AI urine test predicts COPD before symptoms appear – molecular diagnostics

Image: The Headstart test measures five biomarkers in urine (photo courtesy of Global Access Diagnostics)

Chronic obstructive pulmonary disease (COPD) affects more than 400 million people worldwide and is expected to become the third leading cause of death by 2030. COPD is characterized by persistent symptoms, including shortness of breath, cough and wheezing, and ongoing decline in lung function. Sometimes symptoms worsen, leading to flare-ups. These exacerbations are triggered by various factors, such as viral and bacterial infections, as well as a number of small changes that lead to destabilization of the disease. Exacerbations of COPD can be harmful, causing further lung damage, so prevention is essential to improve quality of life and reduce the risk of death. Current methods for measuring inflammation in COPD typically involve blood or sputum tests taken during exacerbations, as opposed to visits in stable patients that occur several weeks apart. However, these approaches have not led to widespread adoption of the test due to limitations in sensitivity and specificity, as well as difficulties in obtaining samples in a timely manner for adequate treatment. The challenge now is to develop tests that are performed close to the patient and that are capable of detecting and analyzing the heterogeneous inflammatory response that precedes an exacerbation. Now researchers have applied artificial intelligence (AI) to urine samples to predict when COPD symptoms will appear.

Global Access Diagnostics (Bedford, UK) has developed a test prototype called Headstart, a remote patient monitoring platform that measures five biomarkers in urine. Headstart testing is currently underway to detect early signs of exacerbation reliably enough to help patients determine whether they need to seek medical attention. The study, conducted by the University of Leicester (Leicester, UK), involved patients who used this simple daily dipstick test to monitor their urine and sent the results to researchers via their mobile phones. Using artificial intelligence to analyze the data, the researchers were able to predict worsening symptoms a week in advance, allowing them to adjust treatment to prevent or reduce flare-ups.

The study began by analyzing urine samples from 55 COPD patients to identify any changes in urine composition that may have preceded worsening symptoms. This led to the identification of a set of biomarkers: molecules that change as COPD worsens. The researchers then asked an additional 105 COPD patients to use the Headstart device daily for six months and send their results to the researchers via mobile phones. AI, specifically artificial neural network, has been used to study fluctuations in these biomarkers and predict when an outbreak will occur. Research published in ERJ Open Studyshowed that AI analysis can predict exacerbations about seven days before symptoms appear.

“Our study was the first to examine many substances in urine samples from people with COPD during an outbreak and when they were stable,” said Professor Chris Brightling from the University of Leicester. “We found that small amounts of these substances can detect an outbreak. We then followed a group of people with COPD and tested five substances daily. This allowed us to develop a risk forecasting or forecasting tool using artificial intelligence. We found that this AI tool can reliably predict worsening symptoms up to seven days before diagnosis.”