Abstract
Middle- and inner-ear surgery is a vital treatment option in hearing loss, infections, and tumors of the lateral skull base. Segmentation of otologic structures from computed tomography (CT) has many potential applications for improving surgical planning but can be an arduous and time-consuming task. We propose an end-to-end solution for the automated segmentation of temporal bone CT using convolutional neural networks (CNN). Using 150 manually segmented CT scans, a comparison of 3 CNN models (AH-Net, U-Net, ResNet) was conducted to compare Dice coefficient, Hausdorff distance, and speed of segmentation of the inner ear, ossicles, facial nerve and sigmoid sinus. Using AH-Net, the Dice coefficient was 0.91 for the inner ear; 0.85 for the ossicles; 0.75 for the facial nerve; and 0.86 for the sigmoid sinus. The average Hausdorff distance was 0.25, 0.21, 0.24 and 0.45 mm, respectively. Blinded experts assessed the accuracy of both techniques, and there was no statistical difference between the ratings for the two methods (p = 0.93). Objective and subjective assessment confirm good correlation between automated segmentation of otologic structures and manual segmentation performed by a specialist. This end-to-end automated segmentation pipeline can help to advance the systematic application of augmented reality, simulation, and automation in otologic procedures.
Highlights
Middle- and inner-ear surgery is a vital treatment option in hearing loss, infections, and tumors of the lateral skull base
This necessitates a mental translation of the data back into the three-dimensional (3D) relationships expected at the time of surgery
An automated pipeline of medical image segmentation for temporal bone computed tomography (CT) (TBCT) scans might expand the application of simulation, planning, and procedural automation
Summary
Middle- and inner-ear surgery is a vital treatment option in hearing loss, infections, and tumors of the lateral skull base. Safe and effective middle- and inner-ear surgery requires extensive training and knowledge of radiological and surgical anatomy Procedures such as cochlear implantation, tympanomastoidectomy, and superior semicircular canal dehiscence repair depend on the pre- and intra-operative identification of critical structures and an appreciation of their complex interrelationships[1]. Individualized preoperative planning and the implementation of augmented reality systems may assist in such surgery given the intricacy and variability of anatomy involved Such efforts require specialized anatomical and radiological knowledge of the key structures, which takes considerable time and effort to acquire. Efforts to enhance preoperative planning using innovative tools such as 3D simulations and augmented reality offer promise for improving operative safety and efficiency These efforts are limited by the labor intensive step of manual segmentation of imaging data[6,7] by highly trained specialists (Fig. 1). These methods require significant manual input, and may be limited in their scalability and subject to user-dependent variability
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
