Synapse
⌘+K
Synapse
PulseExploreJournal ClubResearchersJournals
Instagram
HomeJournal ClubExplore
May 24, 2026Open Access

Performance of CAD4TB artificial intelligence technology in TB screening programmes

View Full Paper
Ask AI
Bookmark
Share

Authors

HCHuman Sciences Research Council

Discussion

Loading...

Member takes

Overview

Randomized trial investigates CAD4TB effectiveness in TB screening, indicating significant variability in performance accuracy.

Key Points

  • This research examines how various factors impact the performance accuracy of CAD4TB in tuberculosis screening.
  • Participants included 6,524 chest images from national prevalence surveys and TB screening studies.
  • Two versions of CAD4TB (v6 and v7) analyzed digital chest radiography results.
  • Threshold and subgroup analyses were performed based on age, X-ray hardware, and HIV status.
  • CAD4TB v7 showed improved performance accuracy compared to v6 (p < 0.01).
  • ROC curve area for v6 was 0.833 (95% CI 0.808–0.859); for v7, it was 0.865 (95% CI 0.842–0.889).
  • At 90% sensitivity, v7 achieved a specificity of 65%, while v6 achieved 54%.

Cite This Study

Human Sciences Research Council (2026) studied this question.

synapsesocial.com/papers/6a12960648a0ea1665672845https://doi.org/10.14749/32361069
View Full Paper
Ask AI
Bookmark
Share