Automated diagnosis and classification of metacarpal and phalangeal fractures using a convolutional neural network: a retrospective data analysis study.

IF 2.5 2区 医学 Q1 ORTHOPEDICS
Michael Axenhus, Anna Wallin, Jonas Havela, Sara Severin, Ablikim Karahan, Max Gordon, Martin Magnéli
{"title":"Automated diagnosis and classification of metacarpal and phalangeal fractures using a convolutional neural network: a retrospective data analysis study.","authors":"Michael Axenhus, Anna Wallin, Jonas Havela, Sara Severin, Ablikim Karahan, Max Gordon, Martin Magnéli","doi":"10.2340/17453674.2024.42702","DOIUrl":null,"url":null,"abstract":"<p><strong>Background and purpose: </strong> Hand fractures are commonly presented in emergency departments, yet diagnostic errors persist, leading to potential complications. The use of artificial intelligence (AI) in fracture detection has shown promise, but research focusing on hand metacarpal and phalangeal fractures remains limited. We aimed to train and evaluate a convolutional neural network (CNN) model to diagnose metacarpal and phalangeal fractures using plain radiographs according to the AO/OTA classification system and custom classifiers.</p><p><strong>Methods: </strong> A retrospective analysis of 7,515 examinations comprising 27,965 images was conducted, with datasets divided into training, validation, and test datasets. A CNN architecture was based on ResNet and implemented using PyTorch, with the integration of data augmentation techniques.</p><p><strong>Results: </strong> The CNN model achieved a mean weighted AUC of 0.84 for hand fractures, with 86% sensitivity and 76% specificity. The model performed best in diagnosing transverse metacarpal fractures, AUC = 0.91, 100% sensitivity, 87% specificity, and tuft phalangeal fractures, AUC = 0.97, 100% sensitivity, 96% specificity. Performance was lower for complex patterns like oblique phalangeal fractures, AUC = 0.76.</p><p><strong>Conclusion: </strong> Our study demonstrated that a CNN model can effectively diagnose and classify metacarpal and phalangeal fractures using plain radiographs, achieving a mean weighted AUC of 0.84. 7 categories were deemed as acceptable, 9 categories as excellent, and 3 categories as outstanding. Our findings indicate that a CNN model may be used in the classification of hand fractures.</p>","PeriodicalId":6916,"journal":{"name":"Acta Orthopaedica","volume":"96 ","pages":"13-18"},"PeriodicalIF":2.5000,"publicationDate":"2025-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11714779/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Acta Orthopaedica","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.2340/17453674.2024.42702","RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"ORTHOPEDICS","Score":null,"Total":0}
引用次数: 0

Abstract

Background and purpose:  Hand fractures are commonly presented in emergency departments, yet diagnostic errors persist, leading to potential complications. The use of artificial intelligence (AI) in fracture detection has shown promise, but research focusing on hand metacarpal and phalangeal fractures remains limited. We aimed to train and evaluate a convolutional neural network (CNN) model to diagnose metacarpal and phalangeal fractures using plain radiographs according to the AO/OTA classification system and custom classifiers.

Methods:  A retrospective analysis of 7,515 examinations comprising 27,965 images was conducted, with datasets divided into training, validation, and test datasets. A CNN architecture was based on ResNet and implemented using PyTorch, with the integration of data augmentation techniques.

Results:  The CNN model achieved a mean weighted AUC of 0.84 for hand fractures, with 86% sensitivity and 76% specificity. The model performed best in diagnosing transverse metacarpal fractures, AUC = 0.91, 100% sensitivity, 87% specificity, and tuft phalangeal fractures, AUC = 0.97, 100% sensitivity, 96% specificity. Performance was lower for complex patterns like oblique phalangeal fractures, AUC = 0.76.

Conclusion:  Our study demonstrated that a CNN model can effectively diagnose and classify metacarpal and phalangeal fractures using plain radiographs, achieving a mean weighted AUC of 0.84. 7 categories were deemed as acceptable, 9 categories as excellent, and 3 categories as outstanding. Our findings indicate that a CNN model may be used in the classification of hand fractures.

使用卷积神经网络自动诊断和分类掌骨和指骨骨折:回顾性数据分析研究。
背景与目的:手部骨折常见于急诊科,但诊断错误持续存在,导致潜在的并发症。人工智能(AI)在骨折检测中的应用已经显示出前景,但针对手掌骨和指骨骨折的研究仍然有限。我们的目的是训练和评估卷积神经网络(CNN)模型,根据AO/OTA分类系统和自定义分类器使用x线平片诊断掌骨和指骨骨折。方法:回顾性分析包括27,965张图像的7,515次检查,数据集分为训练、验证和测试数据集。CNN架构基于ResNet,使用PyTorch实现,并集成了数据增强技术。结果:CNN模型对手部骨折的平均加权AUC为0.84,敏感性为86%,特异性为76%。该模型对掌骨横断骨折的诊断效果最好,AUC = 0.91,敏感性100%,特异性87%;对簇状指骨骨折的诊断效果最好,AUC = 0.97,敏感性100%,特异性96%。对于复杂类型如斜指骨折,AUC = 0.76,表现较差。结论:我们的研究表明,CNN模型可以有效地利用x线平片对掌骨和指骨骨折进行诊断和分类,平均加权AUC为0.84。7个类别为可接受,9个类别为优秀,3个类别为优秀。我们的研究结果表明,CNN模型可用于手部骨折的分类。
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
Acta Orthopaedica
Acta Orthopaedica 医学-整形外科
CiteScore
6.40
自引率
8.10%
发文量
105
审稿时长
4-8 weeks
期刊介绍: Acta Orthopaedica (previously Acta Orthopaedica Scandinavica) presents original articles of basic research interest, as well as clinical studies in the field of orthopedics and related sub disciplines. Ever since the journal was founded in 1930, by a group of Scandinavian orthopedic surgeons, the journal has been published for an international audience. Acta Orthopaedica is owned by the Nordic Orthopaedic Federation and is the official publication of this federation.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信