ASSOCIATION BETWEEN KRAS/STK11/KEAP1 MUTATIONS AND OUTCOMES IN PATIENTS FROM THE POSEIDON STUDY: DURVALUMAB ± TREMELIMUMAB + CHEMOTHERAPY IN MNSCLC

Introduction: In POSEIDON, a Phase 3 study (NCT03164616) of metastatic EGFR/ALK wild-type NSCLC (mNSCLC), first-line tremelimumab + durvalumab + chemotherapy (T+D+CT) demonstrated statistically significant improvements in both progression-free survival (PFS) and overall survival (OS) vs CT alone. Mutations in STK11 and KEAP1 correlate with poor prognosis and are associated with chemorefractory and immunologically “cold” tumours that are less responsive to therapy. KRAS is frequently co-mutated with STK11 and/or KEAP1. The triplet regimen of T+D+CT may improve clinical outcomes for hard-to-treat subgroups of patients with mNSCLC. Here we describe post hoc exploratory analyses of survival outcomes in POSEIDON according to KRAS, STK11 and KEAP1 mutational status. Methods: 1013 patients were randomised 1:1:1 to T+D+CT, D+CT or CT alone. Patient tumours were molecularly characterised via sequencing of tumour tissue DNA and/or circulating tumour DNA (ctDNA) samples. OS, PFS and response were analysed in patients with mutations (m) in KRAS, STK11 or KEAP1. Analyses of KRAS and STK11 were conducted in patients with non-squamous (NSQ) histology, while KEAP1 was analysed irrespective of histology due to the small sample size. Results: 637 patients (63% of ITT) had NSQ histology, of whom 612 (96% of NSQ ITT) were mutation evaluable either by tissue, ctDNA or both. NSQ mutation-evaluable patients had 30%, 14% and 6% KRASm, STK11m and KEAP1m respectively. OS hazard ratios (HRs) favoured T+D+CT vs CT in KRASm (HR 0.56 [95% CI 0.36-0.88]), STK11m (0.56 [0.30-1.03]) or KEAP1m (0.43 [0.16-1.25]) patients. Both median OS and landmark 24-month OS rates were higher with T+D+CT vs CT across the KRASm, STK11m and KEAP1m subgroups, suggesting sustained benefit with the triplet regimen. Additionally, higher objective response rates and duration of responses were observed with the triplet regimen vs CT in these patients. Similar trends were observed for PFS (data will be presented). Conclusion: OS HRs, as well as median and 24-month landmark values, favoured T+D+CT vs CT in KRASm, STK11m or KEAP1m patients, although the small sample size of some subgroups limits interpretation. These post hoc exploratory analyses support T+D+CT as a potential first-line treatment option for mNSCLC in harder-to-treat patient subgroups, such as those with KRASm, STK11m or KEAP1m tumours.