Related Experiment Video
Updated: Sep 13, 2026
Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model
Published on: April 18, 2025
The coefficient of variation of CT values for predicting lymph node metastasis in extrahepatic cholangiocarcinoma
Naokazu Chiba1, Yuta Abe2, Hirotaka Kojima1
1Department of Digestive and Transplantation Surgery, Tokyo Medical University Hachioji Medical Center, Tokyo, Japan.
Abstract:
Preoperative prediction of lymph node metastasis (LNM) in extrahepatic cholangiocarcinoma (ECC) is crucial for surgical strategies and prognosis, as current imaging methods are insufficient. A total of 149 ECC patients who underwent curative resection were retrospectively analyzed. We measured the Coefficient of Variation (CV) of CT values, ADC of MRI, and SUVmax of PET/CT from multiple lymph nodes. Diagnostic performance for LNM prediction was evaluated using ROC curve analysis. The CV of CT values was the most powerful independent predictor of LNM. Its AUC (0.829) was superior to MRI ADC (0.666) and PET/CT SUVmax (0.701). The CV of CT values showed a sensitivity of 72.8%, specificity of 72.9%, and accuracy of 72.9%. In conclusion, The CV of CT values is a promising and effective parameter for preoperatively predicting LNM in ECC. This can potentially improve patient selection for neoadjuvant therapies and surgical planning.
