Precise evaluation of the tympanic membrane (TM) is required for accurate diagnosis of middle ear diseases. However, making an accurate assessment is sometimes difficult. Artificial intelligence is often employed for image processing, especially for performing high level analysis such as image classification, segmentation and matching. In particular, convolutional neural networks (CNNs) are increasingly used in medical image recognition. This study demonstrates the usefulness and reliability of CNNs in recognizing the side and perforation of TMs in medical images. CNN was constructed with typically six layers. After random assignment of the available images to the training, validation and test sets, training was performed. The accuracy of the CNN model was consequently evaluated using a new dataset. A class activation map (CAM) was used to evaluate feature extraction. The CNN model accuracy of detecting the TM side in the test dataset was 97.9%, whereas that of detecting the presence of perforation was 91.0%. The side of the TM and the presence of a perforation affect the activation sites. The results show that CNNs can be a useful tool for classifying TM lesions and identifying TM sides. Further research is required to consider real-time analysis and to improve classification accuracy.
Background: Although CT has been used widely, the role of preoperative CT findings including other factors in tympanoplasty has not been elucidated comprehensively. Aims/objectives: To evaluate relationship of CT findings with other factors and audiological results in type 1 tympanoplasty. Material and methods: A cohort of consecutive 175 patients with type 1 tympanoplasty was enrolled. Addition of mastoidectomy was based on the presence of soft tissue in antrum on CT. Postoperative air–bone gap (ABG) and reperforation rate were analyzed. Results: Positive soft tissue in antrum on CT was found in 52 (29.7%) patients and showed larger preoperative ABG than the negative group. Successful ABG closure (≤20 dB) was obtained in 97% when preoperative ABG ≤20 dB, but it decreased as the preoperative ABG increased (83% with preoperative ABG of 21–30 dB, and 0% with preoperative ABG >30 dB). Postoperative reperforation rate was positively related to the preoperative ABG, but not the presence of soft tissue in the antrum, the size, or locations of preoperative perforations. Conclusions and significance: Our findings showed that temporal bone CT was helpful in determining addition of mastoidectomy and the presence of soft tissue in the antrum was associated with large ABG.
1 January 2015 in Korean Journal of Otorhinolaryngology-Head and Neck Surgery