The micro made macro: Microsurgery at human scales

PROJECT SUMMARY Microsurgery is at the forefront of modern surgical practice. The quest for ultra-precise resection of invading tumors, durable anastomosis of severed neurovascular structures, and overall minimization of invasiveness has driven the scale of surgical maneuvers below a millimeter and firmly into the microsurgical domain. Applications of microsurgery abound, commensurate with the potential patient benefits, including tissue sparing, better margins, function restoration, and reduced pain. Uniting its diverse procedures is the defining characteristic of microsurgery and the source of its challenges: the sub-millimeter scale of surgical maneuvers and target tissues. The human motor and perception systems are ill-equipped to operate in this regime, and microsurgeons thus reach intrinsic limits of human performance and encounter shortcomings of surgical aids at microsurgical scales. Fortunately, advances in robotics, virtual reality (VR), and artificial intelligence (AI) promise to solve these fundamental challenges from manipulation, visualization, and workflow standpoints, respectively. These incredible advances and, in some cases, commercial products fail microsurgeons from the perception standpoint, however. The project therefore proposes to develop an adaptive perception system based on optical coherence tomography (OCT) to satisfy this unmet need in microsurgery. Specifically, adaptive perception leverages an AI-derived understanding of the surgical scene and its dynamics to build a high-fidelity 3D reconstruction from intelligently distributed image acquisition effort. Critically, OCT is the only live volumetric medical imaging modality routinely acquired during surgery and thus presents a unique opportunity in microsurgery for 3D surgical feedback. Leveraging adaptive perception to guide surgery for natural and artificial agents, this project will create an OCT-integrated robotic microsurgery system capable of autonomous and VR tele-microsurgery modes. Its core capabilities include an OCT-based adaptive perception system that predicts and responds to the microsurgeon’s or AI’s visualization needs, immersion of the microsurgeon in a life-size VR environment constructed from live volumetric OCT upsized 100-1000x to human scale, and AI-guided automated microsurgical suturing without fiducial markers. This project further proposes to compare this system to conventional techniques in simulated and ex vivo microsurgical procedures. This platform promises to empower microsurgeons to complete their tasks faster, commit fewer technical errors, require less training to reach proficiency, and enjoy improved ergonomics. These benefits would translate into the downstream patient benefits of improved outcomes, reduced anesthesia time, faster recoveries, and easier access to care for hundreds of thousands of microsurgery patients each year. These innovations and applications promise to transform microsurgery in ways that could not be envisioned even a few years ago.