Spectral Ct-Based Non-Invasive Prediction of Neoadjuvant Chemotherapy Response and Postoperative Lymph Node Metastasis in Esophageal Squamous Cell Carcinoma

Background Esophageal squamous cell carcinoma (ESCC) is a highly aggressive malignancy, and neoadjuvant chemotherapy (nCRT) is the standard treatment for locally advanced cases. However, nCRT response varies significantly among individuals, and postoperative lymph node metastasis (LNM) remains a crucial prognostic factor. Traditional assessment methods rely on postoperative pathology, which is invasive and time-consuming. Therefore, there is an urgent need for a non-invasive predictive tool to evaluate nCRT response and LNM risk before surgery. Methods We enrolled 300 ESCC patients undergoing nCRT. Preoperative spectral CT was used to obtain iodine concentration (IC), effective atomic number (Zeff), and extracellular volume (ECV), combined with clinical factors such as NLR and tumor staging. Patients were classified based on nCRT response and LNM status. Independent predictors were identified using logistic regression. A nomogram was developed with machine learning and validated in a training and test set. Model performance was assessed with ROC curves, AUC, and calibration tests. Results Multivariate analysis identified venous phase Zeff (OR = 2.85, p < 0.001), normalized IC (nIC) (OR = 2.20, p = 0.002), NLR (OR = 0.65, p = 0.015), and clinical staging (OR = 0.18, p < 0.001) as independent predictors of nCRT response. For LNM prediction, venous phase IC (OR = 1.95, p = 0.008) and tumor size (OR = 1.78, p = 0.012) were significantly associated. The nomogram achieved AUCs of 0.832 (95% CI: 0.768–0.896) in the training set and 0.801 (0.712–0.890) in the validation set, significantly outperforming traditional imaging methods (DeLong test, p < 0.05). Conclusion This study developed a novel non-invasive predictive model by integrating spectral CT quantitative parameters with clinical characteristics, significantly improving the accuracy of preoperative evaluation. The multi-energy imaging properties of spectral CT provide valuable quantitative information that better reflects tumor microstructural changes and biological behavior. The nomogram is an intuitive and practical tool for clinicians, enabling personalized treatment strategies while reducing ineffective therapies and recurrence risks. These findings fill a critical gap in predicting nCRT response and LNM in ESCC and promote the broader clinical application of spectral CT in oncology.