Advanced Software for Enabling Quantitative 3D Stress Echocardiography

DESCRIPTION (provided by applicant): Through more than 5 years of research, we have developed a new cardiovascular procedure termed quantitative three-dimensional (3D) stress echocardiography (echo). Because of exceedingly high computational demands, the procedure remains clinically nonviable despite our mature academic research. The objective of this STTR project is to remove the computation bottleneck and thus enable routine use of quantitative 3D stress echo. This new stress procedure overcomes many of the limitations of conventional stress echo by utilizing more powerful real-time 3D (RT3D) ultrasound instead of standard 2-dimensional (2D) ultrasound. The new 3D procedure allows a clinician to make diagnoses through comprehensive visualization of the heart and quantitative data on normal or abnormal left ventricular wall motion. Volumetric data gathered by RT3D ultrasound are the basis for these novel diagnostic capabilities. We have reported previously the ability: to (a) correct for the well-documented problem of misaligned views; (b) visualize any (dynamic) cross-section through the aligned pre- and post-stress data sets interactively on a personal computer; and (c) automatically identify the left ventricular myocardium to compute local, segment-wise measurements. These advanced image analysis algorithms also require up to an hour's execution time. We propose at least a 10-fold acceleration of these novel algorithms (net execution time of 5 min or less) by re- implementing those in graphics processor units (GPUs), which are now widely available and are powerful enough to solve complex computational tasks. Our specific aims for the Phase I project, therefore, are to: (1) develop GPU-accelerated implementation of pre- and post-stress RT3D ultrasound image registration; and (2) develop GPU-accelerated implementation of myocardial segmentation in RT3D ultrasound images. Efficient data processing, a prerequisite for clinical viability, will justify advancing to Phase II, in which we will (1) develop a complete, GPU-based software suite for visualizing and quantitatively interpreting pre- and post-stress RT3D ultrasound images; and (2) conduct multicenter clinical trials. As hundreds of cardiology departments look to adopt RT3D ultrasound, our timely research will accelerate the phasing in of quantitative 3D stress echo and its resulting benefits to millions with CAD. PUBLIC HEALTH: Three-dimensional (3D) acquisition is a new development in ultrasound imaging. We have suggested a new stress procedure based on 3D ultrasound and shown it to be more accurate than the conventional procedure. The new procedure, despite its benefits, remains limited to research. We propose converting our academic research into clinically and commercially viable solutions that will allow us to perform the new 3D stress procedure routinely. Because millions of patients suffer from cardiovascular disease, even a modest 1% increase in diagnostic accuracy by the new procedure could save many lives and reduce health care costs.