Artificial intelligence as a proficient tool in detecting pulmonary tuberculosis in massive population screening programs: a case study in Chennai, India
- VernacularTitle:Artificial intelligence as a proficient tool in detecting pulmonary tuberculosis in massive population screening programs: a case study in Chennai, India
- Author:
Prabakaran JAYARAMAN
1
;
Sangeetha S
2
;
Saumit PAUL
3
;
Richa PANT
3
;
Tanveer GUPTE
3
;
Viraj KULKARNI
3
;
Amit KHARAT
3
Author Information
- Keywords: tuberculosis; artificial intelligence; computer-aided detection; Genki; x-ray
- From:Journal of Rural Medicine 2025;20(1):13-19
- CountryJapan
- Language:English
- Abstract: Objective: To evaluate the performance of Genki, a computer-aided detection (CADe) software, in detecting tuberculosis (TB) using chest radiography in a mobile TB screening program in Chennai, India.Materials and Methods: Genki, an AI-based CADe software, was employed in four mobile diagnostic units in remote areas of Chennai, India for screening TB. Patients from remote areas of Chennai who visited the vans and registered in the screening program underwent chest radiography, and the acquired X-ray scans were analyzed using Genki, which provided an assessment of each scan as either “TB suggestive” or “TB not suggestive”. Subsequently, sputum or swab from the patients with “TB suggestive” results was collected to confirm the diagnosis.Results: In total, 25,598 patients were screened between January and December 2022. When the annotations from the expert radiologists were considered to be true, Genki demonstrated an aggregated sensitivity of 98%, specificity of 96.9%, and accuracy of 96.9% in detecting TB from chest X-ray scans of the screened population. Furthermore, it exhibited a sensitivity, specificity, and accuracy of >95%, >94%, and >94%, respectively, for both sexes (male and female) and all age groups (14–35, 36–60, and ≥61 years).Conclusion: Genki demonstrated excellent value as a TB screening tool in remote locations in Chennai, India. Employing a CADe-based approach for systematic TB screening is cost-effective and reduces workload in high-burden and low-resource settings.
