Use of artificial intelligence for sepsis risk prediction after flexible ureteroscopy: a systematic review

Abstract

ABSTRACT Introduction: flexible ureteroscopy is a minimally invasive surgical technique used for the treatment of renal lithiasis. Postoperative urosepsis is a rare but potentially fatal complication. Traditional models used to predict the risk of this condition have limited accuracy, while models based on artificial intelligence are more promising. The objective of this study is to carry out a systematic review regarding the use of artificial intelligence to detect the risk of sepsis in patients with renal lithiasis undergoing flexible ureteroscopy. Methods: the literature review is in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA). The keyword search was performed in MEDLINE, Embase, Web of Science and Scopus and resulted in a total of 2,496 articles, of which 2 met the inclusion criteria. Results: both studies used artificial intelligence models to predict the risk of sepsis after flexible uteroscopy. The first had a sample of 114 patients and was based on clinical and laboratory parameters. The second had an initial sample of 132 patients and was based on preoperative computed tomography images. Both obtained good measurements of Area Under the Curve (AUC), sensitivity and specificity, demonstrating good performance. Conclusion: artificial intelligence provides multiple effective strategies for sepsis risk stratification in patients undergoing urological procedures for renal lithiasis, although further studies are needed.

Keywords

• Ureteroscopy
• Sepsis
• Artificial Intelligence
• Nephrolithiasis
• Machine Learning