Project Grant R01HL181107

The National Heart, Lung, and Blood Institute awarded the University of Virginia $760,465 on August 1, 2026, under the Cardiovascular Diseases Research program (CFDA 93.837) to develop deep learning methods that advance cardiac magnetic resonance strain imaging for personalized cardiac resynchronization therapy treatment planning.

The award funds development of deep strain networks to rapidly and accurately predict myocardial displacements from cardiac magnetic resonance imaging, addressing the clinical challenge that 30–40 percent of heart failure patients receiving cardiac resynchronization therapy through traditional biventricular pacing are non-responders. Current strain imaging methods—feature tracking software accessible in commercial platforms but limited in segmental accuracy, or specialized techniques such as displacement encoding with stimulated echoes that produce highly accurate data but require extended scan times and are not widely available—create barriers to incorporating strain guidance into clinical practice. The project will leverage deep learning to overcome these limitations and establish strain-derived imaging features that can predict patient response and guide decisions between biventricular pacing and conduction system pacing approaches, particularly left bundle branch area pacing.

Work is performed in Charlottesville, Virginia. The period of performance extends from August 1, 2026, through May 31, 2031. The award is classified as a Project Grant, an assistance type supporting discrete, circumscribed research projects under NIH peer-review competition.