Implementing upfront mobile digital chest x-ray for tuberculosis diagnosis in India—feasibility and benefits

Abstract Background The Tuberculosis (TB) Control Program in India changed the TB diagnostic algorithm and recommended sputum testing and chest x-ray (CXR) for presumptive TB up front. There is no experience of testing this algorithm in routine field settings. Methods In a public–private partnership (PPP), a private hospital provided mobile digital CXR services (mounted on a van) to complement the existing diagnostic services of sputum microscopy and GeneXpert testing. All presumptive TB patients (cough >2 weeks) underwent CXR and sputum microscopy, and GeneXpert testing if eligible (smear-negative CXR suggestive of TB). Results All 2973 presumptive TB patients underwent CXR and sputum microscopy; 471 (15.8%) had abnormal CXR findings suggestive of TB, 129 (4.3%) were smear positive and 17 were extrapulmonary TB. Of the remaining 325 with smear-negative and CXR suggestive of TB, 147 (45.2%) underwent GeneXpert testing, yielding 32 positives (21.8%). Of the remaining 178 with no GeneXpert test done, 106 (60.0%) had CXR definitely suggesting TB (clinically diagnosed TB). Thus a total of 284 cases of TB (161 microbiologically confirmed, 106 clinically diagnosed, 17 extrapulmonary TB) were identified, giving a potential diagnostic yield of 19.6%. Conclusions Systematic screening with mobile digital X-ray service via a PPP model integrated into the national program is feasible and scalable with a high yield.

Download Full-text

Early chest X-ray in persons with presumptive tuberculosis increases Xpert® MTB/RIF diagnostic yield and efficiency

Public Health Action ◽

10.5588/pha.19.0052 ◽

2020 ◽

Vol 10 (1) ◽

pp. 17-20

Author(s):

Z. Nadiah ◽

R. C. Koesoemadinata ◽

S. M. McAllister ◽

G. Putriyani ◽

L. Chaidir ◽

...

Keyword(s):

Mycobacterium Tuberculosis ◽

Diagnostic Yield ◽

Content Type ◽

X Ray ◽

Acid Fast Bacilli ◽

Number Needed To Screen ◽

Sputum Microscopy ◽

Chest X Ray

Adult presumptive tuberculosis (TB) patients (n = 1690) were screened for TB using a questionnaire, chest X-ray (CXR) and sputum microscopy for acid-fast bacilli (AFB); <named-content content-type="genus-species">Mycobacterium tuberculosis</named-content> culture was performed for 74% of the patients and Xpert® MTB/RIF was done for 17.2%. Among patients recruited, 943 (55.8%) were diagnosed with TB, of whom 870 (92.3%) were bacteriologically confirmed and 73 (7.7%) were clinically diagnosed on the basis of CXR. Using CXR prior to culture or Xpert testing reduces the number needed to screen from 7.6 to 5.0. Using CXR to triage for culture or Xpert testing reduces the number of missed cases and increases the efficiency of culture and Xpert testing.

Download Full-text

Compliance of St Joseph’s Hospital Roma, Lesotho with the National Tuberculosis Programme of Lesotho, 2007 and 2008

African Journal of Primary Health Care & Family Medicine ◽

10.4102/phcfm.v6i1.586 ◽

2014 ◽

Vol 6 (1) ◽

Author(s):

Oladoyinbo O. Samuel ◽

Pierre J.T. De Villiers

Keyword(s):

Treatment Failure ◽

Sputum Smear ◽

Smear Microscopy ◽

Sputum Smear Microscopy ◽

X Ray ◽

National Tuberculosis Programme ◽

National Tuberculosis ◽

Tuberculosis Programme ◽

Chest X Ray ◽

Smear Negative

Background: In 2009 Lesotho had an estimated TB prevalence of 696 cases/100 000 population − the 4th highest in the world. This epidemic was characterised by high rates of death, treatment failure and unknown treatment outcomes. These adverse outcomes were attributable to a high rate of TB and/or HIV co-infection and weaknesses in the implementation of Lesotho’s National Tuberculosis Programme (NTP). This study was conducted in St Joseph’s Hospital, Roma (SJHR) to assess the implementation of the NTP.Method: Records of 993 patients entered into the SJHR TB register between 2007 and 2008 were reviewed. Patients’ treatment details were extracted from the register, validated and analysed by STATA 10.0.Results: Of 993 patients registered: 88% were new patients, 37% were diagnosed on sputum smear microscopy alone, 35% were diagnosed on sputum smear microscopy with chest X-ray, whilst 25% were diagnosed on chest X-ray alone. In addition: 33% were sputum smear positive, 45% were sputum smear negative, and 22% had extra-pulmonary TB. As to treatment outcome: 26% were cured, 51% completed treatment, and 51% converted from sputum smear positive to sputum smear negative over six months, whilst 16% died. Regarding HIV, 77% of patients were tested for HIV and 59% had TB and/or HIV co-infection. Of ten NTP targets only the defaulter and treatment failure rate targets were met.Conclusion: Whilst only two out of ten NTP targets were met at SJHR in 2007–2008, improvements in TB case management were noted in 2008 which were probably due to the positive effects of audit on staff performance.

Download Full-text

Chest X-ray vs. Xpert® MTB/RIF assay for the diagnosis of sputum smear-negative tuberculosis in Uganda

The International Journal of Tuberculosis and Lung Disease ◽

10.5588/ijtld.13.0464 ◽

2014 ◽

Vol 18 (2) ◽

pp. 216-219 ◽

Cited By ~ 5

Author(s):

C. Wekesa ◽

B. J. Kirenga ◽

M. L. Joloba ◽

F. Bwanga ◽

A. Katamba ◽

...

Keyword(s):

Sputum Smear ◽

X Ray ◽

Chest X Ray ◽

Smear Negative

Download Full-text

Automated chest radiograph diagnosis: A Twofer for Tuberculosis and Covid-19

10.1101/2020.10.13.20178483 ◽

2020 ◽

Author(s):

Mitushi Verma ◽

Deepak Patkar ◽

Madhura Ingalharikar ◽

Amit Kharat ◽

Pranav Ajmera ◽

...

Keyword(s):

Artificial Intelligence ◽

X Ray ◽

Recent Onset ◽

Radiology Reports ◽

Fast Spreading ◽

Digital Chest ◽

Chest X Ray ◽

End To End ◽

The Times

AbstractCoronavirus disease (Covid 19) and Tuberculosis (TB) are two challenges the world is facing. TB is a pandemic which has challenged mankind for ages and Covid 19 is a recent onset fast spreading pandemic. We study these two conditions with focus on Artificial Intelligence (AI) based imaging, the role of digital chest x-ray and utility of end to end platform to improve turnaround times. Using artificial intelligence assisted technology for triage and creation of structured radiology reports using an end to end platform can ensure quick diagnosis. Changing dynamics of TB screening in the times of Covid 19 pandemic have resulted in bottlenecks for TB diagnosis. The paper tries to outline two types of use cases, one is COVID-19 screening in a hospital-based scenario and the other is TB screening project in mobile van setting and discusses the learning of these models which have both used AI for prescreening and generating structured radiology reports.

Download Full-text

DIAGNOSTIC YIELD AND CLINICAL IMPACT OF CHEST X-RAY (CXR) AFTER PERCUTANEOUS TRACHEOSTOMY (PT).

Critical Care Medicine ◽

10.1097/00003246-200512002-00621 ◽

2005 ◽

Vol 33 ◽

pp. A176

Author(s):

Samir H Haddad ◽

Abdulaziz Al-Dawood ◽

Yaseen Arabi

Keyword(s):

Diagnostic Yield ◽

Percutaneous Tracheostomy ◽

Clinical Impact ◽

X Ray ◽

Chest X Ray

Download Full-text

The Diagnostic Yield and Clinical Impact of a Chest X-Ray after Percutaneous Dilatational Tracheostomy: A Prospective Cohort Study

Anaesthesia and Intensive Care ◽

10.1177/0310057x0703500313 ◽

2007 ◽

Vol 35 (3) ◽

pp. 393-397 ◽

Cited By ~ 10

Author(s):

S. H. Haddad ◽

A. S. Aldawood ◽

Y. M. Arabi

Keyword(s):

Diagnostic Yield ◽

Perioperative Complications ◽

Percutaneous Dilatational Tracheostomy ◽

Chest Physiotherapy ◽

Radiographic Findings ◽

X Ray ◽

Dilatational Tracheostomy ◽

New Finding ◽

Chest X Ray ◽

Tube Malposition

A chest X-ray (CXR) is routinely performed after percutaneous dilatational tracheostomy (PDT). The purpose of this study was to evaluate the diagnostic yield of routine CXR following PDT and its impact on patient management and to identify predictors of post-PDT CXR changes. Two-hundred-and-thirty-nine patients who underwent PDT in a 21-bed intensive care unit were included prospectively in the study. The following data were collected: patient demographics, APACHE III scores, pre-PDT FiO2 and PEEP, PDT technique, perioperative complications and the use of bronchoscopic guidance. We compared post-PDT CXR with the last pre-PDT CXR. We documented any post-PDT new radiographic findings including atelectasis, pneumothorax, pneumomediastinum, surgical emphysema, pulmonary infiltrates or tracheostomy tube malposition. We also recorded management modifications based on post-PDT radiographic changes, including increased PEEP, chest physiotherapy, therapeutic bronchoscopy or chest tube insertion. Atelectasis was the only new finding detected on post-PDT CXRs of 24 (10%) patients. The new radiographic findings resulted in a total of 14 modifications of management in 10 (4%) patients including increased PEEP in six, chest physiotherapy in six and bronchoscopy in two patients. Trauma and pre-PDT PEEP >5 cmH2O were independent predictors of post-PDT CXR changes. Routine CXR following PDT has a low diagnostic yield, detecting mainly atelectasis and leading to a change in the management in only a minority of patients. Routine CXR after apparently uncomplicated PDT performed by an experienced operator may not be necessary and selective use may improve its diagnostic yield. Further studies are required to validate the safety of selective versus routine post-PDT CXR.

Download Full-text

Lung X-ray Segmentation using Deep Convolutional Neural Networks on Contrast-Enhanced Binarized Images

Mathematics ◽

10.3390/math8040545 ◽

2020 ◽

Vol 8 (4) ◽

pp. 545 ◽

Cited By ~ 2

Author(s):

Hsin-Jui Chen ◽

Shanq-Jang Ruan ◽

Sha-Wo Huang ◽

Yan-Tsung Peng

Keyword(s):

Neural Networks ◽

Convolutional Neural Networks ◽

Deep Convolutional Neural Networks ◽

X Ray ◽

Processing Approach ◽

Segmentation Accuracy ◽

Enhancement Method ◽

Digital Chest ◽

Chest X Ray ◽

Model Training

Automatically locating the lung regions effectively and efficiently in digital chest X-ray (CXR) images is important in computer-aided diagnosis. In this paper, we propose an adaptive pre-processing approach for segmenting the lung regions from CXR images using convolutional neural networks-based (CNN-based) architectures. It is comprised of three steps. First, a contrast enhancement method specifically designed for CXR images is adopted. Second, adaptive image binarization is applied to CXR images to separate the image foreground and background. Third, CNN-based architectures are trained on the binarized images for image segmentation. The experimental results show that the proposed pre-processing approach is applicable and effective to various CNN-based architectures and can achieve comparable segmentation accuracy to that of state-of-the-art methods while greatly expediting the model training by up to 20.74 % and reducing storage space for CRX image datasets by down to 94.6 % on average.

Download Full-text