Oded Gonen - US grants

Chemical Shift Imaging (CSI) is the method of choice for in vivo localized spectroscopy. However, several persistent obstacles still remain in its application to nuclei other than protons, specifically 31P and 19F CSI proposed in projects of this application. First, the sensitivity is low. Second and several consequents, the measurements are long, thus, only one nucleus can be observed in a clinical session, even if information from another is significant and sought. Third, the current Fourier reconstruction methods, widely used because of their computational convenience and simplicity, result in an artifact that intermixes signal from one voxel, with its its neighbors. This artifact, known as A voxel bleed, leads to spectral-contamination and impacts CSI in general. We propose to address these difficulties from three angles. First, to use polarization transfer techniques from 1H to the less sensitive 31P to increase the signal of the latter and to be used in project II. Second, to develop simultaneous acquisition techniques that will allow 1H decoupled 31P and 19F CSI acquisition during a single clinical examination. Third, to develop non-Fourier transform techniques that, using prior knowledge from the proton images, will improve the accuracy of CSI reconstruction.

DESCRIPTION (Adapted from Applicant's Abstract): Contrast-enhanced magnetic resonance imaging (MRI) is becoming increasingly important in the evaluation of multiple sclerosis (MS) based on its sensitivity to acute, often subclinical events in the brain and its ability to measure the accumulation of the disease over time. However, to date there are conflicting reports concerning the number and volume of these lesions, "the load of the disease" and the neurological severity. To resolve these conflicts, recent studies employed proton (1H) magnetic resonance spectroscopy (MRS) to investigate the lesions underlying metabolism. They showed that it may be possible to assess the irreversibility of central nervous system injury, discern between edematous and demyelination lesions and (perhaps) predict the time course of their evolution. However, these studies employed the current art of single voxel or small 2D arrays of few tens voxels, which, due to their limited observable volume(s) of interest, must be image-guided onto the pathologies of interest. This restricted them to the study of (i) MRI detectable existing lesions; and (ii) few foci in a single clinical session due to time and voxel-size constraints. The applicants proposed to overcome both these problems with their new three dimensional (3D) 1H-MRS hybrids to achieve simultaneous coverage of most the white-matter volume in clinically feasible time: ~45 min. The 3D hybrids yield ~1000 high resolution, <0.8 cm3 , voxels per exam from 0.5-0.75 liter of brain tissue. Such extensive coverage will enable us to examine: (i) whether lesions develop in white matter regions which are already metabolically abnormal; and (ii) whether 1H-MRS-detected local metabolic changes, primarily neuronal loss in conjunction with demyelination, can predict the onset of a lesion before contrast-enhanced MRI, determining its type and course. These two hypotheses will be investigated by following cohort of 12 relapsing-remitting MS patients over three years with 3D 1H-MRS every 3 months. The MRS data will be compared with (i) localized 1H metabolic levels in age, sex and race matched healthy volunteers; (ii) the contrast enhanced MRI which they undergo every 6 months during that period; and (iii) their own previous local metabolic levels. The 1H-MRS data will be obtained both at the highest, 4 Tesla, magnetic field approved for human use, for maximal sensitivity and at the common clinical filed of 1.5 Tesla.

DESCRIPTION (Verbatim from the Applicant's Abstract): During the past few years, numerous in vivo brain proton (1H) magnetic resonance spectroscopy (MRS) Studies have suggested a correlation between pathology and the observable metabolite levels in cancer, epilepsy, Alzheimer's disease, multiple sclerosis, HIV infection, stroke and other disorders of the central nervous system. Unfortunately 1H-MRS still suffers from (a) low voxel signal-to-noise ratio at the desired, sub 1 ml, spatial resolution obtained at a "clinically reasonable" 40 min. session; (b) spectral contamination by signals from extraneous tissue, due to limitations of current localization methods, especially at high, >3 Tesla, magnetic fields; and (c) scarcity of software to design and interpret results from these experiments. To address these, the first goal of this project is to develop and optimize echo and non-echo 3D localization methods based exclusively on Hadamard spectroscopy imaging (HIS). Using 3D HIS in all directions with either surface or volume-coils will: (a) improve the voxels' profile, hence, intervoxel isolation; (b) intrinsically suppress extraneous contamination; (c) be simple to implement, post-process and display; (d) be suitable for 1H-MRS even at very high, >3 T, magnetic fields, i.e., immune to the chemical shift artifacts associated with selective-pulses; and (e) require fewer encoding steps, thus, execute faster than chemical-shift-imaging based methods. The second goal is to produce platform-independent image-guided radio-frequency pulse-design and data post-processing software for the first goal. The purpose and guiding philosophy is to ensure that the deliverables from this project: pulse-timing sequence templates, pulse design tools, post processing and display software, can be straightforwardly implemented on any modern imager and computing platform. This work will extend both the clinical and biomedical research applications of 1H-MRS by providing increased spatial resolution, shorter acquisition time and 3D localization sequences suited for clinical, <2 Tesla, as well as high, >3 T, magnetic fields. The development will support ongoing studies investigating changes in brain metabolites associated with multifocal and diffuse brain disorders, specifically, multiple-sclerosis, aging, brain trauma, AIDS and pediatric neurodegenerative diseases.

[unreadable] DESCRIPTION (provided by applicant): Changes in metabolic levels observed with proton-magnetic-resonance-spectroscopy ('H-MRS) are frequently used to augment the highly sensitive (but not specific) MRI. At 1.5 Tesla, MRS has so far linked anatomy from MRI with the underlying metabolism in cancer, epilepsy, Alzheimer's disease, multiple sclerosis, HIV infection, stroke and other disorders of the central nervous system (CNS). Therefore, it was anticipated that high-magnetic-field (Bo ) 1.5 T) imagers would provide 'H-MRS a much needed boost in sensitivity. Unfortunately, that has not happened: Translating the most useful 2 and 3 dimensional (3D) 'H-MRS techniques to high-Bos has been stymied by high radio-frequency (BI) power requirements; short TZ reducing the signal-to-noise-ratio (SNR) gain; chemical shift misregistration errors; and scarcity of software to shim, design 3D localization, evaluate and display the large 3D MRS data sets. [unreadable] [unreadable] The long term goal of this competing continuation is to develop and implement 30 'H-MRS methods to address the above issues and perform better at higher BG. They will be extensions of the hybrid techniques developed in the past two cycles. Specific Aim 1 will exploit the shorter T2s for: (a) 3D coverage by optimal interleaving of "slabs," across the volume-of-interest, each thin enough to excite with the available B1 under strong gradients to minimize the misregistration error; and (b) increase the spatial resolution to ((1 ~m )~voxels, and extend coverage all the way to the skull. Specific Aim 2 will introduce 3D localization into our non-echo, non-localized, whole head 'H-MRS, to provide a rapid, imager-side, lower-resolution "metabolic localizer." Specific Aim 3 will utilize the increased sensitivity to produce very high spatial resolution, (0.5 - 0.375 ~m )~ voxels, in restricted regions, e.g., the optic nerve or spinal cord, both of which are currently inaccessible. Finally, Specific Aim 4 will develop novel methods to detect and visualize relationships between different metabolites' spatial distributions to simplify the staggering, often confusing, amount of information generated by 3D 'H MRS. [unreadable] [unreadable] The health relatedness of this project is its extension of increased spatial resolution, volume covered and shorter acquisition time 3D 'H-MRS methods to high, 23 T, magnetic fields, to support ongoing and future studies of CNS metabolism associated with multifocal and diffuse diseases.

DESCRIPTION (provided by applicant): Brain metastases present in 50-80 percent of small-cell-lung-cancer (SCLC) survivors within two years. To reduce this risk and improve their outcome, prophylactic-cranial-irradiation (PCI) is now offered to certain SCLC patients even in the absence of distinct visible brain pathology. Although neurotoxicity is always a concern in brain radiation-therapy (RT), there is currently no direct method to quantify its damage to the central nervous system (CNS). Such knowledge is critical for (a) risk/benefit assessment; and (b) dose determination. Presently, such damage can only be assessed indirectly, using neurocognitive tests. Unfortunately, the results of such tests are often confounded by other factors such as language barriers, patients' state of mind and/or their level of fatigue, fear and depression. Clearly, an objective, i.e., preferably instrumental, non-invasive and, most importantly, sensitive method to quantify RT neurotoxicity is necessary. We propose to quantify the extent of neurona1 cell loss imparted to the brain by RT through the decline of the amino acid derivative N-acetylaspartate (NAA) using state-of-the-art proton magnetic resonance spectroscopy (1H-MRS). Since NAA is believed to be present in neuronal cells only, its amount is proportional to their number and/or integrity. Consequently, we will obtain the amount of whole-brain-NAA (WBNAA) in 40 patients pre, immediately post- (2-3 weeks later) and six months after whole-brain radiation-therapy (WBRT). Since we will evaluate the amount of NAA in the entire brain, its signal-to-noise-ratio (SNR) will be excellent, facilitating short, < 15 min. examinations. It will also not be susceptible to misregistration errors that currently beset serial studies, nor will it be sensitive to the local transient edema common in WBRT. The WBNAA measurements will be augmented by the current tool used to evaluate CNS injury - the mini-mental status examination (MMSE) for correlation and comparison. We will use these observations to test the following three hypotheses, H1- H3: H1: That WBRT induces neuronal injury quantifiable with WBNAA in these patients. H2: That WBNAA is more sensitive than MMSE to detect neuronal injury consequences of WBRT. H3: That this neuronal injury may be transient, in part, and could resolve within several months after WBRT.

Clinical T2 and contrast-enhanced Tl-weighted magnetic resonance imaging (MRI) have become the diagnostic modalities of choice in the evaluation of multiple sclerosis (MS) due to their sensitivity to acute, often subclinical events in the brain and their ability to measure the accumulation of the disease over time. MRI, however, lacks specificity, in that to date, there are conflicting reports concerning the number and volume of these lesions, "the load of the disease," and the associated neurological deficits. Furthermore, clinical MRI is completely blind to occult white matter (WM) and most gray matter (GM) pathology. The need for more reliable surrogate markers frequently leads to proton magnetic resonance spectroscopy 1HMRS being used to probe the underlying metabolism of normal appearing WM (NAWM) and the lesions in it. Since MS pathogenesis starts on molecular cellular levels, we propose to use state of the art, short echotime, three-dimensional high-spatial resolution local and global 1H-MRS methods to examine three hypotheses: HI: That MS lesions develop in WM regions which are already metabolically abnormal; H2: Abnormal metabolic activity persists in NAWM and lesions even absent Gd-enhancement (the current marker of "activity"); and H3: That global whole-brain quantification of the decline rate of the neuronal cell marker N-acetylaspartate (NAA) reflects the aggressiveness of a patient's disease, and therefore, could forecast its future course in that individual. These hypotheses will be tested by following a cohort of 25 relapsingremitting (RR) MS patients and 25 matched controls, for 5 years, with three Specific Aims: Specific Aim 1 is to perform longitudinal follow-up of the overall metabolites levels in the NAWM to establish markers for current MRI-occult disease activity. Specific Aim 2 is to follow localized metabolism in NAWM to determine what focal changes preceded lesion formation and determine those lesions' outcome - to repair or become chronic. Specific Aim 3 is to correlate the NAA levels of the whole brain and its WM and GM fractions with the patients' clinical deficits to establish the NAA as a forecaster of disease severity and future course. The health relatedness of this study is its potential to establish, quantify and validate non-invasive radiological metabolic surrogate markers of RR MS progression that will enable us to: (i) Gauge current level of disease activity, thereby, (ii) increase our capability to forecast its future course for these (young) patients; and consequently (iii) lead to improved monitoring of response in drug and treatment trials.

DESCRIPTION (provided by applicant): Metabolic changes observed with proton-magnetic-resonance-spectroscopy (1H-MRS) often augment the highly sensitive but not specific MRI. Indeed, at 1.5 Tesla, 1H-MRS has so far linked anatomy from MRI with underlying metabolism in cancer, Alzheimer's and Parkinson's diseases, MS, HIV, epilepsy, stroke trauma and other neurological and psychiatric disorders. It was anticipated, therefore, that high, B0 e3 T, magnetic-fields would provide 1H-MRS a much needed boost in sensitivity, spectral and spatial resolution. That unfortunately, did not happen despite their proliferation in number, :300 installed, and field strength, up to 9.4 T. Translation of the most useful two and three dimensional (2D, 3D) 1H-MRS techniques to high-fields has been stymied by: (i) High radio-frequency (B1) power requirements and heat deposition;(ii) short T2s, reducing the signal-to- noise-ratio (SNR) gain;(iii) chemical shift displacement errors;and (iv) lack of software to evaluate and display the large data sets. Consequently, efficient, reliable 3D multivoxel techniques are not offered by instrument manufacturers, who traditionally shift this onus onto publicly-funded academic research. The long term goal of this competing continuation, therefore, is to develop methods to address issues i - iv to perform 3D 1H-MRS at higher B0s, and realize the advantages for clinical research. Our response to these problems is to extend to 3 and 7 T our successful hybrid techniques. Specific Aim 1 is to exploit the shorter T2s to enhance the SNR and acquisition efficiency of 3D coverage by optimal interleaving across the volume-of-interest (VOI), multiple slabs of several slices each. Specific Aim 2 is to overcome the declining B1 fields per watt RF power with shifted-Hadamard pulses that need the B1 of just one slice to sequentially excite several. This will lower the peak and deposited power under very strong selective gradients and reduce the chemical shift displacement. Specific Aim 3, is to recover the SNR lost to shorter T2s at high B0s with non-echo sequences using 3D transverse and longitudinal-Hadamard encoding to define the VOI. Finally, Specific Aim 4 is to develop new post-processing methods to detect and visualize relationships between different metabolites'spatial distributions to simplify the daunting amounts of 3D 1H MRS data.PROJECT NARRATIVE This project will lead to increases in the amount of human brain volume covered in an exam, improve the localization accuracy as well as spatial and spectral resolution and shorten the acquisition time for proton spectroscopy at higher magnetic fields. These capabilities will enhance studies of the underlying metabolism of devastating (but frequently MRI-invisible or of non-specific finding) neurological diseases in the human brain and spine and may also improve our capability to monitor the effectiveness of their treatment(s).

DESCRIPTION (provided by applicant): The conference we propose is entitled "Neuroimaging of MS - State-of-the-art and Future Directions" and will be held on Friday and Saturday, November 6 and 7, 2009 in Washington DC. This conference is relevant to the missions of both the NINDS and NIBIB. The target audience is interdisciplinary: PhD and MD (radiology and neurology) experts in the fields of multiple sclerosis (MS) imaging and management. MS is a chronic disease afflicting the central nervous system (CNS) of young adults and can last decades. MR methods therefore are ideally suited for its study due to their ability to image non-invasively CNS tissue, metabolism and function. MS is studied by PhD's who develop imaging technologies, neuro-radiologists who apply them and neurologists who treat the patients and design treatment trials. This conference is designed, therefore, (1) to acquaint these three diverse groups with the current state-of-the-art MR technology for MS;(2) to identify the techniques that are most likely to become suitable for routine clinical application;and (3) to reach a consensus as to which critical steps are needed to expedite and foster this translation. This would ultimately result in improved care and outcomes for a greater number of MS patients (400,000 currently in the US and 2.5 million worldwide). The conference will span two days. The first day will consist of a series of overview presentations introducing the current state-of-the-art MR imaging of the brain and spine in MS and then focus on the clinical applications in this disease. The second day will commence with one additional series of overview presentations focusing on the current "most ambitious "art, followed by three concurrent workshops. Each workshop will be offered twice to enable attendees to participate in more than one. The workshops will each discuss one of the three most advance imaging modalities - "Molecular imaging", "Quantitative MRI" and "MR at the extremes" (high magnetic fields, difficult CNS regions). For each of the three areas addressed, three sets of recommendations will be clearly stated at the end: 1) Which method(s) is/are most likely to be ready for clinical use within the next 5 years. 2) What the researchers should do to help address the issues raised during the discussions regarding that topic in order to expedite this transition;and 3) what could the NIH do in terms of requests for applications (RFAs). These recommendations will constitute the core of the proceedings from the meeting, and will be published concurrently in the American Journal of Neuroradiology and in Multiple Sclerosis. The organizing committee endeavored to ensure a diversity among the invited speakers who comprise a mix of established and junior clinical and research scientists from various geographic regions in North America and Europe. Selection of invited junior scientists for whom funding is requested was merit based. PUBLIC HEALTH RELEVANCE: The three goals of the conference are: (1) to acquaint MS researchers and clinicians of different disciplines (PhDs, neuroradiologists and neurologists) with the state-of-the-art MR methodology in multiple sclerosis;(2) to reach a consensus which of these methodologies is both needed and is mature enough (in the laboratory) for translation into the clinical environment;and (3) to determine what is needed to standardize these "ready" methodologies to foster and expedite this translation. The outcome will be coherent focused progress that will benefit a greater number of MS patients, sooner and ultimately result in improved care and management for the 400,000 MS patients in the US and the over 2.5 million patients worldwide.

Clinical T2 and contrast-enhanced T1-weighted MRI have become the diagnostic modalities of choice to evaluate multiple sclerosis (MS) due to their sensitivity to acute, often subclinical events in the brain and their ability to measure the accumulation of lesions over time. MRI, however, lacks specificity, in that to date, there are conflicting reports concerning the number and volume of these lesions, “load of the disease,”and the neurological deficits. Furthermore, clinical MRI is blind to occult white matter (WM) and especially most gray matter (GM) pathology. The need for more reliable surrogate markers frequently leads to proton magnetic resonance spectroscopy (1H-MRS) being used to probe the underlying metabolism of normal appearing WM (NAWM) and its lesions. Since MS pathogenesis starts on molecular cellular levels, we propose to use state of the art, short echo-time, three-dimensional high-spatial resolution local and global 1H-MRS methods to examine three hypotheses: H1: That MS lesions develop in WM regions which are already metabolically abnormal;H2: Abnormal metabolic activity persists in NAWM and lesions even absent Gd-enhancement (the current marker of “activity”);and H3: That global whole-brain quantification of the decline rate of the neuronal cell marker N-acetylaspartate (NAA) reflects the aggressiveness of a patient’s disease, and therefore, could forecast its future course in that individual. These hypotheses will be tested by continuing to follow a cohort of 25 relapsing-remitting (RR) MS patients and 25 matched controls, for 5 more years, with three Specific Aims: Specific Aim 1 is to perform longitudinal follow-up of the overall metabolites levels in the NAWM to establish markers for current MRI-occult disease activity. Specific Aim 2 is to follow localized metabolism in NAWM to determine what focal changes preceded lesion formation and determine those lesions outcome - to repair or become chronic. Specific Aim 3 is to correlate the NAA levels of the whole brain and its WM and GM fractions with the patients'clinical deficits to establish the NAA as a forecaster of disease severity. The health relatedness of this study is its prospect to establish, quantify and validate non-invasive metabolic surrogate markers of RR MS progression that will enable us to: (i) Gauge current level of disease activity, thereby, (ii) increase our capability to forecast its future course for these (young) patients;and consequently (iii) lead to improved monitoring of response in drug and treatment trials.