evidoria

← Back to browse

Good practice Imported

AI Computer-Aided Detection for Tuberculosis Screening — Cambodia's CENAT Chest X-ray Programme

Cambodia · Phnom Penh · See the Cambodia profile · See the Phnom Penh profile

Evidence: Observational / pre–post Top 25% 73/100 · Ask Evidence Copilot about this practice

Cambodia's national TB programme uses AI computer-aided detection to read portable chest X-rays in the field. A peer-reviewed active case-finding study found the AI reached 0.86 AUC-ROC and caught 97.7% of confirmed TB cases in a screened cohort of over 8,000 people.

8,386
Cohort screened by chest X-ray
130 1.6%
Confirmed TB-positive (Xpert)
0.86
AI-CAD AUC-ROC
97.7%
Sensitivity at sensitive threshold
96.2%
Sensitivity at stricter triage threshold
~45-50% per decade
National TB prevalence decline
~500,000
People cured since 2000 (whole TB programme)
AI Computer-Aided Detection for Tuberculosis Screening — Cambodia's CENAT Chest X-ray Programme

Details

Maturity
Established
Promoter
National Center for Tuberculosis and Leprosy Control (CENAT), Cambodia Ministry of Health, with WHO, the Global Fund, USAID, RIT/JATA and Fujifilm
Period
Ongoing; national prevalence survey 2023, active case-finding study published 2024
Keywords
tuberculosis screening, computer-aided detection, AI radiology, digital chest X-ray, active case finding

Context

Cambodia's National Center for Tuberculosis and Leprosy Control (CENAT), under the Ministry of Health, uses AI computer-aided detection software on portable digital chest X-ray units, alongside GeneXpert sputum testing, to screen for TB in hard-to-reach communities, with support from WHO, the Global Fund, USAID and JATA/RIT Japan.

Results

A peer-reviewed field study screened 8,386 people, of whom 130 (1.6%) were confirmed TB-positive by Xpert testing; the AI-CAD system achieved an AUC-ROC of 0.86 (95% CI 0.83-0.89), detecting 97.7% of confirmed cases at a sensitive threshold and 96.2% at a stricter triage threshold. National TB prevalence has fallen roughly 45-50% per decade across surveys, with CENAT crediting close to 500,000 people cured since 2000 (whole-programme figure, not AI-CAD alone).

Implementation

Implementation detail (cost, timeline, staffing, conditions for success) is not yet available for this practice.

Commonly funded by

National / regional programmes

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful