Exploring Deep Learning Applications using Ultrasound Single View Cines in Acute Gallbladder Pathologies: Preliminary...

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Gastroenterology

Background:

• Gallbladder conditions like cholelithiasis and acute cholecystitis are common causes of right-upper-quadrant pain.
• Accurate and timely diagnosis is crucial for appropriate patient management, especially differentiating urgent surgical cases from non-urgent ones.

Purpose of the Study:

• To develop and evaluate a deep learning model for classifying gallbladder ultrasound single view cines.
• To differentiate between normal gallbladder imaging, non-urgent cholelithiasis, and acute calculous cholecystitis requiring urgent intervention.

Main Methods:

• A deep learning model was trained on 266 longitudinal-view ultrasound cines from 186 adult patients.
• Cines were categorized as normal, non-urgent cholelithiasis, or acute cholecystitis based on final clinical diagnosis.
• The model was trained to classify cines into normal, cholelithiasis, or acute cholecystitis categories.

Main Results:

• The model achieved 91% accuracy in distinguishing normal from abnormal gallbladder imaging.
• It demonstrated 82% accuracy in classifying urgent (acute cholecystitis) versus non-urgent conditions.
• The model exhibited 100% specificity in identifying abnormal from normal imaging, with no false positives.

Conclusions:

• A deep learning model utilizing single ultrasound view cines can accurately and specifically differentiate between non-urgent gallbladder conditions and acute cholecystitis.
• This AI-driven approach shows promise for improving diagnostic efficiency and patient triage in emergency settings.