基于影像组学特征与机器学习的腰椎MRI中央椎管狭窄严重程度分类
本研究分析1500幅矢状位T2/STIR腰椎MRI,由两名放射科医生手动分割并提取影像组学特征,训练四种机器学习模型对中央椎管狭窄严重程度进行三分类。AdaBoost表现最佳,准确率80%、多分类AUC 93.60%,支持该方法的可行性,但依赖手动分割,属探索性概念验证。
要点
- 影像组学联合机器学习可对腰椎中央椎管狭窄进行三分类,AdaBoost多分类AUC达93.60%。
- ANN模型特异度最高达90%,AdaBoost准确率80%、精确率81%。
- 当前流程依赖放射科医生手动分割,尚不能直接用于临床自动化报告。
核心结论
- 基于腰椎MRI影像组学特征的机器学习模型可对中央椎管狭窄严重程度进行三分类,AdaBoost整体分类效能最高。
- 研究结果支持该方法的可行性,但依赖手动分割,应视为探索性概念验证。
- 需进一步评估泛化性并整合适用于临床的自动化方法。
研究设计与样本
- 研究为影像组学联合机器学习的分类研究,共分析1500幅矢状位T2/STIR加权腰椎MRI图像。
- 狭窄严重程度分为正常/轻度、中度、重度三类。
- 由两名经验丰富的放射科医生手动分割矢状位切片,并从每个感兴趣区提取影像组学特征。
- 使用提取特征训练并评估随机森林、AdaBoost、XGBoost和人工神经网络四种机器学习模型。
诊断标准与关键数值
- AdaBoost分类准确率80%(95% CI: 75.47-84.53%),精确率81%(95% CI: 76.56-85.44%)。
- 多分类AUC:AdaBoost 93.60%(95% CI: 90.83-96.37%),XGBoost 92.80%(95% CI: 89.88-95.72%),随机森林92.60%(95% CI: 89.64-95.56%),ANN 90.70%(95% CI: 87.41-93.99%)。
- ANN模型特异度最高,为90%(95% CI: 86.60-93.40%)。
临床应用与阅片要点
- 阅片时可关注矢状位T2/STIR上中央椎管狭窄的三级分类(正常/轻度、中度、重度),与模型分类任务对应。
- 模型输出可作为辅助参考,但当前依赖手动分割,不应替代放射科医生的独立判读。
- 注意该流程尚未实现自动化,临床落地前需验证泛化性。
局限性
- 当前工作流程依赖手动分割,属探索性概念验证,尚未实现自动化。
- 研究结论的泛化性有待进一步评估。
英文原文(摘要)· 点击展开
Neuroradiology . 2026 Oct 30. doi: 10.1007/s00234-026-04238-8. Online ahead of print. ABSTRACT PURPOSE: Accurate assessment of lumbar degenerative disease is important for treatment planning and surgical decision-making. In this study, we investigated whether radiomics features extracted from lumbar MRI, combined with machine learning models, could be used to classify the severity of lumbar spinal stenosis. MATERIALS AND METHODS: We analyzed 1,500 sagittal T2/STIR-weighted lumbar MRI images representing three categories of central canal stenosis severity: normal/mild, moderate, and severe. Two experienced radiologists manually segmented the sagittal slices, and radiomics features were extracted from each region of interest. These features were used to train and evaluate four machine learning models: random forest (RF), adaptive boosting (AdaBoost), extreme gradient boosting (XGBoost), and an artificial neural network (ANN), for stenosis severity classification. RESULTS: All models demonstrated high classification performance, with AdaBoost achieving the highest overall effectiveness. Specifically, AdaBoost reached an accuracy of 80% (95% CI: 75.47-84.53%) and a precision of 81% (95% CI: 76.56- 85.44%). The multi-class Area Under the Curve (AUC) values were 93.60% (95% CI: 90.83- 96.37%) for AdaBoost, 92.80% (95% CI: 89.88-95.72%) for XGBoost, 92.60% (95% CI: 89.64- 95.56%) for RF, and 90.70% (95% CI: 87.41-93.99%) for ANN. While AdaBoost led in discrimination, the ANN model achieved the highest specificity at 90% (95% CI: 86.60-93.40%). CONCLUSION: The results of this study support the feasibility of using radiomics-based machine learning for severity classification of central canal stenosis on lumbar MRI. Although the findings are encouraging, the current workflow relies on manual segmentation and should be regarded as an exploratory proof of concept. Further work is needed to evaluate generalizability and to integrate automated methods suitable for clinical use. PMID: 42961456 | DOI: 10.1007/s00234-026-04238-8
引用
- Neuroradiology(PubMed 期刊源):MRI-based severity classification of central canal stenosis using radiomics features and machine learning
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。