Upfront osimertinib and as sequential therapy in patients with -mutant non-small cell lung cancer (NSCLC): benefit across patients groups in a real-world retrospective cohort-the smile study.

Background: Osimertinib is the preferred first-line (L1) treatment for epidermal growth factor receptor-mutant (m) advanced non-small cell lung cancer (aNSCLC). Intensification of L1 with chemotherapy or amivantamab has shown improved outcomes at the cost of increased toxicity, raising questions about the optimal patients selection. A sequence involving first-generation tyrosine kinase inhibitor (TKI) (1G) followed by osimertinib might be considered. This study assessed the efficacy of these therapeutic strategies based on the clinical profiles of a real-life cohort.

Methods: Retrospective multicenter study including consecutive patients with m (ex19/ex21) aNSCLC treated with either osimertinib or the sequence of 1G followed by osimertinib ("sequence group"). Central nervous system (CNS) metastases were permitted. We assessed progression-free survival (PFS) of the global strategy (PFSglob) defined as the time between L1 start and progression after L2 treatment or death. Secondary endpoints were overall survival (OS), PFS of the L1 treatment, and tumor response according to each center daily practice [objective response rate (ORR) and disease control rate (DCR)].

Results: A total of 300 patients with m aNSCLC were enrolled (n=161 in the osimertinib group, n=139 in the sequence group). Baseline characteristics in both groups were similar except for baseline CNS involvement (41% in osimertinib-group 25%), poor performance status (PS) ≥2 (21% 10%) and high-tumor burden (TB), defined as >3 metastatic sites or CNS involvement (51% 35%). The osimertinib group had longer median first-line PFS (PFS1; 19.0 16.8 months, P=0.03). The sequence group had improved PFSglob the osimertinib-group (32.4 26.5 months, P=0.04) but this difference was not significant in multivariate Cox analysis (adjusted on age, smoking history, number of metastatic sites, liver, CNS and soft tissue metastasis, and PS) nor after a propensity score matching analysis, osimertinib upfront was associated with better PFSglob in the poor-prognosis groups: high-TB, CNS or liver involvement and poor PS.

Conclusions: In this real-life study we showed that osimertinib upfront demonstrated prolonged PFS1 1G followed by osimertinib, with better PFSglob in patients with poor-prognosis m aNSCLC. This study raises the question of patients selection and treatment tailoring for the first line management of metastatic m non-small cell lung cancer (NSCLC).