Human–Computer Vision Collaborative Measurement of Surgical Exposure and Length in Endonasal Endoscopic Skull Base Surgery

Objective: This study aimed to develop and validate a computer vision (CV)-based system to quantitatively analyze surgical exposure in endonasal endoscopic approach (EEA). Results: The number of pixels of the length or area of interest in the selected frame in the EEA video was measured using a reference instrument. The measured length and area were calibrated by training the current algorithm using EEA videos. A total of 50 EEA operative videos were analyzed, with 95.1%, 95.8%, and 96.2% accuracies in the training, test-1 and test-2 datasets, respectively. The CV-base model was validated using intercarotid distance and sellar height. Compared to neuronavigation, CV-based analysis reduced the time required for area measurement by 89% (p < 0.001). Our CV-based analysis showed that a smaller lateral (p = 0.001) and area (p = 0.024) surgical exposure were associated with residual tumors. Conclusions: CV-based analysis can accurately measure the surgical exposure in EEA videos and reduce the time required to measure surgical areas. The application of AI and CV can expedite quantitative analysis of surgical exposure in EEA surgeries.