Tuberculosis, in countries like India and Bangladesh, has always been a rising concern. As per reports, in India, 10 million people catch the illness every year.
However, only 6 million cases are identified, with the remaining 4 million being “missing”, which means that they are either undiagnosed, untreated or go unreported to the healthcare systems.
This gap occurs mainly because the traditional testing system takes weeks to produce results, making it difficult to reach patients promptly.
Addressing this issue, a Pune-based startup called DeepTek has developed a technology claiming to identify tuberculosis cases within seconds.
Founded by Dr Amit Kharat and Ajit Patil, the startup’s AI-powered tool Genki scans the X-ray reports of the suspected patients within seconds and produces a report.
“The chest X-rays are scanned by the AI and within seconds, depending on the CPU setting and core capacity of the system in the van, camp or hospital, the reports come in within 60 seconds,” Amit informs.
The AI was trained using 10 million data points from various sources and places in the world.
“We have collaborated with medical schools and hospitals for research and collecting data. We have been audited by Stop TB, a WHO body which has given a green signal calling our AI’s results as good as human reading,” he says.
The startup has screened more than 300,000 patients in totality so far.
“We have approval from the United States of America’s FDA and we recently received approval from an Indian regulatory body called Central Drugs Standard Control Organisation (CDSCO) too,” he says.