|
Karnataka has now embarked on Screening for Tuberculosis Using Artificial Intelligence in Remote Radiology (STAR) facilities. The move comes in after the WHO guidelines which also give a huge boost to computer-aided AI detection software as it can better interpret the digital x-rays. This will help triage suspected cases, more accurately than human x-ray readers.
The new AI makes the screening process more efficient, and significantly reduces the time taken to make a diagnosis. This will ensure last mile connectivity and provide seamless screening of presumptive TB patients via x-ray and further confirmation on Nucleic Acid Amplification Test (NAAT).
The National TB Elimination Programme (NTEP) in Karnataka with STAR is a partnership program between State TB Office, Karnataka Health Promotion Trust (KHPT) and Association of Public Health Technologists (APT) with support from Qure AI and India Health Fund.
This AI tool developed by Qure AI which works on Analog x-ray screening is well suited to be used in remote facilities and it not only detects abnormal chest x-rays, but it can identify and localizes 29 common abnormalities in addition to screen for tuberculosis.
This solution provides an opportunity for the health sector, to bring in efficiency, saving resources and accuracy in interpretation and enhancing quality of service delivery.
Based on review of TB data available from Nikshay, 5 districts: Chikaballapur, Koppal, Bellary, Belgaum and Bagalkot were selected for pilot implementation of AI based solution to screen x-rays for tuberculosis.
In Karnataka annually around 8 lakh presumptive TB patients undergo sputum smear examinations and another 2 lakh presumptive undergo rapid molecular testing for TB. WHO recommends that before any confirmation of TB on NAAT it is ideal to screen the presumptive TB patients on x-ray. There are 75 CBNAAT and 101TruNAAT machines in the state deployed for this purpose.
There are approximately 507 x-ray sites in the state in the public sector and most are not co-located with rapid molecular tests.
There are issues with x-ray reading in the periphery due to huge intra and inter observer variations. Hence it is imperative to use a digital x-ray reading technology which can easily identify the abnormal x-rays and pin-point the signs of TB. This called to bring in to use artificial intelligence under TB programme for screening x-rays. This is because the district health facilities do not capture presumptive TB data base from all the diagnostic modalities. This is because there is currently there is no mechanism to capture date from x-ray sites.
There is a high chance of missing presumptive TB patients from x-ray sites. But x-ray screening using AI is an asset in these unprecedented Covid times for TB case finding, said the state authorities.
|