Transparent and Accurate COVID-19 Diagnosis:Integrating Explainable AI with Advanced Deep Learning in CT Imaging

In the current landscape of the COVID-19 pandemic,the utilization of deep learning in medical imaging,especially in chest computed tomography(CT)scan analysis for virus detection,has become increasingly significant.Despite its potential,deep learning’s“black box”nature has been a major impediment to its broader acceptance in clinical environments,where transparency in decision-making is imperative.To bridge this gap,our research integrates Explainable AI(XAI)techniques,specifically the Local Interpretable Model-Agnostic Explanations(LIME)method,with advanced deep learning models.This integration forms a sophisticated and transparent framework for COVID-19 identification,enhancing the capability of standard Convolutional Neural Network(CNN)models through transfer learning and data augmentation.Our approach leverages the refined DenseNet201 architecture for superior feature extraction and employs data augmentation strategies to foster robust model generalization.The pivotal element of our methodology is the use of LIME,which demystifies the AI decision-making process,providing clinicians with clear,interpretable insights into the AI’s reasoning.This unique combination of an optimized Deep Neural Network(DNN)with LIME not only elevates the precision in detecting COVID-19 cases but also equips healthcare professionals with a deeper understanding of the diagnostic process.Our method,validated on the SARS-COV-2 CT-Scan dataset,demonstrates exceptional diagnostic accuracy,with performance metrics that reinforce its potential for seamless integration into modern healthcare systems.This innovative approach marks a significant advancement in creating explainable and trustworthy AI tools for medical decisionmaking in the ongoing battle against COVID-19.

Application of ^(18) F-FDG PET/CT Imaging in Diagnosing Bladder Tumor Metastasis Lesions

Bladder tumor is the most common malignant tumor in urinary system and always com- panied with lymph node metastasis. The accurate staging plays a significant role in treatment for bladder tumor and prognostic evaluation, and the distant metastasis predicts worse prognosis. The objective of this study was to assess the clinical significance of 18F-FDG PET/CT imaging in diagnosing bladder tumor metastasis lesions. A retrospective analysis of 60 patients with bladder tumor from October 2008 to May 2010 was done. The patients were stratified based on the imaging technique. Among all 60 cases, besides the primary lesion, 81 suspected lesions were spotted and 73 confirmed as metastasis, including 50 lymph node metastases, 22 distant metastases, and 1 bone metastasis. For PET/CT imaging, its sensitivity was 94.5%, specificity 87.5%, positive predictive value 98.6%, negative predictive value 63.6% and accuracy 93.8% respectively. For CT, its sensitivity was 82.2%, specificity 50%, positive predictive value 93.8%, negative predictive value 23.5% and accuracy 79% respectively. PET/CT im- aging was superior to CT in sensitivity, specificity and accuracy. In conclusion, 18F-FDG PET/CT imaging is more significant in diagnosing bladder tumor metastasis lesions.

Use of various CT imaging methods for diagnosis of acute ischemic cerebrovascular disease

Thirty-four patients with cerebral infarction and 18 patients with transient ischemic attack were examined by multi-slice spiral CT scan, CT perfusion imaging, and CT angiography within 6 hours after onset. By CT perfusion imaging, 29 cases in the cerebral infarction group and 10 cases in the transient ischemic attack group presented with abnormal blood flow perfusion, which corresponded to the clinical symptoms. By CT angiography, various degrees of vascular stenosis could be detected in 41 patients, including 33 in the cerebral infarction group and eight in the transient ischemic attack group. The incidence of intracranial artery stenosis was higher than that of extracranial artery stenosis. The intracranial artery stenosis was located predominantly in the middle cerebral artery and carotid artery siphon, while the extracranial artery stenosis occurred mainly in the bifurcation of the common carotid artery and the opening of the vertebral artery. There were 34 cases （83%） with convict vascular stenosis and perfusion abnormalities, and five cases （45%） with perfusion abnormalities but without convict vascular stenosis. The incidence of cerebral infarction in patients with National Institutes of Health Stroke Scale scores 〉 5 points during onset was significantly higher than that in patients with National Institutes of Health Stroke Scale scores 〈 5 points. These experimental findings indicate that the combined application of various CT imaging methods allows early diagnosis of acute ischemic cerebrovascular disease, which can comprehensively analyze the pathogenesis and severity of acute ischemic cerebrovascular disease at the morphological and functional levels.

The diagnostic value of 3D spiral CT imaging of cholangiopancreatic ducts on obstructive jaundice

Objective: Computerized tomography (CT) plays an important role in the diagnosis of diseases of biliary tract. Recently, three dimensions (3D) spiral CT imaging has been used in surgical diseases gradually. This study was designed to evaluate the diagnostic value of 3D spiral CT imaging of cholangiopancreatic ducts on obstructive jaundice. Methods: Thirty patients with obstructive jaundice had received B-mode ultrasonography, CT, percutaneous transhepatic cholangiography (PTC) or endoscopic retrograde cholangiopancreatography (ERCP), and 3D spiral CT imaging of cholangiopancreatic ducts preoperatively. Then the diagnose accordance rate of these examinational methods were compared after operations. Results: The diagnose accordance rate of 3D spiral CT imaging of cholangiopancreatic ducts was higher than those of B-mode ultrasonography, CT, or single PTC or ERCP, which showed clear images of bile duct tree and pathological changes. As to malignant obstructive jaundice, this examinational technique could clearly display the adjacent relationship between tumor and liver tissue, biliary ducts, blood vessels, and intrahepatic metastases. Conclusion: 3D spiral CT imaging of cholangiopancreatic ducts has significant value for obstructive diseases of biliary ducts, which provides effective evidence for the feasibility of tumor-resection and surgical options.

Segmentation of Head and Neck Tumors Using Dual PET/CT Imaging:Comparative Analysis of 2D,2.5D,and 3D Approaches Using UNet Transformer

The segmentation of head and neck(H&N)tumors in dual Positron Emission Tomography/Computed Tomogra-phy(PET/CT)imaging is a critical task in medical imaging,providing essential information for diagnosis,treatment planning,and outcome prediction.Motivated by the need for more accurate and robust segmentation methods,this study addresses key research gaps in the application of deep learning techniques to multimodal medical images.Specifically,it investigates the limitations of existing 2D and 3D models in capturing complex tumor structures and proposes an innovative 2.5D UNet Transformer model as a solution.The primary research questions guiding this study are:(1)How can the integration of convolutional neural networks(CNNs)and transformer networks enhance segmentation accuracy in dual PET/CT imaging?(2)What are the comparative advantages of 2D,2.5D,and 3D model configurations in this context?To answer these questions,we aimed to develop and evaluate advanced deep-learning models that leverage the strengths of both CNNs and transformers.Our proposed methodology involved a comprehensive preprocessing pipeline,including normalization,contrast enhancement,and resampling,followed by segmentation using 2D,2.5D,and 3D UNet Transformer models.The models were trained and tested on three diverse datasets:HeckTor2022,AutoPET2023,and SegRap2023.Performance was assessed using metrics such as Dice Similarity Coefficient,Jaccard Index,Average Surface Distance(ASD),and Relative Absolute Volume Difference(RAVD).The findings demonstrate that the 2.5D UNet Transformer model consistently outperformed the 2D and 3D models across most metrics,achieving the highest Dice and Jaccard values,indicating superior segmentation accuracy.For instance,on the HeckTor2022 dataset,the 2.5D model achieved a Dice score of 81.777 and a Jaccard index of 0.705,surpassing other model configurations.The 3D model showed strong boundary delineation performance but exhibited variability across datasets,while the 2D model,although effective,generally underperformed compared to its 2.5D and 3D counterparts.Compared to related literature,our study confirms the advantages of incorporating additional spatial context,as seen in the improved performance of the 2.5D model.This research fills a significant gap by providing a detailed comparative analysis of different model dimensions and their impact on H&N segmentation accuracy in dual PET/CT imaging.

Co-delivery of retinoic acid and miRNA by functional Au nanoparticles for improved survival and CT imaging tracking of MSCs in pulmonary fibrosis therapy

Mesenchymal stem cells(MSCs)have emerged as promising candidates for idiopathic pulmonary fibrosis(IPF)therapy.Increasing the MSC survival rate and deepening the understanding of the behavior of transplanted MSCs are of great significance for improving the efficacy of MSC-based IPF treatment.Therefore,dual-functional Au-based nanoparticles(Au@PEG@PEI@TAT NPs,AuPPT)were fabricated by sequential modification of cationic polymer polyetherimide(PEI),polyethylene glycol(PEG),and transactivator of transcription(TAT)penetration peptide on AuNPs,to co-deliver retinoic acid(RA)and microRNA(miRNA)for simultaneously enhancing MSC survive and real-time imaging tracking of MSCs during IPF treatment.AuPPT NPs,with good drug loading and cellular uptake abilities,could efficiently deliver miRNA and RA to protect MSCs from reactive oxygen species and reduce their expression of apoptosis executive protein Caspase 3,thus prolonging the survival time of MSC after transplantation.In themeantime,the intracellular accumulation of AuPPT NPs enhanced the computed tomography imaging contrast of transplantedMSCs,allowing them to be visually tracked in vivo.This study establishes an Au-based dual-functional platform for drug delivery and cell imaging tracking,which provides a new strategy for MSC-related IPF therapy.

Dual-modal photoacoustic/CT imaging system

Photoacoustic(PA) tomography(PAT) breaks the barrier for high-resolution optical imaging in a strong lightscattering medium, having a great potential for both clinical implementation and small animal studies. However,many organs and tissues lack enough PA contrast or even hinder the propagation of PA waves. Therefore, it is challenging to interpret pure PAT images, especially three-dimensional(3 D) PA images for deep tissues, without enough structural information. To overcome this limitation, in this study, we integrated PAT with X-ray computed tomography(CT) in a standalone system. PAT provides optical contrast and CT gives anatomical information. We performed agar, tissue phantom, and animal studies, and the results demonstrated that PAT/CT imaging systems can provide accurate spatial registration of important complementary contrasts.

MCIF-Transformer Mask RCNN:Multi-Branch Cross-Scale Interactive Feature Fusion Transformer Model for PET/CT Lung Tumor Instance Segmentation

The precise detection and segmentation of tumor lesions are very important for lung cancer computer-aided diagnosis.However,in PET/CT(Positron Emission Tomography/Computed Tomography)lung images,the lesion shapes are complex,the edges are blurred,and the sample numbers are unbalanced.To solve these problems,this paper proposes a Multi-branch Cross-scale Interactive Feature fusion Transformer model(MCIF-Transformer Mask RCNN)for PET/CT lung tumor instance segmentation,The main innovative works of this paper are as follows:Firstly,the ResNet-Transformer backbone network is used to extract global feature and local feature in lung images.The pixel dependence relationship is established in local and non-local fields to improve the model perception ability.Secondly,the Cross-scale Interactive Feature Enhancement auxiliary network is designed to provide the shallow features to the deep features,and the cross-scale interactive feature enhancement module(CIFEM)is used to enhance the attention ability of the fine-grained features.Thirdly,the Cross-scale Interactive Feature fusion FPN network(CIF-FPN)is constructed to realize bidirectional interactive fusion between deep features and shallow features,and the low-level features are enhanced in deep semantic features.Finally,4 ablation experiments,3 comparison experiments of detection,3 comparison experiments of segmentation and 6 comparison experiments with two-stage and single-stage instance segmentation networks are done on PET/CT lung medical image datasets.The results showed that APdet,APseg,ARdet and ARseg indexes are improved by 5.5%,5.15%,3.11%and 6.79%compared with Mask RCNN(resnet50).Based on the above research,the precise detection and segmentation of the lesion region are realized in this paper.This method has positive significance for the detection of lung tumors.

Analysis of Clinical and Imaging Characteristics of Suspected and Confirmed COVID-19 Patients

Objective:To analyze the clinical characteristics and chest CT imaging characteristics of patients with confirmed COVID-19(COVID-19)and patients with suspected COVID-19.Methods:The study time span was from February 2020 to May 2020.The case samples were selected from 72 patients with confirmed covid-19 and suspected covid-19 diagnosed and treated by The First People’s Hospital of Yinchuan and Yinchuan Temporary Emergency Hospital,including 38 patients with confirmed covid-19 and 34 patients with suspected covid-19.All patients underwent laboratory examination and chest CT examination,and the specific examination results were compared and analyzed.Results:There were significant differences in number of white blood cell,percentage of lymphocytes,creatine kinase and erythrocyte sedimentation rate between confirmed and suspected COVID-19 patients(P<0.05).The CT imaging characteristics of COVID-19 patients were compared with those of suspected COVID-19 patients.The lesions of COVID-19 patients were mostly characterized by mixed ground glass density and pure ground glass density.There were vascular thickening and interstitial thickness increase,and accompanied by bronchiectasis or air bronchogram.The distribution of lesions was mostly subpleural without pleural effusion.The lesion area of suspected COVID-19 patients mostly showed solid density and mixed ground glass density.The lesion was distributed along bronchovascular and pleural effusion was observed.Conclusion:There are some differences in biochemical indexes and chest CT images between confirmed and suspected covid-19 patients,which can be used for differential diagnosis.

Research Progress on the Application of Three-dimensional CT Imaging Technology in Oral and Maxillofacial Region and Radiation Protection

The anatomic relationship of oral and maxillofacial region is very com-plex,due to the large number of sinuses,cavities and spaces,and also closely related to the brain.The diagnosis of oral and maxillofacial lesions usually depends on the imaging examination.The conventional imaging methods are common CT and X-ray plain films.In recent years,with the rapid development of medical science and technology,more intuitive and vivid three-dimensional images have been applied in the diagnosis and treatment of oral and maxillofacial diseases.Therefore,CT three-dimensional imaging technology has been widely used in clinical practice.This paper reviews this topic.

Predicting 3D Radiotherapy Dose-Volume Based on Deep Learning

Cancer is one of the most dangerous diseaseswith highmortality.One of the principal treatments is radiotherapy by using radiation beams to destroy cancer cells and this workflow requires a lot of experience and skill from doctors and technicians.In our study,we focused on the 3D dose prediction problem in radiotherapy by applying the deeplearning approach to computed tomography(CT)images of cancer patients.Medical image data has more complex characteristics than normal image data,and this research aims to explore the effectiveness of data preprocessing and augmentation in the context of the 3D dose prediction problem.We proposed four strategies to clarify our hypothesis in different aspects of applying data preprocessing and augmentation.In strategies,we trained our custom convolutional neural network model which has a structure inspired by the U-net,and residual blocks were also applied to the architecture.The output of the network is added with a rectified linear unit(Re-Lu)function for each pixel to ensure there are no negative values,which are absurd with radiation doses.Our experiments were conducted on the dataset of the Open Knowledge-Based Planning Challenge which was collected from head and neck cancer patients treatedwith radiation therapy.The results of four strategies showthat our hypothesis is rational by evaluating metrics in terms of the Dose-score and the Dose-volume histogram score(DVH-score).In the best training cases,the Dose-score is 3.08 and the DVH-score is 1.78.In addition,we also conducted a comparison with the results of another study in the same context of using the loss function.

Sinogram fusion-based metal artifact correction method

To solve the problem that metal artifacts severely damage the clarity of the organization structure in computed tomography（CT） images, a sinogram fusion-based metal artifact correction method is proposed. First, the metal image is segmented from the original CT image by the pre-set threshold. The original CT image and metal image are forward projected into the original projection sinogram and metal projection sinogram, respectively. The interpolation-based correction method and mean filter are used to correct the original CT image and preserve the edge of the corrected CT image, respectively. The filtered CT image is forward projected into the filtered image sinogram. According to the position of the metal sinogram in the original sinogram and filtered image sinogram, the corresponding sinograms PM^D （ in the original sinogram） and PM^C （ in the filtered image sinogram）can be acquired from the original sinogram and filtered image sinogram, respectively. Then, PM^D and PM^C are fused into the fused metal sinogram PM^F according to a certain proportion.The final sinogram can be acquired by fusing PM^F , PM^D and the original sinogram P^O. Finally, the final sinogram is reconstructed into the corrected CT image and metal information is compensated into the corrected CT image.Experiments on clinical images demonstrate that the proposed method can effectively reduce metal artifacts. A comparison with classical metal artifacts correction methods shows that the proposed metal artifacts correction method performs better in metal artifacts suppression and tissue feature preservation.

Primary stratification and identification of suspected corona virus disease 2019(COVID-19) from clinical perspective by a simple scoring proposal

In this Commentary,we would like to comment on the article titled"A rapid advice guideline for the diagnosis and treatment of 2019 novel coronavirus(2019-nCoV)infected pneumonia(standard version)"as a featured article in Military Medical Research.In the guideline,except for"confirmed cases","suspected cases","close contact"and"suspicious exposure"were defined by clinical perspective based on epidemiological risk,clinical symptoms and auxiliary examination.Combined with our experience,we introduced a simple scoring proposal additionally based on not only CT imaging as strongly recommended by the guideline but also blood routine test,especially for primary screening of such patients in the out-patient department.

Morphological evolution and flow conduction characteristics of fracture channels in fractured sandstone under cyclic loading and unloading

In coal mining,rock strata are fractured under cyclic loading and unloading to form fracture channels.Fracture channels are the main flow narrows for gas.Therefore,expounding the flow conductivity of fracture channels in rocks on fluids is significant for gas flow in rock strata.In this regard,graded incremental cyclic loading and unloading experiments were conducted on sandstones with different initial stress levels.Then,the three-dimensional models for fracture channels in sandstones were established.Finally,the fracture channel percentages were used to reflect the flow conductivity of fracture channels.The study revealed how the particle size distribution of fractured sandstone affects the formation and expansion of fracture channels.It was found that a smaller proportion of large blocks and a higher proportion of small blocks after sandstone fails contribute more to the formation of fracture channels.The proportion of fracture channels in fractured rock can indicate the flow conductivity of those channels.When the proportion of fracture channels varies gently,fluids flow evenly through those channels.However,if the proportion of fracture channels varies significantly,it can greatly affect the flow rate of fluids.The research results contribute to revealing the morphological evolution and flow conductivity of fracture channels in sandstone and then provide a theoretical basis for clarifying the gas flow pattern in the rock strata of coal mines.

EVALUATION OF SMALL ANIMAL IMAGING UNDER DIFFERENT SCAN CONDITIONS

To evaluae small animal imaging with individual different high voltage, filter thickness and tube current, an animal X-ray micro-computed tomography （micro-CT） system based on panel detector is developed and a rat is scanned by using the system with individual high voltage, tube current, filter thickness, and exposure time. A model is presented based on the Monte Carlo code PENELOPE for generating the X-ray spectra of X-ray tube used in the micro-CT system. A platform developed based on Matlab allows for calculating beam quality parameters, including the average energy of X-ray beam, the change of transmition rate and the input X-ray fluence. The factors affecting the signal difference to noise ratio （SDNR） of micro-CT are investigated and the relationship between SDNR and scan combinations is analyzed. A series of tools and methods are developed for small animal imaging and imaging performance evaluation in the field of small animal imaging.

A Deep Learning Interpretable Model for Novel Coronavirus Disease (COVID-19) Screening with Chest CT Images

In this article, we propose a convolutional neural network (CNN)-based model, a ResNet-50 based model, for discriminating coronavirus disease 2019 (COVID-19) from Non-COVID-19 using chest CT. We adopted the use of wavelet coefficients of the entire image without cropping any parts of the image as input to the CNN model. One of the main contributions of this study is to implement an algorithm called gradient-weighted class activation mapping to produce a heat map for visually verifying where the CNN model is looking at the image, thereby, ensuring the model is performing correctly. In order to verify the effectiveness and usefulness of the proposed method, we compare the obtained results with that obtained by using pixel values of original images as input to the CNN model. The measures used for performance evaluation include accuracy, sensitivity, specificity, positive predictive value, negative predictive value, F1 score, and Matthews correlation coefficient (MCC). The overall classification accuracy, F1 score, and MCC for the proposed method (using wavelet coefficients as input) were 92.2%, 0.915%, and 0.839%, and those for the compared method (using pixel values of the original image as input) were 88.3%, 0.876%, and 0.766%, respectively. The experiment results demonstrate the superiority of the proposed method. Moreover, as a comprehensible classification model, the interpretability of classification results was introduced. The region of interest extracted by the proposed model was visualized using heat maps and the probability score was also shown. We believe that our proposed method could provide a promising computerized toolkit to help radiologists and serve as a second eye for them to classify COVID-19 in CT scan screening examination.

A Robust Zero-Watermarking Based on SIFT-DCT for Medical Images in the Encrypted Domain

Remote medical diagnosis can be realized by using the Internet,but when transmitting medical images of patients through the Internet,personal information of patients may be leaked.Aim at the security of medical information system and the protection of medical images,a novel robust zero-watermarking based on SIFT-DCT(Scale Invariant Feature Transform-Discrete Cosine Transform)for medical images in the encrypted domain is proposed.Firstly,the original medical image is encrypted in transform domain based on Logistic chaotic sequence to enhance the concealment of original medical images.Then,the SIFT-DCT is used to extract the feature sequences of encrypted medical images.Next,zero-watermarking technology is used to ensure that the region of interest of medical images are not changed.Finally,the robust of the algorithm is evaluated by the correlation coefficient between the original watermark and the attacked watermark.A series of attack experiments are carried out on this method,and the results show that the algorithm is not only secure,but also robust to both traditional and geometric attacks,especially in clipping attacks.

Acoustic emission characterization of microcracking in laboratory-scale hydraulic fracturing tests

Understanding microcracking near coalesced fracture generation is critically important for hydrocarbon and geothermal reservoir characterization as well as damage evaluation in civil engineering structures. Dense and sometimes random microcracking near coalesced fracture formation alters the mechanical properties of the nearby virgin material. Individual microcrack characterization is also significant in quantifying the material changes near the fracture faces （i.e. damage）. Acoustic emission （AE） monitoring and analysis provide unique information regarding the microcracking process temporally, and infor- mation concerning the source characterization of individual microcracks can be extracted. In this context, laboratory hydraulic fracture tests were carried out while monitoring the AEs from several piezoelectric transducers. In-depth post-processing of the AE event data was performed for the purpose of under- standing the individual source mechanisms. Several source characterization techniques including moment tensor inversion, event parametric analysis, and volumetric deformation analysis were adopted. Post-test fracture characterization through coring, slicing and micro-computed tomographic imaging was performed to determine the coalesced fracture location and structure. Distinct differences in fracture characteristics were found spatially in relation to the openhole injection interval. Individual microcrack AE analysis showed substantial energy reduction emanating spatially from the injection interval. It was quantitatively observed that the recorded AE signals provided sufficient information to generalize the damage radiating spatially away from the injection wellbore.

An Efficient CNN-Based Hybrid Classification and Segmentation Approach for COVID-19 Detection

COVID-19 remains to proliferate precipitously in the world.It has significantly influenced public health,the world economy,and the persons’lives.Hence,there is a need to speed up the diagnosis and precautions to deal with COVID-19 patients.With this explosion of this pandemic,there is a need for automated diagnosis tools to help specialists based onmedical images.This paper presents a hybrid Convolutional Neural Network(CNN)-based classification and segmentation approach for COVID-19 detection from Computed Tomography(CT)images.The proposed approach is employed to classify and segment the COVID-19,pneumonia,and normal CT images.The classification stage is firstly applied to detect and classify the input medical CT images.Then,the segmentation stage is performed to distinguish between pneumonia and COVID-19 CT images.The classification stage is implemented based on a simple and efficient CNN deep learning model.This model comprises four Rectified Linear Units(ReLUs),four batch normalization layers,and four convolutional(Conv)layers.TheConv layer depends on filters with sizes of 64,32,16,and 8.A2×2windowand a stride of 2 are employed in the utilized four max-pooling layers.A soft-max activation function and a Fully-Connected(FC)layer are utilized in the classification stage to perform the detection process.For the segmentation process,the Simplified Pulse Coupled Neural Network(SPCNN)is utilized in the proposed hybrid approach.The proposed segmentation approach is based on salient object detection to localize the COVID-19 or pneumonia region,accurately.To summarize the contributions of the paper,we can say that the classification process with a CNN model can be the first stage a highly-effective automated diagnosis system.Once the images are accepted by the system,it is possible to perform further processing through a segmentation process to isolate the regions of interest in the images.The region of interest can be assesses both automatically and through experts.This strategy helps somuch in saving the time and efforts of specialists with the explosion of COVID-19 pandemic in the world.The proposed classification approach is applied for different scenarios of 80%,70%,or 60%of the data for training and 20%,30,or 40%of the data for testing,respectively.In these scenarios,the proposed approach achieves classification accuracies of 100%,99.45%,and 98.55%,respectively.Thus,the obtained results demonstrate and prove the efficacy of the proposed approach for assisting the specialists in automated medical diagnosis services.