Chest Tube Learning Synthesis and Evaluation Assistant (CheLSEA): A Prospective Observational Trial of an Intelligent...

Area of Science:

• Medical Informatics
• Artificial Intelligence in Medicine
• Thoracic Surgery

Background:

• Chest tube management is critical post-pulmonary resection surgery.
• Optimizing chest tube removal timing balances risks of complications and prolonged drainage.
• Clinical decision support systems can aid in complex management decisions.

Purpose of the Study:

• To evaluate the performance of an artificial intelligence (AI) predictive clinical decision support system, CheLSEA.
• To assess CheLSEA's accuracy in generating chest tube management recommendations.
• To compare AI-driven recommendations with standard chest tube care.

Main Methods:

• A single-arm, double-anonymized, observational study of 50 adult patients undergoing pulmonary resection.
• Prospective collection of clinical status, digital pleural drainage data, and chest X-ray data.
• CheLSEA provided recommendations for chest tube removal or maintenance, including removal time predictions.

Main Results:

• CheLSEA was queried 174 times; its safeguard system activated in 21% of cases, primarily due to subcutaneous emphysema.
• CheLSEA recommended chest tube removal in 9% of requests, with 83% deemed safe and 17% premature by ≤6 hours.
• For maintenance recommendations, 93% of chest tube removal time predictions were accurate.

Conclusions:

• The CheLSEA system demonstrates safety in providing chest tube management recommendations.
• CheLSEA can potentially enhance patient care by reliably emulating expert clinical guidance.
• AI-powered decision support shows promise in optimizing thoracic surgical patient management.