Chest X-ray as an alternative method of making a preliminary diagnosis in patients with susceptible or drug-resistant pulmonary tuberculosis.

Background: Making a preliminary diagnosis using X-ray methods for the study of resistant and resistant tuberculosis (TB) will help to make a preliminary diagnosis and determine further tactics for the treatment of TB, even with limited resources for microbiological diagnosis of drug resistance of TB. The present study was aimed at identifying chest X-ray differences between susceptible and resistant TB. Methods: A prospective cohort study of data from all consecutive patients with culture-confirmed pulmonary TB admitted during the year to the Kharkiv TB Dispensary No. 1 in Kharkiv, Ukraine. Results: One hundred and sixty-eight patients with lung TB were examined. Patients were divided into two groups: 1st patients with pulmonary TB with resistance of Mycobacterium tuberculosis (MTB) to at least isoniazid and rifampicin (resistant TB) and 2nd pulmonary TB with preserved susceptibility of MTB to anti-TB drugs (susceptible-TB). Patients of 1st group often had lesions in two lobes of the lungs 31.1% and one lung 43.3% versus 15.4% and 2.6% of patients with susceptible TB (P < 0.001). In addition, more than 3 cavities in the lungs 45.5% were significantly more often observed in patients with resistant TB versus 7.9%-the 2nd group (P < 0.001). Smaller cavities were observed in patients with susceptible TB up to 1.99 cm 74% versus 35.2% in 1st group (P < 0.001). We did not observe any significant radiological features depending on the right or left lung, as well as the lobar localization of the TB process. Conclusions: For resistant forms of TB, radiologically, a more widespread TB process in the lungs with the presence of a larger number of cavities and their larger size against a background of a more pronounced clinical picture and mycobacterium excretion than with susceptible TB is characteristic.

Radiomics analysis of lung CT for multidrug resistance prediction in active tuberculosis: a multicentre study.

OBJECTIVES: Multidrug-resistant TB (MDR-TB) is a severe burden and public health threat worldwide. This study aimed to develop a radiomics model based on the tree-in-bud (TIB) sign and nodules and validate its predictive performance for MDR-TB. METHODS: We retrospectively recruited 454 patients with proven active TB from two hospitals and classified them into three training and testing cohorts: TIB (n = 295, 102), nodules (n = 302, 97), and their combination (n = 261, 81). Radiomics features relating to TIB and nodules were separately extracted. The maximal information coefficient and recursive feature elimination were used to select informative features per the two signs. Two radiomics models were constructed to predict MDR-TB using a random forest classifier. Then, a combined model was built incorporating radiomics features based on these two signs. The capability of the models in the combined training and testing cohorts was validated with ROC curves. RESULTS: Sixteen features were extracted from TIB and 15 from nodules. The AUCs of the combined model were slightly higher than those of the TIB model in the combined training cohort (0.911 versus 0.877, p > 0.05) and testing cohort (0.820 versus 0.786, p < 0.05) and similar to the performance of the nodules model in the combined training cohort (0.911 versus 0.933, p > 0.05) and testing cohort (0.820 versus 0.855, p > 0.05). CONCLUSIONS: The CT-based radiomics models hold promise for use as a non-invasive tool in the prediction of MDR-TB. CLINICAL RELEVANCE STATEMENT: Our study revealed that complementary information regarding MDR-TB can be provided by radiomics based on the TIB sign and nodules. The proposed radiomics models may be new markers to predict MDR in active TB patients. KEY POINTS: • This is the first study to build, validate, and apply radiomics based on tree-in-bud sign and nodules for the prediction of MDR-TB. • The radiomics model showed a favorable performance for the identification of MDR-TB. • The combined model holds potential to be used as a diagnostic tool in routine clinical practice.

Machine learning and radiomics for the prediction of multidrug resistance in cavitary pulmonary tuberculosis: a multicentre study.

OBJECTIVES: Multidrug-resistant tuberculosis (MDR-TB) is a major challenge to global health security. Early identification of MDR-TB patients increases the likelihood of treatment success and interrupts transmission. We aimed to develop a predictive model for MDR to cavitary pulmonary TB using CT radiomics features. METHODS: This retrospective study included 257 consecutive patients with proven active cavitary TB (training cohort: 187 patients from Beijing Chest Hospital; testing cohort: 70 patients from Infectious Disease Hospital of Heilongjiang Province). Radiomics features were extracted from the segmented cavitation. A radiomics model was constructed to predict MDR using a random forest classifier. Meaningful clinical characteristics and subjective CT findings comprised the clinical model. The radiomics and clinical models were combined to create a combined model. ROC curves were used to validate the capability of the models in the training and testing cohorts. RESULTS: Twenty-one radiomics features were selected as optimal predictors to build the model for predicting MDR-TB. The AUCs of the radiomics model were significantly higher than those of the clinical model in either the training cohort (0.844 versus 0.589, p < 0.05) or the testing cohort (0.829 versus 0.500, p < 0.05). The AUCs of the radiomics model were slightly lower than those of the combined model in the training cohort (0.844 versus 0.881, p > 0.05) and testing cohort (0.829 versus 0.834, p > 0.05), but there was no significant difference. CONCLUSIONS: The radiomics model has the potential to predict MDR in cavitary TB patients and thus has the potential to be a diagnostic tool. KEY POINTS: • This is the first study to build and validate models that distinguish MDR-TB from DS-TB with clinical and radiomics features based on cavitation. • The radiomics model demonstrated good performance and might potentially aid in prior TB characterisation treatment. • This noninvasive and convenient technique can be used as a diagnosis tool into routine clinical practice.

Intelligence Classification Algorithm-Based Drug-Resistant Pulmonary Tuberculosis Computed Tomography Imaging Features and Influencing Factors.

The drug resistance and influencing factors of patients with pulmonary tuberculosis were investigated, and a dual attention dilated residual network (DADRN) algorithm was proposed. The algorithm was applied to process and analyze lung computed tomography (CT) images of 400 included patients with pulmonary tuberculosis. Besides, sparse code book algorithm and bag of visual word (BOVW) algorithms were introduced and compared, and the influencing factors of pulmonary tuberculosis drug resistance were analyzed. The results demonstrated that the localization precision of lung consolidation, nodules, and cavities by the DADRN algorithm reached 91.2%, 92.5%, and 93.8%, respectively. The recall rate of the three algorithms amounted to 83.55%, 84.5%, and 86.4%, respectively. Both localization precision and recall rate of the DADRN algorithm were higher than those of other two algorithms (P < 0.05). The drug resistance rate of streptomycin, isoniazid, and rifampin of the patients aged between 40 and 59 was all higher than those of the patients in other age groups. The drug resistance rate of streptomycin, isoniazid, and rifampin of retreated patients was all higher than those of patients initially treated. The drug resistance rate of streptomycin, isoniazid, and rifampin of the patients with tuberculosis contact was all higher than those of the patients without tuberculosis contact (P < 0.05). Based on the above results, the accuracy of CT images processed by dual attention-based dilated residual classification network algorithm was higher than that processed by other two algorithms. Age, medical history, and history of exposure to tuberculosis were the influencing factors of the drug resistance of patients with pulmonary tuberculosis.

Identifying Drug-Resistant Tuberculosis in Chest Radiographs: Evaluation of CNN Architectures and Training Strategies.

Tuberculosis (TB) is a serious infectious disease that mainly affects the lungs. Drug resistance to the disease makes it more challenging to control. Early diagnosis of drug resistance can help with decision making resulting in appropriate and successful treatment. Chest X-rays (CXRs) have been pivotal to identifying tuberculosis and are widely available. In this work, we utilize CXRs to distinguish between drug-resistant and drug-sensitive tuberculosis. We incorporate Convolutional Neural Network (CNN) based models to discriminate the two types of TB, and employ standard and deep learning based data augmentation methods to improve the classification. Using labeled data from NIAID TB Portals and additional non-labeled sources, we were able to achieve an Area Under the ROC Curve (AUC) of up to 85% using a pretrained InceptionV3 network.

An Automated Approach to Differentiate Drug Resistant Tuberculosis in Chest X-ray Images Using Projection Profiling and Mediastinal Features.

In this study, an attempt has been made to differentiate Drug Resistant Tuberculosis (DR-TB) in chest X-rays using projection profiling and mediastinal features. DR-TB is a condition which is non-responsive to at least one of anti-TB drugs. Mediastinum variations can be considered as significant image biomarkers for detection of DR-TB. Images are obtained from a public database and are contrast enhanced using coherence filtering. Projection profiling is used to obtain the feature lines from which the mediastinal and thoracic indices are computed. Classification of Drug Sensitive (DS-TB) and DR-TB is performed using three classifiers. Results show that the mediastinal features are found to be statistically significant. Support vector machine with quadratic kernel is able to provide better classification performance values of greater than 93%. Hence, the automated analysis of mediastinum could be clinically significant in differentiation of DR-TB.

Low pectoralis muscle index, cavitary nodule or mass and segmental to lobar consolidation as predictors of primary multidrug-resistant tuberculosis: A comparison with primary drug sensitive tuberculosis.

BACKGROUND: The loss of muscle mass in primary multidrug-resistant tuberculosis (MDR-TB) has not been examined in previous studies. This study aimed to investigate that low pectoralis muscle index and characteristic CT features can help differentiate patients with primary MDR-TB from those with drug-sensitive tuberculosis (DS-TB). MATERIAL AND METHODS: From 2010 to 2016, we retrospectively enrolled 90 patients with primary MDR-TB and 90 age- and sex-matched patients with primary DS-TB. The pectoralis muscle mass was quantitatively measured on axial CT images using density histogram analysis. The pectoralis muscle index (PMI) was defined as the pectoralis muscle mass divided by body mass index. We compared the PMI and characteristic CT features of pulmonary tuberculosis between the two groups. RESULTS: Low PMI, segmental to lobar consolidation, cavity in consolidation, cavitary nodule or mass, and bilateral involvement were more frequently observed in patients with MDR-TB than in those with DS-TB. In stepwise multivariate logistic regression analysis, low PMI (odds ratio, 2.776; 95% confidence interval, 1.450-5.314; p = 0.002), segmental or lobar consolidation (odds ratio, 3.123; 95% confidence interval, 1.629-5.987; p = 0.001), and cavitary nodule or mass (odds ratio, 2.790; 95% confidence interval, 1.348-5.176; p = 0.002) were significant factors for MDR-TB. CONCLUSION: Low pectoralis muscle index, segmental to lobar consolidation and cavitary nodule or mass can help differentiate primary MDR-TB from DS-TB.

Bacterial and host determinants of cough aerosol culture positivity in patients with drug-resistant versus drug-susceptible tuberculosis.

A burgeoning epidemic of drug-resistant tuberculosis (TB) threatens to derail global control efforts. Although the mechanisms remain poorly clarified, drug-resistant strains are widely believed to be less infectious than drug-susceptible strains. Consequently, we hypothesized that lower proportions of patients with drug-resistant TB would have culturable Mycobacterium tuberculosis from respirable, cough-generated aerosols compared to patients with drug-susceptible TB, and that multiple factors, including mycobacterial genomic variation, would predict culturable cough aerosol production. We enumerated the colony forming units in aerosols (≤10 µm) from 452 patients with TB (227 with drug resistance), compared clinical characteristics, and performed mycobacterial whole-genome sequencing, dormancy phenotyping and drug-susceptibility analyses on M. tuberculosis from sputum. After considering treatment duration, we found that almost half of the patients with drug-resistant TB were cough aerosol culture-positive. Surprisingly, neither mycobacterial genomic variants, lineage, nor dormancy status predicted cough aerosol culture positivity. However, mycobacterial sputum bacillary load and clinical characteristics, including a lower symptom score and stronger cough, were strongly predictive, thereby supporting targeted transmission-limiting interventions. Effective treatment largely abrogated cough aerosol culture positivity; however, this was not always rapid. These data question current paradigms, inform public health strategies and suggest the need to redirect TB transmission-associated research efforts toward host-pathogen interactions.

A Benchtop Automated Sputum-to-Genotype System Using a Lab-on-a-Film Assembly for Detection of Multidrug-Resistant Mycobacterium tuberculosis.

Automated genotyping of drug-resistant Mycobacterium tuberculosis (MTB) directly from sputum is challenging for three primary reasons. First, the sample matrix, sputum, is highly viscous and heterogeneous, posing a challenge for sample processing. Second, acid-fast MTB bacilli are difficult to lyse. And third, there are hundreds of MTB mutations that confer drug resistance. An additional constraint is that MTB is most prevalent where test affordability is paramount. We address the challenge of sample homogenization and cell lysis using magnetic rotation of an external magnet, at high (5000) rpm, to induce the rotation of a disposable stir disc that causes chaotic mixing of glass beads ("MagVor"). Nucleic acid is purified using a pipet tip with an embedded matrix that isolates nucleic acid ("TruTip"). We address the challenge of cost and genotyping multiple mutations using 203 porous three-dimensional gel elements printed on a film substrate and enclosed in a microfluidic laminate assembly ("Lab-on-a-Film"). This Lab-on-a-Film assembly (LFA) serves as a platform for amplification, hybridization, washing, and fluorescent imaging, while maintaining a closed format to prevent amplicon contamination of the workspace. We integrated and automated MagVor homogenization, TruTip purification, and LFA amplification in a multisample, sputum-to-genotype system. Using this system, we report detection down to 43 cfu/mL of MTB bacilli from raw sputum.

Clinical predictors of drug-resistant tuberculosis in Mexico.

Drug-resistant tuberculosis (DR-TB) remains a major global health problem. Early treatment of TB is critical; in the absence of rapid- susceptibility testing, the empiric selection of drugs should be guided by clinical data. This study aimed to determine the clinical predictors of DR-TB. From September 2010 to August 2017, sociodemographic and clinical characteristics were collected from 144 patients with tuberculosis at the Hospital Civil de Guadalajara, Mexico. Isolates were subjected to drug-susceptibility testing. Clinical predictors of DR-TB were determined using univariate and multivariate analysis. Any drug, isoniazid, and rifampin resistance rates were 47.7, 23.0, and 11.6%, respectively. The visualization of cavities and nodules through either chest radiography or computed tomography were independent predictors of DR-TB. In conclusion, early detection of DR-TB in this population could be based on multiple cavities being observed using chest imaging. This study's results can be applied to future patients with TB in our community to optimize the DR-TB diagnostic process.

TB DEPOT (Data Exploration Portal): A multi-domain tuberculosis data analysis resource.

The NIAID TB Portals Program (TBPP) established a unique and growing database repository of socioeconomic, geographic, clinical, laboratory, radiological, and genomic data from patient cases of drug-resistant tuberculosis (DR-TB). Currently, there are 2,428 total cases from nine country sites (Azerbaijan, Belarus, Moldova, Georgia, Romania, China, India, Kazakhstan, and South Africa), 1,611 (66%) of which are multidrug- or extensively-drug resistant and 1,185 (49%), 863 (36%), and 952 (39%) of which contain X-ray, computed tomography (CT) scan, and genomic data, respectively. We introduce the Data Exploration Portal (TB DEPOT, https://depot.tbportals.niaid.nih.gov) to visualize and analyze these multi-domain data. The TB DEPOT leverages the TBPP integration of clinical, socioeconomic, genomic, and imaging data into standardized formats and enables user-driven, repeatable, and reproducible analyses. It furthers the TBPP goals to provide a web-enabled analytics platform to countries with a high burden of multidrug-resistant TB (MDR-TB) but limited IT resources and inaccessible data, and enables the reusability of data, in conformity with the NIH's Findable, Accessible, Interoperable, and Reusable (FAIR) principles. TB DEPOT provides access to "analysis-ready" data and the ability to generate and test complex clinically-oriented hypotheses instantaneously with minimal statistical background and data processing skills. TB DEPOT is also promising for enhancing medical training and furnishing well annotated, hard to find, MDR-TB patient cases. TB DEPOT, as part of TBPP, further fosters collaborative research efforts to better understand drug-resistant tuberculosis and aid in the development of novel diagnostics and personalized treatment regimens.

Thoracic Radiography Characteristics of Drug Sensitive Tuberculosis and Multi Drug Resistant Tuberculosis: A Study of Indonesian National Tuberculosis Prevalence Survey.

BACKGROUND: Tuberculosis (TB) remains a burden globally, including Indonesia. The primary objective of this study is to reveal the chest radiography characteristic of drug-sensitive TB (DS-TB) and multi-drug resistant TB (MDR-TB) in the Indonesian national tuberculosis prevalence survey 2013-2014. The secondary objective is to explore the correlation and incidence rate of chest radiography lesion of DS-TB and MDR-TB cases. METHODS: This is a cross-sectional retrospective analytical studies with national and regional coverage. Samples were selected by stratified multi-stage clustering sampling technique in a population aged ≥15 years old. The diagnosis of TB was based on culture and GeneXpert tests. RESULTS: There were 147 DS-TB and 11 MDR-TB patients that were analyzed in this study. The nodule is the only type of lesions that distinguish MDR-TB and DS-TB. In multivariate analysis of DS-TB, there were 3 significant chest radiography lesions, i.e infiltrate, cavity and consolidation with odd-ratio (OR) of 14, 13, and 3, respectively. In MDR-TB, the only significant lesion is a nodule, with OR of 19. CONCLUSION: Nodule is the only type of lesions that distinguish MDR-TB and DS-TB. Infiltrate, cavity and consolidation were the types of chest radiography lesions on DS-TB, meanwhile, a nodule was the only significant lesion for MDR-TB.

Spectrum of central nervous system tuberculosis: An experience from a large tertiary care institution of India.

OBJECTIVES: Central nervous system (CNS) is an important site for extrapulmonary tuberculosis. The present study evaluated the spectrum of CNS tuberculosis in a high tuberculosis endemic region. METHODS: The study included 306 cases of CNS tuberculosis. All cases were assessed for clinical evaluation and neuroimaging. All cases were followed up for 3 months. Modified Barthel index was used to assess the outcome. RESULTS: Out of 306 cases of CNS tuberculosis, 174 (56.86%) had intracranial tuberculosis, 55 (17.97%) had spinal tuberculosis, 15 (4.91%) had both intracranial and spinal pathology. Sixty-two (20.26%) patients had disseminated tuberculosis. Two-hundred and fourteen (69.9%) cases had tuberculous meningitis. Disseminated tuberculosis patients had significantly poor modified Barthel index and 3-month outcome. Culture positivity was significantly higher in the disseminated group. Ten (27.02%) out of 37 culture positive tuberculous meningitis cases had multi-drug-resistant tuberculosis. On multivariate analysis disseminated tuberculosis, baseline modified Barthel index ≤12, and stage 3 predicted poor outcome. Fifty-five patients had spinal tuberculosis. Thirty-four (75.56%) patients with Pott's spine improved with antituberculosis treatment and only 11 (24.44%) patients had modified Barthel index ≤12, after 3 months. CONCLUSIONS: In tuberculosis-endemic areas a varied form of CNS tuberculosis is frequent. CNS tuberculosis is often part of disseminated tuberculosis.

Prediction of Multidrug-Resistant TB from CT Pulmonary Images Based on Deep Learning Techniques.

While tuberculosis (TB) disease was discovered more than a century ago, it has not been eradicated yet. Quite contrary, at present, TB constitutes one of the top 10 causes of death and has shown signs of increasing. To complement the conventional diagnostic procedure of applying microbiological culture that takes several weeks and remains expensive, high resolution computer tomography (CT) of pulmonary images has been resorted to not only for aiding clinicians to expedite the process of diagnosis but also for monitoring prognosis when administering antibiotic drugs. This research undertakes the investigation of predicting multidrug-resistant (MDR) patients from drug-sensitive (DS) ones based on CT lung images to monitor the effectiveness of treatment. To contend with smaller data sets (i.e., hundreds) and the characteristics of CT TB images with limited regions capturing abnormities, patch-based deep convolutional neural network (CNN) allied to support vector machine (SVM) classifier is implemented on a collection of data sets from 230 patients obtained from the ImageCLEF 2017 competition. As a result, the proposed architecture of CNN + SVM + patch performs the best with classification accuracy rate at 91.11% (79.80% in terms of patches). In addition, a hand-crafted SIFT based approach accomplishes 88.88% in terms of subject and 83.56% with reference to patches, the highest in this study, which can be explained away by the fact that the data sets are in small numbers. Significantly, during the Tuberculosis Competition at ImageCLEF 2017, the authors took part in the task of classification of 5 types of TB disease and achieved the top one with regard to averaged classification accuracy (i.e., ACC = 0.4067), which is also premised on the approach of CNN + SVM + patch. On the other hand, when the whole slices of 3D TB data sets are applied to train a CNN network, the best result is achieved through the application of CNN coupled with orderless pooling and SVM at 64.71% accuracy rate.

Primary multidrug-resistant tuberculosis versus drug-sensitive tuberculosis in non-HIV-infected patients: Comparisons of CT findings.

BACKGROUND: Multidrug-resistant tuberculosis has emerged as a global threat. The aim of this work was to compare the CT findings of primary multidrug-resistant tuberculosis and drug-sensitive tuberculosis in non-AIDS adults. MATERIAL AND METHODS: From January 2012 to February 2016, 89 patients with primary multidrug-resistant tuberculosis were retrospectively reviewed, and 89 consecutive drug sensitive TB patients with no history of anti-tuberculous chemotherapy from January 2014 to November 2014 were enrolled as control group. All patients were seronegative for HIV. The patients' demographic data and the locations, frequency and patterns of lung lesions on chest CT were compared. RESULTS: Gender and frequency of diabetes were similar between the two groups. The mean age of primary multidrug-resistant tuberculosis patients was younger than that of drug-sensitive tuberculosis (39.0 vs 47.5, P = 0.005). Lung cavitary nodules or masses were more frequently observed and also showed greater extent in primary multidrug-resistant tuberculosis compared with drug-sensitive tuberculosis. The extent of bronchiectasis was significantly greater in primary multidrug-resistant tuberculosis than in drug-sensitive tuberculosis. Calcification, large nodules and calcified lymph nodes were more frequent in drug-sensitive tuberculosis. CONCLUSION: Characteristic chest CT findings may help differentiate between primary multi-drug resistant tuberculosis and drug-sensitive tuberculosis in patients without HIV infection.

Successful Single-Lung Transplantation in a Patient With a Lung Destroyed by Tuberculosis.

In most countries, patients with lungs destroyed by tuberculosis (TB) are excluded from lung transplantation (LTx) because of concerns about TB recurrence. LTx may be an effective therapeutic option for patients with chronic respiratory failure, but there are no reports of successful LTx in patients with lungs destroyed by TB. We present the case of successful single LTx in a patient with chronic respiratory failure after pneumonectomy with antituberculous chemotherapy. At the 16-month follow-up, he did not show any evidence of TB recurrence and his respiratory problems and quality of life were improved by LTx.

Recurrence after Successful Treatment of Multidrug-Resistant Tuberculosis in Taiwan.

Recurrence after successful treatment for multidrug-resistant tuberculosis (MDR-TB) is challenging because of limited retreatment options. This study aimed to determine rates and predictors of MDR-TB recurrence after successful treatment in Taiwan. Recurrence rates were analyzed by time from treatment completion in 295 M DR-TB patients in a national cohort. Factors associated with MDR-TB recurrence were examined using a multivariate Cox regression analysis. Ten (3%) patients experienced MDR-TB recurrence during a median follow-up of 4.8 years. The overall recurrence rate was 0.6 cases per 1000 person-months. Cavitation on chest radiography was an independent predictor of recurrence (adjusted hazard ratio [aHR] = 6.3; 95% CI, 1.2-34). When the analysis was restricted to 215 patients (73%) tested for second-line drug susceptibility, cavitation (aHR = 10.2; 95% CI, 1.2-89) and resistance patterns of extensively drug-resistant TB (XDR-TB) or pre-XDR-TB (aHR = 7.3; 95% CI, 1.2-44) were associated with increased risk of MDR-TB recurrence. In Taiwan, MDR-TB patients with cavitary lesions and resistance patterns of XDR-TB or pre-XDR-TB are at the highest risk of recurrence. These have important implications for MDR-TB programs aiming to optimize post-treatment follow-up and early detection of recurrent MDR-TB.