Molecular imaging with positron emission tomography (PET) plays an important role in the diagnosis and management of patients with brain tumors and epilepsy. The clinical uses of FDG are discussed, as well as the research applications of novel PET tracers. Where applicable, single-photon emission computed tomography (SPECT) is also discussed.
Molecular imaging with positron emission tomography (PET) plays an important role in the diagnosis and management of patients with brain tumors and epilepsy. Where applicable, single-photon emission computed tomography (SPECT) is also discussed.
Overview of PET and SPECT in neurosurgery: brain tumors
Annually in the United States, more than 20,000 people are diagnosed with malignant brain tumors and approximately 13,000 die of primary brain tumors. Another 22,000 are diagnosed with nonmalignant primary brain tumors. Metastatic brain tumors, predominantly from breast, lung, and colon primary cancers are even more common, with approximately 140,000 patients diagnosed each year, and more than 100,000 deaths per year from symptomatic brain metastases.
Nuclear molecular imaging can be used for the assessment of treated or untreated primary or metastatic brain tumors. PET is commonly used for assessing and grading brain tumors, assessing aggressiveness and prognosis, distinguishing between recurrence and postradiation necrosis, and guiding biopsy. Brain SPECT is also a useful technique in the assessment of brain tumor activity.
PET Technology
PET was developed in the 1970s by Phelps and Hoffman, as an in vivo imaging application of autoradiography, using radioactively labeled glucose. Briefly, a positron-emitting radiotracer is injected into the patient and taken up selectively by cells possessing certain molecular characteristics, such as the presence of glucose or amino acid transporters. In the target tissue, the radiotracer decays, emitting positrons. The emitted positrons collide with nearby electrons and are annihilated, producing 2 high-energy (511 keV) photons, which are emitted 180° apart. The photons are detected by a PET scanner, which is a ring-shaped, high-energy coincidence detector that surrounds the patient. Registration of millions of coincidence events allows localization of the radiotracer distribution within the patient. The spatial resolution of PET is 4 to 10 mm, depending on the scanner type. It is generally accepted that assessment of lesions smaller than 7 to 8 mm in diameter or 0.5 cm 3 may be limited.
Hybrid PET/computed tomographic (CT) scanners have all but replaced the traditional PET-only scanners. The hybrid scanner uses low-dose multislice helical CT (approximately 10–40 mA and 130 kilovolt [peak]) for the dual purpose of anatomic localization and attenuation correction. Once acquired, PET images may also be fused to a patient’s magnetic resonance imaging (MRI). PET/MRI systems are currently under development, in which PET and MRI images are acquired simultaneously.
PET Tracers: FDG
Fluorodeoxyglucose
2-Deoxy-2-( 18 F)fluoro- d -glucose ( 18 F-FDG or FDG) is the most common clinical nuclear medicine imaging tracer used today to assess brain tumors. FDG-PET was initially used for functional brain mapping and then became the first PET tracer used for the assessment of brain tumors. An analogue of glucose, FDG uptake correlates with regional glucose metabolism. FDG readily crosses the blood-brain barrier and is transported intracellularly by glucose transporters. Once intracellular, it is phosporylated by glucose 6-hexokinase and trapped in the cell, because it cannot be metabolized. Brain tumors characteristically have a high concentration of glucose transporters and glucose 6-hexokinase and are therefore typically FDG avid.
The brain uses glucose as its main energy source, and glucose is transported by a group of glucose transporters (GLUTs). GLUT1s are expressed in glia, capillary endothelial cells, choroid plexus, and ependymal cells. GLUT3s are expressed in neurons. Gray matter uses 2 to 4 times more glucose than white matter. In brain tumor imaging, the high rate of glucose transport within physiologically active normal brain can obscure the target to background ratio, particularly when the tumor is adjacent to physiologically active gray matter ( Fig. 1 ).
Tumor grading with FDG
FDG is useful for tumor grading because most high-grade tumors, such as high-grade gliomas, medulloblastoma, and primary central nervous system lymphoma, have high concentrations and activity of GLUTs. Most low-grade tumors have lower concentrations of GLUTs and can be distinguished from high-grade gliomas by the lower FDG uptake on PET. FDG avidity of common benign and malignant primary and metastatic brain tumors is shown in Box 1 .
High
High-grade gliomas (grades III and IV)
Medulloblastoma
Primary central nervous system lymphoma
Pineoblastoma
Central neurocytoma
Pilocytic astrocytoma
Pituitary adenoma
Variable
Metastases
Oligodendroglioma
Meningioma
Low
Most low-grade gliomas
Primitive neurectodermal tumor
Vestibular schwannomas
False positive
Infection
Inflammation
Seizure activity
Malignant tumors are given in italics.
As an exception to the rule, some low-grade tumors have a high FDG avidity, which is because of the high concentrations of GLUTs. In pilocytic astrocytomas, for example, the vascular density is high. In this case, glucose uptake does not represent the blood flow but is thought to represent the metabolic activity of the endothelial cells lining the tumor vasculature. Endothelial proliferation and metabolism account for the high FDG avidity. Oligoastrocytomas also have variable FDG uptakes, even within the same tumor grade.
Measurements of tumor FDG uptake versus physiologic brain FDG uptake can be used to predict high- versus low-grade brain tumors, both primary and metastatic, with certain exceptions as noted earlier. Semiquantification of glucose uptake is performed using a standardized uptake value (SUV). A tumor to gray matter uptake ratio of greater than 0.6 or a tumor to white matter uptake ratio of greater than 1.5 was, in 1 series by Delbeke and colleagues, 100% sensitive and 67% specific for distinguishing high-grade gliomas from low-grade gliomas. Applying the same cutoffs to primary brain tumors and brain metastases yielded 94% sensitivity and 77% specificity. FDG avidity has also been inversely correlated with survival. For example, in a series of 165 patients with highly FDG avid brain tumors, 1-year survival was only 29%, as opposed to 94% in patients with tumors with low FDG avidity.
Recurrence versus Postradiation Necrosis with FDG
In primary or metastatic tumors treated with radiotherapy, tumor growth as seen on contrast MRI is often indistinguishable from postradiation necrosis, both appearing as enlarging, contrast-enhancing lesions. Approximately 70% of radiation necrosis cases are diagnosed in the first 2 years following radiation treatment, which is similar to the expected time frame for tumor recurrence. The incidence of radiation necrosis increases with the radiation dose and is particularly high in cases in which chemotherapy is concurrently administered. FDG-PET is a useful tool in distinguishing postradiation necrosis from tumor progression, both in high-grade gliomas and brain metastases. In general, recurrent tumor is FDG avid ( Fig. 2 ), and radiation necrosis is not FDG avid. In the first 3 to 6 months after radiation, FDG-avid radiation-induced inflammatory cells such as macrophages can have the same appearance as a tumor. Inflammatory cells are seen in the setting of edema, neovascularity, demyelination, and necrosis, which begin to occur in the first hours to months after radiation treatment. As discussed in later sections, techniques such as dual-phase FDG-PET imaging and PET imaging with investigational radiotracers can help to improve the accuracy of PET imaging by increasing the target to background ratio.
The sensitivity of FDG-PET for distinguishing tumor recurrence (primary and metastatic) ranges from 80% to 86%, with a specificity of 40% to 88%. In some centers, PET images are fused to MRI and delayed-phase PET images are obtained. Both methods have been shown to improve the accuracy of FDG-PET. Fusion of FDG-PET images to contrast MRI images is helpful for anatomic correlation and helps to distinguish uptake within a lesion from uptake in an adjacent, functionally active gray matter. This ability may be helpful in cases in which the rounded contour of a rim-enhancing lesion may look similar to the contour of a gyrus, which closely surrounds the lesion, as in Fig. 1 .
In one series, the sensitivity of FDG-PET in assessing for recurrent brain metastasis versus radiation necrosis without fusion to MRI was 65% and 85%, respectively, and after PET/MRI fusion, sensitivity increased to 86% and specificity to 80%.
Dual-Phase FDG-PET Imaging
Dual-phase FDG-PET imaging has been shown to increase accuracy in distinguishing recurrence from postradiation necrosis in gliomas and brain metastases. This method enhances the tumor to background ratio and is based on differences in FDG tracer kinetics between tumor and normal brain parenchyma. When FDG-PET is normally performed at 45 to 60 minutes after tracer injection, FDG tracer kinetics are similar between tumor and gray matter. In other words, transport of the tracer through the blood-brain barrier and into the cells is similar for the tumor and normal brain ( Fig. 3 ). Phosphorylation of the tracer by glucose 6-hexokinase, which traps the FDG intracellularly, may be higher in the tumor than in normal gray matter because tumors may have abnormally high levels of glucose 6-hexokinase. Dephosphorylation via glucose 6-phosphatase is higher in normal brain parenchyma than in tumor. Over time, trapping and retention of the tracer by the tumor is greater than that of normal brain parenchyma. This process is most pronounced at 3 to 6 hours after tracer injection.
In the authors’ clinic, dual-phase imaging was performed on a series of 25 patients with treated brain metastases (predominantly breast, lung, and melanoma primary cancer) and enlarging contrast-enhancing lesions on MRI suspicious for recurrence versus radiation necrosis. Of the 25 patients, 22 received radiation, most often stereotactic radiosurgery. The other 3 received chemotherapy without radiation. SUVmax was calculated for the lesion and normal gray matter at 2 imaging sessions, first at 45 to 60 minutes after tracer injection and the second at a mean of 225 minutes after the initial scan. The ratio of SUVmax of the lesion to gray matter was measured at early and late time points, using the following formula: [(L2/GM2−L1/GM1)/(L1/GM1)] where L1 = early SUVmax lesion; L2 = late SUVmax lesion; GM1 = early SUVmax normal contralateral gray matter; GM2 = late SUVmax normal contralateral gray matter. Using receiver operating characteristic (ROC) analysis, the change in this ratio over time was the most accurate parameter for distinguishing recurrence and posttreatment necrosis. When this increase was 19% or more between the early and late imaging sessions, the sensitivity and specificity for detecting tumor were 95% and 100%, respectively. This ROC cutoff is currently being tested prospectively in brain metastases and primary brain tumors. An example of dual-phase imaging is shown in Fig. 4 .
Several investigational PET tracers provide more specific information about tumor properties such as cellular proliferation, protein synthesis, and hypoxia.
Proliferation: 3-Deoxy-3-[ 18 F]Fluorothymidine
3-Deoxy-3-[ 18 F]fluorothymidine (FLT) is an imaging biomarker of mitotic activity. FLT uptake correlates with the activity of thymidine kinase-1, an enzyme expressed during the S phase (DNA synthesis) of the cell cycle. FLT tracer uptake within normal brain parenchyma is negligible because it does not significantly cross the blood-brain barrier and most normal brain cells do not divide. This scenario allows for an excellent tumor to background ratio. FLT uptake is high in high-grade gliomas because of their high mitotic index and blood-brain barrier permeability. Conversely, FLT uptake is low in low-grade gliomas because of a low mitotic index and an impermeable blood-brain barrier. FLT is more sensitive than FDG in detecting de novo and recurrent high-grade gliomas. Both tracers are useful for glioma grading, as they are generally insensitive in low-grade gliomas ( Fig. 5 ).
FLT-PET shows promise as an early predictor of treatment response. In an interesting pilot study of 19 patients with recurrent glioblastoma multiforme, a greater than 25% decrease in FLT uptake between baseline and posttreatment scans in patients receiving bevacizumab and irinotecan was able to predict outcome, with a 3-fold greater survival in metabolic responders than in nonresponders (mean 10.8 vs 3.4 months).
Optimal assessment of FLT uptake in brain tumors (ie, in baseline vs posttreatment scans) may require kinetic modeling with prolonged dynamic imaging times of up to 90 minutes. Furthermore, metabolite correction optimizes FLT image analysis, which involves collection of blood samples for measurement of radioactive metabolites of FLT. FLT imaging is therefore more cumbersome than FDG for both the patient and the image analyst.
Amino Acid Transport and Protein Synthesis
The essential amino acids are transported into the brain by the large neutral amino acid (LNAA) transport system. Amino acid transport via the LNAA transport system is markedly increased in low- and high-grade gliomas. Advantageously, amino acid tracers cross the blood-brain barrier and label both low- and high-grade gliomas, and uptake in the gray matter is low.
11 C-Methionine ( 11 C-MET) is perhaps the most well-studied amino acid tracer. Increased 11 C-MET uptake is not seen in all low-grade gliomas, and uptake levels are a predictor of biologic behavior. In a series of patients with low-grade gliomas, one-third of patients had a MET uptake that was similar to or lower than that of normal brain parenchyma. When 11 C-MET uptake was equal to or less than that of the normal brain parenchyma, prognosis was significantly better than in patients with a MET uptake higher than that of the normal brain parenchyma. This observation is similar to the results using an analogous 18 F-labeled amino acid radiotracer, O -(2- 18 F-fluoroethyl)- l -tyrosine (FET).
11 C-MET is useful in optimizing stereotactic biopsy guidance of low- and high-grade gliomas. In a group of 32 patients with low- and high-grade gliomas receiving a total of 70 stereotactic biopsies, 100% of 61 11 C-MET-positive biopsy sites were diagnostic for tumor, whereas all 11 C-MET-negative biopsy sites were negative for tumor or nondiagnostic. Because 11 C-MET does not easily discriminate between low- and high-grade tumor, correlation with FDG-PET is helpful in guiding the biopsy to the highest grade portion of the tumor. Amino acid PET also shows promise in optimizing neurosurgical resections of gliomas and radiation planning for gliomas.
The short 11 C radioactive half-life of 20 minutes limits availability of 11 C-MET to hospitals with an on-site cyclotron. 18 F-labeled amino acid tracers such as 18 F-FDOPA (3,4-dihydroxy-6-[ 18 F]-fluoro- l -phenylalanine) and 18 F-FET may be produced off-site because of the longer half-life (110 minutes). These tracers are generally considered to be functionally interchangeable with 11 C-MET because their LNAA transport systems are similar and tumor uptake is primarily due to increased transport.
FDOPA
Before the application of FDOPA to brain tumor imaging, it was primarily used for assessing the striatal dopamine pathway in patients with movement disorders. Applied to gliomas, FDOPA has been shown to be 98% sensitive and 86% specific in distinguishing tumor from radiation necrosis, using a tumor:striatum SUV cutoff ratio of 0.75. Like other amino acid tracers, it is useful in characterizing low-grade gliomas and in distinguishing recurrent tumor from radiation-induced necrosis.
FET
Like other amino acid tracers, FET is useful in distinguishing radiation necrosis from glioma recurrence. FET-PET was used in 45 patients with MRI suspicious for recurrent glioma and was 93% sensitive and 100% specific in making the distinction between recurrence and postradiation necrosis. It has also been shown that analysis of amino acid tracer kinetics using FET and FDOPA accurately distinguishes low- and high-grade gliomas.
Hypoxia: 18 F-Misonidazole
The most well-studied PET hypoxia tracer is 18 F-misonidazole (MISO), a derivative of nitroimidazole. Oxygen consumption of brain tumors is lower than that of the normal brain because of disturbed angiogenesis. MISO is lipophilic, uniformly diffuses throughout the brain, and is trapped only by hypoxic cells. Its uptake is higher in high-grade gliomas because these tumors are typically more hypoxic. Because hypoxia is associated with radioresistance, it is thought that imaging biomarkers of hypoxia may be helpful in treatment planning.
The tumor imaging tracers discussed here are summarized in Table 1 . It should be mentioned that nuclear brain imaging extends beyond PET. SPECT imaging, using a different detector system, is widely clinically available and a valuable complement and/or alternative to FDG-PET. SPECT radiotracers include thallium 201, 99m Tc-sestamibi and 111 Indium-pentreotide (OctreoScan). Each of these tracers demonstrates low background uptake in the brain, which allows for a high target to background ratio. 111 Indium-pentreotide is a somatostatin analogue and is taken up by neuroendocrine tumors, including meningiomas. 111 Indium-pentreotide is most sensitive in well-differentiated meningiomas, which express high levels of somatostatin receptors, but is less useful in dedifferentiated, atypical meningiomas.
| Biomarker | Radiotracer | Clinical/Investigational |
|---|---|---|
| Glucose transport and metabolism | 18 F-FDG | Clinical |
| DNA synthesis | 18 F-FLT | Investigational |
| Protein synthesis and/or transport | 11 C-MET 18 F-FDOPA 18 F-FET | Investigational Investigational Investigational |
| Hypoxia | 18 F-FMISO | Investigational |
Thallium 201, an analogue of potassium, enters viable cells via the Na + , K + -ATPase pump. Thallium does not cross the blood-brain barrier and is therefore used in assessing MRI contrast-enhancing lesions to differentiate tumor recurrence (primary or metastatic) from radiation necrosis. Thallium uptake correlates with the tumor grade and proliferative index and is most sensitive in high-grade lesions. Thallium 201 is reportedly up to 100% sensitive in distinguishing tumor (primary and metastatic) recurrence from postradiation necrosis in as little as 6 to 12 weeks after standard radiation therapy or radiosurgery. Recent hemorrhage, infarct, or the presence of an abscess may give a false positive result.
99m Tc-sestamibi is another useful tracer for distinguishing recurrent tumor (high-grade glioma or metastasis) from radiation necrosis. It is a cation that accumulates in negatively charged mitochondria. It does not cross the blood-brain barrier, and background uptake is minimal, aside from the choroid plexus. For distinguishing tumor recurrence from radiation necrosis, its sensitivity ranges from 67% to 100% and specificity from 91% to 100%, with the highest accuracy in the setting of high-grade gliomas.
111 Indium-pentreotide is a somatostatin analogue and is taken up by neuroendocrine tumors, including meningiomas. 111 Indium-pentreotide is most sensitive in well-differentiated meningiomas that express high levels of somatostatin receptors but is less useful in dedifferentiated, atypical meningiomas.
SPECT imaging is of limited sensitivity for tumors smaller than approximately 10 to 15 mm compared with 7 to 10 mm for FDG-PET. However, if lesions meet this size threshold, SPECT imaging is as sensitive as FDG-PET. SPECT is particularly advantageous in tumors that are not characteristically FDG avid, including some lobular breast cancers and renal cell carcinomas, and in lesions situated near the physiologically active cortex. These tracers are a useful part of the clinical portfolio and will endure as long as PET tracers beyond FDG remain strictly investigational.
Lastly, it should be mentioned that whole-body PET/CT is at times a useful adjunct to brain imaging in patients presenting with a new primary brain tumor versus brain metastasis. Diagnostic biopsy in a more easily accessible site elsewhere in the body may help to preclude a more invasive intracranial biopsy.
Tumor grading with FDG
FDG is useful for tumor grading because most high-grade tumors, such as high-grade gliomas, medulloblastoma, and primary central nervous system lymphoma, have high concentrations and activity of GLUTs. Most low-grade tumors have lower concentrations of GLUTs and can be distinguished from high-grade gliomas by the lower FDG uptake on PET. FDG avidity of common benign and malignant primary and metastatic brain tumors is shown in Box 1 .
High
High-grade gliomas (grades III and IV)
Medulloblastoma
Primary central nervous system lymphoma
Pineoblastoma
Central neurocytoma
Pilocytic astrocytoma
Pituitary adenoma
Variable
Metastases
Oligodendroglioma
Meningioma
Low
Most low-grade gliomas
Primitive neurectodermal tumor
Vestibular schwannomas
False positive
Infection
Inflammation
Seizure activity
Malignant tumors are given in italics.
As an exception to the rule, some low-grade tumors have a high FDG avidity, which is because of the high concentrations of GLUTs. In pilocytic astrocytomas, for example, the vascular density is high. In this case, glucose uptake does not represent the blood flow but is thought to represent the metabolic activity of the endothelial cells lining the tumor vasculature. Endothelial proliferation and metabolism account for the high FDG avidity. Oligoastrocytomas also have variable FDG uptakes, even within the same tumor grade.
Measurements of tumor FDG uptake versus physiologic brain FDG uptake can be used to predict high- versus low-grade brain tumors, both primary and metastatic, with certain exceptions as noted earlier. Semiquantification of glucose uptake is performed using a standardized uptake value (SUV). A tumor to gray matter uptake ratio of greater than 0.6 or a tumor to white matter uptake ratio of greater than 1.5 was, in 1 series by Delbeke and colleagues, 100% sensitive and 67% specific for distinguishing high-grade gliomas from low-grade gliomas. Applying the same cutoffs to primary brain tumors and brain metastases yielded 94% sensitivity and 77% specificity. FDG avidity has also been inversely correlated with survival. For example, in a series of 165 patients with highly FDG avid brain tumors, 1-year survival was only 29%, as opposed to 94% in patients with tumors with low FDG avidity.
Recurrence versus Postradiation Necrosis with FDG
In primary or metastatic tumors treated with radiotherapy, tumor growth as seen on contrast MRI is often indistinguishable from postradiation necrosis, both appearing as enlarging, contrast-enhancing lesions. Approximately 70% of radiation necrosis cases are diagnosed in the first 2 years following radiation treatment, which is similar to the expected time frame for tumor recurrence. The incidence of radiation necrosis increases with the radiation dose and is particularly high in cases in which chemotherapy is concurrently administered. FDG-PET is a useful tool in distinguishing postradiation necrosis from tumor progression, both in high-grade gliomas and brain metastases. In general, recurrent tumor is FDG avid ( Fig. 2 ), and radiation necrosis is not FDG avid. In the first 3 to 6 months after radiation, FDG-avid radiation-induced inflammatory cells such as macrophages can have the same appearance as a tumor. Inflammatory cells are seen in the setting of edema, neovascularity, demyelination, and necrosis, which begin to occur in the first hours to months after radiation treatment. As discussed in later sections, techniques such as dual-phase FDG-PET imaging and PET imaging with investigational radiotracers can help to improve the accuracy of PET imaging by increasing the target to background ratio.
The sensitivity of FDG-PET for distinguishing tumor recurrence (primary and metastatic) ranges from 80% to 86%, with a specificity of 40% to 88%. In some centers, PET images are fused to MRI and delayed-phase PET images are obtained. Both methods have been shown to improve the accuracy of FDG-PET. Fusion of FDG-PET images to contrast MRI images is helpful for anatomic correlation and helps to distinguish uptake within a lesion from uptake in an adjacent, functionally active gray matter. This ability may be helpful in cases in which the rounded contour of a rim-enhancing lesion may look similar to the contour of a gyrus, which closely surrounds the lesion, as in Fig. 1 .
In one series, the sensitivity of FDG-PET in assessing for recurrent brain metastasis versus radiation necrosis without fusion to MRI was 65% and 85%, respectively, and after PET/MRI fusion, sensitivity increased to 86% and specificity to 80%.
Dual-Phase FDG-PET Imaging
Dual-phase FDG-PET imaging has been shown to increase accuracy in distinguishing recurrence from postradiation necrosis in gliomas and brain metastases. This method enhances the tumor to background ratio and is based on differences in FDG tracer kinetics between tumor and normal brain parenchyma. When FDG-PET is normally performed at 45 to 60 minutes after tracer injection, FDG tracer kinetics are similar between tumor and gray matter. In other words, transport of the tracer through the blood-brain barrier and into the cells is similar for the tumor and normal brain ( Fig. 3 ). Phosphorylation of the tracer by glucose 6-hexokinase, which traps the FDG intracellularly, may be higher in the tumor than in normal gray matter because tumors may have abnormally high levels of glucose 6-hexokinase. Dephosphorylation via glucose 6-phosphatase is higher in normal brain parenchyma than in tumor. Over time, trapping and retention of the tracer by the tumor is greater than that of normal brain parenchyma. This process is most pronounced at 3 to 6 hours after tracer injection.
Related posts:
Special Surgical Considerations for Functional Brain Mapping
An Introduction to Diffusion Tensor Image Analysis
Multimodal Image Registration for Preoperative Planning and Image-Guided Neurosurgical Procedures
Preoperative Prediction of Verbal Episodic Memory Outcome Using fMRI
Identification of Neural Targets for the Treatment of Psychiatric Disorders: The Role of Functional Neuroimaging
Development of a Clinical Functional Magnetic Resonance Imaging Service
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
