I-Corps: Translation potential of semiautomatic segmentation of cardiac tomography images for myocardial blood flow quantification

Yale University, New Haven CT

Investigators

Abstract

The broader impact of this I-Corps project is the development of a novel analytic software tool incorporated with a semi-automatic algorithm for the segmentation of cardiac tomography images. The tool will be used to non-invasively quantify myocardial blood flow and flow reserve. This software tool provides a user-friendly image processing mechanism with minimal user intervention and thus, has less variability in quantitative results than other existing commercial software. The current practice of manually drawing regions of interest introduces inter-operator and intra-operator variability and may suffer from relatively low reproducibility between operators. The quantitative analytic tool can be utilized to assess myocardial blood flow and reserve not only for dynamic cardiac positron emission tomography, but potentially for dynamic cardiac single photon emission computed tomography. This I-Corps project utilizes experiential learning coupled with a first-hand investigation of the industry ecosystem to assess the translation potential of the technology. This solution is based on the development of a semi-automatic approach to define the regions of interest of the myocardial and blood pool volumes and to segment myocardial edges from dynamic cardiac tomography images to improve the accuracy and precision of myocardial blood flow and reserve quantification. The semi-automatic method is based on a novel scheme referred as to the triple-factor non-negative matrix factorization. This method is used for dynamic tomography image segmentation of the left ventricle, from which the time activity curves of the left ventricle myocardium and cavity are extracted and used in the kinetic modeling for calculating the myocardial blood flow and reserve. The quantitative software tool incorporated with the semi-automatic segmentation approach has been verified with computer simulations and validated with a large cohort of clinical patients. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.