文献速递骨肌MR

机器学习在粘连性肩关节囊炎诊断中优于深度学习:连接PD-T2 MRI与多模态数据融合的临床-影像组学模型

Machine learning outperforms deep learning in adhesive capsulitis diagnosis: a clinical-radiomics model bridging PD-T2 MRI and multimodal data fusion.
PubMed 文献 · Eur J Radiol · 2025-10-11

本研究回顾性纳入两个医疗中心444例患者,基于PD-COR和T2-SAG MRI序列提取影像组学特征,比较机器学习与深度学习模型对粘连性肩关节囊炎的诊断效能。PD_T2_LightGBM模型测试集AUC为0.886,临床-影像组学联合模型测试集AUC为0.882,深度学习特征模型外部验证准确率下降。

要点

  • PD_T2_LightGBM影像组学模型测试集AUC 0.886,优于深度学习特征模型。
  • 临床-影像组学联合模型训练集AUC 0.981、验证集0.935、测试集0.882。
  • 深度学习特征模型敏感度高,但外部验证准确率下降。

核心结论

  • 整合临床与影像组学特征可提高诊断精度,LightGBM等传统机器学习模型稳定性和可解释性优于深度学习特征模型,适合临床应用;未来应优先扩大数据集和改进融合技术以优化ACS诊断。

研究设计与样本

  • 回顾性研究,共纳入两个医疗中心444例患者,分为主要队列387例和外部测试队列57例;从PD-COR和T2-SAG MRI序列用PyRadiomics提取影像组学特征,深度学习特征来自ResNet-200和ViT模型;机器学习模型采用SVM、XGBoost和LightGBM构建;临床-多序列影像组学模型整合影像组学与临床特征;以AUC、准确率、敏感度、特异度和Brier Score评估性能。

诊断标准与关键数值

  • PD_T2_LightGBM模型AUC:训练集0.975、验证集0.915、测试集0.886;临床-影像组学联合模型AUC:训练集0.981、验证集0.935、测试集0.882。

局限性

  • 原文未报告局限性。

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)。

英文原文(摘要)· 点击展开

BACKGROUND: Adhesive Capsulitis of the Shoulder (ACS) is a chronic inflammatory condition characterized by capsular fibrosis, thickening, and restricted mobility. Early diagnosis remains challenging due to the limited sensitivity of traditional imaging and symptom-based methods.

PURPOSE: This study developed a clinical-multi-sequence radiomics model by integrating clinical data with magnetic resonance imaging (MRI) radiomics to enhance ACS detection and compared machine learning (ML) and deep learning (DL) approaches.

METHODS: A total of 444 patients from two medical centers were retrospectively included and divided into a primary cohort (n = 387) and an external test cohort (n = 57). Radiomic features were extracted from proton density-weighted coronal (PD-COR) and T2-weighted sagittal (T2-SAG) MRI sequences using PyRadiomics, while deep learning features were obtained from ResNet-200 and Vision Transformer (ViT) models. ML models were developed using Support Vector Machine (SVM), Extreme Gradient Boosting (XGBoost), and Light Gradient Boosting machine (LightGBM). The clinical-multi-sequence radiomics model was constructed by integrating radiomic and clinical features, with performance assessed via the area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, specificity, and Brier Score.

RESULTS: The PD_T2_LightGBM model achieved optimal performance (AUC: 0.975 training, 0.915 validation, 0.886 test), surpassing DL features models. The Clinical-Radiomics Combined model showed robust generalization (AUC: 0.981 training, 0.935 validation, 0.882 test). DL features models exhibited high sensitivity but reduced external validation accuracy.

CONCLUSION: Integrating clinical and radiomic features significantly improved diagnostic precision. While DL features models provide valuable feature extraction capabilities, traditional ML models like LightGBM exhibit superior stability and interpretability, making them suitable for clinical applications. Future efforts should prioritize larger datasets and advanced fusion techniques to refine ACS diagnosis.

来源:Eur J Radiol,查看原文。版权归原作者所有。

引用

  1. Yang Y, Pan T, Zhang C, et al. Machine learning outperforms deep learning in adhesive capsulitis diagnosis: a clinical-radiomics model bridging PD-T2 MRI and multimodal data fusion. Eur J Radiol. 2025. PMID: 41092750. doi:10.1016/j.ejrad.2025.112470

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。