A robust ensemble deep learning framework for accurate diagnoses of tuberculosis from chest radiographs.

Introduction: Tuberculosis (TB) stands as a paramount global health concern, contributing significantly to worldwide mortality rates. Effective containment of TB requires deployment of cost-efficient screening method with limited resources. To enhance the precision of resource allocation in the global fight against TB, this research proposed chest X-ray radiography (CXR) based machine learning screening algorithms with optimization, benchmarking and tuning for the best TB subclassification tasks for clinical application. Methods: This investigation delves into the development and evaluation of a robust ensemble deep learning framework, comprising 43 distinct models, tailored for the identification of active TB cases and the categorization of their clinical subtypes. The proposed framework is essentially an ensemble model with multiple feature extractors and one of three fusion strategies-voting, attention-based, or concatenation methods-in the fusion stage before a final classification. The comprised de-identified dataset contains records of 915 active TB patients alongside 1,276 healthy controls with subtype-specific information. Thus, the realizations of our framework are capable for diagnosis with subclass identification. The subclass tags include: secondary tuberculosis/tuberculous pleurisy; non-cavity/cavity; secondary tuberculosis only/secondary tuberculosis and tuberculous pleurisy; tuberculous pleurisy only/secondary tuberculosis and tuberculous pleurisy. Results: Based on the dataset and model selection and tuning, ensemble models show their capability with self-correction capability of subclass identification with rendering robust clinical predictions. The best double-CNN-extractor model with concatenation/attention fusion strategies may potentially be the successful model for subclass tasks in real application. With visualization techniques, in-depth analysis of the ensemble model's performance across different fusion strategies are verified. Discussion: The findings underscore the potential of such ensemble approaches in augmenting TB diagnostics with subclassification. Even with limited dataset, the self-correction within the ensemble models still guarantees the accuracies to some level for potential clinical decision-making processes in TB management. Ultimately, this study shows a direction for better TB screening in the future TB response strategy.

An Upper Bound Visualization of Design Trade-Offs in Adsorbent Materials for Gas Separations: CO2 , N2 , CH4 , H2 , O2 , Xe, Kr, and Ar Adsorbents.

The last 20 years have seen many publications investigating porous solids for gas adsorption and separation. The abundance of adsorbent materials (this work identifies 1608 materials for CO2 /N2 separation alone) provides a challenge to obtaining a comprehensive view of the field, identifying leading design strategies, and selecting materials for process modeling. In 2021, the empirical bound visualization technique was applied, analogous to the Robeson upper bound from membrane science, to alkane/alkene adsorbents. These bound visualizations reveal that adsorbent materials are limited by design trade-offs between capacity, selectivity, and heat of adsorption. The current work applies the bound visualization to adsorbents for a wider range of gas pairs, including CO2 , N2 , CH4 , H2 , Xe, O2 , and Kr. How this visual tool can identify leading materials and place new material discoveries in the context of the wider field is presented. The most promising current strategies for breaking design trade-offs are discussed, along with reproducibility of published adsorption literature, and the limitations of bound visualizations. It is hoped that this work inspires new materials that push the bounds of traditional trade-offs while also considering practical aspects critical to the use of materials on an industrial scale such as cost, stability, and sustainability.

Distinguishing nontuberculous mycobacteria from Mycobacterium tuberculosis lung disease from CT images using a deep learning framework.

PURPOSE: To develop and evaluate the effectiveness of a deep learning framework (3D-ResNet) based on CT images to distinguish nontuberculous mycobacterium lung disease (NTM-LD) from Mycobacterium tuberculosis lung disease (MTB-LD). METHOD: Chest CT images of 301 with NTM-LD and 804 with MTB-LD confirmed by pathogenic microbiological examination were retrospectively collected. The differences between the clinical manifestations of the two diseases were analysed. 3D-ResNet was developed to randomly extract data in an 8:1:1 ratio for training, validating, and testing. We also collected external test data (40 with NTM-LD and 40 with MTB-LD) for external validation of the model. The activated region of interest was evaluated using a class activation map. The model was compared with three radiologists in the test set. RESULT: Patients with NTM-LD were older than those with MTB-LD, patients with MTB-LD had more cough, and those with NTM-LD had more dyspnoea, and the results were statistically significant (p < 0.05). The AUCs of our model on training, validating, and testing datasets were 0.90, 0.88, and 0.86, respectively, while the AUC on the external test set was 0.78. Additionally, the performance of the model was higher than that of the radiologist, and without manual labelling, the model automatically identified lung areas with abnormalities on CT > 1000 times more effectively than the radiologists. CONCLUSION: This study shows the efficacy of 3D-ResNet as a rapid auxiliary diagnostic tool for NTB-LD and MTB-LD. Its use can help provide timely and accurate treatment strategies to patients with these diseases.

A Follow-Up Study of Lung Function and Chest Computed Tomography at 6 Months after Discharge in Patients with Coronavirus Disease 2019.

We aimed to investigate changes in pulmonary function and computed tomography (CT) findings in patients with coronavirus disease 2019 (COVID-19) during the recovery period. COVID-19 patients underwent symptom assessment, pulmonary function tests, and high-resolution chest CT 6 months after discharge from the hospital. Of the 54 patients enrolled, 31 and 23 were in the moderate and severe group, respectively. The main symptoms 6 months after discharge were fatigue and exertional dyspnea, experienced by 24.1% and 18.5% of patients, respectively, followed by smell and taste dysfunction (9.3%) and cough (5.6%). One patient dropped out of the pulmonary function tests. Of the remaining 54 patients, 41.5% had pulmonary dysfunction. Specifically, 7.5% presented with restrictive ventilatory dysfunction (forced vital capacity <80% of the predicted value), 18.9% presented with small airway dysfunction, and 32.1% presented with pulmonary diffusion impairment (diffusing capacity for carbon monoxide <80% of the predicted value). Of the 54 patients enrolled, six patients dropped out of the chest CT tests. Eleven of the remaining 48 patients presented with abnormal lung CT findings 6 months after discharge. Patients with residual lung lesions were more common in the severe group (52.6%) than in the moderate group (3.4%); a higher proportion of patients had involvement of both lungs (42.1% vs. 3.4%) in the severe group. The residual lung lesions were mainly ground-glass opacities (20.8%) and linear opacities (14.6%). Semiquantitative visual scoring of the CT findings revealed significantly higher scores in the left, right, and both lungs in the severe group than in the moderate group. COVID-19 patients 6 months after discharge mostly presented with fatigue and exertional dyspnea, and their pulmonary dysfunction was mostly characterized by pulmonary diffusion impairment. As revealed by chest CT, the severe group had a higher prevalence of residual lesions than the moderate group, and the residual lesions mostly manifested as ground-glass opacities and linear opacities.

Analysis of Chest CT Results of Coronavirus Disease 2019 (COVID-19) Patients at First Follow-Up.

Objective: To investigate the dissipation and outcomes of pulmonary lesions at the first follow-up of patients who recovered from moderate and severe cases of COVID-19. Methods: From January 21 to March 3, 2020, a total of 136 patients with COVID-19 were admitted to our hospital. According to inclusion and exclusion criteria, 52 patients who recovered from COVID-19 were included in this study, including 33 moderate cases and 19 severe cases. Three senior radiologists independently and retrospectively analyzed the chest CT imaging data of 52 patients at the last time of admission and the first follow-up after discharge, including primary manifestations, concomitant manifestations, and degree of residual lesion dissipation. Results: At the first follow-up after discharge, 16 patients with COVID-19 recovered to normal chest CT appearance, while 36 patients still had residual pulmonary lesions, mainly including 33 cases of ground-glass opacity, 5 cases of consolidation, and 19 cases of fibrous strip shadow. The proportion of residual pulmonary lesions in severe cases (17/19) was statistically higher than in moderate cases (19/33) (χ 2 = 5.759, P < 0.05). At the first follow-up, residual pulmonary lesions were dissipated to varying degrees in 47 cases, and lesions remained unchanged in 5 cases. There were no cases of increased numbers of lesions, enlargement of lesions, or appearance of new lesions. The dissipation of residual pulmonary lesions in moderate patients was statistically better than in severe patients (Z = -2.538, P < 0.05). Conclusion: Clinically cured patients with COVID-19 had faster dissipation of residual pulmonary lesions after discharge, while moderate patients had better dissipation than severe patients. However, at the first follow-up, most patients still had residual pulmonary lesions, which were primarily ground-glass opacity and fibrous strip shadow. The proportion of residual pulmonary lesions was higher in severe cases of COVID-19, which required further follow-up.

Machine Learning-Based Differentiation of Nontuberculous Mycobacteria Lung Disease and Pulmonary Tuberculosis Using CT Images.

An increasing number of patients infected with nontuberculous mycobacteria (NTM) are observed worldwide. However, it is challenging to identify NTM lung diseases from pulmonary tuberculosis (PTB) due to considerable overlap in classic manifestations and clinical and radiographic characteristics. This study quantifies both cavitary and bronchiectasis regions in CT images and explores a machine learning approach for the differentiation of NTM lung diseases and PTB. It involves 116 patients and 103 quantitative features. After the selection of informative features, a linear support vector machine performs disease classification, and simultaneously, discriminative features are recognized. Experimental results indicate that bronchiectasis is relatively more informative, and two features are figured out due to promising prediction performance (area under the curve, 0.84 ± 0.06; accuracy, 0.85 ± 0.06; sensitivity, 0.88 ± 0.07; and specificity, 0.80 ± 0.12). This study provides insight into machine learning-based identification of NTM lung diseases from PTB, and more importantly, it makes early and quick diagnosis of NTM lung diseases possible that can facilitate lung disease management and treatment planning.