UNIVERSITE LIBRE DE BRUXELLES
FACULTE DE MEDECINE
Imagebased biomarkers for the invivo evaluation of human brain gliomas
Niloufar SadeghiMeibodi
Service de Radiologie Unité d’IRM
Service de Médecine Nucléaire Unité PET/cyclotron biomédical Service d’Anatomie Pathologique
Thèse présentée en vue de l’obtention du grade académique de Docteur en Sciences Médicales
Promoteur : Professeur Serge Goldman
Copromoteur : Professeur Isabelle Salmon
Je dédie ce travail de thèse à la mémoire de mon cher papa, le Professeur
Djamal SadeghiMeibodi
Professeur à la faculté de Médecine de l’Université de Téhéran
Remerciements
Mes remerciements s’adressent d’abord au Professeur Serge Goldman qui m’a guidée pendant toutes ces années avec tant de gentillesse et d’humanité. Cher Serge, les moments qu’on a passés à discuter sur les différentes étapes de ce projet sont parmi les moments les plus riches de ma vie professionnelle et rien que pour ça je ne te remercierai jamais assez.
Je remercie le Professeur Isabelle Salmon qui m’a accueillie dans son laboratoire et m’a aidée à développer ce projet. Merci Isabelle pour ta disponibilité, ton enthousiasme et tes conseils.
Je remercie le Professeur Julien Struyven de m’avoir donné la chance de devenir radiologue.
Je remercie le Professeur Danielle Balériaux qui m’a accueillie dans son département et m’a donné l’opportunité de développer mes projets et qui m’a soutenue dans les moments les plus difficiles.
Je remercie le Professeur Freddy Avni qui m’a encouragée durant ces années.
Je remercie le docteur Thierry Metens pour tous ses précieux conseils.
Je remercie les docteurs Philippe David, Carine Neugroschl et Isabelle Delpierre pour avoir supporté la surcharge de travail pendant mes années de recherche.
Je remercie tous les membres du Service d’Anatomie Pathologie et en particulier les docteurs Sandrine Rorive, Nicky D’Haene et Calliope Maris et enfin Madame Nathalie Watteau. Je remercie tout particulièrement le Professeur Christine Decaestecker.
Je remercie les Professeurs Jacques Brotchi, Marc Levivier, Florence Lefranc et Nicolas Massager.
Je remercie le docteur Patrice Jissendi pour m’avoir aidée à comprendre les bases de la spectroscopie.
Je remercie le personnel de l’Unité de Résonance Magnétique de même que le personnel du service des Isotopes et en particulier de l’unité PET/Cyclotron biomédical.
Je remercie nos secrétaires Laurence et Diane.
Je remercie le FNRS et le Fonds Erasme pour la recherche médicale.
Merci à ma grande famille en Iran et aux USA, mes sœurs, mon frère, mes neveux et nièces qui m’ont toujours donné de la force malgré la distance. Merci à ma chère maman qui a su seule, soutenir et protéger ses quatre enfants dispersés dans le monde.
Je remercie aussi tous mes amis en particulier, Azadeh, Sima, Najima, Théo. Merci aussi à vous Anne, Maxou, Doris, Michelle, Nathalie, Boris, Frédérique que je considère comme ma petite famille en Belgique.
Merci à toi Marco pour ton soutien et tes encouragements durant toutes ces années.
Et merci à toi Nima, d’être là et de rire à la vie et de toujours me montrer où se trouve l’essentiel.
Et enfin ma gratitude éternelle va à feu Monsieur le Professeur Pierre Lasjaunias.
TABLE
ABBREVIATIONS..………...7
SUMMARY……….8
1. INTRODUCTION………10
1.1. Epidemiology of gliomas………...10
1.2. Histopathologic classification and grading……….11
1.2.1. Astrocytic Tumors………13
1.2.1.1. Pilocytic astrocytoma (WHO Grade I)………13
1.2.1.2. Diffuse astrocytoma (WHO Grade II)……….14
1.2.1.3. Anaplastic astrocytoma (WHO Grade III)………...14
1.2.1.4. Glioblastoma (WHO Grade IV)………...15
1.2.2. Oligodendroglial Tumors……….17
1.2.2.1. Oligodendroglioma (WHO Grade II)………..17
1.2.2.2. Anaplastic oligodendroglioma (WHO Grade III)………..18
1.2.2.3. Oligoastrocytoma (WHO Grade II)………...18
1.2.2.4. Anaplastic oligoastrocytoma (WHO Grade III)………...19
1.2.3. Ependymal Tumors………...19
1.2.3.1. Ependymoma (WHO Grade II)………...20
1.2.3.2. Anaplastic ependymoma (WHO Grade III)………...20
1.2.4. Other tumors of glial origin……….20
1.3. Biology of gliomas………...21
1.3.1. Glioma protein expression………...21
1.3.2. Genetic alterations………21
1.3.3. Cell proliferation………..22
1.3.4. Cell migration………...23
1.3.5. Angiogenesis………..24
1.3.6. Tumour metabolism……….25
1.3.6.1. 18FFluorodeoxyglucose Positron Emission Tomography (FDGPET)………26
1.3.6.2. 11CMethionine Positron Emission Tomography (METPET)……….26
1.3.6.3. Other PET tracers……….26
1.4. Neuroimaging of gliomas ………...27
1.4.1. Morphologic MR brain imaging ………27
1.4.1.1. Tumor characterization and grading………..27
1.4.1.2. Stereotactic biopsy guidance and neuronavigation………28
1.4.2. Physiologybased MR brain imaging………..29
1.4.2.1. Proton MRspectroscopy………..29
1.4.2.1.1. Technique………29
1.4.2.1.2. Clinical applications in gliomas……….30
1.4.2.1.2.1. Grading gliomas………...30
1.4.2.1.2.3. Biopsy and radiotherapy guidance……….31
1.4.2.2. Diffusionweighted imaging ……….32
1.4.2.2.1. Technique………32
1.4.2.2.2. Clinical applications in gliomas………...34
1.4.2.2.2.1. Glioma characterization………..34
1.4.2.2.2.2. Grading gliomas………...34
1.4.2.2.2.3. Radiation necrosis versus tumour recurrence………..35
1.4.2.2.2.4. Prediction of patient outcome……….35
1.4.2.3. Perfusionweighted imaging ………35
1.4.2.3.1. Technique………35
1.4.2.3.2. Clinical applications in gliomas……….37
1.4.2.3.2.1. Grading gliomas………...37
1.4.2.3.2.2. Radiation necrosis versus tumor recurrence………37
1.4.2.3.2.3. Biopsy guidance………..37
1.4.2.3.2.4. Prediction of patient outcome……….38
1.4.3. MR imaging diagnostic strategy for intraaxial brain masses………...38
1.5. Relation between tumor biology and physiologybased MR imaging………..38
1.5.1. Relation between tumor biology and MR Spectroscopy………...39
1.5.2. Relation between tumor biology and ADC………39
1.5.3. Relation between tumor biology and CBV.………40
1.6. Physiologicbased MR imaging of the brain in patients with glioma………...40
1.6.1. Functional MRI………40
1.6.2. Diffusion tensor imaging………..41
2. OBJECTIVE ………..42
3. INVESTIGATIONS………...42
3.1. Assessing the relation between ADC and extracellular matrix hydrophilic components in gliomas represented by hyaluronan………...42
3.1.1. Introduction………..42
3.1.2. Materials and methods……….…42
3.1.3. Results………42
3.1.4. Discussion ……….45
3.2. Assessing the relation between CBV and tumor aminoacid metabolism……….……47
3.2.1. Introduction………..…47
3.2.2. Materials and methods……….…47
3.2.3. Results………47
3.2.4. Discussion………...…...48
3.3. Assessing the regional relation between CBV and both tumor metabolism and histopathological grading features (stereotactic approach)………50
3.3.1. Introduction………..…50
3.3.2. Materials and methods ………50
3.3.3. Results………52
3.3.4. Discussion………..……54
3.4. Assessing the regional relation between ADC and CBV and quantitative histopathology (stereotactic approach)………..……57
3.4.1. Introduction………..………57
3.4.2. Materials and methods……….…57
3.4.3. Results………58
3.4.4. Discussion………..…62
4. CONCLUSIONS AND PERSPECTIVES………...…66
5. REFERENCES………...…68
6. THESE ANNEXE………..…86
ABBREVIATIONS
ADCr: ADC ratio
ADCtum: ADC of the tumor
ADCwm: ADC value of normalappearing white matter Ang: Angiopoietin
BBB: BloodBrain Barrier
BOLD: Blood Oxygenation Level Dependent CBF: Cerebral Blood Flow
CBTRUS: Central Brain Tumor Registry of the United States CBV: Cerebral Blood Volume
Cho: Choline Cr: Creatine
CSF: Cerebrospinal fluid CT: Computed Tomography
DSC: Dynamic SusceptibilityContrast DTI: Diffusion Tensor Imaging DWI: Diffusion Weighted Imaging ECM: Extracellular Matrix
EGFR: epidermal growth factor receptor EPI: EchoPlanar Imaging
FDG: 18FFluorodeoxyglucose
FDGPET: 18FFluorodeoxyglucose Positron Emission Tomography FDOPA: 3,4dihydroxy618FfluoroLphenylalanine
FET: O(218Ffluoroethyl)tyrosine
FLAIR: FluidAttenuated Inversion Recovery FLT: 3’deoxy3’18Ffluorothymidine fMRI: Functional MR imaging
GFAP: Glial Fibrillary Acid Protein H&E: Hematoxylin and Eosin HA: Hyaluronan
HIF: HypoxiaInducible Factor IDH1: Isocitrate dehydrogenase 1 LI: Labelling Index
LIHA: Labelling Index for Hyaluronan LOH: Loss of heterozygosity
MD: Mean Diffusivity MET: 11CMethionine
METPET: 11CMethionine Positron Emission Tomography MGMT: O6methylguanine methyltransferase
mI: Myoinositol
MR: Magnetic Resonance
MRI: Magnetic Resonance Imaging MRS: MR Spectroscopy
NAA: NAcetyl Aspartate NFP: NeuroFilament Protein
PCNA: Proliferating Cell Nuclear Antigen PET: Positron Emission Tomography PWI: Perfusion Weighted Imaging rCBV: relative CBV
ROI: Region of Interest
SPARC: Secreted Protein Acidic and Rich in Cysteine TE: Echo Time
TP53: Tumor Protein 53
VEGF: Vascular Endothelial Growth Factor WHO: World Health Organization
WI: Weighted Imaging
SUMMARY
Gliomas constitute 36% of all primary brain tumors and 81% of all primary malignant brain tumors. The overall prognosis in patients with gliomas depends mainly on the location and histologic grade of the tumor.
The World Health Organization classification of gliomas is the primary basis for guiding therapy and assessing overall prognosis in gliomas. This classification system, based on histologic features, often falls short of predicting therapeutic response of individual tumors within the same histologic grade. Yet, it still remains the grading method for both research and clinical prospects.
Unlike any other organ the brain has multiple protective layers such as the skull that ensure a homeostatic environment. The resulting reduced access to the brain and the absence of plasmatic brain tumor markers bring neuroimaging in a central position in diagnosis and management of brain tumors. Moreover, neuroimaging has evolved from a purely morphologic investigation into a diagnostic tool that allows characterization of particular physical alterations within brain tissue. Understanding the relationship between the physical characteristics of tumor tissue, studied by MR imaging, and biological characteristics of the tumor is therefore important for the appropriate integration of neuroimaging in brain tumor management. The general objective of this work is to define the relationship between physiologybased MR imaging and biological features of glial tumors. Diffusion and perfusionweighted imaging, physiologybased MR techniques provide the data based on physical characteristics of the tissues. Diffusionweighted imaging (DWI) allows the measurement of water molecules diffusivity within the brain tissue by means of apparent diffusion coefficient (ADC) measurements. Perfusionweighted imaging (PWI) is based on changes of MR signal during the passage of contrast material through the intravascular space and allows hemodynamic measurements such as those of cerebral blood volume (CBV) within the brain tissue.
Highgrade diffuse gliomas are currently differentiated from lowgrade diffuse gliomas by increased cellularity with nuclear atypia, mitotic activity, endothelial proliferation and necrosis. Components of the extracellular matrix and angiogenesis constitute some other features of gliomas, which have established links with oncogenic processes that influence the proliferative and infiltrative potentials of these tumors. We have specifically targeted these features in our comparative studies with the working hypothesis that physiologybased MR measurements, obtained in vivo, might provide information that is pertinent in terms of tumor malignancy.
We chose to approach the biology of brain tumors in two ways: in vivo, by means of metabolic imaging techniques such as positron emission tomography (PET) and ex vivo, by means of histological and immunohistochemical analyses of tumor specimens.
Many studies have investigated the relation between ADC values and cellularity in gliomas. The underlining theory is based on the premise that water diffusivity within the
extracellular compartment is inversely related to the content and attenuation of the constituents of the intracellular space. Therefore when cellularity increases, intracellular space volume increases with a relative reduction of the extracellular space, leading to restricted diffusion of water molecules. However other factors may affect the value of ADC in gliomas such as the extracellular matrix which contains various amounts of hydrophilic macromolecules susceptible to change water molecules diffusivity. Hyaluronic acid is one highly hydrophilic component of the extracellular matrix of gliomas and its level of expression changes significantly during the progression to anaplasia in gliomas. Our hypothesis was that hyaluronan may influence ADC values in high and low grade gliomas.
We demonstrated a positive correlation between ADC values and the immunohistochemical level of hyaluronan in glial tumors.
Previous studies have confirmed the utility of positron emission tomography using C
11 Methionine (PETMET) as a prognostic tool in patients with gliomas. Higher MET uptake is associated with greater proliferative potential and a higher level of malignancy in gliomas.
The increased aminoacid uptake in gliomas seems to reflect increased transport mediated by aminoacid carriers located in the endothelial cell membrane. Our hypothesis was that CBV measurements, index of tumor vascularity, may be related to tumor aminoacid metabolism.
We found a positive correlation between maximum CBV values and maximum MET uptake values in gliomas.
A limitation to these preliminary studies was lack of direct correlation between MR
based measurements and histologic and metabolic data. Indeed, glial tumors are known for their remarkable tissue heterogeneity across different grades, within the same grade, and even within a single given tumor. Therefore we used image coregistration and stereotactic biopsies to further assess the relationship between MRbased imaging data and both metabolic and histologic analysis.
The second part of our studies was based on measurements at the exact same localization on both MR and PET images where biopsy specimens were performed. We found a local relationship between CBV and MET uptake values. Both measurements correlated with mitotic activity and endothelial proliferation; two features of tumor aggressiveness.
In order to quantify tumor cellularity and tumor angiogenesis, we respectively measured cell density and vessel density using immunohistochemical markers to identify vessels. We found a regional relationship between CBV and cell density, as well as vessel density in gliomas whereas no correlation was found regionally between ADC and cell density.
We concluded that CBV measurements may be used locally as indices of angiogenesis and cellularity in gliomas; whereas local ADC measurements are more variable and may not be used as a marker of cellularity in heterogeneous tumors such as gliomas.
1. INTRODUCTION
1.1. Epidemiology of gliomas
The annual incidence of primary malignant brain tumors is 3.7 per 100.000 for men and 2.6 per 100.000 for women. Rates are higher in more developed countries than in less developed countries (Bondy ML et al., 2008). Gliomas of astrocytic, oligodendroglial and ependymal origin account for 36% of all primary brain tumors and 81% of all primary malignant brain tumors based on CBTRUS (Central Brain Tumor Registry of the United States, 2008; Figure1). (Website: http://www.cbtrus.org/reports//20072008/2007 report.pdf).
Figure 1: Distribution of all primary brain and CNS gliomas by histology subtypes in the Unite States, study based on 26,630 subjects (CBTRUS 20002004).
Both astrocytoma and glioblastoma peak in incidence at age 65 to 74 years, and oligodendroglioma at age 35 to 44 years (Wrensch M et al., 2002). It seems likely that the duration of exposure to unknown oncogenic factors and poorer immune surveillance with advancing age may account for those tumor types that increase in incidence with age (Wrensch M et al., 2002). The main prognostic factors are patient age, Karnofsky performance status at diagnosis, tumor location, extent of surgical resection, histological type and malignancy grade (Bondy ML et al., 2008). With the exception of pilocytic astrocytoma, the survival of glioma patients is still poor. The poor prognosis of diffuse gliomas is largely related to their high capacity of invasion (Giese A et al., 2003). Patients with glioblastoma (Grade IV) have the poorest survival in all age groups (5year survival rate of 3.3 %), and within any histologic type. Older patients have poorer survival than younger patients (Bondy
ML et al., 2008; CBTRUS, 2008). Highdose ionizing radiation and rare genetic syndromes are the only well established risk factors (Melean G et al., 2004; Sadetzki S et al., 2005).
1.2. Histopathologic classification and grading
The World Health Organization (WHO) classifies gliomas based on main histologic cell differentiation. Classification include: astrocytomas (6070%), oligodendrogliomas (530%) and ependymomas (less than 10%) (Louis DN et al., 2007). For each histological type, a grade of malignancy is also attributed. Table 1 summarizes the main histological types of gliomas and the corresponding grade.
Table 1: WHO classification of tumors of neuroepithelial tissue of the nervous system (Louis DN et al., 2007)
Gliomas Grade I Grade II Grade III Grade IV Astrocytic Tumours
Subependymal giant cell astrocytoma
*
Pilocytic Astrocytoma *
Pilomyxoid Astrocytoma *
Diffuse Astrocytoma *
Pleomorphic
xanthoastrocytoma *
Anaplastic Astrocytoma *
Glioblastoma *
Giant cell Glioblastoma *
Gliosarcoma *
Oligodendroglial Tumours
Oligodendroglioma *
Anaplastic Oligodendroglioma *
Oligoastrocytic Tumours
Oligoastrocytoma *
Anaplastic Oligoastrocytoma *
Ependymal Tumours
Subependymoma *
Myxopapillary Ependymoma *
Ependymoma *
Anaplastic Ependymoma *
Histologic grading of gliomas is a means of predicting the biological behavior of a tumor.
Tumor histological type and malignancy grade influence the choice of therapy (adjuvant radiation, chemotherapy). It is also one criteria used to predict a response to therapy and
outcome. The principal morphological features of malignancy grading are presence of high cellularity, cytological atypia, mitotic activity, endothelial proliferation and necrosis. These features allow grading gliomas from grade II to grade IV. Table 2 summarizes the application of these morphological malignant features in grading diffuse astrocytic tumors. Grade I tumors or pilocytic astrocytomas correspond to lesions with low proliferative potential and the possibility of cure following total surgical resection. The above malignancy grading criteria are not applicable to Grade I astrocytomas (Giannini C and Scheithauer BW, 1997; Louis DN et al., 2007). Grade II astrocytomas are infiltrative and despite low proliferative activity often recur. Infiltrative astrocytomas with only cellular atypia are considered grade II (fibrillary or diffuse astrocytoma). Cellular atypia is defined as variation in nuclear shape or size with accompanying hyperchromasia. Tumors which also show mitotic activity and sometimes endothelial proliferation are considered grade III (anaplastic astrocytoma). Endothelial proliferation is defined as multilayering of endothelium. Mitoses should be unequivocal with no special recognition given to their morphology. The separation between grade II from grade III tumors is however more reliable by determination of MIB1 antibody to the Ki67 antigen (Prayson RA, 2005). Tumors additionally showing endothelial proliferation and necrosis are grade IV (glioblastoma). Necrosis may be of any type, and perinecrotic palissading need not be present (Louis DN et al., 2007).
Table 2: Morphological features of malignancy grading in diffuse astrocytic tumors (Louis DN et al., 2007)
Astrocytomas Grade Cell Density
Cellular Atypia
Mitoses Endothelial Proliferation
Necrosis
Pilocytic Astrocytomas
I Not Applicable
Diffuse Astrocytomas
II + + rare
Anaplasitic
Astrocytomas III ++ ++ + / +
Glioblastomas IV +++ +++ ++ + +
1.2.1. Astrocytic Tumors
There are two main groups of astrocytic tumors. The first group consists of non infiltrating tumors including mainly pilocytic astrocytoma and subependymal giant cell astrocytomas (WHO Grade I). The second group consists of diffusely infiltrating tumors including diffuse astrocytoma (WHO Grade II), anaplastic astrocytoma (WHO Grade III) and glioblastoma (WHO Grade IV). WHO grade I astrocytomas develop more frequently in pediatric population and are usually resected completely. These tumors have a good prognosis. As the surgical resection is rarely total in the second group, the prognosis is poor with eventually a fatal outcome (Louis DN et al., 2007).
1.2.1.1. Pilocytic astrocytoma (WHO Grade I)
These are relatively circumscribed, slowly growing, often cystic astrocytomas occurring in children and young adults (Figure 2). They constitute the most common glioma in children, in whom the majority (67%) arises in the cerebellum (Ohgaki H and Kleihues P, J Neuropathol Exp Neurol. 2005). These tumors are of low to moderate cellularity and rarely mitosis, pleomorphic nuclei, endothelial proliferation, infarctlike necrosis and infiltration of leptomeninges are not signs of malignancy. Pilocytic astrocytomas are highly vascularized.
Long survival times are the rule in these tumors and they may stabilize or even spontaneously regress during their evolution (Gunny R et al., 2005). Supratentorial location is associated with less favorable prognosis (Kidd EA et al., 2006).
A B C D
E F
A B C D
E F
A B C D
E F
Figure 2: Pilocytic astrocytoma. Morphological MR imaging and histopathological illustration: T1WI (A), T2WI (B), FLAIR (C) and T1WI with contrast (D) show a cystic basal ganglia mass with a solid component which enhances intensely (arrow). Corresponding histological slices with Hematoxylin and Eosine coloration (E) and GFAP (F) (200x) show the morphological characteristics of a pilocytic astrocytoma.
1.2.1.2. Diffuse astrocytoma (WHO Grade II)
These tumors typically affect young adults with a peak incidence between ages 30 and 40, and most commonly develop supratentorially. The tumor is composed of well differentiated fibrillary or gemistocytic neoplastic astrocytes with a cellularity which is moderately increased compare to normal brain (Figure 3). Nuclear atypia is a typical feature whereas mitotic activity is generally absent. The mean survival time after surgical resection ranges between 6 to 8 years, with marked individual variations (Louis DN et al., 2007). Young age, total tumor resection and presentation with epilepsy as the single symptom are associated with more favorable prognosis. (Okamoto Y et al., 2004, Peraud A et al., 1998; van Veelen ML et al., 1998) Large tumor size and presentation with neurological deficit are associated with a worse prognosis (Karim AB et al., 1996; Danks RA et al., 1995).
A B C
D
A B C
D
Figure 3: Diffuse astrocytoma. Morphological MR imaging and histopathological illustration: T1WI (A), FLAIR (B) and T1WI with contrast (C) show a left parietal mass which show no enhancement (arrows).
Corresponding histological slice with Hematoxylin and Eosine coloration (D) (200x) show the morphological characteristics of a diffuse astrocytoma.
1.2.1.3.Anaplastic astrocytoma (WHO Grade III)
These tumors arise from diffuse astrocytoma with a tendency of progression to glioblastoma in a mean time interval of approximately 2 years (Ohgaki H et al., 2004). The mean age at diagnosis is between 45 and 51. They affect adults and are preferentially located
in the cerebral hemispheres. They are histologically characterized by nuclear atypia, increased cellularity and significant proliferative activity (Figure 4). Endothelial proliferation may be present in a few cases. As in low grade astrocytomas increasing age is a negative prognostic factor.
A B C D
E F
A B C D
E F
Figure 4: Anaplastic astrocytoma. Morphological MR imaging and histopathological illustration: T1WI (A), T2WI (B), FLAIR (C) and T1WI with contrast (D) show a left frontal mass which show no enhancement (arrows). Corresponding histological slices with Hematoxylin and Eosine coloration (E) and GFAP (F) (200x) show the morphological characteristics of an anaplastic astrocytoma.
1.2.1.4.Glioblastoma (WHO Grade IV)
They constitute the most frequent primary brain tumor and the most malignant with predominant astrocytic differentiation. They preferentially affect adults with a peak incidence between 45 and 75 years of age. Frontotemporal location is typical and tumor infiltration often extends into the adjacent cortex and through the corpus callosum into the contralateral hemisphere. Less than half of the patients survive more than a year despite progress in radio/chemotherapy. Older age represents the most significant adverse prognostic factor (Ohgaki H et al., 2004). Primary glioblastomas (> 90 %) develop very rapidly with a short clinical history (less than 3 months), generally without clinical or histopathological evidence of a preexisting, precursor lesion. Whereas secondary glioblastomas (< 10 %) develop through progression from diffuse astrocytoma or anaplastic astrocytoma (Ohgaki H et al., 2004). Glioblastoma histopathological features include nuclear atypia, cellular
pleomorphism, mitotic activity, endothelial proliferation and necrosis (Figure 5). The regional heterogeneity of glioblastoma is remarkable and therefore the histopathological diagnosis on specimens obtained by stereotaxic biopsies may be challenging (Burger PC et al., 1989). The presence of endothelial proliferation is one of the histopathological hallmarks of glioblastoma.
It typically consists of multilayered, mitotically active endothelial cells together with smooth muscle cells/pericytes. Tumor necrosis is a fundamental feature of glioblastoma, and its presence is one of the strongest predictors of aggressive clinical behavior in diffuse astrocytomas (Burger PC et al., 1987; Homma et al., 2006). There are two forms of necrosis, one which can be seen by neuroimaging as a nonenhancing core representing nonviable tumor tissue, where microscopically necrotic glioma cells can vaguely be identified as well as faded images of large, dilated necrotic tumor vessels. The second form of necrosis can be only noted microscopically and consists of multiple, small, irregularly shaped bandlike foci, surrounded by radially oriented, densely packed, small fusiform glioma cells in a pseudo
palissading pattern. Despite the advances in surgery, radiotherapy and chemotherapy, the overall survival of patients with glioblastoma remains poor. Less than 20 % of patients survive more than one year (Ohgaki H et al., 2004). Age is the most significant prognostic factor with younger patients having a significantly better prognosis (Burger PC et al., 1989).
A B C D
E F G
A B C D
E F G
A B C D
E F G
Figure 5: Glioblastoma. Morphological MR imaging and histopathological illustration: T1WI (A), FLAIR (B), T2WI (C) and T1WI with contrast (D) show a left occipital mass which enhances intensely and heterogeneously (arrows). Corresponding histological slices with Hematoxylin and Eosine coloration (E), GFAP (F) and Ki67 (G) (40x, 200x and 200x) show the morphological characteristics of a Glioblastoma.
1.2.2. Oligodendroglial Tumors
This group of tumors consists of diffusely infiltrating tumors composed predominantly of oligodendroglial cells. They include low grade oligodendrogliomas (WHO grade II) and anaplastic oligodendrogliomas (WHO grade III). Morphological features indicating anaplasia include increased cellularity, marked cytologic atypia, high mitotic activity, vascular proliferation and necrosis with or without pseudopalissading. Anaplastic oligodendrogliomas have a relatively better prognosis than anaplastic astrocytomas because of their high sensitivity to chemotherapy. Oligoastrocytomas correspond to tumors showing a mixture of tumor cells resembling those found in oligodendrogliomas and diffuse astrocytomas. The differential diagnosis of anaplastic oligoastrocytomas versus glioblastoma may be problematic (Louis DN et al., 2007).
1.2.2.1. Oligodendroglioma (WHO Grade II)
These tumors correspond to diffusely infiltrating well differentiated glioma composed of neoplastic cells morphologically resembling oligodendrocytes. They typically are located in the hemispheres and arise in adults with a peak incidence between 40 and 45 years of age. They are moderately cellular, may contain microcalcifications and show a dense network of branching capillaries (Figure 6). They typically grow slowly with relatively long survival times; a 10year survival rate of 51 % (Ohgaki H and Kleihues P J Neuropathol Exp Neurol., 2005). Younger age, frontal location, high postoperative Karnofsky score, and macroscopically complete surgical resection are associated with a more favorable outcome (Shaw EG et al., 1992; Kros JM et al., 1994; Schiffer D et al., 1997).
A B C D
E F G
A B C D
E F G
Figure 6: Low grade oligodendroglioma. Morphological MR imaging and histopathological illustration:
T1WI (A), T2WI (B), FLAIR (C), and T1WI with contrast (D) show a right insular mass with no enhancement (arrows). Corresponding histological slices with Hematoxylin and Eosine coloration (E), GFAP (F) and Ki67 (G) (x40) show the morphological characteristics of low grade oligodendroglioma.
1.2.2.2. Anaplastic oligodendroglioma (WHO Grade III)
These tumors correspond to oligodendrogliomas which present anaplastic features such as high cellularity, marked cytological atypia, high mitotic activity, endothelial proliferation and necrosis (Figure 7). The peak incidence is between 45 and 50 years of age.
They develop by progression from a preexisting WHO grade II oligodendrogliomas. The mean time progression from grade II oligodendroglioma to secondary anaplastic oligodendroglioma is approximately 67 years (Lebrun et al., 2004; Ohgaki H and Kleihues P J Neuropathol Exp Neurol. 2005). Younger patients, those with better performance status and those receiving more extensive resections have a better prognosis (Cairncross G et al., 2006; Shaw EG, 1992).
A B C D
E F G
A B C D
E F G
Figure 7: Anaplastic oligodendroglioma. Morphological MR imaging and histopathological illustration:
T1WI (A), T2WI (B), FLAIR (C), and T1WI with contrast (D) show a left parietal mass wwhich moderately enhances (arrows). Corresponding histological slices with Hematoxylin and Eosine coloration (E), GFAP (F) and Ki67 (G) (x400, x200, x200) show the morphological characteristics of anaplastic oligodendroglioma.
1.2.2.3.Oligoastrocytoma (WHO Grade II)
These tumors are composed of a mixture of two distinct neoplastic cell types morphologically resembling the tumor cells in oligodendroglioma and diffuse astrocytoma.
The median age of patients with oligoastrocytomas ranges between 35 and 45 years. They arise preferentially in the cerebral hemispheres. These are moderately cellularized neoplasms with no or low mitotic activity (Figure 8). Microcalcification and microcystic degeneration may be present. A median survival time of 6.3 to 6.6 years has been reported (Shaw EG et al.,
1992; Okamoto Y et al., 2004). Younger age at operation, total tumor resection and post operative radiation therapy are associated with a better prognosis (Shaw EG et al., 1992).
A B C D
E F G
A B C D
E F G
Figure 8: Low grade oligoastrocytoma. Morphological MR imaging and histopathological illustration: T1
WI (A), T2WI (B), FLAIR (C), and T1WI with contrast (D) show a right frontal mass which moderately enhances (arrows). Corresponding histological slices with Hematoxylin and Eosine coloration (E), GFAP (F) and Ki67 (G) (x40, x200, x200 respectively) show the morphological characteristics of low grade oligoastrocytoma.
1.2.2.4.Anaplastic oligoastrocytoma (WHO Grade III)
These tumors correspond to oligoastrocytomas with features of anaplasia such as increased cellularity, nuclear atypia, increased mitotic activity and endothelial proliferation.
Those with necrosis should be classified as glioblastoma with oligodendroglial component.
The mean age at diagnosis is 44. These tumors are also predominantly hemispheric. The prognosis of patients with anaplastic oligoastrocytomas is better than for patients with glioblastoma. A median survival time of 2.8 years has been reported (Shaw EG et al., 1992).
1.2.3. Ependymal Tumors
Ependymomas arise from ependymal cells covering the ventricles. Infratentorial ependymomas predominate in children whereas supratentorial ependymomas affect pediatric as well as adult patients (Louis DN et al., 2007). Ependymal tumors include myxopapillary ependymomas and subependymomas (WHO Grade I), ependymomas (WHO Grade II) which are the most frequent, and anaplastic ependymomas (WHO Grade III). The prognostic value
of the grade is controversial. The only prognostic factors currently admitted are patient age and tumor location. Complete surgical resection is the treatment of choice in these tumors (Louis DN et al., 2007).
1.2.3.1. Ependymoma (WHO Grade II)
These tumors are generally slowly growing tumors of children and young adults. In the spinal cord, they constitute the most common neuroepithelial neoplasms in adults. The most common pattern is a welldelineated, moderately cellularized glioma with rare or absent mitoses. There are four histopathological variants: cellular ependymoma, papillary ependymoma, clear cell ependymoma and tanycytic ependymoma. The most important prognostic factor is the tumor site. Spinal and supratentorial ependymomas are associated with better survival rates (Ernestus RI et al., 1996). Cerebrospinal dissemination indicates a poor prognosis (Louis DN et al., 2007). Age below 3 years and incomplete tumor resection are both indicators of a poor outcome (Horn B et al., 1999).
1.2.3.2. Anaplastic ependymoma (WHO Grade III)
These tumors are characterized by an ependymal differentiation with accelerated growth and unfavorable clinical outcome. Histologically they present with high mitotic activity, increased cellularity, microvascular proliferation and pseudopalissading necrosis. Anaplastic changes are by far more frequent in childhood intracranial ependymomas than in those of the spinal cord. Age below 3 years, anaplastic histological features, incomplete tumor resection and evidence for CSF metastases have been all proposed as indicators of an adverse outcome in children (Horn B et al., 1999; Jaing TH et al., 2004).
1.2.4. Other tumors of glial origin
Gangliogliomas consist of tumors containing both astrocytic tumor cells and atypic ganglion cells. Glial tumors of uncertain origin include three neoplasms: astroblastoma, gliomatosis cerebri and choroid glioma of the 3rd ventricle (Louis DN et al., 2007).
1.3. Biology of gliomas
Gliomas frequently demonstrate a complex cellular heterogeneity. This heterogeneity is reflected in their growth and dissemination within the neuroaxis. It is this cellular variation between the main types of gliomas, as well as within individual tumors, that explains why the clinical behavior within the same grade and response to treatment cannot always be predicted.
Immunohistochemical analyses (including indices of tumor proliferation) and molecular genetic changes have been investigated to define objective prognostic criteria. In addition to the histological features used in grading, various cellular and extracellular features of gliomas have established links with oncogenic processes that influence the proliferative and infiltrative potentials of these tumors. Extracellular matrix components as well as tumor angiogenesis constitute examples of such features. Tumor hypermetabolism in terms of glucose or amino acid uptake has also been related to a more aggressive potential of gliomas within the same histological grade.
1.3.1. Glioma protein expression
Gliomas in general express glial fibrillary acid protein (GFAP) which is a cytoskeletal protein (Kleihues P et al., 1987). In astrocytomas, decreasing GFAP expression usually accompanies decreasing differentiation (Tascos NA et al., 1982). The expression of S100 protein in astrocytomas is very similar to that of GFAP. As for GFAP, a low S100 protein content correlates with advanced dedifferentiation (Jacque CM et al., 1979; Louis DN et al.
2007). Vimentin shows a pattern of immunoreactivity similar to GFAP. Its presence in astrocytomas indicates a lower degree of differentiation and there is a tendency for vimentin to be expressed in highgrade astrocytomas (Herpers MJ et al., 1986). Vimentin is infrequently expressed in lowgrade oligodendroglioma, but more often found in anaplastic oligodendroglioma (Dehghani et al., 1998). Oligodendroglial tumors with vimentin expression have a poorer prognosis than those lacking vimentin expression (Dehghani et al., 1998).
1.3.2. Genetic alterations
Diffuse glial tumors have the tendency to progress towards a more malignant phenotype and acquire histopathological and biological characteristics of the glioblastoma. It has been demonstrated that the astrocytoma progression to secondary glioblastomas reflects a sequential accumulation of genetic alterations (Furnari FB et al., 1995; Kleihues P et al.,
1994; Ohgaki H et al., 1995). The identification of these genetic alterations is essential for our understanding of the evolution of gliomas and their biological characteristics. Among the genetic alterations classically described, p53 mutation is the genetic hallmark of lowgrade diffuse astrocytomas (60% in fibrillary astrocytomas and 80% in gemistocytic astrocytomas), anaplastic astrocytoma (70%) and secondary glioblastomas (65%). Yet, oligodendroglial tumors are characteristically associated with codeletion of 1p and 19q (6080%) (Okamoto Y et al., 2004). Loss of 1p/19q is associated with longer patient survival in low grade oligodendrogliomas and with positive response to chemotherapy and radiation therapy in anaplastic oligodendrogliomas (Felsberg J et al., 2004; Cairncross JG et al., 1998). There is an amplification of the gene encoding for epidermal growth factor receptor (EGFR) particularly in primary glioblastomas (36%) and less often in secondary glioblastomas (8%) (Ohgaki H et al., 2004). EGFR amplification is associated with poorer survival in adults with glioblastoma (Smith JS et al., 2001). Methylation of the O6methylguanine methyltransferase (MGMT) promoter in glioblastoma is associated with improved outcomes, especially among patients treated with alkylating agents (Hegi ME et al., 2005; Stupp R et al, 2005). Isocitrate dehydrogenase 1(IDH1) mutations constitute a highly selective molecular marker of secondary glioblastomas and are a strong predictor of a more favorable prognosis in diffuse astrocytomas and oligodendroglial tumors (Parsons DW et al., 2008; Nobusawa S et al., 2009; Sanson M et al., 2009).
1.3.3. Cell proliferation
As previously mentioned, increased cellularity is one of the histological features described in anaplastic gliomas. A more quantitative approach for appreciation of cellularity is to measure cell density. Previous reports showed that proliferative activity and cell density are highly correlated in astrocytic tumors. A high level of cell density in high grade astrocytomas has been associated with shorter survival time (Kiss R et al., 1997).
Tumor cell proliferation represents the convergence of many cellular factors. These include cell cycle time, growth fraction, population doubling time and cell loss. Basing the proliferative potential of brain tumors on identification of mitotic figures is subject to inter
observer variability. Moreover, identification of mitotic figures is an assessment of only a small portion of the proliferative phase of cell cycle (Prayson RA, 2005).
Immunohistochemical techniques such as bromodeoxyuridine labeling index (BrdU LI), MIB
1 antibody to the Ki67 antigen (MIB1), proliferating cell nuclear antigen (PCNA) and silver
nucleolar organizing regions (AgNOR) have been developed to assess the proliferative potential of brain gliomas (QuiñonesHinojosa A et al., 2005).
MIB1 is a sensitive marker of the growth fraction that recognizes proliferation
associated nuclear proteins that are present during G1, S, and G2/M phases, but absent in G0 (Gerdes J et al., 1984). In a prospective study comparing predictive power for survival for gliomas, MIB1 was found to be superior to BrdU LI and PCNA (McKeever PE et al., 1998).
MIB1 was found to be an independent and statistically significant prognostic factor of survival in pilocytic and diffuse astrocytomas (Bowers DC et al., 2003; Jaros E et al., 1992) and also in ependymomas (Wolfsberger S et al., 2004). In patients with glioblastoma, MIB1 was not found to be an independent prognostic factor (Moskowitz SI et al., 2006). PCNA is a DNA polymerase delta accessory protein that is associated with DNA replication and is thus reflective of cells in the S phase. Although higher PCNA labeling index are associated with malignancy, the broad range of PCNA values within tumor subgroups can be so marked that low grade gliomas are indistinguishable from highgrade (Louis DN et al., 1991).
In addition to MIB1 determination, cell density has the advantage to integrate cell loss and cell proliferation dimensions, allowing a better evaluation of the net growth of a given tumor (Gordower L et al., 1998). In fact, tumor growth requires a coincidental acquisition of genomic lesions triggering deregulated cell proliferation together with those suppressing the concomitant cell death (Evan G, 1997).
1.3.4. Cell migration
Tumor cell migration refers to multiple mechanisms such as adhesion, motility and invasion (Friedl P and Bröcker EB, 2000). Invasion involves the ability of tumor cells to translocate through extracellular matrix (ECM) barriers by the release of proteolytic enzymes (Tysnes BB et al., 2001). The ECM of the brain represents 20% of the normal adult brain volume (Nicholson C and Sykova E, 1998). Two main classes of extracellular macromolecules make up the ECM. The first group of macromolecules consists of fibrous proteins (collagen and elastin) and fibrous adhesive proteins (fibronectin and laminin). ECM of the normal brain contains less collagen, fibronectine and laminin compared to other organs in the body and they are localized to the perivascular space (Gladson CL, 1999). The second group of macromolecules consists of polysaccharide chains called glycosaminoglycans, which are linked to proteins in the form of proteoglycans. Hyaluronic acid or hyaluronan (HA) differs from other proteoglycans in that it is not attached to a protein. Hyaluronan is a highly
hydrophilic glycosaminoglycan of the ECM with an enormous amount of solvent attached (Bertolotto A et al., 1986; Comper WD and Laurent TC, 1978; Laurent TC and Fraser JRE, 1992). The structure of the ECM of the brain is organized around hyaluronan molecules.
Hyaluronan is attached to other proteoglycans of the lectican family and forms superstructures. These superstructures are connected by different glycoproteins such as tenascins (Rauch U, 2004).
During glioma growth and migration, the tumor cells closely interact with the ECM (Bellail AC et al., 2004). Changes in volume, shape and composition of the extracellular space influence the biological behavior of brain tumors. In astrocytomas, the extracellular volume shows a dramatic increase compared to normal brain cortex (Zamecnik J, 2005). A positive correlation has been reported between the size of the extracellular volume fraction and the malignancy grade of the tumors (Vargova L et al., 2003). In high grade gliomas there is an overproduction of certain ECM glycoproteins which stabilize the extracellular volume and serve as a substrate for adhesion and migration of the tumor cells. In gliomas, glycoproteins such as tenascinC, vitronectin, osteopontin and Secreted Protein Acidic and Rich in Cysteine (SPARC) are frequently overexpressed. Malignant gliomas contain higher amounts of hyaluronan than adult brain tissue (Delpech B et al., 1993, Gladson CL, 1999).
The expression of the HAreceptors, CD44 and Receptor for HA Mediated Mobility (RHAMM), is virtually ubiquitous among glioma cell lines. There is a gradient of expression amongst gliomas; high grade gliomas expressing more CD44 and RHAMM than lower grade lesions (Akiyama Y et al., 2001). Integrins are a large family of cell membrane receptors that mediate interaction between the cell and the components of the extracellular matrix.
Vitronectin and tenascinC are among ECM ligands recognized by the integrins and their interaction stimulates cell migration (D’Abaco GM and Kaye AH, 2007).
1.3.5. Angiogenesis
Angiogenesis consists in the development of new blood vessels from preexisting ones and results from an altered balance of proangiogenic and antiangiogenic factors (Foscher I et al., 2005). The association between tumor growth and angiogenesis is well known (Folkman J, 1971). High grade gliomas and glioblastomas in particular are among the most vascularized tumors in humans. Tumor angiogenesis occurs through several mechanisms. Hypoxia is considered a major driving force of glial tumor angiogenesis and leads to intracellular stabilization of the hypoxiainducible factor 1a (HIF1a) (Acker T and Plate KH, 2004). HIF
1a accumulation leads to activation of many hypoxiaregulated genes such as vascular
endothelial growth factor (VEGF) which appears to be the most important mediator of gliomaassociated vascular dysfunctions. VEGF is produced by perinecrotic palissading cells as a consequence of cellular hypoxia. VEGF induces tumor angiogenesis and increased vascular permeability (edema) (Acker T and Plate KH, 2004). The expression of VEGF and VEGF receptors (VEGFR1 and VEGFR2) correlates with the grade in diffuse astrocytomas (Abdulrauf SI et al., 1998; Zagzag D and Capo V, 2002). Other angiogenic factors are angiopoietins, in particular, Ang1 and Ang2 have been implicated in glioma angiogenesis (Zagzag D et al., 1999). During angiogenesis, endothelial cells must adhere to the ECM in order to proliferate, migrate and maintain an appropriate cell shape. Attachment and adhesion of endothelial cells to the ECM is regulated by cell surface adhesion receptors such as integrins (Wang D et al., 2005). ECM components such as fibronectin and tenascinC have been detected in the ECM surrounding newly formed blood vessels in high grade astrocytomas and are recognized as proangiogenic molecules (Wang D et al., 2005).
Many authors have investigated the possible role of adding vessel quantification to routine histologic grading in gliomas. The protocol of Weidner et al. is the most frequently used in order to quantify tumor vascularity. Histochemical or immunohistochemical stains for endothelium or other vascular wall components (Factor VIIIrelated antigen, CD34, CD31) is performed according to standard methods. Qualitative identification of the area of highest vessel density is then performed by either manual or computerassisted counting (Weidner N et al., 1991; Folkerth RD, 2000). Other parameters such as vessel diameter, perimeter and area may also be evaluated. A significant negative correlation has been reported between increasing microvessel counts and diseasefree survival in supratentorial astrocytomas (Leon SP et al., 1996). Also microvessel density has been identified as independent prognostic marker of survival in fibrillary low grade astrocytomas (Abdulrauf SI et al., 1998).
1.3.6. Tumor metabolism
Positron emission tomography (PET) is a molecular imaging technique that uses radiolabeled molecules to image molecular interactions of biological processes in vivo. PET can provide valuable metabolic information and is used for characterizing brain tumors, evaluating response to therapy and prognosis. In brain tumors, 18FFluorodeoxyglucose (FDG) and 11CMethionine (MET) are the most widely used tracers indicating glucose uptake and amino acid uptake, respectively. Amino acid PET tracers are more sensitive than FDG in imaging recurrent tumors and in particular recurrent lowgrade tumors.
1.3.6.1. 18FFluorodeoxyglucose Positron Emission Tomography (FDGPET) FDGPET depicts the increased capacity for glucose transport in malignant glial cells. In 1982, Di Chiro et al. reported the successful use of FDGPET imaging in evaluating primary brain tumors and radiation necrosis (Patronas NJ et al., 1982; Di Chiro G et al., 1982). Low grade tumors reveal low levels of metabolism whereas high grade tumors appear hypermetabolic compared to normal brain tissue. However, low grade oligodendrogliomas and pilocytic astrocytomas may also appear hypermetabolic on FDGPET images (Kaschten B et al., 1998). Metabolic activity of tumor as shown by FDGPET is also a good indicator of the prognosis in patients with primary brain tumors (Alavi JB et al., 1988; De Witte O et al., 1996). Herholz et al. found that cell density correlated significantly with FDG uptake (Herholz K et al., 1993). However, due to the high rate of physiologic glucose metabolism in normal brain tissue, lowgrade tumors with low FDG uptake may be difficult to detect by FDGPET. Coregistration of FDGPET images with MR images in order to delineate area of interest greatly improves the performance of FDGPET (Wong TZ et al., 2004).
1.3.6.2. 11CMethionine Positron Emission Tomography (METPET)
METPET has the advantage of showing selective uptake in the brain tumor compared with normal brain tissue. Therefore in low grade gliomas METPET offers significantly better contrast than FDGPET relative to surrounding gray matter activity. The mechanisms and biological significance of tumor uptake of MET is not completely understood. Uptake of MET in gliomas is attributed to the activation of carriermediated transport at the level of the bloodbrain barrier. It does not directly reflect protein synthesis, but rather the activity of the transport system and incorporation in proteins (Bustany P et al.
1986; Derlon JM et al. 1989). There is a correlation between MET uptake and the histological grade in gliomas (Kaschten B et al. 1998). High MET uptake is associated with a poor survival time in WHO Grade II and WHO Grade III gliomas (De Witte O et al., 2001).
1.3.6.3. Other PET tracers
Because of the short halflife of 11C (20 minutes), 18Flabeled aromatic amino acid analogs have been developed for tumor imaging. Tumor uptake of O(218Ffluoroethyl)
tyrosine (FET) and 3,4dihydroxy618FfluoroLphenylalanine (FDOPA) has been reported to be similar to that of 11Cmethionine (Weber WA et al., 2000; Becherer A et al., 2003). The thymidine analog 3’deoxy3’18Ffluorothymidine (FLT) PET was developed as a non
invasive method to evaluate tumor cell proliferation (Shields AF et al., 1998). Uptake of 18F
FLT by tumors correlates with Ki67 (Chen W et al., 2005; Jacobs AH et al., 2005). 18F
Fluoromisonidazole is a nitroimidazole derivative that has been developed as a PET agent to image hypoxia (Rasey JS et al., 1996). 18FFluoromisonidazole uptake was found in high grade gliomas but not in low grade gliomas (Cher LM et al., 2006).
1.4. Neuroimaging of gliomas
Unlike any other organ in the body the brain has multiple protection layers such as the skull that ensure a homeostatic environment. The resulting reduced access to the brain and the absence of plasmatic brain tumor markers bring neuroimaging to the forefront for diagnosis, preoperative therapy planning as well as for evaluation of treatments. Unenhanced CT is usually the first line imaging technique used to evaluate patients with suspected brain tumors because it is widely available, fast and well tolerated. CT is very sensitive in detecting lesions associated with brain tumor such as hemorrhage, calcifications, mass effect and hydrocephalus. However MRI is considered the current imaging standard for brain tumor patients because of its higher anatomical resolution and higher sensitivity for lesion detection compared to CT. Moreover CT uses ionizing radiation and the iodinated contrast material used may cause serious allergic reactions. Gadolinium which is used as contrast agent in MRI is better tolerated and has a lower risk profile.
1.4.1. Morphologic MR brain imaging
Morphologic MR imaging is based on anatomic information provided by different MR sequences (T1weighted images (WI), T2WI, Fluidattenuated inversion recovery (FLAIR), T1WI with contrast). This information allows tumor detection and tumor localization with a high sensitivity and some degree of tumor characterization, based on the respective signal intensity on the images.
1.4.1.1. Tumor characterization and grading
Tumors are best detected on T2WI and FLAIR images and usually as areas of high signal intensity. On T1WI they appear with low signal intensity. Low grade gliomas are best depicted on FLAIR images and usually show minimal or no mass effect. Cystic components within or associated with the tumor can also be detected on T2WI and FLAIR images. However the signal intensity of a cyst on FLAIR images is dependent on its protein content and may be quite variable. Calcifications and hemorrhage are best detected on
gradientecho T2WI and may appear as areas of high signal intensity on noncontrast T1WI.
T1WI with contrast is one of the most important MR sequences for the characterization of brain tumors (Felix R et al., 1985). Most high grade gliomas show contrast enhancement due to destruction of the bloodbrain barrier (BBB); whereas lowgrade tumors usually show no or minimal enhancement. MR imaging also depicts signs of increased intracranial pressure and mass effect, as well as edema. High grade tumors usually appear as a heterogeneous mass which is hypointense on T1WI and hyperintense on T2WI and FLAIR images with various degrees of contrast enhancement and edema. Ringlike enhancement surrounding irregularly shaped foci of presumed necrosis is suggestive of glioblastoma. However anaplastic tumors can often present as nonenhancing lesions. Lastly, treatment induced changes such as radiation necrosis can be difficult to distinguish from recurrent tumor by morphologic MR imaging (Levivier M et al., 1996).
1.4.1.2. Stereotactic biopsy guidance and neuronavigation
The first stereotactic frame in humans was used in 1947 (Spiegel EA et al., 1947).
Preoperative images were used to target an instrument towards a precise location within the brain. In order to perform this task a frame was rigidly attached to the head of the patient, followed by image acquisition resulting in images with reference marks related to a frame
based coordinate system. Using the same coordinate system, an instrument could subsequently be guided towards any target depicted on the images. Frames were first adapted to be used with CT and then MRI (Leksell L et al., 1985). Stereotactic frames are still used in the treatment of movement disorders, pain disorders, epilepsy, cyst or abscess drainage and tumor biopsies. Neuronavigation is a technique used for localizing the position of an operative instrument and neurosurgical image guidance without the use of a head mounted frame. Most systems are based on the optical triangulation of infrared light sources fixed to the surgical instrument. Briefly, a navigation system is a threedimensional digitizer that correlates its measurements to a reference data set, such as the preoperatively acquired MRI image stack.
This correlation is achieved through a patienttoimage registration procedure resulting in a mathematical transformation matrix mapping each position in “world space” onto “image space”. Thus, throughout the remainder of the surgical procedure the position of the surgical instrument can be demonstrated on a computer screen, relative to MR images (Willems PW et al., 2006). The standard imaging method for neuronavigation is a high resolution isotropic T1WI sequence.