Toward Precision and Reproducibility of Diffusion Tensor Imaging: A Multicenter Diffusion Phantom and Traveling Volunteer Study

BACKGROUND AND PURPOSE

Precision medicine is an approach to disease diagnosis, treatment, and prevention that relies on quantitative biomarkers that minimize the variability of individual patient measurements. The aim of this study was to assess the intersite variability after harmonization of a high-angular-resolution 3T diffusion tensor imaging protocol across 13 scanners at the 11 academic medical centers participating in the Transforming Research and Clinical Knowledge in Traumatic Brain Injury multisite study.

MATERIALS AND METHODS

Diffusion MR imaging was acquired from a novel isotropic diffusion phantom developed at the National Institute of Standards and Technology and from the brain of a traveling volunteer on thirteen 3T MR imaging scanners representing 3 major vendors (GE Healthcare, Philips Healthcare, and Siemens). Means of the DTI parameters and their coefficients of variation across scanners were calculated for each DTI metric and white matter tract.

RESULTS

For the National Institute of Standards and Technology diffusion phantom, the coefficients of variation of the apparent diffusion coefficient across the 13 scanners was <3.8% for a range of diffusivities from 0.4 to 1.1 × 10^-6 mm²/s. For the volunteer, the coefficients of variations across scanners of the 4 primary DTI metrics, each averaged over the entire white matter skeleton, were all <5%. In individual white matter tracts, large central pathways showed good reproducibility with the coefficients of variation consistently below 5%. However, smaller tracts showed more variability, with the coefficients of variation of some DTI metrics reaching 10%.

CONCLUSIONS

The results suggest the feasibility of standardizing DTI across 3T scanners from different MR imaging vendors in a large-scale neuroimaging research study.

Accurate template-based correction of brain MRI intensity distortion with application to dementia and aging

This paper examines an alternative approach to separating magnetic resonance imaging (MRI) intensity inhomogeneity from underlying tissue-intensity structure using a direct template-based paradigm. This permits the explicit spatial modeling of subtle intensity variations present in normal anatomy which may confound common retrospective correction techniques using criteria derived from a global intensity model. A fine-scale entropy driven spatial normalisation procedure is employed to map intensity distorted MR images to a tissue reference template. This allows a direct estimation of the relative bias field between template and subject MR images, from the ratio of their low-pass filtered intensity values. A tissue template for an aging individual is constructed and used to correct distortion in a set of data acquired as part of a study on dementia. A careful validation based on manual segmentation and correction of nine datasets with a range of anatomies and distortion levels is carried out. This reveals a consistent improvement in the removal of global intensity variation in terms of the agreement with a global manual bias estimate, and in the reduction in the coefficient of intensity variation in manually delineated regions of white matter.

Comparison of methods for measuring longitudinal brain change in cognitive impairment and dementia

PURPOSE

The goal of this project was to compare MRI measures of hippocampal, entorhinal cortex (ERC), and whole brain longitudinal change in cognitively normal elderly controls (C), non-demented subjects with cognitive impairment (CI), and demented (D) subjects.

METHODS

16 C, 6 CI, and 7 D subjects of comparable age were studied with MRI twice, at least 1 year apart. Longitudinal change in total brain size was measured by several methods, including computerized segmentation, non-linear warping, and change in the fluid/tissue boundaries between cerebrospinal fluid (CSF) and brain. Change in hippocampal volume was measured by semi-automated methods, and ERC volumes were manually measured.

RESULTS

The annual rate of atrophy was greater in D versus C and D versus CI for cortical gray matter (cGM) (P=0.009 and 0.002), hippocampus (P=0.0001 and 0.002), and for the change in the fluid/tissue boundary (P=0.03 and 0.03). The annual rate of atrophy of ERC was greater in both CI and D versus C (P=0.01 and 0.0002). No significant differences between groups were found using non-linear warping.

CONCLUSIONS

In CI, the greatest annual rates of atrophy were in ERC, while in D the greatest annual rates of atrophy were in hippocampus and cortex. Progressive ERC atrophy was observed with a greater degree of cognitive impairment, while hippocampal and cortical atrophy were only observed in demented subjects.

Reliability of tissue volumes and their spatial distribution for segmented magnetic resonance images

Before using MRI tissue segmentation in clinical studies as a dependent variable or as a means to correct functional data for differential tissue contribution, we must first establish the volume reliability and spatial distribution reproducibility of the segmentation method. Although several reports of volume reliability can be found in the literature, there are no articles assessing the reproducibility of the spatial distribution of tissue. In this report, we examine the validity, volume reliability, and spatial distribution reproducibility for our K-means cluster segmentation. Validation was examined by classifying gray matter, white matter, and CSF on images constructed using an MRI simulator and digital brain phantom, with percentage volume differences of less than 5% and spatial distribution overlaps greater than 0.94 (1.0 is perfect). We also segmented repeat scan MRIs from 10 healthy subjects, with intraclass correlation coefficients greater than 0.92 for cortical gray matter, white matter, sulcal CSF, and ventricular CSF. The original scans were also coregistered to the repeat scan of the same subject, and the spatial overlap for each tissue was then computed. Our overlaps ranged from 0.75 to 0.86 for these tissues. Our results support the use of K-means cluster segmentation, and the use of segmented structural MRIs to guide the analysis of functional and other images.

Region and tissue differences of metabolites in normally aged brain using multislice 1H magnetic resonance spectroscopic imaging

Quantitative measurements of regional and tissue specific concentrations of brain metabolites were measured in elderly subjects using multislice proton magnetic resonance spectroscopic imaging ((1)H MRSI). Selective k-space extrapolation and an inversion-recovery sequence were used to minimize lipid contamination and linear regression was used to account for partial volume problems. The technique was applied to measure the concentrations of N-acetyl aspartate (NAA), and creatine (Cr)- and choline (Cho)-containing compounds in cortical gray and white matter, and white matter lesions of the frontal and the parietal lobe in 40 normal elderly subjects (22 females and 18 males, 56-89 years old, mean age 74 +/- 8). NAA was about 15% lower in cortical gray matter and 23% lower in white matter lesions when compared to normal white matter. Cr was 11% higher in cortical gray matter than in white matter, and also about 15% higher in the parietal cortex than in the frontal cortex. Cho was 28% lower in cortical gray matter than in white matter. Furthermore, NAA and Cr changes correlated with age. In conclusion, regional and tissue differences of brain metabolites must be considered in addition to age-related changes when interpreting (1)H MRSI data.

Method to correlate 1H MRSI and 18FDG-PET

The in vivo neuronal contribution to human cerebral metabolic rate of glucose (CMRglc), measured by 18FDG-PET, is unknown. Examining the effect of 1H MRSI-derived N-acetyl aspartate (NAA) concentration on positron emission tomography (PET) measures of metabolic activity might indicate the relationship of CMRglc to neuron density. In a population of 19 demented, cognitively impaired, and control subjects, the Miller-Gartner algorithm was applied to whole-brain PET data to isolate the PET signal originating in cortical gray matter alone (GMPET). An analogous procedure applied to multislice proton MRSI data yielded the N-acetyl aspartate concentration in cortical gray matter (GMNAA). In 18 of 19 subjects, a significant linear regression (P < 0.05) resulted when GMPET was plotted against GMNAA, whereby GMPET was higher for higher GMNAA. This suggests that CMRglc rises linearly with increasing neuron density in gray matter. This method may be used to investigate the relationship of CMRglc to neurons in various conditions.

Cortical metabolite alterations in abstinent cocaine and cocaine/alcohol-dependent subjects: proton magnetic resonance spectroscopic imaging

Chronic abuse of cocaine or alcohol is associated with structural, neuropathological and cognitive impairments that have been documented extensively. Little is known, however, about neurobiochemical changes in chronic substance abusers.We performed MRI and multi-slice brain proton magnetic resonance spectroscopic imaging (MRSI) to assess neuronal viability (via N-acetylaspartate (NAA)) and white matter metabolite status in 22 4-months-abstinent individuals dependent on crack cocaine only and on both crack cocaine and alcohol. Compared to 11 non-dependent controls we found (1) significantly lower NAA measures in the dorsolateral prefrontal cortex of the combined cocaine-dependent groups; (2) comparable spatial distribution and magnitude of these NAA effects for both cocaine-dependent groups; (3) higher choline-containing metabolites in frontal white matter of individuals dependent on both cocaine and alcohol; (4) absence of brain atrophy in both abstinent cocaine-dependent samples; and (5) partial recovery from prefrontal cortical NAA loss, primarily with abstinence from alcohol. The MRSI findings suggest preferential neuronal damage to the frontal cortex of both cocaine-dependent samples and gliosis in frontal white matter of individuals dependent on both alcohol and cocaine, conditions that persist for more than 4 months of abstinence.

Magnetization transfer ratio of white matter hyperintensities in subcortical ischemic vascular dementia

BACKGROUND AND PURPOSE

In subjects with subcortical ischemic vascular dementia (SIVD), tissue vacuolization, myelin pallor, and demyelination have been found on pathologic examination of white matter signal hyperintensities (WMSH). Magnetization transfer ratio (MTR) values provide a potential measure of compromised white matter integrity. The purpose of this study was to determine if there were differences in MTR of WMSH between subjects with SIVD and cognitively normal healthy control subjects.

METHODS

Fifteen subjects with SIVD and 16 control subjects of comparable age and sex were studied. MTR images were coregistered to MR images segmented into tissue classes (gray matter, white matter, CSF, WMSH, and lacunar infarcts). MTR of WMSH was compared across groups and examined by WMSH location, size, and total burden.

RESULTS

WMSH burden was greater in SIVD patients than in control subjects (2.4% vs 0.67%). MTR of WMSH did not differ between groups, but MTR of periventricular WMSH was lower in SIVD patients than in control subjects (37.6% vs 39.4%). Even after accounting for covariant effects of lesion burden, there was still a trend toward reduced periventricular WMSH MTR in the group with dementia. There was no correlation between WMSH MTR and WMSH lesion size.

CONCLUSION

These findings are consistent with observations that pathologic changes in vascular dementia are most severe in the periventricular white matter and suggest that insight into the pathophysiology of SIVD might be gleaned from studies of the periventricular region.

Changes of hippocampal N-acetyl aspartate and volume in Alzheimer's disease. A proton MR spectroscopic imaging and MRI study

Hippocampal atrophy detected by MRI is a prominent feature of early Alzheimer's disease (AD), but it is likely that MRI underestimates the degree of hippocampal neuron loss, because reactive gliosis attenuates atrophy. We tested the hypothesis that hippocampal N-acetyl aspartate (NAA: a neuronal marker) and volume used together provide greater discrimination between AD and normal elderly than does either measure alone. We used proton MR spectroscopic imaging (1H MRSI) and tissue segmented and volumetric MR images to measure atrophy-corrected hippocampal NAA and volumes in 12 AD patients (mild to moderate severity) and 17 control subjects of comparable age. In AD, atrophy-corrected NAA from the hippocampal region was reduced by 15.5% on the right and 16.2% on the left (both p < 0.003), and hippocampal volumes were smaller by 20.1% (p < 0.003) on the right and 21.8% (p < 0.001) on the left when compared with control subjects. The NAA reductions and volume losses made independent contributions to the discrimination of AD patients from control subjects. When used separately, neither hippocampal NAA nor volume achieved to classify correctly AD patients better than 80%. When used together, however, the two measures correctly classified 90% of AD patients and 94% of control subjects. In conclusion, hippocampal NAA measured by 1H MRSI combined with quantitative measurements of hippocampal atrophy by MRI may improve diagnosis of AD.

Volumetric method for evaluating magnetization transfer ratio of tissue categories: application to areas of white matter signal hyperintensity in the elderly

An objective technique for analyzing magnetization transfer ratio of segmented tissues was used to evaluate differences between normal-appearing white matter and areas of white matter signal hyperintensity on T2-weighted magnetic resonance images in 23 healthy elderly subjects (mean age, 75 years). Segmented brain images (cerebrospinal fluid, gray matter, white matter, areas of white matter signal hyperintensity) computed from T1- and T2-weighted images were combined with magnetization-transfer-ratio images to produce magnetization-transfer-ratio histograms for each tissue. There was a significant 8.1% reduction in mean magnetization transfer ratio for areas of white matter signal hyperintensity compared with normal-appearing white matter (P < .001). The magnetization transfer ratio for all tissue categories was significantly negatively correlated with age (all, P < .05). This method of measuring magnetization transfer ratio was objective, independent of regional variations, and as reproducible as the segmentation procedure.

Semantic priming impairment in HIV

HIV+ subjects have shown impairment on tests of executive function including automatic attention and verbal tasks. Impairment of semantic priming in HIV patients would suggest a disruption of automatic semantic activation. We examined semantic priming in HIV+ individuals and HIV- control participants with no history of substance abuse, neurologic or psychiatric disorder unrelated to HIV. HIV+ participants were divided into cognitively normal and cognitively impaired subgroups on the basis of a neuropsychological battery of 15 tests. Participants were presented with English words and nonword letter strings and indicated if the stimulus was a word or nonword. The nonwords were orthographically and phonologically correct and were created by rearranging the letter sequence of words ("ulpit"). All words had an obvious antonym ("deep"); two-thirds were presented as sequential antonym pairs ("enter"-"exit"). There were no group differences in speed of response to nonwords, indicating no generalized reaction time deficit. While control and cognitively normal HIV+ participants showed an effect of priming on reaction time to correctly detected words, cognitively impaired HIV+ participants did not. The lack of semantic priming demonstrated by cognitively impaired HIV+ participants suggests that they have lessened activation of automatic semantic networks.

Tissue segmentation of the brain in Alzheimer disease

PURPOSE

To compare brain tissue in patients with Alzheimer disease with that in elderly control subjects by using high-resolution MR imaging and quantitative tissue-segmentation techniques.

METHODS

MR imaging of the brain was performed in 21 patients with Alzheimer disease and 17 control subjects. A computerized segmentation program was used to quantify volumes of ventricular and sulcal cerebrospinal fluid (CSF), white matter, cortical gray matter, and white matter signal hyperintensity. Statistical analysis was performed using analysis of variance.

RESULTS

We found a significant decrease in total brain tissue and cortical gray matter and an increase in the ventricular and sulcal CSF in Alzheimer patients compared with control subjects. There was no difference in the volume of white matter. More white matter signal hyperintensities were found in Alzheimer patients, and a significant interaction between age and group was noted. Neuropsychological test scores correlated significantly with sulcal CSF in patients with Alzheimer disease.

CONCLUSION

Semiautomated segmentation of MR images of the brains of patients with Alzheimer disease reveals significant brain atrophy attributable to loss of cortical gray matter, which is compatible with the pathologic features of Alzheimer disease. There is also a significant increase in white matter signal hyperintensities. Tissue segmentation may increase our understanding of dementia but, as yet, when used alone, it does not play a role in the premorbid diagnosis of Alzheimer disease.

Alzheimer disease and subcortical ischemic vascular dementia: evaluation by combining MR imaging segmentation and H-1 MR spectroscopic imaging

PURPOSE

To determine the association between H-1 magnetic resonance (MR) spectroscopic imaging and MR imaging differences in subjects with Alzheimer disease (AD) or subcortical ischemic vascular dementia (SIVD) versus control subjects and if both studies combined enable discrimination of AD from control subjects better than either study alone.

MATERIALS AND METHODS

Measures were obtained in nine AD, eight SIVD, and 11 control subjects with MR imaging segmentation software.

RESULTS

Statistically significantly lower N-acetylaspartate/choline-containing metabolites (Cho) and higher Cho/creatine-containing metabolites in posterior mesial gray matter in AD versus control subjects were independent of MR imagining differences. Combined measures allowed correct classification of AD and control subjects, but none of the MR measures allowed accurate discrimination between AD and SIVD subjects.

CONCLUSION

Between-group differences in tissue-type contributions to H-1 MR spectroscopic imaging voxels must be accounted for when reporting H-1 MR spectroscopic imaging data in AD, SIVD, and control subjects. Combined studies allowed more accurate discrimination between AD and control subjects than either study alone.

Brain atrophy and cognitive function in older abstinent alcoholic men

We used computer-aided magnetic resonance image analysis and an age-normed battery of neuropsychological tests to measure brain atrophy and cognitive function in 14 older abstinent alcoholic men and 11 older controls in the expectation that these subject groups would show the greatest and most persistent cerebral effects consequent to chronic alcoholism. The abstinent alcoholics exhibited cognitive impairments (primarily in memory and visual-spatial-motor skills) compared with the controls. In contrast, we found no difference in global cerebral atrophy between the groups, although two alcoholics had extensive atrophy compared with all other subjects. However, there was a stronger association between age and ventricular dilation in the alcoholic sample compared with controls. We conclude that a substrate other than magnetic resonance imaging-detectable global atrophy must underlie the persistent cognitive impairments evident in the sampled alcoholics. Furthermore, if there are global atrophic changes in the brain associated with chronic alcoholism, these effects are not ubiquitous and/or may be reversible in most patients with sufficient abstinence.