Rapid, Automatic, AI Driven, Aortic Valve Measurement Software

Project Summary While trans-catheter aortic valve replacements (TAVR) has revolutionized valve replacement, planning for this procedure using CT imaging is a time-consuming manual process requiring hours of advanced cardiac imager’s time. Furthermore, many centers in the United States turn over the job of measurements to in-experienced sales reps who are more focused on the interests of the valve manufacturers they represent without any formal training or knowledge in cardiac anatomy or physiology. DASI Simulations LLC has developed an innovative cloud-based artificial intelligence (AI) platform that promises to automatically perform these complex measurements in a matter of few minutes. The measurements are performed based on AI-predicted landmarks using a pipeline of convolutional neural networks with advanced image sampling techniques. As preliminary work we have already established the product feasibility and have our regulatory approach vetted through the FDA with a successful presubmission completed. In this SBIR Phase 1, we propose a non-hypothesis driven approach with two specific aims to execute on FDA recommendations to finish the verification and validation activities, namely: (1) Develop and validate the DASI Dimensions measurement module to extract TAVR pre-planning procedural measurements; and (2) prepare and submit document package for Food and Drug Administration (FDA) 510k clearance. Because our technology differentiates itself from our competitors who use traditional expensive manual measurement approaches, we are posed to make a significant positive and disruptive impact in how structural heart therapy is planned in the clinic with potentially substantial clinical impact in terms of patient outcomes and hospital efficiency.