Abstract 3170: Different genomic mutations signatures are associated to specific PD-L1/TMB states on lung cancer with potential value for patients screening for immunotherapy

Abstract

Abstract Objective: PD-L1 expression and Tumor Mutation Burden (TMB) have independently emerged as prospective biomarkers of response to anti PD1-/PDL1 checkpoint inhibitors. However, TMB has not fully proven its value as a biomarker of Immunotherapy response in lung cancer. Moreover, FDA-approved CDx PD-L1 expression alone is not an optimal biomarker for checkpoint inhibitors and combined use of MSI, TMB and PD-L1 protein levels has been proposed. Here we present the correlation between genomic landscape, including TMB and MSI with PDL1 IHC in lung cancer specimens to help identify immunogenomic profiles for stratification of patients for check-point inhibitors therapies evaluation. Methods: 874 FFPE clinical samples across cancer types were characterized in our CLIA/CAP accredited clinical laboratory using a CLIA-validated NGS-based assay that interrogates SNVs, indels, TMB and Microsatellite Status (27 MS markers) using a 323 gene panel. TMB (mutations/Mb) is categorized as low (≤7), intermediate (7<TMB≤15) and high: (greater than 15). The de-identified aggregated results paired with PD-L1 IHC data from 424 lung cancer samples were analyzed and correlations between PD-L1 tumor proportion scores (TPS) and TMB results were made. In silico analyses were also performed on 5939 lung cancer samples from public databases. Results: The detected TMB median in lung cancer was 8.6 (range 0-126.4). Median PD-L1 TPS was 7%, with 26% of samples being PD-L1 negative (TPS<1%), 44% Low expressing (≥1-49%) and 31% High (≥50%). All samples were MSS. We found poor correlation between PD-L1 expression and TMB in NSCLC (r2=0.266). We classified samples based on TMB and PD-L1 TPS and found mutational correlations specific to in each of the following groups: PD-L1 High/TMB Low, PD-L1 High/TMB High, PD-L1 Low/TMB High, PD-L1 High/TMB High, and PD-L1 Neg/TMB High. Of note 67% (34/51) of PD-L1 High/TMB Low samples presented mutations either on EGFR (12%), KRAS (23.5%) or in genes from known driver TRK/MAPK pathways, whereas only KRAS was part of the frequently mutated gene signature with 36.5% (13/36) samples mutated on PD-L1 High/TMB High samples. Neither EGFR nor KRAS were found frequently mutated on PD-L1 Low/TMB High group (n=46). Strikingly the top 12 most frequently mutated genes on PD-L1 Neg/TMB High tumors were TP53, LRP1B, SPTA1, SMARCA4, GNAS, ALK, FGFR2, SLIT2, ROS1, AMER1, FAT1, and MED12. We found statistically significant co-occurrences between LRP1B and 7 other genes from this signature in additional 5939 lung cancer patients assessed. Conclusions: TMB and PD-L1 results do not correlate in lung cancer specimens. Common genomic alteration signatures might define subsets of lung cancer tumors with no PD-L1 expression to complement TMB and PD-L1 on the selection criteria for patients whom may benefit from checkpoint inhibitors. Citation Format: Fernando J. Lopez-Diaz, Lauryn Keeler, Forrest Blocker, Shiloh Sprankles, Ryan Bender, Sally Agersborg, Lawrence Weiss, Vincent Funari. Different genomic mutations signatures are associated to specific PD-L1/TMB states on lung cancer with potential value for patients screening for immunotherapy [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 3170.