Incidence and pathology-based predictive modeling of lymph node metastasis in early gastric cancer: evidence from a retrospective single-center Chinese surgical cohort with internal validation
Year: 2026
Type: article
Study design: Observational Study
Abstract: Background: Lymph node metastasis (LNM) is a critical determinant of prognosis and treatment in early gastric cancer (EGC). Current endoscopic submucosal dissection (ESD) indications, largely derived from Japanese guidelines, remain controversial in Chinese populations given reported differences in LNM rates and tumor heterogeneity. This study aimed to comprehensively evaluate LNM risk and refine stratification strategies in a large Chinese cohort. Methods: A retrospective analysis of 918 EGC patients undergoing gastrectomy with systematic lymph node dissection was performed. Clinicopathological variables were assessed, and independent risk factors for LNM were identified using multivariable logistic regression. Subgroup analyses were conducted according to invasion depth and pathological characteristics. A pathology-based nomogram was developed and internally validated. Results: The overall observed pathological LNM rate was 14.3% (131/918). Independent predictors included maximal tumor diameter, undifferentiated histology, mixed histological composition, T1b invasion, ulceration, and lymphovascular invasion. T1a cancers showed an LNM rate of 8.3%, while no LNM was observed in differentiated tumors ≤30 mm [0/170; 95% confidence interval (CI): 0-2.15%]. Differentiated tumors without ulceration showed a low LNM rate of 1.4% (2/148). In T1b cancers, the LNM rate increased to 21.2%. However, submucosal invasion <500 μm (SM1) differentiated tumors without lymphovascular invasion exhibited a relatively low observed LNM rate of 3.9%, with no metastasis observed in lesions ≤20 mm (0/33; 95% CI: 0-10.58%). A nomogram incorporating 6 independent variables demonstrated good discrimination [area under the curve (AUC) 0.810, 95% CI: 0.768-0.851], with an optimism-corrected AUC of 0.801 after 1,000 bootstrap resamples, excellent calibration, and favorable decision-curve performance. The nomogram may support individualized post-resection pathological LNM risk assessment. Conclusions: LNM risk remains substantial in this surgically selected Chinese EGC cohort, particularly in T1b and undifferentiated tumors. These findings highlight the need for caution when interpreting pathological LNM risk and support more individualized post-resection risk stratification. Prospective validation is required before these findings can be used to modify ESD indications.
Language: English
Subfield: Oncology
Field: Medicine
Domain: Health Sciences
Open Access status: diamond
OpenAlex ID: W7224118892