Background Physics-based cardiovascular models are only recently being considered for disease diagnosis or prognosis in clinical settings. These models depend on parameters representing the physical and physiological properties of the modeled system. Personalizing these parameters may give insight into the specific state of the individual and etiology of disease. We applied a relatively fast model optimization scheme based on common local optimization methods to two model formulations of the left ventricle and systemic circulation. One closed-loop model and one open-loop model were applied. Intermittently collected hemodynamic data from an exercise motivation study were used to personalize these models for data from 25 participants. The hemodynamic data were collected for each participant at the start, middle and end of the trial. We constructed two data sets for the participants, both consisting of systolic and diastolic brachial pressure, stroke volume, and left-ventricular outflow tract velocity traces paired with either the finger arterial pressure waveform or the carotid pressure waveform. Results We examined the feasibility of separating parameter estimates for the individual from population estimates by assessing the variability of estimates using the interquartile range. We found that the estimated parameter values were similar for the two model formulations, but that the systemic arterial compliance was significantly different ($$p < 10^{-6}$$ p < 10 - 6 ) depending on choice of pressure waveform. The estimates of systemic arterial compliance were on average higher when using the finger artery pressure waveform as compared to the carotid waveform. Conclusions We found that for the majority of participants, the variability of parameter estimates for a given participant on any measurement day was lower than the variability both across all measurement days combined for one participant, and for the population. This indicates that it is possible to identify individuals from the population, and that we can distinguish different measurement days for the individual participant by parameter values using the presented optimization method.
A recent meta-review on hypertension risk models detailed that the differences in data and study-setup have a large influence on performance, meaning model comparisons should be performed using the same study data. We compared five different machine learning algorithms and the externally developed Framingham risk model in predicting risk of incident hypertension using data from the Trøndelag Health Study. The dataset yielded n = 23722 individuals with p = 17 features recorded at baseline before follow-up 11 years later. Individuals were without hypertension, diabetes, or history of CVD at baseline. Features included clinical measurements, serum markers, and questionnaire-based information on health and lifestyle. The included modelling algorithms varied in complexity from simpler linear predictors like logistic regression to the eXtreme Gradient Boosting algorithm. The other algorithms were Random Forest, Support Vector Machines, K-Nearest Neighbor. After selecting hyperparameters using cross-validation on a training set, we evaluated the models’ performance on discrimination, calibration, and clinical usefulness on a separate testing set using bootstrapping. Although the machine learning models displayed the best performance measures on average, the improvement from a logistic regression model fitted with elastic regularization was small. The externally developed Framingham risk model performed well on discrimination, but severely overestimated risk of incident hypertension on our data. After a simple recalibration, the Framingham risk model performed as well or even better than some of the newly developed models on all measures. Using the available data, this indicates that low-complexity models may suffice for long-term risk modelling. However, more studies are needed to assess potential benefits of a more diverse feature-set. This study marks the first attempt at applying machine learning methods and evaluating their performance on discrimination, calibration, and clinical usefulness within the same study on hypertension risk modelling.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.