Search

This study shows that image-based machine learning techniques can be used to distinguish between disease-specific gray matter atrophy patterns in patients with Alzheimer disease and in patients with behavioral variant frontotemporal dementia by using T1-weighted structural MR imaging for single-subject diagnosis.

To evaluate cerebral metabolism in clinical hydrocephalus with ¹H magnetic resonance spectroscopy (MRS). In 24 children and adults with progressive, arrested, or normal pressure hydrocephalus, long-echo time ¹H MR spectra were acquired from periventricular white matter and intraventricular cerebrospinal fluid (CSF). Metabolite ratios, and the presence of lactate, were compared with 38 age-matched controls. Metabolite ratios of patients were within the 95% confidence interval (CI) of controls. A small lactate resonance was detected in 20% of control and hydrocephalic subjects. Lactate was consistently visible in CSF spectra, though lactate concentrations were normal. The CSF lactate T2 was long in comparison with the known intracellular metabolite T2 relaxation times. In three neonates with hydrocephalus and spina bifida, 3-hydroxybutyrate was detected in CSF in vivo. Within the limits of the present methods, ¹H MRS could not detect cerebral metabolic abnormalities in human hydrocephalus and provided no additional diagnostic information. The long T2 of lactate in CSF explains its high visibility. Hence, the detection of lactate in spectra acquired from voxels that contain CSF does not necessarily imply cerebral ischemia. J. Magn. Reson. Imaging 2003;17:291–299. © 2003 Wiley-Liss, Inc.

The purpose of this study was to assess temporal lobe white matter changes accompanying hippocampal sclerosis on magnetic resonance (MR) imaging using single-voxel ¹H MR spectroscopy and to strengthen the hypothesis that these white matter changes are caused by myelin alterations. In 11 patients with histologically proven hippocampal sclerosis, preoperative coronal fluid-attenuated inversion recovery images were visually assessed by two experienced neuroradiologists for hippocampal signal increase and size decrease, atrophy of collateral white matter, and temporal lobe gray/white matter demarcation loss. Single-voxel ¹H MR spectroscopy of the white matter of each anterior temporal lobe was also performed, excluding the amygdala and hippocampus. The N-acetyl-aspartate (NAA)/choline and NAA/creatine ratios were calculated. In 12 healthy volunteers both temporal lobes were spectroscopically examined. In all patients the excised hippocampi were histologically assessed for the presence of sclerosis, and the excised neocortical temporal lobes were examined for gray and white matter abnormalities. MRI abnormalities were found on the right in six patients, on the left in four, and one scan was normal. Hippocampal signal increase was seen in nine patients, hippocampal size decrease in ten, atrophy of collateral white matter in nine, and gray/white matter demarcation loss in six. A significant decrease in the NAA/choline ratio was found in temporal lobe white matter ipsilateral to the pathologic hippocampus (symptomatic side), compared with the contralateral, asymptomatic side (P < 0.01), and also compared with controls (P < 0.001). The ipsilateral NAA/creatine ratio was also significantly decreased (P < 0.05) compared with the contralateral side and the control subjects (P < 0.001). Histological examination showed hippocampal sclerosis to a different degree in all patients. Neither gliosis nor cortical dysplasia was found in the ipsilateral, symptomatic temporal lobe. Significant decrease in the mean of NAA/choline ratios is found in temporal lobe white matter of patients with histologically confirmed hippocampal sclerosis. As this indicates neuronal loss or dysfunction, the number of axons may be reduced, with associated decrease in myelin density. J. Magn. Reson. Imaging 2000;11:25–31. © 2000 Wiley-Liss, Inc.

We attempted to assess whether magnetic resonance imaging (MRI)-MR angiography (MRA)-MR spectroscopy (MRS) measurements can be used in the differentiation of patients in whom severe carotid lesions result in chronically hypoperfused regions and in whom the collateral capacity is sufficient to maintain a normal cerebral blood flow. Sixty-six patients with severe stenosis of the internal carotid artery (ICA) and 19 control subjects underwent MRI, ¹H MRS, and MRA. Anaerobic metabolic changes in the middle cerebral artery (MCA) territory were studied by assessing N-acetyl-L-aspartate (NAA)/choline and lactate/NAA ratios. Quantitative flow was measured in the ICA, in the basilar artery, and in the MCA. Thirty-four patients had borderzone infarcts, 16 patients had territory infarcts, and 16 patients had no infarcts on MRI. Patients with borderzone infarcts had significantly reduced flow in the ICA (P < 0.001) and in the MCA (P < 0.05) and decreased NAA/choline ratios (P < 0.001) in non-infarcted regions compared with control subjects (P < 0.001) but also compared with patients with territory infarcts (P < 0.05) and patients without infarcts (P < 0.05). Flow measurements in the ICA and MCA and metabolic measurements in the MCA territory can be applied to select patients in whom cerebral perfusion pressure is insufficient to maintain normal cellular integrity. J. Magn. Reson. Imaging 1999;9:1–9 © 1999 Wiley-Liss, Inc.

The purpose of this study was to test the hypothesis that a high lactate signal and a low N-acetyl-aspartate/choline ratio in neonates with postasphyxial encephalopathy indicated a high chance of an adverse outcome in vivowhen proton magnetic resonance spectroscopy was used. Twenty-one full-term asphyxiated neonates were examined at a mean postnatal age of 7.1 d. Five patients died, and five survivors had handicaps. Eleven of the 16 survivors (seven without handicaps and four with handicaps) had a second examination at 3 mo of age. After magnetic resonance imaging, spectra were obtained at 1.5 tesla. A 20-mm-thick slice was selected through the basal ganglia. After optimizing the B-0 field, we used a double spin-echo pulse sequence (90–180-180°) with a time to repeat of 2000 ms and a time to echo of 272 ms. Two-dimensional spectroscopic imaging was performed by 32 ± 32 phase encoding steps in two directions in a 225-mm field of view, resulting in 1-mL volumes, followed by computerized processing. Neuromotor development was examined at 6 wk, 3 mo, and every 3 mo thereafter. Lactate resonances were seen only in the five patients with grade 3 postasphyxial encephalopathy. Lactate was distributed diffusely (n= 4), or localized in areas of infarction (n= 1). N-acetyl-aspartate/choline ratios were significantly lower in the patients with an adverse outcome than in the survivors without handicaps, both neonatally (p< 0.005, Wilcoxon's rank sum test) and at 3 mo (p< 0.05). In conclusion, the presence of cerebral lactate and a low N-acetyl-aspartate/choline ratio demonstrated in vivousing proton magnetic resonance spectroscopy in full-term neonates with postasphyxial encephalopathy indicate a poor outcome.