An exploratory deep learning approach to investigate tuberculosis pathogenesis in nonhuman primate model: Combining automated radiological analysis with clinical and biomarkers data.

BACKGROUND: Tuberculosis (TB) kills approximately 1.6 million people yearly despite the fact anti-TB drugs are generally curative. Therefore, TB-case detection and monitoring of therapy, need a comprehensive approach. Automated radiological analysis, combined with clinical, microbiological, and immunological data, by machine learning (ML), can help achieve it. METHODS: Six rhesus macaques were experimentally inoculated with pathogenic Mycobacterium tuberculosis in the lung. Data, including Computed Tomography (CT), were collected at 0, 2, 4, 8, 12, 16, and 20 weeks. RESULTS: Our ML-based CT analysis (TB-Net) efficiently and accurately analyzed disease progression, performing better than standard deep learning model (LLM OpenAI's CLIP Vi4). TB-Net based results were more consistent than, and confirmed independently by, blinded manual disease scoring by two radiologists and exhibited strong correlations with blood biomarkers, TB-lesion volumes, and disease-signs during disease pathogenesis. CONCLUSION: The proposed approach is valuable in early disease detection, monitoring efficacy of therapy, and clinical decision making.

A deep learning-based algorithm for pulmonary tuberculosis detection in chest radiography.

In tuberculosis (TB), chest radiography (CXR) patterns are highly variable, mimicking pneumonia and many other diseases. This study aims to evaluate the efficacy of Google teachable machine, a deep neural network-based image classification tool, to develop algorithm for predicting TB probability of CXRs. The training dataset included 348 TB CXRs and 3806 normal CXRs for training TB detection. We also collected 1150 abnormal CXRs and 627 normal CXRs for training abnormality detection. For external validation, we collected 250 CXRs from our hospital. We also compared the accuracy of the algorithm to five pulmonologists and radiological reports. In external validation, the AI algorithm showed areas under the curve (AUC) of 0.951 and 0.975 in validation dataset 1 and 2. The accuracy of the pulmonologists on validation dataset 2 showed AUC range of 0.936-0.995. When abnormal CXRs other than TB were added, AUC decreased in both human readers (0.843-0.888) and AI algorithm (0.828). When combine human readers with AI algorithm, the AUC further increased to 0.862-0.885. The TB CXR AI algorithm developed by using Google teachable machine in this study is effective, with the accuracy close to experienced clinical physicians, and may be helpful for detecting tuberculosis by CXR.

Development and validation of a preoperative CT­based radiomics nomogram to differentiate tuberculosis granulomas from lung adenocarcinomas: an external validation study.

BACKGROUND: An accurate and non-invasive approach is urgently needed to distinguish tuberculosis granulomas from lung adenocarcinomas. This study aimed to develop and validate a nomogram based on contrast enhanced-compute tomography (CE-CT) to preoperatively differentiate tuberculosis granuloma from lung adenocarcinoma appearing as solitary pulmonary solid nodules (SPSN). METHODS: This retrospective study analyzed 143 patients with lung adenocarcinoma (mean age: 62.4 ± 6.5 years; 54.5% female) and 137 patients with tuberculosis granulomas (mean age: 54.7 ± 8.2 years; 29.2% female) from two centers between March 2015 and June 2020. The training and internal validation cohorts included 161 and 69 patients (7:3 ratio) from center No.1, respectively. The external testing cohort included 50 patients from center No.2. Clinical factors and conventional radiological characteristics were analyzed to build independent predictors. Radiomics features were extracted from each CT-volume of interest (VOI). Feature selection was performed using univariate and multivariate logistic regression analysis, as well as the least absolute shrinkage and selection operator (LASSO) method. A clinical model was constructed with clinical factors and radiological findings. Individualized radiomics nomograms incorporating clinical data and radiomics signature were established to validate the clinical usefulness. The diagnostic performance was assessed using the receiver operating characteristic (ROC) curve analysis with the area under the receiver operating characteristic curve (AUC). RESULTS: One clinical factor (CA125), one radiological characteristic (enhanced-CT value) and nine radiomics features were found to be independent predictors, which were used to establish the radiomics nomogram. The nomogram demonstrated better diagnostic efficacy than any single model, with respective AUC, accuracy, sensitivity, and specificity of 0.903, 0.857, 0.901, and 0.807 in the training cohort; 0.933, 0.884, 0.893, and 0.892 in the internal validation cohort; 0.914, 0.800, 0.937, and 0.735 in the external test cohort. The calibration curve showed a good agreement between prediction probability and actual clinical findings. CONCLUSION: The nomogram incorporating clinical factors, radiological characteristics and radiomics signature provides additional value in distinguishing tuberculosis granuloma from lung adenocarcinoma in patients with a SPSN, potentially serving as a robust diagnostic strategy in clinical practice.

Artificial intelligence-based radiographic extent analysis to predict tuberculosis treatment outcomes: a multicenter cohort study.

Predicting outcomes in pulmonary tuberculosis is challenging despite effective treatments. This study aimed to identify factors influencing treatment success and culture conversion, focusing on artificial intelligence (AI)-based chest X-ray analysis and Xpert MTB/RIF assay cycle threshold (Ct) values. In this retrospective study across six South Korean referral centers (January 1 to December 31, 2019), we included adults with rifampicin-susceptible pulmonary tuberculosis confirmed by Xpert assay from sputum samples. We analyzed patient characteristics, AI-based tuberculosis extent scores from chest X-rays, and Xpert Ct values. Of 230 patients, 206 (89.6%) achieved treatment success. The median age was 61 years, predominantly male (76.1%). AI-based radiographic tuberculosis extent scores (median 7.5) significantly correlated with treatment success (odds ratio [OR] 0.938, 95% confidence interval [CI] 0.895-0.983) and culture conversion at 8 weeks (liquid medium: OR 0.911, 95% CI 0.853-0.973; solid medium: OR 0.910, 95% CI 0.850-0.973). Sputum smear positivity was 49.6%, with a median Ct of 26.2. However, Ct values did not significantly correlate with major treatment outcomes. AI-based radiographic scoring at diagnosis is a significant predictor of treatment success and culture conversion in pulmonary tuberculosis, underscoring its potential in personalized patient management.

Lung Lesions: Differential Diagnoses Beyond Cancer.

BACKGROUND Lung cancer is the most common malignant neoplasm diagnosed worldwide. Early diagnosis and treatment are of great importance for patient's prognosis. A wide variety of pulmonary conditions display clinical and radiological presentation similar to that of lung cancer, and the awareness of their existence can help in making correct diagnoses. CASE REPORT This article presents a description of 4 patients with an insidious type of lesions mimicking pulmonary carcinomas. The first patient was referred to Department with a tumor-like lesion in the right lung. After CT of the chest and core-needle biopsy, the lesion turned out to be an ectopic thyroid tissue. The second patient reported a dry cough and weight loss. A lung nodule mass was revealed in chest CT and the patient was diagnosed with pulmonary tuberculoma. The remaining 2 patients, despite the suspicion of lung cancer, were subsequently diagnosed with a post-traumatic pleural hematoma and diffuse large B cell lymphoma. CONCLUSIONS Low-dose computed tomography of the chest plays a significant role in the diagnosis of newly detected lesions in the lungs. However, due to the similarity of the image of cancer to that of other diseases, the ultimate diagnosis should be based on the interpretation of full imaging diagnostic tests, clinical presentation, and histopathological examination of the material obtained from the lesion. Analysis of cases enables us to expand our understanding of the diseases that need to be considered in differential diagnosis of a patient with a detected tumor-like lesion in the lungs.

Integrating molecular and radiological screening tools during community-based active case-finding for tuberculosis and COVID-19 in southern Africa.

OBJECTIVES: To evaluate diagnostic yield and feasibility of integrating testing for TB and COVID-19 using molecular and radiological screening tools during community-based active case-finding (ACF). METHODS: Community-based participants with presumed TB and/or COVID-19 were recruited using a mobile clinic. Participants underwent simultaneous point-of-care (POC) testing for TB (sputum; Xpert Ultra) and COVID-19 (nasopharyngeal swabs; Xpert SARS-CoV-2). Sputum culture and SARS-CoV-2 RT-PCR served as reference standards. Participants underwent ultra-portable POC chest radiography with computer-aided detection (CAD). TB infectiousness was evaluated using smear microscopy, cough aerosol sampling studies (CASS), and chest radiographic cavity detection. Feasibility of POC testing was evaluated via user-appraisals. RESULTS: Six hundred and one participants were enrolled, with 144/601 (24.0%) reporting symptoms suggestive of TB and/or COVID-19. 16/144 (11.1%) participants tested positive for TB, while 10/144 (6.9%) tested positive for COVID-19 (2/144 [1.4%] had concurrent TB/COVID-19). Seven (7/16 [43.8%]) individuals with TB were probably infectious. Test-specific sensitivity and specificity (95% CI) were: Xpert Ultra 75.0% (42.8-94.5) and 96.9% (92.4-99.2); Xpert SARS-CoV-2 66.7% (22.3-95.7) and 97.1% (92.7-99.2). Area under the curve (AUC) for CAD4TB was 0.90 (0.82-0.97). User appraisals indicated POC Xpert to have 'good' user-friendliness. CONCLUSIONS: Integrating TB/COVID-19 screening during community-based ACF using POC molecular and radiological tools is feasible, has a high diagnostic yield, and can identity probably infectious persons.

Comparison of Computed Tomography Findings in Lung Tuberculosis in Diabetic and Nondiabetic Patients.

BACKGROUND: Tuberculosis (TB) is one of the leading infectious causes of mortality globally. The purpose of this research is to examine the clinical and radiological characteristics of patients with TB and diabetes. METHODS: The research comprised 276 TB patients, 52 of whom were diabetic and 224 of whom were not. During the evaluation of the patients' clinical histories, age, gender, diagnostic indicator, and whether or not they had undergone prior treatment were questioned, as were the requirement of inpatient treatment and the existence of drug resistance. Radiographically, they were questioned in terms of bilateral-unilateral extent, percentage of parenchymal involvement, cavitation, tree-in-bud appearance, the presence of ground glass, consolidation, miliary involvement, sequela fibrotic changes, parenchymal calcification, mediastinal lymphadenopathy, pleural effusion, and pleural calcification. In addition, segmenting was used to assess involvement in the affected lobes. RESULTS: When we look at the results of 276 patients, 182 males and 94 females, the mean age is 46.01 ± 17.83. Diabetes and TB coexistence are more prevalent in male individuals (P = 0.029). Smear positivity and the need for inpatient treatment were found to be higher in the clinical features of diabetic patients (P = 0.05 and P = 0.01, respectively). Radiologically, diabetes individuals are more likely to have larger mediastinal lymph nodes (P = 0.032). CONCLUSION: In the coexistence of both TB and diabetes, there are variations in radiological findings, complexity in treatment response, and patient management.

Chest Radiograph Screening for Detecting Subclinical Tuberculosis in Asymptomatic Household Contacts, Peru.

The World Health Organization's end TB strategy promotes the use of symptom and chest radiograph screening for tuberculosis (TB) disease. However, asymptomatic early states of TB beyond latent TB infection and active disease can go unrecognized using current screening criteria. We conducted a longitudinal cohort study enrolling household contacts initially free of TB disease and followed them for the occurrence of incident TB over 1 year. Among 1,747 screened contacts, 27 (52%) of the 52 persons in whom TB subsequently developed during follow-up had a baseline abnormal radiograph. Of contacts without TB symptoms, persons with an abnormal radiograph were at higher risk for subsequent TB than persons with an unremarkable radiograph (adjusted hazard ratio 15.62 [95% CI 7.74-31.54]). In young adults, we found a strong linear relationship between radiograph severity and time to TB diagnosis. Our findings suggest chest radiograph screening can extend to detecting early TB states, thereby enabling timely intervention.

Mycobacterium tuberculosis y virus linfotrópico humano, una co-infección fatal / Mycobacterium tuberculosis and human lymphotropic virus, a fatal co-infection

La tuberculosis es una infección de alta incidencia en Latinoamérica. Su presentación como infección activa está determinada por factores de riesgo del hospedero. Comunicamos el caso clínico de una mujer joven que presentó una forma grave de tuberculosis pulmonar. Al explorar sus factores de riesgo se confirmó un estado de inmunosupresión profundo, causado por un linfoma de células T, asociada a una co-infección por virus linfotrópico T humano tipo 1. Se destacan los aspectos microbiológicos y de pronóstico de la co-infección de Mycobacterium tuberculosis y HTLV-1

Tuberculosis is a high-incidence infection in Latin America. Its presentation as an active infection is determined by risk factors in the host. We report the case of a young woman who presented a severe form of pulmonary tuberculosis. When exploring her risk factors, a profound state of immunosuppression was found, caused by T-cell lymphoma, associated with co-infection with human lymphotropic virus. Microbiological and prognostic aspects of Mycobacterium tuberculosis and HTLV-1 co-infection are highlighted.

Added value of spectral computed tomography quantitative parameters for differentiating tuberculosis-associated fibrosing mediastinitis from endobronchial lung cancer: initial results.

OBJECTIVE: The objective of this study was to explore the added value of spectral computed tomography (CT) parameters to conventional CT features for differentiating tuberculosis-associated fibrosing mediastinitis (TB-associated FM) from endobronchial lung cancer (EBLC). METHODS: Chest spectral CT enhancement images from 109 patients with atelectasis were analyzed retrospectively. These patients were divided into two distinct categories: the TB-associated FM group (n = 77) and the EBLC group (n = 32), based on bronchoscopy and/or pathological findings. The selection of spectrum parameters was optimized with the least absolute shrinkage and selection operator regression analysis. The relationship between the spectrum parameters and conventional parameters was explored using Pearson's correlation. Multivariate logistic regression analysis was used to build spectrum model. The spectrum parameters in the spectrum model were replaced with their corresponding conventional parameters to build the conventional model. Diagnostic performances were evaluated using receiver operating characteristic curve analyses. RESULTS: There was a moderate correlation between the parameters ㏒(L-AEFNIC) - ㏒(L-AEFC) (r= 0.419; p< 0.0001), ㏒(O-AEF40KeV) - ㏒(O-AEFC) (r= 0.475; p< 0.0001), [L-A-hydroxyapatite {HAP}(I)] - (L-U-CT) (r= 0.604; p< 0.0001), {arterial enhancement fraction (AEF) derived from normalized iodine concentration (NIC) of lymph node (L-AEFNIC), AEF derived from CT40KeV of bronchial obstruction (O-AEF40KeV), arterial-phase Hydroxyapatite (Iodine) concentration of lymph node [L-A-HAP(I)], AEF derived from conventional CT (AEFC), unenhanced CT value (U-CT)}. Spectrum model could improve diagnostic performances compared to conventional model (area under curve: 0.965 vs 0.916, p= 0.038). CONCLUSION: There was a moderate correlation between spectrum parameters and conventional parameters. Integrating conventional CT features with spectrum parameters could further improve the ability in differentiating TB-associated FM from EBLC.

Chest X-ray evaluation using machine learning to support the early diagnosis of pulmonary TB.

TB is a public health problem, and its diagnosis can be challenging. Among imaging methods, chest X-ray (CXR) is the leading choice for assessing pulmonary TB (PTB). Recent advancements in the field of artificial intelligence have stimulated studies evaluating the performance of machine learning (ML) for medical diagnosis. This study validated a new original Brazilian tool, the XmarTB, applied to CXR images to support the early diagnosis of PTB.An ML model was trained on 3,800 normal images, 3,800 abnormal CXRs without PTB and 1,376 with PTB manifestations from the publicly available TBX11K database.The binary classification can distinguish between normal and abnormal CXR with a sensitivity of 99.4% and specificity of 99.4%. The XmarTB tool had a sensitivity of 98.1% and a specificity of 99.7% in detecting TB cases among CXRs with abnormal CXRs; sensitivity was 96.7% and specificity 98.7% in detecting TB cases among all samples.This diagnostic tool can accurately and automatically detect abnormal CXRs and satisfactorily differentiate PTB from other pulmonary diseases. This tool holds significant promise in aiding the proactive detection of TB cases, providing rapid and accurate support for early diagnosis..

Patient delay of patients with pulmonary tuberculosis in Yantai, China, 2013 to 2022: Analysis of a case series.

Patient delay increases the morbidity and mortality due to tuberculosis (TB). This study aimed to assess patient delay among patients with pulmonary tuberculosis in Yantai from 2013 to 2022, and to analyze factors related to patient delay. Data of patients with pulmonary tuberculosis in Yantai City from 2013 to 2022 were obtained from the Tuberculosis Management Information System of the Chinese Disease Prevention and Control System. Statistical analyses were performed using the SPSS.26.0 software. The trend in patient delay rate was tested using the chi-square trend test. Univariate analyses were performed using the chi-square test, and factors with statistically significant differences in the univariate analysis were included in the binary logistic regression analysis to identify the factors affecting patient delay. Patient delay was defined as an interval of more than 14 days between the onset of clinical symptoms and the patient first visit to a healthcare facility. From 2013 to 2022, the median delay time for patients with pulmonary tuberculosis in Yantai was 28 ±â€…52 days and the patient delay rate was 69.5%. There was an overall increasing trend in the rate of patient delay as the number of years increased. Univariate analyses revealed statistically significant differences in patient delay in terms of age, occupation, patient source, domicile, pathogenetic results, and the presence of comorbidities (all P < .05). The results of logistic regression analysis showed that the age was 20 to 39, 40 to 59, and ≥ 60 years (OR = 1.365, 95%CI: 1.156-1.612; OR = 1.978, 95%CI: 1.660-2.356; OR = 1.767, 95%CI: 1.480-2.110), occupation was domestic and un-employed (OR = 1.188, 95%CI: 1.071-1.317), domicile as mobile population (OR = 1.212, 95%CI: 1.099-1.337), and positive pathogenic results (OR = 1.242, 95%CI: 1.015-1.520) were risk factors for patient delay. Patient delays were serious among pulmonary tuberculosis patients in Yantai City, 2013 to 2022, and patient delay was related to factors such as age, occupation, domicile, patient source, and pathogenetic results.

Adapted generative latent diffusion models for accurate pathological analysis in chest X-ray images.

Respiratory diseases have a significant global impact, and assessing these conditions is crucial for improving patient outcomes. Chest X-ray is widely used for diagnosis, but expert evaluation can be challenging. Automatic computer-aided diagnosis methods can provide support for clinicians in these tasks. Deep learning has emerged as a set of algorithms with exceptional potential in such tasks. However, these algorithms require a vast amount of data, often scarce in medical imaging domains. In this work, a new data augmentation methodology based on adapted generative latent diffusion models is proposed to improve the performance of an automatic pathological screening in two high-impact scenarios: tuberculosis and lung nodules. The methodology is evaluated using three publicly available datasets, representative of real-world settings. An ablation study obtained the highest-performing image generation model configuration regarding the number of training steps. The results demonstrate that the novel set of generated images can improve the performance of the screening of these two highly relevant pathologies, obtaining an accuracy of 97.09%, 92.14% in each dataset of tuberculosis screening, respectively, and 82.19% in lung nodules. The proposal notably improves on previous image generation methods for data augmentation, highlighting the importance of the contribution in these critical public health challenges.

Symbolic Artificial Intelligence to Diagnose Tuberculosis Using Ontology.

Pulmonary Tuberculosis (PTB) is an infectious disease caused by a bacterium called Mycobacterium tuberculosis. This paper aims to create Symbolic Artificial Intelligence (SAI) system to diagnose PTB using clinical and paraclinical data. Usually, the automatic PTB diagnosis is based on either microbiological tests or lung X-rays. It is challenging to identify PTB accurately due to similarities with other diseases in the lungs. X-ray alone is not sufficient to diagnose PTB. Therefore, it is crucial to implement a system that can diagnose based on all paraclinical data. Thus, we propose in this paper a new PTB ontology that stores all paraclinical tests and clinical symptoms. Our SAI system includes domain ontology and a knowledge base with performance indicators and proposes a solution to diagnose current and future PTB also abnormal patients. Our approach is based on a real database of more than four years from our collaborators at Pondicherry hospital in India.

Clinical and Imaging Characteristics of Smear Negative Pulmonary Tuberculosis Patients: A Comparative Study.

Background: Prevalence surveys in Ethiopia indicate smear negative pulmonary tuberculosis (SNPTB) taking the major share of the overall TB burden. It has also been a diagnostic dilemma worldwide leading to diagnostic delays and difficulty in monitoring treatment outcomes. This study determines and compares the clinical and imaging findings in SNPTB and smear positive PTB (SPPTB). Methodology. A case-control study was conducted on 313 PTB (173 SNPTB) patients. Data and sputum samples were collected from consented patients. Smear microscopy, GeneXpert, and culture analyses were performed on sputum samples. Data were analyzed using Stata version 17; a P value < 0.05 was considered statistically significant. Results: Of the 173 SNPTB patients, 42% were culture positive with discordances between test results reported by health facilities and Armauer Hansen Research Institute laboratory using concentrated smear microscopy. A previous history of TB and fewer cavitary lesions were significantly associated with SNPTB. Conclusions: Though overall clinical presentations of SNPTB patients resemble those seen in SPPTB patients, a prior history of TB was strongly associated with SNPTB. Subject to further investigations, the relatively higher discrepancies seen in TB diagnoses reflect the posed diagnostic challenges in SNPTB patients, as a higher proportion of these patients are also seen in Ethiopia.

Automatic bright-field smear microscopy for diagnosis of pulmonary tuberculosis.

In recent decades, many studies have been published on the use of automatic smear microscopy for diagnosing pulmonary tuberculosis (TB). Most of them deal with a preliminary step of the diagnosis, the bacilli detection, whereas sputum smear microscopy for diagnosis of pulmonary TB comprises detecting and reporting the number of bacilli found in at least 100 microscopic fields, according to the 5 grading scales (negative, scanty, 1+, 2+ and 3+) endorsed by the World Health Organization (WHO). Pulmonary TB diagnosis in bright-field smear microscopy, however, depends upon the attention of a trained and motivated technician, while the automated TB diagnosis requires little or no interpretation by a technician. As far as we know, this work proposes the first automatic method for pulmonary TB diagnosis in bright-field smear microscopy, according to the WHO recommendations. The proposed method comprises a semantic segmentation step, using a deep neural network, followed by a filtering step aiming to reduce the number of false positives (false bacilli): color and shape filtering. In semantic segmentation, different configurations of encoders are evaluated, using depth-wise separable convolution layers and channel attention mechanism. The proposed method was evaluated with a large, robust, and annotated image dataset designed for this purpose, consisting of 250 testing sets, 50 sets for each of the 5 TB diagnostic classes. The following performance metrics were obtained for automatic pulmonary TB diagnosis by smear microscopy: mean precision of 0.894, mean recall of 0.896, and mean F1-score of 0.895.

National cross-sectional cluster survey of tuberculosis prevalence in Timor-Leste: a study protocol.

INTRODUCTION: Timor-Leste has one of the world's highest estimated tuberculosis (TB) incidences, yet the data which informs this estimate is limited and the true burden of TB disease is not known. TB prevalence surveys offer the best means of determining robust estimates of disease burden. This study aims to provide an estimate of the prevalence of bacteriologically confirmed pulmonary TB in Timor-Leste and provide additional insights into diagnostic coverage and health-seeking behaviour of persons with symptoms suggestive of TB. METHODS AND ANALYSIS: A national population-based cross-sectional cluster survey will be conducted in which participants aged 15 years and older will be screened for pulmonary TB using an algorithm consisting of symptom screening and digital X-ray of the chest with computer-aided detection software for X-ray interpretation. Xpert Ultra and liquid culture methods will be used to confirm survey TB cases. Additional data will be collected from persons reporting symptoms suggestive of TB to assess health-seeking behaviour and access to TB diagnosis and care. The survey aims to screen a target sample population of 20 068 people, living within 50 clusters, representing every municipality of Timor-Leste. Bacteriologically confirmed pulmonary TB prevalence will be estimated using WHO-recommended methods. ETHICS AND DISSEMINATION: Research ethics approval has been granted by the human research ethics committee of the Northern Territory, Australia, and the Instituto Nacional da Saúde, Timor-Leste. The results will be published in a peer-reviewed scientific journal and disseminated with relevant stakeholders. TRIAL REGISTRATION NUMBER: ACTRN12623000718640.

TB or not TB? Diagnostic Sensitivity, Specifity and Interobserver Agreement in the Radiological Diagnosis of Pulmonary Tuberculosis in Children.

BACKGROUND: The differentiation between latent tuberculosis infection (LTBI) and tuberculosis (TB) relies on radiological changes. Confirming the diagnosis remains a challenge because typical findings are often missing in children. This study evaluates diagnostic sensitivity, specifity and interobserver agreement on the radiological diagnosis of TB by chest-x-rays in accordance to professional specialization and work experience. METHODS: Chest x-rays of 120 children with proven tuberculosis infection were independently evaluated by general radiologists, paediatric radiologists and paediatric pulmonologists. Results were compared to a reference diagnosis created by group of experienced paediatric radiologists and paediatric pulmonologists. Primary endpoints were diagnostic sensitivity and specificity and interobserver variability defined as Krippendorfs alpha of thesel groups compared to the reference diagnosis. RESULTS: Of the 120 chest x-rays 33 (27,5%) were diagnosed as TB by the reference standard . Paediatric pulmonologist had the highest diagnostic sensitivity (90%) but were less specific (71%) whereas general radiologist were less sensitive (68%) but more secific (95%). The best diagnostic accuracy was achieved by pediatric radiologists with a diagnostic sensitivity of 77% and specificity 95% respectively. CONCLUSIONS: We demonstrated significant interobserver variability and relevant differences in sensitivity and specificity in the radiological diagnosis of TB between the groups. Paediatric radiologists showed the best diagnostic performance. As the diagnosis of pulmonary TB has significant therapeutic consequences for children they should be routinely involved in the diagnostic process.

Diagnosis of pulmonary tuberculosis with 3D neural network based on multi-scale attention mechanism.

This paper presents a novel multi-scale attention residual network (MAResNet) for diagnosing patients with pulmonary tuberculosis (PTB) by computed tomography (CT) images. First, a three-dimensional (3D) network structure is applied in MAResNet based on the continuity and correlation of nodal features on different slices of CT images. Secondly, MAResNet incorporates the residual module and Convolutional Block Attention Module (CBAM) to reuse the shallow features of CT images and focus on key features to enhance the feature distinguishability of images. In addition, multi-scale inputs can increase the global receptive field of the network, extract the location information of PTB, and capture the local details of nodules. The expression ability of both high-level and low-level semantic information in the network can also be enhanced. The proposed MAResNet shows excellent results, with overall 94% accuracy in PTB classification. MAResNet based on 3D CT images can assist doctors make more accurate diagnosis of PTB and alleviate the burden of manual screening. In the experiment, a called Grad-CAM was employed to enhance the class activation mapping (CAM) technique for analyzing the model's output, which can identify lesions in important parts of the lungs and make transparent decisions.