Clinical correlates of response to anti-PD-1 (aPD1)-based therapy in patients (pts) with metastatic melanoma (MM)

e21000 Background: aPD1, alone or in combination with ipilimumab (IPI), produces durable responses in a subset of MM. Tumor features that correlate with treatment response, including size, number, and location of metastases (mets) are not well defined. Methods: We collected clinical data from mm pts treated at one center who received aPD1 (n = 185) or aPD1 + IPI (n = 42). We correlated number of mets, size of largest tumor, and organ involvement with response rate (RR), progression-free survival (PFS) and overall survival (OS). Results: Among all pts, RR was 67% for aPD1 + IPI and 41% for aPD1 alone. In univariate analyses, responders to aPD1 had lower diameter of largest tumor (4cm vs. 5.5cm; p = 0.02) whereas aPD1 + IPI had equivalent largest tumor diameters (p = 0.65). Regarding sites of mets, liver mets were associated with lower RR in aPD1 treated pts (26% vs. 46%), lower PFS (median 138 vs. 326 days, p = 0.02), and lower OS (median 334 vs. 1080 days, p < 0.01). No associations with RR, PFS, or OS were observed with liver mets in aPD1 + IPI treated pts. We also did not observe any differences between pts who did or did not have lung, lymph node, or brain mets for either aPD1 or aPD1 + IPI. Interestingly, superior RR to aPD1 + IPI was observed in pts with bone mets compared to those without bone mets (91% vs. 58%, p = 0.048). Regarding number of sites, RR to aPD1 was greater in pts with ≤10 mets compared with those with > 10 (46% vs. 28%, p = 0.02), although no consistent relationship was observed at lower cutoffs. In multivariable analyses, diameter of largest tumor (tumor bulk) was independently associated with PFS (OR, 1.11, p < 0.001) and OS (OR 1.08, p < 0.001) whereas AJCC stage, lactate dehydrogenase, liver mets, ECOG performance status, number of mets, and prior therapies were not significant. Tumor bulk and other risk factors were not associated with PFS or OS in aPD1 + IPI. Conclusions: Tumor bulk was strongly and independently associated with clinical outcomes in aPD1 but not IPI + aPD1. Other associations with disease sites (liver and bone) need further validation. In conjunction with molecular biomarkers, clinical predictors may help guide selection of aPD1 or aPD1 + IPI.