Hybrid PET/MRI in Cerebral Glioma

Advanced MRI methods and PET using radiolabelled amino acids provide valuable information in addition to conventional MR imaging for brain tumour diagnostics. The advent of hybrid PET/MRI has allowed a convergence of the methods, but up-to-date simultaneous imaging has reached little relevance in clinical neuro-oncology. A key factor for the benefit of PET/MRI in neuro-oncology is a multimodal approach that provides decisive improvements in the diagnostics of brain tumours compared with a single modality.

brain tumour diagnosis
cerebral glioma
PET
radiolabelled amino acids

1. Introduction

Currently, the diagnosis of brain tumours is primarily based on contrast-enhanced MRI. Structural imaging using T1- and T2-weighted sequences provides high-resolution imaging of brain tumours and allows a differential diagnosis in a large fraction of lesions [1]. Differentiating tumour tissue from non-specific tissue changes, however, can be difficult, especially in cases of gliomas with diffusely infiltrating tumour growth, lack of contrast enhancement, and reactive tissue changes after surgery, radiotherapy, alkylating chemotherapy, or other experimental therapy approaches. In this situation, PET using radiolabelled amino acids can provide important additional diagnostic information [2]. The Response Assessment in Neuro-Oncology (RANO) Working Group has recommended the use of amino acid PET, in addition to MRI, in all stages of brain tumour management [3,4,5,6,7,8]. O-(2-[¹⁸F]-fluoroethyl)-L-tyrosine (¹⁸F-FET) was developed in our institution in the 1990s in order to provide a fluorine-18-labelled amino acid PET tracer with a longer half-life (110 min), which provides logistical advantages compared with shorter-lived carbon-11 labelled amino acids (half-life 20 min) such as [¹¹C]-methyl-L-methionine [9,10,11].

2. PET Tracers for Brain Tumour Imaging

Today, radiolabelled amino acids are the preferred PET tracers in neuro-oncology [1]. Amino acid PET is helpful regarding differential diagnosis, classification and the prognostication of newly diagnosed brain tumours, the delineation of brain tumour extent for treatment planning, the assessment of treatment response and the differentiation of tumour recurrence or progression from treatment-related changes [1]. The most widely used amino acid tracers are [¹¹C]-methyl-L-methionine (MET), ¹⁸F-FET and 3,4-dihydroxy-6-[¹⁸F]-fluoro-L-phenylalanine (¹⁸F-FDOPA), as described in previous publications from the RANO Group [3,4]. Furthermore, the synthetic amino acid analogue anti-1-amino-3-[¹⁸F]fluorocyclobutane-1-carboxylic acid (FACBC or Fluciclovine) has gained clinical interest for brain tumour imaging in recent years [25,26,27]. The uptake of these tracers in brain tumours is primarily dependent on the increased expression and functionality of large neutral amino acid transporters of the L-type (LAT, subtypes LAT1 and LAT2) [1]. In contrast to radiolabelled amino acids, the most widely used PET tracer 2-[¹⁸F]-fluorodeoxyglucose (¹⁸F-FDG) has a limited use in brain tumours because of the high glucose metabolism in normal brain tissue. The proliferation tracer [¹⁸F]-3′-deoxy-3′-fluorothymidine accumulates in cerebral gliomas in relation to the grade of malignancy and prognosis [29,30], but uptake is usually restricted to contrast-enhancing tumour parts on MRI and the tumour volume is smaller than that observed with amino acid tracers [31]. [¹¹C]-choline or [¹⁸F]-fluoro-choline are markers of cell membrane phospholipids in brain tumours, but tracer uptake is also restricted to tumour parts with the disruption of the blood–brain barrier (BBB) [32]. A correlation of tracer uptake with the grade of malignancy has been reported [33,34], but the role of choline tracers in the primary diagnosis of brain tumours is limited, as the accumulation is not tumour-specific [35,36,37].

Another important approach for brain tumour imaging is the use of ligands for the mitochondrial translocator protein (TSPO), such as [¹¹C]-PK11195, [¹⁸F]-GE-180 and [¹⁸F]-DPA-714 [45]. TSPO is overexpressed in activated microglia and macrophages, but also in glioma cells [46]. PET imaging of gliomas using TSPO ligands depicts tumours with high contrast compared with the normal brain [47], but discrimination between tumour mass and brain tissue appears to be critical at the tumour rim, where glia-associated microglia/macrophages may also show high tracer binding [48,49,50]. TSPO ligands accumulate in brain areas with intact BBB, but differences exist in the visualisation of tumour extent compared with amino acid PET [51].

3. Advanced MRI Methods in Neuro-Oncology

Advanced MRI methods can provide functional, physiologic and molecular information beyond conventional MRI, which may be helpful in equivocal findings [53]. A detailed description of these methods is beyond the scope of this article, and therefore only a brief overview of the most important methods from this area is given. PWI either via dynamic susceptibility contrast (DSC) MRI, dynamic contrast-enhanced (DCE) MRI or arterial spin labeling (ASL) MRI provides several surrogate markers of tissue perfusion, such as relative cerebral blood flow (rCBF), relative cerebral blood volume (rCBV), and other perfusion metrics [1,54,55]. In particular, rCBV mapping is a valuable supplement to conventional MRI in the differentiation of tumour progression or recurrence from treatment-related changes [56].

Proton MR spectroscopy (MRS) enables the non-invasive measurement of the signals of selected metabolites in vivo. Important metabolites for the characterisation of brain tumours are the neuronal marker N-acetyl-aspartate (NAA) and choline-containing compounds as cell membrane markers (Cho). MR spectroscopic imaging (MRSI) provides parameter maps, which visualise heterogenous distributions of different metabolites, or ratios thereof, in larger volumes of the brain [57]. Diffusion-weighted imaging (DWI) is based upon the random Brownian motion of water molecules within a voxel of tissue, which can be quantified, for example, by the apparent diffusion coefficient (ADC) [58]. In brain tumours, the ADC is inversely correlated with cell density, probably due to reduced water mobility from dense cellular packing. Diffusion kurtosis imaging (DKI) is an advanced neuroimaging modality that is an extension of diffusion tensor imaging by estimating the kurtosis (skewed distribution) of water diffusion based on a probability distribution function [59].

4. Hybrid PET/MRI in Animal Research

Hybrid PET/MRI has been successfully used in preclinical neuroimaging to correlate changes in neuronal activity using fMRI and changes in receptor expression and neurotransmitter binding [66,67,68,69]. Simultaneous PET/MRI imaging is essential for these examinations, as neuronal activations in temporally separate examinations are not comparable and do not permit any reliable conclusions. Moreover, several studies have used combined PET and MRI in animal brain tumour models to explore novel PET tracers and advanced MR methods for brain tumour diagnosis, but the investigations have used mainly sequential PET/MRI [70,71,72,73,74].

Previous review articles have made suggestions as to the expectation that simultaneous hybrid PET/MRI will be used for the modelling of physiological and biochemical processes, because during the simultaneous acquisition one can be sure the prevailing physiological conditions such as blood flow and perfusion, pertain to both the PET and MRI measurements [75]. However, there has been little implementation in experimental brain tumour research to date.

5. Hybrid PET/MRI in Newly Diagnosed Cerebral Gliomas

In brain lesions suspicious for neoplasms, conventional MRI is frequently inconclusive and additional imaging methods can be helpful. This concerns differential diagnosis, the definition of an optimal biopsy site, and the detection of tumour infiltration, especially in tumours without contrast enhancement in MRI. Furthermore, the non-invasive classification of tumours and the assessment of molecular features and prognostication can be valuable if neuropathological assessment is not possible. Pyka et al. investigated the value of combined ¹⁸F-FET PET and MRS in a series of 67 patients with newly diagnosed gliomas [77]. Static ¹⁸F-FET PET allowed the differentiation of low-grade and high-grade gliomas with an area under the curve (AUC) in receiver operating characteristics analysis (ROC) of 0.86 and MRS using the Cho/NAA with an AUC of 0.72. The combination of ¹⁸F-FET PET and MRS achieved an AUC of 0.97. Furthermore, the multimodal approach was able to differentiate glioblastoma from non-glioblastoma with an AUC of 0.97. In the survival analysis, PET parameters (but not spectroscopy) were significantly correlated with overall survival. Song et al. reported that the combination of ¹⁸F-FET PET and DSC-PWI increased the diagnostic accuracy to differentiate gliomas with and without IDH mutation (AUC 0.90) compared with the single modalities (¹⁸F-FET PET and rCBV, each AUC 0.80), but none of the parameters discriminated between oligodendrogliomas and astrocytomas [78].

Summarising, there is some evidence that combined amino acid PET and advanced MRI is helpful in improving the non-invasive characterisation of suspected gliomas. Concerning tumour delineation, amino acid PET appears to be the most reliable method to identify metabolically active tumour tissue, and so far there is little evidence that the combination with advanced MR methods leads to superior results.

6. Hybrid PET/MRI in Patients with Recurrent Gliomas

Most studies investigating multimodal PET/MRI to differentiate brain tumour progression or recurrence from treatment-related changes have compared PWI with amino acid PET. While some older publications reported the superiority or equivalence of rCBV mapping compared with amino acid PET [94,95,96], more recent publications consistently observed the superiority of amino acid PET [97,98,99]. Recently, scholars analysed the additive value of ¹⁸F-FET PET and perfusion-weighted MRI in a group of 104 patients with suspected glioma recurrence [100]. Eighty-three patients had tumour progression and 21 patients had treatment-related changes. The combination of ¹⁸F-FET PET and PWI did not increase the diagnostic power, but an rCBVmax > 2.85 reached a positive predictive value of 100% so that 44 patients could be correctly classified using rCBVmax alone. In the remaining patients, ¹⁸F-FET PET still achieved an accuracy of 78%, so that 87% of the patients could be correctly diagnosed, in total. These results support the sequential use of PWI and amino acid PET, particularly when a more economical use of the diagnostic methods has priority. In contrast, one study using ¹¹C-MET PET reported on an additive value of amino acid PET and DSC-PWI [101]. While both the maximum tumour-to-brain ratio (TBR_max) of ¹¹C-MET uptake and mean rCBV achieved an AUC of 0.85, the combination of the parameters yielded an AUC of 0.95 in the differentiation tumour recurrence from radiation injury. Furthermore, a number of studies have reported the additive value of amino acid PET and MRI when including advanced MRI methods other than rCBV in patients with suspected tumour recurrence. Jena et al. achieved the highest accuracy (97%) in differentiating recurrent tumours from radiation necrosis when combining the TBR_max of ¹⁸F-FET uptake and MRS using the Cho/Cr ratio [102]. An identical accuracy of 97% was achieved by Sogani et al. with a combination of ¹⁸F-FET PET, MRS, PWI and DWI [103], and a hybrid PET/MRI study achieved an accuracy of 95% using ¹⁸F-FDOPA as the amino acid tracer [104]. Another hybrid PET/MRI study compared dynamic ¹⁸F-FET PET, PWI, and DWI in 47 patients with suspected glioma recurrence [105]. Static ¹⁸F-FET PET alone achieved an AUC of 0.86 for differentiating recurrent tumour and treatment-related changes, which could be increased to an AUC of 0.89 when combined with PWI and DWI. Lohmeier et al. reported the highest AUC by using a combination of static ¹⁸F-FET PET and ADC (0.90) versus ¹⁸F-FET PET (0.81) or ADC alone (0.82) [106]. These results could not be confirmed by Werner et al., who reported the highest accuracy using static and dynamic ¹⁸F-FET PET parameters (93%), which could not be further improved by ADC mapping [107].

Few data exist concerning the additive value of amino acid PET and advanced MR methods in terms of response assessment. A recent study reported that the simultaneous evaluation of ¹⁸F-FET PET and ADC metrics using PET/MRI allowed the early and reliable identification of treatment responses and predicted overall survival in recurrent glioblastoma patients treated with regorafenib [111]. A key aspect in this study was the fact that the authors used pathological ¹⁸F-FET uptake to define the region of interest (ROI) on the ADC maps. The authors emphasised that radiological recommendations do not provide a strategy for identifying the ROI on the DWI-ADC images or how to define the threshold for pathological ADC values. Thus, a PET-guided evaluation strategy for advanced MRI methods is another important aspect for the use of PET/MRI and also played a decisive role in the combined use of ¹⁸F-FET PET and DKI mentioned above [109].

7. Hybrid PET/MRI in Paediatric Brain Tumours

The use of hybrid PET/MRI appears particularly advantageous in paediatric patients, in order to reduce the examination time, to avoid radiation exposure from the CT scanner, and to prevent repeated general anaesthesia in separate measurements [112,113]. Furthermore, the fusion of separately acquired PET and MRI data may cause more problems in children than in adults owing to the fact that paediatric tumours are frequently located in the cerebellum and medulla or by high extra cerebral ¹⁸F-FET uptake in the cranial bone marrow [18]. On the other hand, the logistics of anaesthesia in the hybrid scanner are challenging, especially in younger children, and attenuation correction in children causes problems [18] as MR-based attenuation methods often are built upon reference data sets acquired in adult subjects [114,115].

This entry is adapted from the peer-reviewed paper 10.3390/cancers15143577

© Text is available under the terms and conditions of the Creative Commons Attribution (CC BY) license; additional terms may apply. By using this site, you agree to the Terms and Conditions and Privacy Policy.