Real-world experience and molecular features of response to immune checkpoint blockade in patients with recurrent small cell lung cancer.

Abstract

8556 Background: Immune checkpoint blockade (ICB) is now a routine component of treatment in recurrent small cell lung cancer (SCLC). We evaluated the response to ICB in patients (pts) with recurrent SCLC and genomic features of response using next-generation sequencing (NGS). Methods: Pts with recurrent SCLC treated with ICB were identified. The majority of pts were treated outside of a clinical trial to focus emphasis on the real-world experience. Tumor mutation burden (TMB) and the landscape of somatic variants were determined by targeted NGS using MSK-IMPACT. Objective response rate (ORR) to ICB was determined using RECIST v1.1. PFS and OS were measured from the start of ICB and analyzed using Kaplan-Meier. Results: Between December 2013 and October 2018, 108 pts with SCLC were treated with ICB (57 subjected to NGS). Pts received PD-1 monotherapy alone (n = 28) or in combination with CTLA-4 blockade (n = 80). Median line of therapy was 2 (range 1-6). ORR was 14% (15/108, 95% CI 8-22%). From the start of ICB, median PFS was 1.4 months in non-responders and 10.8 months in responders (HR 0.2; 95% CI 0.13-0.32). Median OS was 6.3 months in non-responders and undefined in responders (range 8-44 months) (HR 0.26, 95% CI 0.16-0.44). Four responders remain on ICB treatment. TMB in the ICB-treated cohort was similar to that of an unselected cohort (n = 233) of SCLC (median 8.8 Mt/MB vs 8.2 Mt/MB, p = 0.71). Clinical benefit was enriched among those with a higher TMB (upper vs middle/lower tertile PFS HR 0.48, 95% CI 0.28-0.84, p = 0.01 and ORR 26% [5/19] vs ORR 8% [3/38]). Rates of whole genome duplication and commonly altered genes in SCLC ( TP53, RB1, KMT2C/D, NOTCH1/2/4, PTPRD, APC) were similarly distributed across responders and non-responders. Completion of whole-exome sequencing and PD-L1 testing is in progress. Conclusions: In pts with recurrent SCLC receiving routine clinical care, the ORR to ICB is comparable to reports from clinical trials. A high TMB was associated with a longer median PFS and better response. Further investigation into the genomic landscape of recurrent SCLC is needed to identify biomarkers predictive of response to ICB.