Brain tumors can appear anywhere in the brain and have vastly different sizes and morphology. Additionally, these tumors are often diffused and poorly contrasted. Consequently, the segmentation of brain tumor and intratumor subregions using magnetic resonance imaging (MRI) data with minimal human interventions remains a challenging task. In this paper, we present a novel fully automatic segmentation method from MRI data containing in vivo brain gliomas. This approach can not only localize the entire tumor region but can also accurately segment the intratumor structure. The proposed work was based on a cascaded deep learning convolutional neural network consisting of two subnetworks: (1) a tumor localization network (TLN) and (2) an intratumor classification network (ITCN). The TLN, a fully convolutional network (FCN) in conjunction with the transfer learning technology, was used to first process MRI data. The goal of the first subnetwork was to define the tumor region from an MRI slice. Then, the ITCN was used to label the defined tumor region into multiple subregions. Particularly, ITCN exploited a convolutional neural network (CNN) with deeper architecture and smaller kernel. The proposed approach was validated on multimodal brain tumor segmentation (BRATS 2015) datasets, which contain 220 high-grade glioma (HGG) and 54 low-grade glioma (LGG) cases. Dice similarity coefficient (DSC), positive predictive value (PPV), and sensitivity were used as evaluation metrics. Our experimental results indicated that our method could obtain the promising segmentation results and had a faster segmentation speed. More specifically, the proposed method obtained comparable and overall better DSC values (0.89, 0.77, and 0.80) on the combined (HGG + LGG) testing set, as compared to other methods reported in the literature. Additionally, the proposed approach was able to complete a segmentation task at a rate of 1.54 seconds per slice.

Intracranial aneurysms (IA) are lethal, with high morbidity and mortality rates. Reliable, rapid, and accurate segmentation of IAs and their adjacent vasculature from medical imaging data is important to improve the clinical management of patients with IAs. However, due to the blurred boundaries and complex structure of IAs and overlapping with brain tissue or other cerebral arteries, image segmentation of IAs remains challenging. This study aimed to develop an attention residual U-Net (ARU-Net) architecture with differential preprocessing and geometric postprocessing for automatic segmentation of IAs and their adjacent arteries in conjunction with 3D rotational angiography (3DRA) images. The proposed ARU-Net followed the classic U-Net framework with the following key enhancements. First, we preprocessed the 3DRA images based on boundary enhancement to capture more contour information and enhance the presence of small vessels. Second, we introduced the long skip connections of the attention gate at each layer of the fully convolutional decoder-encoder structure to emphasize the field of view (FOV) for IAs. Third, residual-based short skip connections were also embedded in each layer to implement in-depth supervision to help the network converge. Fourth, we devised a multiscale supervision strategy for independent prediction at different levels of the decoding path, integrating multiscale semantic information to facilitate the segmentation of small vessels. Fifth, the 3D conditional random field (3DCRF) and 3D connected component optimization (3DCCO) were exploited as postprocessing to optimize the segmentation results. Comprehensive experimental assessments validated the effectiveness of our ARU-Net. The proposed ARU-Net model achieved comparable or superior performance to the state-of-the-art methods through quantitative and qualitative evaluations. Notably, we found that ARU-Net improved the identification of arteries connecting to an IA, including small arteries that were hard to recognize by other methods. Consequently, IA geometries segmented by the proposed ARU-Net model yielded superior performance during subsequent computational hemodynamic studies (also known as "patient-specific" computational fluid dynamics [CFD] simulations). Furthermore, in an ablation study, the five key enhancements mentioned above were confirmed. The proposed ARU-Net model can automatically segment the IAs in 3DRA images with relatively high accuracy and potentially has significant value for clinical computational hemodynamic analysis.

Obtaining accurate displacement estimates along both axial (parallel to the acoustic beam) and lateral (perpendicular to the beam) directions is an important task for several clinical applications such as shear strain imaging, modulus reconstruction and temperature imaging, where a full description of the two or three-dimensional (2D/3D) deformation field is required. In this study we propose an improved speckle tracking algorithm where axial and lateral motion estimations are simultaneously performed to enhance motion tracking accuracy. More specifically, using conventional ultrasound echo data, this algorithm first finds an iso-contour in the vicinity of the peak correlation between two segments of the pre- and post-deformation ultrasound radiofrequency echo data. The algorithm then attempts to find the center of the iso-contour of the correlation function that corresponds to the unknown (sub-sample) motion vector between these two segments of echo data. This algorithm has been tested using computer-simulated data, studies with a tissue-mimicking phantom, and in vivo breast lesion data. Computer simulation results show that the method improves the accuracy of both lateral and axial tracking. Such improvements are more significant when the deformation is small or along the lateral direction. Results from the tissue-mimicking phantom study are consistent with findings observed in computer simulations. Using in vivo breast lesion data we found that, compared to the 2D quadratic subsample displacement estimation methods, higher quality axial strain and shear strain images (e.g. 18.6% improvement in contrast-to-noise ratio for shear strain images) can be obtained for large deformations (up to 5% frame-to-frame and 15% local strains) in a multi-compression technique. Our initial results demonstrated that this conceptually and computationally simple method could improve the image quality of ultrasound-based strain elastography with current clinical equipment.

Accurate tracking of tissue motion is critically important for several ultrasound elastography methods. In this study, we investigate the feasibility of using three published convolution neural network (CNN) models built for optical flow (hereafter referred to as CNN-based tracking) by the computer vision community for breast ultrasound strain elastography. Elastographic datasets produced by finite element and ultrasound simulations were used to retrain three published CNN models: FlowNet-CSS, PWC-Net, and LiteFlowNet. After retraining, the three improved CNN models were evaluated using computer-simulated and tissue-mimicking phantoms, and in vivo breast ultrasound data. CNN-based tracking results were compared with two published two-dimensional (2D) speckle tracking methods: coupled tracking and GLobal Ultrasound Elastography (GLUE) methods. Our preliminary data showed that, based on the Wilcoxon rank-sum tests, the improvements due to retraining were statistically significant (p < 0.05) for all three CNN models. We also found that the PWC-Net model was the best neural network model for data investigated, and its overall performance was on par with the coupled tracking method. CNR values estimated from in vivo axial and lateral strain elastograms showed that the GLUE algorithm outperformed both the retrained PWC-Net model and the coupled tracking method, though the GLUE algorithm exhibited some biases. The PWC-Net model was also able to achieve approximately 45 frames/second for 2D speckle tracking data investigated.