Validating the Results of a Matching Adjusted Indirect Comparison (MAIC) in MET Exon 14 (METex14) Skipping Non-Small Cell Lung Cancer (NSCLC)

OBJECTIVES: Tepotinib is approved in multiple countries for METex14 skipping NSCLC based on the Phase II VISION study (NCT02864992). As this study did not contain a control arm, real-world data (RWD) from five pooled sources in patients with METex14 skipping NSCLC were used to estimate the comparative effectiveness of tepotinib against pembrolizumab. However, uncertainty exists in the comparisons due to the non-randomized nature of RWD. This study aimed to confirm the findings of the RWD analysis using published clinical trial data.

METHODS: RWD for pembrolizumab were available for patients with METex14 skipping NSCLC, and published data from KEYNOTE-024 were available for pembrolizumab in NSCLC without EGFR/ALK mutations (wild-type). Tepotinib (treatment-naïve patients from VISION) was compared with pembrolizumab RWD using propensity scoring and with KEYNOTE-024 using MAIC. For validation, pembrolizumab RWD (METex14 skipping) were reweighted for comparison with KEYNOTE-024 (wild-type) to assess differences in pembrolizumab outcomes between wild-type and METex14 populations.

RESULTS: A benefit was seen for tepotinib when comparing to pembrolizumab using METex14 NSCLC RWD (median investigator progression-free survival [PFS]: 8.5 versus 3.3 months). In the KEYNOTE-024 comparison, a benefit was also observed for tepotinib (median PFS 13.5 versus 8.3 months). Overall survival (OS) results were similar between tepotinib and pembrolizumab. After accounting for differences in patient characteristics, pembrolizumab RWD presented a poor match in outcomes versus the wild-type population from KEYNOTE-024 (median PFS: 6.2 vs 8.3; median OS: 15.6 vs 26.0 months).

CONCLUSIONS: Multiple datasets with different levels of data access can be leveraged to reduce the inherent uncertainty of indirect comparisons. In this analysis, tepotinib improved PFS compared with pembrolizumab regardless of data source. After adjusting for prognostic characteristics, the pembrolizumab RWD showed poorer outcomes than pembrolizumab clinical trial data, suggesting a negative prognostic impact of METex14 skipping. However, other differences in the data sources limit conclusions.