| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Neuroradiology |
1 From the Departments of Radiology (D.M.M., C.R.T., J.A.A., W.R.B.), Pathology (V.R.C.), and Public Health Sciences (C.D.L.), Wake Forest University School of Medicine, Medical Center Blvd, Winston-Salem, NC 27157. From the 2000 RSNA scientific assembly. Received August 5, 2002; revision requested September 24; final revision received February 6, 2004; accepted March 16. Study supported by National Institutes of Health grants NS20618 and NS36780 to D.M.M. Address correspondence to D.M.M. (e-mail: dmmoody@wfubmc.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Internal review board approval or informed consent from next of kin was not required. Brains of 21 subjects (mean age, 72.5 years; 12 men, nine women) were evaluated at autopsy with MR imaging. The presence of LA was indicated by confluent or patchy areas of hyperintensity in deep WM. Microvascular density (percentage of vessel area divided by total area) in subjects with LA was measured with computerized morphometric analysis in LA lesions, healthy-appearing WM at MR imaging, and the cortex. These measurements were compared with each other and with measurements from corresponding areas in healthy subjects. Afferent vasculature was stained with alkaline phosphatase in celloidin sections. Hypotheses were tested with computation of a series of repeated-measures linear mixed models.
RESULTS: Autopsy brains from 12 subjects with LA (mean age, 72 years; six men, six women) and nine subjects without LA (mean age, 73 years; six men, three women) were studied. Afferent microvascular density ± standard deviation in LA lesions in deep WM (2.56% ± 1.56) was significantly lower than that in corresponding deep WM of healthy subjects (3.20% ± 1.82) (P = .018). Subjects with LA demonstrated decreased afferent vascular density at early ages in all three areas of the brain when compared with healthy subjects of the same age.
CONCLUSION: Findings of decreased afferent vascular density in the area of LA and outside the lesion indicate that LA is a generalized cerebrovascular disease process rather than one confined to deep WM.
© RSNA, 2004
Index terms: Arteries, abnormalities, 17.70 • Brain, diseases, 10.70 • Brain, ischemia, 17.70 • Brain, white matter
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
In a small study in which an exogenous contrast material–based MR method was used in subjects with LA, Markus et al (14) found reduced blood flow in cerebral WM but not in cerebral gray matter. In further studies, this group emphasized that blood flow is reduced in WM that appears healthy at MR in subjects with LA lesions elsewhere (22).
The association of LA with ischemia and the discovery that LA is associated with characteristic abnormal alterations in the structure of subependymal veins has prompted us to continue to examine the vascular bed for additional alterations associated with LA. We have used 100-µm-thick celloidin sections of brain stained with alkaline phosphatase enzyme histochemistry to analyze the brain vasculature. Alkaline phosphatase is concentrated on endothelial cells lining afferent brain vessels and has been used to distinguish afferent from efferent vessels in studies of development, normal aging, and pathologic changes (8,21,23–29). The combination of thick celloidin sections, alkaline phosphatase histochemistry, digital imaging, and computerized morphometry has been used to measure changes in vascular density (23). Use of these protocols yields a robust method that facilitates quantification of density of the afferent vascular bed.
Our a priori hypotheses were as follows: (a) There is decreased vascular density in the area of LA lesions in the deep WM of subjects with LA compared with the deep WM of subjects who did not have LA, (b) there is decreased vascular density in the healthy-appearing WM outside the LA lesions compared with similar WM areas in subjects without LA, (c) there is decreased vascular density in the cortex of subjects with LA compared with the cortex of subjects without LA, and (d) brain vascular density changes due to increasing age differ between subjects with and those without LA. Thus, the purpose of our study was to investigate vessel density changes due to increasing age in three areas of the brain and to correlate these changes with LA on the basis of MR signals and location in the deep WM.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
At autopsy, brains were cooled in a plastic bag by immersing them in ice slush and then refrigerated to improve slicing consistency before gross cutting. An oblique-coronal slice approximately 1.5 cm thick and angled to parallel the penetrating vessels was obtained through the frontal lobes. This slice provided tissue samples from frontal gyri, anterior centrum semiovale, basal ganglia, and anterior insula. The thick sections were stored overnight in a 24-mmol barbiturate-buffered aqueous solution containing 90 mmol CaCl2 and 1% formalin. In preparation for MR imaging, the brain slice was placed between previously cooled lucite sheets within a special lucite holder (30). MR imaging of the brain slice was performed with a 1.5-T MR imager (Signa; GE Medical Systems, Milwaukee, Wis). Intermediate-weighted and T2-weighted MR images were acquired parallel to the brain section in the following manner: A T1-weighted scout view was obtained initially to align the plane of imaging in accordance with the tissue slice. A two-dimensional Fourier transmission spin-echo sequence was then performed (repetition time msec/echo time msec, 2500/20, 80; number of signals acquired, one; matrix, 256 x 192; section thickness, 3 mm; no section gap; field of view, 17 cm). Acquisition of the MR image was performed with the direction of one of the authors (D.M.M.).
After MR imaging, brain slices were fixed for 48 hours in several changes of 70% ethanol at 4°C. The fixed tissue was then dehydrated with a graded ethanol series and embedded in celloidin. The celloidin-embedded tissue was sectioned on a base sledge microtome at 100 µm and stained with alkaline phosphatase (18,25,31,32). According to the study protocol, six of 18 sections from each block were also counterstained separately with Congo red (for amyloid), Masson trichrome (for collagen), Kultchitsky hematoxylin and Luxol fast blue (for myelin), and cresyl violet acetate plus light green and Gill hematoxylin (for structure). These counterstained sections were analyzed for the presence of brain tissue abnormalities by one of the authors (V.R.C.). The histochemistry and histologic aspects of this study were supervised by one of the authors (W.R.B.).
Image Evaluation
Specimen MR imaging was scored with clinical reading, the LA area—which was characterized by hyperintensity—was mapped, and the areas on the alkaline phosphatase–stained slides corresponding to the mapped LA lesions were outlined with glass-fast ink. To be considered an LA lesion, the area had to be (a) hyperintense on T2-weighted MR images, (b) equally or more intense than the cortex on the intermediate-weighted image, (c) larger than 5 mm in diameter, and (d) located in the cerebral WM. No attempt was made to confine the LA to a standard area; its mapping was determined by the location and size of the abnormal MR signal. In the subjects with LA, control areas of WM that appeared to be healthy at MR imaging were arbitrarily outlined with glass-fast ink. The subcortical U fibers, areas with a dual blood supply (18), were not selected for control. In the subjects with no abnormal MR signal, standard areas (range, 16–99 mm2) were outlined on the histologic slide in the periventricular deep WM for quantification of vascular density. Areas of healthy WM apart from the periventricular deep WM were also outlined (range, 12–81 mm2) for computerized morphologic analysis in all subjects. Analysis of MR images and mark-up of particular areas on histologic slides was performed by one of the authors (D.M.M.).
After slide marking, the selected areas were visualized with x10 magnification with a microscope (Eclipse E600; Nikon, Melville, NY), and images were captured digitally with a 2.2-megapixel digital camera (SPOT RT; Diagnostic Instru-ments, Sterling Heights, Mich). Color images were 1600 x 1200 pixels (1.92 x 106 pixels total area), and each image encompassed an area of 1.2 x 0.9 mm on the tissue section. During image acquisition, the stage was positioned systematically within the selected areas without regard to field of view with one exception. In cases where large arteries (>200 µm) were in the field of view, the stage was repositioned only enough to displace the vessel from the field. For this reason, the data in this study exclusively relate to afferent vessels smaller than 200 µm in diameter (eg, small arteries, arterioles, and capillaries).
Acquisition of the digital micrographic images and image processing were performed by one of the authors (C.R.T.) according to the following scheme: Multiple digital images were captured systematically throughout the selected areas. Because the numbers and sizes of LA lesions varied, we analyzed between five and seven areas per slide and between two and fifteen images per selected area. The cap of tissue immediately adjacent to the lateral angle of the lateral ventricle, which consists of a loose matrix of glial processes, is rich in subependymal veins and devoid of axons; thus, it was not included in the analysis.
Vascular Density Quantification
Afferent microvascular density, as characterized in this study, is defined as the area fraction occupied by alkaline phosphatase–stained vessels that are smaller than 200 µm in diameter. The area fraction is calculated as follows: the vessel area (measured in pixels) is divided by the total image area (measured in pixels) and multiplied by 100; the product is expressed as a percentage. Adobe Photoshop (Adobe Systems, San Jose, Calif) and Image Pro Tool Kit (Reindeer Games, Asheville, NC) were used to measure vascular density on digital images.
Image processing was similar to the procedure used for analysis of vessel density in neonates (23) and consisted of the following steps: Pixel intensity data were filtered from color images, hue and saturation information were discarded, and the image was converted to a gray-scale image. Background staining was then removed by a process of background subtraction, in which the filtered gray-scale image was duplicated, and the alkaline phosphatase–positive vessels were removed from the duplicate image with repetitive iterations of a rank-opening algorithm (33).
The number of iterations was kept constant for each image. The resulting background image was subtracted from the original grayscale image, which left only alkaline phosphatase–positive vessels on the image. Conversion to a black-and-white image in binary mode permitted measurement of the percentage of vessel density. Specifically, the analysis program was used to count the number of black pixels (eg, the area occupied by vessels) in the binary image. The number of black pixels was then divided by the total area of the image field (measured in pixels) and multiplied by 100, which yielded the percentage vascular density for that image field. Factors that influence these measurements are depth of field and binary algorithm settings. Depth of field was controlled by collecting all digital images with the same objective lens and Kohler illumination settings.
The optics, combined with the mechanics of the morphometric method, resulted in measurement of the vascular density in a 3-µm-thick plane within the tissue and was uniform for all counts obtained. The Shannon algorithm (34) was used to convert grayscale background-subtracted images to binary mode prior to quantification. The area fraction occupied by afferent vessels in the image was measured with standard Image Pro Tool Kit (Reindeer Games) algorithms. Quantification of vascular density from the histologic sections was performed by one of the authors (C.R.T.).
Statistical Analyses
As described previously, multiple observations were available to maximize the quality of the estimate of vascular density for the respective beds. To test the above hypotheses while accounting for multiple measurements from each individual, we computed a series of repeated-measures linear mixed models (35). Specifically, we assumed that the individual was a random effect under a normal model, with an exchangeable correlation structure for repeated measurements performed in the same individual. Simple contrasts were computed to test the particular hypothesis of interest (eg, WM that appeared healthy at MR imaging in an individual with LA vs WM in an individual without LA and deep WM lesion from an individual with LA vs deep WM from an individual without LA). Analyses were computed with and without adjustment for the age of the individual. Finally, we computed the age-diagnosis interaction to test whether the effect of the location differed as a function of age. To best approximate the conditional normality and homogeneity of variance assumptions of our models, vascular density was transformed with the square root. Residuals from our models were examined for outliers, and influence measures were calculated to ensure that no individual was dominating the model. Statistical analysis software (SAS; SAS Institute, Cary, NC) was used for statistical analyses and comparisons. A P value of .05 or less was considered to indicate statistical significance. Statistical analysis resulted from a consensus between the statistician (C.D.L.) and others (C.R.T., W.R.B., J.A.A.).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
|
Both the LA lesion and the deep WM in healthy subjects contain afferent vessels ranging in size from arteries and arterioles to capillaries (Fig 2, A, B). Actual binary images (Fig 2, C, D) were derived from the digital images (Fig 2, A, B) and demonstrate the effectiveness of the image processing method. Data derived from binary images verify that vessel density is significantly decreased in LA lesions in comparison with that in healthy deep WM (Fig 3, A). The vessel density of alkaline phosphatase–positive vessels in the LA lesion is significantly less than that measured in healthy deep WM (P = .018). The mean afferent vascular density, standard deviation, and other statistical information are shown in the Table.
|
|
|
Similar measurements were also made in more superficial areas of the centrum semiovale–WM areas that were judged to be healthy according to MR imaging criteria (Fig 1, C, D). While the mean vessel density in WM outside the LA lesion in subjects with LA is less than that in healthy WM of subjects without LA (Fig 3, C; Table), the difference is not significant (P = .30). However, when these same data were plotted in relation to age (Fig 3, D), there was a significant difference (P < .006) in the slopes of the lines describing the age-associated vascular density in subjects with LA versus healthy subjects. Vessel density in this area of WM decreases significantly with age in healthy subjects (P < .045). It declines approximately 28.8% per decade between years 57 and 90, which is similar to the age-related decline observed in healthy deep WM. In contrast, the vessel density in the healthy-appearing WM of subjects with LA elsewhere decreased by 47% in those subjects who died before reaching 60 years of age, and it does not decrease significantly thereafter (P = .299), resembling the pattern seen in deep WM. Thus, in subjects with LA, both the area of the lesion and the healthy-appearing WM differ significantly from corresponding areas in healthy subjects (Fig 3, B–D). Since these data reveal apparent vascular density changes in areas away from the lesion, vascular densities were compared in the cortex of healthy subjects versus those with LA (Fig 3, E, F).
All of the cortical areas examined in both healthy subjects and subjects with LA appeared healthy at both MR and pathologic microscopic examination. While there is no significant difference (P = .82) in the average vessel density of the cortex in a healthy subject compared with that of a subject with LA (Fig 3, E; Table), there is a difference in the slopes of the lines that describe the vessel density and age interaction (P = .051) (Fig 3, F). In subjects who are 60 years of age or younger, afferent vessel density in the LA cortex in subjects with LA is diminished by 38% in comparison with that in the cortex of healthy subjects. Age-associated vessel density in the cortex of healthy subjects decreases 21.2% per decade between years 57 and 90, while it remains constant in subjects with LA. Thus, a healthy-appearing cortex in subjects with LA exhibits the same pattern as is seen in the LA lesion and healthy-appearing WM (Fig 3, B, D, F).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
A single hypotensive or metabolic catastrophe can mimic the appearance of LA on MR images (eg, patchy or confluent hyperintensity seen on a T2-weighted MR image in the deep WM) (37,38). However, it is believed that the WM disease in patients with the MR imaging appearance of LA and without history of such a catastrophe results from chronic ischemia (9–11,14,39–41). This imaging feature is predictive of increased morbidity (42). Results of studies with positron emission tomography (40,41) and MR imaging (14,22) have shown decreased blood flow in the WM of subjects with LA. Caution must be used in comparing blood flow with the morphologically determined vascular density because the entire vascular bed in a focal area may not be filled at the same time, and nothing is known about the velocity of blood flow through individual microvessels.
It is known that the afferent vascular supply to the deep WM is derived from the convexity border zone, which is an area with a precarious blood supply, and these WM arterioles are long and travel for considerable distances in the brain parenchyma before they reach their ultimate destination (18). These arterioles are often tortuous, theoretically leading to further losses in hemodynamic kinetic energy (15,43). Thus, the vascular supply of the cerebral deep WM is vastly different than that of the corpus callosum WM, which is characterized by short penetrating arterioles (44).
In the present study, we quantified the local afferent cerebral microvasculature density with nonsubjective, automated reproducible morphometric analysis. Other methods of visualizing brain vessels have been used in the past, but all possess disadvantages in comparison with the alkaline phosphatase histochemical technique. Some methods require staining of blood that remains in the vessels, while others require postmortem injections of barium-containing gelatin, India ink, or plastic (45,46). These methods can cause artifacts due to rupture of vessels, incomplete filling, and inclusion of air bubbles, all of which are avoided with the alkaline phosphatase method. The combination of the afferent vessel marker, alkaline phosphatase, with digital imaging and analysis results in a method to quantify the afferent cerebrovascular bed volume. The computerized methods we have developed and adapted for quantifying afferent vessels are considerably faster than alternative methods (25,47,48) that require postmortem injection, followed by noncomputerized quantitative processes that are tedious and slow compared with the automated method we used (25,48).
When using this method, we find that LA lesions are associated with a significantly decreased vascular density and that this decrease is especially apparent in the youngest subjects we studied (eg, those who died between 55 and 60 years of age). Across the age spectrum examined, the LA lesion averages 20% less afferent vessel concentration than does healthy deep WM. The fact that vascular density is decreased in LA lesions is not surprising, in view of known LA-associated abnormalities, such as loss of oligodendrocytes, axonopathy, apoptosis, and vacuolization that leads to spongiosis. Blood vessel loss could be a response to reduced metabolic requirements; however, the alternative, that vascular loss precedes parenchymal cell loss evident in either MR images or routine neuropathologic preparations, cannot be excluded. We found evidence of this alternative.
We compared tissue from similar regions in subjects without and those with LA. In the latter group, we found decreased afferent vascular density in the LA lesion, healthy appearing WM, and cortex. Despite an obvious loss of afferent vasculature, there is no alteration to the neuropil of the cortex in subjects with LA that is detectable with MR imaging or histologic analysis. Similarly, a significant loss of afferent vasculature in healthy-appearing WM in subjects with LA is not associated with a histologically detectable abnormality. Thus, our findings indicate that the vessel loss appears to precede any visible damage to the parenchyma and that the areas most severely affected are those most susceptible to ischemia.
Our measurements of decreased afferent vascular density indicate that a 55-year-old subject with deep WM hyperintensities on a T2-weighted MR image has the afferent cerebrovascular bed volume similar to that of a 70- or 80-year-old subject. The diagnosis of LA, based on findings at MR imaging, has been viewed as an indicator of localized deep WM degeneration restricted to older individuals. Our findings of decreased afferent vascular density in the area of the LA lesion and outside the lesion in healthy-appearing WM and the cortex led us to identify LA as a general cerebrovascular disease process.
The finding of a substantial age-related decline in afferent vascular density in the brains of healthy subjects is in itself an important discovery. In addition, it is important to note that use of age-matched controls is not always sufficient to identify changes in parameters that may vary with increasing age. Subjects with LA who die before 80 years of age show substantially decreased vascular density, whereas those dying around 80 years of age show no difference compared with age-matched controls. Thus, heavy sampling in the older subjects could obscure the changes, and graphs showing changes over time give a better understanding than a simple bar graph showing the mean and standard deviation. Another potentially significant finding is the apparent floor on vascular density in the brain. The flat line in afferent vascular density in subjects with LA suggests that vascular densities below a certain level may be incompatible with viability.
LA is a generalized cerebrovascular disease, the initial stages of which can be observed morphometrically before abnormalities are apparent at MR imaging. The presence of deep WM hyperintensities on MR images may be used as a clinical sign to alert physicians to the existence of influences that act to the detriment of brain vasculature. A precise knowledge of the disease process could facilitate the development of promising interventional therapy that could be applied before the disease inflicts neurologic disability.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Authors stated no financial interest to disclose.
Author contributions: Guarantors of integrity of entire study, D.M.M., C.R.T., J.A.A., W.R.B., V.R.C.; study concepts, D.M.M., C.R.T., J.A.A., W.R.B., V.R.C.; study design, C.D.L., D.M.M., C.R.T., W.R.B.; literature research, J.A.A., C.R.T., W.R.B., D.M.M.; experimental studies, D.M.M., C.R.T., J.A.A., W.R.B., V.R.C.; data acquisition, C.R.T., D.M.M.; data analysis/interpretation, C.D.L., C.R.T., D.M.M., J.A.A., W.R.B.; statistical analysis, C.D.L., C.R.T., W.R.B.; manuscript preparation, C.D.L., D.M.M., W.R.B., C.R.T., J.A.A.; manuscript definition of intellectual content, D.M.M., W.R.B., C.R.T., J.A.A.; manuscript editing, D.M.M., C.R.T., J.A.A., W.R.B., C.D.L.; manuscript revision/review and final version approval, all authors
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A.J. Bastos-Leite, J.P.A. Kuijer, S.A.R.B. Rombouts, E. Sanz-Arigita, E.C. van Straaten, A.A. Gouw, W.M. van der Flier, P. Scheltens, and F. Barkhof Cerebral Blood Flow by Using Pulsed Arterial Spin-Labeling in Elderly Subjects with White Matter Hyperintensities AJNR Am. J. Neuroradiol., August 1, 2008; 29(7): 1296 - 1301. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. N. K. Nandigam, K. Park, and N. Yasuda Significant association between leukoaraiosis and metabolic syndrome in healthy subjects Neurology, July 15, 2008; 71(3): 227 - 228. [Full Text] [PDF]
|
|
|
|
 |
|
 H. Ay, E. M. Arsava, J. Rosand, K. L. Furie, A. B. Singhal, P. W. Schaefer, O. Wu, R. G. Gonzalez, W. J. Koroshetz, and A. G. Sorensen Severity of Leukoaraiosis and Susceptibility to Infarct Growth in Acute Stroke Stroke, May 1, 2008; 39(5): 1409 - 1413. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Pieperhoff, L. Homke, F. Schneider, U. Habel, N. J. Shah, K. Zilles, and K. Amunts Deformation Field Morphometry Reveals Age-Related Structural Differences between the Brains of Adults up to 51 Years J. Neurosci., January 23, 2008; 28(4): 828 - 842. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. D. Nave, S. Foresti, A. Pratesi, A. Ginestroni, M. Inzitari, E. Salvadori, M. Giannelli, S. Diciotti, D. Inzitari, and M. Mascalchi Whole-Brain Histogram and Voxel-Based Analyses of Diffusion Tensor Imaging in Patients with Leukoaraiosis: Correlation with Motor and Cognitive Impairment AJNR Am. J. Neuroradiol., August 1, 2007; 28(7): 1313 - 1319. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Lin, L. Svensson, R. Gupta, B. Lytle, and D. Krieger Chronic ischemic cerebral white matter disease is a risk factor for nonfocal neurologic injury after total aortic arch replacement J. Thorac. Cardiovasc. Surg., April 1, 2007; 133(4): 1059 - 1065. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R Peters Ageing and the brain Postgrad. Med. J., February 1, 2006; 82(964): 84 - 88. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
.05. B, D, and F show vessel densities in healthy subjects (
and solid line) and those with LA (
and dashed line) as a function of age. Exponential decay curves were plotted with nonlinear regression analysis of data. Deep WM is hyperintense on T2-weighted MR images obtained in patients with LA; all other brain areas are not. A and B show afferent vessel density in deep WM of healthy subjects (Normal DWM) and in deep WM lesion of subjects with LA (LA DWM). In A, density of afferent vessels in subjects with healthy deep WM is significantly (P = .018) greater than that in deep WM lesions in subjects with LA. In B, normal vessel density decreases with age from approximately 6% at 55-60 years of age to 2% at 85-90 years of age. Afferent vessel density in the LA lesion in deep WM is 2%-3% at 55-60 years of age, which is less than half the level observed for healthy deep WM, and does not decline further with increasing age. The difference in the slope of lines for the two groups is significant (P = .009). C and D show comparison of afferent vascular density in WM that appeared healthy on T2-weighted MR images in subjects with LA lesions elsewhere (LA NAWM) and corresponding areas in healthy subjects (Normal WM). In C, average density of afferent vessels in healthy WM is not significantly different than that in healthy-appearing WM in subjects with LA. In D, vessel density decreases from approximately 8% at 55-60 years to 2% at 85-90 years in healthy subjects. In subjects with LA, afferent vessel density in healthy-appearing WM is 3%-4% at 55-60 years of age, which is almost half that seen in similar areas in healthy subjects, and does not decline significantly with increasing age. The difference in the slope of lines for the groups is significant (P = .006). E and F show comparison of afferent vascular density in cortex of healthy subjects (Normal Cortex) and those diagnosed with LA (LA Cortex). In E, density of afferent vessels in cortex of healthy subjects is not significantly different from that in cortex of subjects with LA. In F, afferent vessel density in cortex of healthy subjects decreases from approximately 16% at 55-60 years of age to approximately 6% at 90 years of age. While vessel density in cortex of subjects with LA is already reduced by almost 40% at 55-60 years of age, it does not decrease further. When afferent vascular density for the groups is plotted relative to increasing age, difference in the slopes of the lines approaches statistical significance (P = .051).