Molecular subtypes and clinical outcomes to initial systemic treatment in patients with small cell lung cancer.

Abstract

Listen

Translate article

9018 Background: Investigators have proposed that differential expression of the transcription regulators ASCL1 and NeuroD1 can be used to define molecular subtypes of small cell lung cancers (SCLCs). Here we evaluate SCLC subtypes based on ASCL1 and NeuroD1 expression in patients (pts) treated with first-line (1L) chemotherapy profiled with targeted next-generation sequencing (NGS). Methods: We used NGS (MSK-IMPACT) to profile tumors from pts with SCLCs. We performed IHC to assess ASCL1 (A) or NeuroD1 (N). Objective response rate (ORR) to therapy was determined using RECIST. PFS and OS were analyzed using Kaplan-Meier. Results: 281 pts with SCLCs were profiled with NGS (102 LS-SCLC; 179 ES-SCLC). Most frequently mutated genes were TP53 (90%), RB1 (68%), KMT2D (22%), NOTCH1 (15%), FAT1 (14%), PTPRD (12%). Mutations in BIRC3, FOXL2, TENT5C, TET1, NRAS, KIT, TSHR, ESR1 were enriched in ASCL1-/NeuroD1+ (A-/N+), and mutations in KMT2D and EP300 were enriched in A-/N- (p<0.05). Copy number alterations in WWTR1, ATR, IKZF1, PALB2, PIK3CB were enriched in A-/N+ (p<0.05). IHC for ASCL1 and NeuroD1 was performed on 78 samples: 11 A-/N-, 32 A+/N-, 4 A-/N+, 31 A+/N+. Overall survival at 1 year based on subtype was 25% in A-/N- (2/9), 60% in A-/N+ or A+/N- (13/32), and 55% in A+/N+ (10/25). For the 10 pts who survived 2 years, 5 were A+/N- and 5 were A+/N+. 146 pts treated with 1L platinum had RECIST-evaluable disease. ORR was 75% (110/146; 95% CI 68-82%). Median PFS was 7 months with CR/PR and 3.5 months with SD/PD (HR 0.32; 95% CI 0.18-0.56). Median OS was 17 months with CR/PR and 11 months with SD/PD (HR 0.55; 95% CI 0.34-0.9). Mutations in RUNX1, EPHA7, CDKN2A, FLT1 and copy number alterations in FGFR1, CCND1 were enriched in patients with SD/PD (p<0.05). PFS rate at 6 months was 25% in A-N- (1/4), 60% in A-/N+ or A+/N- (9/15), and 55% in A+/N+ (6/11). For the 7 pts who survived 2 years, 3 were A+/N- and 4 were A+/N+. Conclusions: Molecular subtypes defined by ASCL1 and NeuroD1 encompass molecular characteristics that may predict patient outcomes. Further investigation is needed to delineate the underlying biological differences among the various subtypes to help define therapeutic vulnerabilities of each subtype of SCLC. Completion of IHC for ASCL1, NeuroD1 and additional key transcription factors POU2F3 and YAP1 are in progress for the entire cohort. WES and RNA sequencing are occurring in parallel and will be correlated with IHC results and clinical outcomes.