Development and validation of a nomogram for preoperative prediction of tumor deposits in colorectal cancer

BACKGROUND: Based on the clinical data of colorectal cancer (CRC) patients who underwent surgery at our institution, a model for predicting the formation of tumor deposits (TDs) in this patient population was established. AIM: To establish an effective model for predicting TD formation, thus enabling clinicians to identify CRC patients at high risk for TDs and implement personalized treatment strategies. METHODS: CRC patients (n = 645) who met the inclusion criteria were randomly divided into training (n = 452) and validation (n = 193) cohorts using a 7:3 ratio in this retrospective analysis. Least absolute shrinkage and selection operator regression was employed to screen potential risk factors, and multivariable logistic regression analysis was used to identify independent risk factors. Subsequently, a predictive model for TD formation in CRC patients was constructed based on the independent risk factors. The discrimination ability of the model, its consistency with actual results, and its clinical applicability were evaluated using receiver-operating characteristic curves, area under the curve (AUC), calibration curves, and decision curve analysis (DCA). RESULTS: Thirty-four (7.5%) patients with TDs were identified in the training cohort based on postoperative pathological specimens. Multivariate logistic regression analysis identified female sex, preoperative intestinal obstruction, left-sided CRC, and lymph node metastasis as independent risk factors for TD formation. The AUCs of the nomogram models constructed using these variables were 0.839 and 0.853 in the training and validation cohorts, respectively. The calibration curve demonstrated good consistency, and the training cohort DCA yielded a threshold probability of 7%-78%. CONCLUSION: This study developed and validated a nomogram with good predictive performance for identifying TDs in CRC patients. Our predictive model can assist surgeons in making optimal treatment decisions.