Lung conditions: New AI test set to transform diagnosis

A rapid test harnessing the power of AI is set to transform the diagnosis of a lung condition that affects 3 million people in the UK.

Two thirds of those with chronic obstructive pulmonary disease (COPD) don’t even know they’ve got it, but it’s the second most common cause of emergency hospital admissions.

Until now, confirmation has required a spirometry test involving forced exhalation. It can take half an hour or more, and that’s once you’ve managed to get an appointment.

Dr Simon Rudland, a GP in Suffolk, is excited about the potential to provide quick, accessible diagnosis.

“Spirometry is delivered out of a hub, so patients have to travel,” he said.

“We need to employ experts within those hubs that require specialist training. And the biggest issue is the wait.”

Instead, the handheld device called N-Tidal Diagnose will enable a patient to walk into their doctor’s surgery and find out in as little as five minutes if their symptoms like breathlessness and a persistent cough are caused by COPD.

“It completely changes the whole pathway,” said Dr Rudland.

“With that diagnosis I can initiate treatment. I don’t have to wait. They don’t get put on the wrong drugs. It’s revolutionary.”

COPD isn’t curable but it is treatable, but the sooner that treatment starts, the better it is long-term for the patient.

Flare-ups account for one in eight people being admitted to hospital and a third won’t have been diagnosed before that.

‘Only five minutes’

Colin Best, 67, from Stowmarket in Suffolk, was gasping for breath at the slightest exertion when he went to the doctor’s eight years ago.

He had multiple different tests before it was decided he had the disease, including spirometry four or five times.

“The spirometry test is exhausting. You have to have several tests and each one has a gap of five [to] 10 minutes between to let you relax,” he said.

“It took me a year to convince them there was something seriously wrong whereas with that machine it’s only five minutes.”

The patient breathes normally into the handset, which logs carbon dioxide and sends data to a cloud platform where it is analysed in real-time by AI software.

Dr Ameera Patel, chief executive of TidalSense which makes the device, said: “In total, we’ve collected over 2.5 million patient breaths from every cardio-respiratory condition that you can think of and then we’ve trained the machine learning model.

“We’ve taught it to discriminate COPD from everything else that can look like COPD. So that’s asthma, heart failure, lung cancer.”

And it only takes 10 minutes to train staff to use it.

The sensor has been approved for use across the EU and is expected to start being used in doctors’ surgeries in a matter of weeks.

The hope is it will transform diagnosis and cut costs for a cash-strapped NHS.