Machine learning based model to diagnose obstructive coronary artery disease using calcium scoring, PET imaging, and...

Area of Science:

• Cardiology
• Medical Imaging
• Artificial Intelligence

Background:

• Accurate risk stratification is crucial for managing stable coronary artery disease (CAD).
• Machine learning (ML) offers potential for improving diagnostic accuracy in cardiology.
• Obstructive CAD (oCAD) diagnosis requires reliable risk assessment tools.

Purpose of the Study:

• To develop and validate a machine learning (ML) model for diagnosing obstructive coronary artery disease (oCAD).
• To compare the diagnostic performance of the ML model against expert readers.

Main Methods:

• Retrospective analysis of 1007 patients without prior CAD history.
• Utilized CT-based calcium scoring (CACS) and Rubidium-82 PET scan data.
• Developed an ML model using stratified cross-validation; compared performance against expert readers using invasive coronary angiography (ICA) as the gold standard.

Main Results:

• The ML model achieved an AUC of 0.92 on the training set and 0.89 on the test set.
• Diagnostic performance was similar to expert readers, with no significant differences in accuracy (89% vs. 88%), sensitivity (68% vs. 69%), and specificity (92% vs. 90%).

Conclusions:

• The ML model demonstrates comparable diagnostic performance to expert readers for oCAD.
• This ML-based tool holds promise as a risk stratification method for obstructive CAD.
• Machine learning utilization is a promising approach for the diagnosis of obstructive CAD.