Development of a Nomogram to Predict Postoperative Peritoneal Metastasis of Colon Cancer.

Abstract

The aim of this study was to determine the clinicopathological and radiological risk factors for postoperative peritoneal metastasis and develop a prediction model for the early detection of peritoneal metastasis in patients with colon cancer. We included 174 patients with colon cancer. The clinicopathological and radiological data were retrospectively analyzed. A Cox proportional hazards regression model was used to identify risk factors for postoperative peritoneal metastasis. Based on these risk factors, a nomogram was developed. At a median follow-up of 63 months, 43 (24.7%) patients developed peritoneal metastasis. Six independent risk factors (hazards ratio [95% confidence interval]) were identified for postoperative peritoneal metastasis: abdominopelvic fluid (2.12 [1.02-4.40]; P = 0.04), longer maximum tumor length (1.02 [1.00-1.03]; P = 0.02), pN1 (2.50 [1.13-5.56]; P = 0.02), pN2 (4.45 [1.77-11.17]; P = 0.02), nonadenocarcinoma (2.75 [1.18-6.38]; P = 0.02), and preoperative carcinoembryonic antigen levels ≥5 ng/mL (3.08 [1.50-6.30]; P < 0.01). A clinicopathological-radiological model was developed based on these factors. The model showed good discrimination (concordance index, 0.798 [0.723-0.876]; P < 0.001) and was well-calibrated. The developed clinicopathological-radiological nomogram may assist clinicians in identifying patients at high risk of postoperative peritoneal metastasis.