Preoperative prediction of lymph node metastasis from colorectal cancer using computed tomography

Background Preoperative diagnosis of lymph node metastasis from colorectal cancer is crucial; guidance for its diagnosis based on radiological findings is lacking. Aims In this study, we validated the diagnostic accuracy of preoperative computed tomography for lymph node metastasis from colorectal cancer. Methods Patients who underwent surgery for colorectal cancer from January 2015 to August 2020 and from January 2007 to December 2014 were enrolled in the derivation ( n = 250) and validation cohorts ( n = 141), respectively. Computed tomography was performed using a contrast agent and 5-mm-thick sections. The characteristics of the largest lymph nodes detected through computed tomography were estimated. Results In the derivation cohort, receiver operating curve analysis revealed optimal cutoff values of 7.4 and 4.9 mm for the long-axis and short-axis diameters, respectively, based on the large area under the curve values (0.76 and 0.78, respectively). The sensitivity, specificity, and accuracy at a cutoff of 7 mm for the long-axis diameter were 65%, 78%, and 72%, respectively. In the validation cohort, univariate and multivariate analyses revealed that a long-axis diameter ≥ 7 mm was a significant predictor of lymph node metastasis. Conclusions A lymph node long-axis diameter ≥ 7 mm is a predictive factor of lymph node metastasis in patients with colorectal cancer.