ALK testing patterns in early-stage non-small cell lung cancer: A real-world evidence study.

Abstract

8074 Background: Comprehensive biomarker testing in advanced non-small cell lung cancer (aNSCLC) is standard of care, but disparities exist. As new clinical data emerge in early disease (eNSCLC), the set of actionable lung driver mutations (LDMs) and therefore the testing landscape in that setting will change. In particular, following the results of the ALINA trial, therapies that target ALK fusions are anticipated to become part of routine treatment of eNSCLC in 2024.This study thus sought to identify the factors associated with receipt of timely ALK testing in eNSCLC, prior to this shift. Methods: We retrospectively analyzed patients with resected stage I-IIIa NSCLC diagnosed between 2018-2023 using the expanded nationwide (US-based) de-identified Flatiron Health EHR-derived database; data were consolidated from approximately 280 US cancer clinics (≅800 sites of care), using natural language processing with machine learning (ML) to extract patient characteristics. The primary endpoint was ALK testing results (by any modality) within 90 days of initial diagnosis. We used multivariable logistic regression to assess the association between clinico-demographic characteristics and likelihood of testing. Results: The sample contained 21,982 stage I-IIIa patients. Of these, 6,685 (30%) were tested within 90 days, and 15,297 (70%) were not. Patients were more likely to be tested within 90 days if they were diagnosed in more recent years (2023 vs. 2018 odds ratio (OR) 2.12); had no history of smoking (OR 1.21); had highest socioeconomic status (SES) quintile (OR vs. lowest SES 1.14); had more advanced disease (stage IIIa vs. stage I OR 2.43); or were treated in a community setting (OR vs. academic 1.13), all p < 0.05. Patients with Medicare insurance (OR vs. commercial insurance 0.89, p = 0.014) or squamous histology (OR vs. non-squamous 0.46, p < 0.01) were less likely to be tested. There were no significant differences between black and white patients (OR 0.91, p = 013). Conclusions: In this real-world analysis of patients diagnosed between 2018-2023 most patients did not have testing within 90 days of diagnosis. Our data show variation in the baseline probability of testing with respect to clinico-demographic factors. Given that barriers to testing persist in aNSCLC, this suggests that once ALK inhibitors become part of routine care in eNSCLC, some groups might be less likely to be tested and therefore to receive efficacious targeted treatment. Preemptive measures such as education of patients, providers, payers, and policy makers could help prevent disparities from emerging.