Response to ICI per RECIST 1.1 is durable and associated with improved overall survival (OS). Benefit extends beyond responders to pts with stable disease (SD), who constitute a heterogeneous group with increase or decrease in tumor size without meeting response or progression criteria. We hypothesize that ART may better discriminate benefit translating to improved OS. We included 1216 pts with advanced non-small cell lung cancer (NSCLC), melanoma, urothelial carcinoma (UC), esophagogastric cancer (EGC), renal cell carcinoma (RCC), head and neck squamous cell cancer (HNSCC), and colorectal cancer (CRC) treated with ICI. Using Neutrophil-Lymphocyte Ratio (NLR), sites of metastasis, ECOG-performance status, tumor mutation burden, and prior therapy, we conducted Cox proportional hazards models to examine association of ART and RECIST1.1 with OS. Results were validated using bootstrap analyses of 1000 replicates. We included pts with NSCLC (n=522), melanoma (n=262), UC (n=140), EGC (n=93), RCC (n=83), HNSCC (n=71) and CRC (n=43). Median age was 66 (range 58-73) and 58% were male. Per RECIST, 68 pts (5.6%) had complete response (CR), 277 (22.9%) partial response (PR), 579 (47.7%) progressive disease (PD), and 290 (23.9%) SD. 175 (60.3%) pts with SD had ART. CR+PR (vs SD+PD) was associated with OS (HR=0.3, 95%CI: 0.2-0.4, p<0.001). ART (vs no ART) was significantly prognostic for pts with SD on univariate (HR=0.3, 95%CI: 0.2-0.3, p<0.001) and multivariate analysis (Table). Bootstrap validation corroborated these results (HR=0.5, 95%CI: 0.4-0.7, p<0.001) in NSCLC, melanoma, and UC where >25 pts had SD.Table: 116PVariableHR (95% CI)p-valueECOG1.6 (1.1-2.2)0.009NLR1.4 (1.1-1.7)0.003Bone Metastasis0.7 (0.5-0.9)0.013ART0.5 (0.3-0.6)<0.001 Open table in a new tab ART is a robust endpoint capturing OS benefit in a larger proportion of pts with advanced solid tumors receiving ICI compared to RECIST 1.1. ART warrants study as a surrogate endpoint in pts without RECIST-measurable disease to enable inclusive trials that expedite drug development and biomarker discovery.

Purpose/Objective(s)Radiation associated muscle-invasive bladder cancer (RA-MIBC) is a rare but potentially lethal late toxicity following definitive radiation therapy for prostate cancer. RA-MIBC is associated with worse prognosis compared to primary MIBC, but it remains unknown whether worse outcomes are driven primarily by difficulty with delivering definitive treatment in a previously irradiated pelvis or whether RA-MIBCs harbor specific aggressive biological features. We sought to characterize the genomic features of RA-MIBC to assess whether the underlying genetic alterations differ from primary (non-radiation associated) MIBC.Materials/MethodsWe performed whole-exome sequencing of 27 bladder tumors in patients previously treated with definitive radiation therapy for prostate cancer, and we correlated genomic with clinical data. We compared the mutation profile of RA-MIBC with primary MIBC from The Cancer Genome Atlas (TCGA). We used the Kaplan-Meier method to estimate progression-free survival (PFS) and overall survival (OS).ResultsThe median age of included patients was 79.2 years, and the median time from prostate radiation to MIBC diagnosis was 11.5 years. Median PFS was 12.4 months and median OS was 20.2 months. Among RA-MIBCs, we found mutation signatures characteristic of BRCA1/2 mutations, impaired DNA mismatch repair, and microsatellite instability. When compared to de novo MIBC from TCGA, RA-MIBCs also had significantly higher mutation rates of DNA repair genes, such as FANCA (38.1% vs 5.1%, P < 0.001), CHEK2 (26.1% vs 2.5%, P < 0.001) and MSH6 (26.1% vs 2.5%, P < 0.001). The most commonly mutated genes were TP53 (50%) and KDM6A (36%), which harbored mutations at similar rates as de novo MIBC samples from TCGA.ConclusionIn this study, we characterize the mutational profiles and clinical outcomes of radiation-associated bladder cancer. These patients exhibited significantly worse PFS and OS compared to prior series of de novo MIBC. Genomically, these tumors exhibited higher rates of mutation signatures characteristic of impaired DNA damage repair compared to previously unirradiated tumors. These genomic differences suggest that some RA-MIBCs may represent a biologically distinct entity compared to de novo MIBC and may indicate an opportunity for the use and development of novel therapeutic strategies.