A NEW study demonstrates how artificial neural networks (ANN) and urine biomarkers can accurately predict chronic obstructive pulmonary disease (COPD) exacerbations days before symptoms appear. This innovative approach offers a new way to improve early intervention and reduce the burden of COPD exacerbations, which significantly contribute to patient morbidity and mortality.
Researchers identified a panel of 10 urine biomarkers, including NGAL, TIMP1, and CRP, that distinguish between stable and exacerbation states in COPD. In a retrospective discovery study involving paired urine samples from 55 COPD patients, this biomarker panel achieved an area under the curve (AUC) of 0.84 in identifying exacerbation states. The panel was validated in a prospective study of 105 COPD patients over six months, achieving a similar AUC of 0.81.
Building on this, the research team developed an ANN model using biomarkers from 85 patients to predict exacerbation risk. The model achieved an impressive AUC of 0.89 and was able to identify exacerbations a median of seven days before clinical diagnosis, offering a critical window for early intervention.
The biomarkers were also integrated into a prototype dipstick test with an opto-electronic reader for near-patient use, providing a practical tool for ongoing monitoring and timely intervention. This technology could revolutionize how COPD exacerbations are managed, potentially reducing hospitalizations and improving patient outcomes.
Reference
Yousuf A et al. Artificial neural network risk prediction of chronic obstructive pulmonary disease (COPD) exacerbations using urine biomarkers. EMJ. 2024;DOI: 10.1183/23120541.00797-2024.
