Search

Chest computed tomography (CT) scans of coronavirus 2019 (COVID-19) disease usually come from multiple datasets gathered from different medical centers, and these images are sampled using different acquisition protocols. While integrating multicenter datasets increases sample size, it suffers from inter-center heterogeneity. To address this issue, we propose an augmented multicenter graph convolutional network (AM-GCN) to diagnose COVID-19 with steps as follows. First, we use a 3-D convolutional neural network to extract features from the initial CT scans, where a ghost module and a multitask framework are integrated to improve the network's performance. Second, we exploit the extracted features to construct a multicenter graph, which considers the intercenter heterogeneity and the disease status of training samples. Third, we propose an augmentation mechanism to augment training samples which forms an augmented multicenter graph. Finally, the diagnosis results are obtained by inputting the augmented multi-center graph into GCN. Based on 2223 COVID-19 subjects and 2221 normal controls from seven medical centers, our method has achieved a mean accuracy of 97.76%. The code for our model is made publicly. 1 .

Objective: This study aimed to compare the clinical effects of uterine artery embolization (UAE) with those of high-intensity focused ultrasound (HIFU) ablation for the treatment of symptomatic uterine myomas., Data Sources: We searched PubMed, EMBASE, Web of Science, Cochrane Library, Google Scholar, and ClinicalTrials.gov for studies from January 2000 to August 2020. Related articles and relevant references of the included studies were also searched., Methods of Study Selection: Two researchers independently performed the data selection. We included comparative studies that compared the clinical outcomes of UAE with those of HIFU ablation in women with myomas., Tabulation, Integration, and Results: We assessed the study quality using the Cochrane Handbook for Systematic Reviews of Interventions for evaluating the risk of bias. Two independent researchers performed the article selection according to the screening criteria and rated the quality of evidence for each article. We calculated pooled mean difference with 95% confidence interval (CI) for continuous data and relative risk (RR) with 95% CI for dichotomous data. The systematic review registration number is CRD42020199630 on the International Prospective Register of Systematic Reviews. A total of 7 articles (5 trials), involving 4592 women with symptomatic uterine myomas, were included in the meta-analysis. Compared with the HIFU ablation group, the decrease in "uterine fibroid symptom" scores as well as the increase in quality-of-life scores at the time of follow-up were higher in the UAE group, with overall mean difference 19.54 (95% CI, 15.21-23.87; p <.001) and 15.72 (95% CI, 8.30-23.13; p <.001), respectively. The women in the UAE group had a significantly lower reintervention rate (RR 0.25; 95% CI, 0.15-0.42; p <.001). The women undergoing UAE had a significantly lower pregnancy rate than those undergoing HIFU ablation (RR 0.06; 95% CI, 0.01-0.45; p = .006). The difference in the incidence of adverse events between the 2 groups was not statistically significant (p = .53)., Conclusion: Compared with HIFU ablation, UAE provided more significant alleviation of symptoms and improvement in quality of life, lower postoperative reintervention rate, and lower pregnancy rate for women with uterine myomas. However, we cannot conclude that HIFU ablation is more favorable for desired pregnancy than UAE because of the confounding factors., (Copyright © 2020 AAGL. Published by Elsevier Inc. All rights reserved.)

Mitotic count is an important indicator for assessing the invasiveness of breast cancers. Currently, the number of mitoses is manually counted by pathologists, which is both tedious and time-consuming. To address this situation, we propose a fast and accurate method to automatically detect mitosis from the histopathological images. The proposed method can automatically identify mitotic candidates from histological sections for mitosis screening. Specifically, our method exploits deep convolutional neural networks to extract high-level features of mitosis to detect mitotic candidates. Then, we use spatial attention modules to re-encode mitotic features, which allows the model to learn more efficient features. Finally, we use multi-branch classification subnets to screen the mitosis. Compared to existing related methods in literature, our method obtains the best detection results on the dataset of the International Pattern Recognition Conference (ICPR) 2012 Mitosis Detection Competition. Code has been made available at: https://github.com/liushaomin/MitosisDetection.

Alzheimer's disease (AD) is a neurodegenerative disease with an irreversible and progressive process. To understand the brain functions and identify the biomarkers of AD and early stages of the disease [also known as, mild cognitive impairment (MCI)], it is crucial to build the brain functional connectivity network (BFCN) using resting-state functional magnetic resonance imaging (rs-fMRI). Existing methods have been mainly developed using only a single time-point rs-fMRI data for classification. In fact, multiple time-point data is more effective than a single time-point data in diagnosing brain diseases by monitoring the disease progression patterns using longitudinal analysis. In this article, we utilize multiple rs-fMRI time-point to identify early MCI (EMCI) and late MCI (LMCI), by integrating the fused sparse network (FSN) model with parameter-free centralized (PFC) learning. Specifically, we first construct the FSN framework by building multiple time-point BFCNs. The multitask learning via PFC is then leveraged for longitudinal analysis of EMCI and LMCI. Accordingly, we can jointly learn the multiple time-point features constructed from the BFCN model. The proposed PFC method can automatically balance the contributions of different time-point information via learned specific and common features. Finally, the selected multiple time-point features are fused by a similarity network fusion (SNF) method. Our proposed method is evaluated on the public AD neuroimaging initiative phase-2 (ADNI-2) database. The experimental results demonstrate that our method can achieve quite promising performance and outperform the state-of-the-art methods.

The automated whole breast ultrasound (AWBUS) is a new breast imaging technique that can depict the whole breast anatomy. To facilitate the reading of AWBUS images and support the breast density estimation, an automatic breast anatomy segmentation method for AWBUS images is proposed in this study. The problem at hand is quite challenging as it needs to address issues of low image quality, ill-defined boundary, large anatomical variation, etc. To address these issues, a new deep learning encoder-decoder segmentation method based on a self-co-attention mechanism is developed. The self-attention mechanism is comprised of spatial and channel attention module (SC) and embedded in the ResNeXt (i.e., Res-SC) block in the encoder path. A non-local context block (NCB) is further incorporated to augment the learning of high-level contextual cues. The decoder path of the proposed method is equipped with the weighted up-sampling block (WUB) to attain class-specific better up-sampling effect. Meanwhile, the co-attention mechanism is also developed to improve the segmentation coherence among two consecutive slices. Extensive experiments are conducted with comparison to several the state-of-the-art deep learning segmentation methods. The experimental results corroborate the effectiveness of the proposed method on the difficult breast anatomy segmentation problem on AWBUS images., Competing Interests: Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2020 Elsevier B.V. All rights reserved.)

Gliomas are the most dominant and lethal type of brain tumors. Growth prediction is significant to quantify tumor aggressiveness, improve therapy planning, and estimate patients' survival time. This is commonly addressed in literature using mathematical models guided by multi-time point scans of multi/single-modal data for the same subject. However, these models are mechanism-based and heavily rely on complicated mathematical formulations of partial differential equations with few parameters that are insufficient to capture different patterns and other characteristics of gliomas. In this paper, we propose a 3D generative adversarial networks (GANs) for glioma growth prediction. Specifically, we stack 2 GANs with conditional initialization of segmented feature maps. Furthermore, we employ Dice loss in our objective function and devised 3D U-Net architecture for better image generation. The proposed method is trained and validated using 3D patch-based strategy on real magnetic resonance images of 9 subjects with 3 time points. Experimental results show that the proposed method can be successfully used for glioma growth prediction with satisfactory performance.

Quality control/assessment of ultrasound (US) images is an essential step in clinical diagnosis. This process is usually done manually, suffering from some drawbacks, such as dependence on operator's experience and extensive labors, as well as high inter- and intra-observer variation. Automatic quality assessment of US images is therefore highly desirable. Fetal US cardiac four-chamber plane (CFP) is one of the most commonly used cardiac views, which was used in the diagnosis of heart anomalies in the early 1980s. In this paper, we propose a generic deep learning framework for automatic quality control of fetal US CFPs. The proposed framework consists of three networks: (1) a basic CNN (B-CNN), roughly classifying four-chamber views from the raw data; (2) a deeper CNN (D-CNN), determining the gain and zoom of the target images in a multi-task learning manner; and (3) the aggregated residual visual block net (ARVBNet), detecting the key anatomical structures on a plane. Based on the output of the three networks, overall quantitative score of each CFP is obtained, so as to achieve fully automatic quality control. Experiments on a fetal US dataset demonstrated our proposed method achieved a highest mean average precision (mAP) of 93.52% at a fast speed of 101 frames per second (FPS). In order to demonstrate the adaptability and generalization capacity, the proposed detection network (i.e., ARVBNet) has also been validated on the PASCAL VOC dataset, obtaining a highest mAP of 81.2% when input size is approximately 300 × 300.

Accurate segmentation of pediatric echocardiography is an essential preprocessing step for a wide range of analysis tasks. Currently, it highly relies on sonographer's manual segmentation, which is time-consuming and redundant, and therefore might lead to mistakes. In this paper, we present a deep learning method based on Bilateral Segmentation Network (BiSeNet) to fully automatic segment pediatric echocardiography images in 4 chamber view. BiSeNet consists of two paths, a spatial path for capturing low-level spatial features, and a context path for exploiting high-level context semantic features. In addition, a feature fusion module is used to fuse features learned by both the two paths. Experiments based on our selfcollected dataset shows that our method achieves 0.932, and 0.908 in term of Dice index in the left ventricle and left atrium segmentation task, which outperforms different state-of-the-art U-Net architectures.

Detection of mild cognitive impairment (MCI) is important, and appropriate interventions can be taken to delay or prevent its progression to Alzheimer's disease (AD). The construction of brain networks based on brain image data to depict the interaction of brain functions or structures at the level of brain connections has been widely used to identify individuals with MCI/AD from the normal control (NC). Exploring the structural and functional connections and interactions between brain regions is beneficial to detect MCI. For this reason, we propose a new model for automatic MCI diagnosis based on this information. Firstly, a new functional brain network estimation method is proposed. Self-calibration is introduced using quality indicators, and functional brain network estimation is performed at the same time. Then we integrate the functional and structural connected neuroimaging patterns into our multitask learning model to select informative feature. By identifying synergies and differences between different tasks, the most discriminative features are determined. Finally, the most relevant features are sent to the support vector machine classifier for diagnosis and identification of MCI. The experimental results based on the public Alzheimer's disease neuroimaging (ADNI) show that our method can effectively diagnose different stages of MCI and assist the physician to improve the MCI diagnostic accuracy. At the same time, compared with the existing classification methods, the proposed method achieves relatively high classification accuracy. In addition, it can identify the most discriminative brain regions. These findings suggest that our approach not only improves classification performance, but also successfully identifies important biomarkers associated with disease.

Alzheimer's disease (AD), the most common type of the dementia, is a progressive neurodegenerative disease that mainly affects elderly. It causes a high financial burden for patients and their families. For effective treatment of AD, it is important to identify the AD progression of clinical disease over time. As the cognitive scores can effectively indicate the disease status, the prediction of the scores using the longitudinal magnetic resonance imaging (MRI) data is highly desirable. In this paper, we propose a joint learning and clinical scores prediction method for AD diagnosis via longitudinal MRI data. Specifically, we devise a novel feature selection method that consists of a temporally constrained group LASSO model and the correntropy. The baseline MRI data is used to jointly select the most discriminative features. Then, we use the stacked long short-term memory (SLSTM) to effectively capture useful information in the input sequence to predict the clinical scores of future time points. Extensive experiments on the Alzheimer's disease Neuroimaging Initiative (ADNI) database are conducted to demonstrate the effectiveness of the proposed model. Our model can accurately describe the relationship between MRI data and scores, and thus it can be effective in predicting longitudinal scores.

Breast cancer is one of the leading causes of cancer death worldwide. Recently, the computer-aided diagnosis and detection technique has been developed for the early diagnosis of breast cancer, but the diagnostic efficiency has still been a challenging issue. For this reason, we aim to improve the breast cancer diagnostic accuracy and reduce the workload of doctors in this paper by devising a deep learning framework based on histological image. Therefore, we develop a model of multi-level feature of dual-network combined with sparse multi-relation regularized learning method, which enhances the classification performance and robustness. Specifically, first, we preprocess the histological images using scale transformation and color enhancement methods. Second, the multi-level features are extracted from preprocessed images using InceptionV3-ML and ResNet-50 networks. Third, the feature selection method via sparse multi-relation regularization is further developed for performance boosting and overfitting reduction. We evaluate the proposed method based on the public ICIAR 2018 Challenge dataset of breast cancer histology images. Experimental results show that our method has achieved promising performance and outperformed the related works.

Breast cancer grading is important for patient prognosis, and the mitosis count is one of the most important indicators for breast cancer grading. Traditional methods use handcraft features and deep learning based methods to detect mitosis in a classified model. These methods are time-consuming and difficult for practical clinical practice application. For this reason, this paper proposes an improved object detection method for automatic mitosis detection from histological images. First, we use a convolutional neural network (CNN) to automatically extract mitosis features. Then, we use the region proposed network (RPN) to locate a set of class-agnostic mitosis proposals. Finally, we use the improved R-CNN subnet to screen for mitosis from these proposals. Our approach achieved the best results in the ICPR2012 mitosis detection competition test dataset. Additionally, our proposed method is fast enough to be potentially used in clinical and health centers.

Melanoma is one of the most deadly skin lesion, which often uses the skin dermoscopy to detect it. However, the low interclass variations between melanoma images make manual dermoscopic detection time-consuming and laborious. Therefore, an automatic recognition algorithm of skin image is highly desirable. However, the traditional methods still have the limitations (e.g., weak robustness and generalization ability). To meet the challenge, we propose an effective architecture based on residual - squeeze - and - excitation -Inception-v4 network (MelanomaNet) to detect melanoma. Specifically, Inception-v4 structure is utilized to get the rich spatial features and increase feature diversity. We also consider the relationship between feature channels by adding residual-squeeze-and-excitation (RSE) blocks in Inception- v4 network using the feature recalibration strategies. Finally, we use the support vector machine (SVM) as the classifier for the skin lesion classification. We evaluate our proposed method on the public available ISIC skin lesion challenge datasets in 2018 for training and evaluation. The experimental results show that the proposed method has achieved better performance over the state-of-the-arts methods.

Neuroimaging data has been widely used to predict clinical scores for automatic diagnosis of Alzheimer's disease (AD). For accurate clinical score prediction, one of the major challenges is high feature dimension of the imaging data. To address this issue, this paper presents an effective framework using a novel feature selection model via sparse learning. In contrast to previous approaches focusing on a single time point, this framework uses information at multiple time points. Specifically, a regularized correntropy with the spatial-temporal constraint is used to reduce the adverse effect of noise and outliers, and promote consistent and robust selection of features by exploring data characteristics. Furthermore, ensemble learning of support vector regression (SVR) is exploited to accurately predict AD scores based on the selected features. The proposed approach is extensively evaluated on the Alzheimer's disease neuroimaging initiative (ADNI) dataset. Our experiments demonstrate that the proposed approach not only achieves promising regression accuracy, but also successfully recognizes disease-related biomarkers.

Alzheimer's disease (AD) is a neurodegenerative disease with an irreversible and progressive process. Close monitoring of AD is essential for making adjustments in the treatment plan. Since clinical scores can indicate the disease status effectively, the prediction of the scores based on the magnetic resonance imaging (MRI data is highly desirable. Different from previous studies at a single time point, we propose to build a model to explore the relationship between MRI data and scores, thereby predicting longitudinal scores at future time points from the corresponding MRI data. The model incorporates three parts, correntropy regularized joint learning-based feature selection, deep polynomial network based feature encoding, and finally, support vector regression. The regression process is carried out for two scenarios. One is to use baseline data for predictions at future time points, and the other is to combine all the previous data for the prediction at the next time point. Meanwhile, the missing scores are filled in the second scenario to address the incompleteness presented in the data. The simulation results demonstrate that the proposed model accurately describes the relationship between MRI data and scores, and thus it can be effective in predicting longitudinal scores.

Tumor growth modeling at macroscopic level from multimodal images can help in predicting the future evolution of tumor and the treatment planning. This can be achieved using mathematical models where multi time-point images are available. In this paper, we propose a coupled modified reaction diffusion model that measures tumor invasion and infiltration with biomechanical model to consider tumor mass effect. In addition, our model considers treatment effects from radiotherapy and/or chemotherapy if any. The chemotherapy effect is included via a modified log-kill method to consider tissue heterogeneity while radiotherapy effect is considered using the linear quadratic model. We test the proposed model on both synthetic and 6 real datasets of low grade glioma cases with and without treatments. Experimental results of the proposed model on the clinical magnetic resonance images show that our model can simulate the tumor growth with good accuracy and effectively include the treatment effects.

Cystocele is a common disease in woman. Accurate assessment of cystocele severity is very important for treatment options. The transperineal ultrasound (US) has recently emerged as an alternative tool for cystocele grading. The cystocele severity is usually evaluated with the manual measurement of the maximal descent of the bladder (MDB) relative to the symphysis pubis (SP) during Valsalva maneuver. However, this process is time-consuming and operator-dependent. In this study, we propose an automatic scheme for csystocele grading from transperineal US video. A two-layer spatio-temporal regression model is proposed to identify the middle axis and lower tip of the SP, and segment the bladder, which are essential tasks for the measurement of the MDB. Both appearance and context features are extracted in the spatio-temporal domain to help the anatomy detection. Experimental results on 85 transperineal US videos show that our method significantly outperforms the state-of-the-art regression method.

Currently, the evaluation of placental maturity has mainly focused on subjective measure, which highly depends on the observation and experiences of the clinicians and not reliable. This paper proposes a new method for grading placenta maturity in B-mod ultrasound (US) images automatically based on local intensity order pattern (LIOP) and fisher vector (FV). After extracting invariant LIOP feature from the affine covariant region, the feature is encoded by FV to improve the classification accuracy and reduce the processing time. Experimental results show the effectiveness of the proposed method with an accuracy of 0.9375, a sensitivity of 0.9804 and a specificity of 0.9375 for the placental maturity grading. Moreover, experimental results demonstrate that the LIOP feature outperforms the traditional SIFT feature for grading.

In this paper, a superpixel and convolution neural network (CNN) based segmentation method is proposed for cervical cancer cell segmentation. Since the background and cytoplasm contrast is not relatively obvious, cytoplasm segmentation is first performed. Deep learning based on CNN is explored for region of interest detection. A coarse-to-fine nucleus segmentation for cervical cancer cell segmentation and further refinement is also developed. Experimental results show that an accuracy of 94.50% is achieved for nucleus region detection and a precision of 0.9143±0.0202 and a recall of 0.8726±0.0008 are achieved for nucleus cell segmentation. Furthermore, our comparative analysis also shows that the proposed method outperforms the related methods.