Machine learning-based prediction of prolonged air leak after uniportal video-assisted thoracic surgery segmentectomy - Takeaways - MDSpire

Machine learning-based prediction of prolonged air leak after uniportal video-assisted thoracic surgery segmentectomy

  • By

  • Liang Chen

  • Ting Yu

  • Yanqing Pan

  • Guodong Ma

  • June 19, 2026

  • 0 min

Share

  • 1

    Prolonged air leak (PAL) is a common complication after uniportal video-assisted thoracic surgery (uVATS) segmentectomy, affecting 12.46% of patients in the study.

  • 2

    The study developed machine learning models to predict PAL using clinical, surgical, and physiological variables from 610 patients undergoing uVATS segmentectomy.

  • 3

    The XGBoost model demonstrated the highest performance for predicting PAL, achieving an AUC of 0.874 in the internal test set.

  • 4

    Key predictive factors for PAL identified included low body mass index (BMI), prolonged operative time, reduced DLCO%, diabetes, and complex segmentectomy.

  • 5

    The study emphasizes the need for external validation and prospective evaluation of machine learning models before clinical implementation.

Original Source(s)

Machine learning-based prediction of prolonged air leak after uniportal video-assisted thoracic surgery segmentectomy

  • by Liang Chen, Ting Yu, Yanqing Pan, Guodong Ma

  • June 19, 2026

Related Content

Remimazolam Is Associated with Superior Cerebral and Pulmonary Protection over Propofol in Elderly Thoracic Surgery: A Real-World Study Validated by Propensity Score Matching

Lobectomy versus sublobar resection for stage I non-small cell lung cancer: an umbrella review of evidence quality, overlap, surgical-extent heterogeneity, and survival outcomes

Surgeon Sleep Timing Tied to Risk

Severe social jet lag among surgeons was associated with higher rates of major adverse events, independent of sleep duration, workload, and patient risk.