Related Experiment Video
Updated: May 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
Comparative performance of 3 lymph node staging schemes in duodenal adenocarcinoma: A retrospective analysis of
Zhenxing Jiang1, Jianwei Dai1, Wei Li2
1Department of Gastrointestinal Surgery, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China.
Background:
The comparative prognostic performance of lymph node-related prognostic parameters-positive lymph nodes, log odds of positive lymph nodes, and lymph node ratio-in duodenal adenocarcinoma remains undefined. This study aims to compare their predictive performance in both Western and Eastern cohorts and develops a prognostic nomogram based on the optimal indicator.
Methods:
Data were obtained from the Surveillance, Epidemiology, and End Results (N = 1,693) database and Wuhan Union Hospital (N = 295). Using 7 evaluation indicators, this study assessed the prognostic performance of 3 lymph node-related prognostic parameters in both their continuous and categorical form. A nomogram was constructed based on the optimal lymph node-related prognostic parameter and independent prognostic factors. External validation was carried out using the Wuhan Union Hospital cohort.
Results:
Among continuous lymph node-related prognostic parameters, standardized log odds of positive lymph nodes demonstrated superior predictive accuracy in both Surveillance, Epidemiology, and End Results cohort (c-index: 0.639; 5-year area under the curve: 0.685) and Wuhan Union Hospital cohort (c-index: 0.679; 5-year area under the curve: 0.697). Similarly, its categorical counterpart, log odds of positive lymph nodes classification, also exhibited robust prognostic discrimination (Surveillance, Epidemiology, and End Results c-index: 0.619, 5-year area under the curve: 0.666; Wuhan Union Hospital c-index: 0.637, 5-year area under the curve: 0.640). The nomogram, incorporating log odds of positive lymph nodes classification, age, T-stage, retrieved lymph nodes, positive lymph nodes, and chemotherapy, demonstrated an area under the curve of 0.735, 0.708, and 0.698 for 1-, 3-, and 5-year predictions, respectively, in the Surveillance, Epidemiology, and End Results cohort, and 0.734, 0.715, and 0.691, respectively, in the Wuhan Union Hospital cohort. Calibration curves and decision curve analysis confirmed the model's accuracy and clinical utility. The online tool facilitates clinical use.
Conclusion:
Log odds of positive lymph nodes classification is identified as the most powerful lymph node-related prognostic parameter for prognostic prediction in duodenal adenocarcinoma. The proposed nomogram, incorporating log odds of positive lymph nodes classification, provides an accurate and clinically useful tool for individualized survival estimation.
