Abstract
OBJECTIVE: The primary aim of this study was to investigate the accuracy of a semi-automatic algorithm in assessing the feasibility and complexity of endoscopic stapes surgery preoperatively. METHODS: A semi-automatic algorithm was developed to simulate endoscopic stapes surgery in 3D. To test the accuracy of the algorithm, five fresh-frozen cadaveric heads (ten ears) were used. Each head was CT-scanned, followed by segmentation and 3D reconstruction of the bones including the ear canal, middle ear, and ossicular chain. Two interventions were tested on each ear. Initially, the algorithm was digitally employed to simulate the endoscopic ear surgery. Subsequently, the actual endoscopic ear surgery was performed. Primary outcomes for each intervention included the amount of scutum removal (postoperative 3D model subtracted from preoperative 3D model), visibility of important landmarks, feasibility score, and a complexity questionnaire. Finally, the outcomes of the pre-operative planning and the actual procedure were directly compared to evaluate the algorithm's accuracy and usability. RESULTS: The preoperative planning method achieved an accuracy rate of 70% in predicting the need for scutum removal. The mean volume of the removed scutum was 0.5 mm(3). Endoscopic surgery was feasible in all ten ears, with all relevant anatomical landmarks adequately visualized as estimated by the preoperative planning algorithm. CONCLUSION: Preoperative planning models can assist ENT surgeons in evaluating the feasibility and complexity of endoscopic stapes surgery. With adequate training and testing on clinical cases, these models can significantly improve their predictive accuracy and thereby improve patient outcomes.