For the heart, life is easy. It beats for as long as it can, then it stops.
– Karl Ove Knausgaard, My Struggle
At its most basic level the cardiovascular system is a collection of tubes filled with fluids and a pump to move them. The whole point of this setup is to deliver volumes of blood to regions of the body primarily for the purposes of transporting dissolved gases, nutrients, hormones, and metabolic wastes to places they are needed from places they are not, and regulating the pH balance and ion concentrations of interstitial fluids to ensure a consistent biochemistry throughout the body. Blood coursing through the body performs many other functions including stabilizing body temperatures, defending against the introduction of toxins and the invasion of pathogens, and restricting fluid losses at sites of injury, but for the investigation reported herein, our attention is best focused at on the transportation and the regulation of fluids. The pump and the tubes shall concern us here.
Even for such a simple system, many problems can arise. Problems can arise with the heart (the pump) including coronary artery disease, arrhythmia, heart valve disease, congenital heart disease, cardiomyopathies, pericarditis, aorta disease, Marfan syndrome, and myocardial infarction. Problems can arise with the vasculature (the tubes) including peripheral vascular disease (one of the fast growing causes of mortality world wide (1)), aneurysm, renal artery disease, Raynaud’s syndrome, Bueger’s disease, blood clotting disorders, and lymphedema. Problems can arise with the fluids themselves in the form of anemia, lymphoma, and hemophilia, how the fluids are regulated as is the case in chronic kidney disease, diabetes, and end stage renal disease, and how the fluids are distributed from hypertension to edema. These are but a small fraction of all possible problems, but together these problems represent tens of millions of lives lost and hundreds of billions of dollars spent annually (1). Such costs – in money, in life, in suffering – are high, and growing higher.
Treatments for these and many other conditions are predicated on clinicians’ abilities to diagnose the presence and extent of the conditions themselves. Without a semblance of what is wrong, one would be at great pains to know how to make it right. Thus stands the entire edifice of clinical decision-making on the foundation of tools to measure the regularity and irregularity of patient states. Disorders, diseases, and pathologies manifest themselves in many and multifaceted ways, affecting the body’s anatomy, physiology, chemistry, and functions. The detection of such anomalous behavior lies at the heart of clinical monitoring.
Elsewhere I present one solution within a subset of clinical monitoring. It is a mixed modal approach linking the physiology detecting methods of bioimpedance with the anatomy imaging methods of ultrasound. In their combination is a hybrid method of evaluating parts of the human body in a way heretofore unavailable to clinicians and researchers. This mixed method technique allows for the measurement of long-term structural changes to anatomy (as seen in a clinical study presented [Forthcoming]), rapid, short-term dynamic changes in in anatomy and physiology (as seen in the clinical trials presented in [Forthcoming]), the development of a wearable clinical monitor utilizing clinical results to continuously monitor these structural-functional changes (as discussed in [Forthcoming]), and ultimately in a proposed melding of the two imaging modalities (as presented in the theoretical work of [Forthcoming]). The whole of this work is intended to demonstrate the relatedness of the signals derivable from electrical impedance (bioimpedance) and acoustical impedance (ultrasonography) based techniques. Though the signals correspond in many ways, such as the results determined from the ultrasound measurement of the inferior vena cava and the measure of electrical impedance in the limb, they differ greatly in the ways they are collected, examined, and interpreted. The goal of this work is to demonstrate a pathway towards hybridization of the techniques.
The path taken for this demonstration began with the development and implementation of separate signal and image processing techniques for both bioimpedance and ultrasound. Algorithms and software were created in parallel for each modality to independently verify their accuracy before then synthesizing their results.
I first designed an ultrasound imaging toolbox to work with and analyze post-clinical images in the standard Digital Imaging and Communications in Medicine (DICOM) format. I chose this approach given the need to test this work in many different clinical environments where few researchers and clinicians have access to the raw underlying radiofrequency (RF) data of their ultrasound machines. The other tangible benefit to such an approach is in the creation of a software toolbox that can be used by researchers around the world to identify and measure parameters of interest freely and without the need to modify their existing imaging capabilities. The specifications and validation of this toolbox will be discussed at length in [Forthcoming] and it will suffice to state here that the software was created to make possible the measurement of wall distensibility, local and global tissue strain, and the shear rate of blood across vessel walls – three of the most prominent and significant dynamically varying aspects of the vasculature measurable by ultrasound.
Once mature, the ultrasound software was used to validate the measurements of the algorithms to assess dynamic changes in bioimpedance, ultimately correlating the two with changes in blood volume. The theoretical work proposed in the final chapter suggests how a 2D map of electrical conductivities could be modified by ultrasound imaging to give a holistic picture of the anatomy and physiology of a system under examination.
To understand any measurement technique, one must first understand the system being measured. Only in such a manner can the signals derived from the measurement be understood. That is to say, without context a signal is without meaning. For us to impart meaning on the work presented here, let us first understand the environment in which it is collected. Broadly speaking this work was conducted in two general environments: the pre-clinical and the clinical. The preclinical work consisted of computation fluid dynamic simulation of blood flowing through peripheral vasculature, particle image velocimetric validation of such simulations as a standard for our developed software algorithms, and in the design and calibration of a wearable monitoring device. The software and hardware components developed were then tested in one of four clinics at the University of Michigan depending on the study: the transplant unit (for follow up ultrasound measurements of fistula), the emergency medicine department (for the initial ultrasound correlation with DRIVE), the in-hospital dialysis unit (to track large volume shifts via ultrasound and impedance), and the intensive care unit (to validate converse pressure relationships). In so doing, I have endeavored to move these techniques and this technology from the bench to the bedside as quickly as possible, validating the two modes in each environment. For all of this, ultimately generating each and every signal examined here – from the strain and wall shear measurements in the fistula maturation study ([Forthcoming]) to the volume status monitoring in the dialysis clinic ([Forthcoming]) – were the hearts of each and every patient. Let us begin then at the heart of the matter.