| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, Hôpital Erasme, Université Libre de
Bruxelles, 808, Route de Lennik, 1070, Brussels, Belgium.
2 Laboratory of Histopathology, Hôpital Erasme, Faculty of Medicine,
Université Libre de Bruxelles, 1070, Brussels, Belgium.
3 PET/Biomedical Cyclotron Unit, Hôpital Erasme, Université Libre
de Bruxelles, 1070, Brussels, Belgium.
4 Institute of Physiological Chemistry, Ludwig-Maximilians University, 13,
Veterinarstraat, 80539 Munich, Germany.
5 Department of Pathology, Hôpital Erasme, Université Libre de
Bruxelles, 1070, Brussels, Belgium.
Received September 30, 2002;
accepted after revision December 6, 2002.
N. Sadeghi was supported by an FNRS (Fond National de Recherche
Scientifique) grant.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Nineteen patients with primary glial brain tumors were included in the study. Mean ADC values were calculated in all tumors and were normalized with the ADC values of the contralateral normal-appearing brain ratios. All tumors underwent surgical resection, and the histologic diagnosis was based on the analysis of the surgical specimen. Mean values of the labeling index of hyaluronan (LI-HA) were calculated to determine quantifiably the histochemical expression of hyaluronan in the tumor. The mean ADC values and the mean ADC ratios (ADCratio) of the tumors were then correlated to the mean values of the LI-HA.
RESULTS. The mean ADC (93 x 10–5
mm2/sec) and the mean ADCratio (1.25) of the high-grade
glial tumors were significantly lower than the mean ADC (123 x
10–5 mm/sec) and the mean ADCratio (1.64) of the
low-grade glial tumors (p < 0.01). The mean LI-HA (72.8%) was also
significantly lower in the high-grade gliomas than the mean LI-HA (93.4%) in
the low-grade gliomas (p < 0.001). A positive correlation was
found between mean ADC values and the mean LI-HA (
= 0.35, p
< 0.05) and also between the mean ADCratio and the mean LI-HA
( = 0.33, p < 0.05).
CONCLUSION. Hyaluronan as one of the main hydrophilic components of the extracellular matrix in gliomas likely contributes to differences in the ADC values between high- and low-grade glial tumors.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The exact role of this new technique in the evaluation of intraaxial brain tumors is still undergoing evaluation. After the initial enthusiasm for this technique, inherent limitations of diffusion imaging are now being defined [11]. In fact, no obvious advantage of diffusion-weighted imaging is found with respect to the evaluation of tumor extension because the contrast between tumor and white matter is generally lower in diffusion-weighted imaging and ADC maps as compared with conventional MR imaging [12]. Also, a large variability exists in each group of tumors, and no consistent significant difference in intensity between groups is found on diffusion-weighted imaging [11, 13]. Nevertheless, quantifiable diffusion imaging techniques with ADC measurements reveal diagnostic information on human gliomas that is not detectable with the qualitative visualization of diffusion images [11]. Indeed, significant differences have been reported between the ADC values of low- and high-grade gliomas [13, 14]. These differences appear to be partly related to the differences in cell density of these tumors [13–16].
A substantial portion of the tumor volume is made up of extracellular matrix, which is composed of various proteins and polysaccharides that are expressed by the tumor cells and the tumor vessels.
Two main classes of extracellular macromolecules make up the extracellular matrix. The first group of macromolecules consists of polysaccharide chains called glycosaminoglycans that are linked to proteins in the form of proteoglycans. The second group of macromolecules consists of fibrous proteins (collagen and elastin) and fibrous adhesive proteins (fibronectin and laminin). Several glycosaminoglycans have been localized to the tumor cell–associated extracellular matrix of astrocytic glial tumors in vivo. The glycosaminoglycans are highly hydrophilic and tend to accumulate a negative charge that attracts cations such as sodium, which are osmotically active, causing the shift of large amounts of water. Therefore, glycosaminoglycans are thought to influence the water content of the extracellular matrix and thus the value of ADC.
Hyaluronan or hyaluronic acid is a highly hydrophilic glycosaminoglycan of the extracellular matrix that surrounds tumoral astrocytes [17]. Hyaluronan is a uniformly repetitive, linear glycosaminoglycan with the two monosaccharides—glucuronic acid, which introduces the negative charge, and N-acetylglucosamine—acting as building blocks and extending to lengths from 2 to 25 µm. Hyaluronan's unique viscoelastic properties with an enormous amount of solvent attached to it is one reason for its wide occurrence in the body [18–20]. In addition to these properties, the saccharide chain is also a carrier of biologic information decoded by specific receptor proteins, making hyaluronic acid a passive and active component in tissue assembly, homeostasis, and tumor progression [21–24].
Hyaluronan has already been proven to play a number of important roles in the biologic behavior of human glial tumors involving cell migration and proliferation [25, 26]. Hyaluronan plays a critical role in the control of cell proliferation and migration because its level of expression changes significantly during the progression of malignancy in human astrocytic tumors [17]. Low-grade astrocytomas have a concentration of glycosaminoglycans three times higher than that of the normal white matter, partly due to a higher hyaluronan content. Oligodendrogliomas also have a high concentration of glycosaminoglycans but with a proportion of hyaluronan less than that of astrocytomas. The concentration of glycosaminoglycans decreases in glioblastomas and particularly in anaplastic astrocytomas [17]. Our hypothesis is that hyaluronan as one of the main hydrophilic extracellular components of the glial tumors may influence the measured values of ADC in high- and low-grade glial tumors. The purpose of our study is therefore to search for a possible relationship between the histochemical level of hyaluronan and the ADC values measured in high- and low-grade glial tumors.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following sequences were acquired for each patient: axial T1-weighted
imaging with parameters of a TR/TE of 422/9, a slice thickness of 5 mm, a
field of view of 180 x 240, and a matrix of 179 x 256; axial
T2-weighted spin-echo MR imaging with parameters of 5081/100, a slice
thickness of 5 mm, a field of view of 180 x 240, and a matrix of 224
x 256; and axial turbo spin-echo fluid-attenuated inversion recovery
(FLAIR) imaging with parameters of 6500/150, an inversion time of 2100 msec, a
slice thickness of 5 mm, a field of view of 180 x 240, and a matrix of
140 x 256. Diffusion-weighted images were acquired using a multislice
single-shot spin-echo echoplanar imaging sequence in the transverse plane with
parameters of 4784/129, an echoplanar imaging factor of 63, a slice thickness
of 6 mm, a field of view of 180 x 230, and a matrix of 78 x 128.
The diffusion gradient was applied in the three orthogonal directions of the
space (the x-, y-, and z-axes). Diffusion sensitivity was
determined with motion-probing gradient factor b according to the equation,
 |
is the gyromagnetic ratio, G is the gradient amplitude,
is the gradient duration, and 
is the interval between
diffusion gradient pulses. Diffusion-weighted MR images were obtained with b factors of 0 and 1000 sec/mm2 for each section in the same sequence. ADC maps were then constructed automatically for all the diffusion-weighted MR images by means of pixel-by-pixel calculation.
We finally obtained contrast-enhanced axial T1-weighted images after an IV
injection of 0.1 mmol of gadopentetate dimeglumine (Dotarem, Guerbet
Laboratories, Aulnay-sous-Bois, France) per kilogram of body weight. Uniform
regions of interest (ROIs) of 49–51 mm2 were manually drawn
on ADC maps of each tumor with an electronic cursor. The solid portions of the
tumor were defined on the basis of conventional MR imaging (T1-weighted,
T2-weighted, FLAIR, and T1-weighted imaging with contrast material). ROIs were
carefully placed in solid areas of each tumor so that they did not include
identified cystic–necrotic or hemorrhagic regions that might influence
ADC values. The ADC values for each lesion were measured at five different
locations of the tumor, and the value for each ROI was included in the
computation; an average value was used for the analyses. For each ROI placed
in the tumor, same-sized uniform ROIs were drawn in matching structures in the
contralateral hemisphere to obtain ADC values of normal-appearing white matter
for the purpose of normalization. If the contralateral hemisphere was involved
with tumor, a comparable normal structure was considered for measurements. The
normalization was applied to deal with intersubject variability and to allow
data comparison between studies. The ADCs were calculated using the following
equation:
 |
Histochemical Data Acquisition
The histochemical procedures were carried out as detailed previously
[28,
29]. Five 1-µm-thick
sections were obtained from each surgical specimen. Incubation with the two
histochemical probes was carried out at 25°C ± 1°C for 60 min.
The extent of the specifically bound antibody (antihyaluronan) was visualized
by avidin-biotin-peroxidase (Vector Labs, Burlingame, CA) complex kit reagents
with diaminobenzidine and H2O2 as chromogenic
substrates. The antihyaluronan antibody is a monoclonal mouse antibody raised
against native trophoblast membranes on which hyaluronic acid is the only type
of glycosaminoglycan (AnaWa Trading, Wangen Zürich, Switzerland) present.
One variable quantifiably characterizing the histochemical staining for
hyaluronan (by means of the antihyaluronan antibody) was determined by a 2005
computer-assisted microscope system (Samba Technologies, Grenoble, France)
with a 20x (aperture, 0.50) magnification lens.
The labeling index of hyaluronan (LI-HA) refers to the percentage of tissue area specifically stained by a given histochemical marker. The labeling index was calculated as follows. A tri-CCD color camera (DONPISHA, Brussels, Belgium) was used to compute 768 x 512 (i.e., 393216) pixels. The segmentation procedure automatically associated with the computer-assisted microscope-related procedure, which registered as 0r for each pixel that did not correspond to the tissue and as 1r for each pixel that corresponded to the tissue, was analyzed in the red channel of the tri-CCD camera. The software labeled (in the green channel of the tri-CCD camera) each 1r pixel as 0g if the tissue did not exhibit a specific histochemical staining and as 1g if it did. Thus, 1r0g pixels represented unstained tissue and 1r1g, specifically stained tissue. The labeling index, therefore, represents the sum of the 1r1g pixels (of the 393216 pixels computed for each of the 10 histologic fields calculated for each case) divided by the sum of the 1r0g + 1r1g pixels (and expressed as percentages). The tri-CCD camera delivered an 8-bit message (i.e., 256 levels) of gray values.
The way in which the computer-assisted system was used to quantify the histochemical staining is detailed elsewhere, as are all standardization procedures dealing with the manner in which the computer-assisted microscopy was used [28, 29]. A negative finding on a histologic control slide (from which the primary antihyaluronan antibody was omitted, as we mentioned previously) was analyzed for each case under study. The software used on the computer-assisted microscope automatically subtracted the LI-HA values of the negative control sample from each corresponding positive one.
Only tumor areas were analyzed on the histologic slides. Areas identified as necrosis, cystic changes, or hemorrhage were excluded from analysis. This method was possible for the software that we had set up and that enables specific tissue of interest to be analyzed on a given histologic slide by means of a mouse link. In each case, 10 fields of between 60,000 and 120,000 mm2 were scanned for tumor tissues.
Data Analysis
For each ROI, the ADC ratio (ADCratio) was obtained by dividing
the measured ADC of the tumor (ADCtum) by the ADC value of
normal-appearing white matter in the contralateral hemisphere in the same
patient. The mean ADCtum, mean ADCratio, and mean LI-HA
were calculated for each tumor. These variables were then compared between
high- and low-grade gliomas. The statistical comparisons of the data were
carried out using the nonparametric Mann-Whitney test (for two groups). The
mean ADCtum and the mean ADCratio of each lesion were
then correlated with the mean LI-HA. The rank correlation test involved the
nonparametric Kendall's coefficient test. A p value of less than 0.05
was considered as statistically significant. All statistical analyses were
carried out using Statistical (Statsoft, Tulsa, OK).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
|
|
|
|
Histochemical Analysis
Measured LI-HA in the high-grade group (n = 11) varied from 54.7%
to 89.4%, with a mean of 72.8% ± 11.8%. The LI-HA in the low-grade
group (n = 8) varied from 81.5% to 98%, with a mean of 93.4% ±
5.4%. This difference in mean LI-HA was also statistically significant
(p < 0.001). Figures
1E and
2E illustrate histomorphologic
features of low-grade and high-grade glial brain tumors, respectively, whereas
Figures 1F and
2F depict the morphologic
appearance of hyaluronan immunoreactivity in the corresponding tumors.
|
|
|
|
Comparison of ADC with Histochemistry
Table 1 summarizes the mean
values of ADC, ADCratio, and LI-HA in relation to each tumor grade.
Values of p are given comparing each of the three variables between
the two groups of tumors. The most significant difference was found between
high- and low-grade gliomas when considering mean ADC values and LI-HA.
Measured values of mean ADCtum, mean ADCratio, and mean
LI-HA for each group of high- and low-grade tumors are plotted in Figures
3,
4,
5. Figures
3 and
4 show that both mean
ADCtum and mean ADCratio are lower in high-grade gliomas
compared with low-grade gliomas. Figure
5 shows that mean LI-HA is also lower in high-grade tumors
compared with low-grade tumors. A positive correlation was found between mean
ADCtum values and mean LI-HA ( = 0.35, p < 0.05)
(Fig. 6) and also between mean
ADCratio and mean LI-HA ( = 0.33, p < 0.05).
|
|
|
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Previous studies concerned mostly astrocytic tumors, and less information exists about ADC values in oligodendroglial tumors. It has been previously shown [13–16] that the differences in the ADC values that can be observed between low- and high-grade astrocytic tumors relate partly to the cell density of these tumors. Higher cell densities are observed in high-grade rather than in low-grade astrocytic tumors, and an inverse correlation was found between cell density and ADC values in astrocytic tumors [13–16].
In addition to cell density, components of the extracellular matrix may influence the ADC values in glial tumors and especially those components such as glycosaminoglycans that are hydrophilic and whose levels of expression differ according to whether a tumor is of low as opposed to high histopathologic grade. Hyaluronan is a glycosaminoglycan that is highly expressed in glial tumors. The percentages of hyaluronan are 56% for low-grade astrocytomas, 34% for oligodendrogliomas, 32% for anaplastic astrocytomas, and 29% for glioblastomas, respectively [17]. Considering this previously reported information on hyaluronan content of glial tumors, we included in our present study a relatively heterogeneous group of tumors of astrocytic and oligodendrocytic origin to search for a possible relation between ADC values and the level of hyaluronan in the extracellular matrix of glial tumors. Considering the fact that a different method of measurement was used in our series, we confirmed a previous demonstration of higher levels of hyaluronan expression in the low-grade gliomas as opposed to the high-grade gliomas. Furthermore, a positive correlation was shown between ADC values of the tumors (both for ADCtum and ADCratio) and the level of hyaluronan expression in our group of human glial tumors of mixed astrocytic and oligodendrocytic types. Thus, the data from our study suggest that the ADC value in glial tumors is determined at least partly by the expression of hyaluronan in the tumor matrix.
High levels of hyaluronan in low-grade gliomas could cause a high degree of viscosity in the extracellular matrix. On this basis, one would expect lower ADC values in low- as opposed to high-grade tumors. However, in our series a positive correlation was found between ADC and the hyaluronan expression. We therefore assumed that this positive correlation may actually reflect a high level of water mobility in this hydrophilic environment produced by high levels of hyaluronan, leading to a net increase in water content of the extracellular matrix.
The differences in hyaluronan values between tumor groups may partly relate to a relative difference in the amount of extracellular matrix. In other words, higher cellular density could lead to a smaller amount of extracellular matrix and therefore a smaller amount of hyaluronan per unit of volume. This could be assessed by determining the ratio of hyaluronan to the amount of extracellular space. However, the amount of extracellular space not only relates to cell density, assessed by nuclear counting, but also relates to cell volume. Thus, both nuclear count and cell volume must be known to measure the relative amount of extracellular space volume. These measurements were not performed in our study and may be the subject of further studies.
Regions of high- and low-grade tumor can exist in the same tumor, and presumably regions of different hyaluronan expression are heterogeneous as well. This is one of the reasons for which five ROIs for ADC measurements and 10 fields for HA-LI measurements were used to calculate a mean value for the purpose of comparison. One of the limitations of this study was that the MR imaging examination was not coregistered with the surgical specimen. Thus, the histochemical analysis could not be performed exactly on the ROIs where the ADC measurements were performed. However, analyses were focused on the solid portion of the tumor both macroscopically on the MR images and microscopically on the histologic specimens, with exclusion of areas of hemorrhage and cystic or necrotic changes. For a better recognition of the extracellular components that may potentially influence the value of ADC in brain tissue in general and in glial tumors in particular, anatomic correlation between the MR imaging and histopathologic analyses should be improved. To do so, ADC measurements should be performed at the precise location of imaging-guided biopsies, as in previously reported metabolic–histopathologic correlation studies [30, 31] in brain tumors. This is the subject of our ongoing investigations.
The other limitation of this study is the univariate statistical analysis between ADC values on MR imaging and the hyaluronan expression in tissue samples. Other variables such as tumor cellularity, capillary density, calcification, tumor necrosis, and tumor hemorrhage are known to or can affect the ADC values. The imperfect correlation between ADC values and the hyaluronan expression level is probably related to these other variables that have not been considered in our present study. The present preliminary results should prompt a multivariate statistical analysis with inclusion of multiple histologic covariables in a larger set of data.
In addition to hyaluronan, other extracellular matrix components such as heparan sulfate and dermatan sulfate should also be evaluated in future studies to define a more precise relationship between ADC values in a given glial tumor and the individual properties of its extracellular matrix. Further study of these variables with comparison of ADC values would help to provide a greater understanding of the individual effects of each extracellular matrix component on observed ADC values.
Another limitation of our study is the low number of low-grade astrocytic tumors precluding a refined analysis of the relationship between the histologic type and the histologic determinants of ADC values.
In conclusion, our study shows a positive correlation between ADC values and the immunohistochemical level of hyaluronan in glial tumors. As such, our data are helpful in the comprehension of measured ADC values in vivo and can be used as an initial step for further studies in the evaluation of extracellular matrix determinants of ADC values in normal and abnormal brain tissue.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  N. Sadeghi, N. D'Haene, C. Decaestecker, M. Levivier, T. Metens, C. Maris, D. Wikler, D. Baleriaux, I. Salmon, and S. Goldman Apparent Diffusion Coefficient and Cerebral Blood Volume in Brain Gliomas: Relation to Tumor Cell Density and Tumor Microvessel Density Based on Stereotactic Biopsies AJNR Am. J. Neuroradiol., March 1, 2008; 29(3): 476 - 482. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Provenzale, S. Mukundan, and D. P. Barboriak Diffusion-weighted and Perfusion MR Imaging for Brain Tumor Characterization and Assessment of Treatment Response. Radiology, June 1, 2006; 239(3): 632 - 649. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. K. Koeller and E. J. Rushing From the Archives of the AFIP: Oligodendroglioma and Its Variants: Radiologic-Pathologic Correlation RadioGraphics, November 1, 2005; 25(6): 1669 - 1688. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. J. Fabiano, M. A. Horsfield, and R. Bakshi Interhemispheric Asymmetry of Brain Diffusivity in Normal Individuals: A Diffusion-Weighted MR Imaging Study AJNR Am. J. Neuroradiol., May 1, 2005; 26(5): 1089 - 1094. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. R. Jager, A. D Waldman, C. Benton, N. Fox, and J. Rees Differential Chemosensitivity of Tumor Components in a Malignant Oligodendroglioma: Assessment with Diffusion-Weighted, Perfusion-Weighted, and Serial Volumetric MR Imaging AJNR Am. J. Neuroradiol., February 1, 2005; 26(2): 274 - 278. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE |
