Fatemeh Zabihollahy, Nicola Schieda, Satheesh Krishna, Eranga Ukwatta
Department of Systems and Computer Engineering, Carleton University, Ottawa, ON, Canada.
Department of Radiology, University of Ottawa, Ottawa, ON, Canada.
Med Phys. 2020 Sep;47(9):4032-4044. doi: 10.1002/mp.14193. Epub 2020 Jul 28.
Detection and accurate localization of renal masses (RM) are important steps toward future potential classification of benign vs malignant RM. A fully automated algorithm for detection and localization of RM may eliminate the observer variability in the clinical workflow.
In this paper, we describe a fully automated methodology for accurate detection and segmentation of RM from contrast-enhanced computed tomography (CECT) images. We first determine the boundaries of the kidneys on the CECT images utilizing a convolutional neural network-based method to be used as a region of interest to search for RM. We then employ a homogenous U-Net-based ensemble learning model to identify and delineate RM. We used an institutional dataset comprised of CECT images in 315 patients to train and evaluate the proposed method. We compared results of our method to those of three-dimensional (3D) U-Net for RM localization and further evaluated our algorithm using the kidney tumor segmentation (KiTS19) challenge dataset.
The developed algorithm reported a Dice similarity coefficient (DSC) of 95.79% ± 5.16% and 96.25 ± 3.37 (mean ± standard deviation) for segmentation accuracy of kidney boundary from 125 and 60 test images from institutional and KiTS19 datasets, respectively. Using our method, RM were detected in 125 and 52 test cases, which corresponds to 100% and 86.67% sensitivity at patient level in institutional and KiTS19 test images. Our ensemble method for RM localization yielded a mean DSC of 88.65% ± 7.31% and 87.91% ± 6.82% on the institutional and KiTS19 test datasets, respectively. The mean DSC for RM delineation from CECT institutional test images using 3D U-Net was 85.95% ± 1.46%.
We describe a method for automated localization of RM using CECT images. Our results are important in terms of clinical perspective as fully automated detection of RM is a fundamental step for further diagnosis of cystic vs solid RM and eventually benign vs malignant solid RM, that has not been reported previously.
检测并准确定位肾肿块（RM）是对RM未来进行良性与恶性分类的重要步骤。一种用于检测和定位RM的全自动算法可消除临床工作流程中观察者的变异性。
在本文中，我们描述了一种从增强计算机断层扫描（CECT）图像中准确检测和分割RM的全自动方法。我们首先利用基于卷积神经网络的方法在CECT图像上确定肾脏的边界，将其用作搜索RM的感兴趣区域。然后，我们采用基于均匀U-Net的集成学习模型来识别和描绘RM。我们使用了一个由315例患者的CECT图像组成的机构数据集来训练和评估所提出的方法。我们将我们方法的结果与三维（3D）U-Net用于RM定位的结果进行了比较，并使用肾脏肿瘤分割（KiTS19）挑战数据集进一步评估了我们的算法。
所开发的算法在分别来自机构和KiTS19数据集的125张和60张测试图像上，对肾脏边界分割准确性的骰子相似系数（DSC）分别为95.79%±5.16%和96.25±3.37（平均值±标准差）。使用我们的方法，在125例和52例测试病例中检测到了RM，在机构和KiTS19测试图像中，患者水平的灵敏度分别为100%和86.67%。我们用于RM定位的集成方法在机构和KiTS19测试数据集上的平均DSC分别为88.65%±7.31%和87.91%±6.82%。使用3D U-Net从CECT机构测试图像中描绘RM的平均DSC为85.95%±1.46%。
我们描述了一种使用CECT图像自动定位RM的方法。从临床角度来看，我们的结果很重要，因为RM的全自动检测是进一步诊断囊性与实性RM以及最终良性与恶性实性RM的基本步骤，此前尚未有相关报道。
