IGNITE

A prospective observational study developing a smartphone-based AI algorithm that detects early pulmonary congestion in heart failure from subtle changes in voice, cough and breathing sounds. Two-part study — a 50-patient pilot followed by a 200-patient definitive cohort — to enable earlier community-based intervention and reduce emergency admissions.

A deep-learning system that estimates cardiac stroke volume directly from the audio picked up by ordinary in-ear earbuds. Data from 23 healthy participants yielded a mean absolute error of 5.24 ml, a Pearson correlation of r = 0.94 versus a clinically validated reference, and a percentage-error limit-of-agreement of 11.05% — within accepted clinical range. Published open-access in Nature Communications, 17 July 2026.

A three-centre UK observational study evaluating a wearable ECG patch as a remote, non-invasive tool to capture multidomain physiological endpoints in patients recovering from acute decompensated heart failure. Findings will inform the design of future heart failure drug-development programmes.

Prospective mechanistic experimental-medicine trial in oncology patients receiving pazopanib, a vascular endothelial growth factor receptor (VEGFR) tyrosine-kinase inhibitor, to interrogate the vascular mechanisms of anti-angiogenic drug-induced hypertension. Combined invasive haemodynamics, endothelial-function testing and AI-based retinal image analysis (in collaboration with the Cambridge Image Analysis Group at DAMTP) to quantify microvascular structural change. Published in Hypertension, 2021;77(5):1591-1599.