Abstract
Background Lymph node status is a crucial prognostic determinant in gastric cancer (GC). Conventional haematoxylin-eosin (H&E) assessment overlooks the micro volume of nodal disease. This study explored the prevalence and prognostic significance of lymph-node micrometastases (Mic) and isolated tumour cells (ITC) detected by cytokeratin AE1/AE3 immunohistochemistry (IHC) in patients undergoing gastrectomy with curative intent. Methods A retrospective, single-centre analysis of patients with gastric adenocarcinoma treated between 2014 and 2019 by curative intent gastrectomy with D2 lymphadenectomy and with IHC assessment of lymph nodes (LN). Nodal involvement was classified as macrometastases (>2 mm), Mic (>0.2-2 mm), or ITC (≤0.2 mm). Clinicopathological characteristics and survival outcomes were obtained from the institutional database. Data were analysed using χ²/Fisher's exact and Wilcoxon tests. Overall survival (OS) was analysed using Kaplan-Meier curves and log-rank tests. Results Eighty-six patients were included (mean age 67.2±11.2 years). A total of 1,974 lymph nodes were examined (mean 22.95±7.43 per patient): 256 (12.97%) contained macrometastases, 21 (1.06%) Mic and 31 (1.57%) ITC. Sixteen patients (18.6%) had Mic and/or ITC (ITC+/Mic+). Tumours larger than 20 mm with lymphovascular invasion were significantly associated with the presence of ITC+/Mic+. In node-negative (pN0) patients, IHC uncovered occult nodal disease in six cases (13.6%). In the overall cohort, five-year OS was 55.8%, and the presence of ITC+/Mic+ did not significantly affect OS across tumour, nodes, metastases (TNM) and pathological lymph node (pN) categories. In the subgroup treated with perioperative chemotherapy, the presence of ITC+/Mic+ was associated with a lower five-year OS (p=0.023). OS was comparable between pN0 patients with ITC+/Mic+ and those with node-positive (pN1) disease (p=0.624). Conclusion Micrometastases and ITC are detected in a relevant proportion of gastric cancer patients and are associated with adverse pathological features, but they did not independently influence OS in this clinical cohort. Their prognostic impact might be relevant in patients receiving perioperative chemotherapy. Routine IHC assessment is not sustained by these data and may be better suited for selected high-risk patients.