Tânia Pereira

Novel Methods for Pulse Wave Velocity Measurement

The great incidence of cardiovascular (CV) diseases in the world spurs the search for new solutions to enable an early detection of pathological processes and provides more precise diagnosis based in multi-parameters assessment. The pulse wave velocity (PWV) is considered one of the most important clinical parameters for evaluate the CV risk, vascular adaptation, and therapeutic efficacy. Several studies were dedicated to find the relationship between PWV measurement and pathological status in different diseases, and proved the relevance of this parameter. The commercial devices dedicate to PWV estimation make a regional assessment (measured between two vessels), however a local measurement is more precise evaluation of artery condition, taking into account the differences in the structure of arteries. Moreover, the current devices present some limitations due to the contact nature. Emerging trends in CV monitoring are moving away from more invasive technologies to non-invasive and non-contact solutions. The great challenge is to explore the new instrumental solutions that allow the PWV assessment with fewer approximations for an accurately evaluation and relatively inexpensive techniques in order to be used in the clinical routine.

Characterization of a double probe for local pulse wave velocity assessment.

Local pulse-wave velocity (PWV) is an accurate indicator of the degree of arteriosclerosis (stiffness) in an artery, providing a direct characterization of the properties of its wall. Devices currently available for local PWV measurement are mainly based on ultrasound systems and have not yet been generalized to clinical practice since they require high technical expertise and most of them are limited in precision, due to the lack of reliable signal processing methods. The present work describes a new type of probe, based on a double-headed piezoelectric (PZ) sensor. The principle of PWV measurement involves determination of the pulse transit time between the signals acquired simultaneously by both PZs, placed 23 mm apart. The double probe (DP) characterization is accomplished in three main studies, carried out in a dedicated test bench system, capable of reproducing a range of clinically relevant properties of the cardiovascular system. The first study refers to determination of the impulse response (IR) for each PZ sensor, whereas the second one explores the existence of crosstalk between both transducers. In the last one, DP time resolution is inferred from a set of three different algorithms based on (a) the maximum of cross-correlation function, (b) the maximum amplitude detection and (c) the zero-crossing point identification. These values were compared with those obtained by the reference method, which consists of the simultaneous acquisition of pressure waves by means of two pressure sensors. The new probe demonstrates good performance on the test bench system and results show that the signals do not exhibit crosstalk. A good agreement was also verified between the PWV obtained from the DP signals (19.55 ± 2.02  ms(-1)) and the PWV determined using the reference method (19.26 ± 0.04  ms(-1)). Although additional studies are still required, this probe seems to be a valid alternative to local PWV stand-alone devices.

Machine learning techniques for arterial pressure waveform analysis

The Arterial Pressure Waveform (APW) can provide essential information about arterial wall integrity and arterial stiffness. Most of APW analysis frameworks individually process each hemodynamic parameter and do not evaluate inter-dependencies in the overall pulse morphology. The key contribution of this work is the use of machine learning algorithms to deal with vectorized features extracted from APW. With this purpose, we follow a five-step evaluation methodology: (1) a custom-designed, non-invasive, electromechanical device was used in the data collection from 50 subjects; (2) the acquired position and amplitude of onset, Systolic Peak (SP), Point of Inflection (Pi) and Dicrotic Wave (DW) were used for the computation of some morphological attributes; (3) pre-processing work on the datasets was performed in order to reduce the number of input features and increase the model accuracy by selecting the most relevant ones; (4) classification of the dataset was carried out using four different machine learning algorithms: Random Forest, BayesNet (probabilistic), J48 (decision tree) and RIPPER (rule-based induction); and (5) we evaluate the trained models, using the majority-voting system, comparatively to the respective calculated Augmentation Index (AIx). Classification algorithms have been proved to be efficient, in particular Random Forest has shown good accuracy (96.95%) and high area under the curve (AUC) of a Receiver Operating Characteristic (ROC) curve (0.961). Finally, during validation tests, a correlation between high risk labels, retrieved from the multi-parametric approach, and positive AIx values was verified. This approach gives allowance for designing new hemodynamic morphology vectors and techniques for multiple APW analysis, thus improving the arterial pulse understanding, especially when compared to traditional single-parameter analysis, where the failure in one parameter measurement component, such as Pi, can jeopardize the whole evaluation.

An automatic method for arterial pulse waveform recognition using KNN and SVM classifiers

AbstractThe measurement and analysis of the arterial pulse waveform (APW) are the means for cardiovascular risk assessment. Optical sensors represent an attractive instrumental solution to APW assessment due to their truly non-contact nature that makes the measurement of the skin surface displacement possible, especially at the carotid artery site. In this work, an automatic method to extract and classify the acquired data of APW signals and noise segments was proposed. Two classifiers were implemented: k-nearest neighbours and support vector machine (SVM), and a comparative study was made, considering widely used performance metrics. This work represents a wide study in feature creation for APW. A pool of 37 features was extracted and split in different subsets: amplitude features, time domain statistics, wavelet features, cross-correlation features and frequency domain statistics. The support vector machine recursive feature elimination was implemented for feature selection in order to identify the most relevant feature. The best result (0.952 accuracy) in discrimination between signals and noise was obtained for the SVM classifier with an optimal feature subset .

Pulse pressure waveform estimation using distension profiling with contactless optical probe

The pulse pressure waveform has, for long, been known as a fundamental biomedical signal and its analysis is recognized as a non-invasive, simple, and resourceful technique for the assessment of arterial vessels condition observed in several diseases. In the current paper, waveforms from non-invasive optical probe that measures carotid artery distension profiles are compared with the waveforms of the pulse pressure acquired by intra-arterial catheter invasive measurement in the ascending aorta. Measurements were performed in a study population of 16 patients who had undergone cardiac catheterization. The hemodynamic parameters: area under the curve (AUC), the area during systole (AS) and the area during diastole (AD), their ratio (AD/AS) and the ejection time index (ETI), from invasive and non-invasive measurements were compared. The results show that the pressure waveforms obtained by the two methods are similar, with 13% of mean value of the root mean square error (RMSE). Moreover, the correlation coefficient demonstrates the strong correlation. The comparison between the AUCs allows the assessment of the differences between the phases of the cardiac cycle. In the systolic period the waveforms are almost equal, evidencing greatest clinical relevance during this period. Slight differences are found in diastole, probably due to the structural arterial differences. The optical probe has lower variability than the invasive system (13% vs 16%). This study validates the capability of acquiring the arterial pulse waveform with a non-invasive method, using a non-contact optical probe at the carotid site with residual differences from the aortic invasive measurements.

Cardiovascular risk analysis by means of pulse morphology and clustering methodologies

The purpose of this study was the development of a clustering methodology to deal with arterial pressure waveform (APW) parameters to be used in the cardiovascular risk assessment. One hundred sixteen subjects were monitored and divided into two groups. The first one (23 hypertensive subjects) was analyzed using APW and biochemical parameters, while the remaining 93 healthy subjects were only evaluated through APW parameters. The expectation maximization (EM) and k-means algorithms were used in the cluster analysis, and the risk scores (the Framingham Risk Score (FRS), the Systematic COronary Risk Evaluation (SCORE) project, the Assessing cardiovascular risk using Scottish Intercollegiate Guidelines Network (ASSIGN) and the PROspective Cardiovascular Münster (PROCAM)), commonly used in clinical practice were selected to the cluster risk validation. The result from the clustering risk analysis showed a very significant correlation with ASSIGN (r=0.582, p<0.01) and a significant correlation with FRS (r=0.458, p<0.05). The results from the comparison of both groups also allowed to identify the cluster with higher cardiovascular risk in the healthy group. These results give new insights to explore this methodology in future scoring trials.

Empirical mode decomposition for self-mixing Doppler signals of hemodynamic optical probes.

A new type of optical probe based on laser Doppler self-mixing technology, for a truly non-contact measurement in a single location, and extraction of the temporal features of the distension wave in the arterial wall, was developed. The monitoring of temporal features allows the assessment of cardiovascular function when measurement is carried out at the carotid artery. An algorithm based on the short-time Fourier transform and empirical mode decomposition was applied to the test setup self-mixing signals for the determination of waveform features, with an accuracy of a few milliseconds and a root mean square error less than 3 ms. In vivo testing signals show great consistency in the measured pulse pressure waveform.

Visible and infrared optical probes for hemodynamic parameters assessment

Four optical probes were developed to measure the arterial distension waveform generated by the ventricular contraction and assess clinically relevant information. The pressure wave propagates through the arterial tree and can be measured in the peripheral arteries. The probes make use of two distinct photo-detectors: planar and avalanche photodiodes. Independently, two different light sources were tested: visible and infrared light. Performance of the probes was evaluated in a test setup that simulates the fatty deposits commonly seen in the obese, between skin and the artery. The probes show good overall performance in the test setup with less than 8% root mean square error (RMSE). However, the probes lit with IR sources show better results for the more extreme cases, with a better resolution in the waveform, higher definition of notable points and higher SNR when compared to the visible source signals. In vivo, the IR probes allow easier waveform detection, even more relevant with the increasing of the deposit structures.

Comparison of Low-Cost and Noninvasive Optical Sensors for Cardiovascular Monitoring

New optical probes are developed for carotid distention waveform measurements, in order to assess the risk of cardiovascular diseases. The probes make use of two distinct photodetectors: planar and avalanche photodiodes. Their performance is compared for visible and infrared (IR) light wavelengths. The test setup designed for the evaluation of the probes simulates the fatty deposits commonly seen in the obese people, between skin and the artery. The performed tests show that the attenuation of the signal is lower for the IR light, with higher penetration and better resolution in the captured distension waveform, with higher definition in morphological features on the wave and higher signal-to-noise ratio when compared to the visible source signals. The probes show good overall performance in the test setup with a root mean square error lower than 8%. In vivo, the IR probes allow easier waveform detection, even more relevant with the increasing deposit structures.

Programmable Test Bench for Hemodynamic Studies

The non-invasive assessment of hemodynamic parameters has been a permanent challenge posed to the scientific community. The literature shows many contributions to this quest expressed as algorithms dedicated to revealing some of its characteristics and as new probes or electronics, featuring some enhanced instrumental capability that can improve their insight.