Rapid and reliable alpha angle measurement using a smartphone-based artificial intelligence system for cam-type femoroacetabular impingement

IF 2.7 Q2 ORTHOPEDICS

Journal of Experimental Orthopaedics Pub Date : 2025-09-15 DOI:10.1002/jeo2.70418

Masayoshi Saito, Hiroyuki Ogawa, Takuya Kusakabe, Naoyuki Hirasawa, Sachiyuki Tsukada

{"title":"Rapid and reliable alpha angle measurement using a smartphone-based artificial intelligence system for cam-type femoroacetabular impingement","authors":"Masayoshi Saito, Hiroyuki Ogawa, Takuya Kusakabe, Naoyuki Hirasawa, Sachiyuki Tsukada","doi":"10.1002/jeo2.70418","DOIUrl":null,"url":null,"abstract":"<div>\n \n \n <section>\n \n <h3> Purpose</h3>\n \n <p>Manual alpha angle measurement for diagnosing cam-type femoroacetabular impingement (FAI) is time-consuming and variable. We hypothesised that a smartphone-based artificial intelligence (AI) system would demonstrate a high correlation with the conventional manual method.</p>\n </section>\n \n <section>\n \n <h3> Methods</h3>\n \n <p>We trained an AI model using 300 of the 45° Dunn view radiographs in a Python-based environment. The model was implemented using a smartphone application that semi-automatically extracted the femoral head from radiographs, detected at least four rim points, and calculated the femoral head centre and radius using a least-squares method. It also identified the narrowest part of the femoral neck, and allowed manual adjustment of the alpha angle deviation. The images were directly captured and transferred without time loss. We evaluated 130 hips of patients with hip pain who underwent 45° Dunn view radiography. The alpha angle was measured using both the AI system and the manual method. Pearson correlation coefficient was used to assess agreement between the two methods. Measurement time and intra- and interobserver reliability were assessed using intraclass correlation coefficients (ICC).</p>\n </section>\n \n <section>\n \n <h3> Results</h3>\n \n <p>Pearson correlation coefficient for alpha angle measurement between the two methods was 0.91, indicating a strong correlation. The average alpha angle measured by the AI system (52.7 ± 6.0°) was not significantly different from that of the conventional method (53.9 ± 6.5°, <i>p</i> = 0.12). The AI-powered system showed intra- and interobserver agreement ICCs of 0.96 and 0.93, respectively, demonstrating excellent reliability. The mean measurement time was 5.0 ± 2.4 and 46.7 ± 10.1 s for the AI-powered system and manual method, respectively, indicating significant time-saving.</p>\n </section>\n \n <section>\n \n <h3> Conclusion</h3>\n \n <p>The AI-powered system showed a strong correlation with the manual method and significantly reduced measurement time. This smartphone-based tool offers a rapid and reliable approach for real-time assessment of alpha angles and may help standardise the evaluation of cam-type FAI in clinical practice.</p>\n </section>\n \n <section>\n \n <h3> Level of Evidence</h3>\n \n <p>Level III, retrospective comparative study.</p>\n </section>\n </div>","PeriodicalId":36909,"journal":{"name":"Journal of Experimental Orthopaedics","volume":"12 3","pages":""},"PeriodicalIF":2.7000,"publicationDate":"2025-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://esskajournals.onlinelibrary.wiley.com/doi/epdf/10.1002/jeo2.70418","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Experimental Orthopaedics","FirstCategoryId":"1085","ListUrlMain":"https://esskajournals.onlinelibrary.wiley.com/doi/10.1002/jeo2.70418","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"ORTHOPEDICS","Score":null,"Total":0}

引用次数: 0

Abstract

Purpose

Manual alpha angle measurement for diagnosing cam-type femoroacetabular impingement (FAI) is time-consuming and variable. We hypothesised that a smartphone-based artificial intelligence (AI) system would demonstrate a high correlation with the conventional manual method.

Methods

We trained an AI model using 300 of the 45° Dunn view radiographs in a Python-based environment. The model was implemented using a smartphone application that semi-automatically extracted the femoral head from radiographs, detected at least four rim points, and calculated the femoral head centre and radius using a least-squares method. It also identified the narrowest part of the femoral neck, and allowed manual adjustment of the alpha angle deviation. The images were directly captured and transferred without time loss. We evaluated 130 hips of patients with hip pain who underwent 45° Dunn view radiography. The alpha angle was measured using both the AI system and the manual method. Pearson correlation coefficient was used to assess agreement between the two methods. Measurement time and intra- and interobserver reliability were assessed using intraclass correlation coefficients (ICC).

Results

Pearson correlation coefficient for alpha angle measurement between the two methods was 0.91, indicating a strong correlation. The average alpha angle measured by the AI system (52.7 ± 6.0°) was not significantly different from that of the conventional method (53.9 ± 6.5°, p = 0.12). The AI-powered system showed intra- and interobserver agreement ICCs of 0.96 and 0.93, respectively, demonstrating excellent reliability. The mean measurement time was 5.0 ± 2.4 and 46.7 ± 10.1 s for the AI-powered system and manual method, respectively, indicating significant time-saving.

Conclusion

The AI-powered system showed a strong correlation with the manual method and significantly reduced measurement time. This smartphone-based tool offers a rapid and reliable approach for real-time assessment of alpha angles and may help standardise the evaluation of cam-type FAI in clinical practice.

Level of Evidence

Level III, retrospective comparative study.

Abstract Image

查看原文本刊更多论文

基于智能手机的人工智能系统对凸轮型股髋臼撞击进行快速可靠的α角测量

目的人工测量α角诊断凸轮型股髋臼撞击（FAI）费时且易变。我们假设基于智能手机的人工智能（AI）系统将与传统的手动方法表现出高度相关性。方法在基于python的环境中使用300张45°Dunn视图x线片训练AI模型。该模型使用智能手机应用程序实现，该应用程序可以半自动地从x线片中提取股骨头，检测至少四个边缘点，并使用最小二乘法计算股骨头中心和半径。它还确定了股骨颈最窄的部分，并允许手动调整α角偏差。图像直接捕获和传输，没有时间损失。我们评估了130例髋关节疼痛患者，他们接受了45°邓恩透视片。α角测量采用人工智能系统和人工方法。使用Pearson相关系数来评估两种方法之间的一致性。采用类内相关系数（ICC）评估测量时间和观察者内部和观察者之间的信度。结果两种方法测量α角的Pearson相关系数为0.91，相关性较强。人工智能系统测得的平均α角（52.7±6.0°）与常规方法测得的平均α角（53.9±6.5°）无显著差异（p = 0.12）。人工智能驱动的系统显示观察者内部和观察者之间的一致性icc分别为0.96和0.93，显示出出色的可靠性。人工智能系统和人工方法的平均测量时间分别为5.0±2.4 s和46.7±10.1 s，节省了大量时间。结论人工智能系统与人工方法具有较强的相关性，可显著缩短测量时间。这种基于智能手机的工具提供了一种快速可靠的方法来实时评估α角，并可能有助于在临床实践中标准化对cam型FAI的评估。证据等级III级，回顾性比较研究。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊