AI-Powered Smartphone Application for Measuring Hallux Valgus Angle From Radiographs Displayed on a Monitor

一款利用人工智能技术的智能手机应用程序,可通过显示器上显示的X光片测量拇外翻角度

阅读:2

Abstract

INTRODUCTION: We developed a smartphone application capable of automatically measuring the hallux valgus angle (HVA) and various intermetatarsal angles by capturing radiographic images displayed on a monitor. This study aimed to evaluate the accuracy of these measurements using the application. METHODS: Three users-a board-certified orthopedic surgeon, a resident, and a nonhealthcare professional (Users 1, 2, and 3)-independently used the application to measure angles on 92 radiographs from 92 consecutive patients. Mean absolute errors (MAEs) between the application-based measurements and the median of manual measurements performed by three experienced foot and ankle surgeons using a DICOM viewer were calculated for each user. To evaluate whether the measurement errors were acceptably small, one-sided t-tests were conducted to determine whether the MAEs were significantly less than 3°. Differences in MAEs among the three users were also assessed using analysis of variance. RESULTS: The MAEs of HVA by the three users were 1.1°, 1.3°, and 1.4°, respectively, all significantly below the 3° threshold (95% CI upper limit; 1.2°, 1.5°, and 1.5°). Comparable accuracy was observed for intermetatarsal angles, which have slightly greater variability for more lateral metatarsals. All measurements met the accuracy criterion of < 3°, except for the intermetatarsal angles between the first and fifth metatarsals measured by the nonhealthcare user. No significant difference in MAE was found among users for HVA (p = 0.13), whereas significant differences were noted for some intermetatarsal angles. CONCLUSIONS: The developed smartphone application accurately measured the HVA and various intermetatarsal angles, with performance comparable to that of experienced foot and ankle surgeons. Importantly, sufficient accuracy was achieved even when used by individuals without clinical training. The application may be useful as a practical tool in clinical and research settings.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。