| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Review |
Authors' Affiliations: 1 Cancer Imaging Program, Division of Cancer Treatment and Diagnosis, National Cancer Institute, NIH; 2 Molecular Tracer, LLC, Bethesda, Maryland; 3 Division of Nuclear Medicine, Department of Radiology, University of Washington, Seattle, Washington; 4 Memorial Sloan-Kettering Cancer Center, New York, New York; 5 Center for Molecular Imaging Research, Massachusetts General Hospital, Harvard Medical School, Charlestown, Massachusetts; 6 CCS Associates, Mountain View, California; 7 Baylor Charles A. Sammons Cancer Center, Dallas, Texas; 8 Lombardi Comprehensive Cancer Center, Georgetown University, Washington, District of Columbia; and 9 Office of Drug Evaluation, Division of Medical Imaging and Radiopharmaceutical Drug Products, and 10 Office of the Commissioner, Food and Drug Administration, Rockville, Maryland
Requests for reprints: Gary J. Kelloff, Cancer Imaging Program, Division of Cancer Treatment and Diagnosis, National Cancer Institute, NIH, EPN 6130 Executive Boulevard, Suite 6058, Bethesda, MD 20892. Phone: 301-594-0423; Fax: 301-480-3507; E-mail: kelloffg{at}mail.nih.gov.
 |
Abstract |
|---|
 Top Abstract Value of Imaging-Based...Basic Scientific Themes for...Imaging Fundamental Properties...Clinical Applications of Imaging...Opportunities for Probe...SummaryReferences |
|---|
|
Value of Imaging-Based Biomarkers in Drug Development |
|---|
|
TopAbstractValue of Imaging-Based...Basic Scientific Themes for...Imaging Fundamental Properties...Clinical Applications of Imaging...Opportunities for Probe...SummaryReferences |
|---|
Many other molecularly targeted agents, despite robust scientific rationale and promising preclinical and preliminary clinical results, have failed to show efficacy in definitive clinical trials (9). As an example, monotherapy with the EGFR inhibitor gefitinib did not improve survival compared with best supportive care in a recent Phase 3 trial in lung cancer patients. A small (10%) but encouraging response rate in refractory non–small cell lung cancer had supported its accelerated approval under Subpart H, with some patients experiencing striking and durable responses. Emerging data suggest that certain EGFR mutations are more frequent in the patients responding to gefitinib (and other EGFR inhibitors), particularly nonsmokers, women, and those with adenocarcinoma histology or of Asian ethnicity (10–13). However, neither the phosphorylation status nor the expression of EGFR is predictive of response, and EGFR polymorphisms do not seem to correlate with relevant EGFR mutations (14). Thus, selection of non–small cell lung cancer patients likely to respond to EGFR inhibitors remains problematic.
Together, these difficulties highlight the need for faster, more efficient, and more cost-effective development of cancer therapeutics and for better definition of patients likely to benefit from treatment. As addressed by the recent Food and Drug Administration (FDA) Critical Path Initiative, collaborative interactions among such scientific knowledge areas as bioinformatics, genomics, materials science, and imaging technologies are needed to design and implement better drug development tools.11 Important among these tools are molecular imaging probes that image specific molecular pathways in vivo, enabling visualization of phenotypic expression of key targets in the cancer disease processes. Unlike anatomic imaging, oncologic molecular imaging probes display biochemical and physiologic abnormalities underlying the cancer rather than the structural consequences of these abnormalities. Imaging-based biomarkers have many potential uses in all phases of the drug development process, from target discovery and validation to pivotal clinical trials for drug registration (see Fig. 1; see also ref. 15). First, as disease biomarkers, imaging end points can be employed to define, stratify, and enrich study groups. One such approach is to apply imaging-based probes in molecular target identification, characterization, and quantification to identify appropriate patient populations in which to test targeted agents. Second, clinical imaging studies of the labeled drug (e.g., in microdosing protocols; ref. 16) have potential to facilitate early clinical pharmacokinetic/pharmacodynamic assessments, including target interaction and modulation, particularly in patients where traditionally there are no direct measures of pharmacokinetics/pharmacodynamics throughout the tissues of the body and at the target. These approaches could be used in early studies comparing lead candidates designed to interact with the same target. A third area where imaging-based biomarkers have promise for speeding drug evaluation is by replacing or supplementing time- and labor-intensive dissection and histologic analyses in both preclinical and clinical testing. These noninvasive approaches may enable longitudinal preclinical studies with greater relevance to future clinical study designs. Finally, as biomarkers of tumor response, imaging end points (apoptosis, proliferation, angiogenesis, etc.) can also serve as early surrogates of therapy success.
|
This article outlines the scientific basis of oncology imaging probes and delineates areas where application of noninvasive or minimally invasive molecular imaging techniques could facilitate progress. Example probes are highlighted and settings in which imaging can meet clearly defined needs are discussed. These opportunities include enhancement of clinical risk stratification due to improved diagnostic capabilities, optimization of disease therapy based on molecular target characterization, and improved efficacy assessments. The current regulatory landscape for new probe development is also reviewed, with a focus on recent FDA guidance to facilitate early clinical development of promising probes. Specifically, the topics covered include the following:
|
Basic Scientific Themes for Cancer Imaging Probes |
|---|
|
TopAbstractValue of Imaging-Based...Basic Scientific Themes for...Imaging Fundamental Properties...Clinical Applications of Imaging...Opportunities for Probe...SummaryReferences |
|---|
|
Approaches being explored to address these hurdles include employing small peptides and antibody fragments that have limited antigenicity. Developments in peptide chemistry have improved targeting and facilitated synthesis, radiolabeling, and linkage to chelators of small peptides or peptide fragments. Likewise, small-molecule probe discovery has been facilitated by novel chemical synthesis techniques, such as in vitro click chemistry. This approach employs the intended biological target (e.g., an enzyme) in the chemical assembly of inhibitors from complementary building block reagents (21, 22). Strategies, such as high-throughput screening, phage display, and nanotechnology, are also being applied in imaging probe development. Progress has been achieved in developing lower thresholds of detection, multivalency to improve target affinity, and cellular internalization and biological trapping of imaging ligands. As an example, small-molecule probes that are trapped on activation at the active site of the enzyme have been applied to image targets of the many drugs that bind to the ATP-binding pocket of kinases. Practical limitations include competitive binding of the probe, which is present in low concentrations, by high ATP levels. Such was found to be the case with ML01, a 18F-labeled reversible inhibitor of EGFR (23). ML03, a 11C-labeled irreversible EGFR inhibitor, was not subject to washout by ATP due to covalent binding of the probe at the tyrosine kinase domain of the receptor; however, nonspecific chemical reactivity limited the bioavailability and tumor accumulation of the probe (24). A newer generation of irreversible EGFR inhibitor probes has greater stability and thus improved potential for imaging EGFR (25). A separate issue to consider is the promiscuity of such inhibitors for other kinases as was addressed recently by determining the in vitro binding affinity of 20 inhibitors using a panel of 119 related protein kinases (26). The study identified a wide range of specificity that correlated with neither the chemical structure nor the intended drug target.
Together, these findings highlight the difficulties in imaging tyrosine kinases and other key molecular targets. However, there have been several significant successes. For example, several chemotherapeutic drugs have been labeled to assess drug pharmacokinetics as well as multidrug resistance (see Table 1). In general, these examples involve tracer doses of nontargeted drugs. A comparable approach employing radiolabeled analogues of molecularly targeted therapies also has promise for elucidating not only pharmacokinetic properties but also the specificity of the drug for the purported molecular target. Examples include a recent in vitro study of the inhibitory properties of biotinylated, [125I]iodophenylated, and fluorescent analogues of the phosphatidylinositol 3-kinase inhibitor wortmannin (27). Peptide and antibody ligands linked to radionuclides and cytotoxins have been successful in cancer therapy (e.g., in treatment of hematologic and neuroendocrine tumors), and the radiolabeled peptide and antibody probes have been applied in oncologic imaging (see Table 1). Primarily, these agents have comprised antibodies and peptides that bind with high affinity to key receptors (e.g., somatostatin, bombesin, gastrin-releasing peptide, vasoactive intestinal peptide, and cholecystokinin receptors) or antigens [carcinoembryonic antigen (CEA), prostate-specific membrane antigen, and CD20]. Another promising area in which considerable progress has been made in target selection and signal amplification is for probes that image proteases. Protease cleavage activates the probe, resulting in high target-to-background signal ratios. Several proteases are up-regulated in cancer, including cathepsins, matrix metalloproteinases, urokinase-type plasminogen activator, etc. (28–30). These proteases play key roles in disease progression related to invasion, metastasis formation, high growth rates, and microenvironment host response and may represent key targets for imaging to detect cancers, measure their aggressiveness, and report on therapeutic efficacy of protease inhibitors. Optical protease imaging agents were first synthesized in the late 1990s, with magnetic resonance agents appearing more recently (31). The most efficient optical preparations have a 10- to 1,000-fold signal amplification on enzyme activation in preclinical in vivo studies.
A further consideration in molecular imaging is that up-regulation of some key oncologic molecules, such as EGFR, may be minimal or not highly correlated with outcome, limiting the value of measuring overexpression. Indeed, as shown in Fig. 2, other mechanisms, such as mutation and amplification, contribute more prominently to the anomalies seen in cancer, are more predictive of outcome, and are important targets for current and future imaging probe development (32). Another possibility is to image downstream molecules and events that may better reflect drug activity. Increasingly, molecular imaging probes are being developed to target proliferation, apoptosis, angiogenesis, hypoxia, and other cellular processes that are essential to carcinogenesis (Table 1). These in turn reflect tumor cell turnover, physiology, vitality, and blood supply. Such probes can be used in combination with those that enhance imaging of key molecular targets (e.g., proteases and receptor tyrosine kinases) that are prominently up-regulated in carcinogenesis. This dual-probe methodology would examine both molecular target modulation and downstream effect (see Fig. 3). For example, a decrease in phosphorylated extracellular signal-regulated kinase will occur when EGFR tyrosine kinase is inhibited and may in fact be a better indicator of inhibitor activity. Moreover, imaging the downstream cellular consequence of inhibition of phosphorylated extracellular signal-regulated kinase (e.g., increased apoptosis and reduced proliferation) may provide additional, complementary information about drug activity. This approach could also be used to improve selection of patients for EGFR inhibitor therapy and may prove more successful than imaging the actual drug target.
|
|
|
Imaging Fundamental Properties of Neoplasia |
|---|
|
TopAbstractValue of Imaging-Based...Basic Scientific Themes for...Imaging Fundamental Properties...Clinical Applications of Imaging...Opportunities for Probe...SummaryReferences |
|---|
Proliferation imaging. The nucleoside thymidine can be incorporated into DNA by either the salvage pathway or the de novo pathway of DNA synthesis. The salvage pathway directly reflects proliferative activity and entails uptake, sequential phosphorylation, and ultimately DNA incorporation of exogenous thymidine. For >40 years, [3H]thymidine incorporation has been the gold standard for assessing proliferation in vitro. Thymidine labeled with positron-emitting nuclides can enable in vivo imaging using PET. Like the tritiated derivative, thymidine labeled with 11C in the pyrimidine ring provides the authentic substrate for the salvage pathway transporters, thymidine kinase, and DNA polymerase and is incorporated into DNA. However, the 11C derivatives are also substrates for the catabolic pathway, yielding 11C-CO2 and other labeled derivatives (33). Validated kinetic models of [11C]thymidine that account for this metabolism have been used to characterize human tumors, including response to chemotherapy directed at the de novo DNA synthesis pathway (34). [124I]Iododeoxyuridine has also been used to image proliferation (35) and response to therapy (36), but the approach is limited by the long half-life of the isotope, radiation dose, and hydrolysis of the iodine label.
The deoxyribose group of thymidine can also be labeled with 18F at either 2'-arabino position or 3'-deoxy position. The former approach provides 1-(2'-deoxy-2'-fluoro-ß-D-arabinofuranosyl) thymine or 2'-deoxy-2'-fluoro-5-fluoro-1-ß-D-arabinofuranosyluracil (37, 38), whereas the latter yields 3'-deoxy-3'-fluorothymidine (FLT; ref. 39). The resulting fluorinated derivatives have practical advantages, including the longer half-life of 18F versus 11C (110 versus 20 minutes, respectively). However, these derivatives interact differently with nucleoside transporters and are poorer substrates than thymidine for thymidine kinase. In addition, 1-(2'-deoxy-2'-fluoro-ß-D-arabinofuranosyl) thymine reacts preferentially with the mitochondrial rather than the cytosolic thymidine kinase isoform, which lacks specificity for the S phase of the cell cycle. Although FLT is not degraded by thymidine phosphorylase (40), it is glucuronidated. FLT is trapped intracellularly in a manner analogous to [18F]fluorodeoxyglucose (FDG) because, following initial phosphorylation by thymidine kinase, the resulting monophosphate cannot be incorporated into DNA due to lack of a hydroxyl group. In general, FLT provides complementary information to FDG based on differences in regional distribution (39, 41). FLT uptake [expressed as the maximum standard uptake value (SUV)] at 60 minutes is roughly correlated with Ki-67 index, an immunohistochemical indication of proliferation (42). Although not likely to replace FDG for tumor detection and staging, FLT offers improved sensitivity or specificity in certain clinical situations (e.g., for distinguishing radionecrosis from recurrent brain tumor, for detecting indolent lymphomas, or for discriminating inflammation from tumor). FLT also has great potential utility in following response to therapy. A responding tumor cell may continue to metabolize FDG to maintain ion gradients or to provide energy for the P-glycoprotein (P-gp) pump function or protein biosynthesis; however, it will not synthesize new DNA and will not accumulate FLT. A decrease in DNA synthesis is likely following either cytostatic or cytotoxic therapy, highlighting the general utility of FLT PET for detecting response. Early clinical studies to quantify response will benefit from using rigorously quantitative methods to distinguish thymidine delivery and transport from thymidine kinase enzyme activity (43, 44). Comparative studies with SUV should lead to efficient protocols to assess proliferation in oncologic drug development.
Apoptosis imaging. Programmed cell death is an essential component of normal human growth and development, immunoregulation, and homeostasis. Cancer is as much a failure of apoptosis as it is a result of unchecked proliferation (45). Apoptosis also likely plays a significant role in cancer response to therapy. Many radiation and chemotherapy regimens kill cells by inducing apoptosis, and the development of resistance to apoptosis commonly limits response to cancer treatment (46). For this reason, a probe to noninvasively measure apoptotic cell death could prove useful for assessing the clinical response to chemotherapy.
When tumor cells detect DNA damage beyond the repair capability of the cell, programmed cell death pathways are triggered. Many of these are p53 dependent; the p53 gene is mutated in the majority of human cancers (47). Apoptotic cell death can be initiated through an extrinsic pathway involving activation of cell surface death receptors or by an intrinsic pathway via the mitochondria (48). Both pathways lead to activation of effector caspases that trigger a proteolytic cascade resulting in fragmentation of intracellular components. One of the earliest effects of caspase activation is the disruption of the translocase system that normally maintains phosphatidylserine on the interior of the cell membrane. Together with up-regulation of a scramblase activity that also occurs on caspase activation, this results in the redistribution of phosphatidylserine to the outer membrane leaflet, where it serves as a signal to phagocytic cells to engulf and digest the membrane-enclosed apoptotic cells (49).
Annexin V is a 36-kDa protein that binds with high affinity (Kd, 
10 nmol/L) to externalized phosphatidylserine. Annexin V staining has become a standard histopathologic measure of apoptosis. The 99mTc-labeled Annexin V imaging probe lacked specificity in early clinical testing (50, 51). In addition to concerns about probe formulation, a separate issue was the apparent lack of specificity of phosphatidylserine expression for apoptotic cells. Positron-emitting Annexin V probes (labeled with 124I and 18F) were developed to take advantage of the higher resolution and improved quantitation with PET (52, 53). In initial preclinical validation studies, [18F]Annexin V correlated with the terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling in vitro assay (54). Fluorophore and iron oxide Annexin V derivatives are also being explored (55–59).
Active research is ongoing to measure apoptosis using other molecular effectors or inhibitors, including members of the inhibitor of apoptosis family of proteins, which function by binding and suppressing caspases (60, 61). Certain cancers overexpress the various inhibitors of apoptosis proteins (e.g., survivin, X-linked inhibitor of apoptosis, and livin), and small-molecule suppressors are being tested in early in vitro and in vivo studies. Short peptides that can reverse the resulting caspase inhibition are also being explored. Other targets include the Fas-associated death domain–like interleukin-1ß-converting enzyme inhibiting protein, which seems to be an important determinant of resistance to apoptosis induction. Following validation in animals and successful early clinical testing, apoptosis imaging probes could be applied in the evaluation of cancer therapies, particularly for lymphoma and leukemia (62, 63). They could be especially useful in helping sequence multiagent treatment strategies.
Angiogenesis imaging. Angiogenesis is a key oncologic process that is essential for tumor growth and for the initiation of metastasis (64). Imaging modalities for detecting angiogenesis include methods to assess blood volume and flow and to derive semiquantitative and quantitative kinetic hemodynamic variables. In addition, targeted probes can be used to visualize molecular effectors of angiogenesis, such as VEGF and 
Vß3 integrin (65). One approach for steady-state imaging of blood flow is to use magnetic nanoparticles, which have a long-lived intravascular nature. Because the images cover large areas of the body, both primary malignancies and metastases could potentially be evaluated (66–69). In tumor xenograft models of varying degrees of angiogenesis, magnetic nanoparticles selectively enhanced the vascularity without significant leakage into tumor interstitium (68). The models were characterized by determining microvessel counts, VEGF production, and global tumor intravascular volume fraction using a validated 99mTc marker (70, 71). In these experimental tumor models, steady-state measures of vascular volume fraction with MRI provided a volumetric, in vivo, noninvasive assay of microvascular density. Experiments are ongoing to determine the in vivo sensitivity of this steady-state technique for investigating antiangiogenic therapies in animal models and humans.
Neovascular density can also be imaged by direct or indirect specific molecular targeting. One preferred target has been the integrin Vß3. Although also expressed on endothelial cells, Vß3 integrins are found on a wide range of tumor cells, including the MDA-MB-435 breast and B16B15b melanoma cells (72) as well as human lung carcinoma (73) and melanoma (74). This receptor is up-regulated in angiogenic endothelium and has been imaged using an Arg-Gly-Asp–containing peptide with high affinity for V integrins or using antibody-conjugated nanoparticles (72, 75–79). Using a nanoparticle comprising an antibody to the integrin Vß3 ligand, micrometastases could be detected and characterized in a melanoma mouse model (80). The same ligand has been exploited in other MRI-targeted approaches either using antibodies conjugated to liposome nanoparticles sequestering gadolinium or other direct antibody conjugations to nanoparticles (81).14 E-selectin offers another target that has been exploited by MRI using either paramagnetic (82, 83) or superparamagnetic nanoparticle approaches (84). Another exciting target that has recently been explored is vascular cell adhesion molecule 1.
Hypoxia imaging. In solid tumors, hypoxia may result from unregulated cellular growth, but it is also a common attribute of the tumor phenotype and may even be a factor in tumorigenesis. Hypoxia induces tissue changes that result in selection of cells with mutant p53 expression (85–89). Indeed, after DNA damage, hypoxic cells do not readily undergo apoptosis (90, 91). Hypoxia enhances expression of endothelial cytokines, such as VEGF, interleukin-1, tumor necrosis factor-, and transforming growth factor-ß, and a cellular O2 sensing mechanism triggers production of hypoxia-inducible factor-1 (92). A subunit of the basic-helix-loop-helix transcription factor hypoxia-inducible factor-1 that is activated by redox-dependent stabilization, hypoxia-inducible factor-1 induction initiates a cascade of events culminating in angiogenesis. Tumor hypoxia and hypoxia-inducible factor-1 activation may also contribute to the metabolic switch to glycolysis that characterizes tumor cells (93). However, many cancer cells use glycolysis for energy production regardless of the availability of oxygen, suggesting that the two processes are independent (94).
Hypoxia is prevalent in nearly all tumors studied and predicts radiation response in sarcoma, glioma, and cancers of the uterine cervix, lung, and head and neck (for recent reviews, see refs. 95, 96). In addition to significant interpatient differences in the distribution of hypoxia, microscopic heterogeneity also occurs within a tumor (97). The level of oxygenation, even in well-perfused tissues, is extremely variable. In patients with head and neck cancer, oxygenation at one site correlated well with other sites (98). Hypoxia in tumors does not depend on tumor size, grade, and extent of necrosis or blood hemoglobin status and seems to be an independent predictor of outcome (96, 99, 100).
In addition to inducing radioresistance, hypoxia promotes resistance to several chemotherapeutic agents potentially through three mechanisms. Hypoxia impedes drugs from reaching the cells from blood vessels, slows proliferation, and promotes gene expression changes that enable cellular rescue from severe damage (101, 102). To address the clinical problem of tumor resistance associated with hypoxia, the focus for several decades has been to improve O2 levels in the tumor environment or to use O2-independent irradiation (neutrons), altered fractionation, or radiosensitizers. However, these techniques have been associated with problems of low availability or serious clinical toxicity. An alternative strategy is to selectively target hypoxic cells with hypoxia-activated prodrugs. The introduction of the relatively nontoxic, bioreductive, hypoxia-activated prodrug, tirapazamine, has rekindled interest in identifying patients with tumor hypoxia (100). A second way to exploit hypoxia is to employ therapies that take advantage of hypoxia-inducible factor-1 activation under hypoxic conditions (103).
PET imaging is an ideal modality for evaluating hypoxia. It images the entire tumor and is less operator-dependent than oxygen electrodes. Its safety profile and noninvasiveness make it useful in patient follow-up. The PET imaging agent fluoromisonidazole is a 18F-labeled fluorinated derivative of misonidazole, an azomycin hypoxic cell sensitizer introduced two decades ago, which binds covalently to intracellular molecules at a rate inversely proportional to intracellular O2 concentration. Its uptake in hypoxic cells is dependent on the reduction of the nitro group on an imidazole ring (96). Fluoromisonidazole is easily synthesized (104), has a long record of use in humans, involves only modest radiation exposure (105), and has undergone extensive clinical validation. Second-generation agents include 2-(2-nitro-1H-imidazol-1-yl)-N-(2,2,3,3,3-pentafluoropropyl) acetamide (106), fluoroerythronitroimidazole, a more hydrophilic variant of fluoromisonidazole, and fluoroetanidazole, which has binding characteristics similar to fluoromisonidazole but with decreased hepatic retention and fewer metabolites (107, 108). The nitroimidazole derivatives have similar blood clearance characteristics despite different partition coefficients. A 64Cu-labeled acetyl derivative of pyruvaldehyde diacetyl-bis(N4-methylthiosemicarbazone) copper (II) and 124I-labeled iodo-azomycin-galactoside have the potential advantage of longer half-lives for clinical use (109–111).
Because hypoxia is associated with poor response to both radiation and chemotherapy, identifying hypoxia should have prognostic value. Recent advances in patient-specific radiation treatment planning, such as intensity-modulated radiotherapy, have enabled customization of radiation delivery based on physical conformity, but it can also incorporate variables, such as hypoxia, proliferation, and tumor burden to generate a biological profile of the tumor (112). In addition to its potential for defining the biological microenvironment of a tumor, hypoxia imaging also can help in selecting and directing the appropriate treatment, both radiation and chemotherapy (95). Hypoxia imaging during treatment might also enable advantageous treatment modifications.
|
Clinical Applications of Imaging Probes |
|---|
|
TopAbstractValue of Imaging-Based...Basic Scientific Themes for...Imaging Fundamental Properties...Clinical Applications of Imaging...Opportunities for Probe...SummaryReferences |
|---|
Molecular imaging of estrogen receptor for breast cancer. Hormonal therapy of breast cancer is one of the earliest successes of targeted therapy, with the estrogen receptor (ER) as the therapeutic target (113). ER is expressed in most breast cancers; in most cancers expressing ER, interrupting estrogen-stimulated proliferation halts tumor growth and leads to tumor regression. Current hormonal therapies function by competing with estrogen metabolites for the receptor (tamoxifen), by inducing ER degradation (fulvestrant), or by lowering agonist concentration (aromatase inhibitors; ref. 114). The success of aromatase inhibitor therapy has increased the clinical utility of hormonal therapy in both primary and recurrent diseases. It has also spurred interest in novel hormonal agent development and in overcoming clinical hormone resistance. In parallel, there has been increased interest in characterizing ER expression and estrogen binding in vivo with PET to guide hormonal therapy in clinical trials and clinical practice.
Although no compound is currently in clinical use, several agents have been tested for PET ER imaging (115), and new compounds continue to be evaluated (116). The close analogue of estradiol, 16-fluoroestradiol-17ß (FES; ref. 117), has had the best results for imaging and quantifying the functional ER status of breast cancer by PET. The quantitative level of FES uptake in primary tumors correlates with the level of ER expression measured by in vitro radioligand-binding assay (118) and, in preliminary studies, by immunohistochemistry (119). FES PET provides sufficient image quality to visualize metastatic lesions with high sensitivity in patients with ER-positive tumors (120, 121).
As a quantitative, noninvasive measure of regional estradiol binding to ER, FES PET provides capability not possible by biopsy and in vitro assay. FES PET can provide specific characterization of sites identified by nonspecific methods, such as CT or FDG-PET. In a diagnostic sense, documentation of estradiol binding at sites suspicious for breast cancer recurrence or metastasis provides highly specific evidence of breast cancer (121). In assessing ER expression, FDG and FES PET are complementary. FDG identifies active sites of cancer and can be used to indicate where to interrogate the FES PET scan for tumor ER expression. This approach is especially important for identifying active areas of disease with low or absent ER expression but requires two different scanning days, because both are 18F-labeled compounds. Image interpretation and analysis is facilitated by FDG/FES PET image coregistration. Combined PET/CT devices, which fuse functional and anatomic imaging, may also be helpful in this regard.
Perhaps most importantly, FES PET may provide a predictive assay, in analogy to in vitro assay of ER expression in biopsy material, for directing hormonal therapy in drug trials and in clinical practice. FES PET can assess ER expression at all sites of disease in patients with large primary breast cancers and/or metastatic lesions and overcome potential sampling error inherent to in vitro assays (122). As shown in Fig. 4, this could be especially valuable when evaluating recurrent disease. Dehdashti et al. (120) showed that, by FES/FDG-PET, 15% of patients with metastatic disease from an ER-positive primary tumor have one or more ER-negative metastatic sites. This is consistent with recent studies based on biopsy (123). Low or absent ER expression predicts a very low likelihood of response to hormonal therapy (124). Mortimer et al. (125) showed that, for patients with locally advanced or metastatic breast cancer, low FES uptake (SUV < 2) before primary tamoxifen therapy predicted nonresponse. In a recent study, Linden et al. (126) showed similar results for a heavily pretreated population with recurrent or metastatic breast cancer and found that only patients with at least modest pretherapy FES uptake (SUV > 1.8) had an objective response to hormonal therapy, mostly aromatase inhibitors. In the Linden et al. study, a hypothetical treatment algorithm using FES PET to select patients for hormonal therapy would have increased the response rate from 25% to 50%, without inappropriately withholding therapy from patients destined to respond.
|
In summary, PET ER imaging using FES has been validated and tested in preliminary patient studies as a tool for measuring ER expression at all sites of disease and as a predictive assay for hormonal therapy. Formal validation in prospective clinical trials is required. Future applications as a tool for hormonal agent drug development and testing as well as for use in clinical practice to guide hormonal therapy seem quite promising and likely.
Molecular imaging of androgen receptor in prostate cancer. Blockade of androgen action by medical or surgical means is standard front-line therapy for patients with advanced prostate cancer. Most traditional approaches are focused on reducing the production or action of the native ligand, testosterone. Testosterone is converted in the prostate to dihydrotestosterone, which in turn binds to and activates the androgen receptor. The activated receptor regulates the transcription of a range of target genes. The clinical benefits of castration, as practiced traditionally, are focused on the ligand rather than on the receptor. Response is a function of the degree to which the tumor is dependent on androgens for growth and survival. It is not, however, a curative strategy; over time, virtually all tumors progress. In the clinical state model of prostate cancer progression, patients who have failed androgen ablative therapies are separated based on the measured testosterone concentrations in the blood as either noncastrate or castrate (130). Castration-resistant disease represents the lethal variant of prostate cancer, as the majority of patients eventually die of disease.
Much research has focused on the mechanisms of tumor progression despite castrate levels of testosterone. The androgen receptor seems to remain functional, and signaling is sufficient to support continued tumor growth and progression. The evidence includes increased prostate-specific antigen levels at the time of progression, the responses to second-line and third-line hormonal manipulations that further reduce the ligand by blocking adrenal androgen synthesis, the steroid hormone withdrawal responses, and the benefits of nilutamide and bicalutamide (agents that act exclusively by androgen receptor binding).
Recent research by several groups suggests several nonmutually exclusive mechanisms of continued androgen receptor signaling, including the following: mutations in the ligand-binding domain leading to promiscuous activation by a range of steroidal hormones and nonsteroidal antiandrogens; increased levels of wild-type receptors; increased intratumoral levels of adrenal androgens, which may be related in part to an increase in adrenal androgen synthetic enzymes; and ligand-independent activation of the receptor by growth factors, such as receptor tyrosine kinases and cytokines. Although a detailed discussion of these mechanisms is beyond the scope of this article, they do provide a rationale for the development of therapies targeting the receptor directly and for the ability to visualize the receptor in vivo. The clinical development of such therapies (including selecting patients most likely to respond to a given treatment, determining a dose and schedule, etc.) would be enhanced significantly if it were possible to assess the presence and the level of receptor within a tumor and to measure changes in the receptor following treatment.
16-Fluoro-5-dihydrotestosterone (FDHT) was developed and characterized in the mid-1990s (131). This radiotracer binds with high affinity to androgen receptor and has been employed to study androgen receptor expression in animals and humans. This agent, as with FES, has been used only in experimental clinical trials. Initial clinical studies included a few patients with progressing, androgen-independent prostate cancer manifested by rising prostate-specific antigen values documented on three or more occasions, castrate (<50 ng/mL) levels of testosterone, and metastatic disease visible by conventional imaging (CT or MRI and/or bone scan). The characteristics of androgen receptor expression in prostate cancer metastases were compared with [99mTc]medronate bone scan and [18F]FDG metabolism (132, 133). A secondary goal was to characterize biodistribution, metabolism, and radiation dosimetry (133). Of the 59 lesions found by conventional imaging methods, 57 (97%) were positive with FDG, with an average SUV of 5.2. Of these metabolically active lesions, 48 (80%) were FDHT positive, with an average SUV of 5.3. In two patients treated with testosterone, FDHT tumor uptake was reduced (Fig. 5). This small study suggests that FDHT is actively concentrated in metastatic tumor sites of patients with androgen-independent prostate cancer, that uptake can be blocked by circulating androgen, and that this radiotracer is likely to be a useful tool for helping to understand the role that the androgen receptor may play in prostate cancer progression. The study also found a rapid conversion of the radiotracer to inactive metabolites, with 80% of FDHT disappearing from the plasma within 10 minutes.
|
In summary, FDHT-PET imaging is feasible in patients with advanced prostate cancer with testosterone levels that are in the castrate range. There is active uptake of FDHT in the majority of metastatic lesions that are detected by conventional imaging. These recent data suggest that FDHT will be a valuable tracer for studying the biology of prostate cancer metastasis and for determining the effectiveness of androgen receptor blockade in these patients.
Imaging multidrug resistance. Molecularly targeted imaging probes have a potential role in identifying acquired resistance to chemotherapy, which is important in managing the treatment of most cancer patients. Doxorubicin (Adriamycin), vincristine (Oncovin), etoposide (Vepesid), taxanes (Taxol, Taxotere), and other chemotherapy drugs are subject to drug resistance as a result of increased expression of P-gp or related membrane transporter proteins (135–137). Imaging probes can quantify efflux of these drugs via the P-gp and other transporters.
One approach to imaging multidrug resistance in experimental studies measures the tumor efflux level of [11C]verapamil, a well-characterized substrate for P-gp (138–141). In one study, the tumor clearance half-life of [11C]verapamil was relatively fast (high P-gp; ref. 140), which could explain the inability of [11C]verapamil coadministration to modulate resistance to chemotherapeutics, such as taxanes in solid tumors. In nonhuman primates, brain uptake increased 2-fold when P-gp was inhibited by cyclosporine A (141). These results were confirmed in a clinical study, showing that [11C]verapamil PET can be used to measure P-gp inhibition (142). Another effective imaging substrate for multidrug resistance that has been well studied in several human cancer settings involving chemotherapy resistance is [99mTc]sestamibi.
A standardized PET methodology for assessing the activity of the P-gp would provide important information about drug resistance in cancer patients at the time of diagnosis. It could be used to test the variability of this mechanism among patients and the extent to which P-gp resistance increases with exposure to chemotherapy. Institution of multidrug chemotherapy following early recognition of P-gp expression is called risk-adapted therapy, an extremely important concept for managing care of individual cancer patients that has increased survival in osteosarcoma patients (137).
Peptide probes in neuroendocrine tumor imaging. Oncologic peptide research has focused on identification of suitable targets (e.g., overexpressed peptide receptors) as well as the discovery and development of radiolabeled agents that can interact with these targets for therapeutic and imaging purposes. New targets include the gastrin-releasing peptide receptors in prostate and breast cancer and neuropeptide Y receptors in breast cancer. As noted above (see Fig. 2), deeper insight into receptor pathogenesis in cancer is needed, particularly with regard to the mechanisms of up-regulation in primary tumor as well as metastases, functional activity of receptors and receptor-ligand binding in tumors and peritumoral tissues, effect of receptor homodimerization and heterodimerization, and activity of endogenous peptide substrates. Several peptide receptor-binding compounds have been radiolabeled and are currently undergoing in vitro testing, in vivo validation, and clinical trials for radiotherapy or imaging; examples in clinical use are summarized in Table 1. Limitations of imaging specificity for these peptide probes include physiologic uptake in many inflammatory and granulomatous conditions (143). In addition, the signal intensity correlates better with receptor density than tumor size but is highly predictive of response to peptide analogue or radioligand therapy. A reduction in intensity during therapy could signal differentiation rather than response; combined FDG-PET scans can help discriminate these two processes.
One prominent successful group of peptide probes is the small (1.5 kDa) radiolabeled peptide analogues of somatostatin, the primary imaging agent for neuroendocrine tumors (144). Somatostatin receptors are expressed at high density in most neuroendocrine tumors as well as in pancreatic (insulinomas) and other cancers. Five receptor subtypes have been identified, which variously affect multiple cellular signaling cascades (including mitogen-activated protein kinase, phospholipase A2, and cyclic AMP protein kinase A pathways). Tumors frequently express two or more subtypes, and homodimerization and heterodimerization can result in a range of functional characteristics. Based on the known mechanisms of action of somatostatin analogues, the receptors directly control proliferation and apoptosis as well as mediate indirect effects. The latter range from growth factor and hormone secretion to angiogenic and immunomodulatory effects. Radiolabeled somatostatin analogues include the approved agent diethylenetriamine pentaacetic acid-D-Phe(1)-pentetreotide (Octreoscan), widely used in clinical practice (see Table 1). Metal chelators (e.g., 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid) have improved the stability of the radioconjugates, enabling incorporation of various radionuclides, such as 68Ga for PET or 90Y for radiotherapy.
Radiolabeled somatostatin analogues are well tolerated and have high affinity for the receptor. The most common somatostatin analogues (octreotide, lanreotide, and vapreotide) bind mainly with receptor subtypes 2 and 5. However, radiolabeled depreotide (Neotect) also has affinity for subtype 3 and can image tumors expressing these receptors, such as insulinomas. Most (85%) carcinoid neuroendocrine tumors, which have an 80% to 90% incidence of subtype 2A receptors, can be imaged with 111In-labeled pentetreotide. Somatostatin receptor imaging can also identify and localize metastatic lesions and is thus useful in staging and treatment and surgical planning. In one study, 111In-labeled pentetreotide altered patient management in 47% of 122 gastrinoma patients (145). Further, as also discussed below, imaging (e.g., using 86Y-labeled somatostatin analogues; ref. 146) to define patient-specific dosimetry is being explored. In addition to their role in neuroendocrine tumors, which express a high density of somatostatin receptors, these imaging agents also have application in a wide range of tumors. These include meningiomas and medulloblastomas as well as cancers with low and heterogeneous receptor expression, such as small cell lung cancer, breast cancer, gastrointestinal cancers (including colorectal cancer), lymphoma, and renal cell carcinomas. Ongoing effort is focused on developing radiolabeled somatostatin analogues with enhanced resistance to proteases, specificity for certain somatostatin receptor subtypes or with broad specificity, or enhanced uptake.
Antibody probes: progress in clinical imaging. Radiolabeled antibodies take advantage of the natural biochemical specificity of the immune system to provide molecularly targeted agents for cancer imaging (147–150). In principle, because antibodies can be made against virtually any biomolecule, there is opportunity to develop radiotracers that target many physiologic and pathophysiologic processes of importance to cancer biology, diagnosis, and therapy. Early pioneering work in this field, which began in the 1950s, was reviewed by Pressman (151). Despite the long history of active research, development of radiolabeled antibodies as practical diagnostic and therapeutic radiotracers has been slow to mature. Impediments to progress include the following:
Several first-generation radiolabeled antibody products have received new drug application approval by the FDA as suitable for diagnosis or therapy of human tumors (see Table 1). Differentiation antigens that can serve as tumor burden markers have had the greatest success to date. For example, CEA was one of the earliest targets for detection and radiotherapy. Extensive work in the late 1970s and early 1980s used a variety of antibodies primarily labeled with iodine radionuclides (147, 156). Despite high rates of detection, the methods depended on 
-imaging and were considered too insensitive for routine use. The commercial product arcitumomab (CEAscan) is a 99mTc-labeled F(ab') fragment that addresses several limitations of the radioiodine-labeled whole IgG products. First, compared with whole IgG, the ab' fragment localizes much faster to tumor and is more rapidly cleared, thus providing greater tumor to normal tissue contrast. Secondly, F(ab') fragments are significantly less immunogenic than whole IgG; thus, repeated imaging is possible even when the product is based on a murine antibody (157). Finally, the 99mTc label is more ideally suited for single photon emission CT imaging, which affords better contrast than conventional -imaging and is therefore superior for detecting tumors deep within the body. Several successful clinical trials have been reported with arcitumomab (158), including for detection of occult colorectal cancer when conventional imaging methods, such as CT, are negative (159). However, FDG-PET is preferentially used for detecting recurrent colorectal cancer due to its improved accuracy (160) and ease of implementation compared with the single photon emission CT arcitumomab technique. Arcitumomab is also effective in detecting medullary cancer of the thyroid (161). These tumors are difficult to diagnose and may be missed by FDG-PET if well differentiated, but FDG-PET and arcitumomab scans have not been directly compared in medullary thyroid cancer.
A second approved antibody probe is capromab pendetide (ProstaScint). An IgG1 murine monoclonal antibody (7E11-C5.3) conjugated to the linker-chelator, glycyl-tyrosyl-lysyl-diethylenetriamine pentaacetic acid, capromab pendetide reacts with an intracellular or internal epitope of prostate-specific membrane antigen. Prostate-specific membrane antigen is widely expressed in prostate and prostate cancer but not in other tissues. 111In-labeled capromab pendetide is most useful in patients with high-risk, primary prostate cancer (Gleason 
7 with prostate-specific antigen > 10 times the normal; refs. 162, 163) or suspected recurrent prostate cancer after prostatectomy. In a large clinical trial, the advantages derived primarily from detection of soft tissue and especially lymph node metastases (164). Nevertheless, the benefits of this technique remain controversial, as bone metastases are the primary cause of morbidity and mortality, and these lesions are more accurately detected by bone scanning. The imaging technique is difficult to perform, and fusion imaging of the capromab pendetide single photon emission CT with CT or MRI is a highly valuable adjunct (164). In addition, it has been reported that capromab pendetide lacks utility for predicting response to salvage radiation therapy (165).
Radiolabeled CD20-reactive antibodies have been used to treat non-Hodgkin's lymphoma. These include 131I-labeled tositumomab (Bexxar), to which responses are seen in about two thirds of patients who have failed rituximab (Rituxan) and/or chemotherapy. Half are complete responses, and the response duration is 1 year (166). The treatment regimen entails immediate prior administration of nonradioactive tositumomab together with a tracer amount of its radiolabeled counterpart. This dosimetric dose is used to assess the patient-specific pharmacokinetics at three time points using -imaging, from which the subsequent therapeutic dose is calculated. 131I-labeled tositumomab is not approved for use in first-line therapy, but initial results in this setting are promising, with most patients in long-term complete remission (167). 90Y-labeled ibritumomab tiuxetan (Zevalin) is also a CD20-reactive antibody with activity in non-Hodgkin's lymphoma. The first step in the treatment regimen is imaging of 111I-labeled ibritumomab tiuxetan combined with the unlabeled antibody, which is administered following rituximab, a nonradioactive antibody of the same reactivity to CD20 antigen. A therapeutic dose of 90Y-labeled ibritumomab tiuxetan is then given together with the unlabeled antibody. Compared with 131I, the advantages of 90Y include its significantly better imaging characteristics and the improved radiation safety for health care workers. Although the products have not been directly compared, equivalent long-term response rates for 90Y-labeled ibritumomab tiuxetan and 131I-labeled tositumomab are reported in patients with non-Hodgkin's lymphoma (168). These radiolabeled antibody therapies are actively being studied in non-Hodgkin's lymphoma, and novel agents targeting CD22, CD80, and other antigens are being developed.
Third-generation products for radioimmunodetection and targeted radiotherapy may be able to take advantage of key advances in targeting technology, such as multistep techniques (169, 170). One such example that has recently been applied to CEA imaging in preclinical validation tests is pretargeting, which addresses the limitation of the slow blood clearance of IgG. The technique involves signal amplification of the radiotracer in situ by pretargeting a multivalent, bispecific antibody with reactivity to both CEA and a small compound to which the radionuclide is attached (171). Improved imaging methods, such as PET, as well as labeling with near-IR-emitting fluorochromes also offer advantages (172–174), and rapid imaging using pretargeting and other techniques could enable application of PET radionuclides with short half-lives (e.g., 18F and 68Ga). A spectrum of antibodies and antibody derivatives, including minibodies and diabodies (175, 176), are now available based on advanced molecular engineering techniques, and these forms are likely to have targeting advantages when optimized. These new antibody forms are also likely to lack immunogenicity, permitting repeated diagnostic use, or therapeutic administration (177). Other possible applications include assessment of drug pharmacodynamics and treatment response based on in vivo antigen detection (178). In some cancers, particularly leukemias and lymphomas, -emitters with higher linear energy transfer rates may join ß-emitters (179, 180). These diverse advances in the technology of imaging and radiolabeling, and in molecular engineering of antibodies, will together provide opportunities to develop increasingly effective molecularly targeted agents for imaging and therapy.
|
Opportunities for Probe Development and Testing: Development Path and Regulations |
|---|
|
TopAbstractValue of Imaging-Based...Basic Scientific Themes for...Imaging Fundamental Properties...Clinical Applications of Imaging...Opportunities for Probe...SummaryReferences |
|---|
312; FDA recently updated its 1998 draft guidance regarding the development of imaging drug and biological products for approval indications.15 Limited enrollment, basic science clinical research projects with radiolabeled drugs, and imaging probes that are not new molecular entities can also be conducted in a non-IND setting as per the RDRC regulations set forth in 21 CFR 361.1. (Generally, a "new molecular entity" is an agent that has not previously been tested in humans.)
FDA has recognized that the traditional design of first-in-humans trials, which are intended to facilitate rapid ascertainment of dose-limiting toxicity and movement into Phase 2 testing, is not well suited for low-dose imaging, pharmacokinetic or pharmacodynamic assessments. Early information from such exploratory trials may significantly improve the subsequent drug development program. Therefore, FDA has developed draft guidance regarding exploratory (early Phase 1) clinical trials of new molecular entities conducted under the IND regulations.12 The guidance addresses approaches consistent with the flexibility of the IND regulations (21 CFR 312) regarding initial clinical testing in exploratory studies. Specifically, guidance is provided on the types and amount of preclinical data and chemistry, manufacturing, and controls information needed for IND applications if an exploratory study of limited expected exposure and duration (e.g., 7 days) is planned. In keeping with the FDA Critical Path Initiative, the FDA draft guidance aims to reduce barriers to initial clinical testing of promising agents and to decrease the resources required to develop agents with a higher potential risk of failure. The FDA guidance outlines several exploratory clinical study designs and the attendant preclinical data requirements as examples; additional approaches that are consistent with the IND regulations could be developed for specific needs.
For imaging probes, preclinical toxicology studies, chemistry, manufacturing, and controls requirements, and attendant scaleup synthesis are the greatest perceived hurdles to development and early clinical testing. In addition, because preclinical studies cannot always predict the clinical potential of an imaging probe, early exploratory clinical testing (including comparisons among multiple related candidates) is critical for early go/no-go decisions. Thus, microdosing and other exploratory studies provide an especially important opportunity for the development of novel imaging probes. For imaging agents that are not new molecular entities [i.e., agents with prior human exposure either endogenously or exogenously (e.g., drugs studied under a prior IND)], basic scientific research questions can be addressed in small clinical studies under the RDRC regulations. In addition, such established imaging probes can play a valuable role in several types of exploratory studies of novel oncologic drug products. The following sections discuss the U.S. regulations and draft guidances addressing the IND and non-IND pathways pertinent to early clinical imaging probe development. The European Agency for the Evaluation of Medicinal Products position article regarding the nonclinical safety studies required to support clinical microdose studies is also considered.
Exploratory studies
Imaging probe development using exploratory studies. Before the conduct of traditional dose escalation and safety studies during drug development, exploratory studies can provide early indications of the feasibility or promise of new molecular entities. Exploratory studies are defined as those of limited duration (
