Home » Digital Health » Fast, accurate, safe and autonomous clinical diagnosis is provided by AI technology
Sponsored

Simon Rasalingham

Chairman and CEO, Behold.ai

Artificial intelligence innovations are being introduced to help the NHS tackle a massive – and growing – elective care backlog resulting from the COVID-19 pandemic.


More than six million people are now on hospital waiting lists for planned treatments, while the NHS faces chronic staff shortages and too few radiologists.

Artificial intelligence

One area where artificial intelligence (AI) is being applied is to read chest x-rays for suspected lung cancer patients. At present, all scans are reported manually by radiologists, which take up considerable time and valuable resource, particularly given the lack of radiologists.

However, a new AI algorithm ‘trained’ on 150,000 images to identify normal lung scans is dramatically speeding up the process and freeing up radiologists’ time to focus on abnormal cases.

With a backlog of 6 million patients, if we remove the normal chest x-rays, that adds significant extra radiological reporting capacity back into the system.

Autonomous diagnosis

Simon Rasalingham, CEO of British start-up Behold.ai, outlined the support his company is providing the NHS in highlighting normal chest x-rays using AI.

Currently, a diagnosis of lung cancer can take over 30 days, whereas the new approach sees chest x-rays read within 30 seconds with an initial diagnosis, helping to quickly rule out those who are not at risk and significantly reduce patient anxiety. As a result, patients with suspected lung cancer can be prioritised for a CT scan which can reduce the whole diagnostic process to under two hours.

Behold.ai, has won an AI award to introduce this novel innovation – initially to 22 sites, with a plan to roll-out to a further 100 this year. The Behold.ai algorithm is 40-times better than the human eye at ruling out normal chest x-rays.

Increased accuracy

With 8.3 million chest x-rays performed each year in the NHS, Behold.ai estimates the algorithm can rapidly report on 4 million of those with increased accuracy to significantly help reduce misdiagnosis. “With a backlog of 6 million patients, if we remove the normal chest x-rays, that adds significant extra radiological reporting capacity back into the system,” adds Rasalingham.

Behold.ai is also registered with the Care Quality Commission as the first and only AI diagnostic provider of its kind, delivering a safe framework for AI services to patients. “Our mission is to save lives from lung cancer and stroke, by significantly reducing the time to diagnosis and treatment,” he says.

Next article