OBJECTIVES : Initiating treatment without molecular profiling results may lead to metastatic NSCLC patients receiving therapies that provide suboptimal benefit. There is limited literature comparing patients with results prior to treatment initiation and received recommended therapy vs. those initiating treatment without results or did not receive recommended therapy.

METHODS : This retrospective analysis identified patients (≥18 years) in the ConcertAI Oncology Dataset initiating first line (1L) treatment for non-squamous mNSCLC diagnosed 01/01/2015 or later. Baseline demographic and clinical characteristics and progression-free survival (PFS) were compared between patients with ≥1 test result (EGFR, ALK, ROS1, BRAF, or NTRK) prior to 1L and received recommended therapy (cohort A) and patients without results prior or did not receive recommended therapy (cohort B). Recommended therapy was targeted therapy (with associated positive result), immunotherapy (EGFR- or ALK- prior to 1L), or chemotherapy (EGFR- or ALK- prior to 1L and PD-L1- prior to or during 1L). Kaplan-Meier methods and Cox regression analysis, controlling for baseline characteristics, were used to evaluate PFS from start of 1L treatment.

RESULTS : Of 238 patients, 94 (39.5%) were in cohort A. Overall, 60.9% were ≥65 years, 51.3% male, and 77.7% white. Characteristics were similar by cohort. Cohort A had less stage IV at initial diagnosis (79.8% vs. 96.5%; p=0.001) and longer median time from mNSCLC diagnosis to 1L (36.5 vs. 27.5 days). Median PFS was 12.06 months (95% CI: 7.5-16.2) for cohort A and 5.06 months (95% CI: 4.1-7.1) for cohort B (p=0.023). Controlling for characteristics, cohort A had a lower risk of disease progression (HR=0.665; p=0.023).

CONCLUSIONS : Patients with mNSCLC may benefit from waiting for results prior to initiating 1L treatment. Those with ≥1 result prior and received recommended therapy had a longer median PFS and lower risk of progression. This study supports the current guidelines recommending molecular profiling at diagnosis.