The development of patient-specific 3D anatomical models in minimally invasive parathyroidectomy

Surgery is the preferred treatment for primary hyperparathyroidism. Minimally invasive parathyroidectomy is only feasible with accurate preoperative localisation. Virtual 3D anatomical models can be constructed from patient-specific CT scans using segmentation software.

We aimed to create virtual 3D models from 4D-CT scans of parathyroid tumours using segmentation technology. We designed a small pilot study to assess the utility of 3D models within surgical practice. We assessed surgeon, trainee and patients' opinion and satisfaction with the models. The NASA TLX survey was the primary data collection tool.

Creation of novel 3D models was achieved, these featured a 360-degree axis of rotation and transparency mode to assist in surgical planning. Models were used intraoperatively with the HoloLens 2 headset to locate parathyroid tumours real time before surgery. Total mean workloads for surgery planning when averaged revealed a decrease workload (39.45 vs 27.45) points with adjunctive use of models (p=0.002). Mental demand showed the greatest decrease in mean workload out of all the 6 subscales tested for in the NASA TLX (210.3 vs 136.7) points. Patient satisfaction score was statistically significant for the difference before and after seeing the 3D model regarding anatomical location (p=≤0.001).

In this work, we developed patient-specific virtual 3D anatomical models of parathyroid tumours for use in surgery using novel techniques, previously never applied to parathyroidectomy. Our initial success in model construction and subsequent opinion of surgeons, trainees and patients contributes to the developing body of literature in favour of virtual modelling for parathyroidectomy.