Normal Chest X-ray Research Articles

The Detection of COVID-19 in Chest X-rays Using Ensemble CNN Techniques

Advances in the field of image classification using convolutional neural networks (CNNs) have greatly improved the accuracy of medical image diagnosis by radiologists. Numerous research groups have applied CNN methods to diagnose respiratory illnesses from chest X-rays and have extended this work to prove the feasibility of rapidly diagnosing COVID-19 with high degrees of accuracy. One issue in previous research has been the use of datasets containing only a few hundred images of chest X-rays containing COVID-19, causing CNNs to overfit the image data. This leads to lower accuracy when the model attempts to classify new images, as would be clinically expected. In this work, we present a model trained on the COVID-QU-Ex dataset containing 33,920 chest X-ray images, with an equal share of COVID-19, Non-COVID pneumonia, and Normal images. The model is an ensemble of pre-trained CNNs (ResNet50, VGG19, and VGG16) and GLCM textural features. The model achieved a 98.34% binary classification accuracy (COVID-19/no COVID-19) on a test dataset of 6581 chest X-rays and 94.68% for distinguishing between COVID-19, Non-COVID pneumonia, and normal chest X-rays. The results also demonstrate that a higher 98.82% three-class test accuracy can be achieved using the model if the training dataset only contains a few thousand images. However, the generalizability of the model suffers due to the smaller dataset size. This study highlights the benefits of both ensemble CNN techniques and larger dataset sizes for medical image classification performance.

P42 Are chest X-rays still necessary for prebiologic tuberculosis screening? A retrospective audit

Abstract Biologics are commonly used in autoimmune dermatological conditions as an option if other treatments are unsuccessful or contraindicated. However, one major side-effect of biologics is immunosuppression, which may lead to the reactivation of latent tuberculosis (TB) in patients. The British Association of Dermatology has implemented guidelines that require doctors to screen for TB before prescribing biologics to patients. The 2020 guidelines advise screening questions, performance of a chest X-ray (CXR) and an interferon-γ release assay (IGRA). Current studies show that CXR has a sensitivity of 15% when diagnosing latent TB. While there is no gold standard for diagnosing latent TB, the IGRA is a commonly used test and is shown to be four times more effective than the tuberculin skin test. This study questions the diagnostic value of CXR in the screening of TB in asymptomatic patients, and whether it should be removed from the guidelines for such patients. A retrospective audit of dermatological patients in our trust was done over a 24-month period (January 2020–January 2022). Patients selected for the audit were those diagnosed with psoriasis or hidradenitis suppurativa and started on biologics during this 24-month period. Patients were assessed on the following: patient history, CXR results, IGRA results and screening outcomes. The data were quantitatively analysed to uncover the necessity of performing a CXR on patients starting biologics. Between January 2020 and January 2022, 77 patients were scheduled for prebiologic TB screening; CXRs were performed in 74 patients (96%), while 72 (94%) of patients had IGRA tests. During this period, three patients (4%) were diagnosed with latent TB based on positive IGRA results and normal CXR findings. They were subsequently treated for latent TB infection prior to biologic therapy. None of the patients with negative IGRA results had positive CXR findings for TB. This finding suggests that CXR is unnecessary at the initial stage of prebiologic TB screening for asymptomatic patients. IGRA can detect both latent and active TB, while CXR is recommended for the diagnosis of active TB. Mandating CXR comes at a cost to the National Health Service and subjects patients to unnecessary radiation exposure. We propose that CXR be done only for patients who are either symptomatic for TB or have a positive IGRA result.

Survey on deep learning techniques for thorax disease classification task

In recent years, medical image analysis has been widely studied, largely because it provides useful tools to support daily clinical routines. Thorax disease classification (TDC) is an important one among these tools since Chest X-ray (CXR) has become prevalent in daily clinical examinations. It aims to distinguish between normal and abnormal CXR images according to different diseases. To this aim, many thorax disease classification systems have been proposed to increase performance. Most of them adopt deep learning techniques to train a deep neural network (DNN) by using labeled CXR datasets. The TDC performance has been continually refined. Aiming to provide future researchers with the work being done on TDC to date, we review the DNN models for this task. A brief review of each method along with their evaluations on a set of benchmark datasets is included. Moreover, we give a detailed comparison of these methods as well as a conclusion.

Classification of COVID-19 Chest X-Ray Images Based on Speeded Up Robust Features and Clustering-Based Support Vector Machines

Abstract Due to the worldwide deficiency of medical test kits and the significant time required by radiology experts to identify the new COVID-19, it is essential to develop fast, robust, and intelligent chest X-ray (CXR) image classification system. The proposed method consists of two major components: feature extraction and classification. The Bag of image features algorithm creates visual vocabulary from two training data categories of chest X-ray images: Normal and COVID-19 patients’ datasets. The algorithm extracts salient features and descriptors from CXR images using the Speeded Up Robust Features (SURF) algorithm. Machine learning with the Clustering-Based Support Vector Machines (CB-SVMs) multiclass classifier is trained using SURF features to classify the CXR image categories. The careful collection of ground truth Normal and COVID-19 CXR datasets, provided by worldwide expert radiologists, has certainly influenced the performance of the proposed CB-SVMs classifier to preserve the generalization capabilities. The high classification accuracy of 99 % demonstrates the effectiveness of the proposed method, where the accuracy is assessed on an independent test sets.

PSO Optimized CNN-SVM Architecture for Covid -19 Classification

This paper presents a hybrid model that utilizes PSO particle swarm optimization, Convolution Neural Networks (CNN) and (SVM) Support Vector Machine architecture for recognition of Covid19.The planned model extracts optimized structures with particle swarm optimization then passes to Convolution Neural Network for automatic feature extraction, while the SVM serves as a Multi classifier. The dataset comprises Covid 19, Pneumonia and Normal Chest X-Ray pictures used to hone and evaluate the suggested algorithm. The most distinct traits are automatically extracted by the algorithm from these photographs. Experimental results show that the suggested framework is effective, with an average recognition accuracy of 97.42%.The most successful SVM Kernel was RBF.

Chest computed tomography scan utilization and diagnostic outcomes in chronic cough patients with normal chest X-rays: analysis of routinely collected data of a tertiary academic hospital.

The role of chest computed tomography (CT) scan is controversial in the management of chronic cough patients with normal chest X-rays. We investigated the utilization pattern and diagnostic outcomes of chest CT scans using institutional routinely collected data (RCD) in South Korea. This is a retrospective analysis of adults with chronic cough (>8 weeks in duration) identified from routinely collected electronic health records (EHRs). Structured data were retrieved, including demographics, medical history, symptoms, and diagnostic test results (including chest X-rays and CT scans). Chest CT scan outcomes were classified into major abnormal findings (malignancy, infectious diseases, or other critical conditions that warrant immediate treatment decisions), minor abnormal findings (other abnormal findings), or normal CT. A total of 5,038 chronic cough patients with normal chest X-rays were analyzed. Chest CT scans were performed in 1,006 patients. Prescription of CT scans was significantly associated with older age, male sex, smoking history, and physician-diagnosed history of lung disease. Only 8 of 1,006 (0.8%) patients had major abnormal findings (4 pneumonia, 2 pulmonary tuberculosis, and 2 lung cancer), while 367 (36.5%) had minor findings, and 631 (62.7%) had normal CT scans. However, no baseline parameters were significantly associated with major CT findings. Chest CT scans were frequently prescribed for chronic cough patients with normal chest X-rays, and abnormal findings were frequently found (37.3%). However, the diagnostic yield for malignancy or infectious disease were low (<1%). Given the potential radiation harm, a routine chest CT scan may not be warranted in chronic cough patients with normal chest X-rays.

Why do emergency department clinicians miss acute aortic syndrome? A case series and descriptive analysis

Objectives: To understand why the diagnosis of AAS is missed in the ED, and to characterise the presenting features of cases in which a diagnosis of AAS was missed.&#x0D; Methods: A retrospective case series cohort study was performed, identifying and analysing cases where AAS was misdiagnosed in three UK EDs between 1st January 2011 and 31st December 2020.&#x0D; Results: 43 cases were included, 22 of which were type A aortic dissections. The most common incorrect presumed diagnoses made were acute coronary syndrome (28%), pulmonary embolism (12%) and ‘non-specific chest pain’ (12%). In 31 cases (72%) there was no evidence from the notes that the clinician had considered AAS in the differential diagnosis. In 10 cases (23%), AAS was considered, but the clinician was falsely reassured by atypical or resolved symptoms, clinical examination, or normal chest x-ray.&#x0D; Conclusions: ED clinicians may miss AAS by not considering it as a possibility, being falsely reassured by atypical or resolved symptoms, or mistaking it for other more common conditions. Further prospective work is necessary to establish the role of diagnostic aids and biomarkers in UK EDs.

Intermediate Learning-Based Attention Regulated Densenet for Diagnosis of Covid-19

Objectives: The AI-based Computer-Aided-Diagnosis of Chest X-Rays related to COVID-19 is very essential. Here we present an Attention regulated Pre-trained DenseNet-121with intermediate transfer learning as a Chest X-Ray image classifier to classify images according to three labels: COVID-19, pneumonia, and normal. Methods: We are proposing a new Attention regulated ImageNet pre-trained DenseNet-121 neural network architecture, which is retrained on NIH ChestX-ray14 data as an intermediate database before the actual COVID- 19 database. We also fine-tuned the last layer of this neural network with a suggested novelty called the output neuron-keeping technique. Before feeding the Covid-19 data we removed all other neurons corresponding to Chest X-ray14 pathology classes except the “No finding” and “Pneumonia” classes. A new third neuron with random weights and bias is created in the final layer to detect Covid-19 pathology. A DenseNet-121 is supported by a GRAD-CAM-based attention mechanism in achieving detection accuracy and localization of pathologies. The used Covid-19 dataset is a combination of 370 Pneumonia, 1255 Normal, and 439 COVID-19 Chest X-Ray images. We randomly took 50 pictures from each class for testing purposes, the remaining images are augmented more to improve DenseNet performance on a small COVID- 19 dataset. Findings: Our state-of-art model achieved 98.6% test accuracy since it misclassified one out of 50 Covid-19 images, and one out of 50 Pneumonia images, but all 50 normal Chest X-Ray images are classified with 100% accuracy. We compared our model with the other three state-of-theart models, particularly under three-class classification problems (Pneumonia,Covid-19, and Normal). The experimental result shows that the proposed model outperforms the existing three models by obtaining an accuracy of 98.6%. The GRAD-CAM generated heatmaps improved the interpretation of the infections. Novelty: The intermediate transfer learning, neuron keeping, and attention-regulated DenseNet are the highlight of the proposed work in achieving detection accuracy. The GRAD-CAM generated heatmaps indicate future research on COVID-19 diagnosis to analyze the severity of the infection. We believe this model may bridge the medical community and Artificial Intelligence gap. Keywords: Chest XRay14; Covid19 diagnosis; DenseNet121; GradCam; Neuron keeping

Down syndrome patients with normal hearts: are they really normal?

Even though congenital heart disease is a common finding in down syndrome (DS) patients, some of them have anatomically normal hearts. However, the term "normal" might not be suitable, as these patients usually suffer from functional cardiac dysfunction. Several research highlighted that despite the absence of anatomical heart defects, subtle cardiac function derangements are present in DS patients. We aim to assess cardiac functions by Two-dimensional echocardiography and tissue Doppler imaging (TDI) in pediatric DS patients who have anatomically normal hearts. One hundred seventy-two patients with karyotyping confirmed DS with anatomically normal hearts and 165 healthy normal control children were enrolled in the current study. Their cardiac functions were assessed using both 2-dimensional echocardiography and TDI. Both patients and controls had structurally and anatomically normal hearts. In DS patients, the right side of the heart showed a significant reduction in both systolic and diastolic functions. Systolic dysfunction was evident by significantly decreased levels of Tricuspid annular plane systolic excursion and systolic wave by TDI. Diastolic dysfunction of the right ventricle was evident by prolonged deceleration time by conventional echocardiography and a significant decrease in annular tissue doppler velocity during early diastole/late diastole ratio by TDI. The E/De ratio was significantly increased. Even with anatomically normal hearts, DS patients should undergo cardiac function assessment by echocardiography & TDI. TDI is superior to conventional echocardiography in detecting subtle cardiac dysfunction especially left ventricular diastolic dysfunction in DS patients. TDI showed a significant decrease in the early/atrial ratio of mitral valve annulus and prolongation of left ventricle isometric relaxation time in DS children. Also, the left ventricle E/De ratio was prolonged denoting elevated filling pressures and diastolic dysfunction. This indicates that the TDI has higher sensitivity to detect diastolic dysfunction than conventional Echocardiography. Biventricular TDI-derived myocardial performance index was found to be significantly increased in DS children.

Epidemiology and clinical characteristics of pediatric sepsis in PICUs of China: A national cross-sectional study.

Epidemiology and clinical characteristics of pediatric sepsis in PICUs of China: A national cross-sectional study.

Interpretable Differential Diagnosis of Non-COVID Viral Pneumonia, Lung Opacity and COVID-19 Using Tuned Transfer Learning and Explainable AI.

The coronavirus epidemic has spread to virtually every country on the globe, inflicting enormous health, financial, and emotional devastation, as well as the collapse of healthcare systems in some countries. Any automated COVID detection system that allows for fast detection of the COVID-19 infection might be highly beneficial to the healthcare service and people around the world. Molecular or antigen testing along with radiology X-ray imaging is now utilized in clinics to diagnose COVID-19. Nonetheless, due to a spike in coronavirus and hospital doctors' overwhelming workload, developing an AI-based auto-COVID detection system with high accuracy has become imperative. On X-ray images, the diagnosis of COVID-19, non-COVID-19 non-COVID viral pneumonia, and other lung opacity can be challenging. This research utilized artificial intelligence (AI) to deliver high-accuracy automated COVID-19 detection from normal chest X-ray images. Further, this study extended to differentiate COVID-19 from normal, lung opacity and non-COVID viral pneumonia images. We have employed three distinct pre-trained models that are Xception, VGG19, and ResNet50 on a benchmark dataset of 21,165 X-ray images. Initially, we formulated the COVID-19 detection problem as a binary classification problem to classify COVID-19 from normal X-ray images and gained 97.5%, 97.5%, and 93.3% accuracy for Xception, VGG19, and ResNet50 respectively. Later we focused on developing an efficient model for multi-class classification and gained an accuracy of 75% for ResNet50, 92% for VGG19, and finally 93% for Xception. Although Xception and VGG19's performances were identical, Xception proved to be more efficient with its higher precision, recall, and f-1 scores. Finally, we have employed Explainable AI on each of our utilized model which adds interpretability to our study. Furthermore, we have conducted a comprehensive comparison of the model's explanations and the study revealed that Xception is more precise in indicating the actual features that are responsible for a model's predictions.This addition of explainable AI will benefit the medical professionals greatly as they will get to visualize how a model makes its prediction and won't have to trust our developed machine-learning models blindly.

Nosocomial COVID-19: A Nationwide Spanish Study

Introduction: SARS-CoV-2 is a highly contagious virus, and despite professionals’ best efforts, nosocomial COVID-19 (NC) infections have been reported. This work aimed to describe differences in symptoms and outcomes between patients with NC and community-acquired COVID-19 (CAC) and to identify risk factors for severe outcomes among NC patients. Methods: This is a nationwide, retrospective, multicenter, observational study that analyzed patients hospitalized with confirmed COVID-19 in 150 Spanish hospitals (SEMI-COVID-19 Registry) from March 1, 2020, to April 30, 2021. NC was defined as patients admitted for non-COVID-19 diseases with a positive SARS-CoV-2 test on the fifth day of hospitalization or later. The primary outcome was 30-day in-hospital mortality (IHM). The secondary outcome was other COVID-19-related complications. A multivariable logistic regression analysis was performed. Results: Of the 23,219 patients hospitalized with COVID-19, 1,104 (4.8%) were NC. Compared to CAC patients, NC patients were older (median 76 vs. 69 years; p < 0.001), had more comorbidities (median Charlson Comorbidity Index 5 vs. 3; p < 0.001), were less symptomatic (p < 0.001), and had normal chest X-rays more frequently (30.8% vs. 12.5%, p < 0.001). After adjusting for sex, age, dependence, COVID-19 wave, and comorbidities, NC was associated with lower risk of moderate/severe acute respiratory distress syndrome (ARDS) (adjusted odds ratio [aOR]: 0.72; 95% confidence interval [CI]: 0.59–0.87; p < 0.001) and higher risk of acute heart failure (aOR: 1.40; 1.12–1.72; p = 0.003), sepsis (aOR: 1.73; 1.33–2.54; p < 0.001), and readmission (aOR: 1.35; 1.03–1.83; p = 0.028). NC was associated with a higher case fatality rate (39.1% vs. 19.2%) in all age groups. IHM was significantly higher among NC patients (aOR: 2.07; 1.81–2.68; p < 0.001). Risk factors for increased IHM in NC patients were age, moderate/severe dependence, malignancy, dyspnea, moderate/severe ARDS, multiple organ dysfunction syndrome, and shock; odynophagia was associated with lower IHM. Conclusions: NC is associated with greater mortality and complications compared to CAC. Hospital strategies to prevent NC must be strengthened.

Abnormality Detection in Chest X-Ray via Residual-Saliency From Normal Generation

In this paper, we propose a novel disease detection framework based on translating a given chest X-ray (CXR) to a corresponding normal CXR image. To train a model for normal CXR translation, we synthesize a paired image dataset from existing public CXR datasets. More specifically, we synthesize new images that appear to be from the same patient but without abnormalities, i.e., matching-pairs, that are not present in the dataset. Here, we search for normal local regions in images contained within the dataset that have a similar appearance, and we apply a blending process to effectively erase the abnormal regions. A conditional GAN can then be trained on the synthetic matching-pair images for a normal CXR translator model that transforms the given CXR images into their corresponding normal CXRs. We incorporate this pretrained translator model into a deep learning based object detection model. Here we define the difference between the translated normal CXR and the given input as the residual-saliency map. As the residual-saliency map will be most active at diseased regions, we use this as the attention within the detection model. We also utilize the matching-pair synthesis pipeline to synthesize abnormal images for data augmentation when we train the detector. We evaluate the proposed approach on two public chest X-ray datasets: RSNA and VinBigData.

Early detection of tuberculosis using hybrid feature descriptors and deep learning network.

To detect tuberculosis (TB) at an early stage by analyzing chest X-ray images using a deep neural network, and to evaluate the efficacy of proposed model by comparing it with existing studies. For the study, an open-source X-ray images were used. Dataset consisted of two types of images, i.e., standard and tuberculosis. Total number of images in the dataset was 4,200, among which, 3,500 were normal chest X-rays, and the remaining 700 X-ray images were of tuberculosis patients. The study proposed and simulated a deep learning prediction model for early TB diagnosis by combining deep features with hand-engineered features. Gabor filter and Canny edge detection method were applied to enhance the performance and reduce computation cost. The proposed model simulated two scenarios: without filter and edge detection techniques and only a pre-trained model with automatic feature extraction, and filter and edge detection techniques. The results achieved from both the models were 95.7% and 97.9%, respectively. The proposed study can assist in the detection if a radiologist is not available. Also, the model was tested with real-time images to examine the efficacy, and was better than other available models.

Circulating markers of neutrophil activation and lung injury in pediatric pneumonia in low-resource settings

ABSTRACT Diagnostic biomarkers for childhood pneumonia could guide management and improve antibiotic stewardship in low-resource settings where chest x-ray (CXR) is not always available. In this cross-sectional study, we measured chitinase 3-like protein 1 (CHI3L1), surfactant protein D (SP-D), lipocalin-2 (LCN2), and tissue inhibitor of metalloproteinases-1 (TIMP-1) in Ugandan children under the age of five hospitalized with acute lower respiratory tract infection. We determined the association between biomarker levels and primary end-point pneumonia, indicated by CXR consolidation. We included 89 children (median age 11 months, 39% female). Primary endpoint pneumonia was present in 22 (25%). Clinical signs were similar in children with and without CXR consolidation. Broad-spectrum antibiotics (ceftriaxone) were administered in 83 (93%). Levels of CHI3L1, SP-D, LCN2 and TIMP-1 were higher in patients with primary end-point pneumonia compared to patients with normal CXR or other infiltrates. All markers were moderately accurate predictors of primary end-point pneumonia, with area under receiver operator characteristic curves of 0.66-0.70 (p<0.05 for all markers). The probability of CXR consolidation increased monotonically with the number of markers above cut-off. Among 28 patients (31%) in whom all four markers were below the cut-off, the likelihood ratio of CXR consolidation was 0.11 (95%CI 0.015 to 0.73). CHI3L1, SP-D, LCN2 and TIMP-1 were associated with CXR consolidation in children with clinical pneumonia in a low-resource setting. Combinations of quantitative biomarkers may be useful to safely withhold antibiotics in children with a low probability of bacterial infection.

Exhaled Pneumocystis jirovecii output and detection of asymptomatic exhalation by facemask sampling in HIV-uninfected, immunocompromised patients

Pneumocystis jirovecii pneumonia (PJP) transmission is poorly defined. Previous studies have sampled air of rooms occupied by HIV-infected patients with PJP, while natural and direct exhalations of HIV-uninfected subjects remain under-investigated. Here, clinical facemasks were used to examine and quantify potential P.jirovecii exhalations from HIV-uninfected patients with suspected PJP and to determine whether pathogen exhalation was definable clinically or radiologically. Forty-five patients in Leicester (England), highly suspected of having PJP based on European Conference on Infections in Leukaemia (ECIL-5) guidelines, each wore one facemask carrying a gelatine/PVA sampling matrix for 1h while respiring normally. Mask contamination with P.jirovecii was assessed using a modified quantitative polymerase chain reaction targeting mitochondrial large subunit (MtLSU). Radiological findings on chest X-ray (CXR) and computed tomography (CT) were graded and analysed for correlation with P.jirovecii signals alongside relevant clinical and laboratory findings. P.jirovecii was detected in seven of 20 patients diagnosed with PJP and three of 19 patients with suspected but undiagnosed PJP. The median captured signal was 8.59× 104 MtLSU copies/mask (interquartile range (IQR)= 3.01× 105-1.81× 104). Blood β-D-glucan test results correlated with the mask detection data (r= 0.65; P<0.0001) but other clinical indices and radiological features did not. Five of the 10 P.jirovecii-exhalers exhibited normal CXR with a median exhalation burden 1.28× 105 copies/mask (IQR= 1.51× 105-2.27× 104). Two P.jirovecii-exhalers (7.64× 104 copies/mask) were asymptomatic. P.jirovecii was exhaled sufficiently during normal respiration to be detectable in facemasks worn by HIV-uninfected patients. Neither clinical nor radiological features correlated with P.jirovecii exhalation.

Image Translation by Ad CycleGAN for COVID-19 X-Ray Images: A New Approach for Controllable GAN.

We propose a new generative model named adaptive cycle-consistent generative adversarial network, or Ad CycleGAN to perform image translation between normal and COVID-19 positive chest X-ray images. An independent pre-trained criterion is added to the conventional Cycle GAN architecture to exert adaptive control on image translation. The performance of Ad CycleGAN is compared with the Cycle GAN without the external criterion. The quality of the synthetic images is evaluated by quantitative metrics including Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Peak Signal-to-Noise Ratio (PSNR), Universal Image Quality Index (UIQI), visual information fidelity (VIF), Frechet Inception Distance (FID), and translation accuracy. The experimental results indicate that the synthetic images generated either by the Cycle GAN or by the Ad CycleGAN have lower MSE and RMSE, and higher scores in PSNR, UIQI, and VIF in homogenous image translation (i.e., Y → Y) compared to the heterogenous image translation process (i.e., X → Y). The synthetic images by Ad CycleGAN through the heterogeneous image translation have significantly higher FID score compared to Cycle GAN (p < 0.01). The image translation accuracy of Ad CycleGAN is higher than that of Cycle GAN when normal images are converted to COVID-19 positive images (p < 0.01). Therefore, we conclude that the Ad CycleGAN with the independent criterion can improve the accuracy of GAN image translation. The new architecture has more control on image synthesis and can help address the common class imbalance issue in machine learning methods and artificial intelligence applications with medical images.

Pneumonia of Chest X-Ray Images Detection using VGG Architectures

Pneumonia is one of the world's most frequent diseases. Pneumonia, a serious respiratory infection, damages the lungs and makes it difficult to breathe. Chest X-ray scans are one of the most significant and common procedures for diagnosing pneumonia condition. An effective model for the identification of pneumonia based on digital chest X-ray pictures is proposed in this paper, which could help radiologists make better decisions. This is a supervised learning technique in which the proposed model predicts the outcome depending on the dataset's image quality. In this research, Kaggle dataset has been used to train the model. To improve training and validation accuracy, fine-tuning of Visual Geometry Group (VGG) pre-trained model are conducted with different hyperparameters. The deep Convolutional Neural Networks (CNNs) based VGG-16 and VGG-19 architectures are used to extract features from given chest X-ray images. These features are then used for classification. In addition to the accuracy and F1-score as an evaluation matrix, the results of pre-trained models are compared using the model loss, and model accuracy graphs. The model's performance in detecting pneumonia demonstrates that the proposed VGG pre-trained model can efficiently categorize normal and pneumonia chest X-rays images in practice. As a result, the proposed model can be utilized to make a speedy diagnosis of pneumonia and can assist radiologists with their work.

Biometric Data of Adults’ Aortic Knob Diameter in Posteroanterior Chest Radiograph, Correlation to Age and Normative Heart Diameter: A Cross-Sectional Study

Background: The aortic knob (AK) is an essential feature on a chest x-ray. It could be the first sign of a cardiovascular problem if there is any deformation or enlargement of the knob. This study aimed to measure the normal AK diameter (AKD) on a posteroanterior chest radiograph in healthy adult Sudanese. Methods and Results: The study was conducted in the Department of Radiology and Imaging in Ribat Hospital (Sudan) between Jun 2019 and Jan 2020. A total of 113 participants of both sexes (45.1% males and 54.9% females) with a normal chest x-ray and no history of diabetes, blood hypertension, cardiovascular disease, or skeletal abnormality were selected. Participants' age fluctuated from 18 to 75 years. The measurements (AK, heart diameter [HD], cardiothoracic ratio [CTR]) were carried out with the measuring tools available on the software of the computed radiography system. The mean AKD was 2.8±0.8 cm (2.94±0.8 cm in males and 2.51±0.77 cm in females, P=0.005)). The mean HD was 9.22±2.8 cm (9.8±3.3 cm in males and 8.7±0.2.1 cm in females, P=0.005). The mean CTR was estimated as 46.6±7.7% with a significant difference between males and females and significantly correlated with HD and BMI (P&lt;0.05). The AKD increased by 0.019 cm with an increase of one year of age (AKD = 0.0199(age)+1.9469), and there was a strong positive correlation between age and AKD (P&lt;0.001). Conclusion: The study found a significant positive correlation between age and AKD. Increased heart sizes increase AKD. The AKD value is greater in males than in females.

LWSNet - a novel deep-learning architecture to segregate Covid-19 and pneumonia from x-ray imagery.

Automatic detection of lung diseases using AI-based tools became very much necessary to handle the huge number of cases occurring across the globe and support the doctors. This paper proposed a novel deep learning architecture named LWSNet (Light Weight Stacking Network) to separate Covid-19, cold pneumonia, and normal chest x-ray images. This framework is based on single, double, triple, and quadruple stack mechanisms to address the above-mentioned tri-class problem. In this framework, a truncated version of standard deep learning models and a lightweight CNN model was considered to conviniently deploy in resource-constraint devices. An evaluation was conducted on three publicly available datasets alongwith their combination. We received 97.28%, 96.50%, 97.41%, and 98.54% highest classification accuracies using quadruple stack. On further investigation, we found, using LWSNet, the average accuracy got improved from individual model to quadruple model by 2.31%, 2.55%, 2.88%, and 2.26% on four respective datasets.