From the Department of Clinical Medicine, Center of Functionally Integrative Neuroscience and MINDLAB, Aarhus University, Denmark (A.N., M.B.H., A.T., K.M.)
Cercare Medical ApS, Aarhus, Denmark (A.N.).
Stroke. 2018 Jun;49(6):1394-1401. doi: 10.1161/STROKEAHA.117.019740. Epub 2018 May 2.
Treatment options for patients with acute ischemic stroke depend on the volume of salvageable tissue. This volume assessment is currently based on fixed thresholds and single imagine modalities, limiting accuracy. We wish to develop and validate a predictive model capable of automatically identifying and combining acute imaging features to accurately predict final lesion volume.
Using acute magnetic resonance imaging, we developed and trained a deep convolutional neural network (CNN) to predict final imaging outcome. A total of 222 patients were included, of which 187 were treated with rtPA (recombinant tissue-type plasminogen activator). The performance of CNN was compared with a shallow CNN based on the perfusion-weighted imaging biomarker Tmax (CNN), a shallow CNN based on a combination of 9 different biomarkers (CNN), a generalized linear model, and thresholding of the diffusion-weighted imaging biomarker apparent diffusion coefficient (ADC) at 600×10 mm/s (ADC). To assess whether CNN is capable of differentiating outcomes of ±intravenous rtPA, patients not receiving intravenous rtPA were included to train CNN to access a treatment effect. The networks' performances were evaluated using visual inspection, area under the receiver operating characteristic curve (AUC), and contrast.
CNN yields significantly better performance in predicting final outcome (AUC=0.88±0.12) than generalized linear model (AUC=0.78±0.12; =0.005), CNN (AUC=0.72±0.14; <0.003), and ADC (AUC=0.66±0.13; <0.0001) and a substantially better performance than CNN (AUC=0.85±0.11; =0.063). Measured by contrast, CNN improves the predictions significantly, showing superiority to all other methods (≤0.003). CNN also seems to be able to differentiate outcomes based on treatment strategy with the volume of final infarct being significantly different (=0.048).
The considerable prediction improvement accuracy over current state of the art increases the potential for automated decision support in providing recommendations for personalized treatment plans.
急性缺血性脑卒中患者的治疗选择取决于可挽救组织的体积。目前，这种体积评估基于固定阈值和单一成像方式，限制了准确性。我们希望开发和验证一种能够自动识别和组合急性成像特征的预测模型，以准确预测最终病变体积。
使用急性磁共振成像，我们开发并训练了一个深度卷积神经网络（CNN）来预测最终成像结果。共纳入 222 例患者，其中 187 例接受重组组织型纤溶酶原激活剂（rtPA）治疗。CNN 的性能与基于灌注加权成像生物标志物 Tmax 的浅层 CNN（CNN）、基于 9 种不同生物标志物组合的浅层 CNN（CNN）、广义线性模型以及将扩散加权成像生物标志物表观扩散系数（ADC）阈值设定为 600×10 mm/s（ADC）进行比较。为了评估 CNN 是否能够区分±静脉 rtPA 的结果，我们纳入未接受静脉 rtPA 治疗的患者来训练 CNN 以评估治疗效果。使用视觉检查、受试者工作特征曲线下面积（AUC）和对比度评估网络性能。
CNN 在预测最终结果方面的表现明显优于广义线性模型（AUC=0.78±0.12；=0.005）、CNN（AUC=0.72±0.14；<0.003）和 ADC（AUC=0.66±0.13；<0.0001），与 CNN（AUC=0.85±0.11；=0.063）相比，性能也有显著提高。通过对比度测量，CNN 显著改善了预测结果，优于所有其他方法（≤0.003）。CNN 似乎还能够根据治疗策略区分结果，最终梗死体积存在显著差异（=0.048）。
相对于当前的技术水平，预测准确性有了显著提高，为提供个性化治疗计划的建议提供了自动化决策支持的潜力。
