Joseph A. Maldjian - US grants

DESCRIPTION (provided by applicant): The objective of this Phase I project is to provide the neuroscience research community with a unique investigative tool for the seamless real-time integration of data from emerging imaging modalities (fMRI, PET, SPECT, MRS, and diffusion imaging). This tool will allow analysis of an enormous array of biological processes involving multiple cross-platform modalities recording activation, stimulation, biochemistry and tissue contrast. The proposed tool is IDLSPM, a software solution created by the PI incorporating an advanced computational architecture for distributed processing, analysis and visualization of multimodal imaging data. In this application we propose to incorporate Biologic Parametric Mapping (BPM), a new form of functional imaging data analysis that will integrate the information available from multiple functional imaging modalities and combine them into a user-friendly highly extensible and scaleable software environment. The basic idea behind BPM is to allow the probing of functional imaging data using the results of any other form of functional imaging data. The IDLSPM project represents a complete brain imaging solution that will provide the functional imaging community with unparalleled computational capabilities, integrated neuroanatomic and cytoarchitectonic atlases, ease of use, and tight integration to native SPM, allowing rapid advancement in both research and clinical implementations of functional imaging. In the biologic portion of our effort we will apply the neuroinformatics tools developed in this project towards studying the neurofunctional correlates of dyslexia. Developmental dyslexia is a common neurobehavioral disorder affecting a large number of people. Despite the sizeable prevalence of the disorder, there is relatively little known about its neural basis. Recent multimodal imaging data has implicated the left occipitotemporal and temporoparietal areas as regions of abnormality in dyslexics. The tools developed in this project will be used to uncover the neural functional and chemical relationships of dyslexia using crossmodal sensory tasks, diffusion tensor imaging, and 3D spectroscopy. Our informatics goal is to l) extend the functionality of this tool into a multimodal imaging environment for fMRI, PET, SPECT, diffusion imaging, and 3D spectroscopy data, and 2) incorporate biologic parametric mapping capabilities for probing functional MR data with multimodal imaging data. The goal of the biologic element of our project is to use BPM to determine if diffusion and spectroscopy changes in the occipitotemporal and temporoparietal regions are related to functional MRI changes observed in cross-modal sensory processing in dyslexics.

[unreadable] DESCRIPTION (provided by candidate): [unreadable] The objective of this supplement is to provide a structured year of research training to a clinical resident [unreadable] under the mentorship of a NIBIB funded investigator. The parent grant for this supplement is composed of two components: a neuroinformatics component and a biologic component. The neuroinformatics [unreadable] component of the parent grant will develop multiple toolboxes including a biologic parametric mapping [unreadable] toolbox (BPM) for the statistical analysis of multimodal image data to be used in the current supplement. In this supplement, the BPM toolbox will be used to combine information from diffusion imaging and perfusion imaging in a diabetic patient population. The overarching hypothesis for Dr. Kaufman's research project is that changes in white matter in diabetic patients will be correlated with decreased perfusion in a regionally specific manner. Dr. Kaufman will be trained in the use of the BPM toolbox and he will implement the analyses of the data using the neuroinformatics tools from the parent project. The educational training plan proposed for Dr. Kaufman has four elements: 1) development of knowledge through didactic course work in statistics, neuroscience, as well as attending journal club meetings and institutional seminars; 2) mastery of brain imaging research skills through study design, data collection, and data analyses under mentor guidance; 3) introduction to translational genomics research through participation in a practical laboratory rotation in the Human Genomics Center at Wake Forest University; and 4) training in research ethics and medical conduct. [unreadable] [unreadable]

[unreadable] DESCRIPTION (provided by candidate): [unreadable] The objective of this supplement is to provide a structured year of research training to a clinical resident under the mentorship of a NIBIB funded investigator. The parent grant for this supplement is composed of two components: a neuroinformatics component and a biologic component. The neuroinformatics component of the parent grant will develop multiple toolboxes including a biologic parametric mapping toolbox (BPM) for the statistical analysis of multimodal image data to be used in the current supplement. In this supplement, the BPM toolbox will be used to perform a multiple variable regression analysis to evaluate age-related relationships between functional MRI activation and structural anatomy. The overarching hypothesis for Dr. Deibler's research project is that age-related changes in grey matter volume are predictive of functional brain activity in a regionally specific manner. This analysis will control for differences in grey matter concentration. Dr. Deibler will be trained in the use of the BPM toolbox and he will implement the [unreadable] analyses of the data using the neuroinformatics tools from the parent project. The educational training plan proposed for Dr. Deibler has four elements: 1) development of knowledge through didactic course work in statistics, neuroscience, as well as attending journal club meetings and institutional seminars; 2) mastery of brain imaging research skills through study design, data collection, and data analyses under mentor guidance; 3) introduction to basic neurophysiologic research in the Department of Neurobiology and Anatomy at Wake Forest University; and 4) training in research ethics and medical conduct. [unreadable] [unreadable]

[unreadable] DESCRIPTION (provided by applicant): The WFU_Pickatlas software toolbox (pickatlas) is a highly successful package written in Matlab for ROI-based (region of interest) analyses of brain imaging data. It provides a powerful method of probing virtually any type of brain imaging data using automatically generated masks based on lobar anatomy, cortical and subcortical anatomy, and Brodmann areas. The pickatlas was the first tool available to the neuroimaging community for performing automated hypothesis-based ROI analyses, and it remains a highly used tool. It was designed to be used with SPM, and requires the statistical analyses to be processed through SPM. Since the original release of the pickatlas, there have been rapid advances in ROI-based methods of analyzing data, including in other popular fMRI processing environments (e.g., FSL, AFNI, etc). In this R03, we plan to extend the pickatlas package to be usable by anyone regardless of the original software used to analyze the data, as long as a T, F, or Z-map is generated, and information on degrees of freedom and smoothness are available. In specific aim 2 we will add functionality often requested by pickatlas users including the addition of two non- human primate atlases and NifTi format compatibility. In specific aim 3 we will generate complete documentation, including an updated user manual, and a developer's manual. All software, source code, manuals, and atlas templates will be distributed freely through the NITRC (Neuroimaging Informatics Tools and Resources Clearinghouse, www.nitrc.gov). We will participate in community activities with the NITRC, including meetings and phone conferences. PUBLIC HEALTH RELEVANCE: Hypothesis driven analysis based on neuroanatomy is critical to brain imaging studies of human brain disorders and disease, as well as animal studies geared towards understanding brain disease. This project will build on the pickatlas software to provide automated neuroanatomical region of interest analyses for human imaging studies and studies of non-human primates. [unreadable] [unreadable]

DESCRIPTION (provided by applicant): The incidence rates of cerebral vascular disease (CBVD) and its associated cognitive impairment are rapidly increasing around the world. African Americans (AA) and patients with diabetes have higher rates of CBVD and associated cognitive dysfunction. However, the impact of race on the development of CBVD assessed by Magnetic Resonance Imaging (MRI) of the brain and cognitive testing remains poorly studied. Controversy exists regarding the severity of cerebral small blood vessel disease in AA, relative to European Americans (EA). Small blood vessel disease appears as white mater hyperintensities (WMH) on brain MRI. Several studies have reported more WMH, and some less WMH, in AA relative to EA. Existing studies failed to account for the effects of potentially modifiable environmental risk factors that impact development of WMH, such as access to healthcare, socioeconomic status, blood pressure and blood sugar control. This application proposes to determine the environmental and genetic factors that underlie cerebral structural and functional changes in 600 AA with type 2 diabetes. We plan to re-evaluate 600 AA in the ongoing AA-Diabetes Heart Study (AA-DHS) with cerebral MRI and a battery of cognitive tests; these individuals have previously been tested for the presence of silent (sub-clinical) heart disease and all have DNA available for genetic testing. We plan to test for relationships between alterations in brain structure (including WMH) and brain function with cognitive performance in these AA subjects. To determine whether racial differences in the relationships between CBVD and cognitive performance are present, results in AA would be contrasted with similar data in EA who have T2D from the related Diabetes Heart Study (DHS-MIND). This proposal will attempt to detect the inherited and environmental factors that produce susceptibility to CBVD in AA. We expect that modifiable environmental risk factors may explain the higher rates of CBVD reported in existing studies of AA. Furthermore, attention to these modifiable risk factors may reduce the rates of CBVD and cognitive dysfunction in the high risk AA population.

DESCRIPTION (provided by applicant). Football head injuries involve significant forces and can result in mild to severe traumatic brain injuries. While this has received increasing attention at the professional, collegiate, and high school levels, there is scarce if any data available for participants in the youth leagues (8-12 years old). This study will relate information about head motion during a hit in youth football to neurocognitive and imaging data to determine the effects of subconcussive impacts, and the true incidence of cognitive and objective imaging changes. All elements of this study focus on the objective to increase understanding of pediatric mild Traumatic Brain Injury (mTBI) and prospectively collect biomechanical, imaging, functional, and computational modeling data on a scale never before attempted. This project integrates neuroinformatics work and the computational modeling techniques developed by Drs. Maldjian and Stitzel at Wake Forest University Health Sciences (WFUHS). It leverages the investments by WFUHS which has identified, instrumented, and begun recruiting the target population and funding the initial year as part of our ongoing work in this area of critical public importance. Th long term benefit of the research will be to allow equipment designers, researchers, and clinicians to better prevent, mitigate, identify and treat injuries to help make youth league football a safer activity for millions of children.

DESCRIPTION (provided by applicant): Title: Sports-related sub concussive impacts in children: MRI & biomechanical correlates Football head injuries involve significant forces and can result in mild to severe traumatic brain injuries. While this has received increasing attention at the professional and collegiate levels, there is little data at the high school level and scarceif any data available for participants in the youth leagues (8-12 years old). This study will relate information about biomechanical forces experienced during a hit in youth football to neurocognitive and imaging data to determine the effects of subconcussive impacts. All elements of this study focus on the objective to increase understanding of pediatric mild Traumatic Brain Injury (mTBI). In this R03 we will investigate the biomechanical and cognitive correlates of diffusion MRI scalar measures of mTBI in children. This project leverages the investments by Wake Forest University Health Sciences which has identified, instrumented and enrolled youth football players, and funded the data collection as part of our ongoing work in this area of critical public importance. We already have acquired a very rich data set of ~30 youth players (8- 12 years old) followed over the course of an entire season (2012 football season). All players were instrumented with the Head Impact Telemetry System (HITS) for collection of real-time impact data during all practices and games. Baseline, post-season and post-concussion multimodal magnetic resonance imaging (MRI), including structural T1-weighted, and diffusion tensor and kurtosis imaging (DTI/DKI) were performed in all subjects. Neurocognitive assessments using the pediatric Immediate Post-concussion Assessment and Cognitive Testing (ImPACT) were obtained at baseline, post-injury, and post-season. Data acquisition is now continuing into the 2013 season with 30 new players already being monitored with HITs, and having all received baseline imaging and neuropsychologic testing including IMPACT testing and additional NIH toolbox measures. We have also recently been awarded internal institutional funding for acquiring a non-contact sport control data set (25 participants), including imaging and cognitive testing at baseline and post-season. This R03 will help fund a graduate student in Biomedical Engineering to assist in the MRI diffusion analyses of the acquired youth football and control data. The utilization of state-of-the-art HITs technology and the acquisition of HITs data at all practices and games provides an extremely thorough characterization of the biomechanical forces experienced by the brains of the young athletes. This unique data set represents the largest of its type in the world, combining imaging, cognitive testing and biomechanical measures in the youth age group. The long term benefit of the research will be to allow equipment designers, researchers, and clinicians to better prevent, mitigate, identify and treat injuries to help make football a safer activity for millions of children.

? DESCRIPTION (provided by applicant): Football head injuries involve significant forces and can result in mild to severe traumatic brain injuries. While this has received increasing attention at the professional and collegiate levels, there is less data available for the millions of participants in high school leagues (14-18 years old) during this time of rapid brain development. The purpose of this study is to relate information about cumulative head impact exposure over a season of high school football with neurocognitive and neuroimaging data to determine the effects of sub-concussive impacts on the brain. All elements of this study focus on the objective to increase understanding of pediatric mild Traumatic Brain Injury (mTBI) and prospectively collect biomechanical, imaging, functional, and computational modeling data on a scale never before attempted in adolescents involved in contact sports. This project integrates neuroinformatics work and the computational modeling techniques developed by Drs. Whitlow, Maldjian and Stitzel at Wake Forest University Health Sciences.

Project Summary Recent advances in the understanding of cancer biology have led to an increased number of cancer therapies. The evaluation of these new therapies in human clinical trials is associated with high cost and potential risks. Imaging approaches can play an important role in this evaluation by identifying patients who respond to treatments. The current radiological assessment of treatment outcomes predominantly relies on changes in tumor size. This is a major limiting factor as the effects of many therapeutic agents at the microscopic level precede changes in tumor size. One such tumor property that has been extensively targeted for new cancer therapies is tumor angiogenesis (or perfusion), which has been shown to support tumor proliferation and infiltration. We have recently developed a quantitative magnetic resonance imaging (MRI) technique, called Arterial Spin Labeling (ASL) that can measure tumor perfusion non-invasively and without the administration of exogenous contrast agent. ASL MRI uses highly permeable water as a tracer, by magnetically labeling the water proton in the arterial blood and measuring their accumulation in the tissue of interest. We have used ASL to monitor therapy response in multiple clinical trials and have shown that ASL measured tumor perfusion decreased as early as 8 days after the initiation of antiangiogenic therapy in patients with renal cell carcinoma (RCC), much earlier than the tumor size changes. However, ASL has not undergone a robust and rigorous validation process to be established as a quantitative imaging method. In this project, we will validate ASL measured perfusion as a quantitative imaging marker to evaluate treatment response in patients with brain tumors (glioblastoma multiforme, GBM) and metastatic RCC, two known highly vascularized tumors. The specific aims of the project are: 1) To demonstrate the reliability and precision of ASL measured perfusion in the brain and kidneys of 30 normal volunteers; 2) To predict clinical outcomes based on baseline (pre- treatment) perfusion and early changes in post-treatment perfusion in 40 patients with newly diagnosed GBM undergoing chemoradiation therapy; and 3) To predict long-term outcomes based on baseline (pre-treatment) perfusion and early changes in post-treatment perfusion in 40 patients with metastatic RCC undergoing antiangiogenic therapies. In the first aim, we will also develop quality-control protocols using a novel 3D printed perfusion phantom, currently available at UT Southwestern (UTSW) Medical Center to measure the reliability and precision of ASL measured flow. In the second and third aims, we will incorporate automated and semi- automated methods for tumor segmentation and analysis, such that these results can be replicated elsewhere. The patients for aims 2 and 3 will be recruited from ongoing clinical trials at UTSW. In both these aims, we will test our hypothesis that greater reduction in tumor perfusion immediately after treatment, compared to baseline correlates with improved progression free survival and overall survival. Such early changes in ASL measured perfusion may predict tumor responsiveness better than anatomical imaging, thereby affecting patient management in a timely manner by changing treatments that may be ineffective and potentially toxic.

Project Summary This is a Bioengineering Research Grant (BRG) proposal in response to PAR-19-158 to further develop and validate a non-invasive panel of the most critical glioma molecular markers (IDH, 1p/19q, MGMT) using standard clinical MRI T2-weighted images and deep learning, and extend the performance to tissue-level accuracies. Currently, the only reliable way of obtaining molecular marker status is through direct tissue sampling of the tumor, requiring either a craniotomy and stereotactic biopsy or a large open surgical resection. Noninvasive determination of molecular markers with tissue-level accuracy would be transformational in the management of gliomas, reducing or eliminating the risks and costs associated with a neurosurgical procedure, accelerating the time to definitive treatment, improving patient experience and ultimately patient outcomes and survival time. Artificial intelligence such as deep learning has emerged as a powerful method for classification of imaging data that can exceed human performance. Preliminary work using our novel voxel-wise classification-segmentation approach with the NIH/NCI TCIA glioma database has outperformed any prior noninvasive methods for determination of IDH, 1p/19q, and MGMT methylation, achieving accuracies of 97%, 93%, and 95%, respectively. The approach however, needs to be validated beyond the TCIA and accuracies need to be extended in order to achieve tissue level performance. This will be accomplished by using our top-performing voxel-wise classification framework, leveraging marker-specific targeted sample sizes, and gaining a final boost from deep-learning artifact correction networks. In Aim 1 we will curate a database of over 2000 gliomas including 500 subjects from our institution, 1200 subjects from our external collaborators, and over 300 subjects from the TCIA. We will train our voxel-wise deep learning classifiers to determine molecular status based on clinical T2-weighted MR images with target accuracies of 97%. In Aim 2 we will rigorously evaluate the motion and noise sensitivity of the networks and create an artifact correction network with the goals of 1) recovering accuracies in the setting of large amounts of motion/noise and 2) further boosting accuracy to tissue-level performance even in the absence of visible artifact. In Aim 3 we will deploy a complete end-to-end clinical workflow and evaluate real-world live performance of the AI tool on 300 prospectively acquired brain tumor cases and 300 subjects from our external collaborators. The AI tool will be made available for deployment at other medical centers. The developed framework can also be extended to additional markers in a straightforward fashion. In summary, this BRG proposal will further develop, refine and validate a non-invasive MRI-based method for determining the most critical glioma molecular markers rivaling tissue-level accuracies to significantly reduce and in many cases eliminate the need for stereotactic biopsy.