**4. The tissue around the tumor — Cell invasion and edema**

Beside resistance to chemo-and radiotherapy, the failure of a local control of GBM by therapies is due to the modalities of tumor cell invasion into the brain. Surgical resection cannot prevent recurrence because of the occurrence of infiltrating cells; recurrence usually starts from the resection margin. The target volume for radiotherapy, therefore, conventionally includes the tissue within 2 cm from the MRI border of the tumor. This is for sparing normal nervous tissue from irradiation damage, but also for including in it infiltrating cells. Nevertheless, 80% of tumors relapses within 2-cm margin around the enhanced region [58]; another adverse characteristic of infiltrating cells is that they do not proliferate [3,40], escaping thus detection and being less sensitive to treatments.

Tumor invasion is not uniform along its borders. It can be non-existent where the tumor sharply abuts against the normal nervous tissue (Figure 10A), as it may happen with the white matter, or it gradually progresses from the tumor outwards (Figure 10B). Typical is the invasion of the cortex from a tumor located in the white matter, even with the typical picture of perineuronal satellitosis (Figure 10C). The different invasion modalities have been codified [59] and a distinction between diffuse and local tumors has even been proposed [60], but it was not confirmed by the observation of substantially different outcomes.

**Figure 10.** A – Sharp tumor border. Ki-67/MIB.1, DAB, 100x; B – Invasion gradient toward the cortex. H&E, 100x; C – Perineuronal satellitosis. H&E, 400x; D – Infiltration along corpus callosum. H&E, 200x.

**4. The tissue around the tumor — Cell invasion and edema**

74 Tumors of the Central Nervous System – Primary and Secondary

with DAPI.

Beside resistance to chemo-and radiotherapy, the failure of a local control of GBM by therapies is due to the modalities of tumor cell invasion into the brain. Surgical resection cannot prevent recurrence because of the occurrence of infiltrating cells; recurrence usually starts from the resection margin. The target volume for radiotherapy, therefore, conventionally includes the tissue within 2 cm from the MRI border of the tumor. This is for sparing normal nervous tissue from irradiation damage, but also for including in it infiltrating cells. Nevertheless, 80% of

**Figure 9.** Immunofluorescence (IF) for stemness antigens in NS. A – Nestin, 200x; C – CD133, 200x; E – Musashi1, 200x. IF for differentiation antigens in AC. B – GFAP, 400x; D – GalC, 400x; F – β-III Tubulin, 400x. Nuclei are counterstained

> Letting aside the mechanisms of tumor cell migration and invasion of which there is today a good knowledge [61-63], some information about neuropathologic findings on peritumor tissue are relevant. First of all, it has been demonstrated that patients with absence of tumor cells in the adjacent normal nervous tissue had better survival than those with tumor cells [64]. This is not in contrast with the observation that the removal of edematous peritumor tissue

does not improve the outcome of operated patients [65]. In this regard, residual cells after surgery have been interpretated as distinct from the cells found in routinely resected GBM tissue, as if they would represent a distinct, malignant GBM subentity [66]. A second important point is how to recognize invading cells. Beside nuclear abnormalities, there are only the counts of cells, as will be said. Nestin expression [67] and mainly Isocitrate Dehydrogenase isoforms 1 and 2 (IDH1/2) mutations [68] have been proposed, but it must be taken into account that primary GBM cells are IDH1/2 wild-type.

Cell infiltration, as it is usually seen in histological sections of tumor surgically removed, can be very mild and not easily recognizable without cell counts or decidedly evident (Figure 11A,B). Its aspects largely depend on the various modalities of GBM spreading and three main possibilities are recognized [69]: the distant spreading through corpus callosum (Figure 10D), septum pellucidum, *etc.,* the sub-pial invasion (Figure 11C) and the invasion of the cortex from the white matter where the tumor is located (Figures 11C). Also sub-arachnoidal seeding is frequent [70,71], sometimes as small clusters of tumor cells, visible at naked eyes from which tumor cells go down along penetrating vessels to invade the cortex (Figure 11D). It is very important to remark that invading cells do not proliferate, as it has been demonstrated *in vitro* [72,73] and *in vivo* [3,41,74]. Two other cell types can be found in peritumor tissue: macrophages/microglia and reactive astrocytes, both in edematous and non-edematous tissue. The former, independently of their influence on immunoregulation and tumor growth [61], are abundant in both tumor and peritumor tissue [75]; it has been calculated that up to one third of cells in glioma biopsies are represented by macrophages [75,76] (Figure 12B,C). Incidentally, a positive or negative relationship between microglia/macrophages and TICs is today discussed [77]. The same can be said for the possibility that microglia can be exploited in tumor therapy. It remains today "in its infancy" [78] as it happens for the possibility to inhibit microglia/macrophages in order to prevent their promotion of tumor progression [79].

Reactive astrocytes can be sometimes confused with tumor cells, mainly because their phenotype changes over time until complete maturation (Figure 12A). There are analogies between glial reaction and physiological maturation of astrocytes during embryogenesis. In initial phases, the fine processes originate directly from the cell soma and then from the thick and long processes [80]. Nestin and Vimentin would be the main intermediate filaments of immature astrocytes, whereas GFAP of the mature ones [81,82].

may be included in the advancing tumor in which they progressively become no more recognizable from tumor cells. The question is whether they disappear suffocated by the high density of tumor cells or if they remain, unrecognizable from tumor cells, contributing to the

**Figure 11.** A – Mild infiltration. H&E, 200x; B – Strong infiltration. H&E, 200x; C – Sub-pial infiltration and growth. H&E, 100x; D – Infiltration along penetrating vessels from a tumor seeding in sub-arachnoidal space. PCNA, DAB, 100x.

Spatial Relationships of MR Imaging and Positron Emission Tomography with Phenotype, Genotype and...

http://dx.doi.org/10.5772/58391

77

Practically, two important points are that the tumor extends beyond the area of enhancement and that tumor cells can be found in peritumor edema [86]. In 20% of stereotactic biopsies tumor cells have been found in normal areas [87]. With the MRI era, detection of tumor infiltrating cells did not improve and it was shown that they can occur either in the T2 hyperintense areas or beyond them [88] or even in areas apparently normal in T1 or edematous

GBM spreads frequently along white matter tracts and their disruption can be detected by DTI. Observations have been made, but without any histological control. For example, infiltrated white matter shows a decrease of FA and an increase of ADC as when it is edematous. Displacement of white matter tracts with decreased FA can be recognized [89]. Many studies have been dedicated to FA reduction, but it did not appear to be sensitive enough to detect

pleomorphic aspect of gliomas, or if they are transformed into tumor cells [85].

in T2.

It is long debated whether infiltrated tissue can be recognized by MRI, not only when adjacent to tumor, but also at a distance. It has been observed, for example, that low-grade gliomas, which preferentially locate in the insula and the supplementary motor area, spread along distinct sub-cortical fasciculi [83]. Analyzing different peritumor areas with different MRI methods, it has been shown that FA and not apparent diffusion coefficient can be used to evaluate glioma cell invasion. An attempt to classify different peritumoral tissues by a voxelwise analytical solution using serial diffusion MRI has been made [84].

Peritumoral reactive gliosis has a particular importance because of three main characteristics: reactive astrocytes divide by mitosis as tumor cells; they progressively lose Nestin and they increase GFAP expression as during development, and they may exert regionally a series of metabolic and molecular influences [61]. The most important point is that reactive astrocytes Spatial Relationships of MR Imaging and Positron Emission Tomography with Phenotype, Genotype and... http://dx.doi.org/10.5772/58391 77

does not improve the outcome of operated patients [65]. In this regard, residual cells after surgery have been interpretated as distinct from the cells found in routinely resected GBM tissue, as if they would represent a distinct, malignant GBM subentity [66]. A second important point is how to recognize invading cells. Beside nuclear abnormalities, there are only the counts of cells, as will be said. Nestin expression [67] and mainly Isocitrate Dehydrogenase isoforms 1 and 2 (IDH1/2) mutations [68] have been proposed, but it must be taken into account that

Cell infiltration, as it is usually seen in histological sections of tumor surgically removed, can be very mild and not easily recognizable without cell counts or decidedly evident (Figure 11A,B). Its aspects largely depend on the various modalities of GBM spreading and three main possibilities are recognized [69]: the distant spreading through corpus callosum (Figure 10D), septum pellucidum, *etc.,* the sub-pial invasion (Figure 11C) and the invasion of the cortex from the white matter where the tumor is located (Figures 11C). Also sub-arachnoidal seeding is frequent [70,71], sometimes as small clusters of tumor cells, visible at naked eyes from which tumor cells go down along penetrating vessels to invade the cortex (Figure 11D). It is very important to remark that invading cells do not proliferate, as it has been demonstrated *in vitro* [72,73] and *in vivo* [3,41,74]. Two other cell types can be found in peritumor tissue: macrophages/microglia and reactive astrocytes, both in edematous and non-edematous tissue. The former, independently of their influence on immunoregulation and tumor growth [61], are abundant in both tumor and peritumor tissue [75]; it has been calculated that up to one third of cells in glioma biopsies are represented by macrophages [75,76] (Figure 12B,C). Incidentally, a positive or negative relationship between microglia/macrophages and TICs is today discussed [77]. The same can be said for the possibility that microglia can be exploited in tumor therapy. It remains today "in its infancy" [78] as it happens for the possibility to inhibit microglia/macrophages in order to prevent their promotion of tumor progression [79]. Reactive astrocytes can be sometimes confused with tumor cells, mainly because their phenotype changes over time until complete maturation (Figure 12A). There are analogies between glial reaction and physiological maturation of astrocytes during embryogenesis. In initial phases, the fine processes originate directly from the cell soma and then from the thick and long processes [80]. Nestin and Vimentin would be the main intermediate filaments of

It is long debated whether infiltrated tissue can be recognized by MRI, not only when adjacent to tumor, but also at a distance. It has been observed, for example, that low-grade gliomas, which preferentially locate in the insula and the supplementary motor area, spread along distinct sub-cortical fasciculi [83]. Analyzing different peritumor areas with different MRI methods, it has been shown that FA and not apparent diffusion coefficient can be used to evaluate glioma cell invasion. An attempt to classify different peritumoral tissues by a voxel-

Peritumoral reactive gliosis has a particular importance because of three main characteristics: reactive astrocytes divide by mitosis as tumor cells; they progressively lose Nestin and they increase GFAP expression as during development, and they may exert regionally a series of metabolic and molecular influences [61]. The most important point is that reactive astrocytes

primary GBM cells are IDH1/2 wild-type.

76 Tumors of the Central Nervous System – Primary and Secondary

immature astrocytes, whereas GFAP of the mature ones [81,82].

wise analytical solution using serial diffusion MRI has been made [84].

**Figure 11.** A – Mild infiltration. H&E, 200x; B – Strong infiltration. H&E, 200x; C – Sub-pial infiltration and growth. H&E, 100x; D – Infiltration along penetrating vessels from a tumor seeding in sub-arachnoidal space. PCNA, DAB, 100x.

may be included in the advancing tumor in which they progressively become no more recognizable from tumor cells. The question is whether they disappear suffocated by the high density of tumor cells or if they remain, unrecognizable from tumor cells, contributing to the pleomorphic aspect of gliomas, or if they are transformed into tumor cells [85].

Practically, two important points are that the tumor extends beyond the area of enhancement and that tumor cells can be found in peritumor edema [86]. In 20% of stereotactic biopsies tumor cells have been found in normal areas [87]. With the MRI era, detection of tumor infiltrating cells did not improve and it was shown that they can occur either in the T2 hyperintense areas or beyond them [88] or even in areas apparently normal in T1 or edematous in T2.

GBM spreads frequently along white matter tracts and their disruption can be detected by DTI. Observations have been made, but without any histological control. For example, infiltrated white matter shows a decrease of FA and an increase of ADC as when it is edematous. Displacement of white matter tracts with decreased FA can be recognized [89]. Many studies have been dedicated to FA reduction, but it did not appear to be sensitive enough to detect infiltration [90]. The problem has not yet been resolved and it is still under discussion, because new techniques have been proposed [91,92], even though ADC values, lower in the tumor than in peritumor tissue, were not interpreted by others as significant [93]. Nevertheless, DTI is going to be accepted in the evaluation of tumor margins and invasiveness [94].

These findings are in agreement with those indicating that tumor cells could be detected beyond the margin of the tumor by MRSI [97] and this has been substantiated by histopathol‐

Spatial Relationships of MR Imaging and Positron Emission Tomography with Phenotype, Genotype and...

http://dx.doi.org/10.5772/58391

79

**Figure 13.** Axial contrast enhanced T1-weighted image, T2-weighted image, ADC and rCBV maps showing a lesion in the left frontal lobe with heterogeneous signal and diffusion properties, peripheral and irregular contrast enhance‐

**Figure 14.** A – Aquaporin-4 in a peritumor area with astrocyte and vessel positive staining. DAB, 200x; B – Aspect of a

ment. Perfusion is increased in the peripheral portion of the lesion.

gemistocytic astrocytoma found in a T2 hyper-intense area. H&E, 200x.

ogy studies [88,96].

**Figure 12.** A – Reactive astrocytes at regular inter-distance. GFAP; B – Macrophages/microglia in the tumor. CD68; C – *id.* in peritumor area. CD68. All DAB, 200x.

Edema on T2-weighted imaging may have a high Cho/NAA ratio as in tumors [95] and this would indicate the occurrence of tumor cell infiltration [96] (Figure 13). It can be demonstrated by Aquaporin-4 antibody method (Figure 14A), but in the tissue this is not suitable for quantitative assessments. However, the real problem is how to detect mild infiltration, either alone or with edema. Some observations were supported by histological examination of peritumor edematous areas with or without cell infiltration. Three spectral patterns in perienhancing apparently edematous ROIs have been described: high Cho and abnormal Cho/NAA ratio in presumed tumor areas, normal Cho/NAA ratio in presumed edematous areas and Cho levels similar to normal, but with abnormal Cho/NAA ratio in tumor edema. In ROIs on peri-enhancing normal tissue, the patterns therefore could be: presumed infiltration with high Cho and abnormal Cho/NAA ratio and presumed normality with normal values. These findings are in agreement with those indicating that tumor cells could be detected beyond the margin of the tumor by MRSI [97] and this has been substantiated by histopathol‐ ogy studies [88,96].

infiltration [90]. The problem has not yet been resolved and it is still under discussion, because new techniques have been proposed [91,92], even though ADC values, lower in the tumor than in peritumor tissue, were not interpreted by others as significant [93]. Nevertheless, DTI is

**Figure 12.** A – Reactive astrocytes at regular inter-distance. GFAP; B – Macrophages/microglia in the tumor. CD68; C –

Edema on T2-weighted imaging may have a high Cho/NAA ratio as in tumors [95] and this would indicate the occurrence of tumor cell infiltration [96] (Figure 13). It can be demonstrated by Aquaporin-4 antibody method (Figure 14A), but in the tissue this is not suitable for quantitative assessments. However, the real problem is how to detect mild infiltration, either alone or with edema. Some observations were supported by histological examination of peritumor edematous areas with or without cell infiltration. Three spectral patterns in perienhancing apparently edematous ROIs have been described: high Cho and abnormal Cho/NAA ratio in presumed tumor areas, normal Cho/NAA ratio in presumed edematous areas and Cho levels similar to normal, but with abnormal Cho/NAA ratio in tumor edema. In ROIs on peri-enhancing normal tissue, the patterns therefore could be: presumed infiltration with high Cho and abnormal Cho/NAA ratio and presumed normality with normal values.

*id.* in peritumor area. CD68. All DAB, 200x.

going to be accepted in the evaluation of tumor margins and invasiveness [94].

78 Tumors of the Central Nervous System – Primary and Secondary

**Figure 13.** Axial contrast enhanced T1-weighted image, T2-weighted image, ADC and rCBV maps showing a lesion in the left frontal lobe with heterogeneous signal and diffusion properties, peripheral and irregular contrast enhance‐ ment. Perfusion is increased in the peripheral portion of the lesion.

**Figure 14.** A – Aquaporin-4 in a peritumor area with astrocyte and vessel positive staining. DAB, 200x; B – Aspect of a gemistocytic astrocytoma found in a T2 hyper-intense area. H&E, 200x.

The overlapping of tumor cell infiltration and edema remains a major problem in the brain adjacent to tumor (BAT), because of the difficulty of their distinction [98,99], even though somebody supports that white matter fibre tract infiltration can be recognized [97]. In exper‐ imental tumors transplanted into mice, it has been observed by superimposing immunohis‐ tochemistry to MRI that in edema districts around the tumor, reactive astrocytes and activated microglia increased Aquaporin-4 expression and invasive tumor cells coexist [100]. Aquapor‐ in-4 has been observed to correlate in peritumor tissue with edema and in the tumor with Hypoxia-Inducible Factor-1 (HIF-1), Vascular Endothelial Growth Factor (VEGF) and the grade of malignancy [101], whereas NAA seemed to be more suitable to detect low tumor infiltration in peritumor edema [102]. Of course, in the latter a damage to myelin sheaths takes place and it is detectable by MRSI [103].

**Figure 15.** Isolate tumor cells in a white matter bundle. PCNA, DAB, 400x.

considering that they can acquire a stem-like phenotype [113].

In the last decades, the aphorism is that the eradication of the tumor cannot be obtained by directing chemo-and radiotherapies to the entire tumor mass, composed of non-proliferating, differentiated and insensitive cells; on the contrary, such therapies would be successful if addressed to the cells responsible for growth, recurrence and resistance, *i.e.* GSCs. Therefore, the question is whether these cells can be *in vivo* detected by neuro-imaging and where are they located or generated in the tumor. To answer this question, a short discussion on the origin

Spatial Relationships of MR Imaging and Positron Emission Tomography with Phenotype, Genotype and...

http://dx.doi.org/10.5772/58391

81

The hypothesis of GSCs is based on the concept that a rare subset of cells within GBM may have significant expansion capacity and the ability to generate new tumors. The remainder of tumor cells, which predominantly make up GBM, may represent partially differentiated cells with limited progenitor capacity or terminally differentiated cells that cannot form new tumors. Following the model of glioma origin from sub-ventricular zone (SVZ) after nitro‐ sourea derivatives [104], the most important hypothesis on gliomagenesis is today that GSCs derive by the transformation of Normal Stem Cells (NSCs) or progenitors, *i.e.* from B or C cells of the SVZ niche [105]. There is a great similarity between SVZ NSCs or progenitors and TICs and malignant gliomas most probably originate from the SVZ [106,107]. The concept is supported by the observation that GBM is almost always in contact with lateral ventricles [108]. This hypothesis cannot be applied to benign gliomas that should derive from mature glia. According to other hypotheses, also GBMs could derive from mature glial cells by acquiring stemness properties through a dedifferentiation process [109] or from stem cells of the white matter, NG2 cells. This origin would fit better with tumors far from the ventricles or with secondary GBM [110]. Also reactive astrocytes can be candidate for glioma origin [111,112],

GSCs develop in niches that can be perivascular or perinecrotic [114]. In perivascular niches there is a close contact between endothelial cells and Nestin+and CD133+cells [115]; the former would favor the self-renewal of the latter, mainly by Notch, and the opposite would happen

**5. GSCs in the tumor: Target of therapies?**

and nature of GSCs is necessary.

In the recognition of tumor cell infiltration in edematous areas by MRI, histological examina‐ tion of the surgical samples corresponding to the ROIs on rFSE or rFLAIR areas, is of great importance, in spite of the demonstration that removing T2 hyperintense non-enhancing areas and areas possibly containing ITCs, survival did not change [65]. It must be known that a T2 hyper-intense area may well correspond to a tumor (Figure 14B). A distinction would be possible, provided that there is no mismatch between the ROIs and sample removal. Usually, the cells composing edematous areas can be: tumor, normal or endothelial cells, macrophages or inflammatory cells and mainly reactive astrocytes. In our experience, the cell count is of paramount importance, especially when the number of non-tumor cells largely exceeds that of tumor cells, including in the former reactive astrocytes, microglia and endothelial cells. By comparing the number of cells in H&E stained sections and of GFAP+and CD68+cells with MRI variables, it has been found that normal cells, reactive astrocytes and microglia cells represent a rather stable quota of cells, so that variations of the total number of cells of a given area could be attributed to tumor cells. Reactive astrocytes, once no more proliferating, become fibrillogenetic and mature; usually, they do not exceed a certain number *per* field (Figure 12A). Therefore, they may influence the total number of cells only when tumor cell infiltration is mild. When the number of infiltrating cells is high, the astrocytic quota becomes insignificant. The same can be said for microglia/macrophages. Inside the tumor these cells are often found in perivascular or perinecrotic masses, but in peritumor tissue they are more regularly distributed and they too do not exceed a certain number *per* field (Figure 12B,C). CBV or Cho/ NAA values will be influenced by macrophages/microglia only when the total number of cells is very low, *i.e.* when tumor cell infiltration will be really mild, below a certain percentage of the total number of cells, taking into account that the number of reactive astrocytes plus that of microglia/macrophages usually corresponds to the half of that of normal cells (unpublished data).

ITCs can be detected only after a systemic study of the brain at autopsy, as in the whole mounting preparation technique (Figure 15); they cannot be detected in surgical material because this usually cannot include them [4]. ITCs remain as a sword of Damocles in regard to tumor recurrence.

Spatial Relationships of MR Imaging and Positron Emission Tomography with Phenotype, Genotype and... http://dx.doi.org/10.5772/58391 81

**Figure 15.** Isolate tumor cells in a white matter bundle. PCNA, DAB, 400x.
