Visualization for the Planning and Guidance of Minimally Invasive Cement Injection in Orthopaedics

Abstract

Minimally invasive refixation by cement injection is a novel technique, performed at the Leiden University Medical Center for the revision of failed hip implants. During this procedure, the surgeon inserts needles into the periprosthetic space of the femur, to inject new cement for the fixation of the prosthesis. For that, two intra-operative imaging modalities are used: computed tomography (CT) and fluoroscopy. Still, this combination is not optimal, because of the inherent disadvantages of both modalities. The surgeons would appreciate a better workflow, to avoid computed tomography intra-operatively and to follow a fluoroscopic-based procedure, only. In striving towards this goal, we implemented HipRFX, an integrated system for planning and guiding minimally invasive refixation. In our planning approach, the surgeon performs the operation virtually, using data from a single pre-operative CT scan. Then, the system uses the information from the plan, to simulate pseudo-fluoroscopic images. These images realistically imitate how the plan would ideally look on the fluoroscopic monitor. Intra-operatively, we propose a stepwise guidance approach. The surgeon uses the two-dimensional simulated radiographs with additional information embedded, as guidelines for the operation. Post-operatively, HipRFX offers cement filling uncertainty visualization, for the assessment of the outcome of the surgery. In order to test our implementation, we conducted a domain expert review, during which we simulated the visual environment of a refixation procedure. Additionally, we performed other user-in-the-loop validations, concerning the user interface design and, especially, cement filling uncertainty visualization. The final evaluation of the HipRFX workflow was an in situ experiment, on 5 cadaver legs. Our main contribution is a first practical investigation of a complete system, to assist orthopedic surgeons in planning, guiding and assessing minimally invasive refixation procedures. This new workflow has advantages over the current approach, where CT and fluoroscopy are combined, but also over a fluoroscopic-only approach. Our system is based on a single pre-operative CT scan and does not require intra-operative scanning, or new hardware. Moreover, we introduced cement filling uncertainty visualization, as a means of post-operative assessment, based on pre-operative information. To our knowledge, this is a novel domain that can offer new insight into the cement filling of the periprosthetic space. Our secondary contribution emerges from the results of our evaluations, which can be used in further research.