AI technology to examine X-Rays, MRIs, CT scans, identify TB patients

New Delhi: A Mumbai-based health-tech start-up, Qure.ai, has developed an artificial intelligence (AI)-powered technology that can examine X-rays, MRIs and CT scans, identify patients with diseases like tuberculosis or stroke, which can be traced with these diagnostic techniques and prepare accurate reports in no time.

AI technology involves the application of deep learning algorithms to detect and highlight abnormalities in medical imaging like chest X-rays, MRIs and CT scans. It aims to significantly scale up and improve the diagnosis of diseases, especially TB, in the country.

“It is based on artificial neural networks which are trained by showing an algorithm a wide variety of cases, which is in some ways similar to how one would train a radiologist,” said Prashant Warier, Qure.ai's co-founder.

“We showed the algorithm 1.5 million X-rays, some normal and others abnormal. And trained them to understand what an abnormality would look like. Techniques like deep learning have made it possible to automatically identify multiple disease states from a medical image,” he said.

The technology is under evaluation for implementation at public health centres in Maharashtra, claimed the start-up company.

While highly competitively priced diagnostic centres have mushroomed all over the country, medical imaging has multiplied dramatically, the technology can make TB screening in the country faster, effective and economical. It may also help prevent a large number of non-TB patients from having to go through the confirmatory GeneXpert test (which detects TB bacteria and resistance to TB drug Rifampicin) which costs more than Rs3,000 per test.

“Shortage of radiologists in India is huge and hence diagnosis suffers. India has less than 2 radiologists per million population. However, Qure.ai's intelligent and trained algorithms can help a radiologist, especially in remote areas, to scale up productivity by three times and become more efficient,” said Sundeep Reddivari, general manager- telemedicine, Columbia Asia Hospital, Bengaluru.

Qure.ai claims that its algorithms can rule out brain bleed in seconds making the treatment protocol for stroke much more effective. It also claims that its algorithms are unique because they clearly highlight the problem area in the patient's X-rays or CT scans through a heat map.

“The technology will not only easily detect chronic and non-communicable diseases in remote areas, it will also speed up the process of reading an X-ray or CT scan for a radiologist. “It is crucial for the treatment of stroke to assess whether the patient is suffering from a clot or a brain bleed, so that an anti-coagulant can be given,” said Warier.

Currently, India has the highest incidence of TB cases in the world, accounting for nearly 27% of the world's 10 million TB cases.