Related Experiment Video
Updated: Jun 29, 2026
14:08
Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
Published on: April 13, 2013
42.9K
Adhesion-related small bowel obstruction: deep learning for automatic transition-zone detection by CT
Quentin Vanderbecq1,2, Roberto Ardon3, Antoine De Reviers3
1Department of Medical Imaging, Saint Joseph Hospital, 185 rue Raymond Losserand, 75014, Paris, France. q.vanderbecq@gmail.com.
Insights Into Imaging
|January 24, 2022
Summary
Machine learning accurately identifies the transition zone (TZ) in adhesion-related small bowel obstruction (SBO) on CT scans. This automated detection method shows promise for improving radiological diagnosis of SBO.
Area of Science:
- Radiology
- Medical Imaging
- Machine Learning
Background:
- Adhesion-related small bowel obstruction (SBO) poses diagnostic challenges.
- Accurate localization of the transition zone (TZ) is crucial for SBO management.
- Current methods for TZ identification can be time-consuming and subjective.
Purpose of the Study:
- To develop and evaluate a machine-learning model for automated TZ localization in SBO on CT scans.
- To improve the efficiency and accuracy of diagnosing adhesion-related SBO.
Main Methods:
- Utilized 562 CT scans from 404 patients with adhesion-related SBO.
- Trained a neural network model to classify abdominopelvic patches for TZ presence.
- Employed a coupled classification and probabilistic estimation approach.
- Evaluated model performance using Area Under the Receiver Operating Characteristics Curve (AUROC).
Main Results:
- The hypogastric region showed the highest probability of TZ presence (56.9%).
- The combined model achieved an AUROC of 0.93.
- At a 15% volume classification, highlighted patches had a 92% probability of containing a TZ.
Conclusions:
- Coupling convolutional neural network classification with probabilistic estimation enables effective TZ localization.
- This approach offers a potential pathway for automated TZ detection in SBO.
- Automated TZ detection has significant clinical implications for SBO diagnosis and management.
