Validation and refinement of Society of Immunotherapy of Cancer (SITC) definitions for PD-(L)1 resistance: An analysis of more than 1,300 participants from SWOG.

Abstract

2656 Background: New immuno-oncology (IO) agents are commonly used in patients who previously received PD-(L)1 inhibitors. In order to facilitate clinical trial interpretation and better delineate therapeutic contributions of novel IO agents, consensus definitions of PD-(L)1 single-agent and combination immunotherapy resistance were published in 2020 (PMC7174063) and 2022 (PMC10016305) by the Society of Immunotherapy of Cancer (SITC); definitions outlined in Table. Validation of these expert-derived definitions is currently lacking. Herein we analyze two SWOG trials to evaluate the proposed definitions for the advanced and adjuvant settings. Methods: S1609/DART (NCT02834013) was a basket trial for patients with rare cancers treated with ipilimumab (1mg/kg intravenously [IV] every 6 weeks) plus nivolumab (240mg IV every 2 weeks). S1404 (NCT02506153) included an arm where patients with high-risk resectable Stage III melanoma received adjuvant pembrolizumab (200mg IV every 3 weeks for 1 year). In both trials, overall survival (OS) was measured from study registration to death from any cause with those last known to be alive censored. OS was evaluated with Kaplan-Meier and martingale residual plots and Cox regression models. Results: In S1609 (advanced setting), 733 participants were analyzed: 127 (17%) were not evaluable for primary resistance due to death or off treatment before 6 weeks. Among the 570 evaluable, 366 met the SITC primary resistance definition and 204 did not. Martingale residuals plots indicated a positive association between time to progression and OS with no evidence of a threshold. With a 6-month landmark, participants with primary resistance had significantly shorter OS compared to those who did not: hazard ratio (HR)=2.84, 95% confidence interval (CI) 2.28-3.55, p<0.001. In S1404 (adjuvant setting), 626 participants were analyzed with 12 (2%) meeting the definition of early recurrence/primary resistance and 138 (22%) meeting the definition of late recurrence/secondary resistance. Using a 12-month landmark, there was no significant difference in OS between early and late recurrences (HR=0.98, 95% CI:0.35-2.70, p=0.25), however late recurrences were associated with significantly shorter OS than no recurrence (HR=7.69, 95% CI:2.71-20.1,p<0.001). Conclusions: In the advanced cancer cohort, the SITC definitions were validated. In the adjuvant cohort, early recurrences were uncommon and there was no significant difference in OS between early and late recurrences suggesting a 12-month cutoff may be more appropriate than 12 weeks. Additional analyses in other patient cohorts are needed to further understand if the SITC definitions for advanced cancers validate more broadly and if data-drive refinements are needed for the adjuvant setting. Advanced Primary Treatment > 6 weeks; no response or < 6 months Secondary Treatment > 6 months; response/stable > 6 months Adjuvant Early/primary < 12 weeks last dose Late/secondary 12 weeks