Skip to main content

Improving Aortic Stenosis Care with Machine Learning

Victor Chang Cardiac Research Institute & University of New South Wales

  • Cardiovascular Early-Mid Career Researcher Grant
Date Funded:
  • 18 November, 2021
Chief Investigator/s:
  • Dr. Mayooram Namasivayam

Project Summary

Improving patient selection and timing of valve intervention in patients with aortic stenosis using computer vision and machine learning.

What is the issue for NSW?

It is estimated that one in eight people over the age of 75 have moderate or severe aortic stenosis. In NSW there are over 500,000 people aged over 75 years. The two-year mortality for aortic stenosis once symptoms develop is 50%. The cost of surgical or transcatheter valve replacement is estimated at $80,000 per patient. Transcatheter valve replacement is becoming increasingly accessible, but diagnostic uncertainties in aortic stenosis remain. There is an urgent need to better identify patients most likely to benefit, as well as optimal treatment timing.

This study will seek to improve the care of patients with aortic stenosis by improving patient selection and timing of intervention using novel machine learning approaches. These efforts seek to reduce morbidity, mortality and healthcare costs related to aortic stenosis.

What does the research aim to do and how?

This study aims to develop machine learning algorithms to identify likely patient outcomes after aortic valve intervention and determine optimal timing of intervention.  Using imaging and clinical data from leading aortic valve centres around Australia, this study will use computer vision analysis and machine learning algorithms (such as cluster analysis, artificial neural networks and bootstrap lasso regression). These techniques applied to imaging and clinical data will be used to develop a practical risk prediction tool for clinicians which they can use while meeting with the patient.