Visual feedback-based interpretation support for retroperitoneal organ dissection in a VR surgical simulator: a mixed-methods study

In trauma surgery, retroperitoneal organ dissection is a critical but high-risk procedure because surgeons often rely on tactile sensations in an obscured operative field to avoid damaging hidden structures such as the ureter. Although virtual reality surgical simulators provide valuable training opportunities, most existing systems focus on visually dominant tasks and do not adequately support skills that require mapping tactile sensations to underlying anatomical risk. In this study, we developed a VR surgical simulator for retroperitoneal organ dissection training that combines a physical mock-up, force sensing, and X-ray visual feedback. This feedback was designed to visualize the contact state of a ureter-like structure in real time and to help learners associate their own tactile input with hidden contact states. Twenty-three participants were assigned to either a group that received X-ray visual feedback during practice or a group that performed the same task without visual feedback. The evaluation used a mixed-methods approach integrating participant-level force metrics, subjective questionnaires, and exploratory video-based behavioral analysis. Participant-level nonparametric analyses did not show consistent significant improvements attributable to X-ray visual feedback in mean applied force or hazard-zone intrusion rate. In contrast, supplementary linear mixed-effects modeling indicated group-dependent changes in mean applied force during the practice phase. The questionnaire showed between-group differences in the post-test, particularly for the item concerning the ability to return to a safe state after sensing risk. In addition, exploratory video analysis suggested that participants in the X-ray feedback group tended to stop before contact or correct their movement after contact, whereas some participants in the no-feedback group showed excessive caution or intrusion/slipping-through behavior. These results do not provide evidence that X-ray visual feedback uniformly improves objective performance, but they suggest that it may support tactile risk interpretation and self-correction strategies during short-term simulator training. This study highlights the need to combine mechanical metrics, subjective assessment, and behavioral observation when evaluating tactile-dependent VR surgical training.