Genomic alterations associated with response to immune checkpoint inhibitors in rare cancers: A biomarker exploration study from the Australian Molecular Screening and Therapeutics (MoST) Program

2594 Background: Despite immune checkpoint inhibitors (ICI) demonstrating efficacy in over 15 cancer types, evidence in most rare cancers remain scarce, reflecting the challenges of conducting prospective trials. This Australia-wide study reports real-world (RW) objective responses to ICIs targeting the programmed death-(ligand)1 (PD-1/PD-L1) pathway in a large precision oncology program, and associations between response and genomic alterations in rare cancers. Methods: This study was conducted within the framework of the MoST Program (ACTRN12616000908437), where adult patients (pts) with advanced solid tumors were screened using comprehensive genomic profiling (CGP). Only pts who received ICI after CGP were included in this analysis. The best overall response (BOR) to ICI was reported by the treating oncologist during the follow-up period. Best response was categorized as complete (CR), partial (PR), stable (SD), or progressive (PD). Two categories were defined: responders were individuals reported to have experienced CR or PR, while non-progressors were those with CR, PR, or SD. The association between tumor mutational burden (TMB) and response categories was assessed by area under the receiver operator characteristic curve (AUC). Odds ratios (OR) of specific genomic alterations associated with each category were also estimated. Results: Between Dec 2016 and Aug 2022, 4343 pts were screened; 3471 (80%) rare cancer pts, of whom 556 went on to 90 different ICI trials after undergoing CGP. In 273 pts treated with ICI where BOR was reported by the treating oncologist, the median survival was 17.3 months (95% CI 15.0 to 20.4). 13 pts (4.8%) had CR reported, including 8 pts with unapproved indications: 3 sarcomas, and 1 each, small cell carcinoma of unknown primary, adenoid cystic carcinoma, pituitary adenoma, thymoma, and pancreatic adenocarcinoma. 45 (16.5%) and 112 (41.0%) had PR and SD reported, respectively. High TMB, defined as ≥10 non-synonymous tumor mutations/megabase, was identified in 39 (14.3%) pts. TMB was associated with response (AUC 0.715, 0.630 to 0.800) and non-PD (AUC 0.600, 0.528 to 0.671) to ICI. Further exploratory analyses identified that alterations in EP300 (n = 7, OR, 9.8, 1.9 to 52), PTCH1 (n = 6, OR, 15.5, 1.7 to 141) , MSH6 and KMT2D (n = 6 each, OR, 7.7, 1.4 to 43) , and NF1 (n = 10, OR, 3.89, 1.1 to 13.9) were associated with CR/PR. Genomic ARID1A alterations (n = 18) were associated with non-PD (OR 5.16, 1.2 to 22.9), as were alterations in the SWI/SNF complex (n = 20, OR 3.64, 95% CI 1.0 to 12.8) . Conclusions: Notwithstanding their limitations, these real-world data suggest that rare cancer patients may also benefit from ICI. Further prospective investigations into pan-cancer genomic biomarkers of response are warranted. Clinical trial information: ACTRN12616000908437 .