Diffuse low-grade gliomas

, by  Hugues DUFFAU , popularity : 100%
[English] [français]

Metabolic imaging

MRI spectroscopy (MRS) measures major metabolites in the tumor tissue. The spectrum of a typical DLGG shows high choline, reflecting increased cell membrane turnover, reduced N-acetyl-aspartate, reflecting an increase neuronal loss and myoinositol reflecting glial proliferation. However, these spectral anomalies can also be observed in some non-neoplastic lesions. Although it is not possible to diagnose the degree of glioma based only on MRS due to similarities between the low-grade and high-grade gliomas, the presence of lactate and lipids (reflecting necrosis) is associated with a higher proliferative activity and a more aggressive behavior (35). Despite the limitations mentioned above, MRS is useful for guiding a biopsy and for follow up including patients on treatment (36).

DSC-MRI perfusion sequences (Dynamic Susceptibility Contrast Imaging) enable the calculation of relative Cerebral Blood Volume (rCBV), which correlates with the micro-vasculature. An increased VCSr in DLGGs is predictive of malignant transformation even before the onset of contrast enhancement (12). However, these observations seem limited to astrocytomas because of the higher VCSr in oligodendrogliomas.

DCE-MRI (Dynamic Contrast-Enhanced Imaging) sequences measure the permeability of the blood-brain barrier by calculating a transfer coefficient (Ktrans), which is correlated with the grade of the tumor, although the correlation is not as strong that for VCSr (42).

As for MRI diffusion, the values ​​of the apparent diffusion coefficient (ADC) are lower and more variable in oligodendrogliomas than in astrocytomas. There is no correlation between the diffusion coefficient and the level of choline. Quantitative MRI in oligodendrogliomas with deletion 1p19q shows a more heterogeneous T1-T2 signal, less distinct margins and an increased VCSr compared with gliomas without deletion (6).

Positron Emission Tomography (PET) may also provide additional information in the diagnosis and monitoring of a DLGG (70). The [18F] - fluorodeoxyglucose (FDG) has limited value, due to a low uptake by DLGGs with respect to normal cortex. PET-FDG is basically for the detection of anaplastic transformation of astrocytomas and in the differential between radiation necrosis and tumor recurrence. PET with 11C-Methionine (MET) has the advantage of capturing a TEM correlated with proliferative activity of tumor cells. MET uptake in normal brain tissue is less than that of FDG, allowing better contrast and better delineation of DLGG. In addition, DLGGs with oligodendroglial component further capture MET.

MET-PET may be useful for differentiating DLGGs from non-tumor lesions, to guide stereotactic biopsies, to determine the pre-operative volume and to quantify the response to treatment. However, a cyclotron is required. Recently, 18F-fluoro-ethyl-L-tyrosine (FET) was used to guide biopsies and plan treatment for gliomas. FET has the advantage of a longer half-life than the MET, thereby enabling the manufacturing of the tracer in a center with a cyclotron and its transportation to other institutions. The experience of the FET-PET is still limited compared to the MET-PET, but these tracers seem to show a similar uptake and distribution in brain tumors.

In summary, neuroimaging is useful for diagnosis, to guide biopsy or surgical resection, to plan for radiotherapy and to monitor response to treatment.

Functional Imaging

In terms of exploration of brain function, the development of functional neuroimaging techniques, including functional MRI (fMRI), magnetoencephalography, white fiber tractography by diffusion tensor imaging and transcranial magnetic stimulation, has enabled the realization of non-invasive mapping of the whole brain. These methods provide an estimate of the localization of eloquent areas (ie involved in sensorimotor, visual, speech and cognitive) with respect to glial tumor, while providing information about the hemispheric lateralization of speech.

 However, it is crucial to note that functional imaging is currently not sufficiently reliable at the individual level to be used in routine clinical practice. This is mainly due to the fact that this imagery is not a direct reflection of the reality of cerebral function, but a very indirect approximation, based on biomathematical reconstructions - explaining why the results may vary depending on the model used. (22)

In fact, with regard to fMRI correlational studies with intraoperative electrophysiology have shown that the sensitivity of the fMRI varied from 59% to 100% for speech (specificity 97% to 0%) (34). Moreover, fMRI is not able to differentiate between essential functional regions (that must be surgically preserved) from involved but non critical regions for a given function (which can be surgically removed, since a functional compensation is possible).

Diffusion tensor imaging a new technique which allows for the tractography of the main white matter tracts, still requires validation. In fact, the use of different models and software from the same data leads to different reconstructions, showing that tractography is not reliable or reproducible. The correlations between this method and intraoperative electrophysiology (direct subcortical electro-stimulation) showed agreement in only 82% of cases. In other words, a negative tractography does not formally rule out the presence of important fibers within the glioma. Furthermore, this technique is able to provide (indirect) anatomical information but by no means does it provide information on the function of subcortical fibres. Therefore it is not currently reasonable to rely on this method for surgical indications or for the planning of surgery, although it is an excellent tool for both education and research (43 ).

3. Anatomo-pathological Features

As mentioned in the introduction, the WHO classification recognizes grade II astrocytomas, oligodendrogliomas and oligoastrocytoma (44). Morphological criteria differentiate astrocytomas from oligodendrogliomas, even if such a distinction can be problematic, especially in ’mixed’ forms - since there are no specific recommendations on the relative proportion of astrocytic and oligodendroglial tissue differentiation enabling the diagnosis of oligoastrocytoma.

Astrocytomas

Diffuse astrocytomas include fibrillar, (most common) protoplasmic and gemistocytic forms; the latter having been set aside because of an increased risk of malignant transformation. Fibrillary astrocytoma may show some nuclear atypia within a fibrillar matrix.

The gemistocytic variant consists of eosinophilic cytoplasms engorged with eccentric rings in more than 20% of tumor cells. Mitotic activity in the grade II astrocytomas is very low. One mitotic activity should not lead to the diagnosis of anaplastic astrocytoma, although a mitotic activity during stereotactic biopsy should raise suspicion.

The most common molecular alteration in astrocytomas is IDH1 mutation reported in approximately 75% of astrocytomas. However, this alteration is found with a similar frequency in oligodendrogliomas thus represents a marker for WHO grade II and III astrocytomas, oligodendrogliomas and oligoastrocytomas. The development of a specific antibody (H09) of IDH1 R132H mutation-is very useful for the diagnosis of astrocytomas, oligodendrogliomas and oligoastrocytomas. H09 covers over 90% of all IDH1 mutations in diffuse gliomas.

Proliferation index studied using anti-Ki67 / MIB-1 is generally less than 4% in diffuse astrocytoma. Tumor necrosis, capillary endothelial proliferation, vascular thrombosis and high mitotic activity are not consistent with a grade II diffuse astrocytoma. The best immunohistochemical marker is GFAP (Glial Fibrillary Acidic Protein). P53 mutation is present in 50% of diffuse astrocytomas and in 80% of gemistocytic astrocytomas, while the co-deletion 1p19q is rare.

Oligodendrogliomas

Oligodendrogliomas have a moderate cell density and typically have a perinuclear halo giving a ’honeycomb’ or ’fried egg’ appearance. Occasionally, small tumor cells with eosinophilic cytoplasm are encountered and are termed ’mini-gemistocytes’. Oligodendrogliomas have a dense network of capillaries and frequently contain microcalcifications. Occasional mitotic activity and Ki-67 index / MIB-1 of about 5% are consistent with a grade II oligodendroglioma. There is no specific immunohistochemical marker for oligodendrogliomas.

The molecular characteristic of oligodendrogliomas is the 1p19q co-deletion found in 80% of these tumors, while a p53 mutation occurs in only 5% of the cases. IDH1 somatic mutation is present in 80% of oligodendrogliomas.

Oligoastrocytoma

Oligoastrocytomas should be diagnosed upon detection of significant oligodendroglial and astrocytic components but the inter-observer differences for the diagnosis of oligoastrocytoma remains high. Most oligoastrocytoma present an 1p19q deletion or a p53 mutation. These changes tend to be found in both tumor compartments. Up to 80% oligoastrocytomas carry a somatic IDH1 mutation.

Limitations

A number of limitations need to be underscored (40). First, a high inter-and intra-observer variability been shown and should be taken into account due to a lack of reproducibility of the current WHO classification. This explains why morphological neuropathological examination must be associated with the study of molecular characteristics (61).

In addition, in the context of surgical biopsies, especially stereotactic biopsy, even if they are guided by metabolic imaging and staged, there remains a risk of under grading of tumor.

Indeed, DLGGs are heterogeneous tumors with possible macro-foci and even micro-foci of ’malignant transformation’ within a grade II tumor which may not be revealed in the biopsy. Thus, a specimen is only a sample of glioma, a ’false negative’ ’grading’ can lead to inappropriate therapy. Finally, the current WHO classification does not clearly recognize the existence of a continuum between glioma grade II and III gliomas. The number, size and spatial distribution of foci of potential malignant transformation are not taken into account.