Answers are grounded in this paper and cite their sources.
Aim To assess the diagnostic performance of radiomics-based models for prediction of perineural invasion (PNI) in gastric cancer (GC) patients before surgery. Method We performed a comprehensive search up to December 5, 2025, across PubMed, Embase, Web of Science, and the Cochrane Library to identify studies assessing the diagnostic performance of radiomics in PNI prediction in GC. Risk of bias and quality assessment were performed using the QUADAS-2 and METRICS. Using a random-effects bivariate model, diagnostic pooled estimates were calculated. Meta-regression, leave-one-out sensitivity analyses, publication bias tests, and GRADE were also conducted. Results Eleven studies and 3328 patients were included. The four analysis domains were the internal validation radiomics-only cohort (n = 976), internal validation combined (radiomics + clinicopathological) cohort (n = 908), external validation radiomic-only cohort (n = 463), and external validation combined cohort (n = 356). Across internal validation sets, radiomics-only models demonstrated good diagnostic performance, with a pooled AUC of 0.85 (95% CI: 0.82-0.88). Combined models achieved AUC of 0.87 (95% CI: 0.84-0.90). Performance remained stable in external validation cohorts (both AUCs: 0.81). Minimal statistical heterogeneity was observed, but clinical heterogeneity remained substantial. Deeks' tests were non-significant and sensitivity analysis revealed no influential individual studies. Certainty of evidence was very low to low. Conclusion Radiomics models showed good performance for preoperative prediction of PNI in GC patients. With greater protocol standardization and the inclusion of multicenter, prospective, and geographically diverse populations, both radiomics-only and combined models may show potential for preoperative risk stratification in GC patients.
No takes yet. Share an insight, caveat, or question.
Mohammadzadeh et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: