RESUMO
BACKGROUND: The indications for splenic hilar lymph node dissection in advanced proximal gastric cancer without invasion of the greater curvature are controversial. We aimed to develop a preoperative nomogram for individualized prediction of splenic hilar lymph node metastasis in non-greater curvature advanced proximal gastric cancer. METHODS: From January 2014 to December 2021, 558 patients with non-greater curvature advanced proximal gastric cancer who underwent D2 lymphadenectomy (including splenic hilar lymph node) were retrospectively analyzed and divided into a training cohort (n = 361) and validation cohort (n = 197), depending on the admission time. A preoperative predictive nomogram of splenic hilar lymph node metastasis was established based on independent predictors identified by multivariate analysis, and the performance and prognostic value were confirmed. RESULTS: In the training and validation cohorts, 48 (13.3%) and 24 patients (12.2%) had pathologically confirmed splenic hilar lymph node metastasis, respectively. Tumor located in the posterior wall, tumor size ≥5 cm, Borrmann type IV, and splenic hilar lymph node lymphadenectasis on computed tomography were preoperative factors independently associated with splenic hilar lymph node metastasis. The nomogram developed based on these four parameters had a high concordance index of 0.850 (95% confidence interval, 0.793-0.907) and 0.825 (95% confidence interval, 0.743-0.908) in the training and validation cohorts, respectively, with well-fitting calibration plots and better net benefits in the decision curve analysis. In addition, disease-free survival and overall survival were significantly shorter in the high-risk group, with hazard ratios of 3.660 (95% confidence interval, 2.228-6.011; log-rank P < .0001) and 3.769 (95% confidence interval, 2.279-6.231; log-rank P < .0001), respectively. CONCLUSION: The nomogram has good performance in predicting the risk of splenic hilar lymph node metastasis in non-greater curvature advanced proximal gastric cancer preoperatively, which can help surgeons make rational clinical decisions.
