Development and validation of individualized prognostic nomograms based on the tumour‐stroma percentage in patients with proximal and distal gastric cancer: A population‐based study
Yueling Wang, Yan Yang, Liao Zhang, Yitian Xu, Renchao Zhang, Yuan Fang, Chao Han, Chen HuangAbstract
Background
Gastric cancers (GCs) arising in the proximal and distal stomach are increasingly recognized as biologically and clinically heterogeneous subtypes of GC. However, individualized prognostic tools specifically designed for these two tumour locations remain insufficient.
Methods
The clinical records, pathological findings, laboratory indicators, and follow‐up data from 813 patients with GC were retrospectively analysed. Survival‐related variables were examined separately in the proximal GC (PGC) and distal GC (DGC) cohorts using Cox proportional hazards modelling. Factors that remained relevant in the final models were integrated into two tumour location‐based nomograms for overall survival (OS) estimation. The models were examined using receiver operating characteristic (ROC) curves, Harrell's concordance index ( C ‐index), calibration plots and decision curve to assess discrimination, agreement and clinical usefulness.
Results
In patients with PGC, the final model retained tumour‐stroma percentage (TSP) (hazard ratio [HR] = 2.485, 95% confidence interval [CI] 1.258–4.908, p = .009), vascular invasion (HR = 1.911, 95% CI 1.018–3.587, p = .044) and pathological TNM stage (pTNM) (HR = 1.943, 95% CI 1.093–3.453, p = .024) regard as survival‐related predictors. For DGC, nerve invasion (HR = 2.174, 95% CI 1.227–3.851, p = .005) and pTNM stage (HR = 2.070, 95% CI 1.373–4.177, p = .004) were retained in the final model. The location‐specific nomograms demonstrated favourable discriminatory ability for predicting 5‐year OS, with AUC values of 0.924 for PGC and 0.861 for DGC.
Conclusions
TSP was associated with survival outcomes in GC and showed particular prognostic relevance in PGC. Developing separate prognostic nomograms for PGC and DGC may provide a practical approach for improving individualized clinical management.