A Transformer-based deep learning model using non-invasive PPG signals estimated systolic blood pressure with a mean absolute error of 2.52 mmHg, meeting clinical standards.
Does a Transformer-based deep learning model using non-invasive PPG signals accurately estimate arterial blood pressure and oxygen saturation in ICU patients?
1,732 ICU patients from the MIMIC III database with documented bedside waveform records (PPG and ABP signals), mean age 62.0 years, 55.1% male.
Transformer-based deep learning architecture (MLM-Transformer with personalization) utilizing raw Photoplethysmogram (PPG) signals for continuous estimation of blood pressure and oxygen saturation.
Invasive arterial blood pressure (ABP) catheter measurements (ground truth) and baseline machine learning models (basic Transformer, U-Net).
Mean absolute error (MAE) for arterial systolic blood pressure (ASBP), arterial diastolic blood pressure (ADBP), and oxygen saturation (SpO2).surrogate
A Transformer-based deep learning model using non-invasive PPG signals can accurately estimate arterial blood pressure and oxygen saturation, meeting clinical standards for accuracy.
BACKGROUND: Monitoring blood pressure and peripheral capillary oxygen saturation plays a crucial role in healthcare management for patients with chronic diseases, especially hypertension and vascular disease. However, current blood pressure measurement methods have intrinsic limitations; for instance, arterial blood pressure is measured by inserting a catheter in the artery causing discomfort and infection. METHOD: Photoplethysmogram (PPG) signals can be collected via non-invasive devices, and therefore have stimulated researchers' interest in exploring blood pressure estimation using machine learning and PPG signals as a non-invasive alternative. In this paper, we propose a Transformer-based deep learning architecture that utilizes PPG signals to conduct a personalized estimation of arterial systolic blood pressure, arterial diastolic blood pressure, and oxygen saturation. RESULTS: The proposed method was evaluated with a subset of 1,732 subjects from the publicly available ICU dataset MIMIC III. The mean absolute error is 2.52 ± 2.43 mmHg for systolic blood pressure, 1.37 ± 1.89 mmHg for diastolic blood pressure, and 0.58 ± 0.79% for oxygen saturation, which satisfies the requirements of the Association of Advancement of Medical Instrumentation standard and achieve grades A for the British Hypertension Society standard. CONCLUSIONS: The results indicate that our model meets clinical standards and could potentially boost the accuracy of blood pressure and oxygen saturation measurement to deliver high-quality healthcare.
Building similarity graph...
Analyzing shared references across papers
Loading...
Yan Chu
The University of Texas Health Science Center at Houston
Kaichen Tang
City University of Hong Kong
Yu‐Chun Hsu
National Taiwan University Hospital
BMC Medical Informatics and Decision Making
The University of Texas Health Science Center at Houston
San Jose State University
Building similarity graph...
Analyzing shared references across papers
Loading...
Chu et al. (Fri,) conducted a other in ICU patients requiring blood pressure and SpO2 monitoring (n=1,732). Transformer-based deep learning model using PPG signals vs. Invasive arterial blood pressure and SpO2 measurements (ground truth) was evaluated on Mean absolute error (MAE) for systolic blood pressure. A Transformer-based deep learning model using non-invasive PPG signals estimated systolic blood pressure with a mean absolute error of 2.52 mmHg, meeting clinical standards.
synapsesocial.com/papers/6a05146c6c3d07813971bda6 — DOI: https://doi.org/10.1186/s12911-023-02215-2