Fellow or Postdoc
Amir Rouhollahi; Sandra Haltmeier; Hoda Javadikasgari; Kim de la Cruz; Elena Aikawa; Farhad R. Nezami
Farhad R. Nezami
Aortic stenosis (AS) poses a significant health challenge, especially among older populations. Despite its prevalence, knowledge gaps remain about how calcification develops over time and space in relation to disease severity and treatment outcomes. This underscores the potential benefits of employing artificial intelligence (AI) to provide a nuanced understanding of calcium distribution.
In this study, CT images from 1,700 TAVR patients were retrospectively collected in from IRB guidelines. A deep learning tool was employed to produce digital mappings of the aortic root and the calcified valve. This allowed for the detailed quantification of calcium distribution in various directions and provided insights into specific leaflet metrics. To ascertain the value of these metrics for predicting the need for post-TAVR permanent pacemaker (PPM) installation, a combination of principal component analysis and a support vector machine (SVM) model was used.
Initial results, based on 45 severe AS patients, demonstrated the viability of this AI approach to quantify calcium distribution. Notably, certain calcification metrics were found to be significantly linked to PPM outcomes. In conclusion, AI’s ability to practically quantify spatial calcification offers valuable insights into potential post-procedure complications. Such AI-informed strategies can guide interventions, reduce undesirable outcomes, and inspire better prosthetic valve designs.
This research focuses on improving heart treatments using artificial intelligence (AI). The heart has a valve that can get narrow and stiff due to calcium deposits, a condition known as Aortic Stenosis. This is particularly a problem for older adults. Currently, doctors use a procedure called TAVR to replace the damaged valve. However, understanding how and where the calcium builds up can help doctors predict which patients may face complications after the procedure.
The researchers used AI to analyze CT scans of the heart from 1,700 patients who had undergone this valve replacement. The AI could create a detailed map of where calcium had built up in the valve and surrounding areas. This information was then used to predict which patients were more likely to need a pacemaker after the valve replacement.
The study found that the AI’s predictions were accurate and could help doctors plan the valve replacement procedure better, minimizing risks. This could lead to fewer complications and better-designed replacement valves in the future. Overall, the research shows that AI can play a crucial role in personalized healthcare, tailoring treatments to each individual’s unique needs.