**2. Glucose-based probes**

Increased glucose uptake and glycolysis are hallmark characteristics of a variety of neoplasms (Pedersen 2007; Warburg 1956). This makes radiolabeled glucose analogs logical tracers of choice for imaging of tumors. The radiofluorinated analog of glucose, [18F]fluorodeoxyglucose, or [18F]FDG, is the most common PET tracer for clinical PET oncology studies. A practical synthesis, suitable half-life (T1/2=109.7 min), negligible circulating metabolites, and well-established kinetics of uptake and retention of [18F]FDG makes it a preferred imaging probe in cancer imaging (Spence et al. 1998). [18F]FDG is transported into cancer cells by glucose transporters (GLUT1 and GLUT3) and, like glucose, it is phosphorylated via hexokinase to form [18F]fluorodeoxyglucose-6-phosphate. However, in contrast to glucose-6-phosphate, [18F]fluorodeoxyglucose-6-phosphate is very slowly metabolized further and hence is effectively trapped in the cancer cell (Spence et al. 1998). The trapped [18F]FDG-6-phosphate can be detected by [18F]FDG-PET thereby allowing noninvasive evaluation of glucose uptake and glycolysis. In general, [18F]FDG-PET performs well in identifying highly malignant, high-grade gliomas because they typically exhibit higher glycolysis rates than the normal cerebral cortex (Di Chiro et al. 1982, 1987a). Di Chiro et al. (1985, 1987a, 1987b) were first to correlate [18F]FDG uptake with WHO grading of gliomas on the basis of a semiquantitative index of the ratio of [18F]FDG uptake in tumor to the average [18F]FDG uptake in normal cerebral cortex. [18F]FDG uptake ratio in high-grade glioblastoma was almost twice that of low-grade gliomas. In a different study, Delbeke et al. (1995) reported that high-grade gliomas can be detected with high sensitivity of 94% and specificity of 77% when tumor-to-white matter ratios exceed 1.5, and tumor-to-grey matter ratios exceed 0.6. In addition,[18F]FDG-PET was able to indicate anaplastic transformation of grade II gliomas into grade III gliomas by an increase in 18F-FDG uptake (Chen 2007).

appear hypointense on T1-weighted images but hyperintense on T2-weighted images relative to normal brain (Grosu et al. 2002; Sartor 1999). In low-grade gliomas, peritumoral edema is minimal or absent and no contrast enhancement is seen due to intact BBB. Whereas, in high-grade gliomas, peritumoral edema is frequently seen and the tumor lesions usually show contrast enhancement, correlated with the extent of neovascularization and loss of integrity of the BBB owing to tumor infiltration and production of vascular endothelial growth factor. The anatomical features obtained by MRI are not sufficient to differentiate low-grade from high-grade gliomas with intact BBB, tumor lesions from inflammatory or vascular processes, and post-operative residual/relapse from necrosis

In contrast to MRI, positron emission tomography (PET) provides unique functional information of tumors on a range of biological processes such as glucose metabolism, protein/DNA synthesis, cell proliferation, membrane synthesis, angiogenesis and oxygen tension that can reflect the changes in neoplasm (Basu and Alavi 2009). Assessment of the status of these processes in areas of interest in brain has been shown to be helpful in detection and grading of gliomas, delineation of tumor margins, disease prognosis and treatment. PET has also been useful in differentiating post-operative residual tumor from therapy induced necrosis and edema. This review discusses radiopharmaceuticals and progress in the development of PET techniques for imaging of gliomas in the following areas: glucose uptake, amino acid transport, cellular proliferation rate, choline uptake,

Increased glucose uptake and glycolysis are hallmark characteristics of a variety of neoplasms (Pedersen 2007; Warburg 1956). This makes radiolabeled glucose analogs logical tracers of choice for imaging of tumors. The radiofluorinated analog of glucose, [18F]fluorodeoxyglucose, or [18F]FDG, is the most common PET tracer for clinical PET oncology studies. A practical synthesis, suitable half-life (T1/2=109.7 min), negligible circulating metabolites, and well-established kinetics of uptake and retention of [18F]FDG makes it a preferred imaging probe in cancer imaging (Spence et al. 1998). [18F]FDG is transported into cancer cells by glucose transporters (GLUT1 and GLUT3) and, like glucose, it is phosphorylated via hexokinase to form [18F]fluorodeoxyglucose-6-phosphate. However, in contrast to glucose-6-phosphate, [18F]fluorodeoxyglucose-6-phosphate is very slowly metabolized further and hence is effectively trapped in the cancer cell (Spence et al. 1998). The trapped [18F]FDG-6-phosphate can be detected by [18F]FDG-PET thereby allowing noninvasive evaluation of glucose uptake and glycolysis. In general, [18F]FDG-PET performs well in identifying highly malignant, high-grade gliomas because they typically exhibit higher glycolysis rates than the normal cerebral cortex (Di Chiro et al. 1982, 1987a). Di Chiro et al. (1985, 1987a, 1987b) were first to correlate [18F]FDG uptake with WHO grading of gliomas on the basis of a semiquantitative index of the ratio of [18F]FDG uptake in tumor to the average [18F]FDG uptake in normal cerebral cortex. [18F]FDG uptake ratio in high-grade glioblastoma was almost twice that of low-grade gliomas. In a different study, Delbeke et al. (1995) reported that high-grade gliomas can be detected with high sensitivity of 94% and specificity of 77% when tumor-to-white matter ratios exceed 1.5, and tumor-to-grey matter ratios exceed 0.6. In addition,[18F]FDG-PET was able to indicate anaplastic transformation of grade II gliomas into grade III gliomas by an increase in 18F-FDG uptake (Chen 2007).

somatostatin receptor density, angiogenesis and hypoxia.

**2. Glucose-based probes** 

(Chen 2007).

[18F]FDG-PET has also been useful to differentiate hypoglycolytic non-malignant toxoplasmosis common in AIDS patients from hyperglycolytic CNS lymphoma (Hoffman et al. 1993).

Assessment of [18F]FDG uptake in gliomas has high prognostic value (Di Chiro 1987). De Witte et al. (1996) studied 28 patients with histologically proven low-grade gliomas with [18F]FDG–PET and followed progression of disease for a mean of 27 months. All 19 patients with tumors that were hypoglycolytic on PET were alive at the end of the follow-up period, whereas 6 of 9 patients with hyperglycolytic patterns on PET died. The prognostic utility of [18F]FDG–PET has been confirmed in several other studies (Alavi et al. 1988; Barker et al. 1997; Padma et al. 2003; Patronas et al. 1985).

Although [18F]FDG-PET is accurate to detect high-grade gliomas, it has limited usefulness in detection of low-grade gliomas and some high-grade gliomas such as post-operative residual and recurrent glioma (Olivero et al. 1995; Ricci et al. 1998). Since glucose is the preferred fuel in normal brain, high [18F]FDG uptake in surrounding normal tissues in brain is unavoidable (Di Chiro et al. 1982). Low grade gliomas tend to have the same or lower [18F]FDG uptake as compared to average [18F]FDG uptake in white matter, thus resulting in false negative readings (Kawai et al. 2005). This is also true for certain high-grade gliomas, especially hypoglycolytic residual (Padma et al. 2003) and recurrent tumors (Chao et al. 2001) that may exhibit less or similar [18F]FDG uptake to average [18F]FDG uptake in grey matter. In addition, in the case of patients with Alzheimer disease and epilepsy, affected regions in brain can show decreased [18F]FDG uptake compared to background (Fazekas et al. 1989; McGeer et al. 1986). On the other hand, brain regions with abscess or acute necrosis occurring hours of weeks after radiotherapy, chemotherapy can show increased [18F]FDG uptake compared to background leading to false positive readings (Floeth et al. 2006). Thus, low tumor-to-normal background radioactivity concentration (T/N) ratios and difficult to interpret contrasts between normal and pathological regions limit the specificity of [18F]FDG-PET to detect low-grade and residual or relapsed high-grade brain tumors.

Given these concerns, attempts have been made to improve the accuracy of [18F]FDG for imaging of gliomas. In cases where T/N (white or grey matter) ratios for [18F]FDG uptake are not useful for delineating low-grade, residual or relapsed high-grade glioma, two strategies have been reported to help: (1) co-registration of [18F]FDG–PET images with MR images (Chao et al. 2001; Wang et al. 2006) and (2) delayed [18F]FDG–PET imaging (Spence et al. 2004). Co-registration and interpretation of [18F]FDG–PET images with MR images can improve the performance of [18F]FDG–PET **(Figure 1)** for detecting low-grade (Borgwardt et al. 2005; Wong et al. 2004), residual or relapsed high-grade gliomas (Chao et al. 2001; Wang et al. 2006). Low grade gliomas can be identified by similar [18F]FDG uptake to white matter in regions with increased signal on T2-weighted MRI (Borgwardt et al. 2005; Wong et al. 2004), while recurrent high grade gliomas are often indicated as [18F]FDG uptake in regions with contrast enhancement on T1-weighted MRI (Chao et al. 2001; Wang et al. 2006). Delayed PET imaging, as proposed by Spence et al. (2004), is another strategy to improve the contrast between tumor lesion and background. In this study, nineteen patients with gliomas were imaged from 0 to 90 min and once or twice at 3–8 h after injection. In 12 of 19 patients, visual analysis of delayed images up to 8 h afterinjection showed these images to better distinguish relapsed tumors in grey matter (**Figure 2**). Standardized uptake values (SUVs) were also greater in tumors than in normal grey or white matter on delayed imaging. Using kinetic modeling, they demonstrated that the rate constant of

PET Imaging of Gliomas 163

Amino acids play a central role in protein synthesis and intermediary metabolism (Cellarier et al. 2003; Morowitz et al. 2000). The enhanced uptake of essential amino acids into neoplasms through specific amino acid transporters has motivated the design and evaluation of a large number of positron-labeled essential amino acid analogs. In addition, the low uptake of essential amino acids in normal brain tissue relative to tumor tissue renders amino acid tracers advantageous for imaging gliomas (Lilja et al. 1985). The most studied essential amino acid tracers are [11C]methionine ([11C]MET), [18F]fluoroethyl-L-

Increased uptake of methionine by cancer cells results from increased transport flux, primarily by L-amino acid transporters, enhanced protein synthesis, increased need for polyamines, and a high rate of trans-methylation and trans-sulfuration reactions (Leskinen-Kallio et al. 1991). [11C]MET uptake in tumor lesions is not dependent on disruption of the BBB (Roelcke et al. 1995; Sasajima et al. 2004). This is a major advantage compared to MRI where contrast enhancement for detection of tumor lesions is dependent on BBB disruption. Studies with [11C]MET-PET have shown that amino acid tracer, [11C]MET, accumulates in all gliomas, including low-grade glioma that are difficult to detect on contrast-enhanced MRI

[11C]MET-PET can be used to predict histological grades of gliomas. Lilja et al. (1985) evaluated 14 patients with gliomas and found that [11C]MET-PET could differentiate highgrade glioma from low-grade glioma on the basis of T/N ratio. The ratio of the uptake of [11C]MET in high-grade tumors was 1.9-4.8 and low-grade tumor was 0.8-1.0. Derlon et al. (1989) too confirmed positive correlation of T/N ratio with the histological grade of gliomas. Later study with large set of 196 patients, Herholz et al. (1998) showed that [11C]MET could differentiate among high-grade gliomas, low-grade gliomas, and chronic or subacute nontumoral lesions. In this study, [11C]MET-PET was also useful in detecting recurrent or

[11C]MET-PET has been shown to have high prognostic potential. Kaschten et al. (1998) performed [18F]FDG-PET and [11C]MET-PET in 54 patients with gliomas. [11C]MET was superior to [18F]FDG in predicting the histologic grade and prognosis of gliomas. With a larger set of 85 patients, De Witte et al. (2001) applied qualitative and quantitative scoring systems for [11C]MET uptake. Both scoring systems confirmed the prognostic importance of [11C]MET-PET. In this study, gliomas were histologically graded following [11C]MET-PET guided resection (42 cases) or stereotactic biopsy (43 cases). Uptake of [11C]MET was present in 98% of the gliomas studied. The T/N ratio was significantly correlated with the histological grade of glioma. A statistically poor patient outcome was demonstrated during follow-up when this ratio was higher than a threshold of 2.2 for grade II gliomas and 2.8 for grade III gliomas. A high [11C]MET uptake was statistically associated with short survival times. Better prognostic utility of [11C]MET-PET relative to [18F]FDG-PET was also shown in

[11C]MET-PET is also useful to differentiate recurrent tumor from post-operative radiation injury (Gehrke et al. 1991; Ogawa et al. 1991; Sonoda et al. 1998). Tsuyuguchi et al. (2003) examined 21 adult patients with [11C]MET-PET to differentiate radiation necrosis from recurrent metastatic brain tumor following stereotactic radiosurgery. They observed mean

residual tumors as they showed higher [11C]MET uptake than primary gliomas.

tyrosine ([18F]FET) and 3,4-dihydroxy-6-18F-fluoro-L-phenylalanine ([18F]FDOPA).

**3. Amino acid-based probes** 

and [18F]FDG-PET (Ogawa et al. 1993).

other studies (Kim et al. 2005; Van Laere et al. 2005).

**3.1 [11C]MET** 

[18F]fluorodeoxyglucose-6-phosphate degradation (k4) was not significantly different between tumor and normal brain tissue for shorter datasets but was lower in tumor than in normal brain tissue for the longer dataset (8 h), suggesting that higher [18F]Fluorodeoxyglucose-6-phosphate degradation rates are present in normal brain tissue than tumor. Since this report, other studies have shown the utility of delayed PET imaging for delineating brain tumors (Farid et al. 2009; Kim et al. 2010).

Fig. 1. A patient who had received surgery, radiation, and chemotherapy for anaplastic astrocytoma. Axial gadolinium-enhanced T1-weighted image (left) demonstrates nodular enhancement posterior to the surgical resection cavity. Co-registered [18F]FDG-PET image demonstrates increased [18F]FDG activity corresponding to this region, similar to gray matter, compatible with recurrent tumor. Correlation of the MRI and PET imaging findings is necessary to make this determination, and accurate image co-registration is essential.

Fig. 2. A 45-year-old woman with recurrent right temporal GBM. T1-weighted gadoliniumenhanced (T1Gd) MRI showing contrast enhancement in right temporal region of the brain. [18F]FDG–PET scan with much more prominent T/N delineation in this right temporal region at the later time point, 473 min (~8 h), compared to 90 min (1.5 h). *Image reproduced from work by Spence et al. (2004) and used with permission.* 

[18F]fluorodeoxyglucose-6-phosphate degradation (k4) was not significantly different between tumor and normal brain tissue for shorter datasets but was lower in tumor than in normal brain tissue for the longer dataset (8 h), suggesting that higher [18F]Fluorodeoxyglucose-6-phosphate degradation rates are present in normal brain tissue than tumor. Since this report, other studies have shown the utility of delayed PET imaging

Fig. 1. A patient who had received surgery, radiation, and chemotherapy for anaplastic astrocytoma. Axial gadolinium-enhanced T1-weighted image (left) demonstrates nodular enhancement posterior to the surgical resection cavity. Co-registered [18F]FDG-PET image demonstrates increased [18F]FDG activity corresponding to this region, similar to gray matter, compatible with recurrent tumor. Correlation of the MRI and PET imaging findings is necessary to make this determination, and accurate image co-registration is essential.

Fig. 2. A 45-year-old woman with recurrent right temporal GBM. T1-weighted gadoliniumenhanced (T1Gd) MRI showing contrast enhancement in right temporal region of the brain. [18F]FDG–PET scan with much more prominent T/N delineation in this right temporal region at the later time point, 473 min (~8 h), compared to 90 min (1.5 h). *Image reproduced* 

*from work by Spence et al. (2004) and used with permission.* 

for delineating brain tumors (Farid et al. 2009; Kim et al. 2010).

#### **3. Amino acid-based probes**

Amino acids play a central role in protein synthesis and intermediary metabolism (Cellarier et al. 2003; Morowitz et al. 2000). The enhanced uptake of essential amino acids into neoplasms through specific amino acid transporters has motivated the design and evaluation of a large number of positron-labeled essential amino acid analogs. In addition, the low uptake of essential amino acids in normal brain tissue relative to tumor tissue renders amino acid tracers advantageous for imaging gliomas (Lilja et al. 1985). The most studied essential amino acid tracers are [11C]methionine ([11C]MET), [18F]fluoroethyl-Ltyrosine ([18F]FET) and 3,4-dihydroxy-6-18F-fluoro-L-phenylalanine ([18F]FDOPA).
