Baseline mutational profiles of patients with carcinoma-of-unknown-primary-origin (CUP) enrolled onto CUPISCO

Abstract

Aims: NCCN guidelines consider next-generation sequencing important in therapeutic decision-making in CUP. CUPISCO (NCT03498521) is an ongoing, phase II, randomized study of targeted therapy/cancer immunotherapy vs platinum-based chemotherapy in ESMO-guideline-defined unfavourable CUP. We present a preliminary, descriptive molecular analysis of ∼50% of patients designated for enrollment. Methods: Upon enrollment, comprehensive genomic profiling, including determination of microsatellite instability and tumour mutational burden (TMB), was performed on formalin-fixed, paraffin-embedded issues using the F1CDx assay. Gene alterations (GAs) found in ≥3% of patients were analyzed using multiple correspondence analyses and hierarchical clustering to identify co-occurrences. Results: Median age was 61.5 years (n = 346 [Apr 2021]; range: 22– 84); median TMB, 2.5 mutations/Mb (0–63.0). In our analysis, 30% of patients carried a potentially targetable GA. Most-frequent GAs were TP53 (44%), CDKN2A (32%), KRAS (21%; 2% G12C alterations), CDKN2B (21%), ARID1A (13%), STK11 (13%), MTAP (12%), PIK3CA (10%), MYC (8%), PBRM1 (8%), BAP1 (8%), and FGFR2 (8%). Beyond PIK3CA and FGFR2, other targetable GAs were identified in EGFR (2%), ERBB2 (6%), ALK (0.3%), ROS1 (1%), MET (2%), NTRK1 (1%), and BRAF (6%). The frequency of microsatellite instability- and TMB-high (> 16 mutations/Mb) samples was 3% and 9%, respectively. Based on hierarchical clustering of co-mutational profiles, multiple clusters were identified in the study cohort, each characterized by specific GA cooccurrences. Conclusions: This descriptive analysis sheds further light on the molecular landscape in patients with poor-prognosis CUP. Our analyses demonstrate that CUP cases can be clustered based on molecular profiling; further studies are needed to determine if these clusters carry clinical relevance. Our early results suggest that comprehensive genomic profiling of CUP samples identifies therapeutically relevant GAs in a significant proportion of patients and could thus guide personalized treatment of these tumors.