Researchers have known for a long time that more information was available through X-ray data but lacked a reliable method for analyzing it in a way that scales.
Technology-driven initiatives in healthcare are often designed to increase early detection rates for better patient outcomes. Now, a research team has created a deep learning algorithm to predict the ten-year likelihood of heart disease.
These deep learning algorithms are able to train on medical data to uncover patterns that are frequently presented in a positive heart disease diagnosis. The risk is calculated using the atherosclerotic cardiovascular disease (A.S.C.V.D.) risk score. This is a statistical model that considers many different variables.
The current recommendation from the medical community is to establish the ten-year risk factor of things like stroke or heart attack based on these assessments. A.I. could help healthcare providers become more accurate in analyzing the risk and provide care sooner.
The algorithm learned on over 40,000 patient X-rays. Researchers have known for a long time that more information was available through X-ray data but lacked a reliable method for analyzing it in a way that scales. These models could help patients receive a faster risk assessment without jumping through hoops at their local healthcare provider.
The researchers compared the result to the established clinical pattern and are hopeful of the results. However, additional research is necessary to continue refining the model and measuring its efficacy. This includes the need for a randomized trial.
These tools could be vital to expand healthcare services efficiently and affordably for healthcare facilities. It could also point to further support for providers as they manage more significant patient caseloads and reduced time with patients.
The study was conducted by Vineet Raghu, Ph.D., Kaavya Paruchuri, M.D., Pradeep Natarajan, M.D., M.M.S.C., Hugo Aerts, Ph.D., and Michael T. Lu, M.D., M.P.H. The National Academy of Medicine and the American Heart Association also supported the study with additional funding.