CoChat

Accuracy of computer-aided chest X-ray in community-based tuberculosis screening: Lessons from the 2016 Kenya National Tuberculosis Prevalence Survey

Open access

The DOI matches its Crossref record; 26 of the 26 references it cites were found in a registry or an index; 10 of the 20 sources its claims cite were checked for support.

Abstract

Community-based screening for tuberculosis (TB) could improve detection but is resource intensive. We set out to evaluate the accuracy of computer-aided TB screening using digital chest X-ray (CXR) to determine if this approach met target product profiles (TPP) for community-based screening. CXR images from participants in the 2016 Kenya National TB Prevalence Survey were evaluated using CAD4TBv6 (Delft Imaging), giving a probabilistic score for pulmonary TB ranging from 0 (low probability) to 99 (high probability). We constructed a Bayesian latent class model to estimate the accuracy of CAD4TBv6 screening compared to bacteriologically-confirmed TB across CAD4TBv6 threshold cut-offs, incorporating data on Clinical Officer CXR interpretation, participant demographics (age, sex, TB symptoms, previous TB history), and sputum results. We compared model-estimated sensitivity and specificity of CAD4TBv6 to optimum and minimum TPPs. Of 63,050 prevalence survey participants, 61,848 (98%) had analysable CXR images, and 8,966 (14.5%) underwent sputum bacteriological testing; 298 had bacteriologically-confirmed pulmonary TB. Median CAD4TBv6 scores for participants with bacteriologically-confirmed TB were significantly higher (72, IQR: 58-82.75) compared to participants with bacteriologically-negative sputum results (49, IQR: 44-57, p<0.0001). CAD4TBv6 met the optimum TPP; with the threshold set to achieve a mean sensitivity of 95% (optimum TPP), specificity was 83.3%, (95% credible interval [CrI]: 83.0%-83.7%, CAD4TBv6 threshold: 55). There was considerable variation in accuracy by participant characteristics, with older individuals and those with previous TB having lowest specificity. CAD4TBv6 met the optimal TPP for TB community screening. To optimise screening accuracy and efficiency of confirmatory sputum testing, we recommend that an adaptive approach to threshold setting is adopted based on participant characteristics.

Overview

CoChat Summary

TL;DR

This retrospective analysis of the 2016 Kenya National TB Prevalence Survey found that CAD4TBv6 performed well for community TB screening and met the paper’s target product profile at an overall threshold of 55. Accuracy varied by participant characteristics, with lower specificity in older people and those with previous TB, supporting adaptive thresholding.

Key findings

  • Of 62,484 CXR images uploaded, 61,848 (99.0%) were analysable and included; 8,996/61,848 (14.5%) had sputum collected, and 298 had bacteriologically-confirmed TB.
  • The median CAD4TBv6 score was higher in bacteriologically-confirmed TB than in bacteriologically-negative sputum results: 72 vs 49.
  • Overall model estimates for CAD4TBv6 were AUC 96.7% (95% CrI 95.9%–97.4%); at a sensitivity target of 95% (optimum TPP), specificity was 83.2%–83.3% and the threshold was 55.
  • Accuracy varied by subgroup: sensitivity was highest among those aged 41 years and older, those with cough >2 weeks, and those with previous TB; specificity was lowest among older participants and those with previous TB. Clinical Officer CXR interpretation as “suggestive of TB” had overall sensitivity 43.7% and specificity 89.2% for bacteriologically-confirmed TB.

Methods

  • Retrospective analysis of cross-sectional individual-level data from adults (aged 15+ years) in the 2016 Kenya National TB Prevalence Survey, a population-based national survey conducted in 2015–2016 across 100 clusters.
  • Analysed 61,848 participants with CXR images; CAD4TBv6 processed anonymised DICOM CXR images and generated scores from 0 to 99. Reference standard was bacteriologically-confirmed TB defined as Xpert and/or culture positive with MTB speciation.
  • Primary analysis used Bayesian latent class modelling in Stan/R to estimate TB prevalence and CAD4TBv6 sensitivity, specificity, and AUC across thresholds 0–99; secondary analyses stratified by age, sex, cough >2 weeks, and previous TB history.

Limitations

  • The paper states it included participants aged 15 years and above, so it could not assess CAD4TBv6 accuracy in children.
  • The study only included bacteriologically-confirmed TB and did not assess accuracy in sputum-negative, clinically diagnosed TB; the paper states this explicitly.
  • HIV status could not be used for stratified analysis because testing was not systematically conducted and there was substantial missing data; this is stated in the paper.
  • Evident from the design: sputum collection was conditional on symptoms or abnormal CXR, so latent class modelling was needed and some reference-standard information was missing for part of the cohort.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1371/journal.pgph.0001272

Cites

26 of 26 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Fig 1 Participant flowchart and results of prevalence survey investigations.
Fig 1 Participant flowchart and results of prevalence survey investigations.CC BY · PubMed Central
Fig 2 Distribution of CAD4TBv6 scores in Kenya National TB prevalence survey. A) Distribution (median and 95% highest density interval) of CAD4TBv6 scores by whether prevalence survey participant’s sputum was tested or not. B) Distribution (median and 95% highest density interval) of CAD4TBv6 scores by sputum bacteriological results. 95%HDI: 95% highest density interval.
Fig 2 Distribution of CAD4TBv6 scores in Kenya National TB prevalence survey. A) Distribution (median and 95% highest density interval) of CAD4TBv6 scores by whether prevalence survey participant’s sputum was tested or not. B) Distribution (median and 95% highest density interval) of CAD4TBv6 scores by sputum bacteriological results. 95%HDI: 95% highest density interval.CC BY · PubMed Central
Fig 3 Model-based sensitivity and specificity of CAD4TBv6 for bacteriologically-confirmed pulmonary TB at minimum and optimum target product profile thresholds.
Fig 3 Model-based sensitivity and specificity of CAD4TBv6 for bacteriologically-confirmed pulmonary TB at minimum and optimum target product profile thresholds.CC BY · PubMed Central
Fig 4 Sensitivity and specificity of CAD4TBv6 by prevalence survey participant characteristics, with threshold set at optimal target product profile to achieve overall sensitivity of 95% (CAD4TBv6 = 55).
Fig 4 Sensitivity and specificity of CAD4TBv6 by prevalence survey participant characteristics, with threshold set at optimal target product profile to achieve overall sensitivity of 95% (CAD4TBv6 = 55).CC BY · PubMed Central

Verification

Claims the paper makes

7 Valid · 0 Not valid · 4 Not checked · 41 Own results and interpretations

27 more claims

Create a free account to see all 52 claims and to verify the 3 references still open.

See all claims — free account
  • Probably not a substantial claim. 8
  • It may state two separate claims as one. 17

Create a free account to see these claims.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

1 of 1

Citations

Cites

Research works printed but not linked: 9

  1. World Health Organization, 2020, Global tuberculosis report 2020
  2. World Health Organization, 2020, Tuberculosis key facts
  3. World Health Organization, 2015, The END TB Strategy
  4. World Health Organization, 2014, High-priority target product profiles for new tuberculosis diagnostics: report of a consensus meeting.
  5. World Health Organization, 2011, Tuberculosis prevalence surveys: a handbook.
  6. DI Systems, 2019, Xray systems
  7. World Health Organization, 1968, Principles and practice of screening for disease / J. M. G. Wilson, G. Jungner
  8. D. Imaging, 2021, Delft Imaging
  9. B Hochhegger, 2012, The chest and aging: radiological findings., Jornal brasileiro de pneumologia: publicacao oficial da Sociedade Brasileira de Pneumologia e Tisilogia, 38, 656

Cited by

See all 25

Metadata

Authors
Mungai, B., Ong‘angò, J., Ku, C. C., Henrion, M. Y. R., Morton, B., Joekes, E., Onyango, E., Kiplimo, R., Kirathe, D., Masini, E., Sitienei, J., Manduku, V., Mugi, B., Squire, S. B., MacPherson, P., & The IMPALA Consortium
Affiliations
  • Liverpool School of Tropical Medicine (GB)
  • Imperial College London (GB)
  • Global Fund to Fight AIDS, Tuberculosis and Malaria (CH)
  • Kenya Medical Research Institute (KE)
  • Kenyatta National Hospital (KE)
Corresponding author
Mungai, Brenda
Venue
PLOS Global Public Health
Published
2022-11-23
Volume
2
Issue
11
Pages
e0001272
ISSN
2767-3375 (ISSN-L)
OpenAlex
W4309796546
Cited by
24
References
26
Citations by year
2026: 5 · 2025: 8 · 2024: 7 · 2023: 3 · 2022: 1
FWCI
2.32
Type
journal article
Language
en
Open access
Locations
  • PLOS Global Public Health (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • LSTM Online Archive (Liverpool School of Tropical Medicine) (acceptedVersion)
  • ENLIGHTEN (Jurnal Bimbingan dan Konseling Islam) (acceptedVersion, cc-by-sa)
  • LSHTM Research Online (London School of Hygiene and Tropical Medicine) (acceptedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
  • National Institute for Health Research (16/136/35)
  • Wellcome (200901/Z/16/Z)
  • Malawi Liverpool Wellcome Trust Programme (206545/Z/17/Z)
Topics
  • Health Sciences › Medicine › Infectious Diseases › Tuberculosis Research and Epidemiology
  • Health Sciences › Medicine › Radiology, Nuclear Medicine and Imaging › COVID-19 diagnosis using AI
  • Health Sciences › Medicine › Pulmonary and Respiratory Medicine › Lung Cancer Diagnosis and Treatment
Keywords
tuberculosis screening, computer-aided detection, pulmonary tuberculosis, target product profile, screening accuracy, sensitivity and specificity, Bayesian latent class model, community-based screening

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.