Chest CT Image based Lung Disease Classification - A Review.

Computed tomography (CT) scans are widely used to diagnose lung conditions due to their ability to provide a detailed overview of the body's respiratory system. Despite its popularity, visual examination of CT scan images can lead to misinterpretations that impede a timely diagnosis. Utilizing technology to evaluate images for disease detection is also a challenge. As a result, there is a significant demand for more advanced systems that can accurately classify lung diseases from CT scan images. In this work, we provide an extensive analysis of different approaches and their performances that can help young researchers to build more advanced systems. First, we briefly introduce diagnosis and treatment procedures for various lung diseases. Then, a brief description of existing methods used for the classification of lung diseases is presented. Later, an overview of the general procedures for lung disease classification using machine learning (ML) is provided. Furthermore, an overview of recent progress in ML-based classification of lung diseases is provided. Finally, existing challenges in ML techniques are presented. It is concluded that deep learning techniques have revolutionized the early identification of lung disorders. We expect that this work will equip medical professionals with the awareness they require in order to recognize and classify certain medical disorders.

Influence of broader geographic allograft sharing on outcomes and cost in smaller lung transplant centers.

OBJECTIVE: On November 24, 2017, Organ Procurement and Transplantation Network implemented a change to lung allocation replacing donor service area with a 250 nautical mile radius around donor hospitals. We sought to evaluate the experience of a small to medium size center following implementation. METHODS: Patients (47 pre and 54 post) undergoing lung transplantation were identified from institutional database from January 2016 to October 2019. Detailed chart review and analysis of institutional cost data was performed. Univariate analysis was performed to compare eras. RESULTS: Similar short-term mortality and primary graft dysfunction were observed between groups. Decreased local donation (68% vs 6%; P < .001), increased travel distance (145 vs 235 miles; P = .004), travel cost ($8626 vs $14,482; P < .001), and total procurement cost ($60,852 vs $69,052; P = .001) were observed postimplementation. We also document an increase in waitlist mortality postimplementation (6.9 vs 31.6 per 100 patient-years; P < .001). CONCLUSIONS: Following implementation of the new allocation policy in a small to medium size center, several changes were in accordance with policy intention. However, concerning shifts emerged, including increased waitlist mortality and resource utilization. Continued close monitoring of transplant centers stratified by size and location are paramount to maintaining global availability of lung transplantation to all Americans regardless of geographic residence or socioeconomic status.

Relación genética entre las enfermedades pulmonares de origen ambiental u ocupacional y la osteoporosis: un enfoque bioinformático / Genetic relationship between pulmonary diseases of environmental or occupational origin and osteoporosis: a bioinformatic approach

Objetivo: Identificación de biomarcadores que relacionan la osteoporosis con enfermedades pulmonares ocupacionales y ambientales. Material y métodos: Mediante bases de datos de terminología médica unificada se obtuvieron enfermedades relacionadas con enfermedades pulmonares que, junto con la osteoporosis, fueron analizadas en DisGeNET para obtener los genes asociados a cada enfermedad y formar una red de interacción proteína-proteína (PPI) mediante el uso de STRING dentro de Cytoscape. A través de la aplicación de diferentes algoritmos de centralidad utilizando CythoHubba en Cytoscape, se seleccionaron las 5 proteínas de la red con el mayor grado de centralidad. Resultados: 9 enfermedades fueron incluidas en el grupo de enfermedades pulmonares. Se obtuvieron 2.698 genes asociados a enfermedades pulmonares y a osteoporosis. Los genes vinculados con osteoporosis y con al menos dos de las enfermedades pulmonares incluidas dieron lugar a una red PPI con 152 nodos y 1.378 ejes. Las proteínas con mayor grado de centralidad de la red fueron AKT1, ALB, IL6, TP53 y VEGFA. Conclusiones: Existe una elevada relación entre la osteoporosis y las enfermedades pulmonares ambientales estudiadas, a través de genes con una implicación dual. Nosotros proponemos cinco genes importantes que vinculan estas enfermedades y que podrían constituir una base coherente para investigaciones más profundas en este campo. (AU)

Objetives: Identifying biomarkers that relate osteoporosis to occupational and environmental lung diseases. Material and methods: Using integrated medical terminology databases, diseases related to lung diseases were obtained which, together with osteoporosis, were analyzed in DisGeNET to obtain the genes associated with each disease and form a protein-protein interaction network (PPI) through the Cytoscape StringApp. Applying different centrality algorithms using CythoHubba in Cytoscape, the 5 network proteins with the highest degree of centrality were selected. Results: 9 diseases were included in the group of pulmonary diseases. 2,698 genes associated with lung diseases and osteoporosis were obtained. Genes associated with osteoporosis and with at least two of the included lung diseases resulted in a PPI network with 152 nodes and 1,378 axes. The proteins with the highest degree of network centrality were AKT1, ALB, IL6, TP53 and VEGFA. Conclusions: There is a significant relationship between osteoporosis and the environmental lung diseases studied, through genes with dual involvement. We propose five important genes that link these diseases. This could provide a coherent basis for further research in this field. (AU)

Adverse Cardiovascular and Pulmonary Events Associated With Chimeric Antigen Receptor T-Cell Therapy.

BACKGROUND: Pivotal trials of chimeric antigen receptor T-cell (CAR-T) have identified common toxicities but may have been underpowered to detect cardiovascular and pulmonary adverse events (CPAEs). OBJECTIVES: This study sought to investigate CPAEs associated with commercial CD19-directed CAR-T therapy. METHODS: In this retrospective, pharmacovigilance study, the authors used the Food and Drug Administration adverse event reporting system to identify CPAEs associated with axicabtagene-ciloleucel and tisagenlecleucel. The authors evaluated disproportionate reporting by the reporting odds ratio (ROR) and the lower bound of the information component 95% credibility interval (IC025 >0 is deemed significant). Significant associations were further adjusted to age and sex (adj.ROR). RESULTS: The authors identified CAR-T reports of 2,657 patients, including 546 CPAEs (20.5%). CPAEs overlapped with cytokine release syndrome in 68.3% (373 of 546) of the reports. Compared with the full database, CAR-T was associated with overreporting of tachyarrhythmias (n = 74 [2.8%], adj.ROR = 2.78 [95% CI: 2.21-3.51]), cardiomyopathy (n = 69 [2.6%], adj.ROR = 3.51 [2.42-5.09]), pleural disorders (n = 46 [1.7%], adj.ROR = 3.91 [2.92-5.23]), and pericardial diseases (n = 11 [0.4%], adj.ROR = 2.26 [1.25-4.09], all IC025 >0). Venous thromboembolic events (VTEs) were associated only with axicabtagene-ciloleucel therapy (n = 28 [1.6%], adj.ROR = 1.80 [1.24-2.62], IC025 >0). Atrial fibrillation (n = 55) was the leading tachyarrhythmia, followed by ventricular arrhythmias (n = 14). Tachyarrhythmias and VTEs were reported more often following axicabtagene-ciloleucel than tisagenlecleucel in an age- and sex-adjusted model (adj.ROR = 1.82 [1.04-3.18] and adj.ROR = 2.86 [1.18-6.93], respectively). Finally, the fatality rate of CPAEs was 30.9%. CONCLUSIONS: In this largest post-marketing study to date, the authors identified an association between CAR-T and various CPAEs, including tachyarrhythmias, cardiomyopathy, pericardial and pleural disorders, and VTEs. These findings should be considered in the multidisciplinary assessment for and monitoring of CAR-T therapy recipients.

NHLBI-CMREF Workshop Report on Pulmonary Vascular Disease Classification: JACC State-of-the-Art Review.

The National Heart, Lung, and Blood Institute and the Cardiovascular Medical Research and Education Fund held a workshop on the application of pulmonary vascular disease omics data to the understanding, prevention, and treatment of pulmonary vascular disease. Experts in pulmonary vascular disease, omics, and data analytics met to identify knowledge gaps and formulate ideas for future research priorities in pulmonary vascular disease in line with National Heart, Lung, and Blood Institute Strategic Vision goals. The group identified opportunities to develop analytic approaches to multiomic datasets, to identify molecular pathways in pulmonary vascular disease pathobiology, and to link novel phenotypes to meaningful clinical outcomes. The committee suggested support for interdisciplinary research teams to develop and validate analytic methods, a national effort to coordinate biosamples and data, a consortium of preclinical investigators to expedite target evaluation and drug development, longitudinal assessment of molecular biomarkers in clinical trials, and a task force to develop a master clinical trials protocol for pulmonary vascular disease.

Integrative Omics to Characterize and Classify Pulmonary Vascular Disease.

Advances in high-throughput biotechnologies have facilitated omics profiling, a key component of precision phenotyping, in patients with pulmonary vascular disease. Omics provides comprehensive information pertaining to genes, transcripts, proteins, and metabolites. The resulting omics big datasets may be integrated for more robust results and are amenable to analysis using machine learning or newer analytical methodologies, such as network analysis. Results from fully integrated multi-omics datasets combined with clinical data are poised to provide novel insight into pulmonary vascular disease as well as diagnose the presence of disease and prognosticate outcomes.

The Significance of Subpleural Sparing in CT Chest: A State-of-the-Art Review.

The subpleural sparing pattern is a common finding on computed tomography (CT) of the lungs. It comprises of pulmonary opacities sparing the lung peripheries, typically 1cm and less from the pleural surface. This finding has a variety of causes, including idiopathic, inflammatory, infectious, inhalational, cardiac, traumatic, and bleeding disorders. Specific disorders that can cause subpleural sparing patterns include nonspecific interstitial pneumonia (NSIP), organizing pneumonia (OP), pulmonary alveolar proteinosis (PAP), diffuse alveolar hemorrhage (DAH), vaping-associated lung injury (VALI), cracked lung, pulmonary edema, pneumocystis jirovecii pneumonia (PJP), pulmonary contusion, and more recently, Coronavirus disease 2019 (COVID-19) pneumonia. Knowledge of the many etiologies of this pattern can be useful in preventing diagnostic errors. In addition, although the etiology of subpleural sparing pattern is frequently indistinguishable during an initial radiologic evaluation, the differences in location of opacities in the lungs, as well as the presence of additional radiologic findings, patient history, and clinical presentation, can often be useful to suggest the appropriate diagnosis. We did a comprehensive search on Pubmed and Google Scholar database using keywords of "subpleural sparing," "peripheral sparing," "sparing of peripheries," "CT chest," "chest imaging," and "pulmonary disease." This review aims to describe the primary differential diagnosis of subpleural sparing pattern seen on chest imaging with a strong emphasis on clinical and radiographic findings. We also discuss the pathogenesis and essential clues that are crucial to narrow the differential diagnosis.

Discriminative Feature Learning for Thorax Disease Classification in Chest X-ray Images.

This paper focuses on the thorax disease classification problem in chest X-ray (CXR) images. Different from the generic image classification task, a robust and stable CXR image analysis system should consider the unique characteristics of CXR images. Particularly, it should be able to: 1) automatically focus on the disease-critical regions, which usually are of small sizes; 2) adaptively capture the intrinsic relationships among different disease features and utilize them to boost the multi-label disease recognition rates jointly. In this paper, we propose to learn discriminative features with a two-branch architecture, named ConsultNet, to achieve those two purposes simultaneously. ConsultNet consists of two components. First, an information bottleneck constrained feature selector extracts critical disease-specific features according to the feature importance. Second, a spatial-and-channel encoding based feature integrator enhances the latent semantic dependencies in the feature space. ConsultNet fuses these discriminative features to improve the performance of thorax disease classification in CXRs. Experiments conducted on the ChestX-ray14 and CheXpert dataset demonstrate the effectiveness of the proposed method.

Human T-cell leukaemia virus type 1 associated pulmonary disease: clinical and pathological features of an under-recognised complication of HTLV-1 infection.

The lung is one of several organs that can be affected by HTLV-1 mediated inflammation. Pulmonary inflammation associated with HTLV-1 infection involves the interstitium, airways and alveoli, resulting in several clinical entities including interstitial pneumonias, bronchiolitis and alveolitis, depending on which structures are most affected. Augmentation of the inflammatory effects of HTLV-1 infected lymphocytes by recruitment of other inflammatory cells in a positive feedback loop is likely to underlie the pathogenesis of HTLV-1 associated pulmonary disease, as has been proposed for HTLV-1 associated myelopathy. In contrast to the conclusions of early case series, HTLV-1 associated pulmonary disease can be associated with significant parenchymal damage, which may progress to bronchiectasis where this involves the airways. Based on our current understanding of HTLV-1 associated pulmonary disease, diagnostic criteria are proposed.

Pediatric Cystic Lung Lesions: Where Are We Now?

Pediatric cystic lung lesions have long been a source of confusion for clinicians, radiologists, and pathologists. They encompass a wide spectrum of entities with variable prognostic implications, including congenital lung malformations, pulmonary neoplasms, and hereditary conditions. As our understanding of the developmental and genetic origins of these conditions has evolved, revised nomenclature and classifications have emerged in an attempt to bring clarity to the origin of these lesions and guide clinical management. This review discusses cystic lung lesions and the current understanding of their etiopathogenesis.

Classifying the diagnosis of study participants in clinical trials: a structured and efficient approach.

BACKGROUND: A challenge in imaging research is a diagnostic classification of study participants. We hypothesised that a structured approach would be efficient and that classification by medical students, residents, and an expert panel whenever necessary would be as valid as classification of all patients by experts. METHODS: OPTIMACT is a randomised trial designed to evaluate the effectiveness of replacing chest x-ray for ultra-low-dose chest computed tomography (CT) at the emergency department. We developed a handbook with diagnostic guidelines and randomly selected 240 cases from 2,418 participants enrolled in OPTIMACT. Each case was independently classified by two medical students and, if they disagreed, by the students and a resident in a consensus meeting. Cases without consensus and cases classified as complex were assessed by a panel of medical specialists. To evaluate the validity, 60 randomly selected cases not referred to the panel by the students and the residents were reassessed by the specialists. RESULTS: Overall, the students and, if necessary, residents were able to assign a diagnosis in 183 of the 240 cases (76% concordance; 95% confidence interval [CI] 71-82%). We observed agreement between students and residents versus medical specialists in 50/60 cases (83% concordance; 95% CI 74-93%). CONCLUSIONS: A structured approach in which study participants are assigned diagnostic labels by assessors with increasing levels of medical experience was an efficient and valid classification method, limiting the workload for medical specialists. We presented a viable option for classifying study participants in large-scale imaging trials (Netherlands National Trial Register number NTR6163).

Classification of pulmonary lesion based on multiparametric MRI: utility of radiomics and comparison of machine learning methods.

OBJECTIVES: We develop and validate a radiomics model based on multiparametric magnetic resonance imaging (MRI) in the classification of the pulmonary lesion and identify optimal machine learning methods. MATERIALS AND METHODS: This retrospective analysis included 201 patients (143 malignancies, 58 benign lesions). Radiomics features were extracted from multiparametric MRI, including T2-weighted imaging (T2WI), T1-weighted imaging (TIWI), and apparent diffusion coefficient (ADC) map. Three feature selection methods, including recursive feature elimination (RFE), t test, and least absolute shrinkage and selection operator (LASSO), and three classification methods, including linear discriminate analysis (LDA), support vector machine (SVM), and random forest (RF) were used to distinguish benign and malignant pulmonary lesions. Performance was compared by AUC, sensitivity, accuracy, precision, and specificity. Analysis of performance differences in three randomly drawn cross-validation sets verified the stability of the results. RESULTS: For most single MR sequences or combinations of multiple MR sequences, RFE feature selection method with SVM classifier had the best performance, followed by RFE with RF. The radiomics model based on multiple sequences showed a higher diagnostic accuracy than single sequence for every machine learning method. Using RFE with SVM, the joint model of T1WI, T2WI, and ADC showed the highest performance with AUC = 0.88 ± 0.02 (sensitivity 83%; accuracy 82%; precision 91%; specificity 79%) in test set. CONCLUSION: Quantitative radiomics features based on multiparametric MRI have good performance in differentiating lung malignancies and benign lesions. The machine learning method of RFE with SVM is superior to the combination of other feature selection and classifier methods. KEY POINTS: • Radiomics approach has the potential to distinguish between benign and malignant pulmonary lesions. • Radiomics model based on multiparametric MRI has better performance than single-sequence models. • The machine learning methods RFE with SVM perform best in the current cohort.

A multi-level similarity measure for the retrieval of the common CT imaging signs of lung diseases.

The common CT imaging signs of lung diseases (CISLs) which frequently appear in lung CT images are widely used in the diagnosis of lung diseases. Computer-aided diagnosis (CAD) based on the CISLs can improve radiologists' performance in the diagnosis of lung diseases. Since similarity measure is important for CAD, we propose a multi-level method to measure the similarity between the CISLs. The CISLs are characterized in the low-level visual scale, mid-level attribute scale, and high-level semantic scale, for a rich representation. The similarity at multiple levels is calculated and combined in a weighted sum form as the final similarity. The proposed multi-level similarity method is capable of computing the level-specific similarity and optimal cross-level complementary similarity. The effectiveness of the proposed similarity measure method is evaluated on a dataset of 511 lung CT images from clinical patients for CISLs retrieval. It can achieve about 80% precision and take only 3.6 ms for the retrieval process. The extensive comparative evaluations on the same datasets are conducted to validate the advantages on retrieval performance of our multi-level similarity measure over the single-level measure and the two-level similarity methods. The proposed method can have wide applications in radiology and decision support. Graphical abstract.

Occupational lung diseases in the 21st century: the changing landscape and future challenges.

PURPOSE OF REVIEW: Occupational exposures remain an underrecognized and preventable cause of lung disease in high-income countries. The present review highlights the emergence of cleaning-related respiratory disease and the re-emergence of silicosis as examples of trends in occupational lung diseases in the 21st century. RECENT FINDINGS: Employment trends, such as the shift from large-scale manufacturing to a service economy, the growth of the healthcare sector, and changing consumer products have changed the spectrum of work-related lung diseases. Following decades of progress in reducing traditional hazards such as silica in U.S. workplaces, cases of advanced silicosis have recently re-emerged with the production of engineered stone countertops. With growth in the healthcare and service sectors in the United States, cleaning products have become an important cause of work-related asthma and have recently been associated with an increased risk of chronic obstructive pulmonary disease (COPD) in women. However, these occupational lung diseases largely go unrecognized by practicing clinicians. SUMMARY: The present article highlights how changes in the economy and work structure can lead to new patterns of inhalational workplace hazards and respiratory disease, including cleaning-related respiratory disease and silicosis. Pulmonary clinicians need to be able to recognize and diagnose these occupational lung diseases, which requires a high index of suspicion and a careful occupational history.

Label Co-Occurrence Learning With Graph Convolutional Networks for Multi-Label Chest X-Ray Image Classification.

Existing multi-label medical image learning tasks generally contain rich relationship information among pathologies such as label co-occurrence and interdependency, which is of great importance for assisting in clinical diagnosis and can be represented as the graph-structured data. However, most state-of-the-art works only focus on regression from the input to the binary labels, failing to make full use of such valuable graph-structured information due to the complexity of graph data. In this paper, we propose a novel label co-occurrence learning framework based on Graph Convolution Networks (GCNs) to explicitly explore the dependencies between pathologies for the multi-label chest X-ray (CXR) image classification task, which we term the "CheXGCN". Specifically, the proposed CheXGCN consists of two modules, i.e., the image feature embedding (IFE) module and label co-occurrence learning (LCL) module. Thanks to the LCL model, the relationship between pathologies is generalized into a set of classifier scores by introducing the word embedding of pathologies and multi-layer graph information propagation. During end-to-end training, it can be flexibly integrated into the IFE module and then adaptively recalibrate multi-label outputs with these scores. Extensive experiments on the ChestX-Ray14 and CheXpert datasets have demonstrated the effectiveness of CheXGCN as compared with the state-of-the-art baselines.

Fusing learned representations from Riesz Filters and Deep CNN for lung tissue classification.

A novel method to detect and classify several classes of diseased and healthy lung tissue in CT (Computed Tomography), based on the fusion of Riesz and deep learning features, is presented. First, discriminative parametric lung tissue texture signatures are learned from Riesz representations using a one-versus-one approach. The signatures are generated for four diseased tissue types and a healthy tissue class, all of which frequently appear in the publicly available Interstitial Lung Diseases (ILD) dataset used in this article. Because the Riesz wavelets are steerable, they can easily be made invariant to local image rotations, a property that is desirable when analyzing lung tissue micro-architectures in CT images. Second, features from deep Convolutional Neural Networks (CNN) are computed by fine-tuning the Inception V3 architecture using an augmented version of the same ILD dataset. Because CNN features are both deep and non-parametric, they can accurately model virtually any pattern that is useful for tissue discrimination, and they are the de facto standard for many medical imaging tasks. However, invariance to local image rotations is not explicitly implemented and can only be approximated with rotation-based data augmentation. This motivates the fusion of Riesz and deep CNN features, as the two techniques are very complementary. The two learned representations are combined in a joint softmax model for final classification, where early and late feature fusion schemes are compared. The experimental results show that a late fusion of the independent probabilities leads to significant improvements in classification performance when compared to each of the separate feature representations and also compared to an ensemble of deep learning approaches.

The Effect of Pulmonary Rehabilitation on Non-chronic Obstructive Pulmonary Disease Patients.

BACKGROUND: Pulmonary rehabilitation has shown significant benefit for patients with chronic obstructive pulmonary disease (COPD). The effect on non-COPD pulmonary patients is less well established. OBJECTIVES: To determine whether pulmonary rehabilitation is also beneficial for non-COPD pulmonary patients. METHODS: Clinical and demographic data on non-COPD pulmonary patients who participated in our institutional pulmonary rehabilitation program between January 2009 and December 2016 were collected. Participants engaged in a 60-minute, twice-weekly, ambulatory hospital-based program lasting 12 to 24 sessions. Sessions included both endurance and muscle training as well as healthy lifestyle educational activities. The six-minute walk test (6MWT) and the St. George's Respiratory Questionnaire (SGRQ) were conducted before and after the rehabilitation program. RESULTS: We recruited 214 non-COPD patients, of whom 153 completed at least 12 sessions. Of these, 59 presented with interstitial lung disease (ILD), 18 with non-ILD restrictive lung defects, 25 with asthma, 30 with lung cancer, and 21 with other conditions (e.g., pulmonary hypertension, bronchiectasis) The groups demonstrated significant improvement in 6MWT and in SGRQ scores. Non-COPD patients gained a 61.9 meter (19%) improvement in the 6MWT (P < 0.0001) and 8.3 point reduction in their SGRQ score (P < 0.0001). CONCLUSIONS: Pulmonary rehabilitation is effective in non-COPD pulmonary patients. As such, it should be an integral part of the treatment armament provided to the vast majority of those suffering from chronic respiratory disease.