Ivan I. Kirov - US grants

Project Summary Traumatic brain injury (TBI) is the world's leading cause of neurological disability in the young adult and middle-aged population. The role of computed tomography (CT) and magnetic resonance imaging (MRI) in patient management is limited due to the widespread presence of clinically-relevant, but imaging-occult injury. We propose new metrics of TBI damage obtained from sodium (23Na) MRI and proton MR spectroscopy (1H MRS). The central hypothesis is that conjoint 23Na MRI/1H MRS will provide metabolic imaging markers for loss of ion homeostasis and neurodegeneration, which will predict patients? long-term clinical outcome. Sodium MRI is a non-invasive modality based on the direct detection of Na+ ions using dedicated MRI software and hardware. It can assess loss of Na+ homeostasis, which can be disturbed by processes where ionic imbalance is the driving force behind the cascade of cell damage. The Na+/K+ exchange pump can be affected by energy deficits due to mitochondrial dysfunction, as well as by diffuse axonal injury, the histopathological signature of TBI. This would result in variations of the intracellular sodium concentration (C1), while inflammation, gliosis and cell loss would be reflected in variations of the extracellular volume fraction (alpha2). We propose to combine the 23Na metrics C1 and alpha2 with 1H MRS, from which quantification of N-acetylaspartate (NAA), creatine (Cr), choline (Cho) and myo-inositol (mI) can be used to infer neuronal health/density (NAA), cellular energy/density (Cr), membrane turnover (Cho) and astrocytic activation (mI). Specific aims are: AIM 1: Methodology. (1.a) To optimize ultrashort echo-time 23Na MRI with and without fluid suppression at 3 T. To optimize image reconstruction with denoising and compressed sensing to increase signal-to-noise ratio, resolution and speed of acquisition. To develop C1 and alpha2 quantification based on 23Na spin dynamics simulation and reference phantoms. (1.b) To implement whole-brain 1H MRS acquisition with 4th order shimming. AIM 2: Clinical application. (2.a) To compare C1 and alpha2 in TBI patients and controls at baseline (?10 days after TBI) and their rates of change over time (2- month and 1-year follow-ups for TBI, 1-year follow-up for controls), adjusted for MRI (FLAIR, SWI, DTI) and 1H MRS. (2.b) To correlate the baseline values of the MR metrics, and their rates of change over time, with clinical outcome at the two follow-ups. To assess the added value of C1 and alpha2 to MRI and 1H MRS. AIM 3: Mechanistic model. (3.a) Given the interdependence between ATP and both NAA and C1, to determine whether TBI-related changes in C1 are glial or neuronal. (3.b) To test whether, based on a mechanistic model of neurodegeneration: (i) baseline 23Na MRI and 1H MRS findings will be moderated by injury severity; (ii) longitudinal 23Na MRI and 1H MRS changes consistent with neurodegeneration will correlate with worse clinical outcome at the two follow-ups. (3.c) To determine which MR measurements (or combinations thereof) are best predictors of long-term TBI outcome, based on the mechanistic model and the longitudinal data.

Project Summary Neurodegeneration, which underlies the progression of Multiple Sclerosis (MS), is poorly reflected by standard clinical and radiological metrics. Brain volume, the most widely used MRI marker for neurodegeneration, measures only irreversible neuronal loss (atrophy). Complementary markers for monitoring the earlier stages of the neurodegeneration process are therefore needed, in order to better predict disease progression and assess the efficacy of treatment with neuroprotective agents. By satisfying a number of biomarker requirements, N-acetyl-aspartate (NAA), a metabolite quantified by proton MR spectroscopy (1H MRS), has the potential to fulfil this role. Unsatisfactory reproducibility and sensitivity, however, have hampered its transition to clinical use. The goal of this work is to address these shortcomings by introducing a robust multiparametric NAA biomarker for neurodegeneration in early-stage MS. Using MRS Fingerprinting (MRSF), a novel method which represents a departure from conventional 1H MRS (cMRS), we will assess neurodegeneration by: (i) more accurate estimates of NAA concentrations, compared to those from cMRS; (ii) measurements of the neuronal microenvironment, in the form of NAA?s T1 and T2 relaxation times. Unlike cMRS, MRSF produces relaxation times (T1, T2) and transmitter inhomogeneity (B1+) per subject for NAA and water. These variables were previously not collected, due to prohibitively long protocols, forcing cMRS studies into using average values from patient cohorts evaluated in other studies, or more commonly, into assuming unchanged values across all subjects and all brain regions. Our group and others, however, have shown that relaxation times of water and metabolites in normal-appearing MS tissue are not only different from control values, but also exhibit regional and inter-subject variations. We therefore hypothesize that by accounting for these factors, and by integrating the relaxation times into a multiparametric classifier, MRSF will provide superior sensitivity compared to cMRS in measuring neurodegeneration. The hypothesis will be tested for recently-diagnosed patients with a relapsing-remitting course, in normal-appearing white matter, as well as in the thalamus, a region recently shown to be sensitive to early neurodegeneration in MS.