Automatic prostate segmentation using deep learning on clinically diverse 3D transrectal ultrasound images
Automatic prostate segmentation using deep learning on clinically diverse 3D transrectal ultrasound images
复制标题
DOI：
10.1002/mp.14134
复制
发表时间：
2020-04-08
期刊：
影响因子：
3.8
通讯作者：
Fenster, Aaron
中科院分区：
文献类型：
--
作者：
Orlando, Nathan;Gillies, Derek J.;Fenster, Aaron
目的用于诊断和治疗前列腺癌的基于针的程序，例如活检和近距离放射治疗，已经结合了三维（3D）经直肠超声（TRUS）成像以改善针引导。有效地使用这些图像通常需要医生手动分割前列腺以定义用于准确配准、靶向和其他引导技术的边缘。然而，手动前列腺分割是一个耗时且困难的术中过程，通常发生在患者处于镇静（活检）或麻醉（近距离放射治疗）状态时。使用3D TRUS前列腺分割方法最大限度地缩短手术时间可以为医生提供快速准确的前列腺分割，并允许有效的工作流程，提高患者吞吐量，使患者能够更快地获得护理。本研究的目的是开发一种基于监督深度学习的方法，用于分割来自不同设施的3D TRUS图像中的前列腺，这些图像使用多种采集方法和商业超声机器模型生成，以创建用于针式前列腺癌手术的通用算法。方法我们提出的三维分割方法涉及预测的二维（2D）切片采样径向周围的近似中心轴的前列腺，然后重建成一个3D表面。使用临床活检和近距离放射治疗过程中采集的84个端射和122个侧射3D TRUS图像对2D U-Net进行了修改、训练和验证。对标准U-Net的扩展部分的修改包括增加50%的辍学率，以及使用转置卷积代替标准上采样，然后进行卷积，以分别减少过拟合和提高性能。手动轮廓提供了训练、确认和测试数据集所需的注释，测试数据集由20个端射和20个侧射不可见3D TRUS图像组成。由于使用2D图像进行预测可能会丢失空间和结构信息，因此在对不同损失函数进行研究后，对3D重建和优化的3D网络（包括3D V-Net，Dense V-Net和High-resolution 3D-Net）进行了比较。计算绝对和符号误差指标的扩展选择，包括像素图比较[骰子相似系数（DSC）、召回率和精度]、体积百分比差异（VPD）、平均表面距离（MSD）和Hausdorff距离（HD），以评估3D分割准确度。结果总体而言，我们提出的重建的改良U-Net的中位[第一四分位数，第三四分位数]绝对DSC，召回率，精确度，VPD，MSD和HD分别为94.1 [92.6，94.9]%，96.0 [93.1，98.5]%，93.2 [88.8，95.4]%，5.78 [2.49，11.50]%，0.89 [0.73，1.09] mm和2.89 [2.37，4.35] mm。与性能最佳的优化3D网络（即，3D V-Net加上交叉熵损失函数），我们提出的方法在几乎所有指标上都有显着改进。的计算时间
Purpose Needle-based procedures for diagnosing and treating prostate cancer, such as biopsy and brachytherapy, have incorporated three-dimensional (3D) transrectal ultrasound (TRUS) imaging to improve needle guidance. Using these images effectively typically requires the physician to manually segment the prostate to define the margins used for accurate registration, targeting, and other guidance techniques. However, manual prostate segmentation is a time-consuming and difficult intraoperative process, often occurring while the patient is under sedation (biopsy) or anesthetic (brachytherapy). Minimizing procedure time with a 3D TRUS prostate segmentation method could provide physicians with a quick and accurate prostate segmentation, and allow for an efficient workflow with improved patient throughput to enable faster patient access to care. The purpose of this study was to develop a supervised deep learning-based method to segment the prostate in 3D TRUS images from different facilities, generated using multiple acquisition methods and commercial ultrasound machine models to create a generalizable algorithm for needle-based prostate cancer procedures. Methods Our proposed method for 3D segmentation involved prediction on two-dimensional (2D) slices sampled radially around the approximate central axis of the prostate, followed by reconstruction into a 3D surface. A 2D U-Net was modified, trained, and validated using images from 84 end-fire and 122 side-fire 3D TRUS images acquired during clinical biopsies and brachytherapy procedures. Modifications to the expansion section of the standard U-Net included the addition of 50% dropouts and the use of transpose convolutions instead of standard upsampling followed by convolution to reduce overfitting and improve performance, respectively. Manual contours provided the annotations needed for the training, validation, and testing datasets, with the testing dataset consisting of 20 end-fire and 20 side-fire unseen 3D TRUS images. Since predicting with 2D images has the potential to lose spatial and structural information, comparisons to 3D reconstruction and optimized 3D networks including 3D V-Net, Dense V-Net, and High-resolution 3D-Net were performed following an investigation into different loss functions. An extended selection of absolute and signed error metrics were computed, including pixel map comparisons [dice similarity coefficient (DSC), recall, and precision], volume percent differences (VPD), mean surface distance (MSD), and Hausdorff distance (HD), to assess 3D segmentation accuracy. Results Overall, our proposed reconstructed modified U-Net performed with a median [first quartile, third quartile] absolute DSC, recall, precision, VPD, MSD, and HD of 94.1 [92.6, 94.9]%, 96.0 [93.1, 98.5]%, 93.2 [88.8, 95.4]%, 5.78 [2.49, 11.50]%, 0.89 [0.73, 1.09] mm, and 2.89 [2.37, 4.35] mm, respectively. When compared to the best-performing optimized 3D network (i.e., 3D V-Net with a Dice plus cross-entropy loss function), our proposed method performed with a significant improvement across nearly all metrics. A computation time