Nomogram Model Based on MRI and Clinical Characteristics for Predicting Lymph Node Metastasis in Rectal Cancer

Objective To explore the diagnostic value of a predictive model that combines MRI features and clinical features for detecting lymph node metastasis (LNM) in patients with rectal cancer (RC). Methods A retrospective analysis was conducted on 170 patients who had been pathologically diagnosed with RC by a pathologist. Of these, 74 were in the LNM group and 96 were in the non-metastatic group. The relationship between LNM and clinical and MRI features was analyzed using univariate and multivariate binary logistic regression. Based on these results, a clinical prediction model was constructed, and its diagnostic effectiveness was analyzed using a receiver operating characteristic (ROC) curve. Results Univariate binary logistic regression analysis showed statistically significant differences in MR T-stage, CEA, CA199, lymphovascular invasion (LVI), perineural invasion (PNI) and circumferential resection margin (CRM) between the LNM group and the non-metastasis group. Multivariate regression analysis revealed that CEA levels greater than 5 ng/mL, positive LVI and positive CRM were independent risk factors for LNM in RC. ROC curve analysis revealed an area under the curve (AUC) of 0.850 (95% CI: 0.791–0.909) for the model assessing LNM in RC. Conclusions The clinical prediction model, which was constructed using a combination of MRI and clinical features (including CEA, LVI and CRM), has high diagnostic value.