胸部CT中与COVID-19相关的CT模式的自动定量分析
Automated Quantification of CT Patterns Associated with COVID-19 from Chest CT.
作者信息
Chaganti Shikha, Grenier Philippe, Balachandran Abishek, Chabin Guillaume, Cohen Stuart, Flohr Thomas, Georgescu Bogdan, Grbic Sasa, Liu Siqi, Mellot François, Murray Nicolas, Nicolaou Savvas, Parker William, Re Thomas, Sanelli Pina, Sauter Alexander W, Xu Zhoubing, Yoo Youngjin, Ziebandt Valentin, Comaniciu Dorin
机构信息
Hôpital Foch, Suresnes, France (P.G., F.M.), Donald and Barbara Zucker School of Medicine, Feinstein Institutes for Medical Research, Northwell Health, Manhasset, NY, USA (S.C., P.S.), Siemens Healthinners, Bangalore, India (A.B.), Siemens Healthineers, Forchheim, Germany (T.F., V.Z.), Siemens Healthineers, Princeton, NJ, USA (S.C., B.G., S.G., S.L., T.R., Z.X., Y.Y., D.C.), Siemens Healthineers, Paris, France (G.C.), University Hospital Basel, Clinic of Radiology & Nuclear medicine, Basel, Switzerland (A.W.S.), Vancouver General Hospital, Vancouver, Canada (N.M., S.N., W.P.).
出版信息
Radiol Artif Intell. 2020 Jul 29;2(4):e200048. doi: 10.1148/ryai.2020200048. eCollection 2020 Jul.
PURPOSE
To present a method that automatically segments and quantifies abnormal CT patterns commonly present in coronavirus disease 2019 (COVID-19), namely ground glass opacities and consolidations.
MATERIALS AND METHODS
In this retrospective study, the proposed method takes as input a non-contrasted chest CT and segments the lesions, lungs, and lobes in three dimensions, based on a dataset of 9749 chest CT volumes. The method outputs two combined measures of the severity of lung and lobe involvement, quantifying both the extent of COVID-19 abnormalities and presence of high opacities, based on deep learning and deep reinforcement learning. The first measure of (PO, PHO) is global, while the second of (LSS, LHOS) is lobe-wise. Evaluation of the algorithm is reported on CTs of 200 participants (100 COVID-19 confirmed patients and 100 healthy controls) from institutions from Canada, Europe and the United States collected between 2002-Present (April 2020). Ground truth is established by manual annotations of lesions, lungs, and lobes. Correlation and regression analyses were performed to compare the prediction to the ground truth.
RESULTS
Pearson correlation coefficient between method prediction and ground truth for COVID-19 cases was calculated as 0.92 for PO ( < .001), 0.97 for PHO ( < .001), 0.91 for LSS ( < .001), 0.90 for LHOS ( < .001). 98 of 100 healthy controls had a predicted PO of less than 1%, 2 had between 1-2%. Automated processing time to compute the severity scores was 10 seconds per case compared to 30 minutes required for manual annotations.
CONCLUSION
A new method segments regions of CT abnormalities associated with COVID-19 and computes (PO, PHO), as well as (LSS, LHOS) severity scores.
目的
提出一种自动分割和量化2019冠状病毒病（COVID-19）中常见的异常CT模式的方法，即磨玻璃影和实变。
材料与方法
在这项回顾性研究中，所提出的方法以非增强胸部CT作为输入，基于9749例胸部CT容积数据集在三维空间中分割病变、肺和肺叶。该方法基于深度学习和深度强化学习输出两个衡量肺和肺叶受累严重程度的综合指标，量化COVID-19异常的范围和高密度影的存在情况。第一个指标（PO，PHO）是全局的，而第二个指标（LSS，LHOS）是按肺叶的。对200名参与者（100例COVID-19确诊患者和100名健康对照）的CT进行算法评估，这些CT来自加拿大、欧洲和美国的机构，收集时间为2002年至今（2020年4月）。通过对病变、肺和肺叶进行手动标注来确定金标准。进行相关性和回归分析以将预测结果与金标准进行比较。
结果
COVID-19病例的方法预测值与金标准之间的Pearson相关系数计算如下：PO为0.92（<0.001），PHO为0.97（<0.001），LSS为0.91（<0.001），LHOS为0.90（<0.001）。100名健康对照中有98名预测的PO小于1%，2名在1%-2%之间。计算严重程度评分的自动处理时间为每例10秒，而手动标注需要30分钟。
结论
一种新方法可分割与COVID-19相关的CT异常区域并计算（PO，PHO）以及（LSS，LHOS）严重程度评分。
Zotero 插件