Tushar Fakrul Islam, D'Anniballe Vincent M, Hou Rui, Mazurowski Maciej A, Fu Wanyi, Samei Ehsan, Rubin Geoffrey D, Lo Joseph Y
Center for Virtual Imaging Trials, Carl E. Ravin Advanced Imaging Laboratories, Department of Radiology and Department of Electrical and Computer Engineering, Duke University, 2424 Erwin Rd, Studio 302, Durham, NC 27705 (F.I.T., R.H., M.A.M., W.F., E.S., J.Y.L.); Department of Radiology, Duke University, Durham, NC (V.M.D.); and Department of Medical Imaging, University of Arizona, Tucson, Ariz (G.D.R.).
Radiol Artif Intell. 2021 Dec 1;4(1):e210026. doi: 10.1148/ryai.210026. eCollection 2022 Jan.
To design multidisease classifiers for body CT scans for three different organ systems using automatically extracted labels from radiology text reports.
This retrospective study included a total of 12 092 patients (mean age, 57 years ± 18 [standard deviation]; 6172 women) for model development and testing. Rule-based algorithms were used to extract 19 225 disease labels from 13 667 body CT scans performed between 2012 and 2017. Using a three-dimensional DenseVNet, three organ systems were segmented: lungs and pleura, liver and gallbladder, and kidneys and ureters. For each organ system, a three-dimensional convolutional neural network classified each as no apparent disease or for the presence of four common diseases, for a total of 15 different labels across all three models. Testing was performed on a subset of 2158 CT volumes relative to 2875 manually derived reference labels from 2133 patients (mean age, 58 years ± 18; 1079 women). Performance was reported as area under the receiver operating characteristic curve (AUC), with 95% CIs calculated using the DeLong method.
Manual validation of the extracted labels confirmed 91%-99% accuracy across the 15 different labels. AUCs for lungs and pleura labels were as follows: atelectasis, 0.77 (95% CI: 0.74, 0.81); nodule, 0.65 (95% CI: 0.61, 0.69); emphysema, 0.89 (95% CI: 0.86, 0.92); effusion, 0.97 (95% CI: 0.96, 0.98); and no apparent disease, 0.89 (95% CI: 0.87, 0.91). AUCs for liver and gallbladder were as follows: hepatobiliary calcification, 0.62 (95% CI: 0.56, 0.67); lesion, 0.73 (95% CI: 0.69, 0.77); dilation, 0.87 (95% CI: 0.84, 0.90); fatty, 0.89 (95% CI: 0.86, 0.92); and no apparent disease, 0.82 (95% CI: 0.78, 0.85). AUCs for kidneys and ureters were as follows: stone, 0.83 (95% CI: 0.79, 0.87); atrophy, 0.92 (95% CI: 0.89, 0.94); lesion, 0.68 (95% CI: 0.64, 0.72); cyst, 0.70 (95% CI: 0.66, 0.73); and no apparent disease, 0.79 (95% CI: 0.75, 0.83).
Weakly supervised deep learning models were able to classify diverse diseases in multiple organ systems from CT scans. CT, Diagnosis/Classification/Application Domain, Semisupervised Learning, Whole-Body Imaging© RSNA, 2022.
利用从放射学文本报告中自动提取的标签，为三种不同器官系统的身体CT扫描设计多疾病分类器。
这项回顾性研究共纳入12092例患者（平均年龄57岁±18[标准差]；6172例女性）用于模型开发和测试。使用基于规则的算法从2012年至2017年期间进行的13667例身体CT扫描中提取19225个疾病标签。使用三维密集连接网络（DenseVNet）对三个器官系统进行分割：肺和胸膜、肝脏和胆囊、肾脏和输尿管。对于每个器官系统，使用三维卷积神经网络将其分类为无明显疾病或存在四种常见疾病，三个模型共有15个不同标签。相对于来自2133例患者（平均年龄58岁±18；1079例女性）的2875个手动得出的参考标签，对2158个CT容积的子集进行测试。性能以受试者操作特征曲线（AUC）下的面积报告，使用德龙方法计算95%置信区间（CI）。
对提取标签的人工验证证实，15个不同标签的准确率为91%-99%。肺和胸膜标签的AUC如下：肺不张，0.77（95%CI：0.74，0.81）；结节，0.65（95%CI：0.61，0.69）；肺气肿，0.89（95%CI：0.86，0.92）；胸腔积液，0.97（95%CI：0.96，0.98）；无明显疾病，0.89（95%CI：0.87，0.91）。肝脏和胆囊的AUC如下：肝胆钙化，0.62（95%CI：0.56，0.67）；病变，0.73（95%CI：0.69，0.77）；扩张，0.87（95%CI：0.84，0.90）；脂肪肝，0.89（95%CI：0.86，0.92）；无明显疾病，0.82（95%CI：0.78，0.85）。肾脏和输尿管的AUC如下：结石，0.83（95%CI：0.79，0.87）；萎缩，0.92（95%CI：0.89，0.94）；病变，0.68（95%CI：0.64，0.72）；囊肿，0.70（95%CI：0.66，0.73）；无明显疾病，0.79（95%CI：0.75，0.83）。
弱监督深度学习模型能够根据CT扫描对多个器官系统中的多种疾病进行分类。CT，诊断/分类/应用领域，半监督学习，全身成像©RSNA，2022年。
