Autonomous X-ray acquisition and mixed-reality fixation guidance
An autonomous, AI-driven X-ray navigation system for percutaneous pelvic fracture fixation. It uses machine learning and robotic imaging to optimize surgical trajectories and provide real-time, mixed-reality guidance, aiming to reduce radiation exposure and operating time while improving placement accuracy for surgeons.
Overview
This autonomous X-ray acquisition and interpretation system is designed to streamline percutaneous pelvic fracture fixation. By integrating robotic X-ray imaging with machine learning, the system automates image acquisition and provides real-time guidance, significantly reducing surgical time and radiation exposure. It assists surgeons by optimizing imaging viewpoints, detecting bony corridors and surgical instruments, and providing immersive mixed-reality visualization, which helps prevent complications like cortical breaches.
Technical specifications
The system employs a deep neural network to detect instruments (such as K-wires) and target bony corridors in radiographs. Key features include:
- Automated viewpoint planning: Robotic C-arm movement is triggered by initial image analysis to select optimal trajectories, minimizing the need for manual C-arm adjustments.
- Mixed-reality guidance: A spatially registered optical see-through head-mounted display allows clinicians to visualize both the planned trajectory and current instrument position in real-time.
- Breach detection: The system automatically estimates the risk of cortical breach by comparing the current trajectory against the ideal anatomical corridor.
- Compatibility: Designed for seamless integration into standard orthopaedic workflows, leveraging common mobile C-arm imaging systems.
Technology readiness level
TRL 4-5. The technology has been validated through in silico evaluations, phantom studies with attending surgeons, and cadaveric testing. These studies have demonstrated millimeter-scale accuracy (e.g., approximately 2.8 mm and 2.7 degrees in cadaver specimens) and reduced procedural imaging requirements compared to standard clinical practice. Research code and data have been published, and the system is currently undergoing further refinement and evaluation for clinical translation.
About Johns Hopkins University
Johns Hopkins University is a comprehensive private research university headquartered in Baltimore with a large research enterprise and a culture of discovery and translation. Integrated with a major academic health system, the university offers access to clinicians, patient populations, biorepositories, and audited trial operations. A research park near the medical campus, plus startup incubators and corporate engagement teams, enable co-location, pilots, and fast iteration. A major university‑affiliated research center expands applied R&D and prototyping capacity, and proximity to Washington, DC links partners to federal agencies and regulators. Research is supported by federal funding from NIH, NSF, DoD, and others, and a dedicated technology transfer office manages IP, licensing, and startup formation.
delivered to your inbox.