Continuous cardiovascular monitoring can play a key role in precision health. However, some fundamental cardiac biomarkers of interest, including stroke volume and cardiac output, require invasive measurements, e.g., arterial pressure waveforms (APW). As a non-invasive alternative, photoplethysmography (PPG) measurements are routinely collected in hospital settings. Unfortunately, the prediction of key cardiac biomarkers from PPG instead of APW remains an open challenge, further complicated by the scarcity of annotated PPG measurements. As a solution, we propose a hybrid approach that uses hemodynamic simulations and unlabeled clinical data to estimate cardiovascular biomarkers directly from PPG signals. Our hybrid model combines a conditional variational autoencoder trained on paired PPG-APW data with a conditional density estimator of cardiac biomarkers trained on labeled simulated APW segments. As a key result, our experiments demonstrate that the proposed approach can detect fluctuations of cardiac output and stroke volume and outperform a supervised baseline in monitoring temporal changes in these biomarkers.

Published in Oct. in the Journal of Pharmacokinetics and Pharmacodynamics, the model assesses metabolic and cardiovascular biomarkers – measurable biological processes such as cholesterol levels, body mass index, glucose and blood pressure – to calculate health status and disease risks across a patient's lifespan. The findings are critical due to the increased risk of developing metabolic and cardiovascular diseases with aging, a process that has adverse effects on cellular, psychological and behavioral processes. "There is an unmet need for scalable approaches that can provide guidance for pharmaceutical care across the lifespan in the presence of aging and chronic co-morbidities," says lead author Murali Ramanathan, PhD, professor of pharmaceutical sciences in the UB School of Pharmacy and Pharmaceutical Sciences. "This knowledge gap may be potentially bridged by innovative disease progression modeling." The model could facilitate the assessment of long-term chronic drug therapies, and help clinicians monitor treatment responses for conditions such as diabetes, high cholesterol and high blood pressure, which become more frequent with age, says Ramanathan.