Sign in Subscribe
Machine learning

Artificial Intelligence-enhanced Perfusion Scoring Improves Diagnostic Accuracy of Myocardial Perfusion Imaging

Strategic Insights

— 1 min read

A new report from Cedars Sinai Medical Center discusses the use of deep learning models to improve the accuracy of myocardial perfusion imaging for detecting obstructive coronary artery disease. Researchers aimed to develop an artificial intelligence (AI) approach that could be integrated into clinical practice to enhance diagnostic performance. The study found that the AI-enhanced perfusion scoring improved the diagnostic accuracy of myocardial perfusion imaging, with a higher area under the receiver operating characteristic curve (AUC) compared to traditional stress total perfusion deficit (TPD) and AI prediction alone.

Key Takeaways:

  • The study used a cohort of 555 patients undergoing myocardial perfusion imaging within 180 days of invasive coronary angiography.
  • The median age of the patients was 65 years, and 69% were male.
  • Obstructive coronary artery disease was present in 59% of the patients.
  • The AI-enhanced perfusion scoring improved the diagnostic accuracy of myocardial perfusion imaging, with an AUC of 0.837 compared to 0.795 for AI prediction alone and 0.737 for traditional stress TPD.
  • The study concluded that the AI approach may facilitate the integration of new AI methods into clinical practice.

Statistics:

  • 555 patients were used for testing in the study.
  • The median age of the patients was 65 years (interquartile range, 57-73 years).
  • 381 (69%) of the patients were male.
  • 329 (59%) of the patients had obstructive coronary artery disease.
  • The AI-enhanced perfusion scoring had an AUC of 0.837 (95% CI, 0.804-0.870).
  • AI prediction alone had an AUC of 0.795 (95% CI, 0.758-0.831).
  • Traditional stress TPD had an AUC of 0.737 (95% CI, 0.696-0.778).

Sources:

  • NewsRx. Study Results from Cedars Sinai Medical Center Broaden Understanding of Artificial Intelligence (Artificial Intelligence-enhanced Perfusion Scoring Improves the Diagnostic Accuracy of Myocardial Perfusion Imaging). Robotics & Machine Learning. May 19, 2025; p 85.
  • Artificial Intelligence-enhanced Perfusion Scoring Improves the Diagnostic Accuracy of Myocardial Perfusion Imaging. Journal of Nuclear Medicine, 2025;66(4):648-653.

Read more