Department of Radiology and Biomedical Imaging, University of California, San Francisco, San Francisco, California, USA.
MS in Analytics Program, University of San Francisco, San Francisco, California, USA.
Neuro Oncol. 2022 Apr 1;24(4):639-652. doi: 10.1093/neuonc/noab238.
Diagnostic classification of diffuse gliomas now requires an assessment of molecular features, often including IDH-mutation and 1p19q-codeletion status. Because genetic testing requires an invasive process, an alternative noninvasive approach is attractive, particularly if resection is not recommended. The goal of this study was to evaluate the effects of training strategy and incorporation of biologically relevant images on predicting genetic subtypes with deep learning.
Our dataset consisted of 384 patients with newly diagnosed gliomas who underwent preoperative MRI with standard anatomical and diffusion-weighted imaging, and 147 patients from an external cohort with anatomical imaging. Using tissue samples acquired during surgery, each glioma was classified into IDH-wildtype (IDHwt), IDH-mutant/1p19q-noncodeleted (IDHmut-intact), and IDH-mutant/1p19q-codeleted (IDHmut-codel) subgroups. After optimizing training parameters, top performing convolutional neural network (CNN) classifiers were trained, validated, and tested using combinations of anatomical and diffusion MRI with either a 3-class or tiered structure. Generalization to an external cohort was assessed using anatomical imaging models.
The best model used a 3-class CNN containing diffusion-weighted imaging as an input, achieving 85.7% (95% CI: [77.1, 100]) overall test accuracy and correctly classifying 95.2%, 88.9%, 60.0% of the IDHwt, IDHmut-intact, and IDHmut-codel tumors. In general, 3-class models outperformed tiered approaches by 13.5%-17.5%, and models that included diffusion-weighted imaging were 5%-8.8% more accurate than those that used only anatomical imaging.
Training a classifier to predict both IDH-mutation and 1p19q-codeletion status outperformed a tiered structure that first predicted IDH-mutation, then 1p19q-codeletion. Including apparent diffusion coefficient (ADC), a surrogate marker of cellularity, more accurately captured differences between subgroups.
弥漫性神经胶质瘤的诊断分类现在需要评估分子特征，通常包括 IDH 突变和 1p19q 缺失状态。由于基因检测需要一种侵入性的过程，因此一种替代的非侵入性方法很有吸引力，特别是如果不建议进行切除。本研究的目的是评估训练策略和纳入生物学相关图像对使用深度学习预测遗传亚型的影响。
我们的数据集包括 384 名新诊断为神经胶质瘤的患者，他们接受了术前 MRI 检查，包括标准解剖和弥散加权成像，以及来自外部队列的 147 名患者，这些患者仅接受了解剖成像检查。使用手术期间获得的组织样本，将每个神经胶质瘤分为 IDH 野生型（IDHwt）、IDH 突变/1p19q 非缺失（IDHmut-intact）和 IDH 突变/1p19q 缺失（IDHmut-codel）亚组。在优化训练参数后，使用解剖和弥散 MRI 的组合，使用 3 类或分层结构训练、验证和测试表现最佳的卷积神经网络（CNN）分类器。使用解剖成像模型评估对外部队列的泛化能力。
最佳模型使用包含弥散加权成像的 3 类 CNN，总测试准确率为 85.7%（95%CI：[77.1, 100]），正确分类 IDHwt、IDHmut-intact 和 IDHmut-codel 肿瘤的比例分别为 95.2%、88.9%和 60.0%。一般来说，3 类模型比分层方法的准确率高出 13.5%-17.5%，而包含弥散加权成像的模型比仅使用解剖成像的模型准确率高出 5%-8.8%。
训练一个分类器来预测 IDH 突变和 1p19q 缺失状态的性能优于首先预测 IDH 突变，然后预测 1p19q 缺失状态的分层结构。包括表观扩散系数（ADC），这是细胞密度的替代标志物，更准确地捕捉了亚组之间的差异。
