Background: Sentinel lymph node biopsy (SLNB) is standard for staging high-risk melanoma, but current mapping provides limited spatial guidance. This pilot study assessed whether augmented reality (AR) projected 3D models can improve sentinel node localization compared with the gamma probe.
Methods: In this prospective study, 10 melanoma patients undergoing SLNB at Sheba Medical Center had preoperative sentinel lymphoscintigraphy using SPECT/CT imaging segmented to generate patient-specific 3D models. Models were projected onto the patient via an AR headset before incision. Localization accuracy was measured as the deviation (mm) between AR3D-identified and gamma probe-identified sentinel lymph node positions.
Results: Ten patients were enrolled in this pilot study. The median age was 71 years (range 30-77). Primary tumor sites included upper limb (n = 5), trunk (n = 3), Lower Limb (n = 1), and Head and Neck (n = 1). The median Breslow thickness was 1.1 mm (range 0.8-24 mm), with ulceration in 2 cases. Model generation was successful in all patients, with each model projected onto the patient using the AR headset and evaluated by the surgeon prior to incision. The median deviation between AR3D model and the gamma probe localization was 12 mm (range 0-40 mm), with 43% of cases ≤ 10 mm. No AR-related complications occurred.
Conclusions: AR-based 3D modelling was feasible and safe for SLNB localization in melanoma. Although accuracy varied, in nearly half of the cases, AR3D model was within 10 mm of the gamma probe detection. These preliminary results supported further refinement of the technique and evaluation in larger, multicenter trials.
© 2026 Wiley Periodicals LLC.