Robust Aortic Valve Non-Opening Detection for Different Cardiac Conditions

Abstract

In recent years, extensive studies have been conducted in the area of pumping state detection for implantable rotary blood pumps. However, limited studies have focused on automatically identifying the aortic valve non-opening (ANO) state despite its importance in the development of control algorithms aiming for myocardial recovery. In the present study, we investigated the performance of 14 ANO indices derived from the pump speed waveform using four different types of classifiers, including linear discriminant analysis, logistic regression, back propagation neural network, and k-nearest neighbors (KNN). Experimental measurements from four greyhounds, which take into consideration the variations in cardiac contractility, systemic vascular resistance, and total blood volume were used. By having only two indices, (i) the root mean square value, and (ii) the standard deviation, we were able to achieve an accuracy of 92.8% with the KNN classifier. Further increase of the number of indices to five for the KNN classifier increases the overall accuracy to 94.6%.

Classification of Implantable Rotary Blood Pump States With Class Noise

“…More recent studies in identifying ANO pump states have employed the use of classifiers, with the most notable reports in [12], [11] and [23]. In [12], a classification and regression tree (CART) was applied to the investigation of various physiologically significant pump states.…”

“…The investigation of aortic valve opening covered both experimental data on healthy animals and a numerical model with various hemodynamic alterations [11]. Meanwhile, another study [23] was conducted independently to evaluate the performance of ANO detection using animal experimental data with more variability, including variations in cardiac contractility, systemic vascular resistance and total blood volume. Four different types of classification approaches, namely LDA, logistic regression (LR), MLP and k-NN were compared in the study [23].…”

“…Meanwhile, another study [23] was conducted independently to evaluate the performance of ANO detection using animal experimental data with more variability, including variations in cardiac contractility, systemic vascular resistance and total blood volume. Four different types of classification approaches, namely LDA, logistic regression (LR), MLP and k-NN were compared in the study [23]. Both studies [11], [23] reported that the k-NN classifier achieved higher accuracy compared to other classification approaches.…”

“…Four different types of classification approaches, namely LDA, logistic regression (LR), MLP and k-NN were compared in the study [23]. Both studies [11], [23] reported that the k-NN classifier achieved higher accuracy compared to other classification approaches.…”

“…To date, most studies focusing on the occurrence of the ANO pump state [8], [11], [23] have not included the VC pump state in their data set. The presence of the VC pump state would complicate the identification of the ANO pump state, therefore features that are able to differentiate between ANO and VC (apart from VE) are required.…”