**5. Integrin-based probes**

Angiogenesis is a crucial process for tumor growth and metastasis (Kountouras et al. 2005). This process requires intracellular and extracellular interactions in which integrins play an important role (Brooks et al. 1994). Antagonists against alphav*<sup>3</sup>*integrin have been shown to block angiogenesis and reduce tumor growth in preclinical animal models (MacDonald et al. 2001) and clinical trials (Carter 2010). This integrin is a membrane bound receptor that mediates intracellular signal transduction by recognizing and binding to Arg-Gly-Asp (RGD) containing proteins in the extracellular matrix (Main et al. 1992). On binding to different types of RGD containing protein, it senses the external microenvironment and accordingly regulate cellular shape, mobility and cell cycle progression along with angiogenesis and metastatsis (Brooks et al. 1994). Alphav*<sup>3</sup>*integrin is expressed at low levels on epithelial cells and mature endothelial cells but highly expressed on activated endothelial cells of the neovasculature of gliomas (Liu 2009). Its expression correlates well

than those obtained with [11C]choline in the same patients. A preliminary study by Kwee et al. (2004) in 2 patients suggested that [18F]FCH uptake was significantly higher in glioblastoma multiforme (GBM) than benign demyelinating disease. Subsequently, Kwee et al. (2007) performed a more extensive [18F]FCH PET study on 30 consecutive patients (14 women, 16 men; age range, 26-79 years) with solitary brain lesions defined by MRI. In this study, the order of SUV and T/N ratios were benign lesions < high-grade gliomas < metastases from distant tumors with appreciable separation of these classes. [18F]FCH is also

Pre-clinical studies on the *in vivo* kinetics and metabolism of [18F]FCH and choline in a 9L glioma allograft tumor model showed marked washout from the tumor possibly resulting from the hypoxic nature of these tumors. However, a strong association of [18F]FCH uptake and angiogenesis was found in the C6 glioma xenograft model (Wyss et al. 2007) and increased [18F]FCH uptake was observed in a multidrug resistant U87MG glioma tumor

Fig. 5. [18F]FCH uptake in recurrent GBM. PET/CT (left) and [18F] FCH-PET (right) image shows a 6 mm focus of increased [18F]FCH uptake along a right frontal lobe resection cavity. This lesion was noted to increase in size on serial brain MRI consistent with the diagnosis of recurrent tumor. A post-craniotomy defect is evident on the PET/CT image. [18F] FCH-PET is potentially advantageous for imaging brain tumors such as GBM given the low amounts of physiologic cerebral [18F]FCH uptake. *Image Courtesy of Dr. Sandi A. Kwee, MD, Nuclear* 

Angiogenesis is a crucial process for tumor growth and metastasis (Kountouras et al. 2005). This process requires intracellular and extracellular interactions in which integrins play an

block angiogenesis and reduce tumor growth in preclinical animal models (MacDonald et al. 2001) and clinical trials (Carter 2010). This integrin is a membrane bound receptor that mediates intracellular signal transduction by recognizing and binding to Arg-Gly-Asp (RGD) containing proteins in the extracellular matrix (Main et al. 1992). On binding to different types of RGD containing protein, it senses the external microenvironment and accordingly regulate cellular shape, mobility and cell cycle progression along with

levels on epithelial cells and mature endothelial cells but highly expressed on activated endothelial cells of the neovasculature of gliomas (Liu 2009). Its expression correlates well

*<sup>3</sup>*integrin have been shown to

*<sup>3</sup>*integrin is expressed at low

*Medicine Department, Queen's Medical Center, Honolulu, HI, USA.* 

important role (Brooks et al. 1994). Antagonists against alphav

angiogenesis and metastatsis (Brooks et al. 1994). Alphav

useful in detecting recurrent GBMs (**Figure 5**).

model (Vanpouille et al. 2009).

**5. Integrin-based probes** 

with tumor progression and invasiveness of gliomas (Bello et al. 2001). Therefore, this integrin has received attention as a target for imaging probe development.

The most common integrin-targeting radiopharmaceuticals are radiolabeled RGD containing peptides. Cyclic RGD peptides (cRGD peptides) are preferred over linear RGD peptides due to their higher metabolic stability (Bogdanowich-Knipp et al. 1999). However, radiolabeled cRGD peptides (18F-FB-cRGDs (Chen et al. 2004) and 64Cu-DOTA-cRGDs (Chen et al. 2004)) commonly suffer the drawback of poor tumor retention and high renal and/or hepatic uptake. To improve the tumor imaging properties of these tracers, dimeric (Chen et al. 2004) and tetrameric (Wu et al. 2005) cyclic RGD congeners have been developed. The multimeric cyclic RGD probes showed higher tumor uptake but nevertheless exhibited rapid uptake by liver and kidneys. To decrease uptake in liver and kidneys, modifications such as glycosylation (Schnell et al. 2009) or PEGylation (Chen et al. 2004) have helped to achieve higher uptake in tumors along with decreased uptake in liver and kidney in U87MG glioblastoma models (Chen et al. 2004; Chen et al. 2004; Wu et al. 2005) and in patients with *de novo* or recurrent GBM (Schnell et al. 2009).

New developments to integrin imaging are in progress. Recently, a new generation of RGD containing probes have been designed that show better potential for imaging gliomas. The new approach is based on cystine knot proteins or knottins that are relatively stable in physical, chemical and biological environments due to presence of scaffold of disulfidebonded framework and a triple-stranded ß-sheet fold (Kimura et al. 2009). The Knottin family members also possess one or more surface-exposed loops that can tolerate sequence diversity. RGD peptides have been grafted into these surface-exposed loops, while radiolabeling accomplished with conjugation of PET radionuclide, 18F-FB or 64Cu-DOTA. In a U87MG glioblastoma model, a knottin based probe demonstrated rapid and high tumor accumulation, fast clearance from blood and normal organs, and low uptake in the kidney and liver (Miao et al. 2009).

#### **6. Cellular proliferation probes**

Increased cellular proliferation is an integral part of the cancer phenotype (Bading and Shields 2008). The primary requirement for cell proliferation is replication of nuclear DNA. Of the 4 nucleotides (adenine, guanine, cytosine and thymidine) required for DNA synthesis, thymidine is the only one that is specific for DNA. In cells, thymidine is derived either *de novo* or through the salvage pathway (Bading and Shields 2008). The *de novo* pathway is not a viable alternative for monitoring DNA synthesis in proliferating cells because the relevant precursors (deoxyuridine, uridine, and uracil) are routed into both DNA and RNA (Bading and Shields 2008). Thus, thymidine salvage pathway is a better choice for indication of proliferation. PET radiotracers of thymidine have been developed, including [11C]thymidine ([11C]TdR) (De Reuck et al. 1999) and 3'-deoxy-3'-fluorothymidine ([18F]FLT). These probes are currently being evaluated for their potential for imaging cellular proliferation in gliomas. The thymidine-based tracers are transported into cells and subsequently phosphorylated by thymidine kinase-1 (TK-1), thereby rendering them trapped within the cell. Since TK-1 activity correlates to a significant extent with the cellular proliferation rate, the PET-measurement of tissue retention of radioactivity is a noninvasive indicator of proliferation. The advantage of using the fluorinated thymidine tracer [18F]FLT over [11C]TdR is twofold. First, replacement of the hydroxy group of

PET Imaging of Gliomas 171

(Wheeler et al. 2000). Sigma-receptors are expressed more in proliferating tumor cells than in quiescent tumor cells (Wheeler et al. 2000). Using a preclinical tumor model, Van Waarde et al. (2004) were the first to report feasibility of using sigma-receptor binding ligands, [11C]SA4503 and [18F]FE-SA5845 for detecting gliomas. It is known that tumors are heterogenous in nature, including areas of low and high proliferation rates. The proportion of cells with low proliferation rate increases with increase in tumor size, especially in the necrotic center. In the C6 glioma rat model, tumor uptake of [18F]FE-SA5845 showed a negative correlation with tumor size (P < 0.0001), in contrast to that of [11C]SA4503, suggesting that tissue binding of [18F]FE-SA5845 is solely related to cellular proliferation. Later Van Waarde et al. (2006) compared the bio-distribution of 4 PET tracers ([11C]SA4503, [18F]FE-SA5845, [11C]choline and [11C]MET) with previously published bio-distribution data of [18F]FLT and [18F]FDG in C6 glioma rat tumor model. In their study, sigma-receptor ligands and [18F]FLT were more tumor selective than [18F]FDG, [11C]choline, or [11C]MET in the C6 glioma model. However, [11C]SA4503 and [18F]FE-SA5845 were less sensitive than were [11C]choline, [11C]MET, and [18F]FDG. Clinical PET studies for evaluation of sigma

Meningiomas are the most common non-glial primary tumors of the central nervous system accounting for approximately 15% of all intracranial tumors (Buetow et al. 1991). More than 90% of intracranial meningiomas are slow growing and histopathologically benign but malignant meningiomas are not rare (Buetow et al. 1991; Goldsmith et al. 1994). [18F]FDG uptake in tumor lesions is dependent on glycolytic rate and disruption of blood brain barrier (Roelcke et al. 1995). High expression of the somatostatin receptor (SSTR) subtype 2 (Dutour et al. 1998) in meningiomas offer the possibility of receptor-targeted imaging (Henze et al. 2005; Henze et al. 2001) of meningiomas. In the case of meningiomas with low glycolytic rate and intact BBB (Roelcke et al. 1995), somatostatin receptor based tracers might be more useful for tumor detection and disease management than [18F]FDG. In a clinical study, a somatostatin receptor analog, 68Ga-DOTA-D-Phe1-Tyr3-octreotide (DOTA-TOC) labeled with the positron emitter 68Ga (half-life, 68 min) was evaluated for imaging meningioma (Henze et al. 2001). In contrast to [18F]FDG, this ligand showed higher T/N uptake ratios. The initial results are encouraging but more clinical studies are needed to fully assess the potential of

Hypoxia, a hallmark of aggressive tumor behavior often noted in high grade glioblastomas, is associated with resistance to therapy, poorer survival, invasion and aggressiveness (Szeto et al. 2009). Estimation of hypoxia could be an important determinant of overall survival in several tumors including gliomas. PET imaging with the hypoxia radiotracer [18F]fluoromisonidazole ([18F]FMISO) presents a possible means of noninvasively detecting tumor hypoxia in gliomas (Rasey et al. 2000; Valk et al. 1992). In a preclinical C6 glioma tumor model study (Tochon-Danguy et al. 2002), [18F]FMISO uptake was significantly higher in tumor tissue compared to normal brain and the uptake was independent of tumor size. [18F]FMISO uptake was observed homogeneously throughout viable glioma tissue in tumor sizes ranging from 2 mm to almost 1 cm. Quantitation of uptake of [18F]FMISO

receptor expression in gliomas are ongoing.

somatostatin receptor based tracers for imaging meningiomas.

**7. Somatostatin-based probes** 

**8. Hypoxia-based probes** 

deoxyribose with radiofluorine makes it resistant to degradation (Shields et al. 1998) and second, the longer half-life of 18F is more practical for clinical imaging.

Increased glucose metabolism in inflammatory tissues and other non-specific lesions is the main source of false-positive [18F]FDG-PET findings in oncology. Van Waarde et al. (2004) used a rodent model with C6 glioma tumor and inflammatory lesion to show that [18F]FLT was more specific than [18F]FDG for uptake by glioma lesions relative to inflammation. However, the ability of [18F]FLT to detect tumors was dependent on disruption of the BBB (Muzi et al. 2006), rendering it more suitable for detecting high-grade gliomas than lowgrade ones. In a separate clinical study (Chen et al. 2005), [18F]FLT–PET was more sensitive than [18F]FDG-PET for imaging recurrent high-grade tumors, correlated better with the Ki-67 proliferation index and was a more powerful predictor of tumor progression and survival. The reason for superiority of [18F]FLT–PET over [18F]FDG–PET was low [18F]FLT uptake in the normal brain tissue leading to higher T/N contrast (**Figure 6**). When compared with [11C]MET, Jacobs et al. (2005) showed that [18F]FLT was less sensitive in detecting tumors than [11C]MET, especially for low-grade astrocytomas. Nevertheless, Kawai et al. (2009) found a high correlation between histological tumor grade and [18F]FLT uptake in gliomas. A significant difference in SUVmax of [18F]FLT was observed between grade II (0.27 ± 0.06, n=6) and grade IV (2.18 ± 0.93, n=10) gliomas (P < 0.0001), and grade III (0.70 ± 0.45, n=7) and grade IV gliomas (P < 0.001). Importantly, [18F]FLT uptake correlated significantly better with the Ki-67 index (r = 0.86, P < 0.0001) than did methionine uptake in gliomas. Studies have also reported the use of [18F]FLT-PET in investigating the effectiveness of therapy and prognosis of gliomas (Chen et al. 2007; Kawai et al. 2009). Increased [18F]FLT accumulation is also observed in other brain tumors including malignant lymphoma (Kawai et al. 2009).

Fig. 6. Patient with glioblastoma. FLT-PET scan (A) and contrast-enhanced MRI (B) of biopsy-proven glioblastoma multiforme (GBM) in a 58 year-old female, just prior to the initiation of chemoradiation. Thick black arrow points to the highly proliferative rim of tumor surrounding a photopenic region of central necrosis. FLT uptake in the normal brain parenchyma is minimal, allowing a favorable lesion-to-background ratio. Physiologic uptake is also seen within the bone marrow (short thin arrow) and scalp (long thin arrow) *Image Courtesy of Dr. Laura Horky, Brigham and Women's Hospital, Boston*.

Another strategy for imaging of cell proliferation in tumor lesions involves non-invasive assessment of the concentration of sigma receptors in cells using sigma-receptor ligands

deoxyribose with radiofluorine makes it resistant to degradation (Shields et al. 1998) and

Increased glucose metabolism in inflammatory tissues and other non-specific lesions is the main source of false-positive [18F]FDG-PET findings in oncology. Van Waarde et al. (2004) used a rodent model with C6 glioma tumor and inflammatory lesion to show that [18F]FLT was more specific than [18F]FDG for uptake by glioma lesions relative to inflammation. However, the ability of [18F]FLT to detect tumors was dependent on disruption of the BBB (Muzi et al. 2006), rendering it more suitable for detecting high-grade gliomas than lowgrade ones. In a separate clinical study (Chen et al. 2005), [18F]FLT–PET was more sensitive than [18F]FDG-PET for imaging recurrent high-grade tumors, correlated better with the Ki-67 proliferation index and was a more powerful predictor of tumor progression and survival. The reason for superiority of [18F]FLT–PET over [18F]FDG–PET was low [18F]FLT uptake in the normal brain tissue leading to higher T/N contrast (**Figure 6**). When compared with [11C]MET, Jacobs et al. (2005) showed that [18F]FLT was less sensitive in detecting tumors than [11C]MET, especially for low-grade astrocytomas. Nevertheless, Kawai et al. (2009) found a high correlation between histological tumor grade and [18F]FLT uptake in gliomas. A significant difference in SUVmax of [18F]FLT was observed between grade II (0.27 ± 0.06, n=6) and grade IV (2.18 ± 0.93, n=10) gliomas (P < 0.0001), and grade III (0.70 ± 0.45, n=7) and grade IV gliomas (P < 0.001). Importantly, [18F]FLT uptake correlated significantly better with the Ki-67 index (r = 0.86, P < 0.0001) than did methionine uptake in gliomas. Studies have also reported the use of [18F]FLT-PET in investigating the effectiveness of therapy and prognosis of gliomas (Chen et al. 2007; Kawai et al. 2009). Increased [18F]FLT accumulation is also observed in other brain tumors including malignant

Fig. 6. Patient with glioblastoma. FLT-PET scan (A) and contrast-enhanced MRI (B) of biopsy-proven glioblastoma multiforme (GBM) in a 58 year-old female, just prior to the initiation of chemoradiation. Thick black arrow points to the highly proliferative rim of tumor surrounding a photopenic region of central necrosis. FLT uptake in the normal brain parenchyma is minimal, allowing a favorable lesion-to-background ratio. Physiologic uptake is also seen within the bone marrow (short thin arrow) and scalp (long thin arrow)

Another strategy for imaging of cell proliferation in tumor lesions involves non-invasive assessment of the concentration of sigma receptors in cells using sigma-receptor ligands

*Image Courtesy of Dr. Laura Horky, Brigham and Women's Hospital, Boston*.

**A B**

second, the longer half-life of 18F is more practical for clinical imaging.

lymphoma (Kawai et al. 2009).

(Wheeler et al. 2000). Sigma-receptors are expressed more in proliferating tumor cells than in quiescent tumor cells (Wheeler et al. 2000). Using a preclinical tumor model, Van Waarde et al. (2004) were the first to report feasibility of using sigma-receptor binding ligands, [11C]SA4503 and [18F]FE-SA5845 for detecting gliomas. It is known that tumors are heterogenous in nature, including areas of low and high proliferation rates. The proportion of cells with low proliferation rate increases with increase in tumor size, especially in the necrotic center. In the C6 glioma rat model, tumor uptake of [18F]FE-SA5845 showed a negative correlation with tumor size (P < 0.0001), in contrast to that of [11C]SA4503, suggesting that tissue binding of [18F]FE-SA5845 is solely related to cellular proliferation. Later Van Waarde et al. (2006) compared the bio-distribution of 4 PET tracers ([11C]SA4503, [18F]FE-SA5845, [11C]choline and [11C]MET) with previously published bio-distribution data of [18F]FLT and [18F]FDG in C6 glioma rat tumor model. In their study, sigma-receptor ligands and [18F]FLT were more tumor selective than [18F]FDG, [11C]choline, or [11C]MET in the C6 glioma model. However, [11C]SA4503 and [18F]FE-SA5845 were less sensitive than were [11C]choline, [11C]MET, and [18F]FDG. Clinical PET studies for evaluation of sigma receptor expression in gliomas are ongoing.
