Tuesday, November 28 | 9:40 a.m.-9:50 a.m. | T3-SSIR02-2 | Room S501
CT body composition metrics calculated by an AI model can improve mortality predictions in patients receiving a transjugular intrahepatic portosystemic shunt (TIPS), according to this presentation.
In a retrospective study involving 122 TIPS patients at their institution, researchers led by presenter Tarig Elhakim, MD, of the University of Pennsylvania applied an automated deep-learning algorithm to extract skeletal- and fat-based metrics at the L3 vertebral body level. These metrics included skeletal muscle area, skeletal muscle index, skeletal muscle density, subcutaneous fat area, visceral fat area, visceral-to-subcutaneous fat ratio, and visceral fat index.
The researchers found that several metrics were associated with 90-day mortality after the TIPS procedure. What’s more, these metrics significantly increased the Model for End-Stage Liver Disease (MELD) score’s mortality prediction performance, improving its area under the curve from 0.76 to 0.84.
“Our study demonstrates that CT body composition metrics can predict 90-day mortality post-TIPS, and incorporating these metrics enhances the predictive performance of MELD Score,” the researchers wrote. “Although generalizability is limited to our study population, this research indicates that AI-based CT body composition may serve as a valuable tool for assessing patient risk during TIPS evaluation.”
Get all of the details by attending this Tuesday morning presentation