Comparative effectiveness in <i>MET</i> ex14 NSCLC: Outcomes from a propensity score–based analysis of 8 pooled real-world datasets (the TOGETHER study).

Abstract

e20749 Background: MET ex14 skipping alterations in non-small cell lung cancer (NSCLC), recognized as important drivers of tumorigenesis, present in 3-4% of cases and are associated with poor prognosis. While case series suggest worse real-world outcomes in patients with MET ex14 skipping alterations compared to the broader NSCLC population, there is limited comparative effectiveness evidence on these outcomes for standard treatments prior to the first regulatory approvals of targeted therapy with selective MET receptor tyrosine kinase inhibitors (MET TKIs). Methods: Patient-level data from eight international real-world datasets of MET ex14 skipping NSCLC patients covering the period from 2010 to 2022 were pooled, applying inclusion/exclusion criteria consistent with the tepotinib MET TKI VISION trial. Propensity score (PS) weighting (based on clinical input) was applied to perform comparisons of real-world progression-free survival (PFS) and overall survival (OS) in four treatment classes: chemotherapy (CT), immunotherapy monotherapy (IO), IO plus chemotherapy (IO+CT), and an unapproved TKI (crizotinib). Results: Ultimately, 406 patients, treated across 661 lines of therapy, were incorporated into the analyses. Median PFS was, in months (95% CI): CT 4.8 (3.9–7.3), IO 5.7 (3.2, 10.7), IO+CT 4.9 (3.6, not evaluable), crizotinib 7.4 (4.7, 11.6). Median OS ranged from 11.9 months (8.4, 19.7) for crizotinib, to 18.4 months (11.0, 32.9) for IO. However, OS estimates are confounded by the effects of subsequent treatments and susceptible to immortal time bias due to the retrospective nature of data. Results were consistent across sensitivity analyses, including stratification by treatment line, disease subtype, and bootstrapping by sequential exclusion of datasets. Conclusions: Pooled real-world datasets of traditional therapies for MET ex14 skipping NSCLC demonstrated limited effectiveness. Median PFS was similar across treatment classes. IO (whether alone, or in combination with CT) did not demonstrate its clinical advantage in PFS or OS outcomes generally seen in the wider NSCLC population, while crizotinib did not outperform chemotherapy. These uniformly poor outcomes illustrate the need for further targeted treatment options in MET ex14 skipping alterations NSCLC. Confounding of OS, in part due to the use of subsequent therapies forms an important limitation of the analyses.