Automated detection of otosclerosis with interpretable deep learning using temporal bone computed tomography images

Zheng Wang; Jian Song; Kaibin Lin; Wei Hong; Shuang Mao; Xuewen Wu; Jianglin Zhang

doi:10.1016/j.heliyon.2024.e29670

Automated detection of otosclerosis with interpretable deep learning using temporal bone computed tomography images

Heliyon. 2024 Apr 15;10(8):e29670. doi: 10.1016/j.heliyon.2024.e29670. eCollection 2024 Apr 30.

Authors

Zheng Wang^{1

2}, Jian Song^{3

4}, Kaibin Lin^{1

2}, Wei Hong^{1

2}, Shuang Mao^{3

4}, Xuewen Wu^{3

4}, Jianglin Zhang^{5

6

7}

Affiliations

¹ School of Computer Science, Hunan First Normal University, Changsha, 410205, China.
² Key Laboratory of Informalization Technology for Basic Education in Hunan Province, Changsha, 410205, China.
³ Department of Otorhinolaryngology, Xiangya Hospital Central South University, Changsha, Hunan, China.
⁴ Province Key Laboratory of Otolaryngology Critical Diseases, Changsha, Hunan, China.
⁵ Department of Dermatology, Shenzhen People's Hospital, The Second Clinical Medical College, Jinan University. The First Affiliated Hospital, Southern University of Science and Technology, Shenzhen, 518020, Guangdong, China.
⁶ Candidate Branch of National Clinical Research Center for Skin Diseases, Shenzhen, 518020, Guangdong, China.
⁷ Department of Geriatrics, Shenzhen People's Hospital, The Second Clinical Medical College, Jinan University. The First Affiliated Hospital, Southern University of Science and Technology, Shenzhen, 518020, Guangdong, China.

Abstract

Objective: This study aimed to develop an automated detection schema for otosclerosis with interpretable deep learning using temporal bone computed tomography images.

Methods: With approval from the institutional review board, we retrospectively analyzed high-resolution computed tomography scans of the temporal bone of 182 participants with otosclerosis (67 male subjects and 115 female subjects; average age, 36.42 years) and 157 participants without otosclerosis (52 male subjects and 102 female subjects; average age, 30.61 years) using deep learning. Transfer learning with the pretrained VGG19, Mask RCNN, and EfficientNet models was used. In addition, 3 clinical experts compared the system's performance by reading the same computed tomography images for a subset of 35 unseen subjects. An area under the receiver operating characteristic curve and a saliency map were used to further evaluate the diagnostic performance.

Results: In prospective unseen test data, the diagnostic performance of the automatically interpretable otosclerosis detection system at the optimal threshold was 0.97 and 0.98 for sensitivity and specificity, respectively. In comparison with the clinical acumen of otolaryngologists at P < 0.05, the proposed system was not significantly different. Moreover, the area under the receiver operating characteristic curve for the proposed system was 0.99, indicating satisfactory diagnostic accuracy.

Conclusion: Our research develops and evaluates a deep learning system that detects otosclerosis at a level comparable with clinical otolaryngologists. Our system is an effective schema for the differential diagnosis of otosclerosis in computed tomography examinations.

Keywords: Area under the receiver operating characteristic curve; Computed tomography; Deep learning; Interpretability; Temporal bone computed tomography.