Single-sided magnetic resonance sensors for clinical detection of volume status

Abstract

Several pathological processes affect the body's ability to regulate volume status. In each of these disease states, the body loses some ability to regulate fluid balance and maintain an euvolemic state. Deviations from euvolemia have been shown to increase morbidity and mortality. The ability to detect pre-symptomatic changes in volume status would allow for more responsive management of these conditions and prevention of higher-mortality complications. Direct evaluation and quantification of early-stage changes in volume state is not currently a clinical measure. T2 relaxometry, a magnetic resonance technique, may offer a feasible method to quantify volume status. In this work we explore the design of single-sided magnetic resonance sensors for the quantification of volume status, evaluate the clinical performance of the sensor, and elucidate further physiological considerations for fluid diagnostics with clinical MRI.

The primary research question that motivated this thesis is: can a point-of-care relaxometer accurately distinguish muscle interstitial fluid shifts in a single measurement? Several approaches are used to answer this question including instrumentation development, signal acquisition studies, and clinical studies. We describe the design of a point-of-care, single-sided magnetic resonance relaxometer. The constructed sensor can acquire slice-selective signal from 8mm above the instrument’s surface with a high signal-to-noise ratio. We detail instrument performance on phantoms, ex-vivo tissue, and human participants. Preliminary observational clinical studies of two cohorts— ‘euvolemic’ athletes and hospitalized end-stage kidney disease patients treated with hemodialysis— were conducted and validate the instrument can detect signal selectively from the muscle interstitial compartment and distinguish ‘euvolemic’ adults and those with end stage kidney disease. We discuss the implementation of multi-exponential fitting of acquired data. This enables analysis of individual muscle tissue compartments. We demonstrate strategies to double signal magnitude and improve T2 fitting accuracy through the simulation and implementation of linear frequency swept adiabatic radiofrequency pulses. These decrease the sensitivity of applied RF pulses to B1 and B0 inhomogeneity. Finally, we explore physiological considerations for the instrument’s clinical implementation with an MRI study of patients with chronic kidney disease (CKD) and healthy control participants. This allows for the evaluation of physiological factors which may affect the sensor’s accuracy in clinical outcome prediction and offer further future areas for study.

The single-sided magnetic resonance sensor and signal acquisition and processing techniques described demonstrate high potential for quantitative clinical assessment of volume status. This work focuses exclusively on ‘euvolemic’ adults or adults with CKD, but the principles demonstrated are agnostic to many underlying disease pathologies.