P2.09-11 TMB Estimated with Targeted NGS in Early Stage Squamous Cell Carcinoma: Correlation with PD-L1 Expression and Lymphocyte Density

It has been proposed that a combination of assays may refine the prediction of response to checkpoint inhibitors, with tumour mutation burden (TMB) being lately the strongest biomarker associated with efficacy. Although some targeted next generation sequencing (NGS) assays provide a good estimate of whole exome sequencing TMB, they are only available throughout referral laboratories. We sought to investigate the possibility of profiling TMB in-house with a commercially available targeted NGS assay, and to correlate the results with survival, the status of TP53, PD-L1 overexpression and tumor-infiltrating lymphocytes (TILs). The study included samples from 40 patients diagnosed with early-stage lung squamous cell carcinoma. PD-L1 immunohistochemistry (IHC) was performed with two clones (SP263 and SP142, Ventana Medical Systems). CD8+ TILs were scored with a digital algorithm. The status of TP53 (exons 4–10) was investigated with direct sequencing. Ion TorrentTM OncomineTM Tumor Mutation Load Assay (ThermoFisher Scientific), a targeted NGS assay which covers 1.7Mb across 409 cancer driver genes, was used to assess TMB. NGS was performed on genomic DNA from FFPE tumor samples using the Ion S5TM system. The results were analyzed with the Ion ReporterTM software. According to the manufacturer’s instructions, TMB was calculated based on the number of non-synonymous somatic mutations, after removing polymorphisms and known or predicted driver mutations from all the variants. We investigated the correlation between TMB and PD-L1 expression, lymphocyte density, TP53 status and survival. TMB data was available for all 40 patients. The mean and median number of mutations was 13 and 11, respectively. CD8+/mm2 TILs within the peritumoral stromal compartment ranged from 312 to 4793 and within the intraepithelial compartment ranged from 17 to 2002. Sixty percent and 32% of cases were positive in tumor cells with SP263 or SP142, respectively (1% cut-off). Immune cells PD-L1 expression (1% cut-off) was observed in 92% of samples with both clones. No correlation was found between PD-L1 expression or lymphocyte density and TMB. Tumors with TP53 mutations showed non-significant higher numbers of mutations than those wild-type (p=0.075). Accordingly, there was a similar trend for overall survival using the median number of mutations as a cut-off (p=0.137, HR, 5.29). It is feasible to use targeted NGS to estimate TMB. In our pilot study, there was no obvious correlation with other immunooncology predictive biomarkers. Funding: AES 2017/Feder. ISCIII: PI17/01001