Clinical Report: Assessing CAD4TB for Tuberculosis Screening in Indonesia

Overview

This study evaluates the diagnostic performance of CAD4TB in screening for tuberculosis using chest x-ray images in Indonesia, comparing it to the Timika score. CAD4TB demonstrated a higher specificity and comparable sensitivity, suggesting its potential as a rapid screening tool in resource-limited settings.

Background

Tuberculosis remains a leading cause of global mortality, particularly in Southeast Asia, with Indonesia facing a significant burden. The challenges of misdiagnosis and limited resources in rural areas necessitate effective screening methods. The integration of artificial intelligence, such as CAD4TB, may enhance diagnostic accuracy and efficiency in these settings.

Data Highlights

Key Findings

• CAD4TB achieved an AUC of 0.778 compared to AFB results.
• With a ≤7-day interval, CAD4TB's AUC was 0.767, comparable to the Timika score.
• CAD4TB exhibited superior specificity (71.43%) compared to the Timika score (57.64%).
• Both CAD4TB and the Timika score maintained a fixed sensitivity of 73.91%.
• Misdiagnosis remains a significant challenge in tuberculosis management in Indonesia.

Clinical Implications

The findings indicate that CAD4TB may serve as a reliable tool for tuberculosis screening in resource-limited settings, potentially improving diagnostic accuracy. Its higher specificity could reduce false positives, aiding in better patient management.

Conclusion

CAD4TB shows promise as an effective screening tool for tuberculosis in Indonesia, outperforming the Timika score in specificity while maintaining comparable sensitivity.

Related Resources & Content

1. Open Forum Infectious Diseases, 2022 -- Determining the Most Effective Community-Based Tuberculosis Screening Strategy for Individuals Who Inject Drugs in High-Burden Areas
2. Open Forum Infectious Diseases, 2022 -- Improving Tuberculosis Diagnosis Through Artificial Intelligence (CAD4TB) and Stool Xpert MTB/RIF Testing: A Prospective Study From Oromia, Ethiopia
3. Infection, 2022 -- Evaluation of Creatine Kinase-MB and Hepcidin as Potential Early Biomarkers for Pulmonary Tuberculosis Diagnosis: A Proof-of-Concept Investigation in Lambaréné, Gabon
4. WHO Policy Statement -- Use of computer-aided detection software for tuberculosis screening
5. PMC, 2024 -- A systematic review and meta-analysis of artificial intelligence software for tuberculosis diagnosis using chest X-ray imaging
6. Open Forum Infectious Diseases — Evaluation of the VISITECT CD4 Advanced Disease Test's Accuracy and Acceptability Relative to the PIMA CD4 Test in Point-of-Care Settings Within the Advanced HIV Disease Care Framework: A Mixed-Methods Investigation
7. Use of computer-aided detection software for tuberculosis screening: WHO policy statement
8. A systematic review and meta-analysis of artificial intelligence software for tuberculosis diagnosis using chest X-ray imaging - PMC
9. A simple, valid, numerical score for grading chest x-ray severity in adult smear-positive pulmonary tuberculosis - PubMed