Toni k. Choueiri¹, Mayer n. Fishman², Bernard Escudier³, David f. Mcdermott⁴, Arriet Kluger⁵, Walter m. Stadler⁶, Jose luis Perez-gracia⁷, Douglas Mcneel⁸, Brendan Curti⁹, Michael r. Harrison¹⁰, Elizabeth r. Plimack¹¹, Leonard Appleman¹², Lawrence Fong¹³, Charles g. Drake¹⁴, Tina c. Young¹⁵, Petra Ross-macdonald¹⁵, Jason s. Simon¹⁵, Dana Walker¹⁵, Mario Sznol⁵

¹oncology, Dana-Farber Cancer Institute, Boston, MA, USA, ²Tampa, H. Lee Moffitt Cancer Center & Research Institute, ³Villejuif, Institut Gustave Roussy, ⁴Dana-Farber/Harvard Cancer Center, Beth Israel Deaconess Medical Center, ⁵New Haven, Yale Cancer Center, ⁶Chicago, University of Chicago Medicine, ⁷Pamplona, University Clinic of Navarra, ⁸Department of Medicine, University of Wisconsin-Madison, ⁹Providence Portland Medical Center, Providence Cancer Center, ¹⁰Durham, Duke University Medical Center, ¹¹Philadelphia, Fox Chase Cancer Center, ¹²Pittsburgh, University of Pittsburgh Medical Center (UPMC) Cancer Pavilion, ¹³San Francisco, University of California San Francisco (UCSF) Helen Diller Family Comprehensive Cancer Center, ¹⁴Baltimore, Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins University, ¹⁵Princeton, Bristol-Myers Squibb

Objective:This prospective biomarker study in patients (pts) with mRCC treated with the programmed death-1 (PD-1) inhibitor antibody nivolumab assessed baseline (BL) and changes in serum chemokines, tumor T cell infiltrates (TIL), gene expression, T cell repertoire (TCR), and other biomarkers potentially associated with clinical outcomes (NCT01358721). Method: Pts treated with 1–3 prior therapies received nivolumab 0.3, 2, or 10mg/kg IV Q3W; treatment-naïve pts received 10mg/kg IV Q3W. Biopsies were obtained at BL and cycle 2 day 8. Overall survival (OS) parameters were estimated by Kaplan-Meier method. Tumor PD-L1 expression was measured by immunohistochemistry (28-8 antibody; Dako). PD-L1 positivity was defined as ≥ 5% tumor membrane staining in ≥ 1 biopsy; tumor burden response as ≥ 20% reduction. Gene expression data were obtained on Affymetrix U219. Result: 91 pts were treated. Of 56 evaluable BL biopsies, 32% were PD-L1+. Median OS (95% CI) was 16.4 mo (10.1–not reached [NR]) for 0.3mg/kg, NR for 2mg/kg, 25.2 mo (12.0–NR) for 10mg/kg, and NR for treatment-naïve pts. 1-yr and 2-yr OS rates (95%CI) were 75% (64–83) and 58% (46–68), respectively. OS by PD-L1 status is summarized (table). Pts with tumor burden response (n=13) had ≥ 1.3-fold differential BL expression of 311 genes (P<0.01, false discovery rate < 16%). Cell-mediated immune transcripts were elevated, including effector cell markers GZMB, NKG7, and CD7, NK/CD8-activating ligand MICB, inflammasome component AIM2, and activated macrophage marker IL-1a. Analysis of association between OS and serum chemokine levels, TCR and TIL is ongoing. PD-L1+n = 18PD-L1–n = 38Median OS, mo(95% CI) Overall NR23.4 (13.1–33.3) Previously treated NR22.3 (12.0–27.0) Treatment-naïve NR33.3 (2.0–NR)OS rate(95%CI) 1-yr71 (44–87)71 (52–83)2-yr64 (37–82)48 (30–64). Conclusion:Association of immune markers at BL with subsequent tumor burden response suggests that infiltrating immune activating cells may mediate response to nivolumab in mRCC pts. Consistent with the randomized phase II study of nivolumab in mRCC, OS appears longer in PD-L1+ pts but promising in both PD-L1+ and PD-L1– pts, especially when treatment-naïve. Reused with permission from the American Society of Clinical Oncology (ASCO). This abstract was accepted and previously presented at the 2015 ASCO Annual Meeting. All rights reserved.

Key Words: Nivolumab, Immunomodulatory activity, Renal cell carcinoma