WHO recommends AI reading of chest X-rays for TB screening

WHO's guideline group found that three CAD programs read chest X-rays for TB about as accurately as human readers. In March 2021 WHO conditionally recommended them for people aged 15 and older.

  • Not rated an organization

    The evaluations measured accuracy, and WHO reports no screening program's results with the software.

Depth runs from 1, a convenience or time saved, to 5, a life or a livelihood preserved.

Status
Reported benefit.
Evidence

foundstudiedreportedclaimed

An independent assessment report supports the benefit.

Compared with

Human readers interpreting the same chest X-rays (screening use case). Measured.

Without the system
Sensitivity 0.82–0.93, specificity 0.14–0.63
With it
Sensitivity 0.90–0.92, specificity 0.23–0.66
The system’s part
The evaluations compared the software's accuracy with that of human readers interpreting the same chest X-rays, against a microbiological reference standard.
Limits
Not yet checked.
Costs and harms
Not yet checked.
Later
Not yet checked.

Download data Cite Suggest a correction

Benefit type
Increased Scalability
System
CAD4TB, Lunit INSIGHT CXR and qXR (computer-aided detection software for TB)
Made by
Delft Imaging (CAD4TB), Lunit (Lunit INSIGHT CXR) and Qure.ai (qXR)
Used by
TB screening and triage studies in three independent evaluations reviewed by WHO
Where
Global
Stage
Tested in a laboratory
Date
, when the source was published
Kind
Aggregate result
Broad benefit
Identifying Objects, Actions, and Events
Collections
Scrutiny and findings
Last revised
31 October 2026

Sources

  1. Assessment report, by authors with no tie to the system.

    World Health Organization. WHO consolidated guidelines on tuberculosis. Module 2, screening: systematic screening for tuberculosis disease. Geneva, 22 March 2021. Recommendation 10 and section 3.2.1. ISBN 978-92-4-002267-6.

    In comparing the range of accuracy of CAD to that of human readers interpreting CXRs and noting the variability of readers and the substantial overlap between the two ranges, the data suggested there is little difference between the two. Therefore, the GDG considered that CAD software programmes can be considered accurate when compared with human readers.

Details

AI system

Reported configuration

CAD4TB v6 and qXR v2 (no version given for Lunit INSIGHT CXR)

Workflow

Each program was set to a threshold giving 90% sensitivity for pulmonary TB. The evaluations used image libraries never made available for the software's development or training.

Follow-up

Conditions for replication

Each program calibrates the threshold for its setting and population. The software reads posteroanterior or anteroposterior views of digital chest X-rays.

Entry details

Sector

Healthcare (tuberculosis screening).

Purpose

Delft Imaging, Lunit and Qure.ai built CAD4TB, Lunit INSIGHT CXR and qXR to read digital chest X-rays and score the likelihood of tuberculosis.

Not yet checked: model, provider, role, configuration, inputs and outputs, autonomy, human oversight, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.

Cite this record

Peter H. Vartanian. (2026). WHO recommends AI reading of chest X-rays for TB screening (AIBD-0734). Artificial Intelligence Benefits Database, edition of 31 October 2026. https://petervartanian.xyz/aibd/records/AIBD-0734.html (archived at https://doi.org/10.7910/DVN/AADSII, version 4.0)

BibTeX Data Field definitions