Key result
Clustering-based machine learning cuts cuffless SBP estimation error by ~3.8 mmHg vs unclustered models.
Why the study?
Does applying a k-means clustering step prior to machine-learning regression improve the accuracy of continuous noninvasive cuff-less blood pressure estimation from ECG and PPG signals in ICU patients?
Does applying a k-means clustering step prior to machine-learning regression improve the accuracy of continuous noninvasive cuff-less blood pressure estimation from ECG and PPG signals in ICU patients?
Pre-clustering physiological features (PTT, PIR, and HR) from ECG and PPG signals before applying gradient boosting regression substantially reduces cuff-less blood pressure estimation error and meets BHS Grade A and AAMI accuracy standards.
May improve cuffless BP estimation accuracy in ICU patients; leaves open prospective validation before clinical adoption.
Extensive research has been performed on continuous and noninvasive cuff-less blood pressure (BP) measurement using artificial intelligence algorithms. This approach involves extracting certain features from physiological signals, such as ECG, PPG, ICG, and BCG, as independent variables and extracting features from arterial blood pressure (ABP) signals as dependent variables and then using machine-learning algorithms to develop a blood pressure estimation model based on these data. The greatest challenge of this field is the insufficient accuracy of estimation models. This paper proposes a novel blood pressure estimation method with a clustering step for accuracy improvement. The proposed method involves extracting pulse transit time (PTT), PPG intensity ratio (PIR), and heart rate (HR) features from electrocardiogram (ECG) and photoplethysmogram (PPG) signals as the inputs of clustering and regression, extracting systolic blood pressure (SBP) and diastolic blood pressure (DBP) features from ABP signals as dependent variables, and finally developing regression models by applying gradient boosting regression (GBR), random forest regression (RFR), and multilayer perceptron regression (MLP) on each cluster. The method was implemented using the MIMIC-II data set with the silhouette criterion used to determine the optimal number of clusters. The results showed that because of the inconsistency, high dispersion, and multitrend behavior of the extracted features vectors, the accuracy can be significantly improved by running a clustering algorithm and then developing a regression model on each cluster and finally weighted averaging of the results based on the error of each cluster. When implemented with 5 clusters and GBR, this approach yielded an MAE of 2.56 for SBP estimates and 2.23 for DBP estimates, which were significantly better than the best results without clustering (DBP: 6.27, SBP: 6.36).
No takes yet. Share an insight, caveat, or question.
Farki et al. (2022) studied Blood pressure monitoring in intensive care unit patients (n=12,000). Clustering-based (k-means, 5 clusters) gradient boosting regression algorithm for cuff-less blood pressure estimation using ECG and PPG features (PTT, PIR, HR) vs. Machine-learning regression models (GBR, RFR, MLP) developed on the entire dataset without prior clustering was evaluated on Mean absolute error (MAE) of systolic blood pressure (SBP) and diastolic blood pressure (DBP) estimation. Clustering-based gradient boosting regression (5 clusters) reduced mean absolute error to 2.56 for SBP and 2.23 for DBP compared to 6.36 and 6.27 without clustering.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: