Using machine learning to better diagnose COPD exacerbations

Bojidar Rangelov
Industrial Fellow 2019

Bojidar Rangelov
GSK and University College London

Bojidar is developing an innovative machine learning technique for diagnosing flare-ups of a common respiratory condition known as Chronic Obstructive Pulmonary Disease (COPD). People living with COPD experience shortness of breath and difficulty with physical activity. Many of them also experience episodes where their symptoms get rapidly worse, known as exacerbations. The patho-physiological mechanisms that cause these are not well understood, which makes it difficult for clinicians to diagnose and detect them early.

To better understand exacerbations, Bojidar is applying machine learning techniques to analysing CT scan images of patients’ lungs. As well as establishing new biomarkers based on the CT images this approach could classify exacerbations into different categories that require different treatments. This will potentially help clinicians choose the best treatment for each COPD patient, thus improving outcomes.

“The main benefit of being part of the 1851 family is access to a global network of researchers.”

Bojidar is studying for a PhD in Medical Imaging at UCL. He holds a Masters in Biomedical Engineering from the University of Glasgow and an MRes in Medical Imaging from UCL. During his previous studies and working with his industry supervisor at GSK he developed an interest in finding quantifiable biomarkers to help diagnose and treat COPD patients.

“The Industrial Fellowship programme’s focus on industrial collaboration is very well aligned to my project,” says Bojidar. “The main benefit of being part of the 1851 family is access to a global network of researchers.”