Research & Evidence
Explore scientific research and published studies involving MinXray technology and portable radiography. See how our systems are being evaluated and applied in demanding clinical and real-world environments.
Cost Effectiveness of Active Case Finding with CXR
Achieving a “step change” in the tuberculosis epidemic through comprehensive community-wide intervention: a model-based analysis
This modeling analysis shows that an intensive, short-term TB intervention combining active screening, latent treatment, and system strengthening can cut 10-year TB incidence by over 40% and mortality by nearly two-thirds.
Modeling the impact of case finding for tuberculosis: The role of infection dynamics
We compared models assuming lifelong infection risk ("Conventional") versus infection clearance ("Clearance") to evaluate active case-finding campaigns. Because natural clearance is biologically probable and yields better model fit, traditional models likely underestimate case-finding impact.
Early user experience and lessons learned using ultra-portable digital X-ray with computer-aided detection (DXR-CAD) products: A qualitative study from the perspective of healthcare providers
While ultra-portable DXR with CAD effectively decentralized TB screening, implementation was hindered by programmatic trade-offs, capacity constraints, and missing radiation safety guidelines.
User perspectives on the use of X-rays and computer-aided detection for TB
A 2021 survey of TB project implementers found that CAD-enhanced X-rays significantly improved screening accuracy, throughput, and turnaround times. Respondents favored abnormality score outputs, confirming CAD and X-rays as promising tools to close the TB diagnostic gap.
Trade-offs in diagnostic algorithm, population coverage, and duration of community screening for tuberculosis: A modelling study
Our findings highlight that targeting all TB through symptom-agnostic chest X-ray screening achieves significantly greater burden reductions than symptom-based algorithms, mirroring the historical success of mass radiography.
Subclinical Tuberculosis Disease—A Review and Analysis of Prevalence Surveys to Inform Definitions, Burden, Associations, and Screening Methodology
Analyzing TB prevalence surveys since 1990, we defined subclinical TB and evaluated screening performance. Chest X-rays detected 89% (73%–98%) of confirmed cases, highlighting a key opportunity to improve case-finding policies.
Implementing AI-enabled chest X-ray for community-based integrated screening for tuberculosis, chronic respiratory diseases, and cardiovascular diseases in Nigeria
AI-driven portable chest X-rays effectively detected rural Nigerian TB cases, but optimizing incidental cardiovascular and respiratory findings requires strengthened primary care and integrated diagnostics.
AI Performance, Safety and Effectiveness
Computer-aided detection of tuberculosis from chest radiographs in a tuberculosis prevalence survey in South Africa: external validation and modelled impacts of commercially available artificial intelligence software
Evaluating 12 CAD products in South Africa revealed varied accuracy (AUCs <0.8 to ~0.9), with top software maintaining >90% sensitivity despite shifting thresholds.
Tuberculosis detection from chest x-rays for triaging in a high tuberculosis-burden setting: an evaluation of five artificial intelligence algorithms
Testing five commercial AI algorithms against three radiologists across 23,954 chest X-rays on a novel dataset revealed superior AI performance. All five algorithms (CAD4TB, InferRead, Lunit, JF CXR-1, and qXR) significantly outperformed human readers in TB triage.
A prospective multicenter clinical research study validating the effectiveness and safety of a chest X-ray-based pulmonary tuberculosis screening software JF CXR-1 built on a convolutional neural network algorithm
Across six Chinese hospitals (n=1,161), software matched board radiologists with 94.2% sensitivity and 91.2% specificity for TB screening.
Comparing different versions of computer-aided detection products when reading chest X-rays for tuberculosis
Analyzing 12,890 chest X-rays, this study showed significant accuracy jumps in upgraded CAD software (CAD4TB v7 AUC 0.903 vs. v6 0.823; qXR v3 AUC 0.906 vs. v2 0.872). All versions equaled or outperformed human radiologists, with newer updates demonstrating enhanced triage performance.
Using artificial intelligence to read chest radiographs for tuberculosis detection: A multi-site evaluation of the diagnostic accuracy of three deep learning systems
In Nepal and Cameroon, deep learning chest X-ray triage matched 95%+ sensitivity with superior specificity, potentially reducing costly Xpert testing by 66%.
Radiation Safety
Assessment of scattered and leakage radiation from ultra-portable X-ray systems in chest imaging: An independent study
Ultra-portable X-ray devices can keep public and radiographer radiation doses below the 1 mSv/year limit by using basic safety principles (time, distance, shielding) and proper chest exposure settings.
NCRP Recommendations for Ending Routine Gonadal Shielding During Abdominal and Pelvic Radiography
NCRP (U.S. based) has concluded that in most circumstances GS use does not contribute significantly to reducing risks from exposure and may have the unintended consequences of increased exposure and loss of valuable diagnostic information, and therefore use of GS is not justified as a routine part of radiological protection.
Case Studies
AI-Powered Computer-Aided Detection (CAD) Software
The Stop TB Partnership has been at the forefront of early implementation of AI-powered computer-aided detection (CAD) software for TB detection globally. Stop TB Partnership’s TB REACH Initiative has funded and supported the implementation of CAD products in 14 different countries across Africa, Latin America, and South-East Asia in a diverse range of populations and settings.