, South Korea; Kyushu University Graduate School of Medical Sciences, Fukuoka, Japan; Department of Urology, Fudan University Shanghai Cancer Center, Shanghai, China; Kobe University Graduate School of Medicine, Kobe, Japan; Sungkyunkwan University Samsung Medical Center, Seoul, South Korea; Seoul National University Bundang Hospital, Seoul, South Korea; Asan Medical Center and University of Ulsan College of Medicine, Seoul, South Korea; Winship Cancer Institute of Emory University, Atlanta, GA; Chinese University of Hong Kong, Hong Kong, Hong Kong; SFJ Pharmaceuticals Group, Pleasanton, CA; Pfizer Inc., San Diego, CA; Pfizer, Sandwich, United Kingdom; MD Anderson Cancer Center Madrid, Madrid, Spain; Oncology Department, Centre Hospital- ier Universitaire Saint-Andre, Bordeaux, France; China Medical University Hospital, Taichung, Taiwan

Background: The ATLAS trial compared axitinib vs placebo in patients (pts) with locoregional RCC at risk of recurrence after nephrectomy. ATLAS was stopped due to futility at a pre-planned interim analysis; results showed no difference in disease-free survival (DFS) between the treatment arms. We explored whether pts treated longer with axitinib achieved better outcomes and the impact of axitinib dose reduction or increase on DFS – the trial required dose reduction for toxicity and allowed dose escalation in the event of no or minimal toxicity. Methods: Pts in ATLAS received a maximum of 3 y of study treatment. A landmark analysis was conducted comparing pts treated with axitinib #1yvs. 1 y. Pts who recurred or censored prior to 1 y were excluded. An analysis of daily dose characteristics was undertaken to compare patients whose dose was reduced and whose dose was increased to those with a stable dose of axitinib. Cox proportional hazard model was used for DFS analysis. Toxicity analysis using a 90-day landmark was also conducted. Results: Overall, 264 axitinib-treated pts were included in this analysis. Of these 42 pts were treated for #1 y and 222 pts for . 1 y. Pts remaining on axitinib . 1y vs #1 y did not have different DFS (hazard ratio [HR] = 0.572, 95% confidence interval [CI]: 0.247–1.327, P= 0.1874). Pts with dose reduction had longer DFS than those with a stable dose (HR = 0.458, CI: 0.305–0.687, P= 0.0001). Pts with dose increase did not have DFS different to stable dose pts (HR = 1.936, CI: 0.937–3.997, P= 0.0685). No difference in DFS in pts experiencing grade $2 adverse events (AEs) vs grade , 2 AEs within 90 d of start of treatment was observed (HR = 0.885, CI: 0.419–1.869, P= 0.7488). Pts experiencing grade $3 AEs within 90 d of start of treatment had shorter DFS compared to those that did not (HR = 1.643, CI: 0.963–2.801, P= 0.0653). Conclusions: DFS did not vary based on duration of axitinib treatment, however, pts with dose reductions had longer DFS vs pts with stable dose or dose increases. This difference suggests that there is a relation between axitinib exposure and DFS as seen with sunitinib in the advanced setting and pazopanib in the PROTECT study.

4574 Poster Session (Board #400), Mon, 1:15 PM-4:15 PM

Sequential treatment with pazopanib (PAZO) followed by nivolumab (NIVO) in patients with advanced or metastatic renal cell carcinoma (mRCC): Third interim results of the non- interventional study PAZOREAL.

Martin Boegemann, Jens Bedke, Martin Schostak, Christiane Hering-Schubert, Manfred Welslau, Jan Schleicher, Thomas Wolf, Alexander Petzoldt, Christian Doehn, Carsten Grullich, ¨ Viktor Grunwald, ¨ Thomas Steiner, Anja Rogler, Sven Hanson, Dunja Klein, Tanja Medinger, Peter-Juergen Goebell; University of Muenster Medical Center, Munster, ¨ Germany; Department of Urology, Universitatskli- ¨ nikum Tubingen, ¨ Tubingen, ¨ Germany; University Hospital Magdeburg, Magdeburg, Germany; St. Georg Klinikum Eisenach, Eisenach, Germany; Klinikum Aschaffenburg, Medizinische Klinik II, Aschaffen- burg, Germany; Katharinenhospital, Stuttgart, Germany; Hamatologie ¨ & Intern. Onkologie, Gemein- schaftspraxis Hamatologie ¨ - Onkologie, Dresden, Germany; GP Dres. Wilke/Wagner/Petzoldt, H¨amatologie/Onkologie, Furth, ¨ Germany; Urologikum Lubeck, ¨ Lubeck, ¨ Germany; University of Heidel- berg, Heidelberg, Germany; Innere Klinik und Klinik fur ¨ Urologie, Westdeutsches Tumorzentrum, Universitatsklinikum ¨ Essen, Essen, Germany; Department of Urology, HELIOS Klinikum Erfurt, Erfurt, Germany; Novartis Pharma GmbH, Nurnberg, ¨ Germany; on behalf of Novartis Pharma GmbH, Goettingen, Germany; Molecular Medicine, IOMEDICO AG, Freiburg, Germany; Statistics, IOMEDICO AG, Freiburg, Germany; University Hospital Erlangen, Erlangen, Germany

Background: Randomized clinical trials for the implementation of new therapies include only a selection of patients that are later treated with these new options. Thus, real-world evidence is urgently needed not only to monitor the translation of treatment approaches into routine practice but also to improve cancer treatment and survivorship care on a broader scale. Methods: PAZOREAL is a prospective, multicenter, non-interventional study to evaluate effectiveness [primary time on drug (TD)], tolerability, safety, and quality of life (QoL) in patients (pts) with mRCC, treated with 1st L PAZO followed by 2nd L NIVO or everolimus (EVE). Results: Between Dec. 2015 and Sep. 2017, 421 pts were enrolled and 402 pts started 1st L PAZO treatment (Tx), 127 and 5 pts received NIVO and EVE as 2nd L Tx, resp., 56 entered follow-up. At time of data-cut (08 Nov 2018) median TD was 6.6 months (95%CI 6.0-7.9) for 1st L PAZO and 4.1 months (95%CI 3.2-5.8) for 2nd L NIVO (all pts), 8.1 months (95% CI 6.6-9.5) for PAZO and 3.2 (2.7-6.5) for NIVO Tx for trial eligible pts (39.1% of 402 pts). Median TD for pts with or without prior nephrectomy was 7.6 vs 4.5 months, resp. The clinical benefit rate of 1st L PAZO was 58.2 % (95% CI 53.3-62.9) based on investigator assessment. Median OS of PAZO was 29.5 months (95% CI 23.6-NA) for all pts, 28.2 months (95% CI 22.2-NA) for NIVO in 2nd L. The most commonly reported AEs for PAZO Tx were diarrhea (35%), nausea (20.3%) and fatigue (17.5%). Most common PAZO- related grade 3/4 adverse events were hypertension (5%), hypertensive crisis (2.3%) and GGT increase (1.8%). QoL evaluated by EQ-5D-5L remained stable over different Tx lines. Conclusions: The interim results of the PAZOREAL study confirm a favorable overall survival in pts with mRCC treated with 1st L PAZO in a real-world setting, good benefit-risk profile of PAZO and sustained QoL monitored over several treatment lines. In Germany NIVO as 2nd L Tx is commonly applied after 1st LTxwithPAZO.

4575 Poster Session (Board #401), Mon, 1:15 PM-4:15 PM

Consistent efficacy of nivolumab plus ipilimumab across number of International Meta- static Database Consortium (IMDC) risk factors in CheckMate 214.

Bernard Escudier, Robert J. Motzer, Nizar M. Tannir, Camillo Porta, Yoshihiko Tomita, Sabeen Fatima Mekan, M. Brent McHenry, Brian I. Rini; Gustave Roussy, Villejuif, France; Memorial Sloan Kettering Cancer Center, New York, NY; University of Texas MD Anderson Cancer Center, Houston, TX; University of Pavia, Pavia, Italy; Niigata University, Niigata, Japan; Bristol-Myers Squibb, Princeton, NJ; Cleveland Clinic Taussig Cancer Institute, Cleveland, OH

Background: The IMDC prognostic model was created based on anti-VEGF treatments for advanced renal cell carcinoma (aRCC), and may not be relevant for immunotherapy. Methods: In a post hoc analysis of CheckMate 214, we compared efficacy with nivolumab + ipilimumab (N+I) vs sunitinib (S) by number of IMDC risk factors present. Results: Among 1096 intent-to-treat (ITT) patients (pts) in both arms, 21%, 61%, and 18% had favorable, intermediate (int), or poor-risk, respectively. Of int-risk pts, 58% had 1 factor (most commonly ,1 y from diagnosis [Dx], 52%; Hb , LLN, 27%; or KPS #70%, 10%); and 42% had 2 factors (of these pts, the most common combination of 2 factors was ,1 y from Dx and Hb , LLN, 59%). Of poor-risk pts, 58% had 3 factors, 29% had 4 factors, and few had 5 (10%) or 6 (3%) factors. Due to small numbers, pts with 4–6 factors were pooled. At 30-mo minimum follow-up, RECIST v1.1- confirmed objective response rate (ORR) and complete response (CR) rate per investigator remained consistently higher with N+I vs S across pts with 1–4 factors, although with S, ORR decreased with increasing number of factors (Table). Improved progression-free survival (PFS) and overall survival (OS) were seen with N+I over S irrespective of the number of factors present, including in pts with only 1 risk factor (Table). Conclusions: N+I showed consistent efficacy across number of IMDC risk factors, while S decreasedinefficacywithincreasingnumberoffactors.EfficacyofN+IwassuperiortoSinallint-and poor risk pts. These CheckMate 214 results along with prior CheckMate 025 data showing consistent OS benefit with N monotherapy across IMDC risk categories show a need for improved prognostic models for immunotherapies in aRCC. Clinical trial information: NCT02231749.

S, S, S, S, S, N+I,n=125 n=124 N+I, n=189 n=172 N+I, n =125 n=141 N+I, n=55 n=47 N+I, n=35 n=38 No. of IMDC risk 0011 2 2334–64–6 factors PFS, HR (95% CI) 1.23 0.77 0.83 0.44 0.86 (0.90–1.69) (0.60–0.99) (0.62–1.12) (0.27–0.73) (0.50–1.49) OS, HR (95% CI) 1.22 0.62 0.72 0.50 0.63 (0.73–2.04) (0.44–0.87) (0.51–01.02) (0.31–0.82) (0.36–1.10) ORR, % 39 50 42 38 42 26 44 17 40 16 CR rate, % 8410117113000

4576 Poster Session (Board #402), Mon, 1:15 PM-4:15 PM

Systemic therapy for advanced clear cell renal cell carcinoma (ccRCC) after progression on immune-oncology plus VEGF targeted therapy combinations (IO-VEGF).

Yasser Ged, Ruby Gupta, Cihan Duzgol, Natalie Shapnik, Almedina Redzematovic, Ritesh Kotecha, Martin Henner Voss, Darren R. Feldman, Oguz Akin, Sujata Patil, Robert J. Motzer, Brian I. Rini, Chung-Han Lee; Memorial Sloan Kettering Cancer Center, New York, NY; Cleveland Clinic Taussig Cancer Institute, Cleveland, OH

Background: IO-VEGF combinations are the backbone for current and future therapeutic developments in RCC with several IO-VEGF regimens reporting positive results in phase 3 trials. However, limited data exists on outcomes to subsequent therapy in patients progressing on IO-VEGF regimens. Methods: A retrospective analysis was performed on patients with ccRCC at the Memorial Sloan Kettering Cancer Center and Cleveland Clinic Cancer Institute who initiated systemic therapy post IO-VEGF regimens including combinations with VEGFR tyrosine kinase inhibitors (IO-TKI) and combinations with anti-VEGF monoclonal antibodies (IO-mAB). Patients treated on unreported clinical trials were excluded from the outcomes analysis. The primary objective was to evaluate the overall survival (OS) post IO-VEGF. The secondary objectives included objective response rate (ORR) and progression-free survival (PFS) according to RECIST v1.1. Kaplan-Meier methods and the log-rank test were used to evaluate time from start of systemic therapy post IO-VEGF to the event of interest. Results: Fifty-nine patients were treated after discontinuation of IO-VEGF regimens. Prior IO-VEGF regimens included IO-mAB (n = 35, 59%) and IO-TKI (n = 24, 41%). IMDC scores at the start of next line of therapy were favorable in 20%, intermediate in 60% and poor in 20%. Next line of therapy included VEGFR-TKI monotherapy (n = 45, 76%), VEGFR-TKI based combinations (n = 6, 10%), mTOR inhibitors (n = 3, 5%), and unreported clinical trials (n = 5, 9%). VEGFR-TKI containing regimens (n = 51) included cabozantinib (n = 22), axitinib (n = 17), lenvatinib/everolimus (n = 4), pazopanib (n = 4), and others (n = 4). Median OS was 24.5 months (95% CI 12-NE) with a 12 months OS rate of 63%. The ORR was 27% (14/51) and the median PFS was 6.8 months (95% CI 4.8-11). No difference in post IO-VEGF OS was observed when comparing IO- TKI vs IO-mAB (log rank p = 0.7). Conclusions: Post combination IO-VEGF treatment, most patients received VEGFR-TKIs. In this setting, VEGFR-TKIs continue to show clinical activity similar to historic experiences of patients post VEGF monotherapy.

4577 Poster Session (Board #403), Mon, 1:15 PM-4:15 PM

First-line (1L) immuno-oncology (IO) combination therapies in metastatic renal-cell carcinoma (mRCC): Results from the international mRCC database consortium (IMDC).

Shaan Dudani, Jeffrey Graham, Connor Wells, Ziad Bakouny, Sumanta K. Pal, Nazli Dizman, Frede Donskov, Camillo Porta, Guillermo de Velasco, Aaron Richard Hansen, Marco Adelmo James Iafolla, Benoit Beuselinck, Ulka N. Vaishampayan, Lori Wood, Elizabeth Chien Hern Liow, Flora Yan, Takeshi Yuasa, Georg A. Bjarnason, Toni K. Choueiri, Daniel Yick Chin Heng; Tom Baker Cancer Centre, University of Calgary, Calgary, AB, Canada; University of Manitoba, Winnipeg, MB, Canada; Dana-Farber Cancer Institute, Boston, MA; City of Hope National Medical Center, Duarte, CA; City of Hope Comprehensive Cancer Center, Duarte, CA; Department of Oncology, Aarhus University Hospital, Aarhus, Denmark; University of Pavia, Pavia, Italy; Department of Medical Oncology, University Hospital 12 de Octubre, i + 12, Madrid, Spain, Madrid, Spain; Princess Margaret Cancer Centre, Toronto, ON, Canada; University of Alberta, Edmonton, ON, Canada; University Hospitals Leuven, Leuven Cancer Institute, Leuven, Belgium; Wayne State University, Detroit, MI; Dalhousie University, Halifax, NS, Canada; Eastern Health, Melbourne, Australia; University of Texas Southwestern Medical Center, Dallas, TX; Cancer Institute Hospital of Japanese Foundation for Cancer Research, Tokyo, Japan; Sunnybrook Research Institute, Toronto, ON, Canada; Dana-Farber Cancer Institute, Brigham and Women’s Hospital, and Harvard Medical School, Boston, MA; University of Calgary, Calgary, AB, Canada

Background: In mRCC, ipilimumab and nivolumab (ipi-nivo) is a 1L treatment option. Recent data have also shown efficacy of 1L IO-VEGF (IOVE) inhibitor combinations. Comparative data between these two strategies are limited and the efficacy of subsequent therapies remains unknown. Methods: Using the IMDC dataset, patients (pts) treated with any 1L IOVE combination were compared to those treated with ipi-nivo. Multivariable Cox regression analysis was performed to control for imbalances in IMDC risk factors. Results: 188 pts received 1L IO combination therapy: 113 treated with IOVE combinations and 75 with ipi-nivo. Baseline characteristics and IMDC risk factors were comparable between groups. When comparing IOVE combinations vs ipi-nivo, 1L response rate (RR) was 33% vs 40% (p=0.39), time to treatment failure (TTF) was 14.3 (95% CI 9.2-16.1) vs 10.2 months (95% CI 6.7-15.1, p=0.23), and median overall survival (OS) was not reached (NR) (95% CI 22.3-NR) vs NR (95% CI 35.1-NR, p=0.17). When adjusted for IMDC risk factors, the hazard ratio (HR) for TTF was 0.71 (95% CI 0.46-1.12, p=0.14) and the HR for death was 1.74 (95% CI 0.82-3.68, p=0.14). Second-line (2L) treatments were varied. In pts receiving subsequent VEGF-based therapy, 2L RR was lower in the IOVE (n=20) versus ipi-nivo (n=20) cohort (15% vs 45%; p=0.04), though 2L TTF was not significantly different (3.7 vs 5.4 months, p=0.40, n=55). The use of IO post IOVE was uncommon and 3/5 pts had PD as best response; 2/5 had PR/SD but their 1L IOVE exposure was short at ,3months.Conclusions: There does not appear to be a superior 1L IO combination strategy in mRCC, as lOVE combinations and ipi-nivo have comparable 1L RR, TTF and OS. Most pts received VEGF-based therapy in the 2L. In this group, 2L RR was greater in pts who received ipi- nivo, though there was no difference in 2L TTF.

IO-VEGF (N=113) Ipi-Nivo (N=75) IMDC Risk Groups Favourable 29/92 (32%) 17/64 (27%) Intermediate 49/92 (53%) 33/64 (52%) Poor 14/92 (15%) 14/64 (22%) 2L Treatments Axitinib 5/34 2/30 Cabozantinib 9/34 2/30 Lenvatinib + Everolimus 2/34 0/30 Nivolumab 5/34 0/30 Pazopanib 2/34 9/30 Sunitinib 9/34 15/30 Other 2/34 2/30

*All non-significant (p . 0.05).

4578 Poster Session (Board #404), Mon, 1:15 PM-4:15 PM

Deferred cytoreductive nephrectomy among patients with newly diagnosed metastatic renal cell carcinoma treated initially with sunitinib.

Bimal Bhindi, Jeffrey Graham, Connor Wells, Frede Donskov, Felice Pasini, Jae-Lyun Lee, Naveen S. Basappa, Aaron Richard Hansen, Lori Wood, Christian K. Kollmannsberger, Ravindran Kanesvaran, Takeshi Yuasa, D. Scott Ernst, Sandy Srinivas, Brian I. Rini, I. Alex Bowman, Sumanta K. Pal, Toni K. Choueiri, Daniel Yick Chin Heng; University of Calgary, Calgary, AB, Canada; University of Manitoba, Winnipeg, MB, Canada; Tom Baker Cancer Centre, University of Calgary, Calgary, AB, Canada; Department of Oncology, Aarhus University Hospital, Aarhus, Denmark; Department of Oncology, S Maria della Misericordia Hospital, ULSS 18, Rovigo, Italy; Asan Medical Center and University of Ulsan College of Medicine, Seoul, South Korea; Cleveland Clinic, Cleveland, OH; Princess Margaret Cancer Centre, Toronto, ON, Canada; Dalhousie University, Halifax, NS, Canada; BC Cancer–Vancouver Cancer Centre, Vancouver, BC, Canada; National Cancer Centre Singapore, Singapore, Singapore; Cancer Institute Hospital of Japanese Foundation for Cancer Research, Tokyo, Japan; Division of Medical Oncology, Department of Oncology, London Regional Cancer Program, London Health Sciences Centre and University of Western Ontario,, London, ON, Canada; Stanford University Medical Center, Palo Alto, CA; Cleveland Clinic Taussig Cancer Institute, Cleveland, OH; UT Southwestern Medical Center, Dallas, TX; City of Hope National Medical Center, Duarte, CA; Dana- Farber Cancer Institute, Brigham and Women’s Hospital, and Harvard Medical School, Boston, MA

Background: While the CARMENA trial prompts more caution with upfront cytoreductive nephrectomy (CN) in patients with metastatic renal cell carcinoma (mRCC), 17% of patients in the sunitinib alone arm underwent deferred CN (dCN). Upfront systemic therapy has been proposed as a potential litmus test to identify patients suitable for CN, but data on outcomes are limited. We sought to characterize outcomes of dCN after upfront sunitinib relative to sunitinib alone. Methods: Patients with newly diagnosed mRCC receiving upfront sunitinib were identified from the International mRCC Database Consortium (IMDC) from 2006-2018. All CNs done after initial sunitinib were included, excluding CNs performed after sunitinib failure. The outcomes were overall survival (OS) and time to treatment failure (TTF). Kaplan Meier and multivariable Cox regression analyses were performed; dCN was analyzed as a time-varying covariate to account for immortal time bias. Results: The cohort included 708 patients of whom 53 (7.5%) underwent dCN at a median of 6.5 months (IQR 3.5,10.5) from diagnosis. Patients in the dCN group were more likely to have better Karnofsky performance status (KPS), intermediate IMDC risk, fewer metastatic sites, and response to upfront sunitinib (Table). There were 604 deaths during a median follow-up of 63 months. Median OS and TTF with dCN were 43.5 and 19.8 months vs. 9.4 and 4.3 months without, respectively. Upon multivariable analysis, dCN remained significantly associated with OS (HR 0.45, 95% CI 0.31-0.65; p , 0.001) but not TTF (HR 0.73, 95%CI 0.52-1.01; p = 0.056). Conclusions: Patients who received dCN were carefully selected and achieved long OS. With these benchmark outcomes, optimal selection criteria need to be identified and confirmation of the role of dCN in a clinical trial is warranted.

Variable dCN No CN p-value Age in years, median (IQR) 61 (53-67) 63 (55-70) 0.08 KPS < 80, n (%) 9 (15) 256 (38) , 0.001 Poor IMDC risk (vs intermediate), n (%) 15 (25) 333 (50) , 0.001 > 1 Metastatic site, n (%) 29 (55) 518 (79) , 0.001 Best response, n (%) Stable disease 16 (26) 225 (33) , 0.001 Partial/complete response 28 (46) 110 (16) Progressive disease 12 (20) 218 (32) Unknown 5 (8) 119 (18)

4579 Poster Session (Board #405), Mon, 1:15 PM-4:15 PM

Meta-analysis of randomized clinical trials (RCT) for the adjuvant treatment of renal cell carcinoma (RCC) with vascular endothelial growth factor receptor tyrosine-kinase inhibitors (VEGFR TKIs).

Daniel Vargas Almeida, Cleyton Z. Oliveira, Andrey Soares, RODRIGO COUTINHO MARIANO, Denis L Jardim, Diogo Assed Bastos, Fernando C. Maluf, Gustavo Werutsky, Fernando Sabino M Monteiro, Vinicius Carrera Souza, Guillermo de Velasco, Andre P. Fay, Andre Deeke Sasse, Fabio A. B. Schutz, Latin America Cooperative Oncology Group - Genitourinary (LACOG-GU); BP-A Beneficencia Portuguesa de S~ao Paulo, S~ao Paulo, Brazil; Hospital Israelita Albert Einstein and Centro Paulista de Oncologia, S~ao Paulo, Brazil; Hospital Sirio-Libanes, S~ao Paulo, Brazil; BP-A Beneficencia Portuguesa de S~ao Paulo and Hospital Israelita Albert Einstein, S~ao Paulo, Brazil; PUCRS School of Medicine and Hospital Sao Lucas da PUCRS, Porto Alegre, Brazil; Hospital Santa Lucia and Hospital Universitario de Brasilia, Brasilia, Brazil; Clinica AMO, Salvador, Brazil; 12 de Octubre Universitary Hospital, Madrid, Spain; SONHE - Sasse Oncology and Hematology Group, Campinas, Brazil

Background: Although surgery is the cornerstone in the treatment of most cases of localized kidney cancer, up to 30% of patients will experience disease recurrence at three years of follow-up. Three RCTs with VEGFR TKIs (ASSURE, PROTECT and ATLAS) failed to demonstrate improvement in disease-free survival (DFS). Only S-TRAC trial showed a significant improvement in DFS, and was approved by the Food and Drug Administration (FDA). However, the matter remains controversial among genitourinary oncologists. Therefore, we performed a meta-analysis to better evaluate the potential benefit of adjuvant VEGFR TKIs after curative intent nephrectomy. Methods: Eligible studies were searched in PubMed databases and limited to phase 3 RCT published from January 1996 to December 2018 of US FDA- approved VEGFR TKIs reporting on patients with RCC treated in the adjuvant setting. A summary hazard- ratio (HR) of disease-free survival (DFS) was calculated using 95% CIs by random-effects or fixed-effects models on the basis of the heterogeneity of included studies. Results: Four RCT (ASSURE, S-TRAC, PROTECT and ATLAS trials) were selected for analysis, including a total of 4,820 patients. A VEGFR TKI (sunitinib, sorafenib, pazopanib or axitinib) was administered in 2,737 patients, and 2,083 received placebo. The summary DFS HR for the overall population was 0.89 (95% CI 0.79-1.00; p = 0.06). When including the report of the ASSURE with the sub-group analysis with high-risk patient population (n = 3,946), the summary HR for DFS was 0.84 (95% CI 0.75-0.95, p = 0,0044). No evidence of publication bias was found. Conclusions: This is the first meta-analysis including the four RCTs in RCC adjuvant setting. This meta-analysis failed to demonstrate improvement in DFS for patients receiving a VEGFR TKI after curative intent nephrectomy. A modest benefit in DFS was observed in a selected sub-group of patients with higher risk for recurrence. There is no data regarding overall survival.

4580 Poster Session (Board #406), Mon, 1:15 PM-4:15 PM

Dynamic contrast-enhanced MRI to predict intratumoral molecular heterogeneity in clear cell renal cell carcinoma.

Durga Udayakumar, Ze Zhang, Durgesh Dwivedi, Yin Xi, Tao Wang, Payal Kapur, Qurratulain Yousuf, Allison Joyce, Asghar Hajibeiji, Michael Fulkerson, Alberto Diaz de Leon, Matthew Lewis, Ananth J Madhuranthakam, Jeffrey A Cadeddu, Aditya Bagrodia, Vitaly Margulis, James Brugarolas, Ivan Pedrosa; University of Texas Southwestern, Dallas, TX; UT Southwestern Medical Center, Dallas, TX; University of Texas Southwestern Medical Center, Dallas, TX; The University of Texas Southwestern Medical Center, Dallas, TX

Background: Mutation/inactivation of VHL in clear cell renal cell carcinoma (ccRCC) leads to upregulation of hypoxia inducible factors (HIFs) and angiogenesis. However, ccRCC is characterized by high intra-tumor heterogeneity (ITH). Random small samples such as those in percutaneous biopsies are likely limited for characterization of molecular alterations in heterogeneous ccRCCs. We hypothesize that whole-tumor dynamic contrast-enhanced (DCE) magnetic resonance imaging (MRI) is useful to noninvasively identify ITH in ccRCC. Methods: This IRB-approved, prospective, HIPAA-compliant study, included 62 ccRCCs. 3T DCE MRI was obtained prior to nephrectomy. Surgical specimens were sectioned to match MRI acquisition plane. 182 snap frozen samples (49 tumors) and adjacent uninvolved renal parenchyma (URP) were collected. RNA isolations, cDNA library preparation and mRNA sequencing were performed using standard protocols. RNA expression in 81 tumor samples were correlated (Spearman ranked) with % enhancement in a region of interest (ROI) drawn in the same location of the tumor on pre- and 3 different post-contrast DCE MRI phases. Gene function overrepresentation (OR) analyses were done on top positively and negatively correlated genes. False discovery rate (FDR) , 0.1 was considered statistically significant. Results: Principal component analysis of . 20,000 genes indicated distinct gene expression in tumors from URP. Unsupervised clustering showed enrichment of ccA samples (better prognosis) compared to ccB samples (worse prognosis). Importantly, ccA and ccB samples coexisted in 25% of tumors. DCE-MRI % enhancement correlated with expression of . 300 genes (p , 0.003, FDR , 0.1). OR analyses placed angiogenic pathway gene processes and the immune/inflammatory response processes within the top 5 positively- and negatively-correlated gene functions, respectively. HIF2 target genes correlated positively with % enhancement. Conclusions: DCE MRI detects specific molecular signatures and may help overcome the challenges of ITH in ccRCC. Further research is needed to explore the potential role of DCE MRI to assess response to antiangiogenic and immune-based therapies.

4581 Poster Discussion Session; Displayed in Poster Session (Board #346), Mon, 1:15 PM-4:15 PM, Discussed in Poster Discussion Session, Mon, 4:30 PM-6:00 PM

Comprehensive genomic profiling (CGP) of upper-tract (UTUC) and bladder (BUC) urothelial carcinoma reveals opportunities for therapeutic and biomarker development.

Andrea Necchi, Sumanta K. Pal, Jeffrey S. Ross, Russell Madison, Neeraj Agarwal, Guru Sonpavde, Monika Joshi, Ming Yin, Vincent A. Miller, Petros Grivas, Jon Chung, Siraj Mahamed Ali; Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy; City of Hope National Medical Center, Duarte, CA; SUNY Upstate Medical University, Syracuse, NY; Foundation Medicine, Inc., Cambridge, MA; University of Utah Huntsman Cancer Institute, Salt Lake City, UT; Dana-Farber Cancer Institute, Boston, MA; Penn State Health Milton S. Hershey Medical Center, Hershey, PA; The Ohio State University Comprehensive Cancer Center, Columbus, OH; University of Washington, School of Med- icine, Seattle, WA

Background: To understand the genomic landscape and inform the therapeutic development of UC, 2463 cases were analyzed by CGP for genomic alterations (GAs) and for genome wide signatures. Methods: 479 UTUC and 1984 BUC FFPE tissues (77%/70% primary [PT] and 23%/30% metastatic tumors [MT] from unmatched patients [pts]) underwent hybrid-capture based CGP to evaluate all classes of GA. Tumor mutational burden (TMB) was determined on 1.1 Mbp of sequenced DNA and microsatellite instability (MSI) was determined on 114 loci. Targetable GA and signatures were assessed according to the ESMO Scale for Clinical Actionability of molecular Targets (ESCAT). Results: 48% of UC pts overall harbored $1 tier 1-2 GA suggesting benefit from approved or investigational targeted therapies (TT). Additionally, 17% had a tier 3 GA that provides a strong rationale for clinical trial consideration. Non-FGFR3 kinase fusions were observed in 1% of pts (0.6% UTUC v 1.1% BUC), including BRAF/RAF1 fusions in 0.5%. BRAF mut/fusions were observed in 2% (49/2463) of cases and were mutually exclusive with FGFR3 GA (p = 0.002). In comparing UC from anatomic sites, there were no differences of TMB-H ($20 mut/mb)/MSI-H for PT and MT but UTUC was enriched for MSI-H (3.4%) relative to BUC (0.77%, p , 0.001, all TMB-H). Excluding MSI-H pts, UTUC has lower median TMB (4.35 mut/mb) than BUC (6.96 mut/mb). FGFR3 GA (26% v 19%, p , 0.05) and specifically short variants (SV) (20% v 13%) were enriched in UTUC vs BUC. HRAS SV were also enriched in UTUC vs BC (7.3% v 3.0%), attributed to an enrichment in renal pelvis UC (10.1%) v ureteral UC (1.8%, p , 0.05). RB1 GA were more frequent in BUC vs UTUC (21% v 7.8% p , 0.001). Conclusions: Against a background of 50% actionability in UC with opportunities for immunotherapy, TT, or combinations thereof, the UTUC cohort is enriched for FGFR3 and HRAS SV relative to BUC, with the observation of HRAS mutations predominantly in UC of the renal pelvis, that warrants further investigation into the distinct modes of oncogenesis for UC as stratified by anatomic origin. These results argue strongly for the routine incorporation of CGP prior to systemic therapy initiation in metastatic UC.

4582 Poster Session (Board #408), Mon, 1:15 PM-4:15 PM

Prognostic significance of CD73 expression in localized renal cell carcinoma (RCC).

Abhishek Tripathi, Wanling Xie, Abdallah Flaifel, John A. Steinharter, Emily Stern Gatof, Rebecca B Jennings, Gabrielle Bouchard, Justin Fleischer, Nieves M. Chanza, Connor Gray, Charlene Mantia, Xiao X. Wei, Marios Giannakis, Bradley Alexander McGregor, Toni K. Choueiri, David F. McDermott, Sabina Signoretti, Lauren Christine Harshman; University of Oklahoma, Ste- phenson Cancer Center, Oklahoma City, OK; Dana-Farber Cancer Institute, Boston, MA; Brigham and Women’s Hospital, Boston, MA; Beth Israel Deaconess Medical Center, Boston, MA; Dana–Farber Cancer Institute, Boston, MA; Dana Farber Cancer Institute, Boston, MA; Dana-Farber Cancer Institute, Brigham and Women’s Hospital, and Harvard Medical School, Boston, MA

Background: CD73 or ecto-5’-nucleotidase mediates the de-phosphorylation of adenosine monophos- phate to adenosine which promotes immunosuppression in the tumor microenvironment. Its prognostic significance in localized RCC has not been well characterized. Methods: We assessed CD73 protein expression using immunohistochemistry (Cell Signaling Technology, D7F9A, 1:25) on tissue microarrays (TMAs) containing tumor tissue from patients with pT1-4,N0-1, M0 RCC who underwent nephrectomy.CD73 expression was quantified using a combined score (CS: intensity x percentage of cells staining positive). CD73 positivity was defined as any CD73 expression on tumor cells (CS . 0). Patients were categorized into CD73 negative (CS = 0), low ( , median CS) and high ($ median CS) groups. Clinical data was collected retrospectively.Baseline patient characteristics were compared between CD73 negative and positive (high + low) groups using the Cochran-Armitage trend test. Multivariable Cox regression evaluated associations of CD73 expression with disease-free and overall survival (DFS, OS) after adjusting for other baseline prognostic variables. Results: Of the 112 patients included, clear cell was the most common histology (70%) followed by chromophobe (12.5%), and papillary (12%) RCC. CD73 expression (CS . 0) was noted in 22% (n = 25) of tumors and was associated with more advanced stage (T3-4/N+: 37.5% vs. 25%; p = 0.02) and a trend towards higher nuclear grade ($3: 48% vs. 35%; p = 0.07). Median follow-up from nephrectomy was 9.7 yrs. In multivariable analysis adjusting for nuclear grade ( , 3vs.$3) and stage (T1-2, N0 vs. T3-4/N+), tumors with high CD73 expression had significantly worse DFS and OS (Table). Conclusions: CD73 expression was found in 22% of RCC patients and was associated with adverse pathologic features and poor prognosis independent of tumor stage and histologic grade. Our results provide strong rationale for further investigation of the CD73/adenosine pathway as a therapeutic target in RCC.

DFS OS CD73 expression N 5-year Adjusted HR p 10-year OS Adjusted HR p DFS

Negative 87 75% Reference 64% Reference Low 12 50% 1.2 (0.5-3.0) 0.66 71% 0.8 (0.3-2.3) 0.65 High* 13 42% 2.7 (1.3-5.9) 0.01 22% 2.6 (1.2-5.8) 0.02 *$ median CS.

4583 Poster Session (Board #409), Mon, 1:15 PM-4:15 PM

Atezolizumab plus bevacizumab in non-clear cell renal cell carcinoma (NccRCC) and clear cell renal cell carcinoma with sarcomatoid differentiation (ccRCCsd): Updated results of activity and predictive biomarkers from a phase II study.

Ronan Flippot, Bradley Alexander McGregor, Abdallah Flaifel, Kathryn P. Gray, Sabina Signoretti, John A. Steinharter, Eliezer Mendel Van Allen, Meghara K. Walsh, Katy Gundy, Xiao X. Wei, Lauren Christine Harshman, Ulka N. Vaishampayan, Toni K. Choueiri, Rana R. McKay; Dana- Farber Cancer Institute, Boston, MA; Brigham and Women’s Hospital, Boston, MA; Dana-Farber Cancer Institute/Partners Cancer Care, Boston, MA; Dana Farber Cancer Institute, Boston, MA; Wayne State University, Detroit, MI; Dana-Farber Cancer Institute, Brigham and Women’s Hospital, and Harvard Medical School, Boston, MA

Background: NccRCC and ccRCCsd are aggressive tumors associated with poor prognosis and response to therapy. Combination strategies co-targeting VEGF signaling and inhibitory immune checkpoints are highly active in clear-cell renal cell carcinoma, but data is lacking in NccRCC and ccRCCsd. We conducted a multicenter, open-label, single arm phase II trial of atezolizumab plus bevacizumab in NccRCC and ccRCCsd. Methods: Patients with NccRCC and ccRCCsd ( . 20% sarcomatoid differentiation), and ECOG performance status of 0-2 were eligible. Prior systemic treatment was allowed with the exception of prior PD-1/PD-L1-directed therapy. Atezolizumab 1200mg and bevacizumab 15mg/kg were administered every 3 weeks until progression, unacceptable toxicity, or patient withdrawal. Primary endpoint was objective response rate (ORR) per RECIST 1.1. Exploratory biomarker analyses included PD-L1 expression on tumor (TC) and immune cells (IC), and spatial analysis of the immune infiltrate. Results: Sixty patients received at least 1 cycle of treatment, among whom 56 were evaluable for response (17 ccRCCsd and 39 NccRCC). ORR was 34% in the overall population, 53% in ccRCCsd and 26% in NccRCC. Median progression-free survival was 8.4 months (95%CI, 6.9-16.5). Baseline tumor tissue was available for 36 patients. TC PD-L1 expression $1% was associated with improved ORR (9/14, 64%) compared to patients with PD-L1 expression , 1% (4/20, 20%). Patients with TC PD-L1 expression $1% who experienced progressive disease as best response had shorter average distance between tumor cells and nearest neighboring immune cells at baseline. Further analysis of the immune tumor microenvironment on an expanded cohort, including IC PD-L1 expression and correlation with clinical outcomes, is ongoing and will be updated. Conclusions: The combination of atezolizumab plus bevacizumab is active in NccRCC and ccRCCsd. Candidate predictive biomarkers include PD-L1 expression in TC and topological analysis of the immune infiltrate. Clinical trial information: NCT02724878.

4584 Poster Session (Board #410), Mon, 1:15 PM-4:15 PM

Malignant pheochromocytoma (MP): A comprehensive genomic profiling (CGP) study.

Gennady Bratslavsky, Andrea Necchi, Oleg Shapiro, Joseph Jacob, Julia Andrea Elvin, Jo-Anne Vergilio, Jonathan Keith Killian, Nhu Ngo, Shakti Ramkissoon, Eric Allan Severson, Amanda Hemmerich, Siraj Mahamed Ali, Jon Chung, Prasanth Reddy, Vincent A. Miller, Brian Alexander, Alexa Betzig Schrock, Nick Liu, Jeffrey S. Ross; Urologic Oncology Branch, National Cancer Institute at the National Institutes of Health, Bethesda, MD; Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy; SUNY Upstate Medical University, Syracuse, NY; Foundation Medicine, Inc., Cambridge, MA; Foundation Medicine, Cambridge, MA; SUNY Upstate University Hospital, Syracuse, NY

Background: We CGP to characterize the genomic alterations (GA) in MP and to enable the search for potential therapy targets. Methods: From a series of 201,766 consecutive clinical cases, 44 cases of clinically advanced MP underwent CGP using a hybrid-capture based commercial assay to evaluate all classes of GA. Tumor mutational burden (TMB) was determined on 1.1 Mbp of sequenced DNA and reported as mutations/Mb and microsatellite instability (MSI) was determined on 114 loci. PD-L1 expression was determined by IHC (DAKO 22C3 antibody). Results: All patients had clinically advanced recurrent and/or metastatic disease. 23 patients were females and 21 patients were males. There were 34 (77%) of MP known to have originated in the adrenal gland and 10 (23%) of the MP were sequenced from metastatic site where the exact primary site was unknown. The primary tumor was used for sequencing in 14 (32%) of the MP cases and a non-primary tumor metastatic site (liver, lung, bone, soft tissue, lymph node, kidney, peritoneal cavity, and chest wall) in 30 (68%) of the MP cases. There were 2.3 GA/tumor. The most frequent un-targetable GA were ATRX (25%), TP53 (21%), SDHB (13%), CTNNB1 (7%), VHL (7%), and CDKN2A/2B, PIK3R2, NOTCH2 and MEN1 (all 5%). The most frequent potentially targetable GA included RET (9%), NF1 (9%) and FGFR1 (5%). PBRM1 GA were found in 2% of MAP. Germline mutations in known cancer predisposition genes were predicted in 8 (18%) of cases involving SDHB (5 cases) and BRCA1, MEN1, and MSH2 (1 case each). The genomic signatures of primary MP were not significantly different from that obtained from sequencing of metastatic site biopsies. 0 (0%) of 5 MP stained positively for PD-L1 expression. The mean TMB was 2.95 mutations/ Mb, the median TMB was 2.4 mutations/Mb. There 2 (5%) of MP with TMB $ 10 mutations/MB and 0 (0%) with TMB $ 20 mutations/Mb. 0 (0%) of 33 MP evaluated for MSI had a MSI-High status. Conclusions: Although the GA/tumor is relatively low for MP, CGP can reveal important potential therapy targets including RET, NF1 and FGFR1. MP do not reveal strong potential for immunotherapies with low TMB, absence of MSI-High status and low (2%) PBRM1 mutation frequencies.

4585 Poster Session (Board #411), Mon, 1:15 PM-4:15 PM

Metastatic penile (mPSCC), uterine cervical (mCSCC), and skin (mSSCC) squamous cell carcinomas: A comparative genomic profiling (CGP) study.

Joseph Jacob, Gennady Bratslavsky, Oleg Shapiro, Nick Liu, Julia Andrea Elvin, Jo-Anne Vergilio, Jonathan Keith Killian, Nhu Ngo, Douglas I. Lin, Shakti Ramkissoon, Eric Allan Severson, Siraj Mahamed Ali, Jon Chung, Alexa Betzig Schrock, Prasanth Reddy, Brian Alexander, Kimberly McGregor, Vincent A. Miller, Andrea Necchi, Jeffrey S. Ross; SUNY Upstate Medical University, Syracuse, NY; Urologic Oncology Branch, National Cancer Institute at the National Institutes of Health, Bethesda, MD; SUNY Upstate University Hospital, Syracuse, NY; Foundation Medicine, Inc., Cambridge, MA; Foundation Medicine, Cambridge, MA; Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy

Background: We compared the genomic alteration (GA) profiles of mCSCC, mPSCC and mSSCC to study impact on the targeted and immunotherapy options for the men and women suffering from these refractory cancers. Methods: 78 mPSCC, 604 mCSCC and 338 mSSCC underwent CGP using a hybrid-capture based assay. Tumor mutational burden (TMB) was determined on 1.1 Mbp of sequenced DNA and and microsatellite instability (MSI) was determined on 114 loci. Results: The HPV+/CDKN2A- status was significantly more frequent in the mCSCC than mPSCC or mSSCC (P , 0.0001). The GA/tumor frequencies were similar for mCSCC and mPSCC, but significantly higher in mSSCC (P , 0.0001). TP53 mutations were more common in mSSCC (UV light exposure) and mPSCC (likely due to loss of an original HPV+ status). TERT, NOTCH1 and FAT1 GA were more frequent in mPSCC and mSSCC whereas PIK3CA GA were more common in mCSCC. MTOR pathway targets (GA in STK11, FBXW7 and PTEN)were more common in mCSCC than mPSCC or mSSCC. MSI high status was extremely rare in all cases. TMB $ 10/20 mut/Mb frequencies were noteworthy in mPSCC and mCSCC but extraordinarily high in mSSCC. Examples of patients with mCSCC, mPSCC and mSSCC responding to targeted and immunotherapies will be presented. Conclusions: Although mCSCC, mPSCC and mSSCC share a variety of clinicopathologic features, the 3 tumor types can be sharply differentiated on CGP. The TP53, CDKN2A and HPV status of the tumor types differ significantly with HPV+ higher in the mCSCC group. There are opportunities for targeted therapies in all groups predominantly identified in the MTOR pathway. The relatively high numbers of cases with significantly elevated TMB in all 3 tumor types suggest that immunotherapies would be beneficial in a large subset of patients.

mPSCC (78 cases) mCSCC (604 cases) mSSCC (338 cases) GA/case 5.8 4.9 9.6 TP53 58% 11% 86% CDKN2A 47% 4% 55% TERT 45% 16% 45% FAT1 33% 7% 34% NOTCH1 22% 4% 39% PTCH1 3% 1% 9% PIK3CA 19% 37% 9% CD274 amp 2% 4% 2% PTEN 5% 14% 4% FBXW7 12% 14% 4% STK11 6% 12% 3% MDM2 1% 3% 3% MSI-High 0% 2% 2% TMB (median) 3.6 5.2 43.2 TMB ‡10/‡20 18%/8% 24%/7% 71%/63% HPV+ 29% 69% 5% TP53 Mutated and HPV+ 22% 7% 31% TP53 Mutated and HPV- 73% 93% 89%

TPS4586 Poster Session (Board #412a), Mon, 1:15 PM-4:15 PM

A randomized phase II study of atezolizumab plus recombinant human IL-7 (CYT107) or atezolizumab alone in patients with locally advanced or metastatic urothelial carcinoma (mUC): A Cancer Immunotherapy Trials Network Trial (CITN-14).

Evan Y. Yu, Steven Fling, Bob Salim, Randy F. Sweis, Gurkamal S. Chatta, Rohit K. Jain, Scott Edward Delacroix, Helen Moon, Andreanne Lacroix, Judith C Kaiser, Elad Sharon, Martin A. Cheever, Russell Pachynski; University of Washington, Seattle, WA; Fred Hutchinson Cancer Research Center, Seattle, WA; Axio Research, Seattle, WA; The University of Chicago, Chicago, IL; Roswell Park Comprehensive Cancer Center, Buffalo, NY; H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL; Louisiana State Univ School of Medcn and Stanley S. Scott Cancer Ctr, New Orleans, LA; SCPMG-Kaiser, Riverside, CA; National Cancer Institute, Bethesda, MD; Division of Oncology, Washington University Medical School, St. Louis, MO

Background: Atezolizumab is a regulatory-approved PD-L1 antagonistic antibody for the post-platinum mUC setting. Responses to atezolizumab are highly efficacious in a subset of patients, but suboptimal or absent in most patients. IL-7 (CYT107) is a homeostatic growth factor that promotes proliferation, differentiation, and survival of T lymphocytes. We recently demonstrated CYT107 significantly increases peripheral absolute lymphocyte and T cell numbers in metastatic castration-resistant prostate cancer patients when administered after sipuleucel-T. We hypothesize expansion of T cells by CYT107 may improve responses to PD-L1 inhibition. To test this hypothesis, we designed a randomized trial (NCT03513952) in mUC comparing the combination of CYT107 and atezolizumab to atezolizumab alone. Methods: Patients with ECOG PS #2 and RECIST v1.1 measurable mUC with disease recurrence after platinum-based chemotherapy are eligible. A safety run-in of 6 patients with staggered enrollment to atezolizumab plus CYT107 will be followed by randomization if ,2 patients experience a DLT. An additional 48 patients will then be randomized 1:1 to atezolizumab 1200 mg IV q3wks with or without CYT107 10 ug/kg IM qwk X 4, started 1 wk before atezolizumab. The primary endpoint is RECIST v1.1 ORR, with H0 14.8% and HA 45%, one-sided a 0.10; power 88%. An interim futility analysis will be performed after 24 randomized patients have their first disease assessment; cessation of the trial will occur if an O’Brien-Fleming futility boundary of ,-0.0063 in the ORR scale is observed between the experimental and control arm. Secondary endpoints include clinical benefit rate, PFS, DOR, OS, results by PD-L1 expression stratification, and safety. Exploratory correlative evaluations of tumor-infiltrating immune cells, interferon g expression, inflammatory gene expression, ELISPOT, T cell receptor sequencing, serum metabolite levels, gut microbiome, and PK analyses will be performed. Current state: Trial accrual has begun and is anticipated to complete around mid-2020. Clinical trial information: NCT03513952.

TPS4587 Poster Session (Board #412b), Mon, 1:15 PM-4:15 PM

A phase 3 randomized study of neoadjuvant chemotherapy (NAC) alone or in combination with nivolumab (NIVO) 6 BMS-986205 in cisplatin-eligible muscle invasive bladder cancer (MIBC).

Guru Sonpavde, Andrea Necchi, Shilpa Gupta, Gary D. Steinberg, Juergen E. Gschwend, Michiel Simon Van Der Heijden, Nathalie Garzon, Ayanbola Elegbe, Bradley Raybold, Danny Liaw, Mark Rutstein, Matt D. Galsky; Dana-Farber Cancer Institute, Boston, MA; Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy; Department of Medicine, Masonic Cancer Center, University of Minnesota, Minneapolis, MN; Department of Surgery, The University of Chicago Medicine, Chicago, IL; Department of Urology, Technical University of Munich, Munich, Germany; Department of Medical Oncology, The Netherlands Cancer Institute, Antoni van Leeuwenhoek Hospital, Amsterdam, Nether- lands; Bristol-Myers Squibb, Princeton, NJ; Department of Medicine, Icahn School of Medicine at Mount Sinai, Tisch Cancer Institute, New York, NY

Background: Immuno-oncology (IO) therapies have revolutionized the treatment (tx) of pts with advanced bladder cancer (advBC). For pts with cisplatin-eligible MIBC, the recommended tx regimen is cisplatin- based NAC prior to radical cystectomy (RC). However, since only » 30% of pts achieve a pathologic complete response (pCR) translating to improved long-term outcomes with approved regimens, new therapies are needed. PD-L1 expression is associated with aggressive BC and has been shown to increase in BC after NAC, suggesting that the PD-1/PD-L1 axis is a valid therapeutic target. Additionally, expression of indoleamine 2,3-dioxygenase (IDO) is higher in BC than in normal bladder tissue and is associated with advanced disease and poor clinical outcome. BMS-986205, a selective, potent, once- daily oral IDO1 inhibitor that works early in the IDO1 pathway to reduce kynurenine production, has demonstrated clinical activity in combination with NIVO (anti–PD-1) in pts with IO tx–naive advBC who had $ 1 prior line of therapy (ORR, 37%). Taken together, these data provide a rationale for investigating NAC + NIVO + BMS-986205 in MIBC. Here we describe a randomized, partially blinded, phase 3 study evaluating the efficacy and safety of NAC 6 NIVO 6 BMS-986205 followed by RC and continued IO tx in pts with MIBC (NCT03661320). Methods: Pts aged $ 18 years with previously untreated MIBC (clinical stage T2-T4a, N0, M0), creatinine clearance $ 50 mL/min, and predominant UC histology who are eligible for cisplatin-based NAC and RC will be enrolled. Pts with evidence of positive lymph node; metastatic BC; or prior systemic therapy, radiotherapy, or surgery for BC other than TURBT are not eligible. Pts will be randomized to receive NAC (gemcitabine/cisplatin; arm A), NAC + NIVO + oral placebo (arm B), or NAC + NIVO + BMS-986205 (arm C) followed by RC (all arms); arms B and C will receive continued IO tx. Primary endpoints include pCR after neoadjuvant tx and event-free survival (arms C vs A; arms B vs A). Secondary endpoints are overall survival and safety. This global study in 28 countries began accrual in Nov 2018 and has a target enrollment of 1200 pts. Clinical trial information: NCT03661320.

TPS4588 Poster Session (Board #413a), Mon, 1:15 PM-4:15 PM

DUTRENEO Trial: A phase II randomized trial of DUrvalumab and TREmelimumab as NEOadjuvant approach in muscle-invasive urothelial bladder cancer (MIBC) patients prospectively selected by immune signature scores.

Enrique Grande, Felix ´ Guerrero, Javier Puente, Isabel Galante, Ignacio Duran, Javier Fuentes, Teresa Alonso-Gordoa, Javier Burgos, Albert Font Pous, Oscar Buisan, Alvaro Pinto, Mario Alvarez- Maestro, Oscar Reig Torras, Pablo Maroto, Xavier Garcia del Muro, Patricia Galvan, Nuria Malats, Aleix Prat, Francisco X. Real, Daniel E. Castellano; MD Anderson Cancer Center Madrid, Madrid, Spain; Hospital Universitario 12 de Octubre, Madrid, Spain; Hospital Universitario Clı´nico San Carlos, Madrid, Spain; Hospital Clı´nico San Carlos, Madrid, Spain; Department of Medical Oncology, Hospital Universitario Marques ´ de Valdecilla, Santander, Spain; Hospital Universitario Marques ´ de Valdecilla, Cantabria, Spain; Hospital Universitario Ramon ´ y Cajal, Madrid, Spain; Hospital Ramon ´ y Cajal, Madrid, Spain; Institut Catalad ` ’Oncologia, Hospital Universitari Germans Trias i Pujol (HUGTiP), Badalona, Spain; Hospital Germans Trias i Pujol, Urology Department, Barcelona, Spain; Hospital Universitario La Paz, Madrid, Spain; Hospital Clinic i Provincial, Barcelona, Spain; Hospital de la Santa Creu i Sant Pau, Barcelona, Spain; Instituto Catalan de Oncologia de Hospitalet, Barcelona, Spain; IDIBAPS, Hospital Clinic, Barcelona, Spain; Spanish National Cancer Research Centre, Madrid, Spain; Department of Medical Oncology, Hospital Clinic, Barcelona, Spain; Centro Nacional de Investiga- ciones Oncologicas, ´ Madrid, Spain; Medical Oncology Service, Hospital Universitario 12 de Octubre, Madrid, Spain

Background: Cisplatin-based neoadjuvant chemotherapy (CT) followed by cystectomy improves overall survival in patients (pts) with MIBC. Immune checkpoint inhibitors as single agents are approved in pts with advanced UC. Combination of both programmed cell death ligand-1 (PD-L1) and cytotoxic T-lymphocyte-associated antigen-4 (CTLA-4) checkpoints might be synergistic. Durvalumab (DU) is a selective, engineered, human IgG1 monoclonal antibody (mAb) that blocks PD-L1 binding to PD-1 and CD80. Tremelimumab (TRE) is an anti-CTLA-4 mAb of the lgG2 isotype. Pembrolizumab and atezolizumab, have shown a promising activity (including significant pathologic complete responses (pT0)) in MIBC pts candidate to cistectomy, with better results in those pts with PD-L1 overexpression. It is expected that the dual targeting of the immune system with an anti-PDL1 + antiCTLA4 such as DU and TRE in the neoadjuvant setting may improve these outcomes. In addition, the most precise selection of pts according to a molecular INF-gamma signature is intended to increase the efficacy. Methods: This is a prospective and randomized phase II, open-label study conducted in urothelial MIBC pts diagnosed of T2-T4 and/or N+ candidates to cystectomy, ECOG 0-1 and adequate organ function. Pts will be treated according to the score of a pro-inflammatory signature (PIS) determined with Nanostring technology. Pts with a low PIS will receive standard cisplatin-based neoadjuvant therapy (22 pt). Pts with a high PIS will be randomized 1:1 to receive cisplatin-based neoadjuvant therapy (22 pt) or DU 1500 mg + TRE 75 mg every 4 weeks x 3 cycles (22 pt). If more than 8 responses (pT0) are observed in first 22 pts included in DU+TRE arm, 24 additional pts will be recruited in this arm. Primary objective is to assess the antitumor activity of DU+TRE measured as pT0 rate in pts with a positive PIS. Disease free survival and safety profile will be also evaluated. Tissue, plasma and urine samples will be collected for translational studies. Clinical trial information: NCT03472274.

TPS4589 Poster Session (Board #413b), Mon, 1:15 PM-4:15 PM

ALBAN: An open label, randomized, phase III trial, evaluating efficacy of atezolizumab in addition to one year BCG (bacillus Calmette-Guerin) bladder instillation in BCG-naive patients with high-risk nonmuscle invasive bladder cancer (AFU-GETUG 37).

Morgan Roupret, Yann Neuzillet, Aurelie Bertaut, Geraldine Pignot, Nadine Houede, Stephane Champiat, Soazig Nenan-Le ´ Ficher, Maggy Chausson, Yohann Loriot; Urology University Paris 6, Paris, France; Versailles-Saint-Quentin-en-Yvelines University, Urology Department, Foch Hospital, Suresnes, France; Georges-François Leclerc Cancer Center, Dijon, France; Department of Surgical Oncology, institut Paoli- Calmettes,, Marseille, France; Montpellier University, Departement ´ d’oncologie Médicale, CHU Care- meau,, Nıˆmes, France; Institut Gustave Roussy, Nantes, France; UNICANCER, Kremlin-Bicetre,ˆ France; Unicancer, Kremlin-Bicetre,ˆ France; Gustave Roussy, INSERM U981, Université Paris-Sud, Université Paris-Saclay, Villejuif, France

Background: The majority of patients with bladder cancer are diagnosed with non-muscle invasive bladder cancer (NMIBC). Trans-urethral resection of bladder tumor (TURBT) followed by Bacillus Calmette- Guerin (BCG) therapy is the standard of care for most patients with high-risk NMIBC, although clinical outcomes remain poor. Data from preclinical and clinical studies suggest that programmed death-ligand 1 PD-L1 inhibition may enhance antitumor activity of BCG therapy. Atezolizumab, an anti–PD-L1 monoclonal antibody, is approved in the United States and the European Union as monotherapy when patients are ineligible for cisplatin or after prior platinum therapy. This approval was based on the results from phase II and III IMvigor210/211 trials. We have designed ALBAN, a phase III trial to evaluate the efficacy and safety of atezolizumab given in combination with BCG versus BCG alone in patients with high-risk NMIBC. Methods: In 30 centers in France, ALBAN (NCT03799835) is enrolling patients with histologically documented NMIBC who have not received prior BCG therapy. Eligible patients have to have high-risk features (T1 staging, high grade, or in situ carcinoma), ECOG PS 0-2, and a tissue sample for PD-L1 testing (by Ventana SP142 immunohistochemistry assay). Patients are randomized 1:1 to: (Arm A) BCG alone (six weekly instillations of BCG, followed by three weekly maintenance instillations at 3, 6, 12 months), or (Arm B) atezolizumab (1200 mg; IV; q3w for up to 1 year) combined with BCG given as in arm A. Treatment with atezolizumab and BCG are continued in the absence of unacceptable toxicity, for 1 year. The primary endpoint is recurrence-free survival in the intent-to-treat population. Secondary efficacy endpoints include overall survival, PFS, complete response, disease worsening, and quality of life. Safety, biomarkers, and other exploratory endpoints will also be evaluated. Enrollment began in December 2018 with a target of 614 pts. Clinical trial information: NCT03799835.

TPS4590 Poster Session (Board #414a), Mon, 1:15 PM-4:15 PM

A phase III, randomized, open label, multicenter, global study of first-line (1L) durvalumab in combination with standard of care (SOC) chemotherapy and durvalumab in combination with tremelimumab and SOC chemotherapy versus SOC chemotherapy alone in patients with unresectable locally advanced or metastatic urothelial cancer (UC).

Matt D. Galsky, Andrea Necchi, Srikala S. Sridhar, Osamu Ogawa, Natasha Angra, Stephan Hois, Philip He, Dana C. Ghiorghiu, Joaquim Bellmunt; Department of Medicine, Icahn School of Medicine at Mount Sinai, Tisch Cancer Institute, New York, NY; Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy; Princess Margaret Hospital, Toronto,ON,Canada;KyotoUniversity,Kyoto, Japan; AstraZeneca, Washington, MD; AstraZeneca, Cambridge, United Kingdom; AstraZeneca, Gaithersburg, MD; IMIM-Hospital del Mar Medical Research Institute, Barcelona, Spain

Background: Despite high response rates to 1L SoC for locally advanced or metastatic UC chemotherapy (gemcitabine + cisplatin or gemcitabine + carboplatin for patients who are cisplatin-ineligible [poor performance status, impaired renal function, comorbidities]), most patients experience disease progression. Novel strategies such as combining chemotherapy and immunotherapy offer hope for improving clinical outcomes. Durvalumab is a selective, high affinity, engineered human IgG1 mAb that blocks PD-L1 binding to PD-1 and CD80. Tremelimumab is a human IgG2 mAb directed against CTLA-4. The mechanisms of action of PD-1 and CTLA-4 are nonredundant, so targeting both checkpoint pathways may have additive or synergistic efficacy compared with monotherapy. Studies of platinum-based chemotherapy combined with checkpoint blockade in other tumor types have yielded improved efficacy with acceptable safety and support exploration of this approach for 1L locally advanced or metastatic UC. Methods: NILE (NCT03682068) is a randomized, open-label, multicenter, global trial that will enroll approximately 1265 patients with histologically or cytologically documented, unresectable, locally advanced, or metastatic transitional cell carcinoma of the urothelium. Patients will be randomized (1:1:1) to durvalumab + SoC chemotherapy, durvalumab + tremelimumab + SoC chemotherapy, or SoC chemotherapy as 1L therapy. Primary endpoints are progression-free survival using blinded independent central review assessments per RECIST 1.1 and overall survival (OS). Secondary endpoints include objective response rate, OS at 24 months, proportion of patients alive and progression free at 12 months, duration of response, disease control rate, time from randomization to second progression, and health- related quality of life. Safety, pharmacokinetics, immunogenicity, and biomarkers will also be assessed. The study opened for enrollment in September 2018. 2019 American Society of Clinical Oncology, Inc. Reused with permission. This abstract was accepted and previously presented at the 2019 Genitourinary Cancers Symposium. Clinical trial information: NCT03682068.

TPS4591 Poster Session (Board #414b), Mon, 1:15 PM-4:15 PM

Bacillus Calmette-Guerin (BCG) with or without pembrolizumab (pembro) for high-risk (HR) nonmuscle invasive bladder cancer (NMIBC) that is persistent or recurrent following BCG induction: Phase III KEYNOTE-676 study.

Ashish M. Kamat, Neal D. Shore, Noah M. Hahn, Shaheen Riadh Alanee, Hiroyuki Nishiyama, Shahrokh F. Shariat, Kijoeng Nam, Ekta Kapadia, Tara L. Frenkl, Gary D. Steinberg; The University of Texas MD Anderson Cancer Center, Houston, TX; Carolina Urologic Research Center, Myrtle Beach, SC; Johns Hopkins University School of Medicine, Baltimore, MD; Henry Ford Hospital, Detroit, MI; Department of Urology, University of Tsukuba, Tsukuba, Japan; Medical University of Vienna, Vienna, Austria; Merck & Co., Inc., Kenilworth, NJ; The University of Chicago Medical Center, Chicago, IL

Background: Intravesical instillation of BCG is standard of care for patients (pts) with HR NMIBC. However, many pts have persistent/recurrent HR NMIBC after BCG induction and are at increased risk for progression to muscle-invasive disease. Interim data from the phase 2 KEYNOTE-057 study has shown that the PD-1 inhibitor pembro had promising efficacy in HR NMIBC as monotherapy. KEYNOTE-676 (NCT03711032) is a randomized, comparator-controlled, phase 3 trial to evaluate efficacy and safety of pembro plus BCG in pts with persistent/recurrent HR NMIBC after BCG induction therapy. Methods: Pts are randomly assigned 1:1 to continue BCG therapy alone or receive BCG plus pembro 200 mg every 3 weeks. Treatment is stratified by carcinoma in situ (CIS) histology (presence/ absence), PD-L1 combined positive score ($10/˂10), and timing of NMIBC persistence/recurrence (0 to #6, .6to#12, or .12 to #24 mo). Pts are eligible if they are $18 years of age with histologically confirmed persistent/recurrent HR NMIBC of the bladder after adequate BCG induction therapy, have undergone cystoscopy/transurethral resection of bladder tumor within 12 weeks before randomization, have no concurrent extravesical disease, and have an ECOG PS score of 0-2. Responses are assessed by cystoscopy and blinded independent central review of urine cytology and biopsy (as applicable) every 12 weeks for years 1-2 and every 24 weeks for years 3-5 and by computed tomography urography every 18 months through year 5. Treatment will continue with pembro for up to 2 years and BCG for 3 years or until confirmed HR NMIBC persistence, recurrence, or disease progression, unacceptable toxicity, or pt/physician decision to withdraw. Primary end point is complete response rate in pts with CIS. Secondary end points are event-free survival (EFS), recurrence-free survival, overall survival, disease- specific survival, time to cystectomy, 12-month EFS rate in all pts, duration of response (DOR), 12- month DOR rate in pts with CR and safety and tolerability. Recruitment began in November 2018 and will continue until ~550 pts are enrolled. Clinical trial information: NCT03711032.

TPS4592 Poster Session (Board #415a), Mon, 1:15 PM-4:15 PM

A phase III, randomized, open label, multicenter, global study of efficacy and safety of durvalumab in combination with gemcitabine+cisplatin (G+C) for neoadjuvant treatment followed by durvalumab alone for adjuvant treatment in muscle-invasive bladder cancer (MIBC) (NIAGARA).

Thomas Powles, Joshua J Meeks, Matt D. Galsky, Michiel Simon Van Der Heijden, Hiroyuki Nishiyama, Hikmat Al-Ahmadie, Ashok Kumar Gupta, Jiabu Ye, Sarah E. Donegan, Dana C. Ghiorghiu, Salvatore Ferro, James WF Catto; Barts Cancer Institute, Queen Mary University of London, Royal Free NHS Trust, London, United Kingdom; Northwestern University, Department of Urology, Feinberg School of Medicine, Chicago, IL; Department of Medicine, Icahn School of Medicine at Mount Sinai, Tisch Cancer Institute, New York, NY; Netherlands Cancer Institute, Amsterdam, Netherlands; De- partment of Urology, University of Tsukuba, Tsukuba, Japan; Memorial Sloan Kettering Cancer Center, New York, NY; AstraZeneca, Gaithersburg, MD; AstraZeneca, Gaitherburg, MD; AstraZeneca, Cam- bridge, United Kingdom; Department of Oncology & Metabolism, The Medical School, Sheffield, United Kingdom

Background: Management of MIBC includes both surgery and systemic therapy. Neoadjuvant, cisplatin- based combination chemotherapy has demonstrated improved pathologic complete response (pCR), event-free survival (EFS), and OS compared with radical cystectomy alone. Many patients still develop recurrence, including progression to metastasis. Novel strategies such as combining chemotherapy and immunotherapy in a neoadjuvant setting and consolidating response post cystectomy in the adjuvant setting may improve clinical outcomes. Durvalumab is a selective, high affinity, engineered human IgG1 mAb that blocks PD-L1 binding to PD-1 and CD80. PD-L1 inhibition with durvalumab, in combination with a standard neoadjuvant regimen (G+C), may improve immune-mediated antitumor response and increase the rates of pathologic responses and long-term survival. Methods: NIAGARA (NCT03732677) is a Phase 3, randomized, open-label, multicenter, global study that will enroll ~1050 patients randomized (1:1) to durvalumab and G+C combination (Arm 1) or G+C (Arm 2) as neoadjuvant chemotherapy prior to radical cystectomy. Following radical cystectomy and during adjuvant therapy, patients in Arm 1 will receive durvalumab monotherapy for 8 cycles (8 months); patients in Arm 2 will receive no adjuvant treatment. Patients with resectable MIBC (clinical stage T2N0M0-T4aN0M0) with transitional cell histology planning to undergo a radical cystectomy will be included. Primary endpoints are pCR rates at time of cystectomy following neoadjuvant treatment and EFS. Secondary and exploratory endpoints include proportion of patients who achieve pathologic response ,P2 (stages Pa, P1, and carcinoma in situ) at time of cystectomy following neoadjuvant treatment, EFS at 24 months, metastasis-free survival, proportion of patients who undergo cystectomy, and OS at 5 years. Safety, patient-reported outcomes, pharmacokinetics, immunogenicity, and biomarkers will also be assessed. Enrollment opened in Dec 2018. Clinical trial information: NCT03732677.

TPS4593 Poster Session (Board #415b), Mon, 1:15 PM-4:15 PM

EV-103: Enfortumab vedotin plus pembrolizumab and/or chemotherapy for locally advanced or metastatic urothelial cancer.

Christopher J. Hoimes, Jonathan E. Rosenberg, Daniel Peter Petrylak, Anne-Sophie Carret, Amal Melhem-Bertrandt, Thomas W. Flaig; University Hospitals Cleveland Medical Center/Case Western Reserve University, Cleveland, OH; Memorial Sloan Kettering Cancer Center, New York, NY; Yale School of Medicine, New Haven, CT; Seattle Genetics, Inc., Bothell, WA; Astellas Pharma US, Inc, Northbrook, IL; Division of Medical Oncology, School of Medicine, University of Colorado, Aurora, CO

Background: PD-(L)1-targeted immune checkpoint inhibitors (CPIs), such as pembrolizumab (P), are approved for patients (pts) with locally advanced or metastatic urothelial cancer (la/mUC) who progress after platinum, are ineligible for first-line (1L) cisplatin (PD-L1 positive), or are ineligible for 1L platinum. The ORR for CPI in all treated pts is ~25%, and a combination approach may provide additional benefit. Enfortumab vedotin (EV), an investigational antibody-drug conjugate, delivers the microtubule-disrupting agent monomethyl auristatin E to cells expressing Nectin-4, found in 97% of mUC pt samples (Petrylak ASCO 2017). In a phase 1 study (NCT02091999), EV (1.25 mg/kg) was generally well tolerated with a confirmed ORR of 43% in 112 mUC pts (Rosenberg ASCO-GU 2019). These encouraging results, with the potential for an enhanced immune response, suggest that EV+P may improve response rates and extend response durability. Methods: EV-103, a phase 1b trial for non–resectable la/mUC pts with no prior CPI, added an additional 4 cohorts (Parts 2 and 3) in an Oct 2018 amendment. It is now expected to enroll ~159 pts. Dose escalation pts (EV+P, 1L or 2L) must be ineligible for 1L cisplatin-based chemotherapy or have disease progression during/following treatment with $1platinum-containing regimen. Dose expansion (Part 1) will evaluate EV+P in 1L (Cohort A) and 2L settings (Optional Cohort B). Part 2 will evaluate 1L EV+cisplatin (Cohort D), 1L EV+carboplatin (Cohort E), and 1L or 2L EV+gemcitabine (Optional Cohort F). Part 3 (Cohort G) will evaluate 1L EV+P+cisplatin or carboplatin, depending on pts cisplatin-eligibility. In all cohorts, pts receive EV on Days 1 and 8 of each 3-week cycle. In combination therapies, pts receive P, cisplatin, and carboplatin on Day 1 or gemcitabine on Days 1 and 8 of each cycle. The primary objective is to assess the safety/tolerability of EV+P and/or chemotherapy. Secondary objectives are establishing EV recommended dose for combination therapies, assessing ORR per RECIST for all cohorts and per iRECIST for therapies with pembrolizumab, as well as assessing disease control rate, duration of response, PFS, OS, PK, and biomarkers. The study opened Oct 2017. Clinical trial information: NCT03288545.

TPS4594 Poster Session (Board #416a), Mon, 1:15 PM-4:15 PM

PrE0807 phase Ib feasibility trial of neoadjuvant nivolumab (N)/lirilumab (L) in cisplatin- ineligible muscle-invasive bladder cancer (BC).

Petros Grivas, Maneka Puligandla, Suzanne Cole, Kevin Dale Courtney, Robert Dreicer, Benjamin Adam Gartrell, Jeremy Paul Cetnar, Marc Dall’era, Matt D. Galsky, Rohit K. Jain, Benjamin Louis Maughan, Neeraj Agarwal, Vadim S Koshkin, Noah M. Hahn, Michael Anthony Carducci; University of Washington, School of Medicine, Seattle, WA; Dana-Farber Cancer Institute, Boston, MA; The University of Texas Southwest- ern Medical Center, Dallas, TX; UT Southwestern Medical Center, Dallas, TX; University of Virginia, Charlottesville, VA; Montefiore Medical Center/Albert Einstein College of Medicine, Bronx, NY; Oregon Health and Science University, Portland, OR; University of California Davis Comprehensive Cancer Center, Sacramento, CA; Department of Medicine, Icahn School of Medicine at Mount Sinai, Tisch Cancer Institute, New York, NY; Moffitt Cancer Center, Tampa, FL; Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins University, Baltimore, MD; University of Utah Huntsman Cancer Institute, Salt Lake City, UT; University of California San Francisco, San Francisco, CA; Johns Hopkins University School of Medicine, Baltimore, MD; Sidney Kimmel CancerCenterAtJohnsHopkins,Baltimore,MD

Background: Neoadjuvant cisplatin-based chemotherapy before radical cystectomy (RC) improves outcomes but ~50% of patients (pts) are cisplatin-unfit. Anti-PD(L)1 agents can prolong overall survival (OS) in platinum-resistant advanced BC and have shown high pathologic complete response rate (pCR) and safety as single agent in phase II trials in the neoadjuvant setting. The combination of anti-PD-1 and anti-KIR agents is feasible and very attractive based on complementary and non- overlapping roles in regulating adaptive and innate immune response as well as impacting the function CD8+ T and NK-cells. Higher CD8+ T cell density (TCD) at RC tissue correlates with longer OS. We hypothesize, that combining anti-PD1 (N) with anti-KIR (L) is safe and feasible as neoadjuvant therapy in cisplatin-unfit pts and results in high CD8+ TCD at RC. Methods: Phase Ib multi-institutional trial evaluating 2 doses (4 weeks apart) of N alone or N+L in 2 cohorts; pts will be assigned sequentially to N (Cohort 1), and if there is no negative safety signal after the first 12 pts, subsequent pts will be assigned to N+L (Cohort 2). Key eligibility: cT2-4aN0-1M0 stage, $20% tumor at TURBT, adequate organ function, no autoimmune disease within 2 years, no concurrent invasive upper urinary tract carcinoma or other active cancer. Primary endpoint: safety based on CTCAE v5.0 measured as the rate of $G3 treatment related adverse events (AE). Key secondary endpoints: CD8+ TCD absolute and % change between TURBT and RC, % of pts who do not get RC within 6 weeks after neoadjuvant treatment due to treatment-related AE, % pCR, recurrence-free survival, and evaluation of biomarkers in tumor tissue, blood, urine. Rates of $Grade 3 AE with neoadjuvant treatment will be reported along with 90% exact binomial CI. In Cohort 1, maximum CI width is 0.51; in Cohort 2, it is 0.36. Our hypothesis is that the change in CD8+ TCD between TURBT and RC will be about 3 CD8+ T cells / 100 tumor cells within HPF. Up to 43 pts will be enrolled for 36 eligible, treated pts (12:N, 24:N+L). Cohort 1 and 2 have 81% and 98% power, respectively, to detect the hypothesized difference with 1-sided type I error rate of 0.05. Trial is open to accrual in US. Clinical trial information: NCT03532451.

TPS4595 Poster Session (Board #416b), Mon, 1:15 PM-4:15 PM

A phase III randomized open label study comparing bempegaldesleukin (NKTR-214) plus nivolumab to sunitinib or cabozantinib (investigator’s choice) in patients with previously untreated advanced renal cell carcinoma.

Nizar M. Tannir, Neeraj Agarwal, Sumanta K. Pal, Maria Nirvana Formiga, Jun Guo, Daniel J. George, Mary Ann Tagliaferri, Alison L. Hannah, Jenny Yanzhen Zhang, Bridget A. O’Keeffe, Daniel C. Cho; University of Texas MD Anderson Cancer Center, Houston, TX; University of Utah Huntsman Cancer Institute, Salt Lake City, UT; City of Hope National Medical Center, Duarte, CA; AC Camargo Cancer Center, S~ao Paulo, Brazil; Peking University Cancer Hospital and Institute, Beijing Shi, China; Duke University Cancer Institute, Durham, NC; Nektar Therapeutics, San Francisco, CA; Perlmutter Cancer Center New York University Langone Health, New York, NY

Background: Bempegaldesleukin (NKTR-214) is a CD122-preferential IL-2 pathway agonist that stimulates proliferation and activation of tumor antigen-specific CD8+ T cells and natural killer cells within the tumor microenvironment and increases PD-1/PD-L1 expression. These properties make bempegaldesleukin (NKTR-214) a potentially promising agent for combination therapy with checkpoint inhibitors that target and inhibit the PD-1/PD-L1 pathway. In phase 1 studies, NKTR-214 plus nivolumab demonstrated encouraging objective response rates (ORR) in first-line renal cell carcinoma (RCC) and an acceptable safety profile. Immunotherapy with NKTR-214 plus nivolumab may lead to greater clinical benefit than tyrosine kinase inhibitors (TKIs), standard-of-care agents, in this patient population. Methods: This multicenter, randomized, open-label phase 3 study (NCT03729245) will evaluate the efficacy and safety of bempegaldesleukin (NKTR-214) plus nivolumab compared with investigator’s choice of TKI (sunitinib or cabozantinib) in patients with previously untreated advanced or metastatic RCC with clear cell component. Exclusion criteria include active brain metastasis and autoimmune disease. Approximately 600 patients will be randomized in a 1:1 ratio, stratified by PD-L1 status ($1% vs , 1% or indeterminate), International Metastatic RCC Database Consortium prognostic score (1-2 [intermediate risk] vs 3-6 [poor risk]); and TKI (sunitinib or cabozantinib; cabozantinib percentage to be capped at 50%). Combination therapy will consist of bempegaldesleukin (NKTR-214) 0.006 mg/kg intravenously (IV) every 3 weeks (Q3W) plus nivolumab 360 mg IV Q3W until progression or death or maximum of 2 years. TKI therapy will consist of sunitinib 50 mg orally once daily (QD) for 4 weeks followed by 2 weeks off or cabozantinib 60 mg orally QD. Primary objectives are ORR by blinded independent central radiology (BICR) assessment and overall survival. Secondary objectives are progression-free survival by BICR, safety, predictive value of PD-L1 expression, and quality of life. Enrollment is ongoing. Clinical trial information: NCT03729245.

TPS4596 Poster Session (Board #417a), Mon, 1:15 PM-4:15 PM

PDIGREE: An adaptive phase 3 trial of PD-inhibitor nivolumab and ipilimumab (IPI-NIVO) with VEGF TKI cabozantinib (CABO) in metastatic untreated renal cell cancer (Alliance A031704).

Tian Zhang, Karla V. Ballman, Atish Dipankar Choudhury, Ronald C. Chen, Colleen Watt, Yujia Wen, Tyler Zemla, Hamid Emamekhoo, Shilpa Gupta, Michael J. Morris, Daniel J. George, Toni K. Choueiri; Duke University Medical Center, Durham, NC; Weill Cornell Medicine, New York, NY; Dana-Farber Cancer Institute, Boston, MA; University of North Carolina at Chapel Hill, Chapel Hill, NC; University of Chicago, Chicago, IL; Univ of Chicago, Chicago, IL; Mayo Clinic, Rochester, MN; University of Wisconsin School of Medicine and Public Health, Madison, WI; Department of Medicine, Masonic Cancer Center, University of Minnesota, Minneapolis, MN; Memorial Sloan Kettering Cancer Center, New York, NY; Duke University Cancer Institute, Durham, NC; Dana-Farber Cancer Institute, Brigham and Women’s Hospital, and Harvard Medical School, Boston, MA

Background: First-line treatment of mRCC has rapidly changed to include IPI-NIVO or CABO, with clinical benefit of each based on the Checkmate 214 and CABOSUN trials. Combination immunotherapy with VEGF therapies have shown benefit in the JAVELIN 101 and KEYNOTE 426 trials over sunitinib. It is yet unclear which patients (pts) benefit most from combination immunotherapy-VEGF inhibitors, and the optimal sequence of drugs. Methods: In an adaptive, randomized, multicenter, phase 3 trial (Alliance A031704, PDIGREE), pts will start treatment with induction IPI 1mg/kg and NIVO 3mg/kg intravenously (IV) once every 3 weeks. Key inclusion criteria include clear cell mRCC, IMDC intermediate or poor risk, Karnofsky performance status .70, and no prior treatments for mRCC. Based on 3-month radiographic assessment (after completing IPI-NIVO combination), pts with complete responses (CR) will undergo maintenance NIVO 480mg IV every 4 weeks, pts with progression of disease (PD) will switch to CABO 60mg oral daily, and pts with non-CR/non-PD will be randomized to NIVO 480mg IV every 4 weeks versus NIVO 480mg IV every 4 weeks with CABO 40mg oral daily. Randomization will be stratified by IMDC risk criteria and presence of bone metastases. The primary endpoint of the study is overall survival (OS). We hypothesize that 3-year OS rate will improve to 70% for NIVO-CABO compared to 60% for NIVO alone; to achieve 85% power with a two-sided alpha of 0.05 and exponential distribution, 696 patients will be randomized. Accounting for 30% patients with either CR or PD, and 5% dropout from toxicity, up to 1046 pts will be enrolled. Key secondary endpoints include PFS, 12-month CR rate, ORR based on RECIST 1.1 and irRECIST criteria, and toxicity profiles. Quality of life will be assessed based on the FKSI-19, PROMIS-fatigue, and EQ5D-5L questionnaires. Biomarkers associated with CR and association of IL-6 with treatment benefit will be assessed. Other tissue-based and plasma-based biomarkers are planned. Enrollment will begin this year. Support from UG1CA189823, U24CA196171; https://acknowledgments.alliancefound.org. Clinical trial information: NCT03793166.

TPS4597 Poster Session (Board #417b), Mon, 1:15 PM-4:15 PM

PROSPER: A phase III randomized study comparing perioperative nivolumab (nivo) versus observation in patients with renal cell carcinoma (RCC) undergoing nephrectomy (ECOG- ACRIN 8143).

Lauren Christine Harshman, Maneka Puligandla, Naomi B. Haas, Mohamad Allaf, Charles G. Drake, David F. McDermott, Sabina Signoretti, David Cella, Rajan T. Gupta, Brian M. Shuch, Toni K. Choueiri, Primo Lara, Anil Kapoor, Daniel Yick Chin Heng, Michael A.S. Jewett, Viraj A. Master, M. Dror Michaelson, Bradley C. Leibovich, Deb Maskens, Michael Anthony Carducci, On Behalf of the PROSPER RCC Investigators; Dana-Farber Cancer Institute, Boston, MA; Penn Medicine Abramson Cancer Center, Philadelphia, PA; Johns Hopkins University School of Medicine, Baltimore, MD; Herbert Irving Comprehensive Cancer Center, New York, NY; Beth Israel Deaconess Medical Center, Boston, MA; Brigham and Women’s Hospital, Boston, MA; Robert H. Lurie Comprehensive Cancer Center, Northwestern University, Chicago, IL; Duke University Medical Center, Durham, NC; University of California, Los Angeles, Los Angeles, CA; Dana-Farber Cancer Institute, Brigham and Women’s Hospital, and Harvard Medical School, Boston, MA; University of California, Davis, Sacramento, CA; Juravinski Cancer Centre, McMaster University, Hamilton, ON, Canada; University of Calgary, Calgary, AB, Canada; Princess Margaret Cancer Centre, University Health Network, Toronto, ON, Canada; Winship Cancer Institute of Emory University, Atlanta, GA; Massachusetts General Hospital, Boston, MA; Mayo Clinic, Rochester, MN; IKCC, Guelph, ON, Canada; Sidney Kimmel Cancer Center At Johns Hopkins, Baltimore, MD

Background: The anti-PD-1 antibody nivo improves overall survival (OS) in metastatic RCC and is well tolerated. There is no standard adjuvant (adjuv) systemic therapy that increases OS over surgery alone for non-metastatic RCC. Priming the immune system prior to surgery with anti-PD-1 has shown an OS benefit compared to a pure adjuv approach in mouse solid tumor models. Multiple ph 2 studies in bladder, lung and breast cancers have shown remarkable pathologic responses with neoadjuvant (neoadj) PD-1 blockade. Two ongoing ph 2 studies of perioperative nivo in M0 RCC patients are showing preliminary feasibility and safety with no surgical delays (NCT02575222; NCT02595918). PROSPER RCC (NCT03055013) aims to improve clinical outcomes by priming the immune system prior to nephrectomy with neoadj nivo and continued engagement with adjuv blockade in patients with high risk RCC compared to surgery alone. Methods: This global, unblinded, phase 3 National Clinical Trials Network study is currently accruing patients with clinical stage $T2 or TanyN+ RCC of any histology planned for nephrectomy. Oligometastases are permitted if can be rendered NED. We amended the study to enhance accrual and patient quality of life by changing nivo dosing to 480mg q4 wks and requiring baseline tumor biopsy only in the nivo arm. The investigational arm receives 1 dose of nivo prior to surgery followed by 9 adjuv doses. The control arm undergoes standard nephrectomy followed by observation. Randomized patients are stratified by clinical T stage, node positivity, and M stage. Accrual of 805 patients provides 84.2% power to detect a 14.4% absolute benefit in recurrence-free survival (RFS) at 5 years assuming the ASSURE historical control of ~56% to 70% (HR = 0.70). The study is powered to evaluate a significant increase in OS (HR 0.67). Critical perioperative therapy considerations such as safety, feasibility, and quality of life endpoints have been integrated. PROSPER RCC embeds a wealth of translational work aimed at investigating the impact of the baseline immune milieu, the changes induced by neoadjuvant anti-PD-1 priming, and how both may predict clinical outcomes. Clinical trial information: NCT03055013.