Scholarlink

Purpose <p>Surgical Site Infection (SSI) is the common consequence of hernia surgery. This study aimed to develop an artificial intelligence (AI)-driven predictive model for SSI following open ventral hernia repair using pre- and intra-operative parameters for early detection.</p> Methods <p>A prospective analysis of ventral hernia patients was conducted. Demographic, pre-operative and intra-operative parameters were collected. Feature selection using Recursive Feature Elimination (RFE) identified key predictors. Five models, eXtreme Gradient Boosting (XGBoost), logistic regression, support vector machine (SVM), Adaptive Boosting and random forest (RF), were trained and validated.</p> Results <p>Among 253 patients, 22 were SSI-positive. RF demonstrated the highest predictive power (area under receiver operating characteristic curve = 0.82) and diagnostic odds ratio (DOR = 55). Seven key predictors were identified: defect size area, HbA1c, lymphocytes, neutrophil-to-lymphocyte ratio, blood loss (ml), platelet count and serum albumin. Decision curve analysis showed an estimated risk stratification at an 8.3% cutoff “<a href="https://surgicut-ai.netlify.app/">https://surgicut-ai.netlify.app/</a>”.</p> Conclusion <p>The model’s high accuracy supports its potential for clinical decision-making. External validation is required for real-world application. Integrating this AI model into bedside tools may enhance SSI prevention, especially in resource-limited settings.</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

SurgiCut-AI: an AI-driven tool for predicting surgical site infections following ventral hernia repair using pre- and intra-operative parameters

  • 期刊论文

  • 发表日期：

    2026年1月2日

查看全文
  • Rijhul Lahariya,
  • Ashesh Kumar Jha,
  • Mainak Sinha,
  • Prashant Kumar Singh,
  • Shiv Shankar Paswan,
  • Manasi Manasvi,
  • Manoj Kumar

摘要

Purpose

Surgical Site Infection (SSI) is the common consequence of hernia surgery. This study aimed to develop an artificial intelligence (AI)-driven predictive model for SSI following open ventral hernia repair using pre- and intra-operative parameters for early detection.

Methods

A prospective analysis of ventral hernia patients was conducted. Demographic, pre-operative and intra-operative parameters were collected. Feature selection using Recursive Feature Elimination (RFE) identified key predictors. Five models, eXtreme Gradient Boosting (XGBoost), logistic regression, support vector machine (SVM), Adaptive Boosting and random forest (RF), were trained and validated.

Results

Among 253 patients, 22 were SSI-positive. RF demonstrated the highest predictive power (area under receiver operating characteristic curve = 0.82) and diagnostic odds ratio (DOR = 55). Seven key predictors were identified: defect size area, HbA1c, lymphocytes, neutrophil-to-lymphocyte ratio, blood loss (ml), platelet count and serum albumin. Decision curve analysis showed an estimated risk stratification at an 8.3% cutoff “https://surgicut-ai.netlify.app/”.

Conclusion

The model’s high accuracy supports its potential for clinical decision-making. External validation is required for real-world application. Integrating this AI model into bedside tools may enhance SSI prevention, especially in resource-limited settings.