Abstract:Objective: To develop a machine learning-based risk prediction model for postoperative parastomal hernia (PSH) in colorectal cancer patients undergoing permanent colostomy, assisting nurses in identifying high-risk groups and devising preventive care strategies. Methods: A case-control study was conducted on 495 colorectal cancer patients who underwent permanent colostomy at the Second Affiliated Hospital of Anhui Medical University from June 2017 to June 2023, with a 1-year follow-up period. Patients were categorized into PSH and non-PSH groups based on PSH occurrence within 1-year post-operation. Data were split into training (70%) and testing (30%) sets. Variable selection was performed using Least Absolute Shrinkage and Selection Operator (LASSO) regression, and binary classification prediction models were established using Logistic Regression (LR), Support Vector Classification (SVC), K Nearest Neighbor (KNN), Random Forest (RF), Light Gradient Boosting Machine (LGBM), and Extreme Gradient Boosting (XgBoost). The binary classification label denoted 1 for PSH occurrence and 0 for no PSH occurrence. Parameters were optimized via 5-fold cross-validation. Model performance was evaluated using Area Under Curve (AUC), specificity, sensitivity, accuracy, positive predictive value, negative predictive value, and F1-score. Clinical utility was evaluated using decision curve analysis (DCA), model explanation was enhanced using shapley additive explanation (SHAP), and model visualization was achieved using a nomogram. Results: The incidence of PSH within 1 year was 29.1% (144 patients). Among the models tested, the RF model demonstrated the highest discrimination capability with an AUC of 0.888 (95% CI: 0.881-0.935), along with superior specificity, accuracy, sensitivity, and F1 score. It also showed the highest clinical net benefit on the DCA curve. SHAP analysis identified the top 10 influential variables associated with PSH risk: body mass index (BMI), operation duration, history and status of chronic obstructive pulmonary disease (COPD), prealbumin, tumor node metastasis (TNM) staging, stoma site, thickness of rectus abdominis muscle (TRAM), C-reactive protein CRP, american society of anesthesiologists physical status classification (ASA), and stoma diameter. These insights from SHAP plots illustrated how these factors influence individual PSH outcomes. The nomogram was used for model visualization. Conclusion: The Random Forest model demonstrated robust predictive performance and clinical relevance in forecasting colonic PSH. This model aids in early identification of high-risk patients and guides preventive care.

Establishment of a risk prediction model for bowel necrosis in patients with incarcerated inguinal hernia

A Rapid Preliminary Prediction Model for Intestinal Necrosis in Acute Mesenteric Ischemia: a Retrospective Study

The predictive factors for postoperative intra-abdominal septic complications after intestinal resection in patients with Crohn's Disease

A predictive model of bowel resection for incarcerated inguinal hernia based on the systemic immune-inflammation index

Early Diagnosis of Bowel Necrosis in Incarcerated Hernias

Predicting risk factors for postoperative intestinal stenosis in neonates with necrotizing enterocolitis: development and assessment of a predictive nomogram

Risk factor analysis and nomogram prediction model construction for NEC complicated by intestinal perforation

A prediction model to differentiate transient ischemia from irreversible transmural necrosis in closed-loop small bowel obstruction

Predictors and risk factors for intestinal necrosis in patients with mesenteric ischemia

Nomogram to Predict Postoperative Intra-abdominal Septic Complications after Bowel Resection And Primary Anastomosis for Crohn's Disease.

Prediction Model Based on Blood Urea Nitrogen and the Leukocyte Count for Intestinal Necrosis in Patients with Portal Vein System Thrombosis: a Retrospective Study.

Predictors of Acute Transmural Intestinal Necrosis in Acute Mesenteric Ischemia: A Single-center Retrospective Study

Predictive factors of intestinal necrosis in acute mesenteric vascular occlusive diseases

Developing and validating a nomogram for early predicting the need for intestinal resection in pediatric intussusception

Personalized risk prediction for prolonged ileus after minimally invasive colorectal cancer surgery: in-depth risk factor analysis and model development

Development and validation of a nomogram to predict postsurgical intra-abdominal infection in blunt abdominal trauma patients: A multicenter retrospective study

A predictive model for intraabdominal infection after radical gastrectomy in elderly patients

Application of machine learning algorithms to predict postoperative surgical site infections and surgical site occurrences following inguinal hernia surgery

Risk factors and a simple scoring system for predicting bowel resection in infants with necrotizing enterocolitis

A risk prediction model based on machine learning algorithm for parastomal hernia after permanent colostomy

Nomogram to Predict Postoperative Infectious Complications after Surgery for Colorectal Cancer: a Retrospective Cohort Study in China.