An essential function of the endothelial cells in the cerebral vasculature is the active transport of amino acids from blood to tissue, and endothelial proliferation is associated with an increase of transporter expression and activity (Verrey et al. 2004). Several amino acids are being used as PET tracers, and increased uptake in gliomas has been demonstrated clinically (see review by (Herholz et al. 2012)) and in experimental studies (Miyagawa et al. 1998; Viel et al. 2012).

For many years, 11C-methionine (11C-MET) has been the most widely used amino acid tracer. Compared to normal brain, the uptake is increased in most gliomas, including the majority of low-grade gliomas with intact blood-brain barrier (BBB) (Herholz et al. 1998). On average, uptake is higher in high-grade than in low-grade gliomas, but the difference is quantitatively much less pronounced than with tracers and contrast agents that depend on BBB breakdown, which is present in the majority of high-grade but absent in the majority of low-grade gliomas. In cross-sectional studies, there is no consensus about the utility of 11C-MET for glioma grading (Moulin-Romsee et al. 2007). This may partially be due to the quantitatively small mean difference between grades and is potentially confounded by the fact that oligodendrogliomas have significantly higher uptake than astrocytomas of the same WHO grade (Derlon et al. 1997; Herholz et al. 1998; Ribom and Smits 2005; Shinozaki et al. 2011), even though they tend to have longer survival times. Nevertheless, longitudinal studies demonstrate that malignant progression of gliomas is associated with a significant increase of 11C-MET uptake that is absent in non-progressing tumours (Ullrich et al. 2009).

The very short physical half-life of carbon-11 (20 min) restricts the application of 11C-MET to PET research centres with own cyclotron and thus prevents wider clinical use. Therefore, fluorine-18-labelled amino acid tracers, which have the potential for routine clinical application, have found broad interest. Amongst them, 18F-fluoroethyltyrosine (18F-FET) has been used by several clinical research groups. Similar to 11C-MET, it is also substrate for the large neutral amino acid transporters (Langen and Broer 2004), while, in contrast to 11C-MET, it is not incorporated into proteins. Interestingly, studies comparing the clinical properties of these two tracers directly (Weber et al. 2000) found similar tumour to background ratios (while overall tissue uptake of 18F-FET is somewhat lower than 11C-MET), suggesting that the activated transport is the biologically dominant process rather than further metabolism. Correspondingly, it was also found that the magnitude of 18F-FET uptake is not a particularly strong indicator of tumour grade. This limitation could possibly be overcome by dynamic scans, because the early phase of tracer uptake, which is dependent on blood flow and volume provides a better differentiation between high- and low-grade gliomas than late uptake (Popperl et al. 2007; Jansen et al. 2012). However, it remains to be determined by direct comparison of dynamic 18F-FET studies with measurements of CBV by dynamic contrast-enhanced MR, whether amino acid tracers will ultimately prove to provide additional information for gliomas grading.

As an alternative to 18F-FET, 18F-FDOPA has also been used for imaging of gliomas (Heiss et al. 1996; Ledezma et al. 2009; Tripathi et al. 2009; Walter et al. 2012) with encouraging results. Obviously, there is more variable background than with 18F-FET because of the high physiological accumulation of FDOPA in the basal ganglia. Apparently, involvement of the basal ganglia by gliomas is not that frequent, especially at the time of initial diagnosis, and therefore allows a clear distinction between normal and pathological uptake areas in most patients.

Tracers binding to integrin show promise as more specific markers for endothelial activation and proliferation than amino acids. In a pilot study, uptake of 18F-Galacto-RGD images correlated with immunohistochemical alpha(v)beta(3) integrin expression of corresponding tumour samples in glioblastoma (Schnell et al. 2009).

The standard techniques for imaging of BBB barrier breakdown in malignant gliomas and most other brain tumours are contrast-enhanced CT and MR which, by use of dynamic scanning, also allow quantitation of BBB permeability (Haroon et al. 2004; Herholz et al. 2007). Corresponding nuclear medicine techniques, such as 201Tl SPECT (Dierckx et al. 1994), have attracted only limited interest probably because of their relatively low resolution, even though the sensitivity of the potassium analogue 201Tl for BBB changes may be higher than the large inert molecules used as contrast agents with CT and MR (Gomez-Rio et al. 2008).

It is likely that BBB breakdown also increases amino acid uptake (Roelcke et al. 1996). However, the effect is relatively small as demonstrated by comparable uptake in high-grade gliomas with and without contrast enhancement (Herholz et al. 1998). BBB breakdown does, however, massively increase uptake of labelled nucleic acids and their analogues, such as FLT, because the capacity of their transporters at the BBB is very limited and tumours without BBB damage are negative on a FLT scan (Tripathi et al. 2009).

Several studies demonstrated that FDG uptake in gliomas is related to prognosis (Patronas et al. 1985; Holzer et al. 1993; Kaschten et al. 1998; De Witte et al. 2001; Padma et al. 2003; Pardo et al. 2004; Van Laere et al. 2005; Pauleit et al. 2009). The main limitation for putting that into broader clinical use has probably been the difficulty to identify gliomas on FDG PET scans. FDG uptake in normal grey matter is very high similar to that of malignant gliomas, resulting in poor contrast. White matter FDG uptake is lower (typically about 40–50 % of the grey matter on a FDG PET scan), and FDG uptake of low-grade gliomas is usually to white matter uptake. Thus, the regional heterogeneity of normal brain metabolism is in the same range as the difference between low- and high-grade gliomas. Correspondingly, in most studies comparing FDG with amino acid tracers (Fig. 41.1), the latter appear to provide the more sensitive and distinct signal (Derlon et al. 2000; Kim et al. 2005; Van Laere et al. 2005; Potzi et al. 2007; Pauleit et al. 2009; Li et al. 2012).

The limitations caused by poor image contrast can be overcome by PET-MR coregistration (Padma et al. 2003). While this approach has been limited to research laboratories in the past, there is now standard software provided by the major scanner vendors for PET-MR coregistration based on optimising a measure of mutual information in the two scan modalities, which achieves an accuracy of 1–2 mm and thus considerably better than actual PET scan spatial resolution (Cizek et al. 2004). Reliable classification of FDG uptake in tumour, as localised on the coregistered MR, can be achieved by a grading scale in relation to normal brain tissue, e.g. 0 = no uptake; 1 = uptake less or equal to contralateral white matter; 2 = uptake greater than the contralateral white matter and less than the grey matter; and 3 = equal to or greater than the contralateral grey matter (Padma et al. 2003). Rather than relying on a plain SUV value only, semiquantitative assessment of lesion uptake should be done in the tissue volume that is suspected of tumour on the MR scan and divided by the uptake in a contralateral mirror region, providing a SUV ratio (SUVR) which is equivalent to a simple activity ratio. Normal brain glucose metabolism is reduced in patients with aggressive tumours (DeLaPaz et al. 1983; Holzer et al. 1993), thus enhancing the contrast between malignant gliomas with high FDG uptake and normal brain background. An improvement of the contrast can also be achieved by delayed scanning, which may demonstrate late accumulation of FDG in the tumour (Spence et al. 2004; Prieto et al. 2011).

The most important factors being considered in histological grading of gliomas are the degree of cellular anaplasia and indicators of cellular proliferation. Initial work with 11C-thymidine demonstrated in vitro and in vivo that the activity of thymidine kinase 1 (TK-1) is associated with cellular accumulation of the tracer and tumour proliferation (Eary et al. 1999). In order to facilitate more widespread applicability, the fluorine-18-labelled thymidine analogue FLT has been developed for imaging of tumour proliferation (Shields et al. 1998). Although its affinity to cellular nucleoside transporters and TK-1 is less than that of thymidine (Krohn et al. 2005), a correlation between tracer uptake kinetics and proliferation markers has been demonstrated in malignant glioma (Ullrich et al. 2008; Price et al. 2009). FLT does not cross the intact BBB (at least not at an amount that would allow clinical PET imaging), thus it provides a negative finding in most low-grade gliomas (Chen et al. 2005) and is not suitable for measuring proliferation rates in this tumour group (Tripathi et al. 2009).

Magnetic resonance spectroscopy has demonstrated significant differences with respect to the phospholipid and choline signal between low- and high-grade gliomas (Tedeschi et al. 1997). Correspondingly, there is also interest whether this could also be achieved by 11C-choline (Shinoura et al. 1997). Ohtani et al. (2001) demonstrated increased uptake in high-grade gliomas and extra-axial tumours lacking the BBB, while uptake in most low-grade gliomas was comparable to background. Initial experience with 18F-fluorocholine is very limited (Lam et al. 2011), and interference of the BBB with tracer uptake is likely.