AI in computational medicine

Computational medicine is an emerging field that uses quantitative approaches to solving problems in human health and disease including diagnosis, treatment, and understanding causative mechanisms. In recent years, the advent of big data has sprouted approaches to understanding mechanisms of disease by combining traditional algorithms with data-driven approaches.

Machine learning algorithm to identify stenosis (regions of lumen narrowing) using Machine learning and compared against visual CT data and angiogram [2].

At HeartFlow, we have built one of the largest databases of human coronary artery trees with associated clinical information. I have developed AI based algorithms to predict both anatomic and hemodynamic information in real time with high accuracy compared to annotated “ground truth” data.

Estimation of hemodynamic parameters using machine learning and (right) comparison of time complexity using machine learning compared to three dimensional solutions.

Relevant publications

  1. Nakanishi, R., Sankaran, S.* et al., Automated Estimation of Image Quality for Coronary Computed Tomographic Angiography using Machine Learning, European Radiology, Vol.28/9, pp. 4018-4026, 2018 (* indicates equal contributors).
  2. Sankaran, S., et al., HALE: Healthy area of lumen estimation for vessel stenosis quantification, Lecture notes in Computer Science: MICCAI, Vol. 9902, pp. 380-387, 2016.
  3. Sankaran, S., Grady, L., and Taylor, C., Fast computation of hemodynamic sensitivity to lumen segmentation uncertainty, IEEE Transactions on Medical Imaging, 34 (12), pp. 2562-2571, 2015.
  4. Sankaran, S., Grady, L., and Taylor, C., Impact of geometric uncertainty on hemodynamic simulations using machine learning, Computer Methods in Applied Mechanics and Engineering, 297(1), pp. 167-190, 2015.
  5. Sankaran, S., Grady, L. J., and Taylor, C. A, Real-Time Sensitivity Analysis of Blood Flow Simulations to Lumen Segmentation Uncertainty, Medical Image Computing and Computer-Assisted Intervention, 1, pp. 1-8, 2014.