A Taxonomy for Assessing Real-World Targeted Cancer Therapy Options in the Context of Broad Genomic Profiling

Broad genomic profiling (BGP) is increasingly used to determine eligibility of molecularly targeted cancer therapies. Understanding the population-level impact of BGP on treatment selection requires a clinically oriented framework guiding interpretation of the guideline-directed actionability of BGP results.

We developed the Yale Molecular Alteration Taxonomy for Real-World Individualized Treatments (Y-MATRIX) to classify BGP results in relation to contemporaneous FDA approvals and NCCN Guidelines. We then applied our schema to patients with advanced non-small cell lung cancer (aNSCLC) and assessed temporal changes in the distribution of alteration categories using a large, nationwide database of patients with aNSCLC diagnosed from 2017 to 2023.

Y-MATRIX classifies molecular testing results as categories A (first-line option), B (later-line option), C (off-label, guideline-concordant option), D (off-label, guideline-discordant option), E (no actionable alterations), or X (no alterations detected). In a sample of 8,717 patients with aNSCLC (median age, 69 years; 49.2% female), the proportion of patients in category A/B increased from 15.8% in 2017 to 33.6% in 2023, driven predominantly by an increase in later-line therapies (category B, 0% to 15.2%). Patients in category C/D increased from 21.2% to 39.3%, whereas patients in category E/X decreased markedly from 63.1% to 27.1%. Nonsmoking status, younger age, and nonsquamous histology were associated with having category A alterations.

Y-MATRIX provided a comprehensive taxonomy of BGP results by clinical actionability and demonstrates the increasing actionability of BGP results as the aNSCLC treatment landscape evolved over time. There was an increase in patients with category C and D findings, representing a diverse set of clinical circumstances of unclear actionability with opportunities for both innovative and potentially inappropriate targeted therapy use. The Y-MATRIX schema can be applied to facilitate research on all cancers, including the study of alterations across strata of targetable potential and the clinical utility of molecular profiling in precision oncology over time.