Search

Brain functional connectivity network (FCN) based on resting-state functional magnetic resonance imaging (rs-fMRI) has been widely used to identify neuropsychiatric disorders such as autism spectrum disorder (ASD). Most existing FCN-based methods only estimate the correlation between brain regions of in terest (ROIs), without exploring more informative higher-level inter actions among multiple ROIs which could be beneficial to disease diagnosis. To fully explore the discriminative information provided by different brain networks, a cluster-based multi-view high-order FCN (Ho-FCN) framework is proposed in this paper. Specifically, we first group the functional connectivity (FC) time series into different clusters and compute the multi-order central moment series for the FC time series in each cluster. Then we utilize the correlation of central moment series between different clusters to reveal the high-order FC relationships among multiple ROIs. In addition, to address the phase mismatch issue in conventional FCNs, we also adopt the central moments of the correlation time series as the temporal-invariance features to capture the dynamic characteristics of low-order dynamic FCN (Lo-D-FCN). Experimental results on the ABIDE dataset validate that: 1) the proposed multi-view Ho-FCNs is able to explore rich discriminative information for ASD diagnosis; 2) the phase mismatch issue can be well circumvented by using central moments; and 3) the combination of different types of FCNs can significantly improve the diagnostic accuracy of ASD (86.2%).

Resting-state functional magnetic resonance imaging (rs-fMRI) has become a popular non-invasive way of diagnosing neurological disorders or their early stages by probing functional connectivity between different brain regions of interest (ROIs) across subjects. In the past decades, researchers have proposed many methods to estimate brain functional networks (BFNs) based on blood-oxygen-level-dependent (BOLD) signals captured by rs-fMRI. However, most of the existing methods estimate BFNs under the assumption that signals are independently sampled, which ignores the temporal dependency and sequential order of different time points (or volumes). To address this problem, in this paper, we first propose a novel BFN estimation model by introducing a latent variable to control the sequence of volumes for encoding the temporal dependency and sequential information of signals into the estimated BFNs. Then, we develop an efficient learning algorithm to solve the proposed model by the alternating optimization scheme. To verify the effectiveness of the proposed method, the estimated BFNs are used to identify subjects with mild cognitive impairment (MCIs) from normal controls (NCs). Experimental results show that our method outperforms the baseline methods in the sense of classification performance.

Medical image registration is a vital component of many medical procedures, such as image-guided radiotherapy (IGRT), as it allows for more accurate dose-delivery and better management of side effects. Recently, the successful implementation of deep learning (DL) in various fields has prompted many research groups to apply DL to three-dimensional (3D) medical image registration. Several of these efforts have led to promising results. This review summarized the progress made in DL-based 3D image registration over the past 5 years and identify existing challenges and potential avenues for further research. The collected studies were statistically analyzed based on the region of interest (ROI), image modality, supervision method, and registration evaluation metrics. The studies were classified into three categories: deep iterative registration, supervised registration, and unsupervised registration. The studies are thoroughly reviewed and their unique contributions are highlighted. A summary is presented following a review of each category of study, discussing its advantages, challenges, and trends. Finally, the common challenges for all categories are discussed, and potential future research topics are identified.

Purpose Dynamic contrast-enhanced MRI (DCE-MRI) registration is a challenging task because of the effect of remarkable intensity changes caused by contrast agent injections. Unrealistic deformation usually occurs by using traditional intensity-based algorithms. To alleviate the effect of contrast agent on registration, we proposed a DCE-MRI registration strategy and investigated the registration performance on the clinical DCE-MRI time series of liver. Method We reconstructed the time-intensity curves of the contrast agent through sparse representation with a predefined dictionary whose columns were the time-intensity curves with high correlations with respect to a preselected contrast agent curve. After reshaping 1D-reconstructed contrast agent time-intensity curves into a 4D contrast agent time series, we aligned the original time series to the reconstructed contrast agent time series through traditional free-form deformation (FFD) registration scheme combined with a residual complexity (RC) similarity and an iterative registration strategy. This study included the DCE-MRI time series of 20 patients with liver cancer. Results Qualitatively, the time-cut images and subtraction images of different registration methods did not obviously differ. Quantitatively, the mean (standard deviation) of temporal intensity smoothness of all the patients achieved 54.910 (18.819), 54.609 (18.859), and 53.391 (19.031) in FFD RC, RDDR, Zhou et al.'s method and the proposed method, respectively. The mean (standard deviation) of changes in the lesion volume were 0.985 (0.041), 0.983 (0.041), 0.981 (0.046), and 0.989 (0.036) in FFD RC, RDDR, Zhou et al.'s method and the proposed method. Conclusion Our proposed method would be an effective registration strategy for DCE-MRI time series, and its performance was comparable with that of three advanced registration methods.

Magnetic resonance imaging (MRI) is a major imaging technique for studying neuroanatomy. By applying different pulse sequences and parameters, different modalities can be generated regarding the same anatomical structure, which can provide complementary information for diagnosis. However, limited by the scanning time and related cost, multiple different modalities are often not available for the same patient in clinic. Recently, many methods have been proposed for cross-modality MRI synthesis, but most of them only consider pixel-level differences between the synthetic and ground-truth images, ignoring the edge information, which is critical to provide clinical information. In this paper, we propose a novel edge-preserving MRI image synthesis method with iterative multi-scale feature fusion based generative adversarial network (EP_IMF-GAN). Particularly, the generator consists of a shared encoder and two specific decoders to carry out different tasks: 1) a primary task aiming to generate the target modality and 2) an auxiliary task aiming to generate the corresponding edge image of target modality. We assume that infusing the auxiliary edge image generation task can help preserve edge information and learn better latent representation features through the shared encoder. Meanwhile, an iterative multi-scale fusion module is embedded in the primary decoder to fuse supplementary information of feature maps at different scales, thereby further improving quality of the synthesized target modality. Experiments on the BRATS dataset indicate that our proposed method is superior to the state-of-the-art image synthesis approaches in both qualitative and quantitative measures. Ablation study further validates the effectiveness of the proposed components.

Cortical surface registration is an essential step and prerequisite for surface-based neuroimaging analysis. It aligns cortical surfaces across individuals and time points to establish cross-sectional and longitudinal cortical correspondences to facilitate neuroimaging studies. Though achieving good performance, available methods are either time consuming or not flexible to extend to multiple or high dimensional features. Considering the explosive availability of large-scale and multimodal brain MRI data, fast surface registration methods that can flexibly handle multimodal features are desired. In this study, we develop a Superfast Spherical Surface Registration (S3Reg) framework for the cerebral cortex. Leveraging an end-to-end unsupervised learning strategy, S3Reg offers great flexibility in the choice of input feature sets and output similarity measures for registration, and meanwhile reduces the registration time significantly. Specifically, we exploit the powerful learning capability of spherical Convolutional Neural Network (CNN) to directly learn the deformation fields in spherical space and implement diffeomorphic design with “scaling and squaring” layers to guarantee topology-preserving deformations. To handle the polar-distortion issue, we construct a novel spherical CNN model using three orthogonal Spherical U-Nets. Experiments are performed on two different datasets to align both adult and infant multimodal cortical features. Results demonstrate that our S3Reg shows superior or comparable performance with state-of-the-art methods, while improving the registration time from 1 min to 10 sec.

Most prior studies focused on developing models for the severity or mortality prediction of COVID-19 patients. However, effective models for recovery-time prediction are still lacking. Here, we present a deep learning solution named iCOVID that can successfully predict the recovery-time of COVID-19 patients based on predefined treatment schemes and heterogeneous multimodal patient information collected within 48 hours after admission. Meanwhile, an interpretable mechanism termed FSR is integrated into iCOVID to reveal the features greatly affecting the prediction of each patient. Data from a total of 3008 patients were collected from three hospitals in Wuhan, China, for large-scale verification. The experiments demonstrate that iCOVID can achieve a time-dependent concordance index of 74.9% (95% CI: 73.6-76.3%) and an average day error of 4.4 days (95% CI: 4.2-4.6 days). Our study reveals that treatment schemes, age, symptoms, comorbidities, and biomarkers are highly related to recovery-time predictions.

Accurate segmentation of the prostate is a key step in external beam radiation therapy treatments. In this paper, we tackle the challenging task of prostate segmentation in CT images by a two-stage network with 1) the first stage to fast localize, and 2) the second stage to accurately segment the prostate. To precisely segment the prostate in the second stage, we formulate prostate segmentation into a multi-task learning framework, which includes a main task to segment the prostate, and an auxiliary task to delineate the prostate boundary. Here, the second task is applied to provide additional guidance of unclear prostate boundary in CT images. Besides, the conventional multi-task deep networks typically share most of the parameters (i.e., feature representations) across all tasks, which may limit their data fitting ability, as the specificity of different tasks are inevitably ignored. By contrast, we solve them by a hierarchically-fused U-Net structure, namely HF-UNet. The HF-UNet has two complementary branches for two tasks, with the novel proposed attention-based task consistency learning block to communicate at each level between the two decoding branches. Therefore, HF-UNet endows the ability to learn hierarchically the shared representations for different tasks, and preserve the specificity of learned representations for different tasks simultaneously. We did extensive evaluations of the proposed method on a large planning CT image dataset and a benchmark prostate zonal dataset. The experimental results show HF-UNet outperforms the conventional multi-task network architectures and the state-of-the-art methods.

To better understand early brain development in health and disorder, it is critical to accurately segment infant brain magnetic resonance (MR) images into white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF). Deep learning-based methods have achieved state-of-the-art performance; h owever, one of the major limitations is that the learning-based methods may suffer from the multi-site issue, that is, the models trained on a dataset from one site may not be applicable to the datasets acquired from other sites with different imaging protocols/scanners. To promote methodological development in the community, the iSeg-2019 challenge ( http://iseg2019.web.unc.edu ) provides a set of 6-month infant subjects from multiple sites with different protocols/scanners for the participating methods. T raining/validation subjects are from UNC (MAP) and testing subjects are from UNC/UMN (BCP), Stanford University, and Emory University. By the time of writing, there are 30 automatic segmentation methods participated in the iSeg-2019. In this article, 8 top-ranked methods were reviewed by detailing their pipelines/implementations, presenting experimental results, and evaluating performance across different sites in terms of whole brain, regions of interest, and gyral landmark curves. We further pointed out their limitations and possible directions for addressing the multi-site issue. We find that multi-site consistency is still an open issue. We hope that the multi-site dataset in the iSeg-2019 and this review article will attract more researchers to address the challenging and critical multi-site issue in practice.

Automatic pancreas segmentation is crucial to the diagnostic assessment of diabetes or pancreatic cancer. However, the relatively small size of the pancreas in the upper body, as well as large variations of its location and shape in retroperitoneum, make the segmentation task challenging. To alleviate these challenges, in this article, we propose a cascaded multitask 3-D fully convolution network (FCN) to automatically segment the pancreas. Our cascaded network is composed of two parts. The first part focuses on fast locating the region of the pancreas, and the second part uses a multitask FCN with dense connections to refine the segmentation map for fine voxel-wise segmentation. In particular, our multitask FCN with dense connections is implemented to simultaneously complete tasks of the voxel-wise segmentation and skeleton extraction from the pancreas. These two tasks are complementary, that is, the extracted skeleton provides rich information about the shape and size of the pancreas in retroperitoneum, which can boost the segmentation of pancreas. The multitask FCN is also designed to share the low- and mid-level features across the tasks. A feature consistency module is further introduced to enhance the connection and fusion of different levels of feature maps. Evaluations on two pancreas datasets demonstrate the robustness of our proposed method in correctly segmenting the pancreas in various settings. Our experimental results outperform both baseline and state-of-the-art methods. Moreover, the ablation study shows that our proposed parts/modules are critical for effective multitask learning.

Brain connectivity alterations associated with mental disorders have been widely reported in both functional MRI (fMRI) and diffusion MRI (dMRI). However, extracting useful information from the vast amount of information afforded by brain networks remains a great challenge. Capturing network topology, graph convolutional networks (GCNs) have demonstrated to be superior in learning network representations tailored for identifying specific brain disorders. Existing graph construction techniques generally rely on a specific brain parcellation to define regions-of-interest (ROIs) to construct networks, often limiting the analysis into a single spatial scale. In addition, most methods focus on the pairwise relationships between the ROIs and ignore high-order associations between subjects. In this letter, we propose a mutual multi-scale triplet graph convolutional network (MMTGCN) to analyze functional and structural connectivity for brain disorder diagnosis. We first employ several templates with different scales of ROI parcellation to construct coarse-to-fine brain connectivity networks for each subject. Then, a triplet GCN (TGCN) module is developed to learn functional/structural representations of brain connectivity networks at each scale, with the triplet relationship among subjects explicitly incorporated into the learning process. Finally, we propose a template mutual learning strategy to train different scale TGCNs collaboratively for disease classification. Experimental results on 1, 160 subjects from three datasets with fMRI or dMRI data demonstrate that our MMTGCN outperforms several state-of-the-art methods in identifying three types of brain disorders.

Objective: To estimate a patient-specific reference bone shape model for a patient with craniomaxillofacial (CMF) defects due to facial trauma. Methods: We proposed an automatic facial bone shape estimation framework using pre-traumatic conventional portrait photos and post-traumatic head computed tomography (CT) scans via a 3D face reconstruction and a deformable shape model. Specifically, a three-dimensional (3D) face was first reconstructed from the patient's pre-traumatic portrait photos. Second, a correlation model between the skin and bone surfaces was constructed using a sparse representation based on the CT images of training normal subjects. Third, by feeding the reconstructed 3D face into the correlation model, an initial reference shape model was generated. In addition, we refined the initial estimation by applying non-rigid surface matching between the initially estimated shape and the patient's post-traumatic bone based on the adaptive-focus deformable shape model (AFDSM). Furthermore, a statistical shape model, built from the training normal subjects, was utilized to constrain the deformation process to avoid overfitting. Results and Conclusion: The proposed method was evaluated using both synthetic and real patient data. Experimental results show that the patient's abnormal facial bony structure can be recovered using our method, and the estimated reference shape model is considered clinically acceptable by an experienced CMF surgeon. Significance: The proposed method is more suitable to the complex CMF defects for CMF reconstructive surgical planning.

Multi-modal neuroimage retrieval has greatly facilitated the efficiency and accuracy of decision making in clinical practice by providing physicians with previous cases (with visually similar neuroimages) and corresponding treatment records. However, existing methods for image retrieval usually fail when applied directly to multi-modal neuroimage databases, since neuroimages generally have smaller inter-class variation and larger inter-modal discrepancy compared to natural images. To this end, we propose a deep Bayesian hash learning framework, called CenterHash, which can map multi-modal data into a shared Hamming space and learn discriminative hash codes from imbalanced multi-modal neuroimages. The key idea to tackle the small inter-class variation and large inter-modal discrepancy is to learn a common center representation for similar neuroimages from different modalities and encourage hash codes to be explicitly close to their corresponding center representations. Specifically, we measure the similarity between hash codes and their corresponding center representations and treat it as a center prior in the proposed Bayesian learning framework. A weighted contrastive likelihood loss function is also developed to facilitate hash learning from imbalanced neuroimage pairs. Comprehensive empirical evidence shows that our method can generate effective hash codes and yield state-of-the-art performance in cross-modal retrieval on three multi-modal neuroimage datasets.

Segmenting breast tumors from dynamic contrast-enhanced magnetic resonance (DCE-MR) images is a critical step for early detection and diagnosis of breast cancer. However, variable shapes and sizes of breast tumors, as well as inhomogeneous background, make it challenging to accurately segment tumors in DCE-MR images. Therefore, in this article, we propose a novel tumor-sensitive synthesis module and demonstrate its usage after being integrated with tumor segmentation. To suppress false-positive segmentation with similar contrast enhancement characteristics to true breast tumors, our tumor-sensitive synthesis module can feedback differential loss of the true and false breast tumors. Thus, by following the tumor-sensitive synthesis module after the segmentation predictions, the false breast tumors with similar contrast enhancement characteristics to the true ones will be effectively reduced in the learned segmentation model. Moreover, the synthesis module also helps improve the boundary accuracy while inaccurate predictions near the boundary will lead to higher loss. For the evaluation, we build a very large-scale breast DCE-MR image dataset with 422 subjects from different patients, and conduct comprehensive experiments and comparisons with other algorithms to justify the effectiveness, adaptability, and robustness of our proposed method.

Structured pruning has received ever-increasing attention as a method for compressing convolutional neural networks. However, most existing methods directly prune the network structure according to the statistical information of the parameters. Besides, these methods differentiate the pruning rates only in each pruning stage or even use the same pruning rate across all layers, rather than using learnable parameters. In this article, we propose a network redundancy elimination approach guided by the pruned model. Our proposed method can easily tackle multiple architectures and is scalable to the deeper neural networks because of the use of joint optimization during the pruning procedure. More specifically, we first construct a sparse self-representation for the filters or neurons of the well-trained model, which is useful for analyzing the relationship among filters. Then, we employ particle swarm optimization to learn pruning rates in a layerwise manner according to the performance of the pruned model, which can determine optimal pruning rates with the best performance of the pruned model. Under this criterion, the proposed pruning approach can remove more parameters without undermining the performance of the model. Experimental results demonstrate the effectiveness of our proposed method on different datasets and different architectures. For example, it can reduce 58.1% FLOPs for ResNet50 on ImageNet with only a 1.6% top-five error increase and 44.1% FLOPs for FCN_ResNet50 on COCO2017 with a 3% error increase, outperforming most state-of-the-art methods.

Accurate delineation of individual teeth and alveolar bones from dental cone-beam CT (CBCT) images is an essential step in digital dentistry for precision dental healthcare. In this paper, we present an AI system for efficient, precise, and fully automatic segmentation of real-patient CBCT images. Our AI system is evaluated on the largest dataset so far, i.e., using a dataset of 4,215 patients (with 4,938 CBCT scans) from 15 different centers. This fully automatic AI system achieves a segmentation accuracy comparable to experienced radiologists (e.g., 0.5% improvement in terms of average Dice similarity coefficient), while significant improvement in efficiency (i.e., 500 times faster). In addition, it consistently obtains accurate results on the challenging cases with variable dental abnormalities, with the average Dice scores of 91.5% and 93.0% for tooth and alveolar bone segmentation. These results demonstrate its potential as a powerful system to boost clinical workflows of digital dentistry.

Deep-learning methods (especially convolutional neural networks) using structural magnetic resonance imaging (sMRI) data have been successfully applied to computer-aided diagnosis (CAD) of Alzheimer's disease (AD) and its prodromal stage [i.e., mild cognitive impairment (MCI)]. As it is practically challenging to capture local and subtle disease-associated abnormalities directly from the whole-brain sMRI, most of those deep-learning approaches empirically preselect disease-associated sMRI brain regions for model construction. Considering that such isolated selection of potentially informative brain locations might be suboptimal, very few methods have been proposed to perform disease-associated discriminative region localization and disease diagnosis in a unified deep-learning framework. However, those methods based on task-oriented discriminative localization still suffer from two common limitations, that is: 1) identified brain locations are strictly consistent across all subjects, which ignores the unique anatomical characteristics of each brain and 2) only limited local regions/patches are used for model training, which does not fully utilize the global structural information provided by the whole-brain sMRI. In this article, we propose an attention-guided deep-learning framework to extract multilevel discriminative sMRI features for dementia diagnosis. Specifically, we first design a backbone fully convolutional network to automatically localize the discriminative brain regions in a weakly supervised manner. Using the identified disease-related regions as spatial attention guidance, we further develop a hybrid network to jointly learn and fuse multilevel sMRI features for CAD model construction. Our proposed method was evaluated on three public datasets (i.e., ADNI-1, ADNI-2, and AIBL), showing superior performance compared with several state-of-the-art methods in both tasks of AD diagnosis and MCI conversion prediction.

The coronavirus disease 2019 (COVID-19) pandemic has imposed dramatic challenges to health-care organizations worldwide. To combat the global crisis, the use of thoracic imaging has played a major role in the diagnosis, prediction, and management of COVID-19 patients with moderate to severe symptoms or with evidence of worsening respiratory status. In response, the medical image analysis community acted quickly to develop and disseminate deep learning models and tools to meet the urgent need of managing and interpreting large amounts of COVID-19 imaging data. This review aims to not only summarize existing deep learning and medical image analysis methods but also offer in-depth discussions and recommendations for future investigations. We believe that the wide availability of high-quality, curated, and benchmarked COVID-19 imaging data sets offers the great promise of a transformative test bed to develop, validate, and disseminate novel deep learning methods in the frontiers of data science and artificial intelligence.

Fast and automated image quality assessment (IQA) for diffusion MR images is a crucial step for swiftly making a rescan decision during or after the scanning session. However, learning a model for this task is challenging as the number of annotated data is limited and the annotation labels might not always be correct. As a remedy, we will introduce in this paper an automatic image quality assessment (IQA) method based on hierarchical non-local residual networks for pediatric diffusion MR images. Our IQA is performed in three sequential stages, i.e., 1) slice-wise IQA, where a nonlocal residual network is first pre-trained to annotate each slice with an initial quality rating (i.e., pass/questionable/fail), which is subsequently refined via iterative semi-supervised learning and slice self-training; 2) volume-wise IQA, which agglomerates the features extracted from slices of a volume, and uses a nonlocal network to annotate the quality rating for each volume via iterative volume self-training; and 3) subject-wise IQA, which ensembles the volumetric IQA results to determine the overall image quality pertaining to a subject. Experimental results demonstrate that our method, trained using only samples of modest size, exhibits great generalizability, and is capable of conducting rapid hierarchical IQA with near-perfect accuracy.

The resting-state functional magnetic resonance imaging (rs-fMRI) reflects functional activity of brain regions by blood-oxygen-level dependent (BOLD) signals. Up to now, many computer-aided diagnosis methods based on rs-fMRI have been developed for Autism Spectrum Disorder (ASD). These methods are mostly the binary classification approaches to determine whether a subject is an ASD patient or not. However, the disease often consists of several sub-categories, which are complex and thus still confusing to many automatic classification methods. Besides, existing methods usually focus on the functional connectivity (FC) features in grey matter regions, which only account for a small portion of the rs-fMRI data. Recently, the possibility to reveal the connectivity information in the white matter regions of rs-fMRI has drawn high attention. To this end, we propose to use the patch-based functional correlation tensor (PBFCT) features extracted from rs-fMRI in white matter, in addition to the traditional FC features from gray matter, to develop a novel multi-class ASD diagnosis method in this work. Our method has two stages. Specifically, in the first stage of multi-source domain adaptation (MSDA), the source subjects belonging to multiple clinical centers (thus called as source domains) are all transformed into the same target feature space. Thus each subject in the target domain can be linearly reconstructed by the transformed subjects. In the second stage of multi-view sparse representation (MVSR), a multi-view classifier for multi-class ASD diagnosis is developed by jointly using both views of the FC and PBFCT features. The experimental results using the ABIDE dataset verify the effectiveness of our method, which is capable of accurately classifying each subject into a respective ASD sub-category.

Obtaining accurate segmentation of the prostate and nearby organs at risk (e.g., bladder and rectum) in CT images is critical for radiotherapy of prostate cancer. Currently, the leading automatic segmentation algorithms are based on Fully Convolutional Networks (FCNs), which achieve remarkable performance but usually need large-scale datasets with high-quality voxel-wise annotations for full supervision of the training. Unfortunately, such annotations are difficult to acquire, which becomes a bottleneck to build accurate segmentation models in real clinical applications. In this paper, we propose a novel weakly supervised segmentation approach that only needs 3D bounding box annotations covering the organs of interest to start the training. Obviously, the bounding box includes many non-organ voxels that carry noisy labels to mislead the segmentation model. To this end, we propose the label denoising module and embed it into the iterative training scheme of the label denoising network (LDnet) for segmentation. The labels of the training voxels are predicted by the tentative LDnet, while the label denoising module identifies the voxels with unreliable labels. As only the good training voxels are preserved, the iteratively re-trained LDnet can refine its segmentation capability gradually. Our results are remarkable, i.e., reaching $\sim$ 94% (prostate), $\sim$ 91% (bladder), and $\sim$ 86% (rectum) of the Dice Similarity Coefficients (DSCs), compared to the case of fully supervised learning upon high-quality voxel-wise annotations and also superior to several state-of-the-art approaches. To our best knowledge, this is the first work to achieve voxel-wise segmentation in CT images from simple 3D bounding box annotations, which can greatly reduce many labeling efforts and meet the demands of the practical clinical applications.

Accurate segmentation of organs at risk (OARs) from head and neck (H&N) CT images is crucial for effective H&N cancer radiotherapy. However, the existing deep learning methods are often not trained in an end-to-end fashion, i.e., they independently predetermine the regions of target organs before organ segmentation, causing limited information sharing between related tasks and thus leading to suboptimal segmentation results. Furthermore, when conventional segmentation network is used to segment all the OARs simultaneously, the results often favor big OARs over small OARs. Thus, the existing methods often train a specific model for each OAR, ignoring the correlation between different segmentation tasks. To address these issues, we propose a new multi-view spatial aggregation framework for joint localization and segmentation of multiple OARs using H&N CT images. The core of our framework is a proposed region-of-interest (ROI)-based fine-grained representation convolutional neural network (CNN), which is used to generate multi-OAR probability maps from each 2D view (i.e., axial, coronal, and sagittal view) of CT images. Specifically, our ROI-based fine-grained representation CNN (1) unifies the OARs localization and segmentation tasks and trains them in an end-to-end fashion, and (2) improves the segmentation results of various-sized OARs via a novel ROI-based fine-grained representation. Our multi-view spatial aggregation framework then spatially aggregates and assembles the generated multi-view multi-OAR probability maps to segment all the OARs simultaneously. We evaluate our framework using two sets of H&N CT images and achieve competitive and highly robust segmentation performance for OARs of various sizes.

Multi-modal neuroimages, such as magnetic resonance imaging (MRI) and positron emission tomography (PET), can provide complementary structural and functional information of the brain, thus facilitating automated brain disease identification. Incomplete data problem is unavoidable in multi-modal neuroimage studies due to patient dropouts and/or poor data quality. Conventional methods usually discard data-missing subjects, thus significantly reducing the number of training samples. Even though several deep learning methods have been proposed, they usually rely on pre-defined regions-of-interest in neuroimages, requiring disease-specific expert knowledge. To this end, we propose a spatially-constrained Fisher representation framework for brain disease diagnosis with incomplete multi-modal neuroimages. We first impute missing PET images based on their corresponding MRI scans using a hybrid generative adversarial network. With the complete (after imputation) MRI and PET data, we then develop a spatially-constrained Fisher representation network to extract statistical descriptors of neuroimages for disease diagnosis, assuming that these descriptors follow a Gaussian mixture model with a strong spatial constraint (i.e., images from different subjects have similar anatomical structures). Experimental results on three databases suggest that our method can synthesize reasonable neuroimages and achieve promising results in brain disease identification, compared with several state-of-the-art methods.

Segmentation of prostate in medical imaging data (e.g., CT, MRI, TRUS) is often considered as a critical yet challenging task for radiotherapy treatment. It is relatively easier to segment prostate from MR images than from CT images, due to better soft tissue contrast of the MR images. For segmenting prostate from CT images, most previous methods mainly used CT alone, and thus their performances are often limited by low tissue contrast in the CT images. In this article, we explore the possibility of using indirect guidance from MR images for improving prostate segmentation in the CT images. In particular, we propose a novel deep transfer learning approach, i.e., MR-guided CT network training (namely MICS-NET), which can employ MR images to help better learning of features in CT images for prostate segmentation. In MICS-NET, the guidance from MRI consists of two steps: (1) learning informative and transferable features from MRI and then transferring them to CT images in a cascade manner, and (2) adaptively transferring the prostate likelihood of MRI model (i.e., well-trained convnet by purely using MR images) with a view consistency constraint. To illustrate the effectiveness of our approach, we evaluate MICS-NET on a real CT prostate image set, with the manual delineations available as the ground truth for evaluation. Our methods generate promising segmentation results which achieve (1) six percentages higher Dice Ratio than the CT model purely using CT images and (2) comparable performance with the MRI model purely using MR images.

Early identification of dementia at the stage of mild cognitive impairment (MCI) is crucial for timely diagnosis and intervention of Alzheimer's disease (AD). Although several pioneering studies have been devoted to automated AD diagnosis based on resting-state functional magnetic resonance imaging (rs-fMRI), their performance is somewhat limited due to non-effective mining of spatial-temporal dependency. Besides, few of these existing approaches consider the explicit detection and modeling of discriminative brain regions (i.e., network hubs) that are sensitive to AD progression. In this paper, we propose a unique Spatial-Temporal convolutional-recurrent neural Network (STNet) for automated prediction of AD progression and network hub detection from rs-fMRI time series. Our STNet incorporates the spatial-temporal information mining and AD-related hub detection into an end-to-end deep learning model. Specifically, we first partition rs-fMRI time series into a sequence of overlapping sliding windows. A sequence of convolutional components are then designed to capture the local-to-global spatially-dependent patterns within each sliding window, based on which we are able to identify discriminative hubs and characterize their unique contributions to disease diagnosis. A recurrent component with long short-term memory (LSTM) units is further employed to model the whole-brain temporal dependency from the spatially-dependent pattern sequences, thus capturing the temporal dynamics along time. We evaluate the proposed method on 174 subjects with 563 rs-fMRI scans from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database, with results suggesting the effectiveness of our method in both tasks of disease progression prediction and AD-related hub detection.

It has been recently shown that deep learning models such as convolutional neural networks (CNN), deep belief networks (DBN) and recurrent neural networks (RNN), exhibited remarkable ability in modeling and representing fMRI data for the understanding of functional activities and networks because of their superior data representation capability and wide availability of effective deep learning tools. For example, spatial and/or temporal patterns of functional brain activities embedded in fMRI data can be effectively characterized and modeled by a variety of CNN/DBN/RNN deep learning models as shown in recent studies. However, it has been rarely investigated whether it is possible to directly infer hierarchical brain networks from volumetric fMRI data using deep learning models such as DBN. The perceived difficulties of such studies include very large number of input variables, very large number of training parameters, the lack of effective software tools, the challenge of results interpretation, and etc. To bridge these technical gaps, we designed a novel volumetric sparse deep belief network (VS-DBN) model and implemented it through the popular TensorFlow open source platform to reconstruct hierarchical brain networks from volumetric fMRI data based on the Human Connectome Project (HCP) 900 subjects release. Our experimental results showed that a large number of interpretable and meaningful brain networks can be robustly reconstructed from HCP 900 subjects in a hierarchical fashion, and importantly, these brain networks exhibit reasonably good consistency and correspondence across multiple HCP task-based fMRI datasets. Our work contributed a new general deep learning framework for inferring multiscale volumetric brain networks and offered novel insights into the hierarchical organization of functional brain architecture.

Hippocampus plays an important role in the memory and spatial navigation function of human brain. Study on its growth and change during first year of life would assist the investigation of early brain development as well as the biomarker for neurological disorders. With the help of Magnetic Resonance (MR) imaging techniques, infant brain at different development stage can be acquired with multiple imaging modalities. In this situation, the longitudinal segmentation of infant hippocampus is highly demanded and feasible for the clinical studies regarding to the hippocampal volume changes. However, since the brain structures undergo dynamic appearance, structural changes and various tissue contrast during the first year of life, substantial challenges will be imposed for ensuring the robustness and accuracy of automatic hippocampus segmentation algorithms. In addition, most of the existing hippocampus segmentation methods generally handle each brain development stage independently without considering the potential longitudinal consistency among different stages. In view of the above factors, we propose a longitudinal classification-regression model for segmenting hippocampus in infant brain MRIs. Generally, our model proceeds on a per-timepoint basis, guided by the output of latter timepoint towards the infant hippocampus in the previous timepoint. The key ingredient of our method is a combination of longitudinal context, static context and appearance learning strategies under the classification-regression forest architecture. Specifically, the longitudinal context is borrowed from the mask of prior-timepoint estimation and the static context is from the current-timepoint estimation. Furthermore, we implement the proposed model in a multi-scale and iterative manner to improve the efficiency and effectiveness. The proposed method is evaluated on segmenting infant hippocampi from T1-weighted brain MR images acquired at the age of 2 weeks, 3 months, 6 months, 9 months, and 12 months. Experimental results demonstrate that our method achieves better performance in segmentation accuracy over the state-of-the-art classification and regression random forest model.

Deep learning has shown its great promise in various biomedical image segmentation tasks. Existing models are typically based on U-Net and rely on an encoder-decoder architecture with stacked local operators to aggregate long-range information gradually. However, only using the local operators limits the efficiency and effectiveness. In this work, we propose the non-local U-Nets, which are equipped with flexible global aggregation blocks, for biomedical image segmentation. These blocks can be inserted into U-Net as size-preserving processes, as well as down-sampling and up-sampling layers. We perform thorough experiments on the 3D multimodality isointense infant brain MR image segmentation task to evaluate the non-local U-Nets. Results show that our proposed models achieve top performances with fewer parameters and faster computation.

In various clinical scenarios, medical image is crucial in disease diagnosis and treatment. Different modalities of medical images provide complementary information and jointly helps doctors to make accurate clinical decision. However, due to clinical and practical restrictions, certain imaging modalities may be unavailable nor complete. To impute missing data with adequate clinical accuracy, here we propose a framework called self-supervised collaborative learning to synthesize missing modality for medical images. The proposed method comprehensively utilize all available information correlated to the target modality from multi-source-modality images to generate any missing modality in a single model. Different from the existing methods, we introduce an auto-encoder network as a novel, self-supervised constraint, which provides target-modality-specific information to guide generator training. In addition, we design a modality mask vector as the target modality label. With experiments on multiple medical image databases, we demonstrate a great generalization ability as well as specialty of our method compared with other state-of-the-arts.

Structural magnetic resonance imaging (sMRI) has been widely used for computer-aided diagnosis of neurodegenerative disorders, e.g., Alzheimer’s disease (AD), due to its sensitivity to morphological changes caused by brain atrophy. Recently, a few deep learning methods (e.g., convolutional neural networks, CNNs) have been proposed to learn task-oriented features from sMRI for AD diagnosis, and achieved superior performance than the conventional learning-based methods using hand-crafted features. However, these existing CNN-based methods still require the pre-determination of informative locations in sMRI. That is, the stage of discriminative atrophy localization is isolated to the latter stages of feature extraction and classifier construction. In this paper, we propose a hierarchical fully convolutional network (H-FCN) to automatically identify discriminative local patches and regions in the whole brain sMRI, upon which multi-scale feature representations are then jointly learned and fused to construct hierarchical classification models for AD diagnosis. Our proposed H-FCN method was evaluated on a large cohort of subjects from two independent datasets (i.e., ADNI-1 and ADNI-2), demonstrating good performance on joint discriminative atrophy localization and brain disease diagnosis.

Compared to computed tomography (CT), magnetic resonance imaging (MRI) delineation of craniomaxillofacial (CMF) bony structures can avoid harmful radiation exposure. However, bony boundaries are blurry in MRI, and structural information needs to be borrowed from CT during the training. This is challenging since paired MRI-CT data are typically scarce. In this paper, we propose to make full use of unpaired data, which are typically abundant, along with a single paired MRI-CT data to construct a one-shot generative adversarial model for automated MRI segmentation of CMF bony structures. Our model consists of a cross-modality image synthesis sub-network , which learns the mapping between CT and MRI, and an MRI segmentation sub-network . These two sub-networks are trained jointly in an end-to-end manner. Moreover, in the training phase, a neighbor-based anchoring method is proposed to reduce the ambiguity problem inherent in cross-modality synthesis, and a feature-matching-based semantic consistency constraint is proposed to encourage segmentation-oriented MRI synthesis. Experimental results demonstrate the superiority of our method both qualitatively and quantitatively in comparison with the state-of-the-art MRI segmentation methods.

Brain functional network has been increasingly used in understanding brain functions and diseases. While many network construction methods have been proposed, the progress in the field still largely relies on static pairwise Pearson's correlation‐based functional network and group‐level comparisons. We introduce a “Brain Network Construction and Classification (BrainNetClass)” toolbox to promote more advanced brain network construction methods to the filed, including some state‐of‐the‐art methods that were recently developed to capture complex and high‐order interactions among brain regions. The toolbox also integrates a well‐accepted and rigorous classification framework based on brain connectome features toward individualized disease diagnosis in a hope that the advanced network modeling could boost the subsequent classification. BrainNetClass is a MATLAB‐based, open‐source, cross‐platform toolbox with both graphical user‐friendly interfaces and a command line mode targeting cognitive neuroscientists and clinicians for promoting reliability, reproducibility, and interpretability of connectome‐based, computer‐aided diagnosis. It generates abundant classification‐related results from network presentations to contributing features that have been largely ignored by most studies to grant users the ability of evaluating the disease diagnostic model and its robustness and generalizability. We demonstrate the effectiveness of the toolbox on real resting‐state functional MRI datasets. BrainNetClass (v1.0) is available at https://github.com/zzstefan/BrainNetClass.

Autism spectrum disorder (ASD) is a neurodevelopmental disorder that is characterized by a wide range of symptoms. Identifying biomarkers for accurate diagnosis is crucial for early intervention of ASD. While multi-site data increase sample size and statistical power, they suffer from inter-site heterogeneity. To address this issue, we propose a multi-site adaption framework via low-rank representation decomposition (maLRR) for ASD identification based on functional MRI (fMRI). The main idea is to determine a common low-rank representation for data from the multiple sites, aiming to reduce differences in data distributions. Treating one site as a target domain and the remaining sites as source domains, data from these domains are transformed ( i.e. , adapted) to a common space using low-rank representation. To reduce data heterogeneity between the target and source domains, data from the source domains are linearly represented in the common space by those from the target domain. We evaluated the proposed method on both synthetic and real multi-site fMRI data for ASD identification. The results suggest that our method yields superior performance over several state-of-the-art domain adaptation methods.

Multi-atlas parcellation (MAP) is carried out on a brain image by propagating and fusing labelled regions from brain atlases. Typical nonlinear registration-based label propagation is time-consuming and sensitive to inter-subject differences. Recently, deep learning parcellation (DLP) has been proposed to avoid nonlinear registration for better efficiency and robustness than MAP. However, most existing DLP methods neglect using brain atlases, which contain high-level information (e.g., manually labelled brain regions), to provide auxiliary features for improving the parcellation accuracy. In this paper, we propose a novel multi-atlas DLP method for brain parcellation. Our method is based on fully convolutional networks (FCN) and squeeze-and-excitation (SE) modules. It can automatically and adaptively select features from the most relevant brain atlases to guide parcellation. Moreover, our method is trained via a generative adversarial network (GAN), where a convolutional neural network (CNN) with multi-scale $l_{1}$ loss is used as the discriminator. Benefiting from brain atlases, our method outperforms MAP and state-of-the-art DLP methods on two public image datasets (LPBA40 and NIREP-NA0).

As the population becomes older worldwide, accurate computer-aided diagnosis for Alzheimer's disease (AD) in the early stage has been regarded as a crucial step for neurodegeneration care in recent years. Since it extracts the low-level features from the neuroimaging data, previous methods regarded this computer-aided diagnosis as a classification problem that ignored latent featurewise relation. However, it is known that multiple brain regions in the human brain are anatomically and functionally interlinked according to the current neuroscience perspective. Thus, it is reasonable to assume that the extracted features from different brain regions are related to each other to some extent. Also, the complementary information between different neuroimaging modalities could benefit multimodal fusion. To this end, we consider leveraging the coupled interactions in the feature level and modality level for diagnosis in this paper. First, we propose capturing the feature-level coupled interaction using a coupled feature representation. Then, to model the modality-level coupled interaction, we present two novel methods: 1) the coupled boosting (CB) that models the correlation of pairwise coupled-diversity on both inconsistently and incorrectly classified samples between different modalities and 2) the coupled metric ensemble (CME) that learns an informative feature projection from different modalities by integrating the intrarelation and interrelation of training samples. We systematically evaluated our methods with the AD neuroimaging initiative data set. By comparison with the baseline learning-based methods and the state-of-the-art methods that are specially developed for AD/MCI (mild cognitive impairment) diagnosis, our methods achieved the best performance with accuracy of 95.0% and 80.7% (CB), 94.9% and 79.9% (CME) for AD/NC (normal control), and MCI/NC identification, respectively.

Precise segmentation of teeth from intra-oral scanner images is an essential task in computer-aided orthodontic surgical planning. The state-of-the-art deep learning-based methods often simply concatenate the raw geometric attributes (i.e., coordinates and normal vectors) of mesh cells to train a single-stream network for automatic intra-oral scanner image segmentation. However, since different raw attributes reveal completely different geometric information, the naive concatenation of different raw attributes at the (low-level) input stage may bring unnecessary confusion in describing and differentiating between mesh cells, thus hampering the learning of high-level geometric representations for the segmentation task. To address this issue, we design a two-stream graph convolutional network (i.e., TSGCN), which can effectively handle inter-view confusion between different raw attributes to more effectively fuse their complementary information and learn discriminative multi-view geometric representations. Specifically, our TSGCN adopts two input-specific graph-learning streams to extract complementary high-level geometric representations from coordinates and normal vectors, respectively. Then, these single-view representations are further fused by a self-attention module to adaptively balance the contributions of different views in learning more discriminative multi-view representations for accurate and fully automatic tooth segmentation. We have evaluated our TSGCN on a real-patient dataset of dental (mesh) models acquired by 3D intraoral scanners. Experimental results show that our TSGCN significantly outperforms state-of-the-art methods in 3D tooth (surface) segmentation. Github: https://github.com/ZhangLingMing1/TSGCNet., Comment: 11 pages, 6 figures. arXiv admin note: text overlap with arXiv:2012.13697

Missing data is a common problem in longitudinal studies due to subject dropouts and failed scans. We present a graph-based convolutional neural network to predict missing diffusion MRI data. In particular, we consider the relationships between sampling points in the spatial domain and the diffusion wave-vector domain to construct a graph. We then use a graph convolutional network to learn the non-linear mapping from available data to missing data. Our method harnesses a multi-scale residual architecture with adversarial learning for prediction with greater accuracy and perceptual quality. Experimental results show that our method is accurate and robust in the longitudinal prediction of infant brain diffusion MRI data.

Accurate bone segmentation and anatomical landmark localization are essential tasks in computer-aided surgical simulation for patients with craniomaxillofacial (CMF) deformities. To leverage the complementarity between the two tasks, we propose an efficient end-to-end deep network, i.e., multi-task dynamic transformer network (DTNet), to concurrently segment CMF bones and localize large-scale landmarks in one-pass from large volumes of cone-beam computed tomography (CBCT) data. Our DTNet was evaluated quantitatively using CBCTs of patients with CMF deformities. The results demonstrated that our method outperforms the other state-of-the-art methods in both tasks of the bony segmentation and the landmark digitization. Our DTNet features three main technical contributions. First, a collaborative two-branch architecture is designed to efficiently capture both fine-grained image details and complete global context for high-resolution volume-to-volume prediction. Second, leveraging anatomical dependencies between landmarks, regionalized dynamic learners (RDLs) are designed in the concept of “learns to learn” to jointly regress large-scale 3D heatmaps of all landmarks under limited computational costs. Third, adaptive transformer modules (ATMs) are designed for the flexible learning of task-specific feature embedding from common feature bases.

Landmark localization is an important step in quantifying craniomaxillofacial (CMF) deformities and designing treatment plans of reconstructive surgery. However, due to the severity of deformities and defects (partially missing anatomy), it is difficult to automatically and accurately localize a large set of landmarks simultaneously. In this work, we propose two cascaded networks for digitizing 60 anatomical CMF landmarks in cone-beam computed tomography (CBCT) images. The first network is a U-Net that outputs heatmaps for landmark locations and landmark features extracted with a local attention mechanism. The second network is a graph convolution network that takes the features extracted by the first network as input and determines whether each landmark exists via binary classification. We evaluated our approach on 50 sets of CBCT scans of patients with CMF deformities and compared them with state-of-the-art methods. The results indicate that our approach can achieve an average detection error of 1.47 mm with a false positive rate of 19%, outperforming related methods.

Automatic craniomaxillofacial (CMF) landmark localization from cone-beam computed tomography (CBCT) images is challenging, considering that 1) the number of landmarks in the images may change due to varying deformities and traumatic defects, and 2) the CBCT images used in clinical practice are typically large. In this paper, we propose a two-stage, coarse-to-fine deep learning method to tackle these challenges with both speed and accuracy in mind. Specifically, we first use a 3D faster R-CNN to roughly locate landmarks in down-sampled CBCT images that have varying numbers of landmarks. By converting the landmark point detection problem to a generic object detection problem, our 3D faster R-CNN is formulated to detect virtual, fixed-size objects in small boxes with centers indicating the approximate locations of the landmarks. Based on the rough landmark locations, we then crop 3D patches from the high-resolution images and send them to a multi-scale UNet for the regression of heatmaps, from which the refined landmark locations are finally derived. We evaluated the proposed approach by detecting up to 18 landmarks on a real clinical dataset of CMF CBCT images with various conditions. Experiments show that our approach achieves state-of-the-art accuracy of 0.89 ± 0.64mm in an average time of 26.2 seconds per volume.

Background This study aimed to evaluate the utility of radiomics-based machine learning analysis with multiparametric DWI and to compare the diagnostic performance of radiomics features and mean diffusion metrics in the characterization of breast lesions. Methods This retrospective study included 542 lesions from February 2018 to November 2018. One hundred radiomics features were computed from mono-exponential (ME), biexponential (BE), stretched exponential (SE), and diffusion-kurtosis imaging (DKI). Radiomics-based analysis was performed by comparing four classifiers, including random forest (RF), principal component analysis (PCA), L1 regularization (L1R), and support vector machine (SVM). These four classifiers were trained on a training set with 271 patients via ten-fold cross-validation and tested on an independent testing set with 271 patients. The diagnostic performance of the mean diffusion metrics of ME (mADCall b, mADC0–1000), BE (mD, mD*, mf), SE (mDDC, mα), and DKI (mK, mD) were also calculated for comparison. The area under the receiver operating characteristic curve (AUC) was used to compare the diagnostic performance. Results RF attained higher AUCs than L1R, PCA and SVM. The AUCs of radiomics features for the differential diagnosis of breast lesions ranged from 0.80 (BE_D*) to 0.85 (BE_D). The AUCs of the mean diffusion metrics ranged from 0.54 (BE_mf) to 0.79 (ME_mADC0–1000). There were significant differences in the AUCs between the mean values of all diffusion metrics and radiomics features of AUCs (all P Conclusions The radiomics-based analysis of multiparametric DWI by RF enables better differentiation of benign and malignant breast lesions than the mean diffusion metrics.

Retrospective artifact correction (RAC) improves image quality post acquisition and enhances image usability. Recent machine learning driven techniques for RAC are predominantly based on supervised learning and therefore practical utility can be limited as data with paired artifact-free and artifact-corrupted images are typically insufficient or even non-existent. Here we show that unwanted image artifacts can be disentangled and removed from an image via an RAC neural network learned with unpaired data. This implies that our method does not require matching artifact-corrupted data to be either collected via acquisition or generated via simulation. Experimental results demonstrate that our method is remarkably effective in removing artifacts and retaining anatomical details in images with different contrasts.

Segmentation of infections from CT scans is important for accurate diagnosis and follow-up in tackling the COVID-19. Although the convolutional neural network has great potential to automate the segmentation task, most existing deep learning-based infection segmentation methods require fully annotated ground-truth labels for training, which is time-consuming and labor-intensive. This paper proposed a novel weakly supervised segmentation method for COVID-19 infections in CT slices, which only requires scribble supervision and is enhanced with the uncertainty-aware self-ensembling and transformation-consistent techniques. Specifically, to deal with the difficulty caused by the shortage of supervision, an uncertainty-aware mean teacher is incorporated into the scribble-based segmentation method, encouraging the segmentation predictions to be consistent under different perturbations for an input image. This mean teacher model can guide the student model to be trained using information in images without requiring manual annotations. On the other hand, considering the output of the mean teacher contains both correct and unreliable predictions, equally treating each prediction in the teacher model may degrade the performance of the student network. To alleviate this problem, the pixel level uncertainty measure on the predictions of the teacher model is calculated, and then the student model is only guided by reliable predictions from the teacher model. To further regularize the network, a transformation-consistent strategy is also incorporated, which requires the prediction to follow the same transformation if a transform is performed on an input image of the network. The proposed method has been evaluated on two public datasets and one local dataset. The experimental results demonstrate that the proposed method is more effective than other weakly supervised methods and achieves similar performance as those fully supervised.