Authors:
Demetri, GD, Peters, S, Hibbar, DP, Maund, SL, Veronese, L, Liu, H, Humblet, O, Perez, L
Background
NTRK fusions are actionable biomarkers with efficient and well tolerated TRKis. Clinical profile and outcomes of pts with NTRK+ tumours receiving non-TRKi SoC are unknown and the prognosis of NTRK+ pts vs pts with NTRK fusion-negative (NTRK–) tumours is not well known.
Methods
Demographic and clinical data for TRKi-naïve adults with metastatic/LA solid tumours and ≥1 Foundation Medicine NGS test (1 Jan 2011 – 31 Dec 2019) were extracted from a US electronic health record-derived clinicogenomic database (CGDB; Flatiron Health). NTRK– pts from the CGDB were matched 10:1 to NTRK+ pts based on cancer type and propensity score with preselected prognostic variables (age; smoking status; practice type; lines of therapy from initial diagnosis to NGS report; stage at diagnosis; time from metastatic/LA diagnosis to NGS report; co-mutations). Overall survival (OS), defined as time from index date (metastatic/LA stage diagnosis or start of last treatment) to death/censoring, was compared between NTRK+ and matched NTRK– cohorts. NTRK fusion prognostic value was evaluated via univariate Cox proportional hazard model.
Results
Of 58001 CGDB pts with solid tumours, 28 had NTRK+ metastatic/LA cancers. NTRK+ pts tended to be younger, with less history of smoking, more brain metastases and a shorter time from advanced diagnosis to first NGS report vs all NTRK– pts; no differences were significant (Table). Median (95% CI) OS was 10.2 months (7.2–14.1) in NTRK+ pts vs 10.4 months (6.7–14.3) in matched NTRK– pts (n=280); hazard ratio (HR; 95% CI) for death was 1.6 (1.0–2.5). Results were similar using start of last treatment as index (HR 1.6; 1.0–2.5).
Sources:
ESMO Annual Congress