Tumor immune microenvironment is influenced by frameshift mutations and tumor mutational burden in gastric cancer.

PURPOSE: Immunoscore can effectively predict prognosis in patients with colon cancer; however, its clinical application is limited. We modified the Immunoscore and created a tumor immune microenvironment (TIM) classification system for gastric carcinoma. Unlike previous studies that used small sample sizes or focused on particular immune-cell subtypes, our simplified system enables pathologists to classify gastric carcinomas intuitively using H&E-stained sections. METHODS: Samples from 326 patients with advanced gastric carcinoma were reviewed and analyzed by pathologists using simple determination and digital image analysis. Comprehensive results of cancer-panel sequencing, Epstein-Barr‒virus (EBV) status, and PD-L1, HER2, ATM, PTEN, MET, FGFR2, and EGFR immunohistochemistry were evaluated with respect to the TIM class. RESULTS: The TIM was classified as "hot" (n = 22), "immunosuppressed" (n = 178), "excluded" (n = 83), or "cold" (n = 43). TIM category was significantly associated with numbers of frameshift mutations (P < 0.001) and high tumor mutational burden (P < 0.004), and predicted overall survival. It was also significantly associated with age, histological type, degree of fibrosis, PD-L1 expression, loss of ATM and PTEN expression (P < 0.001), sex, EBV positivity, and HER2 overexpression (P < 0.04). "Hot" tumors were frequent in PD-L1 expressing and EBV-positive samples, and in those with ATM and PTEN loss. "Excluded" tumors were frequent in HER2-positive cases, whereas "cold" tumors were more frequent in younger patients with poorly cohesive histology and high fibrosis levels. CONCLUSIONS: TIM classification system for gastric carcinoma has prognostic significance and results in classes that are associated with molecular characteristics.

The NanoString-based multigene assay as a novel platform to screen EGFR, HER2, and MET in patients with advanced gastric cancer.

INTRODUCTION: Molecular targets are emerging rapidly and the development of clinical tests that simultaneously screen for multiple targets has become especially important. We assessed the gene expression levels of three known targets in advanced gastric cancer, epidermal growth factor receptor (EGFR), human epidermal growth factor 2 (HER2), and N-methyl-N-nitrosoguanidine human osteosarcoma transforming gene (MET), using the nCounter® assay (NanoString Technologies, Seattle, WA, USA) and compared these results with protein overexpression, detected by immunohistochemistry, to evaluate the performance of this new technology. METHODS: We investigated 42 formalin-fixed, paraffin-embedded tumor samples from patients with gastric cancer. A NanoString-based assay containing a 522 kinase gene panel was investigated. We analyzed the correlations between immunohistochemical findings and kinase gene expression levels of EGFR, HER2 and MET to validate this assay. RESULTS: EGFR, HER2, and MET overexpression were observed in 7 (16.6 %), 5 (11.9 %), and 3 (7.1 %) cases, respectively. For EGFR, HER2, and MET, the concordance rates between the NanoString-based assay results and the immunohistochemistry methods were 83.3, 97.6, and 100 %, respectively. Relative to immunohistochemistry findings, the NanoString-based assay sensitivities and specificities were 85.7 and 82.8 % for EGFR, 100 and 97.2 % for HER2, and 100 and 100 % for MET, respectively. CONCLUSIONS: We found a high concordance between immunohistochemistry- and nCounter-based assessments of EGFR, HER2, and MET in advanced gastric cancer. Judged against immunohistochemistry results, the NanoString assay had high sensitivities and high specificities. These results suggest that the nCounter assay provides a reliable, high-throughput assay to simultaneously screen for the overexpression of several target proteins.