Development of a Semi-autonomous Handheld Robotic Device to Enable Rapid Vascular Access By Non-specialists in Austere Environments

The Artificial Intelligence-Guided Ultrasound Intervention Device (AI-GUIDE) is a groundbreaking handheld robotic technology. Non-specialists using AI-GUIDE prototypes in tissue-mimicking phantoms have been shown to achieve femoral vascular cannulation success rates and procedure times on par with medical specialists. In porcine trials, the device demonstrated high success rates, with the pig in both normotensive and hypotensive conditions, further underlining its effectiveness and ease of use. This capability is crucial for addressing noncompressible hemorrhage, the leading cause of preventable battlefield death, by facilitating the delivery of emergency care, including medications, analgesics, blood volume replacement, and/or resuscitative endovascular balloon occlusion of the aorta (REBOA). The AI-GUIDE is a highly portable platform technology designed to readily be adapted to other critical DOD needs, such as peripheral vascular access, tension pneumothorax drainage, and emergent airway cannulation, where rapid and accurate needle placement is crucial. In development since 2019 by MIT Lincoln Laboratory (MIT-LL) and Massachusetts General Hospital with support from the Combat Casualty Care Research Program (CCCRP), AI-GUIDE combines an ultrasound transducer, a low-cost custom handheld robotic assembly, a small computing system running custom on-board artificial intelligence (AI) algorithms, and a modular cartridge that includes intervention-specific components like a needle, guidewire, and cannula. This SBIR project aims to transition AI-GUIDE from the academic laboratories in which it originated to an FDA-cleared medical device. Key steps include (1) auditing, augmenting, and ultimately porting software, design files, documentation, and analysis from the MIT-LL prototype process into the AutonomUS ISO13485-compliant ecosystem, (2) enhancing device performance on an expanded number of vascular access sites through targeted sonographic data acquisition in 80 subjects followed by AI model development, (3) completing Conceptual and Preliminary Design Reviews (CoDR and PDR), and (4) capability demonstration on a perfused cadaver. This work will be conducted with an ISO13485-compliant medical device contract manufacturer, paving the way for further development and ultimately FDA approval.