文献速递神经儿科影像MR

自动机器学习在常规MR成像上鉴别儿童后颅窝肿瘤

Automatic Machine Learning to Differentiate Pediatric Posterior Fossa Tumors on Routine MR Imaging.
PubMed 文献 · AJNR Am J Neuroradiol · 2020-01-01

本回顾性研究纳入288例儿童后颅窝肿瘤(髓母细胞瘤111例、室管膜瘤70例、毛细胞型星形细胞瘤107例)的术前MR图像,从T2加权、增强T1加权和ADC图提取影像组学特征,比较专家手工优化模型与自动机器学习(Tree-Based Pipeline Optimization Tool)的性能。自动机器学习模型在三分类中测试集微平均AUC为0.91、准确率0.83,准确率显著高于专家定性MR阅片(0.83 vs 0.54,P < .001)。

要点

  • 自动机器学习模型三分类测试集微平均AUC 0.91、准确率0.83,专家定性MR阅片平均准确率0.54,差异有统计学意义(P < .001)。
  • 二分类中,毛细胞型星形细胞瘤vs非毛细胞型星形细胞瘤AUC 0.94、准确率0.88,髓母细胞瘤vs非髓母细胞瘤AUC 0.94、准确率0.85,室管膜瘤vs非室管膜瘤AUC 0.84、准确率0.80。
  • 基于χ2特征选择与广义线性分类器的最优化模型三分类测试集微平均AUC 0.92、准确率0.74,AUC略高于自动机器学习模型但准确率较低。

核心结论

  • 基于常规MR成像的自动机器学习能高准确度分类儿童后颅窝肿瘤,性能优于专家手工流程优化和定性MR阅片。

研究设计与样本

  • 回顾性研究,纳入288例儿童后颅窝肿瘤患者的术前MR图像,包括髓母细胞瘤111例、室管膜瘤70例、毛细胞型星形细胞瘤107例;从T2加权图像、增强T1加权图像和ADC图提取影像组学特征;比较机器学习专家标准手工优化模型与Tree-Based Pipeline Optimization Tool自动机器学习模型的性能。

关键影像表现/征象

  • 常规MR成像包括T2加权图像、增强T1加权图像和ADC图,从中提取影像组学特征用于分类。

诊断标准与关键数值

  • 三分类中,Tree-Based Pipeline Optimization Tool自动机器学习模型测试集微平均AUC 0.91、准确率0.83;基于χ2特征选择与广义线性模型分类器的最优化模型测试集微平均AUC 0.92、准确率0.74。
  • Tree-Based Pipeline Optimization Tool模型准确率显著高于专家定性MR阅片平均准确率(0.83 vs 0.54,P < .001)。
  • 二分类中,Tree-Based Pipeline Optimization Tool模型髓母细胞瘤vs非髓母细胞瘤AUC 0.94、准确率0.85;室管膜瘤vs非室管膜瘤AUC 0.84、准确率0.80;毛细胞型星形细胞瘤vs非毛细胞型星形细胞瘤AUC 0.94、准确率0.88。

临床应用与阅片要点

  • 作者结论认为,基于常规MR成像的自动机器学习分类儿童后颅窝肿瘤具有高准确度。

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 AND PURPOSE: Differentiating the types of pediatric posterior fossa tumors on routine imaging may help in preoperative evaluation and guide surgical resection planning. However, qualitative radiologic MR imaging review has limited performance. This study aimed to compare different machine learning approaches to classify pediatric posterior fossa tumors on routine MR imaging.

MATERIALS AND METHODS: This retrospective study included preoperative MR imaging of 288 patients with pediatric posterior fossa tumors, including medulloblastoma (n = 111), ependymoma (n = 70), and pilocytic astrocytoma (n = 107). Radiomics features were extracted from T2-weighted images, contrast-enhanced T1-weighted images, and ADC maps. Models generated by standard manual optimization by a machine learning expert were compared with automatic machine learning via the Tree-Based Pipeline Optimization Tool for performance evaluation.

RESULTS: For 3-way classification, the radiomics model by automatic machine learning with the Tree-Based Pipeline Optimization Tool achieved a test micro-averaged area under the curve of 0.91 with an accuracy of 0.83, while the most optimized model based on the feature-selection method χ2 score and the Generalized Linear Model classifier achieved a test micro-averaged area under the curve of 0.92 with an accuracy of 0.74. Tree-Based Pipeline Optimization Tool models achieved significantly higher accuracy than average qualitative expert MR imaging review (0.83 versus 0.54, P < .001). For binary classification, Tree-Based Pipeline Optimization Tool models achieved an area under the curve of 0.94 with an accuracy of 0.85 for medulloblastoma versus nonmedulloblastoma, an area under the curve of 0.84 with an accuracy of 0.80 for ependymoma versus nonependymoma, and an area under the curve of 0.94 with an accuracy of 0.88 for pilocytic astrocytoma versus non-pilocytic astrocytoma.

CONCLUSIONS: Automatic machine learning based on routine MR imaging classified pediatric posterior fossa tumors with high accuracy compared with manual expert pipeline optimization and qualitative expert MR imaging review.

来源:AJNR Am J Neuroradiol,查看原文。版权归原作者所有。

引用

  1. Zhou H, Hu R, Tang O, et al. Automatic Machine Learning to Differentiate Pediatric Posterior Fossa Tumors on Routine MR Imaging. AJNR Am J Neuroradiol. 2020. PMID: 32661052. doi:10.3174/ajnr.A6621

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。