Related Experiment Video
Updated: Sep 18, 2026
Competing-Risk Nomogram for Predicting Cancer-Specific Survival in Multiple Primary Colorectal Cancer Patients after Surgery
Published on: September 27, 2024
Development and validation of an individualized nomogram to identify occult peritoneal metastasis in patients with
Naiqi Liu1, Zhaohui Wang2, Menglei Li3
1Department of Radiology, Shanghai Municipal Hospital of Traditional Chinese Medicine, Shanghai University of Traditional Chinese Medicine, Shanghai, 200071, PR China.
Objective:
Accurate preoperative prediction of occult peritoneal metastasis (OPM) is crucial for optimizing colorectal cancer (CRC) treatment. We developed and validated a practical nomogram that integrates CT imaging and clinicopathological parameters to assess individualized OPM risk.
Methods:
This retrospective study included 850 pathologically confirmed CRC patients, who were divided into a training cohort (n = 531, from Center 1, 2015-2023) and an external validation cohort (n = 319, from Center 2, 2019-2025). All patients were preoperatively CT-negative for peritoneal metastasis (PM), with OPM being confirmed intraoperatively or postoperatively. Independent predictors were identified using logistic regression to construct a nomogram. Model performance was evaluated by the area under the receiver operating characteristic curve (AUC), calibration plots, decision curve analysis (DCA), and clinical impact curve (CIC), with bootstrap validation for internal validation and the DeLong test for model comparisons.
Results:
Five independent OPM predictors were identified: histopathological type, CA19-9 levels, neutrophil-to-lymphocyte ratio (NLR), tumor enhancement patterns, and CT mesenteric fat-to-subcutaneous fat ratio (CTratio) (all p < 0.05). In the training cohort, the nomogram had an AUC of 0.901 (95% CI: 0.873-0.928), which was superior to the clinicopathological model (AUC = 0.758) and the CT model (AUC = 0.793) (both p < 0.001), with a sensitivity of 0.855 and a specificity of 0.889. Internal validation (AUC = 0.890, 95% CI: 0.860-0.920) and external validation (AUC = 0.889, 95% CI: 0.852-0.925, sensitivity = 0.889, specificity = 0.809) demonstrated robust performance, with no overfitting. DCA and CIC confirmed its clinical utility.
Conclusions:
The nomogram integrating CT features (tumor enhancement, CTratio) and clinicopathological parameters (histopathological type, CA19-9, NLR) showed excellent accuracy and generalizability for preoperative OPM prediction in CRC, aiding preoperative decision-making and individualized treatment.
