Dissection of Progressive Disease Patterns for a Modified Classification for Immunotherapy

Importance Progressive disease (PD) in patients treated with immune checkpoint inhibitors (ICIs) varies widely in outcomes according to the Response Evaluation Criteria in Solid Tumors (RECIST), version 1.1. Efforts to modify RECIST for ICI treatment have not resolved the heterogeneity in PD patterns, posing a clinical challenge. Objective To develop and validate a modified PD classification based on PD patterns and evaluate its association with postprogression survival (PPOS) in patients treated with the programmed cell death protein ligand 1 antibody atezolizumab across various solid tumors. Design, Setting, and Participants This study analyzed data from 5 phase 3 trials (IMmotion151, IMvigor211, OAK, Impower133, and IMspire150) involving patients treated with atezolizumab for renal cell carcinoma (RCC), urothelial carcinoma, small cell lung cancer, non–small cell lung cancer, and melanoma. This post hoc analysis was conducted from March to September 2024. Exposure Treatment with atezolizumab. Main Outcomes and Measures The primary outcome was the association of PD patterns with PPOS. Seven PD patterns were identified based on the enlargement of target and nontarget lesions or new lesions and their combinations. Results A total of 1377 patients were analyzed across the 5 trials. In RCC, 7 PD patterns significantly affected prognosis. The 6-month PPOS probability ranged from 26% for progression in target and nontarget lesions plus new lesions to 90% for progression in either target or nontarget lesions alone. A modified PD classification was developed that categorized PD into 3 risk levels: low risk (progression of existing lesions), intermediate risk (new lesions without progression of existing lesions), and high risk (progression of existing lesions plus new lesions). This score was associated with PPOS in ICI-treated RCC, with hazard ratios of 0.23 (95% CI, 0.13-0.41; P < .001) and 0.39 (95% CI, 0.23-0.66; P < .001) for low-risk and intermediate-risk PD compared with high-risk PD, respectively. Validation in additional trials confirmed the score’s applicability across various tumors. Conclusions and Relevance In this study, a survival score was developed based on PD patterns. The risk classification was associated with PPOS across various solid tumors treated with immunotherapy and may therefore enhance prognostication and clinical decision-making, potentially providing a valuable tool for treating patients with PD who are receiving immunotherapy.