Imaging in Neurooncology

Abstract

Summary: Imaging in patients with brain tumors aims toward the determination of the localization, extend, type, and malignancy of the tumor. Imaging is being used for primary diagnosis, planning of treatment including placement of stereotaxic biopsy, resection, radiation, guided application of experimental therapeutics, and delineation of tumor from functionally important neuronal tissue. After treatment, imaging is being used to quantify the treatment response and the extent of residual tumor. At follow-up, imaging helps to determine tumor progression and to differentiate recurrent tumor growth from treatment-induced tissue changes, such as radiation necrosis. A variety of complementary imaging methods are currently being used to obtain all the information necessary to achieve the abovementioned goals. Computed tomography and magnetic resonance imaging (MRI) reveal mostly anatomical information on the tumor, whereas magnetic resonance spectroscopy and positron emission tomography (PET) give important information on the metabolic state and molecular events within the tumor. Functional MRI and functional PET, in combination with electrophysiological methods like transcranial magnetic stimulation, are being used to delineate functionally important neuronal tissue, which has to be preserved from treatment-induced damage, as well as to gather information on tumor-induced brain plasticity. In addition, optical imaging devices have been implemented in the past few years for the development of new therapeutics, especially in experimental glioma models. In summary, imaging in patients with brain tumors plays a central role in the management of the disease and in the development of improved imaging-guided therapies.

GENERAL AND MOLECULAR ASPECTS

Primary brain tumors may arise form various cell types of the brain including glial cells, neurons, neuroglial precursor cells, pinealocytes, the meninges, choroid plexus, pericytes of the vessels, cells of the hypophysis, and lymphocytes. The incidence of primary brain tumors varies between subtypes. The most common primary brain tumors in adults are gliomas and meningiomas. For gliomas, the incidence is six to eight in 100,000, with approximately 50% belonging to malignant subtypes. Lower-grade gliomas tend to occur in younger patients, whereas higher-grade tumors are more frequent in older patients. Gliomas are divided histologically into astrocytomas, oligodendrogliomas, mixed gliomas, ependymal tumors, and tumors of the choroid plexus. Grading is performed according to the World Health Organization (WHO) criteria, taking into account the presence of nuclear changes, mitotic activity, endothelial proliferation, and necrosis.^1,2 Glioblastoma, corresponding to WHO grade IV, is the most fatal and most common primary brain neoplasm with an incidence of three to four in 100,000. Approximately 50% of all gliomas and 20% of all primary intracranial tumors are glioblastomas. Together with all intracranial neoplasms, the glioblastoma is the second most common cause of death due to an intracranial disease after stroke. Despite aggressive multimodal treatment strategies (surgery, radiation, chemotherapy) median survival of patients with gliomas is limited, depending on grade and age at diagnosis varying from 1 year for glioblastoma, to 2–3 years for grade III and to 5–10 years for a grade II glioma.

A better understanding of glial tumorgenesis is crucial for the development of specific molecular therapeutic targets to overcome current therapeutic limitations. A complex series of molecular changes occurs, which results in dysregulation of the cell cycle, alterations of apoptosis and cell differentiation, in neovascularization as well as tumor cell migration and invasion into the brain parenchyma. Genetic alterations which play an important role in glioma development include a loss, mutation or hypermethylation of the tumor suppressor gene TP53 or other genes involved in the regulation of the cell cycle, such as cyclin-dependent kinase N2A/p16, p14ARF and primitive neuroectodermal tumor (PTEN), as well as activation or amplification of oncogenes and growth factors and/or their receptors, such as MDM2, cyclin-dependent kinase 4, cyclin D1 and D3, epidermal growth factor receptor (EGFR), vascular endothelial growth factor (VEGF), PDGFR [platelet-derived growth factor (PDGF) receptor], and transforming growth factor-β.^3–5 During progression from low-grade astrocytoma (WHO grade II) to anaplastic astrocytoma (WHO grade III) and to glioblastoma multiforme (WHO grade IV) a step-wise accumulation of genetic alterations occurs. Whereas TP53 mutation and PDGF and PDGFR-α overexpression represent early changes during low-grade glioma development, progression to anaplastic astrocytoma is associated with pRB alteration and loss of heterozygocity (LOH) of 19q, further malignant progression to glioblastoma including LOH 10q and mutations of the PTEN gene.⁶ These secondary glioblastomas, which develop from better differentiated astrocytomas, can be distinguished from primary de novo glioblastomas on the basis of molecular genetic findings⁷ with amplification and/or overexpression of the EGFR, p16 deletion, PTEN mutation, pRB alteration, and LOH 10p and 10q associated with primary glioblastoma. Most importantly, molecular alterations have been identified, which indicate therapeutic response of patients and, thus, are prognostically relevant: anaplastic oligodendrogliomas with LOH 1p and/or LOH 19q are characteristically sensitive to PCV (procarbazine, lomustine, and vincristine) chemotherapy, and patients' survival is significantly prolonged.^8–11

An overview of the tumors in the cranial cavity is given in Table 1.^12–50

TABLE 1.

Overview of Tumors in the Cranial Cavity

Tumor Entities (% of all primary brain tumors)	MRI/CT	PET		Ref.
		FDG*	Methionine†
1. Gliomas
Pilocytic astrocytoma WHO I° (<3%)	Cystic tumor with focal contrastenhancement	Variable, focally increased	Up to 2-fold	18
				21
Astrocytoma WHO II° (<5%)	T1: slightly hypointense T2: hyperintense	Decreased	1- to 2-fold	See text
Anaplastic astrocytoma WHO III° (<5%)	T1: hypointense T2: hyperintense Contrast enhancement and perifocal edema	Variable	2- to 3-fold
Glioblastoma WHO IV° (20–25%)	Irregular tumor border T1: central necrosis hypointense T2: perifocal edema hyperintense contrast enhancement	Increased	>2.5-fold
Oligodendroglioma WHO II°/III° (<5 %)	Inhomogenous tumor with focal contrast enhancement and calcifications on CCT in 70–90%	Decreased/increased	>2.5-fold
Oligoastrocytoma WHO II°/III° (<5%)		Decreased/increased	2- to 3-fold
Ependymomas (2–3%)	Characteristic localization in IVth ventricle or intramedullary; heterogenous, cystic, hemorrhages	Decreased	1.3- to 2.7-fold	40
				50
				36
Choroid plexus papilloma (<1%)	Characteristic localization in ventricles; sharp tumor border; gross contrast enhancement	N.A.	N.A.
Gliomatosis cerebri	diffusely infiltrating; hyperintense (T2, FLAIR)	N.A.	N.A.
2. Neuronal and glioneuronal tumors
Dysembryoplastic neuroepithelial tumor (<1%)	Multicystic subcortical tumors with focal contrast enhancement	Decreased benzodiazepine receptor density as possible reason for epileptogenicity		45
				27
				43
Dysplastic gangliocytoma (<1%)		Increased	Increased
Ganglioglioma (<1%)	Cortical localization, solid or cystic with calcifications and little contrast enhancement	Variable, depending on WHO grade	N.A.	26
				44
				34
Central neurocytoma (<1%)	Sharp tumor border, inhomogenous with cysts, necroses, calcifications, positive contrast enhancement	Increased, depending on proliferative activity	increased	35
				41
3. Tumors of the pineal gland (<1%)
Pineocytoma		N.A.	N.A.
Pineoblastoma	Contrast enhancing	Increased	N.A.	12
Germinoma	Isointense with contrast enhancement	N.A.	N.A.
4. Embryogenic tumors
Medulloblastoma (20–25% <15 y.o.; 1% >20 y.o.)	Cystic tumor in 4th ventricle (75%) or cerebellum (25%); T1: hypointense; T2: hyperintense; moderate contrast enhancement	Strongly increased	Increased	24
Primitive neuroecto-dermal tumors (PNET)		Decreased; relatively increased in spinal localization	N.A.	24
				33
5. Meningeal tumors
Meningiomas (25–30%)	Localization: in 50% convexity, parasagital or falx; other localizations: sphenoid bone (15–20%), parasellar (5–10%), frontobasal (5–10%), infratentorial (10%)	Variable (0.2- to 1.8-fold)	Increased (1.3- to 3.6-fold)	14
				31
				17
				25
				38
	CT: hyperdense, calcifications (25%)	[68GA]DOTATOC-PET detects somatostatin receptor expression in meningiomas		20
	MRI: T1 isointense; T2 hypointense; infiltration of bone; dural artery; dural tail sign
	CT/MRI: contrast enhancement and peritumoral edema
Hemangiopericytoma(<0.5%)		Decreased	Increased	28
6. Tumors of the regionof the sella
Craniopharyngioma (<2%)	Intrasellar (30%) and suprasellar (70%); cystic, calcifications, contrast enhancement	-[18F]FDG-PET variable depending on histological type		37
				14
Adenomas of thehypophysis (5–8%)	Microadenomonas on T1 hypo- or isointense; slow contrast enhancement; macroadenomas on T1 and T2 isointense with strong contrast enhancement	-Specific binding to D2-receptors on[18 F]FESP-PET differentiates adenomas of hypophysis from perisellar meningiomas and craniopharyngiomas;		32
		-Specific increase in monoaminooxidase activity on [11C]Deprenyl-PET differentiates adenomas of hypophysis from perisellar meningiomas by specific increased monoaminooxidase activity
7. Tumors of cranialnerves
Neurinoma (6–8%)	Cranial nerve VIII (90%), others: V, VII, IX–XII; sharp tumor border, T1 hypointense, T2 hyperintense, strong contrast enhancement	Iso- orhypo-metabolic	Only slight increase	15
				16
				39
				49
8. Lymphomas
Primary lymphoma of the CNS(2–5%)	Localization around ventricles and in basal ganglia; in 50% multiple sites; on T1 and T2 isointense, homogenous contrast enhancement, perifocal edema	Increased; FDG-PET allows differential diagnosis from toxoplasmosis	Increased	48
				22
				46
				13
9. Metastatic tumors (≈20%of all brain tumors)
Lung, breast,melanoma,gastrointestinal, hypernephroma	Supratentorial (90%) localized at inner border of cortex; infratentorial (10%); 50–70% multiple; 30–50% single; T1 hyperintense, T2 hypointense, ring-like contrast enhancement	Variable; screening for metastasis with [18F]FDG is not recommended	N.A.	19
				29
				30
				42
				47
		[68GA]DOTATOC-PET detects somatostatin receptor positive metastasis of carcinoidtumors		23

^*= [¹⁸F]FDG in comparison to cortical cerebral metabolic rate of glucose (CMRGlc).

^†= [¹¹C]MET in comparison to contralateral control region.

N.A. = Not available.

[¹⁸F]FESP = [¹⁸F]fluoro-ethyl-spiperone.

[⁶⁸GA]DOTATOC = [⁶⁸GA]-1,4,7,10-tetraazacyclododecan-N,N′,N′′,N′′′-tetraacetacid-D-Phe-Tyr-octreoid.

IMAGING FOR PRIMARY DIAGNOSIS

Cranial computed tomography (CT) and magnetic resonance imaging (MRI) with and without contrast media are widely used for primary diagnosis of brain tumors. Standard T1- and T2-weighted MRIs detect brain tumors with high sensitivity. Beside primary information on the size and localization of the tumor, especially MRI provides additional information about secondary phenomena such as mass effect, edema, hemorrhage, necrosis, and signs of increased intracranial pressure at high spatial resolution and with high tissue contrast. A set of various MRI acquisitions parameters, like T1-, T2-, proton-, diffusion-, and perfusion-weighted images as well as fluid attenuated inversion recovery (FLAIR) sequences give a characteristic pattern of each tumor depending on tumor type and grade.

Most brain tumors are hypointense on T1-weighted images and hyperintense on FLAIR, T2-, and proton-weighted images. Highly proliferative active tumors such as glioblastomas lead to a destruction of the blood-brain barrier (BBB) with subsequent leakage of contrast media (FIG. 1), which is being used for diagnostic purpose in CT and T1-weighted MRI. In contrast, low-grade tumors usually have no or minimal enhancement. The contrast-enhancing lesion (T1 + Gd) corresponds histologically to a hypercellular region with neovascularization, a central hypointense area (T1) is mainly caused by tumor necrosis. Already in the CT-era, biopsies from signal changes in areas surrounding the contrast-enhancing tumor revealed the presence of migrating tumor cells. The tumor volume measured as the volume of T2 hyperintensity is the strongest predictor of overall survival in patients with supratentorial diffuse astrocytoma WHO °II and the only predictor of malignant progression.⁵¹

FIG. 1.

Parameters of interest in the noninvasive diagnosis of brain tumors. Alteration of the blood-brain barrier and the extent of peritumoral edema are detected by MRI. Signs of increased cell proliferation can be observed by means of multi-tracer PET imaging using [¹⁸F]FDG, [¹¹C]MET, and [¹⁸F]FLT as specific tracers for glucose consumption, amino acid transport and DNA synthesis, respectively. Secondary phenomena, such as inactivation of ipsilateral cortical cerebral glucose metabolism, may be observed ([¹⁸F]FDG) and are of prognostic relevance. Gd = gadolinium. Reproduced with permission from Jacobs AH. PET in gliomas. In: Neuroonkologie (Schlegel U, Weller M, Westphal M, eds), pp 72–76. Copyright © 2003, Thieme-Verlag. All rights reserved.¹⁸⁶

Studies on the value of diffusion-weighted imaging (DWI) MRI are ongoing. DWI is able to characterize morphological features including edema, necrosis, and viable tumor tissue by measuring differences in the apparent diffusion coefficient (ADC). DWI might be able to detect areas of tumor infiltration which are not visible on other MRIs.⁵² Furthermore, DWI seems to be useful in providing a greater degree of confidence in distinguishing brain abscesses from cystic or necrotic brain tumors than conventional MRI.⁵³

Moreover, dynamic contrast-enhanced MRI is a new and promising imaging tool for measuring physiological tumor properties (e.g., microvascular permeability and plasma volume; FIG. 2). The vascular hyperpermeability of tumor vessels for macromolecular solutes yields a proteinaceous exudate within the tumor interstitium that is considered a favorable milieu for the in-growth of new capillary buds. Changes in tumor vessel permeability and tumor volume as assessed by dynamic contrast enhanced perfusion MRI serve as surrogate marker for angiogenesis and allow the prediction of pathologic tumor grade^54,55 with high sensitivity but limited specificity.^56–58

FIG. 2.

Representative image of an axial T₁-weighted postcontrast sequence (a), a corresponding color coded “relative enhancement map” of a dynamic contrast enhanced three-dimensional T₁-weighted sequence (b), and signal intensity (SI) curves of different tumor areas (c) in a patient with glioblastoma multiforme. Areas with a strong uptake of contrast media show high SI values during the first 2 min with a subsequent wash out phenomena (blue region of interest and blue curve), which is indicative of a substantial microvascular leak with progressive accumulation of contrast agent in the tumor interstitial space. Areas with lower microvascular permeability values show a less pronounced tumor enhancement (pink region of interest and pink curve), whereas necrotic tumor areas only show a minor uptake of contrast media (black region of interest and black curve). Tumor heterogeneity and different areas of microvascular permeability within an individual tumor mass are characteristic findings for malignant tumors and are only visible on dynamic imaging sequences (b) and not on conventional MRIs (a).

It should be pointed out that the different tissue compartments within a glioblastoma giving rise to different magnetic resonance signals have different gene and protein expression patterns.⁵⁹ Thus, there is potential for imaging-guided proteomics and microarray analysis to identify specific markers of tumor behavior.

Because intratumoral heterogeneity of brain tumors is not adequately reflected in conventional MRI because evaluation of the contrast enhancing lesion can either under- or overestimate the presence of active tumor, magnetic resonance spectroscopy (MRS) and positron emission tomography (PET) are being performed to gain additional information on metabolic and molecular tumor markers. MRS gives additional information on the real extent of the tumor and on tissue metabolites, such as N-acetylaspartate (NAA), creatine, choline, and lactate.⁶⁰ The increase of choline-containing compounds and of NAA appears to correlate best with the degree of tumor infiltration.^61,62 The appearance of creatine differentiates gliomas from metastasis, which generally lack creatine.⁶³ An improved automated MRS analysis approach (nosologic imaging) enables correct differentiation between low-grade glioma, high-grade glioma, meningiomas, metastasis, necrosis, and healthy tissue in up to 90% of cases and shall facilitate a noninvasive diagnosis of lesion type.^64,65 Disadvantages of MRS include its low spatial resolution, which cannot fully address the anatomical and contrast heterogeneity of brain tumors observed with MRI.

The role of PET has been primarily investigated in patients with gliomas as the most frequent and most difficult to treat primary brain tumors. PET reveals highly specific quantitative information on the metabolic state of gliomas.^66–68 PET allows the quantitative localization of expression of endogenous or exogenous genes coding for enzymes or receptors by measuring the accumulation or binding of the respective enzyme substrates or receptor binding compounds.^69–71 Depending on the radiotracer, various molecular processes can be visualized by PET, most of them relating to an increased cell proliferation within gliomas (FIG. 1). Radiolabeled 2-[¹⁸F]fluoro-2-deoxy-d-glucose ([¹⁸F]FDG), methyl-[¹¹C]-L-methionine ([¹¹C]MET) and 3′-deoxy-3′-[¹⁸F]fluoro-l-thymidine ([¹⁸F]FLT) are taken up by proliferating gliomas depending on their tumor grade as a reflection of increased activity of membrane transporters for glucose, ([¹⁸F]FDG), amino acids ([¹¹C]MET), and nucleosides ([¹⁸F]FLT) as well as increased expression of cellular hexokinase ([¹⁸F]FDG) and thymidine kinase ([¹⁸F]FLT) genes, which specifically phosphorylate [¹⁸F]FDG and [¹⁸F]FLT, respectively.

[¹⁸F]FDG-PET depicts the rate of glucose uptake and has been used to detect the metabolic differences between normal brain tissue, low-grade and high-grade gliomas, and radionecrosis.^72–74 Increased intratumoral glucose consumption correlates with tumor grade, cell density, biological aggressiveness, and survival of patients in both primary and recurrent gliomas.^74–79 In general, low-grade tumors have a metabolic activity similar to white matter and higher-grade tumors, similar to gray matter (FIG. 3). A tumor-to-white matter ratio greater than 1.5 and tumor-to-gray matter ratio less than 0.6 were found to be indicative of high grade tumors with high sensitivity (94%) and limited specificity (77%).⁸⁰ Relatively benign tumors with a high FDG uptake included pilocytic astrocytoma and ganglioglioma. Pilocytic astrocytomas have a good prognosis despite exhibiting high FDG uptake and positive contrast enhancement (MRI) due to the presence of metabolically active fenestrated endothelial cells. The limited differentiation between tumor and normal gray matter by [¹⁸F]FDG-PET may be improved by scanning at delayed intervals 3–7 h after tracer injection.⁸¹ However, due to the relative high background levels of cortical glucose consumption more specific radiotracers for glioma diagnosis were developed.

FIG. 3.

Noninvasive differentiation between low- and high-grade glioma. In low-grade gliomas (WHO II°) glucose metabolism is similar to white matter (arrows) and amino acid uptake is only moderately increased. In high-grade gliomas (GBM; WHO °IV), both glucose metabolism and amino acid uptake are increased. Reproduced with permission from Jacobs AH. PET in gliomas. In: Neuroonkologie (Schlegel U, Weller M, Westphal M, eds), pp 72–76. Copyright © 2003, Thieme-Verlag. All rights reserved.¹⁸⁶

The radiolabeled amino acids methyl-[¹¹C]-l-methionine ([¹¹C]MET), [¹¹C]-tyrosine, [¹⁸F]fluoro-tyrosine and O-(2-[¹⁸F]-fluoroethyl)-l-tyrosine have been shown to be more specific tracers for tumor detection and tumor delineation due to their low uptake in normal brain.^{21,78,82–86} The increased methionine uptake (factor of 1.3–3.5 in comparison with a contralateral control region) is related to increased transport mediated by type L amino acid carriers.^87,88 [¹¹C]MET-uptake correlates to cell proliferation in vitro, the expression of Ki-67 and proliferating cell nuclear antigen, as well as to microvessel density, explaining its role as a marker for active tumor proliferation.^82,89,90 The intensity of [¹¹C]MET-uptake differentiates between WHO II° and WHO III°/IV° gliomas (FIG. 3).^21,86 Uptake is increased not only in solid parts of the tumor but also in the infiltration area.^91,92 In 80% of gliomas WHO II° [¹¹C]MET-uptake is greater than 1.5-fold of the normal brain tissue,²¹ whereas glucose metabolism is reduced compared with gray matter. Most studies indicated that [¹¹C]MET-uptake is inversely correlated to prognosis,^93–95 but due to significant [¹¹C]-MET uptake also in most low-grade gliomas this relation is less close than with FDG. For WHO III°/IV° gliomas increased [¹¹C]MET-uptake is directly correlated to increased [¹⁸F]FDG uptake^78,96 and exceeds the area of involvement depicted by Gd-enhanced MRI.⁹⁷ It should be pointed out that increased [¹¹C]MET uptake also depends on tumor type, with oligodendrogliomas accumulating more radiotracer than astrocytomas from the same histological grade.^21,78,98 Disadvantages of [¹¹C]methionine are its uptake in acutely ischemic and inflammatory brain tissue⁹⁹ as well as its short half-life.

The third parameter, which can be noninvasively assessed by PET is the incorporation of nucleosides into DNA in proliferating cells. Radiolabeled thymidine ([³H]TdR) is the gold standard for determination of cell proliferation in cell culture, and to date ¹¹C- and ¹⁸F-labeled thymidine compounds have been radiosynthesized to allow a noninvasive assessment of tumor proliferation as well as early response to chemotherapy by PET. 3′-deoxy-3′-[¹⁸F]fluoro-l-thymidine ([¹⁸F]FLT) is stable in vivo and has been used for the evaluation of tumor proliferation primarily in extra-cranial tissues. Unpublished results in patients with gliomas indicate that [¹⁸F]FLT is a promising tracer to study glioma proliferation especially in areas with high [¹⁸F]FDG background. Relative [¹⁸F]FLT uptake within gliomas is greater than relative [¹¹C]MET uptake (FIG. 1), indicating the possible role of [¹⁸F]FLT as a more specific tumor marker than [¹⁸F]FDG and [¹¹C]MET.^100–103 The overall goal is to be able to quantify chemotherapeutic effects early (within days to weeks).

A summary of MRI and PET findings in human brain tumors are given in Table 1.

IMAGING FOR PLANNING RESECTION AND RADIATION THERAPY INCLUDING DELINEATION FROM FUNCTIONAL NEURONAL TISSUE

MRI is the method of choice for tumor localization. Delineation of the tumor in three dimensions (sagittal, coronal, and axial) allows the selection of the best operative procedure. CT is being used for detection of calcifications in oligodendrogliomas, meningiomas, or craniopharyngiomas and for tumors that are located at the base of the skull. Magnetic resonance angiography is being used alone or together with conventional angiography to study the location of the tumor with respect to its vascularization. In selected patients and in specialized institutions, MRS and PET are being used in conjunction with MRI to define the real extend of the tumor.^91,104–107 Most importantly, tumors that are located in eloquent areas require preoperative functional imaging by functional MRI (fMRI) or functional PET (FIG. 4). ^108–115 These combined imaging procedures are especially important in those patients where tumor growth has led to changes within the neuronal network as a result of functional brain plasticity. These changes consist either of a displacement of functional important neuronal tissue (like language areas) surrounding the lesion or in the recruitment of new brain areas that are usually not involved in the performance of a certain task.^114–116 Functional imaging studies alone can provide information about which brain areas are involved in the performance of a certain behavioral task but can not answer the diagnostically important question of whether a certain brain area is essential for that task. This additional information can be obtained by combining functional brain images with transcranial magnetic stimulation, an electrophysiological method to temporarily interfere with normal brain function^116a. A distance of 1 cm or more between the functional relevant cortex, as delineated by fMRI, and the tumor significantly reduces the risk of postoperative loss of function.¹¹⁷ The evaluation of the cerebral reorganization of motor function is an essential step in predicting the risk of motor deficits in patients with an indication for operative treatment.¹¹⁸ Only these combined imaging technologies will allow to maximize the spatial extent of tumor treatment and to simultaneously preserve functional relevant tissue.¹¹⁹ Moreover, metabolic or molecular information derived from PET or SPECT studies is being used in some institutions for the exact planning for radio-¹⁰⁵ and gene therapy.^120,121 This is of special importance because tumor volume as depicted by contrast-enhanced MRI is always smaller than the tumor volume as depicted by [¹¹C]MET-PET.⁹⁷ Whether [¹¹C]MET-PET-guided extended fields of radiation have a significant influence on time to progression and overall survival has to be proven in future studies.

FIG. 4.

Preoperative differentiation of tumor tissue from functionally important neuronal tissue through multimodal and multitracer imaging. These combined imaging procedures shall guide the neurosurgeon to operate as much tumor as possible but at the same time to leave the functionally important tissue intact. Reproduced with permission from Jacobs et al. Molecular imaging of gliomas. Mol Imaging 1:309–355. Copyright © 2002, MIT Press Journals. All rights reserved.⁶⁷

In recent years, diffusion tensor white matter fiber tracking and intraoperative MRI have become promising tools to guide the neurosurgeon intraoperatively. These techniques allow to maximize tumor resection without additional morbidity^122,123 by avoiding lesions to fiber tracts like the pyramidal tract.^124–126

IMAGING FOR PLANNING STEREOTAXIC BIOPSY AND TARGETED APPLICATION OF THERAPEUTICS

Brain tumors may consist of different parts, which are heterogenous with respect to tumor grading; thus, low- and high-grade areas may be present within the same tumor. Stereotaxic biopsy aims at the tumor sites with the highest tumor grade. Therefore, suitable targets for biopsy will have positive contrast enhancement on T1-weighted MRI, a high choline-peak on MRS, hypermetabolism on [¹⁸F]FDG-PET,^127–129 and high accumulation of [¹¹C]methionine.^92,130,131 Moreover, in the development of new experimental therapies, multimodal imaging procedures help identify the most active tumor parts for targeted placement of local infusion catheters (FIG. 5). In a phase I clinical gene therapy trial, imaging the expression of endogenous genes by [¹⁸F]FDG- and [¹¹C]MET-PET as direct measures for the respective gene expression and as surrogate markers for proliferation and tumor cell density was used to identify the biological active tumor portion as proper target tissue and to measure response to gene therapy (FIG. 5).^120,121

FIG. 5.

Multimodal imaging for the establishment of imaging-guided experimental treatment strategies. Coregistration of [¹⁸F]FIAU-, [¹¹C]MET-, [¹⁸F]FDG-PET and MRI before (left column) and after (right column) targeted application (stereotactic infusion) of a gene therapy vector. The region of specific [¹²⁴I]-FIAU retention (68 h) within the tumor after LIPO-HSV-1-tk transduction (white arrow) resembles the proposed tissue dose of vector-mediated gene expression and shows signs of necrosis (cross right column; reduced methionine uptake [MET] and glucose metabolism [FDG]) after ganciclovir treatment. Reproduced with permission from Jacobs et al. Positron-emission tomography of vector-mediated gene expression in gene therapy for gliomas. Lancet 358:727–729. Copyright © 2001, Elsevier Limited. All rights reserved.¹²⁰

In trials employing convection-enhanced delivery of local chemotherapeutics, DWI is being used to assess the convective process and routine diagnostic MR imaging to identify the tumor response.¹³²

IMAGING FOR DETERMINATION OF TREATMENT EFFECT, TUMOR PROGRESSION, AND DIFFERENTIATION OF RECURRENT TUMOR FROM RADIATION NECROSIS

The effects of treatment should ideally be visualized with the same imaging parameters that have been used before therapy (FIG. 5). However, there are several limitations inherent in each imaging modality. On contrast-enhanced MRI, residual tumor and postsurgical changes can both result in abnormal enhancement. Therefore, MRI cannot be used postoperatively after day 3 and for several weeks because the surgical damage of the BBB, with subsequent leakage of contrast media, leads to a false-positive indicator of the presence of residual or recurrent tumor. Moreover, conventional MRI techniques usually fail to detect early effects of radio- and chemotherapy because individual treatment effects are only visible after more than 12 months,^133–135 with a substantial interobserver variability in the assessment of treatment response.¹³⁶ Especially after the application of biologically active agents (gene therapy vectors, toxins), the value of conventional MRI to detect therapy-specific changes of tumor viability is limited¹³⁷ as reviewed previously.¹³⁸

In contrast, dynamic contrast enhanced MRI, as a surrogate marker for angiogenesis, is useful for monitoring antiangiogenic therapies in brain tumors.¹³⁹ Moreover, diffusion-weighted MRI detects therapy-induced water diffusion changes and has been suggested to provide an early surrogate marker for quantification of treatment response.¹⁴⁰ It was found that low values for the ADC indicating high tissue viability imply better response to radiotherapy, whereas high ADC values indicating necrosis correlate with poorer response.¹⁴¹ Assessment of ADC ratios from tumor and contralateral control regions were also useful in the differentiation of radiation effects (high ADC ratios) from tumor recurrence or progression (low ADC ratios).¹⁴² However, it should be kept in mind that dexamethasone treatment significantly reduces the diffusivity of edematous brain,¹⁴³ thus confounding the interpretation of DWIs.

Because MRS can reliably differentiate pure tumor, pure necrosis, and normal tissue, specific changes in tumor metabolite levels as detected by MRS may be predictive for the effectiveness of experimental treatment strategies.¹⁴⁴ However, MRS alone may not be particularly helpful because most patients have mixed histological findings comprised of necrosis and tumor giving rise to inconclusive findings. In contrast, progression from low-grade to high-grade gliomas leads to a characteristically increased concentration of choline and a reduced NAA peak with high diagnostic accuracy.

Due to the relatively high cortical background activity, [¹⁸F]FDG-PET is not suited to detect residual tumor after therapy.^145,146 Similar to structural imaging, the effects of radio- and chemotherapy can be visualized by [¹⁸F]FDG-PET only after several weeks¹⁴⁷ with a possible transient increase of [¹⁸F]FDG-uptake in the initial phase which is most likely due to infiltration of macrophages consuming [¹⁸F]FDG.^148–150 At further follow-up, however, recurrent tumor and progression from low-grade to high-grade glioma can be visualized by a newly appearing hypermetabolism.^151,152 [¹⁸F]FDG-PET has a sensitivity of 75% and a specificity of 81% for the detection of recurrent tumor versus radiation necrosis.¹⁵³ Moreover, in patients after stereotactic radiotherapy for brain metastasis, coregistration of [¹⁸F]FDG-PET images with MRI yields an improvement of the sensitivity for the detection of recurrent tumor from 65–86%. Disadvantages of [¹⁸F]FDG-PET include accumulation of [¹⁸F]FDG in macrophages that may infiltrate the sites having received radiation therapy. Therefore, radiation necrosis may be indistinguishable from recurrent tumor. It should be noted that in patients receiving corticosteroids as symptomatic treatment evaluation of [¹⁸F]FDG-PET may be hampered by a reduced cortex-to-white matter ratio.¹⁵⁴

[¹¹C]MET-PET in contrast is much better suited to follow the effects of radiation therapy, which show as a reduction of relative methionine-uptake,¹⁴⁶ which may also be observed in animal models.¹⁴⁹ Most importantly, [¹¹C]MET-PET successfully differentiates between recurrent tumor and radiation necrosis (FIG. 6) ¹⁵⁵ with the detection of recurrent tumor at high sensitivity and high specificity.^91,156,157 Similar results have been obtained with other tracers for amino acid transport, such as O-(2-[¹⁸F]fluoroethyl)-l-tyrosine and 3-[¹²³I]iodo-α-methyl-l-tyrosine with a reliable differentiation between post-therapeutic benign lesions and tumor recurrence after treatment of low- and high-grade tumors.^158,159

FIG. 6.

Differentiation between recurrent tumor and radiation necrosis. Biopsy of this clinically worsening tumor, taken from the region with positive magnetic resonance contrast enhancement, evidenced only necrosis. However, a second biopsy from the area of increased amino acid uptake (arrowheads) revealed the findings of recurrent tumor. Reproduced with permission from Thiel et al. Enhanced accuracy in differential diagnosis of radiation necrosis by positron emission tomography-magnetic resonance imaging coregistration: technical case report. Neurosurgery 46:232–234. Copyright © 2000, Lippincott Williams & Wilkins. All rights reserved.¹⁵⁵

IMAGING IN EXPERIMENTAL BRAIN TUMOR MODELS

Imaging studies in experimental brain tumor models over the past 10 years aimed toward 1) the development of new radiotracers for cellular proliferation and protein synthesis, 2) characterization of these tracers with respect to their ability to detect responses to radio- and chemotherapy at a relatively early stage, 3) strategies for imaging transcriptional regulation and migration of tumor cells, and 4) imaging the expression of exogenous genes carrying a marker or therapeutic function and introduced into experimental gliomas for the purpose of developing improved gene therapeutic vectors. These experimental strategies have been reviewed in detail previously.⁶⁷

New developments aim toward 1) the detection of tumor cell migration in vivo,¹⁶⁰ 2) the establishment of in vivo assays for direct imaging of tumor-specific signal transduction pathways (e.g., p53-, E2F-1 and HIF-1-α regulated pathways^161–164), 3) the design of labeled peptides binding specifically to the cell adhesion receptor integrin α(v)β³ or other tumor-specific antigens and of labeled bone marrow-derived endothelial precursor cells to allow highly specific tumor visualization and the study of glioma angiogenesis and neovascularization,^165–169 4) the generation and in vivo characterization of transgenic mice with gliomas induced by signaling through Ras and Akt pathways,¹⁷⁰ and 5) the construction of bifunctional imaging marker and therapeutic genes to allow direct assessment of therapeutic gene expression in culture and in vivo models by directly corresponding assays.^171,172 Especially the design of small tumor-specific antibody fragments is an attractive way for specific detection of tumor cells by imaging in vivo as well as for targeted therapy by radioimmunotherapy.

Many of the current experimental protocols investigating new drug and treatment strategies for experimental gliomas include MRI, optical or PET imaging of either the distribution of therapeutic agents,^173–175 or therapy-induced tumor-changes,^{171,176–184} with the overall attempt of designing image-guided treatments.^120,121 Most intriguing for potential clinical application is the design of multifunctional nanoparticles that can be detected both by MRI and fluorescence imaging, allowing for the noninvasive preoperative assessment of the tumor and for the intraoperative visualization of tumor margins by optical imaging.¹⁸⁵

SUMMARY AND CONCLUSION

Multimodal imaging (CT, MRI, PET, optical) and multitracer PET imaging 1) reveal the best set of anatomical, biochemical and molecular information on a specific tumor and, hence, a noninvasive diagnosis of lesion type and grade, 2) guide therapeutic choices, and 3) assess therapy effects. It should be pointed out that these imaging modalities are not competing with each other but give complementary information on various parameters of interest. Not every patient can be studied by these imaging technologies, and it is not necessary to do so. But these imaging technologies should be used together in selected patients to advance model systems and our understanding of the complex mechanism of glioma formation, behavior and migration and to allow the development and assessment of new therapeutic modalities including molecular targeted and gene therapies (“imaging-guided therapies”).