Neurological involvement is frequent in HIV-infected individuals. We have developed methods to quantitatively analyze brain magnetic resonance images (MRI). Parameters that we are currently able to measure include gray and white matter volumes, the ratio of cerebrospinal fluid (CSF) space to total intracranial volume (ICV) as an indicator of cerebral atrophy, and the location and number of pathological, high intensity lesions in the cerebral white matter. We have applied these techniques to patients with AIDS and have found an increased CSF/ICV ratio in AIDS patients. Furthermore, in a patient with AIDS dementia complex (ADC) who recovered after treatment with zidovudine (AZT), we found a dramatic increase in gray matter volume and a resolution of white matter hyperintensities. The current proposal is to extend these studies in order to determine if quantitative measurement of brain MR images will provide an index of neurological response to treatment. We propose to study a larger group of 15 HIV- positive patients who are about to enter open label AZT trials and who demonstrate cognitive impairment at baseline. We will correlate specific MR image parameters with neuropsychological performance before and after 4 months of treatment. We anticipate, based on our preliminary studies and literature values, that a number of these subjects will have abnormal baseline MRI;s with regard to the indices measured here. Based on published studies, many may improve neuropsychologically with treatment with AZT. We propose to follow the response to treatment with our MR image analysis methods and will in addition extend these methods to allow us to determine whether there are regional brain effects, and to quantify the """"""""texture"""""""" of white matter as an index of the extent of pathological tissue. These methods may provide a convenient, accessible, and quantifiable measurement of the effects of treatment for AIDS on the appearance of the brain, an important target for HIV infection.
| Hillman, G R; Chang, C W; Ying, H et al. (1999) A fuzzy logic approach to identifying brain structures in MRI using expert anatomic knowledge. Comput Biomed Res 32:503-16 |
