Mixed-Reality-Assisted Physician-Modified Stent Grafts: An Experimental Pre-Clinical Feasibility Study Using the Valiant Captivia and Endurant II Stent Graft Systems

Objectives: Physician-modified endografts (PMEGs) expand endovascular treatment options in urgent or cost-sensitive settings where industry-provided custom-made devices (CMDs) are not available. Current PMEG manufacturing techniques are time-consuming, lack standardization, and often require repeated adjustments to achieve strut-free fenestration positioning. Mixed reality (MxR) may streamline this process by overlaying virtual templates directly onto the physical stent graft guiding fenestration positioning. Methods: We developed a standardized MxR-assisted workflow for four-fenestrated PMEG preparations and compared it to a conventional marking technique. In this experimental set-up, between May 2025 and July 2025, three stent grafts were evaluated (Endurant II® 28 mm, Valiant Captivia® 30 mm, and Valiant Captivia® 32 mm). Five observers performed fenestration marking on 20 grafts per device type (10 per method), resulting in 60 PMEGs and 240 fenestrations. Outcomes included absolute positional error, relative positional error, number of strut-free fenestrations, number of re-attempts to achieve strut-free configuration, time required, and usability assessed via the System Usability Scale (SUS). Results: Across 240 fenestrations, both methods achieved high accuracy. Median absolute errors ranged from 0 to 1.25 mm for the conventional method and 0 to 1.75 mm for MxR. Relative positional errors were similarly small, with no significant differences between methods. MxR achieved higher rates of strut-free fenestration in the 28 mm Endurant II® device. Re-attempts were fewer with MxR. Median procedure time was significantly reduced for the MxR-assisted workflow in Valiant Captivia 30 mm (5.0 vs. 9.8 min, p = 0.049) and 32 mm (5.6 vs. 8.2 min, p = 0.049) while a trend was observed for Endurant II (7.5 vs. 15.6 min, p = 0.066). SUS scores favored MxR (76.2 vs. 62.6), though not significantly. Conclusions: The MxR-assisted PMEG production workflow seems promising in this pre-clinical, experimental study and warrants continued development and investigation.