**2. Primary tumors**

Gliomas are the most common primary brain tumors. Glioblastoma accounts for 55% of all cases followed by diffuse astrocytoma (8.6%), ependymal tumors (6.9%), anaplastic astrocy‐ tomas (6.1%), oligodenrogliomas (5.7%), pilocytic astrocytomas (5.2%), and otherless frequent glioma types [1]. Incidence is higher in males and in whites than in blacks. For glioblastoma, the median age of diagnosis is 64 years but for low‐grade gliomas (grades I and II), most often occur between 20 and 40 years. The major diagnostic goal in gliomas is the differentiation of low‐grade from high‐grade gliomas and from other pathologies that have similar imaging features.

## **2.1. Diffusion‐weighted imaging**

Apparent diffusion coefficient (ADC) maps alone cannot differentiate between glioma from another neoplasm or glioma type. Malignancy is usually associated with increased cellular density, resulting in decreased signal intensity on ADC images. High‐grade gliomas usually have significant lower ADC values than low‐grade gliomas. A lesion‐to‐normal (L/N) ADC ratio of 1.43 could differentiate low‐grade from high‐grade gliomas with 100% sensitivity and 94.4% specificity [2]. Evaluation of the perilesion area may aid the diagnosis of a pri‐ mary tumor due to its infiltrative nature (**Figure 1**).

**Figure 1.** (A) A case of a grade II astrocytoma and of glioblastoma (D). There is a lower ADC value in astrocytoma (B) than glioblastoma (E). The rCBV map shows increased perfusion in glioblastoma (F) contrary to astrocytoma (C).

#### **2.2. Perfusion imaging**

(1.9%), oligodendrogliomas (1.6%), embryonal tumors (1.1%), oligoastrocytic tumors (0.9%),

Imaging has a fundamental role in intracranial tumor management. Magnetic resonance imaging (MRI) is the imaging modality of choice for establishing diagnosis, classification, surgical planning, and post‐treatment follow‐up. The latest MRI techniques, namely diffu‐ sion, perfusion, and spectroscopy, offer more than the anatomical information that conven‐ tional imaging provides. Diffusion allows the assessment of water displacement within tissue. Diffusion tensor imaging permits the mapping of axonal organization. Perfusion MRI is a technique for the assessment of cerebral perfusion. Dynamic susceptibility contrast imaging (DSC‐MRI) perfusion technique is currently the most widely used and allows the calculation of relative cerebral blood volume (rCBV) and relative cerebral blood flow (rCBF). MR spectroscopy, with single‐voxel or multi‐voxel techniques, can detect metabolites within tissue such as N‐acetyl aspartate (NAA), choline‐containing compounds (Cho), myoinositol (mI), lactate (Lac), creatine (Cr), and other molecules. However, no tumor‐specific metabolite has

Although the discrimination between intra‐axial and extra‐axial lesions is relatively straight‐ forward, for the accurate discrimination of the variety of intra‐axial tumors of several difficulties exist. This is of paramount importance for timely and appropriate patients' management. Herewith, we provide an overview of the latest MR techniques for the differ‐

Gliomas are the most common primary brain tumors. Glioblastoma accounts for 55% of all cases followed by diffuse astrocytoma (8.6%), ependymal tumors (6.9%), anaplastic astrocy‐ tomas (6.1%), oligodenrogliomas (5.7%), pilocytic astrocytomas (5.2%), and otherless frequent glioma types [1]. Incidence is higher in males and in whites than in blacks. For glioblastoma, the median age of diagnosis is 64 years but for low‐grade gliomas (grades I and II), most often occur between 20 and 40 years. The major diagnostic goal in gliomas is the differentiation of low‐grade from high‐grade gliomas and from other pathologies that have similar imaging

Apparent diffusion coefficient (ADC) maps alone cannot differentiate between glioma from another neoplasm or glioma type. Malignancy is usually associated with increased cellular density, resulting in decreased signal intensity on ADC images. High‐grade gliomas usually have significant lower ADC values than low‐grade gliomas. A lesion‐to‐normal (L/N) ADC ratio of 1.43 could differentiate low‐grade from high‐grade gliomas with 100% sensitivity and 94.4% specificity [2]. Evaluation of the perilesion area may aid the diagnosis of a pri‐

and all other less frequent tumors (11.7%) [1].

been recognized to date.

166 Neurooncology - Newer Developments

**2. Primary tumors**

features.

ential diagnosis of intra‐axial tumors.

**2.1. Diffusion‐weighted imaging**

mary tumor due to its infiltrative nature (**Figure 1**).

Perfusion MRI can be performed using a variety of methods. The most common techniques are as follows: T2‐weighted dynamic susceptibility contrast (DSC), T1‐weighted dynamic contrast enhanced (DCE), and arterial spine labeling (ASL). The latter do not require con‐ trast administration. The relative cerebral blood volume (rCBV) is the most frequent report‐ ed metric. This can be calculated by comparing the cerebral blood volume in a region of interest that is drawn over the tumor to the CBV of a mirror region of interest placed over the normal white matter in the contralateral side.

Gliomas are characterized by increased blood vessels formation for the transport of nu‐ trients and oxygen which are essential for tumor growth. Furthermore, apart from glioma infiltration to parenchyma, malignant cells can also migrate using a perivascularroute through microvasculature [3]. Recentreports showed that glioblastoma cells can even differentiate into endothelial cells andpericytes, thus aiding tumor vascularization [4]. High‐grade gliomas have a significantly higher rCBV ratio than low‐grade gliomas (**Figure 1**). A cut‐off ratio of 0.63 has been suggested for the differentiation between them [2]. Furthermore, a significant linear correlation has been reported between rCBV ratio and glioma proliferation potentials as assessed by Ki‐67 index and tumor's cell cycle analysis [5, 6]. High‐grade tumors had higher Ki‐67 index, higher percentage of cells in G2/M phase, and lower percentage of cells in G0/G1 phase.

Oligodendrogliomas contrary to other low‐grade gliomas have significant higher rCBV values (mean 3.68 ± 2.39) [7], overlapping even with high‐grade gliomas. A possible explana‐ tion to this is their increased vascular density and cortical localization [7]. Another impor‐ tant exception is pilocytic astrocytoma, the most common pediatric brain tumor with usually infratentorial localization [8]. The tumor's mural nodule may show increased rCBV ratio in comparison with other low‐grade gliomas. Clinicians should also bear in mind that heman‐ gioblastomas may also demonstrate high rCBV ratios.

#### **2.3. MR spectroscopy**

Initial studies with MR spectroscopy showed promise for the diagnosis of brain lesions and grading; however, recent evidences are controversial. Several metabolites can be measured that correlate with various pathological alterations within lesions. NAA is the acetylated form of the amino acid, aspartate, which is found in increased concentrations in viable neurons. Given that non‐neuronal neoplasms destroy normal neurons there is a reduction in NAA signal. Choline is a marker of cell membrane and can be found elevated in tumors and inflammatory processes. Creatine is a measure of energy stores, whereas lactate increases in cases of ischemia, in which the cell switches to anaerobic glycolysis and lactates accumulates. Thus, lactate is more likely to be present in high‐grade than low‐grade gliomas. Lipids have been recognized as a marker of myelin breakdown. Several studies have evaluated both single‐ voxel and multi‐voxel spectroscopy. Multi‐voxel has the advantage of greater spatial resolu‐ tion and extent of coverage, thus permits the evaluation of different components of heterogeneous masses. Within tumor, some areas may be more metabolically active than others. In brain tumors, there is usually an increased signal of Cho, whereas NAA and Cr are reduced. Cho/Cr ratio tends to increase as glioma malignancy progresses. In a recent meta‐ analysis, Cho/NAA ratio showed a sensitivity of 80% and specificity of 76%, higher than Cho/ Cr ratio and NAA/Cr ratio for the differentiation of high‐grade from low‐grade gliomas [9]. However, both sensitivity and specificity do not enable an accurate diagnosis, thus addition‐ al imaging modalities may be needed. A CHO/NAA ratio >1, in voxels outside of the enhance‐ ment region, suggests tumor infiltration and is indicative of a high‐grade glioma (**Figure 2**).

**Figure 2.** MR proton spectrum of a grade II astrocytoma (A) and of a glioblastoma (B). Contrary to low‐grade glioma, glioblastoma exhibits depression of the NAA and creatine (Cr) peaks and elevation of the choline (Cho) peak.
