Feasibility of Mixed Reality-assisted physician-modified endografts

This study aimed to develop and evaluate a standardized workflow for physician-modified endografts using Mixed Reality (MxR) technology, and to compare it with a conventional method in a phantom model experiment. The experiment consisted of two parts. In the first part, the feasibility of using a virtual MxR overlay to guide fenestration marking on a phantom stent graft was tested. Thirty-two observers each marked four fenestrations (F1-F4), totaling 128 markings. In the second part, 12 observers performed both the MxR-assisted workflow and a conventional method. Outcomes included positional accuracy, procedure time, usability, and number of required reattempts. Accuracy was assessed by comparing absolute and relative distances from ideal positions and fenestration centroids. The required time of the workflow was recorded. Usability was evaluated using the System Usability Scale (SUS) and the Post-Study System Usability Questionnaire (PSSUQ). In the feasibility test, mean deviations from ideal positions were 1.3 mm (F1), 1.0 mm (F2), 1.0 mm (F3), and 1.5 mm (F4). Centroid errors were 0.7 mm (clock position) and 0.5 mm (cranio-caudal). The workflow took 5:31 minutes on average, with step 2 requiring 2:55 minutes. SUS and PSSUQ scores indicated high usability (84.2/100 and 1.8/7, respectively). In the comparative analysis, the MxR group showed comparable accuracy to the conventional method but required fewer reattempts (3 vs 10) and less time (6.7 vs 14.6 minutes; P < .01). Usability ratings were significantly higher for MxR (SUS, 85.1 vs 43.2; P < .01). The MxR-assisted workflow enabled accurate, efficient, and user-friendly physician-modified endografts planning. At least in this experimental setup, it outperformed the conventional method in usability and speed, supporting its potential for broader clinical applications.