Automated monitoring of MRI abnormalities in HIV/AIDS

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): Neuropsychological problems are common in Human Immunodeficiency Virus infection (HIV) and Acquired Immune Deficiency Syndrome (AIDS), even in the absence of clearly visible cerebral lesions. In many cases, the origin of these problems in cerebral function is confounded by other issues, such as medication interactions, substance abuse, or emotional reaction to this devastating illness. Research using hand-delineation of cerebral structures suggests that specific neuropsychological problems are associated with decreased volume of the caudate nucleus. However, the expertise and time required to perform these manual measurements severely limit their application in clinical trials, and prevent them from being used in clinical practice. The overall goal of this project is to develop MRI-analysis software that can automatically and accurately quantify cerebral structures in patients with HIV/AIDS The outcome of this project will be commercialized as an HIV image-processing service for clinical trials and as medical device software providing HIV specific quantification. In a pharmacological research setting, the proposed software will allow non-invasive monitoring of cerebral status, thus providing objective criteria for developing treatment approaches to arrest or prevent AIDS dementia. In the clinical setting, the cerebral quantification provided by this software may eventually be applied to the selection of patients for treatment, and in the monitoring of treatments. Our first specific aim is to optimize our cerebral segmentation and quantification software for subjects with HIV/AIDS. Our second aim is to optimize our cortical labeling and quantification software for subjects with HIV/AIDS. Our third aim is to develop software quantifying tissue property changes in normal appearing white matter of subjects with HIV infection using Diffusion Tensor Imaging. Our fourth aim is to validate the automatic cerebral quantification software against "gold standard" manual segmentation methods in subjects infected with HIV. Fulfillment of these specific aims will demonstrate the feasibility and accuracy of automatic cerebral quantification software for HIV disease monitoring.