Heart disease remains the number one cause of death worldwide. The effectiveness of interventional heart disease treatment requires the use of x-rays, which can be harmful to patients as well as operators who experience repeated exposure over time. In addition, accidental injury to a major nerve near the heart (the phrenic nerve) can cause breathing difficulties after surgery. These two grand challenges will be tackled with the combination of light and sound, known as photoacoustic imaging, which is a promising new method to (1) guide treatment catheters to the heart with robotic assistance and (2) visualize nerves. First, this project will advance robotic photoacoustic imaging technology to introduce a radically new concept that will reduce current reliance on x-rays. Second, this project will expand our knowledge of critical nerve properties for photoacoustic imaging to initiate a paradigm shift for nerve visualization during minimally invasive and open chest heart procedures, which both require careful navigation around the phrenic nerve. This award will also be a vehicle to increase minority participation in computer science and engineering, to mentor the next generation of technological leaders, to develop new course material related to research findings, and to enable cross-disciplinary sharing of research results with engineering and clinical communities.
The objective of this award is to synthesize acoustic models and experimental optical analyses to understand the limits of an innovative robotic photoacoustic imaging system for guiding cardiac surgeries and interventions. The project has three fundamental research aims. First, new acoustic models will be introduced, developed, and refined to predict the likelihood of photoacoustic signal visualization and segmentation during photoacoustic-based robot control, which will be used to advocate a novel paradigm of reduced reliance on x-rays. This prediction will relate fundamental photoacoustic spatial coherence theory to photoacoustic signal-to-noise ratios and associated laser energies in order to avoid the time-consuming, trial-and-error procedures that are currently implemented to determine minimum required energies for the new technology. This theory-based approach will then be compared with deep learning approaches that exhibit the potential to exceed theoretical limits. Second, the currently nonexistent optical properties of multiple phrenic nerves will be characterized in order to create realistic expectations for the challenging task of nerve visualization and avoidance during cardiac procedures. Third, the integration of theoretical acoustic models, deep learning methods, vision-based robot control, and nerve characterization results will be evaluated as a novel interconnected system to guide in vivo cardiac procedures. This research has promising potential to reduce ionizing radiation exposure, to improve current knowledge of nerve optical properties, and to eliminate nerve-related injury complications, while making fundamental contributions to the disciplines of engineering, computer vision, biomedical optics, and medicine.
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.
