Validation of a patient-specific virtual reality preoperative clip planning environment for microsurgical clipping of unruptured anterior circulation aneurysms: a blinded comparison between operative selection and postoperative imaging.
Saemann A, Brayenovitch O, et al. • Neurosurgical focus • 2026
This patient-specific VR environment for clip planning was feasible and fast, but only partially concordant with the operative clip construct, with stronger agreement for broad clip strategy than for exact blade length, recognition of nonstraight clip geometries, multiclip reconstruction, or remnant prediction.
Key Findings
Results
Overall concordance between blinded VR clip selections and the operative clip construct was modest.
Mean VR-Clip Concordance Index (VCCI) was 2.00 ± 1.11 on a 4-point scale.
10 of 30 simulations (33.3%) achieved VCCI ≥ 3 using exact primary blade-length agreement.
The study included 10 patients (6 female, median age 63.5 years [IQR 60.0-69.0]) with unruptured anterior circulation aneurysms and 30 simulations from 3 vascular neurosurgeons.
Raters did one anonymized VR session per case at least 6 months postoperatively, blinded to clinical data, operative documentation, clip records, and postoperative imaging.
Results
Agreement was higher for broad clip strategy than for specific clip parameters, with the lowest agreement for exact primary blade length.
Primary clip type agreement: 17 of 30 (56.7%).
Single- versus multiple-clip strategy agreement: 21 of 30 (70.0%).
Remnant expectation agreement: 18 of 30 (60.0%).
Exact primary blade length agreement: 4 of 30 (13.3%).
Results
Blade-length agreement improved only modestly when tolerances were allowed.
With ±0.5-mm tolerance, agreement was 5 of 30 (16.7%).
With an exploratory ±1.0-mm tolerance, agreement was 10 of 30 (33.3%).
Results
VR planning was fast.
Mean VR planning time was 2.23 ± 1.57 minutes per simulation.
The environment used life-size Yasargil clip models mirroring the institutional clip tray.
Results
Simulated clip positions were fairly close to postoperative clip positions, but rotational deviation was large and variable.
Median positional deviation between simulated and postoperative clip poses was 2.41 mm (IQR 1.67-3.70 mm).
Median rotational deviation was 44.42° (IQR 31.13°-64.95°).
Clip pose validation was based on postoperative day 1 CTA.
Results
Surgical outcomes on postoperative day 1 CTA were favorable.
9 of 10 patients had completely occluded aneurysms.
1 patient had a small dog-ear remnant.
Conclusions
The authors conclude the VR environment is feasible but only partially concordant, and identify engineering targets before prospective clinical use.
Priority targets: exact blade length, recognition of nonstraight clip geometries, multiclip reconstruction, and remnant prediction.
The study was a single-center retrospective validation with only 10 patients, 2021-2025.
What This Means
This study tested a virtual reality (VR) tool that lets neurosurgeons practice choosing the metal clip used to seal off a brain aneurysm before operating. Using scans from 10 patients who had unruptured aneurysms clipped, three neurosurgeons each ran a VR planning session per case, without knowing what clip was actually used or how the surgery turned out. Their choices were then compared with the real operation and with CT scans taken the day after surgery.
The tool was quick (about 2 minutes per case) and agreement was best for broad decisions, such as whether one clip or several would be needed (70%). Agreement was weaker for the type of clip (57%) and predicting a small leftover remnant (60%), and was poor for the exact clip blade length (13%). Simulated clip positions were typically about 2.4 mm from the real ones, but their angles often differed substantially (median about 44°). In the real surgeries, 9 of 10 aneurysms were completely sealed and one had a small remnant.
This research suggests the VR tool is workable and fast but not yet accurate enough in detail to reproduce actual surgical clip choices. The authors point to improving blade-length accuracy, curved clip shapes, multi-clip constructs, and remnant prediction before testing it prospectively. The small sample from one center means the findings should be viewed as preliminary.
Saemann A, Brayenovitch O, Żelechowski M, Faludi B, de Wilde D, Cattin P, et al.. (2026). Validation of a patient-specific virtual reality preoperative clip planning environment for microsurgical clipping of unruptured anterior circulation aneurysms: a blinded comparison between operative selection and postoperative imaging.. Neurosurgical focus. https://doi.org/10.3171/2026.8.FOCUS26434