Pretreatment prediction of neoadjuvant therapy response in locally advanced gastric cancer using an additive ctSLEP score and spectral CT

Background To develop and validate a novel, noninvasive CT-based invasion score (ctSLEP), integrating serosal invasion, lymphovascular invasion, extramural vascular invasion, and occult peritoneal metastasis, combined with spectral CT parameters, for preoperative prediction of neoadjuvant therapy response in locally advanced gastric cancer. Methods In this retrospective study, patients with locally advanced gastric cancer undergoing neoadjuvant therapy followed by surgery (November 2020–November 2025) were included. The ctSLEP score was constructed from CT-detected tumor invasion features. Spectral CT parameters and clinical variables were analyzed. Tumor regression grade by Becker criteria served as the reference standard. Multivariable logistic regression was applied to develop predictive models in training (70%) and validation (30%) cohorts. Results A total of 158 patients (mean age, 57.35 ± 7.89 years; 126 men) were evaluated; 65/158 (41.1%) were responders. Independent predictors of response included neutrophil-to-lymphocyte ratio (OR = 1.680; 95% CI: 1.160–2.567; p = 0.01), iodine concentration in the venous phase (OR = 0.853; 95% CI: 0.731–0.987; p = 0.04), arterial enhancement fraction (OR = 0.195; 95% CI: 0.044–0.812; p = 0.03), and components of the ctSLEP score: extramural vascular invasion ( p < 0.001), occult peritoneal metastasis ( p = 0.03), and lymphovascular invasion ( p = 0.02). The final model incorporating ctSLEP achieved an AUC of 0.880. Conclusions The ctSLEP score, combined with spectral CT parameters and inflammatory markers, may provide a noninvasive tool for preoperative risk stratification and treatment-response prediction in patients with locally advanced gastric cancer.