Real-world pan-cancer landscape of frameshift mutations (FSM) and their role in predicting responses to immune checkpoint inhibitors (ICI) in patients (pts) with tumors with low tumor mutational burden (TMB).

Abstract

2599 Background: Pembrolizumab was recently approved in tumors with TMB ≥10 mut/Mb. FSM can complement TMB in predicting ICI responses. We obtained a real-world dataset of genomic alterations from 250,813 samples to examine the distribution of TMB and FSM across a variety of malignancies. We then conducted a multi-institutional retrospective review of pts treated with ICI. Methods: Database samples were sequenced by Foundation Medicine using hybrid capture genomic profiling to evaluate all classes of genomic alterations in at least 315 genes. The clinical cohort included pts with metastatic solid malignancies who received ICI and had undergone commercial next-generation sequencing (NGS). Pts were classified into four distinct groups: TMB-L ( < 10mut/Mb)/ FS-A (absent FSM), TMB-H (≥10mut/Mb)/ FS-A, TMB-L /FS-P (present, ≥1 FSM) and TMB-H/FS-P. Progression-free survival (PFS), overall survival (OS), and response rate (RR) were compared between the groups. Results: 246,252 MSS and 4,561 MSI-High samples were segregated by histology and divided into four distinct groups based on the TMB and FSM. For the MSS cohort the distribution was: TMB-L/FS-A (N = 111,065, 45%), TMB-H/FS-A(N = 15,313, 6%), TMB-L /FS-P (N = 98,389, 40%) and TMB-H/FS-P (N = 21,485, 9%). In the ICI-treated clinical cohort, there were 230 pts in 12 histology groups; 212 had information on TMB and FSM. The most common primary sites were GI (N = 39), melanoma (N = 37), GU (N = 32) and H&N cancer (N = 21). 159 pts received single ICI and 53 dual ICI. 196 tumors were MSS, 11 MSI, and 5 unknown. Group distribution: TMB-L/FS-A 80 pts (38%), TMB-L/FS-P 57pts (27%), TMB-H/FS-A 36pts (17%), TMB-H/FS-P 39pts (18%). FS-P was associated with higher RR 23.81 vs. 12.8 % (p = 0.02). Regardless of TMB, the median PFS for FS-P vs. FS-A was 7.9 and 4.0 mo, respectively (p < 0.01). TMB-L/FS-P had superior PFS (5.1 mo) compared to TMB-L/FS-A (3.6 mo) group (p < 0.01). The 15-month PFS probability was 12% for TMB-L/FS-A vs. 38% for TMB-L/FS-P. No statistically significant difference was detected in OS between the groups. From the pan-cancer cohort, histologies with more than 40% of samples in the TBM-L/FS-P (MSS) group were: CRC, RCC, PDAC, biliary, breast, esophageal, and endometrial cancers. Additional genomic data will be presented. Conclusions: FSM are frequently found on commercial NGS testing in tumors that are MSS and TMB-L. The presence of FSM may complement TMB in predicting benefit from immunotherapy. If validated in additional cohorts, FSM presence could be utilized to identify pts that may benefit from ICI, particularly for tumors with low TMB.

Association of Baseline Tumor-Specific Neoantigens and CD8+ T-Cell Infiltration With Immune-Related Adverse Events Secondary to Immune Checkpoint Inhibitors

“…47,53 Furthermore, we have previously shown that INDELs can complement TMB in predicting outcomes to ICI, especially in microsatellite stable/TMB-low tumors. 54 Similarly, our previous work, alongside other studies, highlighted the potential predictive value of TILs in identifying both response and survival benefit. 44,55,56 This study serves as a proof of concept but is compounded by multiple limitations.…”

Immunosurveillance in clinical cancer management

“…That said, TMB testing does not account for the immunological alterations imposed by genetic defects other than single nucleotide mutations, such as indels and frameshift mutations. Based on a recent pan‐cancer analysis, frameshift mutations may indeed predict ICI sensitivity in patients with solid tumors bearing a low TMB 43 . Hence, the current approach to measure tumor immunogenicity based on genetic alterations in cancer cells has substantial room for improvement.…”