Targeted Next Generation Sequencing Identifies Markers of Response to PD-1 Blockade

Abstract

Therapeutic antibodies blocking programmed death-1 and its ligand (PD-1/PD-L1) induce durable responses in a substantial fraction of melanoma patients. We sought to determine whether the number and/or type of mutations identified using a next-generation sequencing (NGS) panel available in the clinic was correlated with response to anti-PD-1 in melanoma. Using archival melanoma samples from anti-PD-1/PD-L1-treated patients, we performed hybrid capture-based NGS on 236-315 genes and T-cell receptor (TCR) sequencing on initial and validation cohorts from two centers. Patients who responded to anti-PD-1/PD-L1 had higher mutational loads in an initial cohort (median, 45.6 vs. 3.9 mutations/MB; P = 0.003) and a validation cohort (37.1 vs. 12.8 mutations/MB; P = 0.002) compared with nonresponders. Response rate, progression-free survival, and overall survival were superior in the high, compared with intermediate and low, mutation load groups. Melanomas with NF1 mutations harbored high mutational loads (median, 62.7 mutations/MB) and high response rates (74%), whereas BRAF/NRAS/NF1 wild-type melanomas had a lower mutational load. In these archival samples, TCR clonality did not predict response. Mutation numbers in the 315 genes in the NGS platform strongly correlated with those detected by whole-exome sequencing in The Cancer Genome Atlas samples, but was not associated with survival. In conclusion, mutational load, as determined by an NGS platform available in the clinic, effectively stratified patients by likelihood of response. This approach may provide a clinically feasible predictor of response to anti-PD-1/PD-L1. Cancer Immunol Res; 4(11); 959-67. ©2016 AACR.

Figure 1

Mutational load in responders vs. nonresponders in initial cohort (A) and validation cohort (B). Progression-free survival (C) and overall survival (D) in patients with high, intermediate, and low mutational load.

Figure 2

Genetic alterations observed in responders vs. nonresponders: (A) Total number of mutations; (B) total number of C>T tranversions; (C) types of nucleotide substitutions; (D) mutational load of patients with BRAF mutations, NRAS mutations, NF1 mutations/loss, and “triple WT” (defined as wild-type for BRAF, NRAS, and NF1). BRAF non-V600 mutations were included with in the BRAF cohort except for one patient with concurrent NF1 mutation. One patient with NRAS^Q61R mutation and concurrent NF1 mutation was included in the NRAS cohort.

Figure 3

(A) Mutational load in TCGA skin cutaneous melanoma (SKCM) samples using 315 genes included on hybrid capture NGS panel is highly correlated with mutations assessed by whole exome sequencing. Mutational load groups and survival in the TCGA using (B) WES and (C) 315 FoundationOne (FM) genes. For low, intermediate, and high mutation load groups, we observed a difference in OS using WES (median OS 43.4 months vs. 103.0 months vs. 68.0 months, P = 0.001), and FM genes (median 47.3 vs. 112.5 months vs. 61.5 months, P = 0.008).

Figure 4

Top panel: calculated mutational load per sample. Middle panel: color-coded matrix of individual mutations, copy number alterations, and clinical characteristics. Bottom panel: mutation spectra of individual samples.