Identifying prognostic and predictive value of mutations associated with clinical outcomes in first line (1L) patients with advanced or metastatic non–small cell lung cancer (aNSCLC).

Abstract

9047 Background: Whilst the impact of certain mutations on aNSCLC survival outcomes has been evaluated, real-world (rw) evidence on its predictive and prognostic value remains limited. We aimed to identify the predictive and prognostic value of non-driver mutations that are deemed to be clinically relevant in aNSCLC. Methods: This retrospective cohort study analysed de-identified electronic medical records from the Flatiron Clinico-Genomic (FCGD) and FoundationCORE™ databases of aNSCLC patients who have initiated 1L cancer immunotherapy (CIT, alone or in combination) or chemotherapy (chemo) under routine care between 2016 and 2021. Propensity-based multivariable models were used to determine associations between mutational status and progression-free survival (rwPFS), overall survival (rwOS) and overall response (rwR). Results: Of 10,795 patients identified, 2,999 met inclusion criteria (1,042 chemo, 1,957 CIT). Among these, 53% were men and 65% were ≥65 years old. Patients mostly had non-squamous histology (68%), a previous history of smoking (95%) and ECOG ≤1 (61%). 58% of patients were diagnosed as de novo. A total of 185 different mutations were identified in this population, and variants of the same mutation were grouped, identifying 58 mutation families. STK11, KEAP1 and CDKN2A/B mutations were significantly associated with all effectiveness outcomes (Table), indicating lower response (rwR OR<1) and shorter survival (rwOS and rwPFS HR>1) in patients harbouring mutated vs. wild type genes. APC and KRAS mutations were only significantly associated with lower rwR, while FGRF and HRAS mutations were related to worse rwPFS and rwOS, respectively. In contrast, a significantly higher likelihood of response or PFS was associated with ATM/R/RX and GATA3 mutations, respectively. Moreover, our results suggest that KRAS mutations potentially have predictive value for PFS in CIT-treated patients. Mutations with prognostic value on clinical outcomes (overall population). HR, hazard ratio; PFS, progression-free survival; OR, odds ratio; OS, overall survival; rw, real-world. Conclusions: STK11, KEAP1 and CDKN2A/B mutations were significantly associated with poor prognosis in all effectiveness outcomes. In addition, KRAS mutations resulted in clinically significant differences in terms of PFS. Further analyses should be conducted to confirm this clinical significance and to evaluate the relevance of these mutations in future clinical trials design.[Table: see text]