Publications
Andrew McDonald B S R: Development and validation of AI-Enhanced auscultation for valvular heart disease screening through a multi-centre study. In: Nature npj Cadriovascular Health, vol. 3, iss. 1, no. 5, pp. 1–7, 2026. @article{CAIS,
title = {Development and validation of AI-Enhanced auscultation for valvular heart disease screening through a multi-centre study},
author = {Andrew McDonald, Mark Gales, Bushra S Rana, Matthew Shun-Shin, Benito F Lukban, Rita Adrego, Alexandros Papachristidis, Fatima Hajee, Len Shapiro, Joanna Wilson, Tony Prothero, Andrew Kennedy, Saul Myerson, Bernard Prendergast, Patrik Bachtiger, Mihir A Kelshiker, Nicholas Peters, Richard Steeds, Anurag Agarwal},
doi = {doi.org/10.1038/s44325-026-00103-y},
year = {2026},
date = {2026-02-10},
journal = {Nature npj Cadriovascular Health},
volume = {3},
number = {5},
issue = {1},
pages = {1--7},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
|
Kate Bassil A A: Underestimation of systolic pressure in cuff-based blood pressure measurement. In: PNAS Nexus, vol. 4, iss. 8, pp. pgaf222, 2025. @article{BA25,
title = {Underestimation of systolic pressure in cuff-based blood pressure measurement},
author = {Kate Bassil, Anurag Agarwal},
doi = {doi.org/10.1093/pnasnexus/pgaf222},
year = {2025},
date = {2025-08-12},
journal = {PNAS Nexus},
volume = {4},
issue = {8},
pages = {pgaf222},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
|
A. McDonald N R, Agarwal A: A Flexible Multi-Sensor Device Enabling Handheld Sensing of Heart Sounds by Untrained Users. In: IEEE Journal of Biomedical and Health Informatics, vol. 29, no. 8, pp. 5575–5584, 2025. @article{MNR25,
title = {A Flexible Multi-Sensor Device Enabling Handheld Sensing of Heart Sounds by Untrained Users},
author = {A. McDonald, M. Nussbaumer, N. Rathnayake, R. Steeds and A. Agarwal},
doi = {10.1109/JBHI.2025.3551882},
year = {2025},
date = {2025-08-08},
urldate = {2025-08-08},
journal = {IEEE Journal of Biomedical and Health Informatics},
volume = {29},
number = {8},
pages = {5575--5584},
abstract = {Heart valve disease has a large and growing burden, with a prognosis worse than many cancers. Screening with a traditional stethoscope is underutilised, often inaccurate even in skilled hands, and requires time-consuming, intimate examinations. Here, we present a handheld device to enable untrained users to record high-quality heart sounds without requiring patients to undress. The device incorporates multiple high-sensitivity sensors embedded in a flexible substrate, placed at key chest locations by the user. To address challenges from localised heart sound vibrations and noise interference, we developed time-frequency signal quality algorithms that automatically select the best sensor in the device and reject recordings with insufficient diagnostic quality. A validation study demonstrates the device's effectiveness across a diverse range of body types, with multiple sensors significantly increasing the likelihood of a successful recording. The device has the potential to enable accurate, accessible, low-cost heart disease screening.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Heart valve disease has a large and growing burden, with a prognosis worse than many cancers. Screening with a traditional stethoscope is underutilised, often inaccurate even in skilled hands, and requires time-consuming, intimate examinations. Here, we present a handheld device to enable untrained users to record high-quality heart sounds without requiring patients to undress. The device incorporates multiple high-sensitivity sensors embedded in a flexible substrate, placed at key chest locations by the user. To address challenges from localised heart sound vibrations and noise interference, we developed time-frequency signal quality algorithms that automatically select the best sensor in the device and reject recordings with insufficient diagnostic quality. A validation study demonstrates the device's effectiveness across a diverse range of body types, with multiple sensors significantly increasing the likelihood of a successful recording. The device has the potential to enable accurate, accessible, low-cost heart disease screening. |
A M, MJF G, A A: A recurrent neural network and parallel hidden Markov model algorithm to segment and detect heart murmurs in phonocardiograms. In: PLOS DIGITAL HEALTH, vol. 3, iss. 11, no. e0000436, pp. 1–20, 2024. @article{MGA24,
title = {A recurrent neural network and parallel hidden Markov model algorithm to segment and detect heart murmurs in phonocardiograms},
author = { McDonald A and Gales MJF and Agarwal A},
doi = {10.1371/journal.pdig.0000436},
year = {2024},
date = {2024-11-25},
urldate = {2024-11-25},
journal = {PLOS DIGITAL HEALTH},
volume = {3},
number = {e0000436},
issue = {11},
pages = {1--20},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
|
McDonald A, Matos J N, Silva J, Partington C, Lo E J Y, Fuentes V L, Barron L, Watson P, Agarwal A: A machine-learning algorithm to grade heart murmurs and stage preclinical myxomatous mitral valve disease in dogs. In: Journal of Veterinary Internal Medicine, 2024. @article{canine_jvim,
title = {A machine-learning algorithm to grade heart murmurs and stage preclinical myxomatous mitral valve disease in dogs},
author = {Andrew McDonald and Jose Novo Matos and Joel Silva and Catheryn Partington and Eve J. Y. Lo and Virginia Luis Fuentes and Lara Barron and Penny Watson and Anurag Agarwal},
doi = {https://doi.org/10.1111/jvim.17224},
year = {2024},
date = {2024-10-21},
urldate = {2024-10-21},
journal = {Journal of Veterinary Internal Medicine},
abstract = {Background: The presence and intensity of heart murmurs are sensitive indicators of several cardiac diseases in dogs, particularly myxomatous mitral valve disease (MMVD), but accurate interpretation requires substantial clinical expertise.
Objectives: Assess if a machine-learning algorithm can be trained to accurately detect and grade heart murmurs in dogs and detect cardiac disease in electronic stethoscope recordings.
Animals: Dogs (n = 756) with and without cardiac disease attending referral centers in the United Kingdom.
Methods: All dogs received full physical and echocardiographic examinations by a cardiologist to grade any murmurs and identify cardiac disease. A recurrent neural network algorithm, originally trained for heart murmur detection in humans, was fine-tuned on a subset of the dog data to predict the cardiologist's murmur grade from the audio recordings.
Results: The algorithm detected murmurs of any grade with a sensitivity of 87.9% (95% confidence interval [CI], 83.8%-92.1%) and a specificity of 81.7% (95% CI, 72.8%-89.0%). The predicted grade exactly matched the cardiologist's grade in 57.0% of recordings (95% CI, 52.8%-61.0%). The algorithm's prediction of loud or thrilling murmurs effectively differentiated between stage B1 and B2 preclinical MMVD (area under the curve [AUC], 0.861; 95% CI, 0.791-0.922), with a sensitivity of 81.4% (95% CI, 68.3%-93.3%) and a specificity of 73.9% (95% CI, 61.5%-84.9%).
Conclusion and Clinical Importance: A machine-learning algorithm trained on humans can be successfully adapted to grade heart murmurs in dogs caused by common cardiac diseases, and assist in differentiating preclinical MMVD. The model is a promising tool to enable accurate, low-cost screening in primary care.
},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Background: The presence and intensity of heart murmurs are sensitive indicators of several cardiac diseases in dogs, particularly myxomatous mitral valve disease (MMVD), but accurate interpretation requires substantial clinical expertise.
Objectives: Assess if a machine-learning algorithm can be trained to accurately detect and grade heart murmurs in dogs and detect cardiac disease in electronic stethoscope recordings.
Animals: Dogs (n = 756) with and without cardiac disease attending referral centers in the United Kingdom.
Methods: All dogs received full physical and echocardiographic examinations by a cardiologist to grade any murmurs and identify cardiac disease. A recurrent neural network algorithm, originally trained for heart murmur detection in humans, was fine-tuned on a subset of the dog data to predict the cardiologist's murmur grade from the audio recordings.
Results: The algorithm detected murmurs of any grade with a sensitivity of 87.9% (95% confidence interval [CI], 83.8%-92.1%) and a specificity of 81.7% (95% CI, 72.8%-89.0%). The predicted grade exactly matched the cardiologist's grade in 57.0% of recordings (95% CI, 52.8%-61.0%). The algorithm's prediction of loud or thrilling murmurs effectively differentiated between stage B1 and B2 preclinical MMVD (area under the curve [AUC], 0.861; 95% CI, 0.791-0.922), with a sensitivity of 81.4% (95% CI, 68.3%-93.3%) and a specificity of 73.9% (95% CI, 61.5%-84.9%).
Conclusion and Clinical Importance: A machine-learning algorithm trained on humans can be successfully adapted to grade heart murmurs in dogs caused by common cardiac diseases, and assist in differentiating preclinical MMVD. The model is a promising tool to enable accurate, low-cost screening in primary care.
|
