Visualizing Heart Disease

Accurate heart disese diagnosis with 3D models

(RxWiki News) There's such thing as too much technology when it comes to diagnosing heart disease. A newly developed two dimensional artery model may more accurately predict heart disease than its 3D counterpart.

The newly developed prototype tool called HemoVis has been shown to more accurately predict heart disease, especially when it comes to plaque build up within arteries.

"Talk to a cardiologist about your risk of heart disease."

Michelle Borkin, lead author and a doctoral candidate at the Harvard School of Engineering and Applied Sciences, said researchers found that the attractive and popular 3D visualization is not always the most effective. She said that in 3D models, the more complex and branched the arteries were, the longer it took to diagnosis patients and the less accurate it was. The 2D model, however, provided consistent and quicker results regardless of the number of branches.

Harvard researchers found in tests that the tool can increase diagnostic accuracy by between 39 percent and 91 percent over the traditional 3D rainbow-colored model.

The tool draws data from patient-specific blood flow simulations combined with traditional imaging data, and visually displays a tree diagram of the arteries with highlighted areas of heart disease.

Researchers extensively interviewed doctors during the process to identify popular options for display including accurate layout and coloring of the diseased and non-diseased areas of the arteries. The new tool skips rainbow coloring, opting instead for a graded single-color scheme that goes from red to black.

The 3D models generally require rotating models to get a complete view, but the 2D models do not have such a requirement. Though widely adopting the newer model could aid doctors and patients alike, the 3D tool may still be best in some scenarios, such as surgical planning.

The finding was recently presented at the IEEE Information Visualization Conference in Washington and will be published later this year in journal IEEE Transactions on Visualization and Computer Graphics.

Review Date: 
November 3, 2011