Explainable Knowledge-Guided Algorithm for Contrast Extravasation Detection on Computed Tomography

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Radiology

Background:

• Contrast media extravasation is a critical finding in trauma CT.
• Timely detection is essential for effective trauma triage and patient management.
• Current manual review processes can be time-consuming in emergency settings.

Purpose of the Study:

• To develop an explainable, knowledge-guided framework for automated contrast extravasation detection.
• To evaluate its potential in accelerating time-critical trauma triage.
• To maintain clinically acceptable sensitivity in detection.

Main Methods:

• A mathematical framework encoding expert diagnostic rules was formulated.
• Sequential CT slices were analyzed using a 2.5D approach integrating temporal and spatial information.
• The model outputs a confidence score and binary alert, validated against surgeon assessment.

Main Results:

• The framework achieved clinically acceptable sensitivity for contrast extravasation detection.
• It substantially reduced the time-to-decision compared to manual review.
• Positive cases were identified early in the CT volume, enabling rapid screening.

Conclusions:

• Expert clinical reasoning can be translated into an interpretable computational model.
• The knowledge-guided design allows for rapid, transparent, automated screening.
• The framework shows promise as a decision-support tool for trauma triage acceleration.