Dottorato in NETWORK ONCOLOGY AND PRECISION MEDICINE

Titolo della tesi: Enhancing Pembrolizumab Response in Never-Smoker Non-Small Cell Lung Cancer: Unraveling the Impact of DNA Repair Gene Mutations through Real-World Analysis and Patient Similarity Networks

Abstract Purpose Single-agent immune checkpoint inhibitor (IO) therapy is the standard treatment for non-oncogene addicted advanced non-small cell lung cancer (aNSCLC) with high PD-L1 expression (PD-L1 ≥ 50%). High tumor mutation burden (H-TMB) has been identified as a significant biomarker for IO response. Smoking-induced damage leads to H-TMB in smoking aNSCLC patients (S-pts), while never-smoking patients (NS-pts) typically exhibit low TMB (L-TMB) and are less responsive to IO therapy. However, the molecular characteristics of NS-pts with H-TMB remain poorly defined. Experimental design We conducted a retrospective analysis of clinical data from 142 aNSCLC patients with PD-L1 ≥ 50% who received first-line pembrolizumab therapy. Next-generation sequencing using the FoundationOne®CDx assay was performed to correlate genomic alterations with clinical characteristics and treatment outcomes. Mutations were categorized into eleven main pathways, including cell cycle, Hippo, Myc, Notch, oxidative stress/Nrf2, PI3K, RTK/RAS/MAP, TGF-β, p53, β-catenin/Wnt, and DNA damage response (DDR) pathways. Enrichment analysis was performed on pathways with at least one mutation per patient to identify patient subgroups based on the mutated pathways. Additionally, to further investigate the molecular characterization of patient subgroups, a patient similarity network was constructed based on the mutational profile to determine pairwise patient similarity. These results were validated using publicly available molecular characterization data from the POPLAR and OAK randomized controlled trials, which included 853 NSCLC patients. Results Among the 142 aNSCLC patients, 111 were S-pts and 31 were NS-pts. S-pts exhibited higher TMB compared to NS-pts (median TMB: 8 vs. 4 Mut/Mb). Surprisingly, 11 NS-pts showed H-TMB (median TMB: 16.39 Mut/Mb) and were significantly enriched in β-catenin/Wnt and DDR pathway mutations (FDR adjusted p-value<0.05). The enrichment of DDR pathway mutations in NS-pts with H-TMB was also confirmed using publicly available molecular characterization data from 853 NSCLC patients in the POPLAR and OAK trials. In the real-world cohort, the subgroup of H-TMB/NS-pts with DDR pathway mutations displayed a more favorable response to IO therapy. Furthermore, patient similarity network analysis based on the mutated pathways consistently showed that NS-pts with H-TMB and a prevalence of DDR mutations formed distinct cluster, which was associated with improved overall survival both in the real-world cohort and in the publicly available randomized controlled trials. Conclusions The DDR signature appears to play a crucial role as an additional driver of H-TMB in NS-pts. This specific subgroup of IO-responsive NS-pts may have a better prognosis. Our findings suggest that a mutational profiling approach targeting the DDR pathway could aid in the identification of NS-pts who are likely to benefit from IO therapy. Key messages • Alterations in DNA damage response and repair (DDR) genes are associated with higher genomic instability and H-TMB in cancer and deleterious DDR mutations are associated with improved clinical outcomes in advanced NSCLC patients treated with PD-L1 blockade. • Pathogenic DDR mutations may offer an additional signature for immunotherapy response in advanced, non-oncogene addicted, never-smoker NSCLC with PD-L1 ≥ 50%. • The evaluation of DDR mutation status in addition to other available predictive biomarkers may improve physician confidence in selectin never-smoker patients for immunotherapy. • Patient similarity network analysis has revealed a clear clustering of never-smoker patients exhibiting high TMB and DDR mutations. This particular cluster is associated with significantly improved overall survival outcomes, as observed in both real-world patient data and publicly available clinical trial datasets. This approach enhances patient stratification and contributes to the development of more personalized and effective treatment strategies for cancer patients.