Performance of an algorithm for diagnosing acute cholecystitis using clinical and sonographic parameters

Area of Science:

• Medical Imaging
• Diagnostic Algorithms
• Gastroenterology

Background:

• Acute cholecystitis is a common surgical emergency requiring accurate and timely diagnosis.
• Current diagnostic methods, including ultrasound (US) interpretation by radiologists, may have limitations in sensitivity and diagnostic rate.
• There is a need for improved diagnostic tools to enhance patient outcomes for acute cholecystitis.

Purpose of the Study:

• To develop and validate an algorithm utilizing clinical and ultrasound parameters for the diagnosis of acute cholecystitis.
• To compare the diagnostic performance of this novel algorithm against the impressions of subspecialty radiologists.

Main Methods:

• Retrospective review of emergency department (ED) patients' clinical and ultrasound data.
• Development of an algorithm by identifying predictive parameter combinations using pivot tables.
• Comparison of algorithm predictions with radiologist impressions for sensitivity, specificity, diagnostic rate, and adverse outcomes.

Main Results:

• The developed algorithm demonstrated significantly higher sensitivity (90.0% vs. 55.0%, p < 0.001) and augmented sensitivity (100% vs. 85.0%, p < 0.001) compared to radiologists.
• The algorithm achieved a higher diagnostic rate (96.4% vs. 93.2%, p = 0.04) with a notably lower adverse outcome rate (0.0% vs. 1.6%).
• Specificity was comparable between the algorithm and radiologists (93.6% vs. 94.8%, p = 0.50).

Conclusions:

• An algorithm integrating non-binary ultrasound and clinical assessments offers superior diagnostic performance for acute cholecystitis.
• This algorithm provides enhanced sensitivity and diagnostic accuracy, leading to a reduction in adverse patient outcomes.
• The findings suggest a valuable role for this algorithm in improving the diagnostic pathway for acute cholecystitis in the ED setting.