Jeffrey N. Bryan, DVM, MS PhD, DACVIM (Oncology)
Molecular imaging is the development of a diagnostic imaging by targeting a molecule or a tissue property to deliver a payload that can be detected using a variety of imaging technologies. Most commonly, the images derived from molecular imaging are described as functional because the target is a cellular product or an important player in a metabolic process associated with the cancer. The target could be as simple as a metabolic pathway unique to the tissue of origin of the cancer, such as iodine trapping in thyroid cancer, or it could be a specific antigen that is overexpressed in or even unique to the type of cancer, such as prostate specific membrane antigen (PSMA) in prostate cancer. The goal of molecular imaging has been to improve the specificity of cancer imaging by eliminating background, non-neoplastic cell signal to define purely the tumor burden. The future of molecular imaging is increasingly specific diagnostic assays that might identify the exact type of cancer in vivo and give early evidence of therapeutic efficacy for real-time decision making tailored to individual patient's responses.
The workhorse of nuclear medicine in veterinary practice has been the planar gamma camera. This instrument consists of a collimator filter in front of a scintillation crystal backed by photomultiplier tubes that allow localization of radioactive events of the proper energy in space in front of the camera. Most gamma cameras can perform static or dynamic studies. Static studies identify localization of the imaging agent within the body at a single point in time or as a composite gathered over a defined period of time. Dynamic studies observe the accumulation of agent over a window of time in real-time. They can be used to calculate uptake and kinetics of the agent to define distribution or excretion properties, or flow dynamics.
Nuclear imaging of the thyroid gland and thyroid tumors relies on the unique metabolic property of iodine trapping and organification of the thyroid tissue. Three agents may be used to create a nuclear image of the thyroid using planar scintigraphy. The first and most common is pertechnetate (99mTcO4-) with a single photon emission of 140keV that is excellent for resolution of planar imaging. This nuclide is handled as a halide in the body, causing it to be trapped by the thyroid and excreted by organs that secrete high chloride containing fluids, including the salivary and gastric glands. Pertechnetate is trapped by thyroid cells, but does not undergo organification.1 Little work has been done to correlate pertechnetate uptake with therapeutic 131I uptake in dogs. The iodine radionuclide 123Iemits a 159keV photon and undergoes complete trapping and organification. It is also more specific for thyroid, not accumulating in the salivary and gastric glands. As an isotope of iodine, the uptake of 123I should be completely predictive of therapeutic 131I. Unfortunately, the clinical applicability of this nuclide in the veterinary setting is limited by its very high cost relative to pertechnetate. Finally, 131I itself can be used to image thyroid tumors. The photon energy of 131I is higher than the optimal range for planar scintigraphy at 365keV. In spite of this fact, the author has used small dose injection of 131I to confirm similar uptake with pertechnetate in cases with multiple metastases that were to be treated with the isotope. In published literature, little correlation has been identified between pertechnetate uptake patterns and response to 131I therapy.2 The best that can be said at this time is that it appears that dogs with positive uptake are more likely to respond to therapy, but the pattern of that uptake is not yet instructive. The author has performed a pilot study in dogs with thyroid carcinoma that compared plasma TSH levels with pertechnetate uptake intensity. Humans with thyroid cancer and higher plasma TSH levels are more likely to respond to 131I therapy than those with low plasma TSH levels.3 In dogs that were not measurably hyperthyroid, plasma TSH levels correlated positively with thyroid:salivary gland ratio on pertechnetate scan, but the numbers did not yet reach statistical significance (r2=0.32 with p = 0.245). This relationship was similar to that of plasma T4 and intensity (0.31 with p = 0.042). This area remains an ongoing focus of research in the author's practice.
Unlike thyroid cells, bone cells do not have an easily targetable molecule that is unique to them that can be currently exploited by imaging. Instead, the bone matrix surrounding the cells offers an attractive target for imaging and potential therapy. The hydroxyapatite matrix of bone will readily bind phosphorous-containing molecules. By attaching 99mTc as an imaging nuclide or 153Sm as an imaging and therapy nuclide, the localization of the phosphorus-containing chelator can be identified by planar scintigraphy. The photon of 153Sm is 103keV, and it emits three β- particles with a range of up to 3mm of tissue. Scintigraphic intensity of osteosarcoma has been shown to be prognostic for outcome.4 Bone scintigraphy is also useful in patients with osteosarcoma to identify early osseous metastatic lesions or metachronous primary lesions of bone, which have been reported to occur in as many as 25% of cases.5 In the author's experience, and in several studies, the rate of second bone lesions is probably less than 5%.6 Scintigraphy has also been used to assess tumor margin prior to resection for limb-spare surgery. While plain radiographs tend to underestimate the microscopic extent of tumor, scintigraphy overestimates it, allowing a safe, tumor-free margin to be resected.7 Bone scintigraphy can be used to identify metastatic to bone lesions, and 153Sm-EDTMP has been used to target these lesions with palliative intent.8 Some tumors, like myeloma lesions, tend not to concentrate bone imaging agents, resulting in "cold" spots on scintigraphy.
While not using a targeted agent, cardiac parameters can be measured with high accuracy with dynamic studies performed using a planar gamma camera. The agents used for these studies include rapidly excreted preparations of 99mTc (pertechnetate or 99mTc-DTPA) or 99mTc ex-vivo-labeled red blood cells. Qualitative studies can accurately identify congenital shunting conditions. Quantitative multigated-acquisition (MUGA) studies are extremely sensitive measures of ventricular function, and represent the gold-standard for function evaluation to evaluate cardiac toxicity of drugs like doxorubicin.9 These studies are included here, simply because the equipment for the targeted studies discussed will often allow these cardiac evaluations to be performed as well.
Targeted Monoclonal Antibody Imaging
For several decades, monoclonal antibodies (mAbs) have been evaluated for use in imaging and treating cancer because of their exquisite specificity of molecule recognition. As discussed in the previous lecture, the trade-offs for high specificity are multiple. The molecules are large, having circulating half-lives of several days, resulting in increased radioactivity background on scans due to lack of washout from the blood pool. Also, many of these proteins are xeno-antigens in both humans and dogs, with murine antibodies in common use. The humanized products remain foreign to dogs, limiting the number of times the agents can be administered. Although the development of human anti-mouse antibody (HAMA) reaction is commonly described, the canine anti-human antibody (CAHA) reaction is more obscure!
Agents such as trastuzumab, targeting HER2/neu in breast cancer, and bevacizumab, targeting VEGF, are now FDA approved for therapy of cancer in humans, but are not currently used for imaging. Either could potentially serve as an imaging vector. Capromab pendetide (ProstaScint®) is approved for imaging prostate cancer in humans. It targets the prostate-specific membrane antigen (PSMA) present in the majority of human prostate cancers. Recent literature suggests that this antigen is expressed in most canine prostate cancers and is identifiable by immunohistochemistry in approximately half.10 The products 131I-tositumomab (Bexxar) and 90Y-ibritumomab (Zevalin) target the B-cell antigen CD20 of humans and are FDA approved for therapy.11;12 In each case, however, there is an imaging component. Rather than for staging, the information generated by imaging is used for dosimetry of these products. The visible distribution predicts the total dose delivered to tumors by the radiopharmaceutical. In the case of 90Y-ibritumomab, the Yttrium isotope does not produce an imageable photon emission, so 111In (photon emissions of 171 keV and 245keV) is substituted for the 90Y for imaging purposes. Unfortunately, the canine CD20 is sufficiently different from humans as to render these agents unusable in canine clinical patients.13 However, other mAbs or a canine-specific anti-CD20 mAb may make this feasible in the future. The embryonic antigen associated with human colorectal carcinoma, carcinoembryonic antigen (CEA) has been targeted using mAbs and portions of mAbs for staging purposes for colon cancer.14 The CEA has been identified in canine cholangiocarcinomas and mixed hepatocellular carcinomas, among others.15
Targeted Peptide Imaging
Although of lower binding affinity than mAbs, peptides can be designed to target specific receptors on the surface of cells that have much more favorable circulation kinetics. The result is reasonable targeting efficiency with more rapid loss of background blood pool activity to result in better target to background ratio and a better image. Somatostatin analogues have been used to image endocrine tumors in dogs and both endocrine tumors and lymphoma in humans. Somatostatin analogues labeled with therapeutic radionuclides have been used experimentally to treat human endocrine neoplasia.16 In dogs, radiolabeled somatostatin analogues have been shown to accumulate in both gastrinomas and insulinomas.17-19 In humans, non-Hodgkin's lymphoma expresses somatostatin receptors approximately 80% of the time.20 Both adult and juvenile forms of the disease have been successfully imaged with radiolabeled somatostatin analogues.21-23 This data prompted the authors to perform a pilot study of 111In-DOTA-TATE, a somatostatin analog, as an imaging agent in three dogs with multicentric lymphoma. In each case, affected lymph nodes accumulated activity at a ratio significantly higher than background. Internal organs, primarily liver and spleen, which were infiltrated by lymphoma, could be visualized on the scans as well. This initial data confirmed the presence of the molecular target in canine lymphoma. On the basis of this target identification, a current study involving anti-sense imaging is underway that ultimately could lead to future targeted radiotherapy. Many other peptides are under development for future imaging and therapy use. Bombesin receptors are expressed on the surface of prostate cancer cells, representing an attractive target both for imaging and therapy.24 These are involved in prostate cell division, and appear to play a similar role in canine prostatic epithelium as they do human.25 Processes integral to cancer progression, such as angiogenesis, have been proposed as targets for development of both imaging and therapy agents. The three amino acid motif arginine, glycine, and aspartate (RGD) binds with great affinity to the αvβ3 integrin expressed in tumor vasculature.24 Agents designed to be imaged with a number of technology platforms are being developed using the RGD motif.24
Another strategy designed to improve the kinetics of an imaging agent but retain the specificity of a mAb for targeting is termed pretargeting. Many strategies have been proposed to pretarget a tumor for labeling. The fundamental principle is that a mAb is introduced into the patient to tag a tumor that can be labeled specifically by some property built into that mAb. In this way, the mAb is allowed to circulate, bind, and is cleared from the circulation prior to introducing any radioactivity for imaging. One strategy to accomplish this is to administer a bispecific mAb that binds both a tumor antigen and a peptide that can be labeled with a radionuclide.26 This dual specificity insures that the tumor is effectively labeled and the remaining radioactivity leaves circulation rapidly in the urine, taking the extraneous radionuclide with it. Another strategy is to conjugate the mAb with a system that binds with high affinity and avidity, such as streptavidin. Biotin is then labeled with a radionuclide and injected into the patient following clearance of the remaining mAb from circulation.27 The authors employed such a technique to image metastatic prostate cancer in a canine clinical patient. The experiment was successful in that all known disease was identified on the nuclear scans, and a metastatic location previously unsuspected was identified as well.28
Single-Photon Emission Computed Tomography
Like planar gamma camera imaging, this technology uses collimated scintigraphic crystals to image low- to medium-energy photons. Termed SPECT, the platform moves the crystals around the patient, collecting data from multiple angles, to reconstruct a three-dimensional rendering of the structure of interest. The principal advantage of this technology is that the result can render a more accurate distribution of the imaging agent within the area of interest. While planar images are a compressed view of all radioactive events within a field of interest, the three-dimensional output of SPECT gives more information about spatial distribution of the agent. The author has proposed using this technology to determine tumor targeting homogeneity of novel bone-seeking radiopharmaceuticals. We expect this to offer insight into tumor dosimetry and develop image markers of biological behavior.
Positron emission tomography (PET) imaging has become routine in human cancer staging and monitoring. Limited by the cost of both equipment and imaging agents, this modality has been used largely experimentally in veterinary oncology to this point. The basis of the technology is the particle reaction of a positron (essentially a positively charged electron) annihilating with an electron, resulting in two, diametrically opposed 511keV photon emissions. The simultaneous creation of these photons allows the origin to be identified using high energy detectors surrounding the patient in a ring. The energy of the photons allows them to escape even a very large body very efficiently. The overall result is an image with excellent resolution, limited only by the distance of travel of the positron in tissue prior to annihilation and the resolution of the scanner itself. Modern PET scanning equipment and agents can localize disease within a body with inaccuracy of only a few millimeters. PET imaging is limited to a degree by generation of radionuclides with a half-life sufficient for shipping to a distant location. Flourine-18 is cyclotron produced with a half-life of less than 2 hours. Thus, a producer must be located within a few hours to support the use of a PET scanner. The radionuclide 64Cu has a half-life of 12.7 hours, so can be shipped from a great distance, but the nuclide's use has not yet become routine in clinical imaging. PET imaging is now routinely combined with CT imaging in humans to marry functional information with exact anatomical localization.
The first PET agent approved by the FDA for imaging of cancer, 18FDG relies on the excessive glucose metabolic properties of cancer to localize in tumors. 2-deoxyglucose is a substrate for both the GLUT transporters and hexokinase, which phosphorylates glucose at the 6 position, trapping it irreversibly within the cell. However, once phosphorylated, the lack of a 2 hydroxyl group prevents the molecule from being converted to fructose by phosphoglucose isomerase, preventing its metabolism. As a result, the molecule accumulates in cells actively employing glycolysis, resulting in accumulation of 18F for imaging. The uptake of 18FDG is not unique to cancer cells. Metabolically active tissue, including inflammation, can effectively concentrate the molecule and yield a false-positive scan.29 In human oncology, 18FDG imaging is extremely useful for staging a variety of tumors including breast cancer, hepatocellular carcinoma, lung tumors, head and neck carcinoma, brain tumors, and many sarcomas. Some tumors can not be imaged well with 18FDG. An example of this is prostate cancer, which is why ProstaScint has not been supplanted by 18FDG.26
Nucleoside analogues can be labeled with 18F to provide information about tumor cell proliferation. The most commonly used agent to date is deoxyfluorothymidine (18FLT). This is accumulated by nucleoside transporters into actively dividing cells, where it is trapped to generate an image by PET scan. The utility of this is that metabolically active cells can be distinguished from dividing cells by uptake pattern of 18FLT.24 Following therapy, tumor cells may continue to accumulate glucose under stress, but will no longer be dividing. Early detection of treatment failure will help avoid delays in rescue. Proliferation imaging may also distinguish between radiation necrosis and tumor recurrence, detect indolent disease, such as lymphoma, or discriminate inflammatory lesions from neoplastic.24
Molecular imaging provides a functional image of many processes intimately associated with the neoplastic phenotype that complements the anatomic imaging routinely performed in veterinary oncology. With equipment as simple as a planar gamma camera, thyroid and bone scan imaging can assist clinical decision making about the applicability of radiopharmaceutical therapy, limb sparing surgery, and sensitive staging of bone tumors. It can also provide highly precise information about organ function. The future of molecular imaging is increasingly specific molecules that target cancer cells for identification and, ultimately, destruction. As mAbs and peptides are developed along with strategies to maximize clinical application, these technologies will trickle down to veterinary oncology for more widespread application.
1. Nap AM, et al. J Vet Intern Med 1994;8(4):302.
2. Turrel JM, et al. J Am Vet Med Assoc 2006;229(4):542.
3. Barbaro D, et al. Eur J Surg Oncol 2007;33(5):535.
4. Forrest LJ, et al. J Nucl Med 1992;33(8):1542.
5. Hahn KA, et al. J Am Vet Med Assoc 1990;196(9):1483.
6. Chun R, et al. Vet Clin North Am Small Anim Pract 2003;33(3):491, vi.
7. Leibman NF, et al. Vet Surg 2001;30(3):240.
8. Lattimer JC, et al. J Nucl Med 1990;31(8):1316.
9. Karpawich PP, et al. Am Heart J 1991;121(3 Pt 1):827.
10. Lai CL, et al. Prostate 2008; in press.
11. Witzig TE, et al. J Clin Oncol 2002;20(10):2453.
12. Kaminski MS, et al. Blood 2000;96(4):1259.
13. Impellizeri JA, et al. Vet J 2006;171(3):556.
14. Obradovic V, et al. Hepatogastroenterology 2006;53(70):526.
15. Martin de las Mulas J, et al. Res Vet Sci 1995;59(2):124.
16. Kwekkeboom DJ, et al. J Clin Oncol 2005;23(12):2754.
17. Altschul M, et al. J Small Anim Pract 1997;38(7):286.
18. Garden OA, et al. J Vet Intern Med 2005;19(2):168.
19. Lester NV, et al. Vet Radiol Ultrasound 1999;40(2):174.
20. Ferone D, et al. Semin Nucl Med 2005;35(3):176.
21. Raderer M, et al. Br J Cancer 2001;85(10):1462.
22. Cholewinski W, et al. Eur J Nucl Med Mol Imaging 2004;31(6):820.
23. Lugtenburg PJ, et al. Br J Haematol 2001;112(4):936.
24. Kelloff GJ, et al. Clin Cancer Res 2005;11(22):7967.
25. Allard P, et al. J Androl 2000;21(3):367.
26. Sharkey RM, et al. Clin Cancer Res 2007;13(18 Pt 2):5577s.
27. Lewis MR, et al. Nucl Med Biol 2004;31(2):213.
28. Lewis MR, et al. Annual Conference Proceedings of the Veterinary Cancer Society, Huntington Beach, CA, 2005;42.
29. Ballegeer EA, et al. Vet Radiol Ultrasound 2006;47(2):228.