Molecular-Like Classification in Gastric Adenocarcinomas: Correlation with Clinicopathological Features and Prognosis

<b>Aim:</b> Molecular classifications in gastric cancer are expensive and require experience. This study aimed to establish a molecular-like classification and to investigate the correlation of the classification with clinicopathological features and its prognostic significance. <b>Materials-Methods:</b> In 202 gastric cancer patients who underwent surgery between 2011 and 2019, the clinicopathological features of the tumors were re-evaluated, and the tumors were grouped by molecular-like classification using immunohistochemistry (MLH1, PMS2, MSH2, MSH6, p53, E-cadherin) and chromosomal in situ hybridization (EBER): EBV positive tumors (EBV+), tumors with microsatellite instability (MSI), genomically stable tumors (GS), chromosomal unstable tumors (CUN) and unclassifiable gastric tumors (G-NOS). <b>Results:</b> Twelve of the patients were in the EBV+, 22 in the MSI, 27 in the GS, 76 in the CUN and 65 in the G-NOS group. EBV + tumors were associated with high-grade tubular/papillary morphology and GS tumors were associated with poorly cohesive/mixed type, while tumors in the MSI group were associated with low-grade tubular/papillary morphology (p < .0001). Marked tumor budding was the least in MSI tumors and the most in GS tumors (p = .018). The longest survival time was in the patients of EBV + group, while the shortest survival time was in the patients of GS group (p = .029). Cox regression analysis indicated that age, lymphovascular invasion, positive surgical margins, low-grade tubular and papillary histological types, two groups within molecular- like classification (MSI and GS) were independent prognostic factors (p < .05). <b>Conclusion:</b> Molecular-like classification holds promise as a low-cost, easy-to-implement classification that can be used predictively in prognosis. However, molecular-based studies with large series are still needed to determine whether it is a pretest or alternative of molecular classifications.