文献速递关节外科

人工智能用于全髋关节置换术中植入物自动识别:一项超过两百万张平片的 multicenter 外部验证研究

Artificial Intelligence for Automated Implant Identification in Total Hip Arthroplasty: A Multicenter External Validation Study Exceeding Two Million Plain Radiographs.
PubMed 文献 · J Arthroplasty · 2022-03-07

该研究在3个学术转诊中心、13名外科医生回顾性收集的2954张髋关节正位平片上训练、验证并外部测试了一个深度学习系统,用于将股骨侧全髋关节置换(THA)植入物分类为2家制造商的8种模型之一。外部测试集588张平片显示平均受试者工作特征曲线下面积0.991、准确率97.9%、敏感度88.6%、特异度98.9%,平均分类速度0.02秒/张。作者认为该人工智能软件内部与外部验证表现优异,随着植入物库扩展需持续监测,具有在翻修THA术前规划中即时规模化应用的潜力。

要点

  • 深度学习系统在588张外部测试平片上识别8种股骨侧THA植入物,平均AUC 0.991、准确率97.9%。
  • 外部测试集敏感度88.6%、特异度98.9%,平均分类速度0.02秒/张。
  • 训练集与测试集植入物人群存在统计学差异(P < .001),作者认为需随植入物库扩展持续监测。

核心结论

  • 基于人工智能的软件在内部与外部验证中均表现优异,作者认为这是人工智能负责任且有意义的临床应用,具有在翻修THA术前规划中全球规模化应用的即时潜力,但随植入物库扩展需持续监测。

研究设计与样本

  • 研究为多中心外部验证研究,在3个学术转诊中心、13名外科医生回顾性收集的2954张原始去标识化髋关节正位平片上训练、验证并外部测试深度学习系统,将股骨侧THA植入物分类为2家制造商的8种模型之一。
  • 2954张平片中2117张用于训练、249张用于验证、588张用于外部测试;训练集经增强扩充至2117000张以增加模型稳健性。

主要发现

  • 训练集与测试集来自统计学上不同的植入物人群(P < .001)。
  • 深度学习系统经1000个训练轮次后,在588张正位平片的外部测试数据集上区分8种植入物模型,平均受试者工作特征曲线下面积为0.991、准确率97.9%、敏感度88.6%、特异度98.9%。
  • 软件平均分类速度为0.02秒/张图像。

临床怎么用

  • 作者认为该软件代表了人工智能负责任且有意义的临床应用,具有在翻修THA术前规划中全球规模化应用的即时潜力,但随植入物库扩展需持续监测。

局限性

  • 训练集与测试集来自统计学上不同的植入物人群(P < .001),作者指出随植入物库扩展需持续监测。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

BACKGROUND: The surgical management of complications after total hip arthroplasty (THA) necessitates accurate identification of the femoral implant manufacturer and model. Automated image processing using deep learning has been previously developed and internally validated; however, external validation is necessary prior to responsible application of artificial intelligence (AI)-based technologies. METHODS: We trained, validated, and externally tested a deep learning system to classify femoral-sided THA implants as one of the 8 models from 2 manufacturers derived from 2,954 original, deidentified, retrospectively collected anteroposterior plain radiographs across 3 academic referral centers and 13 surgeons. From these radiographs, 2,117 were used for training, 249 for validation, and 588 for external testing. Augmentation was applied to the training set (n = 2,117,000) to increase model robustness. Performance was evaluated by area under the receiver operating characteristic curve, sensitivity, specificity, and accuracy. Implant identification processing speed was calculated. RESULTS: The training and testing sets were drawn from statistically different populations of implants (P < .001). After 1,000 training epochs by the deep learning system, the system discriminated 8 implant models with a mean area under the receiver operating characteristic curve of 0.991, accuracy of 97.9%, sensitivity of 88.6%, and specificity of 98.9% in the external testing dataset of 588 anteroposterior radiographs. The software classified implants at a mean speed of 0.02 seconds per image. CONCLUSION: An AI-based software demonstrated excellent internal and external validation. Although continued surveillance is necessary with implant library expansion, this software represents responsible and meaningful clinical application of AI with immediate potential to globally scale and assist in preoperative planning prior to revision THA.

来源:J Arthroplasty,查看原文。版权归原作者所有。

引用

  1. Karnuta JM, Murphy MP, Luu BC, et al. Artificial Intelligence for Automated Implant Identification in Total Hip Arthroplasty: A Multicenter External Validation Study Exceeding Two Million Plain Radiographs. J Arthroplasty. 2022. PMID: 35271974. doi:10.1016/j.arth.2022.03.002

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。