targeted therapeutics directed to improve the outcome for patients with other types of cancers, he added.

MOLECULAR IMAGING

Recent advances in biomedical imaging provide potential opportunities to improve the discovery, development, and validation of novel therapeutics. Imaging applications offer the possibility to reduce the time, cost, and workload in drug development. In the session devoted to molecular imaging, five speakers addressed the current and near-horizon opportunities in molecular imaging, particularly how it can be used to detect biomarkers for assessing cancer treatment effectiveness, and the advantages it has over standard imaging. These speakers gave several promising examples of such molecular imaging biomarkers, and showed how they can be used throughout the drug development process. Also addressed in this session were current challenges involved with molecular imaging, suggestions for how to meet those challenges, and a discussion of how the current or potential pitfalls of molecular imaging compare to those of standard imaging.

“In the last 50 years, there have been tremendous advances of imaging,” pointed out Dr. Hedvig Hricak of Memorial Sloan-Kettering Cancer Center (MSKCC). Those advances include the development and modification of positron emission tomography (PET), MRI, and ultrasound, as well as various genetic engineering techniques that enable imaging of specific molecules or molecular processes. All this innovative imaging can reveal, in a dynamic manner, key biological functions within the body related to cancer progression and response to treatment.

Current and Developing Methods

The development of PET and the use of the radioactive tracer FDG ([18F]-2-fluoro-2-deoxy-D-glucose) as a biomarker for the heightened metabolism that occurs in cancer cells paved the way for using an imaging probe to examine a biologic process. “This was a milestone in which a tracer was recognized as showing a particular biologic function,” said Dr. Steven Larson of MSKCC, who noted that there are now more than 900 articles in medical journals related to PET imaging of tumor response. Such imaging is highly sensitive, with resolution down to 1 to 2 mm, and can indicate a response to treatment before standard CT imaging can. As Dr. Larson pointed out, PET imaging can show after just one or two treatments that a tumor is responding, even if it is not yet shrinking in size. The 1997 Food



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 40
0 IMPROVING THE QUALITY OF CANCER CLINICAL TRIALS targeted therapeutics directed to improve the outcome for patients with other types of cancers, he added. MOLECULAR IMAGING Recent advances in biomedical imaging provide potential opportunities to improve the discovery, development, and validation of novel therapeu- tics. Imaging applications offer the possibility to reduce the time, cost, and workload in drug development. In the session devoted to molecular imag- ing, five speakers addressed the current and near-horizon opportunities in molecular imaging, particularly how it can be used to detect biomarkers for assessing cancer treatment effectiveness, and the advantages it has over standard imaging. These speakers gave several promising examples of such molecular imaging biomarkers, and showed how they can be used through- out the drug development process. Also addressed in this session were current challenges involved with molecular imaging, suggestions for how to meet those challenges, and a discussion of how the current or potential pitfalls of molecular imaging compare to those of standard imaging. “In the last 50 years, there have been tremendous advances of imag- ing,” pointed out Dr. Hedvig Hricak of Memorial Sloan-Kettering Cancer Center (MSKCC). Those advances include the development and modifica- tion of positron emission tomography (PET), MRI, and ultrasound, as well as various genetic engineering techniques that enable imaging of specific molecules or molecular processes. All this innovative imaging can reveal, in a dynamic manner, key biological functions within the body related to cancer progression and response to treatment. Current and Developing Methods The development of PET and the use of the radioactive tracer FDG ([18F]-2-fluoro-2-deoxy-D-glucose) as a biomarker for the heightened metabolism that occurs in cancer cells paved the way for using an imaging probe to examine a biologic process. “This was a milestone in which a tracer was recognized as showing a particular biologic function,” said Dr. Steven Larson of MSKCC, who noted that there are now more than 900 articles in medical journals related to PET imaging of tumor response. Such imaging is highly sensitive, with resolution down to 1 to 2 mm, and can indicate a response to treatment before standard CT imaging can. As Dr. Larson pointed out, PET imaging can show after just one or two treatments that a tumor is responding, even if it is not yet shrinking in size. The 1997 Food

OCR for page 40
 WORKSHOP SUMMARY and Drug Administration Modernization Act (FDAMA) and subsequent key FDA approvals of the PET FDG tracer spurred major technologic development over the past decade, according to Dr. Larson. This develop- ment includes combined PET and CT imaging devices at many cancer treatment facilities, as well as the development of new molecular probes that could be imaged with PET or other modalities. Many of the newer cancer drugs target specific growth factors or their receptors that play a key role in the growth and spread of tumors. Conse- quently, researchers have developed several PET imaging probes for these compounds, including the receptors for HER-2, EGF, VEGF, estrogen, and androgen. Unlike standard imaging and non-imaging diagnostic techniques, molecular imaging of these growth factors or their receptors can reveal the heterogeneity of tumors and metastases. Dr. Hricak’s slides showed that imaging with a probe for the estrogen receptor could reveal in breast cancer patients which metastases are estrogen receptor (ER) positive and thus likely to respond to hormonal therapy (Figure 11). Researchers can also now use high-resolution MRI systems to create in reasonable time an anatomical map of the distribution of key metabolites relevant to cancer, pointed out Dr. John Gore of Vanderbilt University. Diffusion-rated MRI, which measures the degree to which water molecules are free to move around within tissue, can be used to measure cell density, which changes rapidly and early after particular cancer treatments, he said. He added that dynamic contrast MRI, by showing changes in blood volume and blood flow into tissues over time, is useful for detecting the abnormal proliferation and leaking blood vessels that typify malignancies. Farther in the future is the possibility of researchers using hyperpolarized carbon-13 as a radioactive label for tracers that can improve the sensitivity of MRI and enable the detection of specific metabolic pathways, as opposed to the heightened overall metabolism that is seen in PET with FDG. Even ultrasound has been adapted to image molecular functions. For example, investigators are experimenting with adding molecular probes to the surface of the microbubbles of air used as contrast agents in ultrasound imaging. One company has labeled these microbubbles with an antibody that will bind to VEGF, whose expression is elevated in various tumors. Such ultrasound imaging may someday provide a low-cost alternative to more expensive techniques such as PET and MRI, Dr. Gore noted. Genetic engineering techniques have enabled researchers to cre- ate probes for imaging the RNA or proteins expressed from specific genes in tumors, as well as to create innovations in optical imaging. Dr. David Piwnica-Worms of Washington University in St. Louis noted

OCR for page 40
 IMPROVING THE QUALITY OF CANCER CLINICAL TRIALS Pre-Treatment Post-Treatment FIGURE 11 Targeted treatment selection: [18F]-fluoroestradiol (FES) in predicting response to hormonal therapy. Both patient A and patient B have ER+ primary tumors and bone metastases. Patient A was strongly FES positive and showed excellent response (size and SUV) after 3 months on letrozole. Patient B was FES negative and showed progression at 6 months. Solid arrows point out tumor locations, and lower posttreat- ment signal density corresponds to positive therapeutic response. Dashed arrows show normal liver FES uptake. fig 11 ACRONYMS: FDG ([18F]-2-fluoro-2-deoxy-D-glucose). SOURCE: Hricak presentation (October 4, 2007) and Linden, H.M., S.A. Stekhova, J.M. Link, J.R. Gralow, R.B. Livingston, G.K. Ellis, P.H. Petra, L.M. Peterson, E.K. Schubert, L.K. Dunnwald, K.A. Krohn, and D.A. Mankoff. 2006. Quantitative fluoroestradiol positron emission tomography imaging predicts response to endocrine treatment in breast cancer. J Clin Oncol 24(18):2793-2799, reprinted with permission from the American Society of Clinical Oncology. that researchers can add genes for proteins that generate bioluminescent compounds in the tumor cells that are introduced into animals. These “reporter” genes can indicate how the tumors are responding to an experi- mental drug. For example, the firefly gene that codes for the luciferase

OCR for page 40
 WORKSHOP SUMMARY enzyme, which causes light to be emitted, is often inserted into tumor cells. Specialized cameras can sensitively detect light emitted from these tumors deep inside an animal’s body in preclinical studies. This optical imaging can be used with automated high-throughput systems that enable as many as 250 mice a day to be imaged, according to Dr. Piwnica-Worms. The luciferase gene can also be linked to a gene for a protein of interest so that it is activated only when this protein is expressed by the gene. With this system, the degree of light emitted will be proportional to the amount of protein produced. This enables optical imaging of molecules related to key cancer pathways in the body. Dr. Gore added that “optical imaging is a tremendously important tool in preclinical models of mice,” and noted that once animal studies show the usefulness of an optical probe, research- ers can then substitute a radiolabeled probe for the optical agent so that it can be imaged by PET in clinical trials. Repetitive, non-invasive molecular imaging can provide the bridge between the genetic studies that are increasingly being done on tumor samples and the radiologic imaging done on patients, by revealing—in a dynamic in vivo fashion—the presence of key genetic biomarkers and pathways in the context of the whole organism over time, Dr. Piwnica- Worms said. In both preclinical drug development as well as in patients, molecular imaging can provide the added fourth dimension of time, which can reveal dynamic processes in the body, he stressed. For example, in one of Dr. Piwnica-Worms’ studies, a luciferase reporter gene was used to reveal tumor production of a key protein targeted by an experimental drug. Repetitive optical imaging essentially provided “a real-time in vivo Western blot of protein content over time” that gave enough pharmacodynamic and pharmacokinetic data to determine an appropriate dosing regimen for a subsequent clinical trial, he said. Researchers have also developed reporter gene imaging probes for key drug metabolizing enzymes. These probes can reveal how experimental drugs are metabolized and, in one animal study, indicated gender differ- ences in drug metabolism (Zhang et al., 2003). In summary, Dr. Piwnica- Worms noted that these innovative probes and molecular imaging can aid both preclinical and clinical studies of experimental cancer therapies by validating mechanisms of action; providing pharmacodynamic and phar- macokinetic information; and humanizing the models so they better predict what will happen in patients. “You can get direct analysis of target-specific pharmacodynamics uncoupled from maximum tolerated dose that can guide the therapeutic clinical trial designs,” he said.

OCR for page 40
 IMPROVING THE QUALITY OF CANCER CLINICAL TRIALS Innovative imaging probes can reveal a number of physiological fea- tures of tumors that might provide earlier and more predictive measures of response to treatment than standard measures of tumor size currently used to determine response rates to treatments in clinical trials of experimental cancer drugs. Several speakers noted the promise of radiolabeled FLT (3′-Deoxy-3′-[F18] fluorothymidine) as a PET imaging probe for increased cell division. Retention of FLT in tumor cells reflects heightened activity of the thymidine kinase enzyme, which is linked to cell proliferation. Another indicator of cell proliferation is heightened production of compounds such as choline, which comprise cell membranes and can be imaged with PET or with higher magnetic field MRI systems, according to Dr. Gore. He added that there are also optical, PET, or single photon emission computed tomography (SPECT) imaging probes for the compound annexin-V, which can indicate cells are undergoing programmed cell death (apoptosis). In addition to FDG, there are other PET probes for the heightened metabo- lism of tumors, including those for amino acids and pH. Researchers have also developed imaging probes for how oxygenated tumors are, which radi- ologists can use to determine radiation dose escalation, Dr. Hricak noted. Dr. Gore said imaging of tumor oxygenation can reveal complementary information to that conveyed by FDG imaging of tumor metabolism as to whether a treatment is having a major biological effect. “We tend to get obsessed with using single biomarkers, but there is tremendous potential in combining different kinds of imaging biomarkers. No single biomarker may be adequate,” he said. This sentiment was echoed by most of the molecular imaging speakers. “We have to image a system of targets, not only one,” said Dr. Hricak. To illustrate the value of combining imaging biomarkers, Dr. Gore showed a slide of animals treated with a drug that targets the epidermal growth factor receptor, which has been linked to several solid tumors. Optical imaging revealed that the EGF receptor density was reduced by the treatment and apoptosis was increased, as indicated by an imaging probe for annexin V. But proliferation was not significantly decreased. “So we have hit the target, we have had some biological effect, but we haven’t had the end result we hoped for,” he said (Figure 12). In addition to combining various biomarker molecular imaging, researchers can also combine different imaging modalities to gain more information on treatment response, Dr. Gore said. Dr. Larson noted that a PET/CT scan can show that a tumor hasn’t changed in size in response to treatment, but has changed markedly in terms of its metabolism. Compa-

OCR for page 40
 WORKSHOP SUMMARY EGFR Density A D E 2.5 NIR-EGF (T/N Ratio) 2.0 T T 1.5 1.0 T T J K 0.5 0.0 Treated Untreated Treated Untreated B F G Apoptosis Frequency NIR-Annexin V (T/N Ratio) T 3.5 K K T 2.5 K K T 1.5 T L M 0.5 -0.5 Treated Untreated Treated Untreated C I H Amount of Proliferation 2.5 [18F]-FLT (T/M Ratio) T T 2.0 1.5 T T 1.0 N O 0.5 Treated Untreated Treated Untreated FIGURE 12 Optical imaging of animals treated with a drug that targets the epidermal growth factor receptor (EGFR), which has been linked to several solid tumors. EGFR density was reduced by the treatment (A) and apoptosis was increased (B), as indicated Figure 12 by an imaging probe for annexin V. But, proliferation was not significantly decreased (C). D-I show representative optical images reflecting data quantified in A-C. Brighter signal corresponds to higher target density. Tumors (T) and kidneys (K) are pointed out with arrows. J-O show representative immunohistochemistry staining in tumor tissue slices of the same markers imaged in D-I. ACRONYMS: FLT (3′-Deoxy-3′-[F18] fluorothymidine), T/M ratio (tumor-uptake to muscle-uptake ratio), T/N ratio (tumor tissue-uptake to normal tissue-uptake ratio). SOURCE: Gore presentation (October 4, 2007).

OCR for page 40
 IMPROVING THE QUALITY OF CANCER CLINICAL TRIALS nies are already making hybrid scanners that combine CT with PET, and MRI–PET scanners are on the horizon, Dr. Gore noted. Challenges of Molecular Imaging Despite their promise, there are several challenges to ensuring that many of these molecular imaging probes meet the basic requirements for imaging biomarkers, several speakers noted. These basic requirements are that they be quantifiable, objective, accurate, sensitive to relevant biologi- cal changes (especially tumor-relevant processes), reproducible, validated, and standardized, Dr. Gore explained. But he also said imaging biomarkers have to be adequate, not perfect. “I think many people in the field criticize themselves for not having a better biomarker; but so long as it is better than the ones we are already using, it has already proven to be somewhat useful,” he said. Another speaker, Dr. Larry Schwartz from MSKCC, noted that current standard imaging also has several shortcomings. He showed how standard imaging of tumor size is not reliable or standardized, and often lacks biological relevancy. For example, there is quite a bit of variability in the measurement of tumor size between tumors measured only by diam- eter and not bidimensionally. There can be a 3-month difference in time to progression when tumors are measured bidimensionally as opposed to unidimensionally, Dr. Schwartz pointed out (Schwartz et al., 2003). A lack of standardization of image acquisition guidelines in clinical trials in regard to whether MRI or CT is used, the timing of contrast administration, and image slice thickness also can create discrepancies in the assessment of tumor response. For example, the size and number of cancer metastases visualized can vary greatly depending on when contrast is administered, and there can be a threefold increase in detection between images acquired at a 10-mm slice thickness versus a 2.5-mm slice thickness, Dr. Schwartz showed. He added that some of the endpoints used in clinical trials, such as a 20- or 30-percent response rate, are rather arbitrary and may not correlate with survival. “Conventional imaging uses poor surrogates as endpoints, and quite frankly biologically irrelevant response parameters,” Dr. Schwartz said. These same issues also pertain to molecular imaging. “Very often, we jump to the modalities that are not mature. When PET came, everybody wanted to use PET in clinical trials, and there were many failures because it was used before the modality was validated, standardized, and reproduc- ible,” Dr. Hricak said. Dr. Schwartz pointed out that different image set-

OCR for page 40
 WORKSHOP SUMMARY tings on a PET scanner can affect the results seen with FDG probes, and molecular imaging probes need to be validated as being clinically relevant. Dr. Hricak cautioned against using a contrast agent in molecular imaging before it is preclinically validated and shown to be sensitive and specific. She also questioned the use of bioluminescence imaging because it is not as quantitative as other molecular imaging techniques, such as PET. But Dr. Piwnica-Worms pointed out that once a target is validated by biolumi- nescence, then it can be quantified using a PET probe. “Although biolu- minescence is semi-quantitative in terms of absolute photon output, it can be absolutely quantitative in terms of the biochemistry,” he said, because changes over time in bioluminescence can reveal pharmacokinetics and pharmacodynamics. Often there is a lack of standardization of imaging protocols in clinical trials, as well as a lack of harmonization between techniques used in dif- ferent centers, Dr. Hricak noted. Imaging protocols are cancer type– and site-specific, she said. For example, MRI can adequately image local breast cancer, but bone metastases are best imaged with PET and an appropriate probe, yet the modality used during a study should not change. She sug- gested including an imaging expert when designing a clinical trial to help ensure the trial’s success. Ideally, tumors should be visualized in volumetric displays, Dr. Hricak and Dr. Schwartz said. A slide of Dr. Schwartz’s showed a nearly 40-fold difference in tumor percentage change following treatment, depending on whether just the tumor diameter was measured versus whether the volume of the tumor was measured. Dr. Schwartz optimistically summed up the discussion of the challenges involved in molecular imaging by saying, “We in imaging view many of these as challenges that are readily achievable by obtaining appropriate image acquisition guidelines which could be stan- dardized in a rational manner.” One additional challenge mentioned by Dr. Larson is how to disperse the bioimaging tracers being developed at individual laboratories to the wider research community and into clinical trials. In a later discussion, Dr. Hricak noted that at MSKCC alone, “there is a menu of radiotracers that have been around for at least 5 to 15 years that are not FDA approved and widely distributed.” Dr. Larson advocated using the nuclear pharmacies that are scattered throughout the world to better distribute molecular imag- ing probes. These pharmacies use automated chemistry to make and ship labeled molecular PET probes. In the panel discussion that followed the molecular imaging presenta-

OCR for page 40
 IMPROVING THE QUALITY OF CANCER CLINICAL TRIALS tions, Dr. Mills echoed Dr. Hricak by stressing the lack of harmonization in bioimaging protocols between centers, which may not have the same hard- ware or software platforms, and which believe they have already optimized the components of their own protocols. “Until we establish harmonization on top of standardization, we are still going to have limitations for applica- tions in clinical trials,” he said. Panel member Dr. Jeff Evelhoch of Amgen agreed that “that is probably the biggest challenge that we have in using imaging in clinical trials,” and that this challenge varies with the imaging modality. He noted that because bioimaging methods for dynamic contrast enhanced (DCE) MRI are continually developing and progressing, harmo- nization is more difficult to achieve than with more standard CT imaging. But there is some consensus on an appropriate acquisition protocol and other standards needed across centers in clinical trials using this technol- ogy, he added. In contrast, FDG PET, even though it is used more often in the clinic than DCE MRI, has fewer agreed-on standards even within the same institution. Pfizer’s Dr. McCarthy noted that a number of initiatives have been made to standardize or harmonize molecular imaging biomarkers, including one by the National Institute of Standards and Technology, and the Oncol- ogy Biomarker Qualification Initiative sponsored by the FDA, the NCI, and Centers for Medicare & Medicaid Services (CMS). But these efforts at standardization and harmonization are going on in parallel with each other without coordination and agreement. Genomic Health’s Dr. Shak pointed out that genomic assays faced the same issues on standardization and harmonization, and successfully met them with financial investment in resources needed to ensure the technologies and the procedures were in place to provide quality control and harmonization. Dr. Mendelsohn added that such standardization and harmonization may not be as critical in clini- cal trials that measure tumor response or other variables over time as long as there is reproducibility at each participating institution. “If the tumor went down 50 percent in a reproducible way at that institution, it might be just as important as in another institution where they might have picked up four other lesions because [their measurements are] more sensitive. But you have still got to go down 50 percent,” he said. “Where you need perfect standardization and harmonization is when it is a one-shot thing—either the androgen receptor is present or absent, and there you would need it.”