Introduction

Neoplasms require a rich vascular network to supply the blood necessary for proliferation. Glioma and meningioma are the two most common types of primary brain tumor. They present very different vascular properties [1Surawicz T.S., McCarthy B.J., Kupelian V., Jukich P.J., Bruner J.M., Davis F.G. Descriptive epidemiology of primary brain and CNS tumors: results from the Central Brain Tumor Registry of the United States, 1990–1994 Neuro-Oncol 1999 ;  1 : 14-25 [cross-ref]

Click here to see the Library]. Meningiomas are the most common extra-axial brain tumors, with a vascular supply that is usually derived from dural vessels [2Dowd C.F., Halbach V.V., Higashida R.T. Meningiomas: the role of preoperative angiography and embolization Neurosurg Focus 2003 ;  15 : E10

Click here to see the Library]. Gliomas, the most common intra-axial primary brain tumor, have variable degrees of neovascularity and blood–brain barrier (BBB) alterations, depending on the grade and histological subtype. Approaching the pathophysiology of peritumoral edema through perfusion imaging with the use of maximum signal drop (MSD) rates as a parameter, we speculated that vascularization within peritumoral edema would be raised in glioblastoma and that the vascular contribution, as revealed by the MSD, would be higher in glioblastoma than its potential mass effect. Although meningioma and glioblastoma both have mass effects, glioblastoma presents with higher rates of angiogenesis. For this reason and because the mass effect may be a confounding co-variable, we chose meningioma as our model of mass effect. The objective of our study was to assess the angiogenesis, using dynamic susceptibility contrast magnetic resonance (MR) perfusion imaging, in the peritumoral tissue, often seen with meningioma and glioblastoma.

	Materials and Methods

	Patients

This prospective study was conducted from December 2003 to March 2005, and was approved by the local ethics committee. Patients were informed of their participation in the study, but their signed consent was not necessary as the imaging protocol was already in place for tumor imaging at our institution. Only patients with meningioma or glioblastoma were recruited, and these two tumor types were selected because of their different vascular properties [1Surawicz T.S., McCarthy B.J., Kupelian V., Jukich P.J., Bruner J.M., Davis F.G. Descriptive epidemiology of primary brain and CNS tumors: results from the Central Brain Tumor Registry of the United States, 1990–1994 Neuro-Oncol 1999 ;  1 : 14-25 [cross-ref]

Click here to see the Library]. Also, only patients with more than 30mm of peritumoral edema were included. All lesions were histologically confirmed. Of 18patients recruited, six were dropped from the study: three because of considerable head movement; two because of minimal spread of edema (<30mm); and one because of very poor-quality images (geometric distortions). Of the 12 included patients (gender ratio M/F: 0.33; mean age: 59.5 years [min: 50, max: 75 years]), six presented with gliolastoma and six with meningioma.

	MRI data acquisition

All patients were examined with a 1.5-T clinical MR imaging (MRI) unit (Signa 9, gradient strength 23mT/m; General Electric Medical Systems, Milwaukee, WI), using the standard head coil, and all underwent the following imaging protocol: axial fast spin-echo (SE) T2-weighted; axial SE T1-weighted; axial FLAIR with automatic injection of gadolinium bolus (0.2mmol/kg); axial echo-planar perfusion [3Knopp E.A., Cha S., Johnson G., and al. Glial neoplasms: dynamic contrast-enhanced T2*-weighted MR imaging Radiology 1999 ;  211 : 791-798

Click here to see the Library] (echo time: 60mins; repetition time: 2300mins; matrix: 128×96; field of view: 24×24cm²; slice thickness: 5mm; intersection gap: 1mm; NEX: 1; 12-slice; 30-scan); and axial SE T1-weighted post-gadolinium.

	Perfusion parameter

To avoid the effects of arterial input function, we used the MSD rate, which corresponds to the decrease in signal intensity induced by a susceptibility magnetic effect when the contrast agent bolus passes through the capillary vessels. The MSD parameter is an useful indicator of the status of brain microcirculation, as it is sensitive to weak signal–noise ratios and faithfully reflects the cerebral blood volume of the brain [4Cha S., Lu S., Johnson G., Knopp E.A. Dynamic susceptibility contrast MR imaging: correlation of signal intensity changes with cerebral blood volume measurements J Magn Reson Imaging 2000 ;  11 : 114-119 [cross-ref]

Click here to see the Library]. This allows the use of this parameter to explore alterations in blood volume related to either a mass effect (meningioma) or angiogenesis extending into the parenchyma (glioblastoma). All signal intensities within the areas of peritumoral brain edema were measured. The MSD rates were then compared with those of the contralateral normal white matter (relative MSD, rMSD) in all patients. FLAIR sequencing was used to ensure that all measures were taken in the surrounding edema present with all tumors.

	Signal analysis

Image processing was performed using PERFTOOL software. The algorithm is based on the intravascular indicator dilution theory [5Rosen B.R., Belliveau J.W., Vevea J.M., Brady T.J. Perfusion imaging with NMR contrast agents Magn Reson Med 1990 ;  14 : 249-265 [cross-ref]

Click here to see the Library]. Instead of deconvolving the signal, we determined the maximum signal decrease from an average baseline of six scans. Using FLAIR images to delineate peritumoral hypersignals, and T1-weighted post-gadolinium images to delineate the edges of the tumor, we drew several rectangular regions of interest (ROI) around each patient’s lesion; each ROI contained 20voxels (468mm³). The ROI extended from close to the lesion, and moved gradually outwards towards areas of distant edema in axial and sagittal planes at intervals of 10, 20 and 30mm. We also drew ROI on the contralateral brain white matter to determine a normal baseline of comparison for our measures (Figure 1).

Figure 1
	Figure 1.  PERFTOOL software. To obtain a perfusion curve, we drew 3 ROI near the lesion and gradually extended outwards to distant edema in a sagittal plane at 10, 20 and 30mm from the tumor. We also drew ROI on contralateral brain white matter to obtain a normal baseline for our measures. Interface du logiciel «  Perftool ». Exemple de courbe de perfusion tissulaire obtenue pour une des régions d’intérêt (ROI). Ces ROIs 3D sont dessinées dans l’œdème péritumoral (l’hémisphère droit) et dans la partie saine controlatérale (l’hémisphère gauche). Zoom

	Statistical analyses

The Mann–Whitney test was used to determine whether or not the calculated MSDs were significantly different between glioblastoma and meningioma in peritumoral tissue. Values were calculated as means and standard deviations, and p <0.05 was considered significant.

	Results

The mean rMSDs found in the surrounding edema of glioblastoma and meningioma were 0.58±0.37 and 0.82±0.43 respectively, and the difference was significant (p <0.001). The rMSDs in the surrounding edema of glioblastoma were, on average, 0.76±0.13 at 10mm, 0.64±0.21 at 20mm and 0.51±0.15 at 30mm from the tumor. The rMSDs in the surrounding edema of meningioma were, on average, 0.48±0.08 at 10mm, 0.69±0.08 at 20mm and 0.81±0.08 at 30mm from the tumor.

	Glioblastoma

We observed a decrease in rMSD with distance from the tumor. The rMSD parameter was significantly different (p =0.01) between the ROI at 10mm and 30mm from the tumor. However, there were no significant differences in rMSD between ROI at 10mm and 20mm, and at 20mm and 30mm (Figure 2).

Figure 2
	Figure 2.  Histogram showing the decrease in relative maximum signal drop (rMSD) rates in the peritumoral edema of glioblastoma with distance from the lesion. Histogramme montrant la diminution de rMSD dans l’œdeme péritumoral des glioblastomes avec l’augmentation de la distance. Zoom

	Meningioma

In contrast, the value of the rMSD increased with distance from the tumor. In addition, the rMSDs were significantly different between ROI at 10mm and 20mm (p =0.006), at 20mm and 30mm (p =0.025) and at 10mm and 30mm (p =0.004) (Figure 3).

Figure 3
	Figure 3.  Histogram showing the increase in relative maximum signal drop (rMSD) rates in the peritumoral edema of meningioma with distance from the lesion. Histogramme montrant l’augmentation du rMSD dans l’œdeme péritumoral des méningiomes avec l’augmentation de la distance. Zoom

	Discussion

Tumor neoplasms need a rich vascular network to supply the blood necessary for proliferation. Angiogenesis is a complex process regulated by multiple stimulatory and inhibitory factors that are able to modulate the migration and/or proliferation of microvascular cells to form new vascular networks from the preexisting vessels. It involves well-coordinated steps, including: the production and release of angiogenic factors; proteolytic degradation of extracellular matrix components to allow formation of capillary branches; proliferation and directional migration of microvascular cells; and the final appearance of new vessels [6Senger D.R. Molecular framework for angiogenesis: a complex web of interactions between extravasated plasma proteins and endothelial cell proteins induced by angiogenic cytokines Am J Pathol 1996 ;  149 : 1-7

Click here to see the Library]. Angiogenesis also depends on tumor type: in benign neoplasms such as meningioma, vascular proliferation occurs by recruiting existing capillaries whereas, in high-grade glioma, the existing capillaries are inadequate to maintain a sufficient blood supply and thus, neoangiogenesis is triggered. Although both meningioma and glioblastoma have mass effects, this was not assessed in our present study as it was not deemed relevant. However, as the mass effect can be considered a confounding co-variable, we chose to use meningioma as a model of mass effect.

Regions of the highest MSD are those that histologically show the greatest hypervascularity [3Knopp E.A., Cha S., Johnson G., and al. Glial neoplasms: dynamic contrast-enhanced T2*-weighted MR imaging Radiology 1999 ;  211 : 791-798

Click here to see the Library]. However, MSD does not measure vascularity directly—it’s a complex function of contrast dose, rate of injection, blood volume, blood flow, cardiac output and MRI parameters. The MSD of the perfusion signal is a clinically useful indicator of brain microcirculation status, and can explore any alterations of cerebral blood volume related to angiogenesis in brain tumors.

Angiogenesis plays an important role in malignant primary tumors [7Cha S., Lupo J.M., Chen M.H., and al. Differentiation of glioblastoma multiforme and single brain metastasis by peak height and percentage of signal intensity recovery derived from dynamic susceptibility-weighted contrast-enhanced perfusion MR imaging AJNR Am J Neuroradiol 2007 ;  28 : 1078-1084 [cross-ref]

Click here to see the Library]. According to the World Health Organization (WHO) classification system [8Kleihues P., Louis D.N., Scheithauer B.W., and al. The WHO classification of tumors of the nervous system J Neuropathol Exp Neurol 2002 ;  61 : 215-225discussion 226–219.

Click here to see the Library], glioblastoma grade IV is the higher-grade histotype that includes endothelial hyperplasia, necrosis or both. Glioblastoma, the most malignant type of glioma in adults, is also among the most angiogenic of all human tumors. Moreover, glioblastoma is an infiltrative lesion that, to spread through edema, needs a vascular supply. The reduced rMSD parameter in the peritumoral edema of glioma with distance from the tumor corresponds to angiogenesis that is greatest nearer the tumor and decreases with distance from the tumor. This corresponds to the angiogenesis front. The lack of a statistically significant difference between the ROI at 10mm and 20mm, and at 20mm and 30mm, is probably related to angiogenesis heterogeneity. Glioblastoma development is not linear, so tumor infiltration and angiogenesis may well vary between two different ‘intermediate’ (20mm) ROI.

In contrast, meningioma is an extra-axial benign tumor that is often associated with cerebral edema. Its vascular supply is most commonly derived from dural vessels [2Dowd C.F., Halbach V.V., Higashida R.T. Meningiomas: the role of preoperative angiography and embolization Neurosurg Focus 2003 ;  15 : E10

Click here to see the Library]. Our patients did not have angiograms in their records but, based on surgical findings and the literature [9Bitzer M., Klose U., Geist-Barth B., and al. Alterations in diffusion and perfusion in the pathogenesis of peritumoral brain edema in meningiomas Eur Radiol 2002 ;  12 : 2062-2076

Click here to see the Library, 10Bitzer M., Wockel L., Luft A.R., and al. The importance of pial blood supply to the development of peritumoral brain edema in meningiomas J Neurosurg 1997 ;  87 : 368-373 [cross-ref]

Click here to see the Library], meningioma with a blood supply of exclusively dural origin are typically free of edema. Our patients presented major edema (>30mm), so their vascularization was considered to be both meningeal and pial.

Various hypotheses have been proposed for edema pathogenesis, including: ischemia, mechanically caused by tumor compression [11Tatagiba M., Mirzai S., Samii M. Peritumoral blood flow in intracranial meningiomas Neurosurgery 1991 ;  28 : 400-404 [cross-ref]

Click here to see the Library], which becomes more likely as the size of the tumor increases [12Go K.G., Wilmink J.T., Molenaar W.M. Peritumoral brain edema associated with meningiomas Neurosurgery 1988 ;  23 : 175-179

Click here to see the Library, 13Stevens J.M., Ruiz J.S., Kendall B.E. Observations on peritumoral oedema in meningioma. Part II: mechanisms of oedema production Neuroradiology 1983 ;  25 : 125-131 [cross-ref]

Click here to see the Library]; stasis, induced by tumor in venous drainage areas, followed by venous congestion [14Hiyama H., Kubo O., Tajika Y., Tohyama T., Takakura K. Meningiomas associated with peritumoral venous stasis: three types on cerebral angiogram Acta Neurochir (Wien) 1994 ;  129 : 31-38 [cross-ref]

Click here to see the Library]; excretory–secretory phenomena, in which substances produced by the tumor appear in adjacent brain tissue [15Philippon J., Foncin J.F., Grob R., Srour A., Poisson M., Pertuiset B.F. Cerebral edema associated with meningiomas: possible role of a secretory-excretory phenomenon Neurosurgery 1984 ;  14 : 295-301 [cross-ref]

Click here to see the Library], such as prostaglandins [16Constantini S., Tamir J., Gomori M.J., Shohami E. Tumor prostaglandin levels correlate with edema around supratentorial meningiomas Neurosurgery 1993 ;  33 : 204-210discussion 211.

Click here to see the Library], expression of hormone receptors [17Benzel E.C., Gelder F.B. Correlation between sex hormone binding and peritumoral edema in intracranial meningiomas Neurosurgery 1988 ;  23 : 169-174 [cross-ref]

Click here to see the Library] and secretion of vascular endothelial growth factor (VEGF) [18Kalkanis S.N., Carroll R.S., Zhang J., Zamani A.A., Black P.M. Correlation of vascular endothelial growth factor messenger RNA expression with peritumoral vasogenic cerebral edema in meningiomas J Neurosurg 1996 ;  85 : 1095-1101 [cross-ref]

Click here to see the Library]; and hydrodynamic processes, whereby agents extravasated from the tumor appear in the immediate surrounding brain tissue [19Bitzer M., Nagele T., Geist-Barth B., and al. Role of hydrodynamic processes in the pathogenesis of peritumoral brain edema in meningiomas J Neurosurg 2000 ;  93 : 594-604 [cross-ref]

Click here to see the Library, 20Ito U., Tomita H., Tone O., Masaoka H., Tominaga B. Peritumoral edema in meningioma: a contrast enhanced CT study Acta Neurochir Suppl (Wien) 1994 ;  60 : 361-364

Click here to see the Library]. We assumed there was no angiogenesis in areas of T2-weighted peritumoral hypersignals because the patients presented with no infiltrative or atypical meningioma. The MSD of the perfusion signal reveals alterations in cerebral blood volume. In our opinion, given the lack of angiogenesis in the T2 peritumoral hypersignals, part of the increased rMSD in the peritumoral edema of meningioma with distance from the tumor suggests compression phenomena. Normal vessels are compressed by the mass effect, which is greater nearer to the lesion. This mass effect probably depends on several associated mechanisms, such as tumor compression, excretory–secretory phenomena or hydrodynamic processes, which probably lead to vasogenic edema.

	Conclusion

Dynamic susceptibility contrast MR perfusion imaging, using the MSD as a parameter, demonstrates the difference in peritumoral tissue between meningioma and glioblastoma. According to our results, the differences observed were related to the angiogenesis process. Indeed, in meningioma, we have found peritumoral edema due to a mass effect, and tumor edema due to an infiltrative malignant lesion, mass effect and the development of tumor angiogenesis.