| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | COVER ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
||||||||
|
||||||||||
Dr. Torigian is Assistant Professor of Radiology, Department of Radiology, University of Pennsylvania School of Medicine, Philadelphia, PA.
Dr. Huang is Nuclear Medicine Resident, Department of Radiology, University of Pennsylvania School of Medicine, Philadelphia, PA.
Dr. Houseni is Visiting Scholar in Nuclear Medicine, Department of Radiology, University of Pennsylvania School of Medicine, Philadelphia, PA.
Dr. Alavi is Professor of Radiology; and Director of Research Education, Department of Radiology, University of Pennsylvania School of Medicine, Philadelphia, PA.
This article is available online at http://CAonline.AmCancerSoc.org
To earn free CME credit for successfully completing the online quiz based on this article, go to http://CME.AmCancerSoc.org
Disclosures: Dr. Huang was supported by a training grant through the National Institutes of Health (NIH) (T32-CA093258–03). Dr. Houseni was supported as a visiting scholar by the Ministry of Higher Education of Egypt.
*Please note that these authors have contributed equally to this manuscript.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRATIONALE FOR FUNCTIONAL IMAGING...FUNCTIONAL IMAGING TECHNIQUES TO...FUNCTIONAL IMAGING TECHNIQUES TO...OTHER APPLICATIONS OF FUNCTIONAL...CONCLUSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRATIONALE FOR FUNCTIONAL IMAGING...FUNCTIONAL IMAGING TECHNIQUES TO...FUNCTIONAL IMAGING TECHNIQUES TO...OTHER APPLICATIONS OF FUNCTIONAL...CONCLUSIONREFERENCES |
|---|
As such, functional imaging has come into existence and is comprised of a multitude of noninvasive, quantitative imaging techniques that are currently in use to study tumor physiology, to probe tumor molecular processes, and to study tumor molecules and metabolites in vitro and in vivo. Functional imaging can be implemented through use of CT, MRI, and US, as well as through positron emission tomography (PET), single-photon emission computed tomography (SPECT), and optical imaging (OI). Such functional imaging techniques can be used in conjunction with structural imaging techniques, often in a synergistic fashion, to detect, diagnose, characterize, or monitor tumors before and after therapeutic intervention. These can also be used to study tumor gene expression, to track cells and therapeutic drugs, to optimize individualized treatment planning for patients with tumors, and to foster new oncologic drug development.
In this article, we review the rationale for functional imaging of cancer, as well as the wide variety of functional imaging techniques that are available for use in the setting of cancer, which are becoming increasingly important for optimal individualized patient treatment in this day and age of "personalized medicine."
|
RATIONALE FOR FUNCTIONAL IMAGING OF CANCER |
|---|
|
TOPABSTRACTINTRODUCTIONRATIONALE FOR FUNCTIONAL IMAGING...FUNCTIONAL IMAGING TECHNIQUES TO...FUNCTIONAL IMAGING TECHNIQUES TO...OTHER APPLICATIONS OF FUNCTIONAL...CONCLUSIONREFERENCES |
|---|
The first is that gross macroscopic changes in tissues or organs due to cancer generally lag in time following alterations at the molecular, subcellular, or cellular levels.1,2 This has significant implications with regard to early detection of cancer for screening purposes, as well as for monitoring response to therapeutic intervention. The second is that macroscopic abnormalities are often nonspecific and often seen in nonneoplastic conditions. For example, enlarged lymph nodes in the setting of known malignancy may either be due to metastatic disease or benign hyperplasia.3 The third is that data regarding physiology, biological processes, and molecular characteristics of tumors are not obtained. As such, structural imaging at a single time point may not provide adequate information regarding patient prognosis, probability of tumor response to therapeutic intervention, and new drug development. This shortcoming is exemplified by the Response Evaluation Criteria in Solid Tumors (RECIST) criteria to assess for tumor response to therapy based on unidimensional lesional size measurements. These criteria completely disregard functional imaging information about tumors that may be relevant to accurate definition of early tumor response.4,5
Fortunately, quantitative functional imaging techniques may be used with structural imaging techniques to maximize the amount and value of data acquired for tumor detection, characterization, staging, prognosis assessment, treatment planning, and early treatment monitoring, as well as for monitoring of cell trafficking (ie, how different types of cells distribute to various organ or tissue locations in the body) and new drug development. These techniques involve specialized applications of CT, MRI, and US, as well as SPECT, PET, and OI.
PET and SPECT are 2 types of emission computed tomography that are used to study the distribution of radiolabeled tracers within the body. PET radiotracers emit positrons that, after encountering electrons, lead to emission of 2 gamma rays directed at 180 degrees from each other. Sites of tracer accumulation in the body are thus determined by detection of these paired gamma rays, which is termed "coincidence detection." SPECT involves the use of radioisotopes that emit single gamma rays in arbitrary directions, thus requiring the presence of a metallic collimator to determine sites of tracer accumulation in the body. While the collimators provide directionality, they screen out most of the emitted photons. As a result of these features, SPECT generally has lower sensitivity than PET and is not as readily quantifiable as PET. Furthermore, in contrast to positron-emitting radionuclides, single photon-emitting radionuclides cannot be labeled to most biologically important compounds. This also limits its use both for clinical and research purposes.2
OI uses visible light to create images following administration of an agent that can be detected through fluorescence or luminescence. However, as biologic tissues are opaque to light at most wavelengths, only small animals that are relatively transparent to light, as well as structures superficially located in the human body within the lumina of hollow organs or along the external bodily surface, are generally amenable to OI.6
|
FUNCTIONAL IMAGING TECHNIQUES TO STUDY TUMOR PHYSIOLOGY |
|---|
|
TOPABSTRACTINTRODUCTIONRATIONALE FOR FUNCTIONAL IMAGING...FUNCTIONAL IMAGING TECHNIQUES TO...FUNCTIONAL IMAGING TECHNIQUES TO...OTHER APPLICATIONS OF FUNCTIONAL...CONCLUSIONREFERENCES |
|---|
MRI has been commonly applied to study tumor perfusion through the dynamic contrast-enhanced MRI (DCE-MRI) technique, performed by acquiring images before, during, and after intravenous administration of a paramagnetic contrast agent (most often a gadolinium [Gd] chelate). This has been used to study various cancers where measurement of tumor microvascularity has been shown to correlate with tumor grade, microvessel density, tendency for metastasis and recurrence, levels of vascular endothelial growth factor (VEGF) expression, and survival outcomes.10 For example, George et al reported that rectal tumors with higher permeability at presentation as measured on DCE-MRI appeared to respond better to chemoradiotherapy than those with lower permeability.11 Perfusion MRI also has utility in the diagnosis and characterization of tumors. For instance, perfusion-weighted MRI has been used to differentiate pyogenic brain abscesses from cystic or necrotic brain tumors, which can be challenging when structural MRI features alone are used for this purpose.12 Furthermore, effects of antiangiogenic and antivascular therapy can also be assessed with this technique.9 As such, DCE-MRI has been used in clinical trials as an indicator of biological activity to assess the effectiveness of such therapies.13
Currently, CT is the most frequently used imaging tool for clinical assessment of the structural features of cancer. However, perfusion techniques can also readily be incorporated into existing CT protocols by acquiring images at multiple time points before and after intravenous administration of iodinated contrast agents.14 Tumor angiogenesis correlates with tumor perfusion and vascular permeability, which can be assessed by determination of peak contrast enhancement and measurement of time-to-peak enhancement on dynamic contrast-enhanced CT (DCE-CT).15 Measurements of tissue perfusion in tumors such as lung cancer have also been shown to correlate with markers of angiogenesis, such as VEGF expression.16 Ng et al demonstrated that in non-small-cell lung carcinoma (NSCLC), quantitative whole-tumor assessment of vascular permeability and blood volume with CT is reproducible and may be useful to monitor effects of antivascular or antiangiogenic agents (Figure 1).17 Perfusion CT analysis can also be applied to characterize lesions, and a primary example of this application is the assessment of the solitary pulmonary nodule. For example, Swensen et al reported that absence of significant enhancement (defined as 
15 Hounsfield units) on CT at 1 to 4 minutes after contrast administration is strongly predictive of a benign histology.
|
Perfusion imaging with SPECT and PET is performed with nontargeted tracers that have very high first-pass extraction, mostly for assessment of large organs such as the heart and brain where detailed spatial resolution is of limited value. With increasing availability of PET/CT and SPECT/CT instrumentation, perfusion and related imaging techniques can be carried out successfully in most clinical settings. However, the value of this approach in cancer diagnosis and management is presently unclear.
Diffusion Imaging
The random translational diffusion of water molecules can be quantitatively measured with diffusion-weighted MRI (DWI).21 DWI provides indirect information about tissue structure, water content, and intra/extracellular space and can be used to estimate tumor cellularity and detect early changes following therapeutic intervention.22 When cellular and subcellular compartmental membranes break down with necrosis or apoptosis, water molecules become less restricted to motion by diffusion.21 However, when cellular swelling or an increase in tumor cellularity are encountered, water molecules become more restricted to motion by diffusion due to reduction in the extracellular space, where most of the translational movement of water molecules occurs (Figure 2).23
|
Through diffusion tensor MRI (DTI), anisotropic diffusion of water molecules in tissues such as the cerebral white matter (due to organization of myelinated axons in bundles) can be quantitatively measured, providing exquisite detail of tissue microstructure.28 This is particularly useful for pretreatment planning for brain tumors through differentiation of peritumoral edema from infiltrative tumor, pretreatment assessment of white-matter tract involvement by tumor, and intraoperative visualization and localization of major white-matter tracts to decrease the chance of injury to normal tissues (Figure 2).29
DWI and DTI have been more difficult to use outside of the brain, in part due to increased motion artifacts.28 For example, investigators have shown that DWI may provide information regarding extent of breast cancer, as well as early response to therapeutic intervention.30,31 However, these applications are still under active investigation.
Functional Lymph Node Imaging
Lymph node imaging is of great importance for the staging of malignant tumors. However, assessment of lymph nodes based on structural imaging features alone is limited in sensitivity and specificity for microscopic metastatic disease. Hence, functional imaging techniques are currently under development to overcome these limitations.
Functional lymph node imaging has largely been performed with radiolabeled particles to identify nodes that most likely represent probable sites of metastasis. The best known is 99mTc sulfur colloid, which is used for conventional sentinel node imaging.32 In the early stages of metastasis when the flow of lymph through an affected lymph node has not yet been altered, conventional sentinel node imaging can correctly identify this lymph node. However, as the number of cells in an affected lymph node increases due to tumor growth, the flow of lymph through the node may be diverted to the next draining lymph node due to increased resistance, leading to lack of depiction of the originally involved sentinel lymph node. At this stage, blood supply to the involved lymph node is primarily provided by capsular arterial vessels to sustain the viability of cancer cells. Therefore, systemically administered compounds such as 18F-2-fluoro-2-deoxy-D-glucose (FDG) will be directly delivered to these cancer cells and permit visualization of these sites of nodal involvement in the whole body. Thus, in such settings, conventional sentinel lymph node imaging is complementary with whole-body FDG-PET nodal imaging.33 However, these observations are based on preliminary results from a limited number of patients and thus will require further validation for clinical utility in the future.
A new agent to evaluate lymphatic drainage using MRI is ferumoxtran-10, a particular ultrasmall superparamagnetic iron oxide-based biodegradable nanoparticle composed of an iron oxide core and a dextran coating, which is an example of a nontargeted contrast agent that leaks into the interstitium and reaches the reticuloendothelial cells of lymph nodes via the lymphatic system, allowing for detection of micrometastases within normal-sized lymph nodes. Normal lymph nodes become dark in signal intensity after ferumoxtran-10 administration due to shortening of T2 and T2* relaxation times of water protons, whereas metastatic deposits within lymph nodes remain bright.34,35 This agent has been used to evaluate lymph nodes in the setting of a wide variety of malignancies. For example, Deserno et al showed that ferumoxtran-10-enhanced MRI significantly improved nodal staging in those with bladder carcinoma by depicting metastases even in normal-sized lymph nodes, with accuracy, sensitivity, specificity, and positive and negative predictive values of 95%, 96%, 95%, 89%, and 98%, respectively.36 Also, Harisinghani et al reported that MRI with lymphotropic superparamagnetic nanoparticles correctly identified all patients with nodal metastases due to prostate carcinoma and that nodal analysis had a significantly higher sensitivity (90.5%) compared with conventional MRI (35.4%) (Figure 3).34 Will et al performed a meta-analysis of prospective studies that compared MRI, with and without ferumoxtran-10, with histological diagnosis after surgery and biopsy and showed that ferumoxtran-10-enhanced MRI is sensitive and specific in detection of nodal metastases for a variety of tumors and offers higher diagnostic precision than unenhanced MRI for detection of lymph node metastases.37 As such, ferumoxtran-10-enhanced MRI, like PET, has the potential to become a useful clinical functional imaging tool to evaluate lymph nodes of the whole body in the setting of cancer.
|
|
FUNCTIONAL IMAGING TECHNIQUES TO PROBE TUMOR MOLECULAR PROCESSES |
|---|
|
TOPABSTRACTINTRODUCTIONRATIONALE FOR FUNCTIONAL IMAGING...FUNCTIONAL IMAGING TECHNIQUES TO...FUNCTIONAL IMAGING TECHNIQUES TO...OTHER APPLICATIONS OF FUNCTIONAL...CONCLUSIONREFERENCES |
|---|
Glucose Metabolism
FDG, a glucose analogue first tested in humans in 1976, is by far the most widely used metabolic tracer in the world.39 Pioneering work since its conception has helped to solidify FDG-PET as a powerful diagnostic imaging modality in oncology and has sparked a growing interest in molecular imaging.2 FDG-PET can be used to assess any malignancy that has accelerated glucose metabolism.
FDG is transported into cells via glucose transporters such as GLUT-1 and subsequently phosphorylated by hexokinase.40 The metabolic product 2-fluorodeoxyglucose-6-phosphate is effectively trapped intracellularly, as it cannot undergo further metabolism through the glycolytic pathway. Persistent glucose uptake in many cancer cells in a low serum insulin state makes FDG an ideal tracer for cancer imaging. In fact, the Centers for Medicare and Medicaid Services (CMS) has approved reimbursements for diagnosis, staging, and restaging of lung cancer, esophageal cancer, colorectal cancer, melanoma, lymphoma, head and neck cancer, and breast cancer. Characterization of the solitary pulmonary nodule, restaging of noniodine-avid thyroid cancer, and evaluation of metastases in cervical cancer before therapy are also reimbursable by Medicare. Recently, the CMS has approved reimbursement for virtually all cancers if conducted in centers participating in certain patient registries such as the National Oncologic PET Registry (NOPR).
FDG-PET provides complementary information to structural imaging techniques for diagnosis and staging of cancer (Figures 4 and 5). For example, meta-analysis by Gould et al reported a sensitivity of 97% and a specificity of 78% for the evaluation of solitary pulmonary nodules.41 In addition, meta-analysis by Toloza et al indicates that FDG-PET has an 84% sensitivity and an 89% specificity for mediastinal staging of NSCLC, whereas CT alone has a sensitivity and specificity of 57% and 82%, respectively.42 The difference in sensitivities between metabolic imaging with FDG-PET and anatomic imaging using CT-size criteria highlights a typical advantage of metabolic imaging when compared with structural imaging. With its high sensitivity for metabolically active lesions, FDG-PET has also been shown to detect unexpected extrathoracic metastases in 10% to 20% of patients and leads to changes in therapeutic management in about 20% of patients with lung cancer. In general, FDG-PET adds value to the staging workup of patients with relatively aggressive and metabolically active cancers such as lymphoma, melanoma, and breast, lung, head and neck, cervical, esophageal, or colorectal cancer by increasing the sensitivity for nodal and distant metastatic disease and by increasing the specificity for malignancy relative to the diagnostic performance of structural imaging alone. However, FDG-PET in general has lower sensitivity for slower growing, less metabolically active tumors such as prostate cancer, thyroid cancer, neuroendocrine tumors, and bronchioloalveolar cell lung carcinoma. FDG-PET is also particularly useful for restaging of malignancy because structural changes caused by surgery or radiation therapy may render anatomic imaging techniques inconclusive at most anatomic sites. Cancer restaging with FDG-PET has a general sensitivity of 80% to 95%, a specificity of 75% to 90%, and an accuracy of 80% to 90% for the detection of persistent or recurrent disease.43 With the introduction of a recent generation of PET and PET/CT scanners, as well as well-established diagnostic criteria for interpretation of FDG-PET scans, the value of FDG-PET in evaluation of patients with cancer has been enhanced even more. Please note that a detailed review of the specific clinical indications for PET is beyond the scope of this article.
|
|
|
|
Amino Acid Metabolism
Analogues of methionine, tyrosine, and L-dihydroxyphenylalanine (L-DOPA) have potential for widespread clinical application and sometimes have better diagnostic performance characteristics in certain PET applications when compared with those utilizing FDG. 11C-L-methionine is chemically identical to nonradioactive L-methionine. Consequently, it is metabolized and trapped in the cell by incorporation into proteins or even RNA molecules.53 Advantages of imaging brain tumors with11C-methionine are well documented, as there is low background uptake of 11C-methionine in the brain. A recent study by Jacobs et al showed a sensitivity of 91% for detecting gliomas, as well as larger detected tumor volumes compared with those detected with Gd-enhanced MRI, which may have an impact on radiation treatment planning.54 However, the major shortcoming of11C-methionine is the 20-minute half-life of the 11C isotope, although 18F-fluorinated amino acid analogues have been developed to overcome this limitation since 18F has a 109-minute half-life.
Tyrosine analogues such as L-3–18F-fluoro-
-methyl tyrosine (FMT) and O-(2–18F-fluoroethyl) L-tyrosine (FET) may be of clinical value in imaging certain tumors. For example, FMT-PET had a sensitivity of 73% and a specificity of 85% in a series of 75 patients with musculoskeletal tumors (whereas the specificity of FDG-PET was lower at 66%).55
18F-6-fluorodihydroxyphenylalanine (18F-FDOPA), an analogue of L-DOPA, accumulates in dopaminergic neurons in the basal ganglia and can be used to study patients with melanoma, carcinoid tumor, medullary thyroid cancer, and brain tumors.56 A recent study with FDOPA-PET in 19 patients with abdominal carcinoid tumors showed 93% sensitivity and 89% accuracy.57 FDOPA-PET has also been shown to be more accurate than FDG-PET for evaluation of low-grade and recurrent brain tumors and for differentiation of low-grade tumors from necrosis.58
Cell Proliferation Imaging
Cell proliferation imaging assesses the uptake and metabolism of nucleosides used in nucleic acid synthesis. The most widely used nucleoside analogue for cell proliferation imaging to date is 3'-deoxy-3'-18F-fluorothymidine (FLT), which enters cells through carrier-mediated diffusion.59 The tracer is trapped when a thymidine kinase (TK) such as TK1 phosphorylates FLT to fluorothymidine-phosphate, which cannot be metabolized further. Shields et al was the first to propose the use of FLT-PET to measure cellular TK activity and thereby infer the rate of cellular proliferation.60 Overall, FLT-PET has lower sensitivity compared with FDG-PET and amino acid analogue PET for the detection of cancer. However, FLT-PET provides images for rates of proliferation, whereas most other tracers do not (Figure 8). The specificity of FLT for assessment of cellular-proliferative activity makes it an ideal tracer for monitoring tumor treatment response, especially when antiproliferative cytostatic agents are employed. For example, Yang et al tested radiation treatment monitoring with FLT-PET in mouse tumor models and found decreased FLT uptake 24 hours after radiation treatment with no detectable change in FDG uptake. This suggests that FLT-PET can shorten the timeframe to detect a therapeutic response and may become an effective tool for monitoring treatment.61
|
Nitroimidazole-based compounds such as18F-fluoromisonidazole (FMISO) are currently the most widely used agents to image hypoxia with PET. Fluoromisonidazole is a lipophilic compound that binds to intracellular proteins in live cells with active mitochondria when oxygen concentration is below 20 mmHg.65 In a recent study of 58 patients with head and neck cancer, pretherapy FMISO-PET had independent prognostic value.66 In addition, FMISO-PET can be used to monitor chemoradiation therapy in patients with advanced head and neck cancer where early resolution of FMISO uptake during treatment is associated with excellent locoregional control.67
64Cu(II)-diacetyl-bis(N4-methylthiosemicarbazone) (64Cu-ATSM) is another tracer that is used to image hypoxia.68 In hypoxic cells, the redox state of the copper atom changes, and 64Cu becomes trapped in mitochondria after dissociation of the Cu-ATSM complex. However, recent animal testing using various tumor models shows that 64Cu-ATSM is a valid PET hypoxia marker in some, but not all, tumor types, implying that it may not be useful as a general hypoxia imaging agent.69
18F-EF5, a 2-nitroimidazole analogue, is another potential positron-emitting hypoxia tracer.70 Hypoxic regions delineated by 18F-EF5-PET have been shown to influence patient survival and lack of response to radiation treatment, and preliminary tests of 18F-EF5 in patients with head and neck and brain tumors show promising results.71,72
Blood oxygen level-dependent (BOLD) functional MRI (fMRI) has been used as a noninvasive method to preoperatively map functional cerebral cortex and to identify eloquent areas of the cerebral cortex in relation to brain cancers, potentially reducing the time of surgery and minimizing intraoperative cortical stimulation methods, which are sometimes used during the surgical resection of brain tumors.73,74 In addition, BOLD fMRI, as well as other functional MRI techniques such as Overhauser-enhanced MRI and electron paramagnetic resonance imaging (EPRI), may be useful to study oxygenation in human tumors and to monitor changes in tumors related to therapy.75,76
Molecular Imaging with Targeted Imaging Agents
Somatostatin Receptors
Somatostatin receptors (SSTRs) are overexpressed in many neuroendocrine tumors, including small-cell lung cancer, carcinoid tumor, insulinoma, VIPoma, gastrinoma, and medullary thyroid cancer. Whole-body SSTR imaging is performed with radiolabeled derivatives of octreotide, which have a preference for subtype 2 SSTRs (ie, SSTR2). The most commonly used agent is 111In-pentetreotide for SPECT imaging, although SSTR-targeted tracers have also been developed for PET, including 68Ga-D-Phe1-Tyr3-octrotide (68Ga-DOTATOC) and Gluc-Lys(18F-fluoropropionyl-TOCA). Meisetschlager et al used Gluc-Lys(18F-fluoropropionyl-TOCA)-PET to study 25 patients with SSTR-positive tumors. Compared with 111In-pentetreotide SPECT, Gluc-Lys(18F-fluoropropionyl-TOCA)-PET revealed more than twice the number of lesions and resulted in near-perfect interobserver agreement. Gluc-Lys(18F-fluoropropionyl-TOCA) also demonstrated better biodistribution, biokinetics, and diagnostic performance in comparison with 68Ga-DOTATOC in the assessment of neuroendocrine tumors.77
Estrogen Receptors
Currently, in vitro examination of biopsy specimens with immunohistochemistry is required to assess levels of estrogen receptor (ER) expression, although not all tumor sites are accessible by biopsy. ER targeting tracers, therefore, may be used to noninvasively assess the ER status of tumors in vivo through use of 18F fluorinated estrogen analogues, and 16-18F-fluoroestradiol-17β (FES) is commonly used for this purpose due to its favorable biodistribution.59 FES-PET may show a faster treatment response than FDG-PET at 7 to 10 days after treatment initiation when monitoring breast cancer therapy using tamoxifen.78 FES-PET may also assist in the selection of the appropriate therapy for breast cancer patients. Use of FES-PET may increase overall response rate following therapy with tamoxifen from 23% to 34%, and for patients with HER2/neu-negative breast cancer, treatment response may improve from 29% to 46% if selection is made using FES-PET.79
Angiogenesis Imaging
Angiogenesis, the formation of new blood vessels, is now the target of many new therapeutic agents.80,81 Perfusion imaging as described above can indirectly provide information about tumor angiogenesis, although some tumors may have dense but mostly quiescent endothelium that is resistant to antiangiogenic therapy. Accurate assessment of angiogenesis thus requires approaches that can directly quantify activated endothelial cells, such as through detection of Vβ3 integrins that are present on activated endothelial cells undergoing angiogenesis.82
Peptides containing the amino acid sequence arginine-glycine-aspartate (RGD) have a general affinity toward integrins. It has been shown that fine-tuning of the stereochemistry adjoining the RGD peptide allows for selective binding to specific integrins.83,84 This led to the development of RGD-based pentapeptide integrin antagonists, such as EMD 121974 (Cilengitide), which can be used to inhibit tumor angiogenesis.85,86 18F-labeled RGD compounds have also been synthesized to allow for rapid visualization of Vβ3 expression in the whole body using PET.87 For example, 18F-galacto-RGD PET was used by Beer et al in 19 patients with various solid tumors. Tumoral uptake of 18F-galacto-RGD correlated with Vβ3 expression, as well as microvessel density as observed at histology.88 Thus, 18F-galacto-RGD PET has potential use for individualized treatment planning and antiangiogenic therapy monitoring, although a drawback is that it binds to Vβ3 integrins present on the membranes of tumor cells, as well as membranes of tumor-associated endothelial cells. As such, imaging agents based on targeting of cell receptors, such as vascular endothelial cell growth factor receptor, that are overexpressed on endothelial cells during angiogenesis are currently under development. Nevertheless, ability to quantify total tumor Vβ3 integrin levels is important, given that Vβ3 expression by tumor cells is linked to increased metastatic potential.89
Apoptosis Imaging
Apoptosis, or programmed cell death, is a characteristic feature of normal human cells that is essential for normal tissue homeostasis and tissue differentiation.90 However, tumor growth is often associated with insufficient apoptosis. In vivo imaging of apoptosis offers another noninvasive approach to monitor therapeutic response independent of changes in glucose metabolism or cell proliferation.91 Annexin V binds to phosphatidylserine (PS) that has moved to the outer surface of cell membranes during apoptosis and can be labeled with either 99mTc, 124I, or 18F to image apoptosis in vivo using SPECT or PET, respectively.92 OI of apoptosis in preclinical animal models or in vitro is also possible using Annexin V labeled with Cy5.5, a near-infrared fluorescent dye.93 The source of light emission for fluorescent imaging is either from fluorescent organic compounds called fluorophores or semiconductor-based nanoparticles called quantum dots.38
Superparamagnetic iron-based contrast agents, as well as other paramagnetic contrast agents in conjunction with targeted carrier nanoparticles, are also under development in animal models to specifically target particular molecules or biologic processes for detection, predominantly using MRI.94,95 For example, Schmieder et al noninvasively characterized melanoma in its earliest phase of growth in mice using Vβ3-targeted paramagnetic nanoparticles to detect Vβ3 integrin expression in tumor neovasculature using MRI, and Schellenberger et al have used nanoparticles conjugated to Annexin V to target apoptosis for noninvasive monitoring using MRI or OI.96,97 However, limitations of such approaches using MRI for detection include low signal-to-noise ratio with low sensitivity, lack of standardization of imaging and analytic methodologies, and inconsistent availability of these new technologies at sites around the country.1,98
Magnetic Resonance Spectroscopy
Magnetic resonance spectroscopy (MRS) is a noninvasive diagnostic tool that allows one to quantitatively assess the amount, type, and location of various small molecular compounds within a tissue or organ of interest at the same time that MRI is performed. It does so by taking advantage of the quantum-mechanical properties of the nuclei of certain atomic isotopes, which permit them to be manipulated through the application of magnetic fields and radiofrequency pulses. The data are typically displayed as a grid of spectra of chemical compound abundances obtained at either single or multiple locations in a tissue or organ of interest. These spectra are collected from spinning nuclei (spins), most often 1H given the abundance of water in tissue, although they can be collected from other nuclei such as 13C or 31P with more difficulty due to lower tissue concentrations.99 Proton 1H MRS enables accurate quantitative assessment of the spatial distribution of tissue metabolites such as creatinine, choline, amino acids, nucleotides, lactate, and lipids.100 Unfortunately, there is no one cancer signature on MRS, but there are MRS findings that are seen more often in cancer.99 For example, increased concentrations of cell membrane metabolites such as phosphocholine may be seen due to increased membrane synthesis, a higher concentration of lactate may be seen due to increased metabolism through the glycolytic pathway, and lower concentrations of normal tissue metabolites (such as N-acetyl aspartate in cerebral tissue or citrate in prostatic tissue) may be encountered (Figure 2).101,102
1H MRS is most often applied clinically to evaluate brain tumors due to the relative lack of motion artifact in this location relative to other body sites. As such, many single-center studies of 1H MRS in the setting of brain cancer have been reported.99 In patients with gliomas, 1H MRS and perfusion-MRI analysis provide complementary information to assess tumor grade and to detect residual or recurrent tumor. 1H MRS enables better discrimination between normal and abnormal tissue than regional cerebral blood-flow maps, whereas regional cerebral blood-flow maps may better reflect the heterogeneity of tumor regions because of their higher resolution.103 In addition, 1H MRS may be useful to predict patient prognosis. Oh et al reported that patients with glioblastoma multiforme had significantly shorter median survival when a large volume of metabolic abnormality was seen at 1H MRS.104 Many investigators have also demonstrated significant spatial discordance between morphologic lesions and metabolic lesions using 1H MRS in brain tumors, which has significant implications with regard to treatment planning for surgery and/or radiotherapy.104,105 Moreover, 1H MRS can be used to distinguish pyogenic brain abscesses from cystic or necrotic brain tumors, gliomas from solitary brain metastases, and residual or recurrent brain tumor from post-therapy radiation necrosis.106–108
1H MRS has occasionally been applied to study tumors in other body locations. Swindle et al reported that 1H MRS has a high sensitivity (97%) and specificity (88%) for prostate carcinoma.109 Furthermore, 1H MRS can lead to more accurate assessment of the presence, location, and extent of prostate cancer; may reveal overall tumor clinicopathologic status and aggressiveness; and can be used to assess therapeutic response.110 1H MRS has also been used to study tumors of the breast, pancreas, cervix, and thyroid gland.111
31P MRS also has potential utility to noninvasively assess tumors. It has shown some promise to differentiate prostate carcinoma from benign prostatic hypertrophy.112 Others have shown that it can differentiate benign from malignant head and neck neoplasms and can predict treatment response and monitor therapeutic effects for head and neck cancer.113 Arias-Mendoza et al reported that 31P MRS pretreatment measurement of phosphoethanolamine and phosphocholine content within non-Hodgkin lymphoma predicts long-term response to treatment and time-to-treatment failure, particularly when combined with the international prognostic index.114
Research efforts directed toward use of hyperpolarized 13C MRI/MRS agents for metabolic and molecular imaging are currently underway, although many technical challenges still exist. 13C MRI/MRS has the potential to provide information about any molecule that contains hyperpolarized 13C and to distinguish between different molecules that contain 13C atoms. This suggests the potential utility of this approach to study carbon-based compounds in vivo for oncologic applications such as tumor perfusion and metabolism assessment, as well as new drug pharmacokinetic and pharmacodynamic assessment.115
Although MRS has been around for several decades and can sensitively detect and characterize metabolites within tumors and nonneoplastic tissue, it is still considered to be a prebiomarker by most investigators as it is not yet fully established as reproducible across multiple institutions. Difficulties in performing multicenter trials with MRS include differences in hardware and software among different sites due to lack of standardization, as well as effects of patient motion.99
|
OTHER APPLICATIONS OF FUNCTIONAL IMAGING TO STUDY TUMORS |
|---|
|
TOPABSTRACTINTRODUCTIONRATIONALE FOR FUNCTIONAL IMAGING...FUNCTIONAL IMAGING TECHNIQUES TO...FUNCTIONAL IMAGING TECHNIQUES TO...OTHER APPLICATIONS OF FUNCTIONAL...CONCLUSIONREFERENCES |
|---|
Several investigators have focused on use of OI to study gene expression. For example, Cao et al used GFP linked to hypoxia-inducible factor (HIF-1) promoter in a tumor cell line that expressed red fluorescent protein (RFP) constitutively and was able to differentiate between normoxic tumor cells and hypoxic tumor cells in a growing tumor environment.117 At present, use of OI to study gene expression in tumors is limited to use in small animals and for superficial human structures such as the breast due to limited light penetration through tissue. Other investigators have focused on PET to study gene expression in tumors. A prototypical reporter protein for use with PET is HSV-tk, which can be imaged following administration of a positron-emitting nucleoside analogue reporter probe that can be phosphorylated and trapped intracellularly by HSV-tk.118
Cell tracking is important to study the behavior of tumor cells in vitro and in animal models and can potentially be used to track cellular therapies for cancer, to track stem cells, and to assess the effectiveness of gene therapy against cancer cells.119 Cell tracking is possible when cells of interest are engineered to constitutively express a reporter protein that is detectable by either OI or PET.120 For example, Jenkins et al developed bioluminescent prostate cancer and breast cancer models to observe treatment response to chemotherapy using OI.121 In addition, investigators have introduced luciferase and GFP into genetically transformed cells during gene therapy to monitor efficiency of transfection using OI or PET.122 Recently, some investigators have successfully performed in vivo cell tracking using MRI in animal models through use of superparamagnetic iron oxide nanoparticles or paramagnetic contrast agents that are introduced into cells, although efforts to translate this approach into humans are currently under active investigation.123,124
Molecular Imaging-based Personalized Treatment Planning
Molecular phenotypes of tumors may affect their susceptibility to chemotherapeutic agents, which may be extremely important to know about for individualized treatment planning. For example, epidermal growth factor receptor (EGFR) mutations are found in about 10% of NSCLC, and patients with these mutations have a statistically significant better outcome following EGFR tyrosine kinase inhibitor therapy with gefitinib (Iressa; AstraZeneca, London, UK).125 In the remaining 90% of patients without these mutations, unnecessary exposure to this drug may be avoided, leading to monetary savings and fewer adverse events from drug toxicity, as well as earlier use of alternative, more effective therapies.
As seen by this example, the "personalized medicine" paradigm is becoming a reality in treatment planning for many patients with malignancies, particularly as therapeutics are increasingly targeted toward specific cellular receptors. As such, molecular imaging will increasingly become a part of individualized treatment planning and pharmacogenomics. Molecular imaging will increasingly supplement in vitro molecular diagnostic testing of tissue or bodily fluid specimens, as it can be used for quantitative, noninvasive serial evaluation, as well as whole-body assessment of multiple lesions, including those that are in difficult to reach locations. Moreover, molecular imaging using radiolabeled preparations of intended drugs may be used to determine appropriate pharmacological doses on an individualized basis.
An example of this concept is the pretreatment evaluation of pharmacokinetics of radiolabeled antibodies or peptides to be used for therapeutic purposes using PET. In order to predict treatment success using radiolabeled antibodies or peptides against a tumor, it is important to demonstrate appropriate targeting to the diseased tissue. Positron-emitting isotopes such as 124I (as a surrogate for 131I) and 86Y (as a surrogate for 90Y) can be utilized for optimal visualization of these agents to targeted sites and may significantly improve the ability to select optimal patient candidates for subsequent treatment with antibodies or peptides radiolabeled with therapeutic beta-emitting radionuclides. Through this approach, one may identify patients who are good candidates for antibody or peptide treatment by comparing images acquired with nonspecific radiotracers such as FDG that reveal overall disease activity with PET images generated with positron-emitting radiolabeled antibodies or peptides for diagnostic purposes. Since radiolabeled antibodies result in a favorable response if they target the entirety of the tumor burden, disparity between uptake within tumor on the FDG-PET images and diagnostic PET images using radiolabeled antibodies or peptides may indicate a poor outcome following treatment with beta-emitting agents.126
New Drug Development
Functional imaging techniques have the potential to accelerate cancer drug development during the preclinical and clinical phases of testing.127,128 This is especially relevant given the difficulties in both designing targeted cancer therapies and monitoring their efficacy. Tumors may be spatially and temporally heterogeneous in terms of gene expression, metabolism, oxygenation status, angiogenesis, cell proliferation, apoptosis, signal transduction, and other phenotypic features, but functional imaging can help investigators to better understand these features.127
As such, functional imaging can aid in molecular target-based drug screening, definition of in vivo target specificity, rapid characterization of in vivo pharmacokinetic and pharmacodynamic properties of candidate drugs, development of animal models of human tumors, and visualization of in vivo effects of therapeutic molecules on target biological processes.128 These approaches are particularly useful as they could reduce the length of time for developing effective drugs (which usually can take more than a decade) and reduce the costs of drug development (often in the range of hundreds of millions of dollars) during preclinical evaluation and clinical trial assessment.129,130 For example, any new therapeutics for GIST may use the post-therapy FDG-PET scan as the surrogate endpoint of the trial since the 8-day post-therapy FDG-PET scan has been shown to predict outcome for imatinib treatment.50 Recently, FLT-PET has been used to monitor the preclinical testing of histone deacetylase inhibitors (HDACI), an expanding class of growth-inhibiting compounds targeting DNA and RNA. In animal models of colon carcinoma, Leyton et al reported that decreasing FLT uptake in tumor correlates with escalating doses of LAQ824, a new HDACI, and that the quantity of FLT uptake correlates with decreased expression of thymidine kinase 1 (TK1).131
Transformation of therapeutic compounds themselves into imaging agents and testing for the presence of the compound at intended target sites is another way functional imaging can be used to facilitate drug development. 18F, 11C, and 13N PET can offer this possibility given its high sensitivity. Similarly, 19F and 31P MRS can be used for the detection, identification, and quantification of fluorine-containing or phosphorous-containing therapeutic agents, and their metabolites in biofluids may play a role in the therapeutic monitoring of fluorinated or phosphorylated chemotherapeutic drugs such as fluorouracil and cyclophosphamide, respectively, in specific groups of patients and may provide further understanding into drug metabolic pathways. However, the main challenge to widespread application of these MRS techniques is a more limited sensitivity in the micromolar-millimolar concentration range.132,133
|
CONCLUSION |
|---|
|
TOPABSTRACTINTRODUCTIONRATIONALE FOR FUNCTIONAL IMAGING...FUNCTIONAL IMAGING TECHNIQUES TO...FUNCTIONAL IMAGING TECHNIQUES TO...OTHER APPLICATIONS OF FUNCTIONAL...CONCLUSIONREFERENCES |
|---|
|
Acknowledgments |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRATIONALE FOR FUNCTIONAL IMAGING...FUNCTIONAL IMAGING TECHNIQUES TO...FUNCTIONAL IMAGING TECHNIQUES TO...OTHER APPLICATIONS OF FUNCTIONAL...CONCLUSIONREFERENCES |
|---|
