Scholarlink

Background <p>Inguinal hernia repair, particularly transabdominal preperitoneal (TAPP) repair, is a common surgical procedure. However, seroma formation remains a frequent postoperative complication, impacting patient recovery and increasing healthcare burden. Accurate prediction of seroma formation is crucial for personalized patient management and improved outcomes.</p> Objective <p>This study aimed to develop and validate machine learning (ML)-based predictive models for seroma formation following TAPP inguinal hernia repair.</p> Methods <p>A retrospective cohort of patients undergoing TAPP repair was analyzed. Demographic, clinical, laboratory, and surgical data were collected. Data preprocessing included handling missing values and encoding categorical variables. Feature selection was performed using LASSO regression and multivariate logistic regression. Five distinct machine learning algorithms—Logistic Regression (LR), Decision Tree (DT), eXtreme Gradient Boosting (XGBoost), Support Vector Machine (SVM), and K-Nearest Neighbors (KNN)—were developed and evaluated. Model performance was assessed using various metrics, including Accuracy, Sensitivity, Specificity, Precision, Recall, F1-score, Receiver Operating Characteristic (ROC) curves, Area Under the Curve (AUC), Precision-Recall (PR) curves, Calibration curves, and Decision Curve Analysis (DCA). SHapley Additive exPlanations (SHAP) were employed to interpret the models and identify key influential features.</p> Results <p>A total of 167 patients were included in the study, divided into training (<i>n</i> = 133) and validation (<i>n</i> = 34) sets. Multivariate analysis identified several independent risk factors for seroma formation, including Diabetes, Smoking, Age, Hemoglobin, and sac diameter. Among the developed ML models, XGBoost consistently demonstrated superior predictive performance with an AUC of 0.979 (0.961–0.997) on the training set and 0.746 (0.557–0.935) on the test set. SHAP analysis revealed that sac diameter, Hemoglobin, Age, Diabetes, and Smoking were the most significant features influencing seroma prediction.</p> Conclusion <p>Machine learning models, particularly XGBoost, can accurately predict seroma formation after TAPP inguinal hernia repair. The identified risk factors and the developed predictive models have the potential to facilitate early risk stratification, guide clinical decision-making, and ultimately improve patient outcomes by enabling targeted preventive strategies.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Development of machine learning-based predictive models for seroma formation after transabdominal preperitoneal inguinal hernia repair

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年3月18日

查看全文
  • Yibo Zhai,
  • Yi Zhao,
  • Chang Chen,
  • Wenzhong Bao,
  • Dawei Tang

摘要

Background

Inguinal hernia repair, particularly transabdominal preperitoneal (TAPP) repair, is a common surgical procedure. However, seroma formation remains a frequent postoperative complication, impacting patient recovery and increasing healthcare burden. Accurate prediction of seroma formation is crucial for personalized patient management and improved outcomes.

Objective

This study aimed to develop and validate machine learning (ML)-based predictive models for seroma formation following TAPP inguinal hernia repair.

Methods

A retrospective cohort of patients undergoing TAPP repair was analyzed. Demographic, clinical, laboratory, and surgical data were collected. Data preprocessing included handling missing values and encoding categorical variables. Feature selection was performed using LASSO regression and multivariate logistic regression. Five distinct machine learning algorithms—Logistic Regression (LR), Decision Tree (DT), eXtreme Gradient Boosting (XGBoost), Support Vector Machine (SVM), and K-Nearest Neighbors (KNN)—were developed and evaluated. Model performance was assessed using various metrics, including Accuracy, Sensitivity, Specificity, Precision, Recall, F1-score, Receiver Operating Characteristic (ROC) curves, Area Under the Curve (AUC), Precision-Recall (PR) curves, Calibration curves, and Decision Curve Analysis (DCA). SHapley Additive exPlanations (SHAP) were employed to interpret the models and identify key influential features.

Results

A total of 167 patients were included in the study, divided into training (n = 133) and validation (n = 34) sets. Multivariate analysis identified several independent risk factors for seroma formation, including Diabetes, Smoking, Age, Hemoglobin, and sac diameter. Among the developed ML models, XGBoost consistently demonstrated superior predictive performance with an AUC of 0.979 (0.961–0.997) on the training set and 0.746 (0.557–0.935) on the test set. SHAP analysis revealed that sac diameter, Hemoglobin, Age, Diabetes, and Smoking were the most significant features influencing seroma prediction.

Conclusion

Machine learning models, particularly XGBoost, can accurately predict seroma formation after TAPP inguinal hernia repair. The identified risk factors and the developed predictive models have the potential to facilitate early risk stratification, guide clinical decision-making, and ultimately improve patient outcomes by enabling targeted preventive strategies.