Grant: Discovery Grants Program - Individual

X-ray fluoroscopy-guided minimally invasive procedures, such as tumour ablation and various cardiovascular procedures, are routinely performed in Canada. These procedures aim to precisely navigate surgical instruments, such as ablation needles or catheters, to treatment sites while minimizing collateral damage to surrounding tissues. Despite the expertise of physicians, these procedures remain complex, necessitating mastery of difficult skill sets and adherence to intricate procedural workflow. Challenges arise from the inherent limitation of the 2D projection and the utilization of ionizing radiation, resulting in minimal soft-tissue contrast and the lack of depth perception, leading to prolonged procedural time and excessive radiation exposure. To address these challenges, we propose an engineering solution of integrating a 3D magnetic spatial tracker into the clinical fluoroscopy imaging system. By leveraging the machine-learning-based video processing and 3D pose tracking of the C-arm fluoroscopy imaging system and surgical instruments, the augmentation of a spatial tracker will significantly enhance surgical outcomes by enabling advanced 3D visualization, registration between 2D intro-procedural fluoroscopy video with 3D pre-procedure imaging data, and monitoring the real-time placement of surgical instruments in 3D. The machine-learning-based video-processing techniques form the pivotal component of our proposal, laying the foundation for subsequent task-specific computer-assisted interventions. The integration of spatial tracking systems into fluoroscopy-guided integration remains underexplored due to several engineering challenges. Optical tracking systems require line-of-sight and are unable to track flexible instruments seated subcutaneously. Conversely, the magnetic tracking system is well-suited for minimally-invasive procedures, but is susceptible to ferromagnetic interference and causes severe imaging artifacts in fluoroscopy. Working with our industrial partner, we are in possession of a novel magnetic tracking system constructed with radiolucent materials. We have developed generalized machine-learning approaches for video processing that removes residual imaging artifacts from fluoroscopy video caused by the tracking system, and real-time pose tracking of the C-arm fluoroscopy using magnetic tracker. Using these approaches, the objective of my research program is to seamlessly integrate this novel magnetic spatial tracker into the fluoroscopy imaging system. The intended contribution of the proposed program will be the development of methods and software tools in support of cross-disciplinary research of fluoroscopy-guided, magnetically-tracked, surgical navigation approach for minimally invasive procedures. Engineering and computer science research that assists clinicians in their surgical execution has the potential to improve quality of life and economical benefits to our society.