Multi-scale Bottleneck Residual Network for Retinal Vessel Segmentation.

Precise segmentation of retinal vessels is crucial for the prevention and diagnosis of ophthalmic diseases. In recent years, deep learning has shown outstanding performance in retinal vessel segmentation. Many scholars are dedicated to studying retinal vessel segmentation methods based on color fundus images, but the amount of research works on Scanning Laser Ophthalmoscopy (SLO) images is very scarce. In addition, existing SLO image segmentation methods still have difficulty in balancing accuracy and model parameters. This paper proposes a SLO image segmentation model based on lightweight U-Net architecture called MBRNet, which solves the problems in the current research through Multi-scale Bottleneck Residual (MBR) module and attention mechanism. Concretely speaking, the MBR module expands the receptive field of the model at a relatively low computational cost and retains more detailed information. Attention Gate (AG) module alleviates the disturbance of noise so that the network can concentrate on vascular characteristics. Experimental results on two public SLO datasets demonstrate that by comparison to existing methods, the MBRNet has better segmentation performance with relatively few parameters.

Effect of 5-Line Signs in the Prediction of Staging, Progression, and Prognosis of Peripheral Lung Carcinoma: Preliminary Observation Report.

BACKGROUND: The single line of the normal interlobar fissure shown on a thin section image can be reconstructed as a 5-line sign on axial maximal intensity projection. The line between the lung nodule and the pleura is called the pleural tail sign on thin image. On the axial maximal intensity projection, it can also be reconstructed as the 5-line sign or fewer than 5 lines. OBJECTIVE: This study aimed to observe the effect of 5-line signs in staging, progression, and prognosis of peripheral lung carcinoma. MATERIALS AND METHODS: This study included 132 patients with peripheral lung carcinoma. Among these patients, 93 were men and 39 were women, with an age range of 27 to 82 years and a lung nodule range of 0.98 to 8.75 cm. Maximal intensity projection was reconstructed based on 1.0 or 1.25 mm of thin-slice images in multislice spiral computed tomography. Five-line signs on the margin of the nodule (mass) were observed and were classified into grades 1 to 4 according to the sharpness of the 5-line signs. RESULTS: Multivariate logistic regression analysis showed that the sharpness of the 5-line signs was correlated with N and TNM staging of peripheral lung carcinoma (P = 0.012, P = 0.016). The lower the sharpness of the 5-line signs, the greater the number of cases of progression of the tumor (P < 0.001), and thus the higher the mortality rate and the lower the survival rate (P = 0.001). The sensitivity and specificity of predicting tumor progression were 56.3% and 93.3%, and those of tumor prognosis were 61.1% and 82.4%, respectively. CONCLUSIONS: The sharpness of the 5-line signs has certain effects on the prediction of invasion, progression, and prognosis of lung carcinoma, particularly of small lung cancer (≤3.0 cm).

[Correlation between the mRNA expression of tissue inhibitor of metalloproteinase-1 and apparent diffusion coefficient on diffusion-weighted imaging in rats' liver fibrosis].

OBJECTIVE: To explore the correlation between the apparent diffusion coefficient (ADC) and mRNA expression of tissue inhibitor of metalloproteinase-1 (TIMP-1) in different stages of liver fibrosis in rats.
 Methods: A model of liver fibrosis in rats was established by intraperitoneal injection of high-fat diet combined with porcine serum. After drug administration for 4 weeks, 48 rats served as a model group and 12 rats served as a control group, then they underwent diffusion weighted imaging (DWI) scanning. The value of ADC was calculated at b value=800 s/mm2. The rats were sacrificed and carried out pathologic examination after DWI scanning immediately. The mRNA expression of TIMP-1 was detected by real time-polymerase chain reaction (RT-PCR). The rats of hepatic fibrosis were also divided into a S0 group (n=4), a S1 group (n=11), a S2 group (n=12), a S3 group (n=10), and a S4 group (n=9) according to their pathological stage. The value of ADC and the expression of TIMP-1 mRNA among the different stage groups of liver fibrosis were compared, and the correlation between ADC and the TIMP-1 mRNA were analyzed.
 Results: The ADC value and the TIMP-1 mRNA expression were significantly different between the control group and the liver fibrosis group (F=46.54 and 53.87, P<0.05). There were significant differences in the value of ADC between every two groups (all P<0.05), except the control group vs the S1 group, the S1 group vs the S2 group, and the S2 group vs the S3 group (all P>0.05). For the comparison of TIMP-1 mRNA, there was no significant difference between the S1 group and the S2 group, the S3 group and the S4 group (both P>0.05). There were significant differences among the rest of the groups (all P<0.05). Rank correlation analysis showed that there was a negative correlation between the ADC value and the TIMP-1 mRNA expression (r=-0.76, P<0.01).
 Conclusion: When the value of ADC decreases in the progress of rats' liver fibrosis, the mRNA expression of TIMP-1 increases gradually, and there is a negative correlation between them.

[Value of liver perfusion imaging of 256-slice CT in evaluation of the cirrhosis].

OBJECTIVE: To investigate the value of liver perfusion imaging of 256-slice CT in evaluating the compensated and decompensated cirrhosis. 
 METHODS: A total of 20 patients with liver cirrhosis, who were confirmed by liver biopsy, clinical symptoms and imaging, were selected from December 2012 to June 2014. According to the results of liver biopsy and the Child-Pugh classification, the patients were divided into a compensated cirrhosis group (n=8) and a decompensated cirrhosis group (n=12). Eleven cases without liver and spleen diseases were served as a control group. All subjects were under the 256-CT liver perfusion (256-CTP). The data of CTP [hepatic arterial perfusion (HAP), portal venous perfusion (PVP), total liver perfusion (TLP), hepatic perfusion index (HPI)] were obtained according to liver perfusion type, and the data of CTP [liver perfusion (LP), peak enhanced (PE), time to peak (TTP), blood volume (BV)] were obtained according to general perfusion type. Spearman rank correlation was used to analyze the correlation of liver cirrhosis with perfusion parameters. The receiver operating characteristic (ROC) curve was used to predict liver cirrhosis, and the maximized Youden index was served as the optimal cutoff value, then the area under curve, sensitivity and specificity were calculated.
 RESULTS: The PVP, TLP and PE values in the control group, the compensated cirrhosis group and the decompensated cirrhosis group were (76.63±37.26), (38.78±16.13) and (36.14±15.31) mL/(100 mL·min); (98.48±43.58), (55.63±14.47) and (54.41±20.81) mL/(100 mL·min); (55.62±18.25), (44.11±5.79) and (41.08±7.74) HU, respectively, showing a gradual downward trend and a significant difference among the 3 groups (all P <0.05). HPI values were (19.50±6.08)%, (31.81±16.48)% and (34.47±16.04)%; TTP values were (37.32±8.59), (47.06±14.61), (59.86±20.87) s, respectively, showing a gradual upward trend and significant difference among the 3 groups ( all P<0.05). There was no significant difference in the HAP, LP and BV among the 3 groups (all P>0.05). PVP, TLP, PE and LP were negatively correlated with the process of liver cirrhosis (r=-0.592, -0.567, -0.409, -0.569, all P<0.05), but HPI and TTP were positively correlated with the process of liver cirrhosis (r=0.434 and 0.538, both P<0.05). 
 CONCLUSION: 256-CTP could provide useful information for the assessment of liver cirrhosis by measuring a plurality of perfusion parameters. The hepatic microvascular changes in patients with liver cirrhosis could be quantitatively assessed by perfusion CT. TTP shows high efficiency in prediction of liver cirrhosis and decompensated liver cirrhosis.

[Feasibility study on liver CT contrast-enhanced scan with low dose of radiation and contrast agent].

OBJECTIVE: To explore the feasibility for liver contrast-enhanced CT scan with low dose of radiation and contrast agent in clinical.
 METHODS: A total of 180 cases were randomly divided into group I (low concentration of contrast agent, 270 mgI/mL of iodixanol) and group II (high concentration of contrast agent, 320 mgI/mL of iodixanol). Three scan conditions (A: 120 kV, 300 mA; B: 100 kV, 400 mA; and C: 100 kV, 300 mA) were randomly distributed in 3 phases (arterial phase, venous phase and delay phase) for liver scans in each group. The effective radiation dose (ED), image CT values and quality of images (image of noise (NI), the image signal to noise ratio (SNR), contrast to noise ratio (CNR), and overall image quality (OIQ) scores were recorded and analyzed. 
 RESULTS: ED values for the group C in the total samples were decreased by 38%, 40% and 41%, respectively compared to the group A in contrast-enhanced scan for 3 phases. The image quality was significantly different (P<0.05); ED values in the group B were decreased by 18%, 20% and 20%, respectively compared to group A in contrast-enhanced scan for 3 phases, and there were no significant differences (P>0.05) in image quality. There were significant differences between the group I and the group II in CT values at the same scanning parameters and scanning phases (P<0.05), but the image quality was not obviously different; there was no significant difference between the group A-II and the group B-I as well as the group C-I in image CT values, and there was also no significant difference between the group A-II and the group B-I in image quality (P>0.05); however the differences in image quality were statistically significant between the group A-II and the group C-I (P<0.05).
 CONCLUSION: Reduction of the tube voltage (to improve the tube current) combined with the low-dose contrast agent can not only reduce the radiation dose and contrast agent dose but also meet the needs of double-low liver contrast-enhanced CT scan.

MR investigation in evaluation of chronic whiplash alar ligament injury in elderly patients.

OBJECTIVE: To observe the imaging features for chronic whiplash alar ligament injury in elderly patients and to provide an effective diagnostic method for long-term neck pain and headaches due to alar ligament injury in elderly patients. METHODS: A total of 134 elderly patients, who engaged in the work or activities related to whiplash motion and suffered from chronic neck pain, were enrolled for the study. All patients were performed comprehensive health examination (CT, MR, ultrasound and laboratory examination) and high resolution PDWI. The patients were divided into 2 groups according to the results of comprehensive health examination: a clear etiology group(CE group, n=96) and an unknown etiology group(UE group, n=38). Th e characteristics of PDWI signal in the ligament were analyzed between the 2 groups. RESULTS: Th e anatomy and signal characteristics of the alar ligament were clearly displayed by high resolution PDWI. Th e alar ligaments were effectively displayed by oblique coronal image. In the CE group, 7 patients (7/96) showed the positive sign of ligament injured, while 21 (21/38) patients showed positive sign of ligament injured in the UE group (P<0.01). Chronic whiplash ligament injury was proved to be the reason for long-term neck pain and headaches in 15.7% patients. CONCLUSION: Th e whiplash injury of alar ligament is an important reason for chronic neck pain in elderly patients. High resolution PDWI is an effective method to evaluate the image features of alar ligament and can provide an accurate diagnosis for chronic neck pain and headaches caused by the alar ligament whiplash injury.

Conditions and mechanism for the appearance of interlobar fissures as 5-line signs in axial maximum intensity projection images.

PURPOSE: To assess the conditions and mechanism for the 5-line sign of normal interlobar fissures revealed on axial maximum intensity projection (MIP). SUBJECTS AND METHODS: Fifty subjects (32 males and 18 females aged 16 to 76 years) whose conditions were diagnosed as normal on computed tomography (CT) of the chest were presented. Images for the fissures with a slice thickness of 1.25 mm, a space thickness of 1.25 mm, and an MIP slab thickness of 6.25 mm were reconstructed. RESULTS: A 5-line shadow of normal interlobar fissures was shown on axial MIP imaging using a slab thickness of 6.25 mm. On axial MIP imaging, 93% of interlobar fissure parts were manifested as clear or barely clear. CONCLUSIONS: On axial MIP imaging using a slab thickness of 6.25 mm, normal interlobar fissures were displayed as 5-line signs, which, similar to the beaded small lung vessels shown on MIP imaging, might be a partial volume-averaging phenomenon.

Cross-domain attention-guided generative data augmentation for medical image analysis with limited data.

Data augmentation is widely applied to medical image analysis tasks in limited datasets with imbalanced classes and insufficient annotations. However, traditional augmentation techniques cannot supply extra information, making the performance of diagnosis unsatisfactory. GAN-based generative methods have thus been proposed to obtain additional useful information to realize more effective data augmentation; but existing generative data augmentation techniques mainly encounter two problems: (i) Current generative data augmentation lacks of the capability in using cross-domain differential information to extend limited datasets. (ii) The existing generative methods cannot provide effective supervised information in medical image segmentation tasks. To solve these problems, we propose an attention-guided cross-domain tumor image generation model (CDA-GAN) with an information enhancement strategy. The CDA-GAN can generate diverse samples to expand the scale of datasets, improving the performance of medical image diagnosis and treatment tasks. In particular, we incorporate channel attention into a CycleGAN-based cross-domain generation network that captures inter-domain information and generates positive or negative samples of brain tumors. In addition, we propose a semi-supervised spatial attention strategy to guide spatial information of features at the pixel level in tumor generation. Furthermore, we add spectral normalization to prevent the discriminator from mode collapse and stabilize the training procedure. Finally, to resolve an inapplicability problem in the segmentation task, we further propose an application strategy of using this data augmentation model to achieve more accurate medical image segmentation with limited data. Experimental studies on two public brain tumor datasets (BraTS and TCIA) show that the proposed CDA-GAN model greatly outperforms the state-of-the-art generative data augmentation in both practical medical image classification tasks and segmentation tasks; e.g. CDA-GAN is 0.50%, 1.72%, 2.05%, and 0.21% better than the best SOTA baseline in terms of ACC, AUC, Recall, and F1, respectively, in the classification task of BraTS, while its improvements w.r.t. the best SOTA baseline in terms of Dice, Sens, HD95, and mIOU, in the segmentation task of TCIA are 2.50%, 0.90%, 14.96%, and 4.18%, respectively.

A new computed tomography score-based staging for melioidosis pneumonia to predict progression.

Background: Melioidosis pneumonia, caused by the bacterium Burkholderia pseudomallei, is a serious infectious disease prevalent in tropical regions. Chest computed tomography (CT) has emerged as a valuable tool for assessing the severity and progression of lung involvement in melioidosis pneumonia. However, there persists a need for the quantitative assessment of CT characteristics and staging methodologies to precisely anticipate disease progression. This study aimed to quantitatively extract CT features and evaluate a CT score-based staging system in predicting the progression of melioidosis pneumonia. Methods: This study included 97 patients with culture-confirmed melioidosis pneumonia who presented between January 2002 and December 2021. Lung segmentation and annotation of lesions (consolidation, nodules, and cavity) were used for feature extraction. The features, including the involved area, amount, and intensity, were extracted. The CT scores of the lesion features were defined by the feature importance weight and qualitative stage of melioidosis pneumonia. Gaussian process regression (GPR) was used to predict patients with severe or critical melioidosis pneumonia according to CT scores. Results: The melioidosis pneumonia stages included acute stage (0-7 days), subacute stage (8-28 days), and chronic stage (>28 days). In the acute stage, the CT scores of all patients ranged from 2.5 to 6.5. In the subacute stage, the CT scores for the severe and mild patients were 3.0-7.0 and 2.0-5.0, respectively. In the chronic stage, the CT score of the mild patients fluctuated approximately between 2.5 and 3.5 in a linear distribution. Consolidation was the most common type of lung lesion in those with melioidosis pneumonia. Between stages I and II, the percentage of severe scans with nodules dropped from 72.22% to 47.62% (P<0.05), and the percentage of severe scans with cavities significantly increased from 16.67% to 57.14% (P<0.05). The GPR optimization function yielded area under the receiver operating characteristic curves of 0.71 for stage I, 0.92 for stage II, and 0.87 for all stages. Conclusions: In patients with melioidosis pneumonia, it is reasonable to divide the period (the whole progression of melioidosis pneumonia) into three stages to determine the prognosis.

Accurate and intelligent diagnosis of pediatric pneumonia using X-ray images and blood testing data.

Computer-aided diagnosis (CAD) methods such as the X-rays-based method is one of the cheapest and safe alternative options to diagnose the disease compared to other alternatives such as Computed Tomography (CT) scan, and so on. However, according to our experiments on X-ray public datasets and real clinical datasets, we found that there are two challenges in the current classification of pneumonia: existing public datasets have been preprocessed too well, making the accuracy of the results relatively high; existing models have weak ability to extract features from the clinical pneumonia X-ray dataset. To solve the dataset problems, we collected a new dataset of pediatric pneumonia with labels obtained through a comprehensive pathogen-radiology-clinical diagnostic screening. Then, to accurately capture the important features in imbalanced data, based on the new dataset, we proposed for the first time a two-stage training multimodal pneumonia classification method combining X-ray images and blood testing data, which improves the image feature extraction ability through a global-local attention module and mitigate the influence of class imbalance data on the results through the two-stage training strategy. In experiments, the performance of our proposed model is the best on new clinical data and outperforms the diagnostic accuracy of four experienced radiologists. Through further research on the performance of various blood testing indicators in the model, we analyzed the conclusions that are helpful for radiologists to diagnose.

Painless and accurate medical image analysis using deep reinforcement learning with task-oriented homogenized automatic pre-processing.

Pre-processing is widely applied in medical image analysis to remove the interference information. However, the existing pre-processing solutions mainly encounter two problems: (i) it is heavily relied on the assistance of clinical experts, making it hard for intelligent CAD systems to deploy quickly; (ii) due to the personnel and information barriers, it is difficult for medical institutions to conduct the same pre-processing operations, making a deep model that performs well on a specific medical institution difficult to achieve similar performances on the same task in other medical institutions. To overcome these problems, we propose a deep-reinforcement-learning-based task-oriented homogenized automatic pre-processing (DRL-HAPre) framework to overcome these two problems. This framework utilizes deep reinforcement learning techniques to learn a policy network to automatically and adaptively select the optimal pre-processing operations for the input medical images according to different analysis tasks, thus helping the intelligent CAD system to achieve a rapid deployment (i.e., painless) and maintain a satisfactory performance (i.e., accurate) among different medical institutes. To verify the effectiveness and advantages of the proposed DRL-HAPre framework, we further develop a homogenized automatic pre-processing model based on the DRL-HAPre framework to realize the automatic pre-processing of key region selection (called HAPre-KRS) in the pneumonia image classification task. Extensive experimental studies are conducted on three pediatric pneumonia classification datasets with different image qualities, and the results show that: (i) There does exist a hard-to-reproduce problem in clinical practices and the fact that having different medical image qualities in different medical institutes is an important reason for the existing of hard-to-reproduce problem, so it is compelling to propose homogenized automatic pre-processing method. (ii) The proposed HAPre-KRS model and DRL-HAPre framework greatly outperform three kinds of state-of-the-art baselines (i.e., pre-processing, attention and pneumonia baseline), and the lower the medical image quality, the greater the improvements of using our HAPre-KRS model and DRL-HAPre framework. (iii) With the help of homogenized pre-processing, HAPre-KRS (and DRL-HAPre framework) can greatly avoid performance degradation in real-world cross-source applications (i.e., thus overcoming the hard-to-reproduce problem).

EFPN: Effective medical image detection using feature pyramid fusion enhancement.

Feature pyramid networks (FPNs) are widely used in the existing deep detection models to help them utilize multi-scale features. However, there exist two multi-scale feature fusion problems for the FPN-based deep detection models in medical image detection tasks: insufficient multi-scale feature fusion and the same importance for multi-scale features. Therefore, in this work, we propose a new enhanced backbone model, EFPNs, to overcome these problems and help the existing FPN-based detection models to achieve much better medical image detection performances. We first introduce an additional top-down pyramid to help the detection networks fuse deeper multi-scale information; then, a scale enhancement module is developed to use different sizes of kernels to generate more diverse multi-scale features. Finally, we propose a feature fusion attention module to estimate and assign different importance weights to features with different depths and scales. Extensive experiments are conducted on two public lesion detection datasets for different medical image modalities (X-ray and MRI). On the mAP and mR evaluation metrics, EFPN-based Faster R-CNNs improved 1.55% and 4.3% on the PenD (X-ray) dataset, and 2.74% and 3.1% on the BraTs (MRI) dataset, respectively. EFPN-based Faster R-CNNs achieve much better performances than the state-of-the-art baselines in medical image detection tasks. The proposed three improvements are all essential and effective for EFPNs to achieve superior performances; and besides Faster R-CNNs, EFPNs can be easily applied to other deep models to significantly enhance their performances in medical image detection tasks.

µ-Net: Medical image segmentation using efficient and effective deep supervision.

Although the existing deep supervised solutions have achieved some great successes in medical image segmentation, they have the following shortcomings; (i) semantic difference problem: since they are obtained by very different convolution or deconvolution processes, the intermediate masks and predictions in deep supervised baselines usually contain semantics with different depth, which thus hinders the models' learning capabilities; (ii) low learning efficiency problem: additional supervision signals will inevitably make the training of the models more time-consuming. Therefore, in this work, we first propose two deep supervised learning strategies, U-Net-Deep and U-Net-Auto, to overcome the semantic difference problem. Then, to resolve the low learning efficiency problem, upon the above two strategies, we further propose a new deep supervised segmentation model, called µ-Net, to achieve not only effective but also efficient deep supervised medical image segmentation by introducing a tied-weight decoder to generate pseudo-labels with more diverse information and also speed up the convergence in training. Finally, three different types of µ-Net-based deep supervision strategies are explored and a Similarity Principle of Deep Supervision is further derived to guide future research in deep supervised learning. Experimental studies on four public benchmark datasets show that µ-Net greatly outperforms all the state-of-the-art baselines, including the state-of-the-art deeply supervised segmentation models, in terms of both effectiveness and efficiency. Ablation studies sufficiently prove the soundness of the proposed Similarity Principle of Deep Supervision, the necessity and effectiveness of the tied-weight decoder, and using both the segmentation and reconstruction pseudo-labels for deep supervised learning.

Aquaporin 1 overexpression may enhance glioma tumorigenesis by interacting with the transcriptional regulation networks of Foxo4, Maz, and E2F families.

BACKGROUND: The glioblastoma has served as a valuable experimental model system for investigating the growth and invasive properties of glioblastoma. Aquaporin-1 (AQP1) in facilitating cell migration and potentially contributing to tumor progression. In this study, we analyzed the role of AQP1 overexpression in glioblastoma and elucidated the main mechanisms involved. METHODS: AQP1 overexpression recombinant vector was introduced into C6 rat glioma cells to construct an AQP1 overexpression C6 cell line, and its effect on cell viability and migration ability was detected by MTT and Transwell. RNA was extracted by Trizol method for gene sequencing and transcriptomics analysis, and the differentially expressed genes (DEGs) were enriched for up- and downregulated genes by Principal component analysis (PCA), and the molecular mechanism of AQP1 overexpression was analyzed in comparison with the control group using the NCBI GEO database. Statistical analysis was performed using Mann-Whitney paired two tailed t test. RESULTS: The cell viability of AQP1-transfected cell lines increased by 23% and the mean distance traveled increased by 67% compared with the control group. Quantitative analysis of gene expression showed that there were 12,121 genes with an average transcripts per million (TPM) value greater than 1. DEGs accounted for 13% of the genes expressed, with the highest correlation with upregulated genes being FOXO4 and MAZ, and the highest with downregulated genes being E2F TFs. CONCLUSIONS: AQP1 may be implicated in glioma formation by interacting with the transcriptional regulation networks involving the FOXO4, MAZ, and E2F1/2. These findings shed light on the potential significance of AQP1 in glioma pathogenesis and warrant further investigations to unravel the underlying molecular mechanisms.

PAC-Net: Multi-pathway FPN with position attention guided connections and vertex distance IoU for 3D medical image detection.

Automatic medical image detection aims to utilize artificial intelligence techniques to detect lesions in medical images accurately and efficiently, which is one of the most important tasks in computer-aided diagnosis (CAD) systems, and can be embedded into portable imaging devices for intelligent Point of Care (PoC) Diagnostics. The Feature Pyramid Networks (FPN) based models are widely used deep-learning-based solutions for automatic medical image detection. However, FPN-based medical lesion detection models have two shortcomings: the object position offset problem and the degradation problem of IoU-based loss. Therefore, in this work, we propose a novel FPN-based backbone model, i.e., Multi-Pathway Feature Pyramid Networks with Position Attention Guided Connections and Vertex Distance IoU (abbreviated as PAC-Net), to replace vanilla FPN for more accurate lesion detection, where two innovative improvements, a position attention guided connection (PAC) module and Vertex Distance IoU Vertex Distance Intersection over Union loss, are proposed to address the above-mentioned shortcomings of vanilla FPN, respectively. Extensive experiments are conducted on a public medical image detection dataset, i.e., Deeplesion, and the results showed that i) PAC-Net outperforms all state-of-the-art FPN-based depth models in both evaluation metrics of lesion detection on the DeepLesion dataset, ii) the proposed PAC module and VDIoU loss are both effective and important for PAC-Net to achieve a superior performance in automatic medical image detection tasks, and iii) the proposed VDIoU loss converges more quickly than the existing IoU-based losses, making PAC-Net an accurate and also highly efficient 3D medical image detection model.

Artificial intelligence-assisted multistrategy image enhancement of chest X-rays for COVID-19 classification.

Background: The coronavirus disease 2019 (COVID-19) led to a dramatic increase in the number of cases of patients with pneumonia worldwide. In this study, we aimed to develop an AI-assisted multistrategy image enhancement technique for chest X-ray (CXR) images to improve the accuracy of COVID-19 classification. Methods: Our new classification strategy consisted of 3 parts. First, the improved U-Net model with a variational encoder segmented the lung region in the CXR images processed by histogram equalization. Second, the residual net (ResNet) model with multidilated-rate convolution layers was used to suppress the bone signals in the 217 lung-only CXR images. A total of 80% of the available data were allocated for training and validation. The other 20% of the remaining data were used for testing. The enhanced CXR images containing only soft tissue information were obtained. Third, the neural network model with a residual cascade was used for the super-resolution reconstruction of low-resolution bone-suppressed CXR images. The training and testing data consisted of 1,200 and 100 CXR images, respectively. To evaluate the new strategy, improved visual geometry group (VGG)-16 and ResNet-18 models were used for the COVID-19 classification task of 2,767 CXR images. The accuracy of the multistrategy enhanced CXR images was verified through comparative experiments with various enhancement images. In terms of quantitative verification, 8-fold cross-validation was performed on the bone suppression model. In terms of evaluating the COVID-19 classification, the CXR images obtained by the improved method were used to train 2 classification models. Results: Compared with other methods, the CXR images obtained based on the proposed model had better performance in the metrics of peak signal-to-noise ratio and root mean square error. The super-resolution CXR images of bone suppression obtained based on the neural network model were also anatomically close to the real CXR images. Compared with the initial CXR images, the classification accuracy rates of the internal and external testing data on the VGG-16 model increased by 5.09% and 12.81%, respectively, while the values increased by 3.51% and 18.20%, respectively, for the ResNet-18 model. The numerical results were better than those of the single-enhancement, double-enhancement, and no-enhancement CXR images. Conclusions: The multistrategy enhanced CXR images can help to classify COVID-19 more accurately than the other existing methods.

[Evaluation of sinus meningiomas in the region of interest by high-resolution three dimensional contrast-enhanced magnetic resonance venography].

OBJECTIVE: To determine the value of high-resolution three dimensional contrast enhanced magnetic resonance venography (3D CE-MRV) in evaluating sinus meningiomas in the region of interest (ROI). METHODS: Twenty patients with sinus meningiomas underwent 3D CE-MRV with ROI preoperatively (including 9 patients postoperatively). We observed the changes of venous sinus adjacent the tumor. RESULTS: All patients received high-resolution image, the single acquisition time was about 11.4 s, and the voxel value was about 1.3 mm3. The images of 20 patients showed the change of the sinus clearly, 6 of which with integral sinus, 14 with sinus invaded at various degrees, including 5 with sinus occlusion and 9 with stenosis. CE-MRV also showed 4 patients with clear sinuses, 1 with narrow sinus, partial interruption, and 4 with sinus removed after the surgery. CONCLUSION: Application of high-resolution 3D CE-MRV in ROI in sinus meningiomas may help obtain a series of high-resolution images in a short time, show the relationship between the tumor and venous sinus, display the degree of invasion of venous sinus clearly, provide information for the surgical treatment, and evaluate the change of sinus after the surgery.

Whole-Body MRI vs. PET/CT for the Detection of Bone Metastases in Patients With Prostate Cancer: A Systematic Review and Meta-Analysis.

Purpose: A recent meta-analysis in patients with non-small cell lung cancer showed no difference between whole-body magnetic resonance imaging (WBMRI) and positron emission tomography/computed tomography (PET/CT), but no such study is available for prostate cancer (PCa). This study aimed to compare WBMRI and PET/CT for bone metastasis detection in patients with PCa. Materials and Methods: PubMed, Embase, and the Cochrane library were searched for papers published up to April 2020. The population was the patients with untreated prostate cancer diagnosed by WBMRI or PET/CT. The outcomes were the true positive and negative and false positive and negative rates for WBMRI and PET/CT. The summarized sensitivity, specificity, positive likelihood ratios (PLR), negative likelihood ratios (NLR), and diagnostic odds ratios (DOR) were calculated with their 95% confidence intervals (CIs). Results: Four prospective and one retrospective study are included (657 patients). Significant differences are observed between WBMRI and PET/CT for sensitivity (WBMRI/PET/CT: 0.896; 95% CI: 0.813-0.987; P = 0.025) and NLR (WBMRI/PET/CT: 2.38; 95% CI: 1.13-5.01; P = 0.023), but not for specificity (WBMRI/PET/CT: 0.939; 95% CI: 0.855-1.031; P = 0.184) and PLR (WBMRI/PET/CT: 0.42; 95% CI: 0.08-2.22; P = 0.305). WBMRI has a similar a DOR compared with PET/CT (WBMRI/PET/CT: 0.13; 95% CI: 0.02-1.11; P = 0.062). The summary area under the receiver operating characteristic curves for WBMRI is 0.88 (standard error: 0.032) and 0.98 (standard error: 0.013) for PET/CT for diagnosing bone metastases in PCa. Conclusion: PET/CT presents a higher sensitivity and NLR for the bone metastasis detection from PCa, whereas no differences are found for specificity and PLR, compared with WBMRI.

Diagnostic Classification of Patients with Dilated Cardiomyopathy Using Ventricular Strain Analysis Algorithm.

Dilated cardiomyopathy (DCM) is a cardiomyopathy with left ventricle or double ventricle enlargement and systolic dysfunction. It is an important cause of sudden cardiac death and heart failure and is the leading indication for cardiac transplantation. Major heart diseases like heart muscle damage and valvular problems are diagnosed using cardiac MRI. However, it takes time for cardiologists to measure DCM-related parameters to decide whether patients have this disease. We have presented a method for automatic ventricular segmentation, parameter extraction, and diagnosing DCM. In this paper, left ventricle and right ventricle are segmented by parasternal short-axis cardiac MR image sequence; then, related parameters are extracted in the end-diastole and end-systole of the heart. Machine learning classifiers use extracted parameters as input to predict normal people and patients with DCM, among which Random forest classifier gives the highest accuracy. The results show that the proposed system can be effectively utilized to detect and diagnose DCM automatically. The experimental results suggest the capabilities and advantages of the proposed method to diagnose DCM. A small amount of sample input can generate results comparable to more complex methods.

AQP4-siRNA alleviates traumatic brain edema by altering post-traumatic AQP4 polarity reversal in TBI rats.

The spatial and temporal distribution of aquaporin-4 (AQP4) expression in rat brain following brain trauma and AQP4-siRNA treatment, as well as corresponding pathological changes, were studied to explore the mechanism underlying the effect of AQP4-siRNA treatment on traumatic brain injury (TBI). The rats in the sham operation group had normal structure, with AQP4 located in the perivascular end-foot membranes and astrocytic membranes in a polarized pattern. The accelerated polarity reversal was observed in the TBI group in 1-12 h after TBI. During this period, AQP4 abundance on the astrocytic membrane is gradually increased, while AQP4 abundance on the perivascular end-foot membrane declined rapidly. Twelve hours after TBI, AQP4 expression was depolarized, showing a shift from the perivascular end-foot membrane to the astrocytic membrane. Pathological observation showed that vasogenic edema occurred immediately after TBI, at which time the extracellular space was expanded, leading to severe intracellular edema. AQP4-siRNA reduced the polarity reversal index at the early stage of TBI recovery and reduced edema, demonstrating the potential benefit of reduced AQP4 expression during recovery from TBI.