Donald E. Thrall, DVM, PhD, DACVR (Radiology, Radiation Oncology)
Cancer treatment methods are becoming sophisticated and many pet owners are undertaking cancer therapy for their pets. As a result, imaging is playing a larger role in the assessment of the cancer patient. Cancer imaging has focused primarily on anatomic derangements but in the future imaging will be used to acquire functional information as well. This functional information will enable categorization of the metabolic and/or physiologic properties of the host and tumor. These properties have value in predicting outcome following therapy, and may also be useful with regard to treatment planning. A brief overview of oncologic imaging and prospects for future evolution will be presented.
Radiography is the mainstay of veterinary imaging and this is unlikely to change, even with the increasing availability of other more sophisticated modalities. For cancer imaging, radiography is used mainly for metastasis screening and assessment of bone tumors. One of the biggest breakthroughs in radiographic imaging of lung cancer was the identification of the importance of lateral recumbency on conspicuity of lesions in the dependent lung lobe1. Though this principle has been known for over 20 years, it is sad that is has yet to be adopted universally by practicing veterinarians in the United States.
Another radiographic breakthrough was the introduction of rare-earth film-screen combinations. This technology requires significantly less radiation for production of diagnostic radiographs. Inherent in this new technology, in comparison to calcium tungstate screens, is decreased image contrast. This decreased contrast has reduced the conspicuity of lung nodules making their identification more difficult. At North Carolina State University we are relying upon fluoroscopic examination of the lung more frequently to assess presumptive solitary lung nodules. Evaluating the mobility of the suspected nodule with relation to the thoracic cage and lung allows a more definitive conclusion to be reached.
The first article describing the use of ultrasound for veterinary imaging appeared in 19762. Use and applications of sonographic imaging have expanded considerably and there is not a body system where sonographic imaging has not been applied successfully. Sonographic imaging of cancer patients has resulted in a dramatic increase in the detection of lesions of parenchymal organs (liver, spleen) and in assessment of lymph node size. Specificity, however, remains poor and as a result patients with sonographically detected masses often undergo invasive needle aspiration or tissue biopsy. This scenario is particularly problematic in dogs with mast cell tumors where liver or spleen nodules are detected sonographically during staging. At NCSU these nodules are routinely aspirated but they are typically neither malignant nor would they have adversely affected the patient. Thus, the poor specificity results in additional time and expense for the owner, and subjects the patient to an unnecessary invasive procedure.
Innovations in sonographic imaging include Doppler and power Doppler imaging, intraoperative sonography, harmonic imaging and microbubble contrast media. Of these, contrast media use may hold the most promise for increasing the specificity of diagnosis of sonographically detected parenchymal lesions. In humans, contrast medium use permits definitive diagnosis of most liver mass lesions3 and there is great opportunity for sonographic contrast medium use in veterinary oncology.
Other than assessing skeletal metastasis, nuclear imaging has not been widely adopted in oncologic imaging. A relationship between absolute radiopharmaceutical uptake in osteosarcoma and time to metastasis has been identified4, but this relationship has not been tested in a prospective setting. Thyroid scintigraphy is widely used in hyperthyroid patients, especially cats, when functional thyroid masses are considered. One untapped area where nuclear medicine may prove beneficial relates to assessing tumor lymphatic drainage. Inclusion of regional lymph nodes within the irradiated volume is often desirable. Scintigraphic mapping of lymphatic drainage from a tumor site, e.g., lymphoscintigraphy, is important for determining the appropriate geometry of radiation fields5. In veterinary medicine, little is known about lymphatic drainage patterns, other than from anatomy texts. Some information, however, suggests that variation from normal is common and could impact treatment outcome. There is great opportunity for more work to be done in lymphoscintigraphy in veterinary oncology.
The contrast resolution and tomographic characteristics of CT imaging have increased the accuracy of tumor staging. Evaluation of the extent of gross tumor and the detection of pulmonary nodules are facilitated by use of CT. At this time, lung CT for metastasis is not a primary screening method. However, for therapies implemented only in metastasis-negative patients, use of lung CT will likely increase. The use of CT imaging for detection of intracranial tumors has largely been supplanted by magnetic resonance imaging because of its superior contrast resolution.
Dynamic CT has been used to assess tumor perfusion, and variability in perfusion between dogs with nasal tumors has been shown6. Tumors with low perfusion might contain hypoxic regions that negatively affect radiation response.
Most veterinary radiation therapy centers use CT images for virtual treatment simulation. CT planning allows configuration of externally applied photon or electron beams to optimize distribution of radiation dose in the patient; optimization involves maximizing tumor dose while limiting dose to critical normal tissues in the treated volume. Inverse planning where dose is specified and a computer algorithm defines treatment beam geometry will be the next generation of CT based radiation planning7. Multislice scanners, with rapid 3D volume imaging and virtual endoscopy, will undoubtedly facilitate treatment planning for tumors whether surgery or radiation are used8.
MAGNETIC RESONANCE IMAGING
MRI has revolutionized neural imaging. It allows precise identification of masses affecting the neuropil. Specificity is a problem. MR images have not yet replaced CT images for radiation planning but this will likely occur.
One powerful aspect of MR technology is the ability to characterize physiologic processes using spectroscopy. Parameters of high-energy phosphorus metabolism and assessment of intracellular pH are two examples of how MR may be used to characterize the tumor. In canine sarcomas, intracellular pH, measured using MR spectroscopy, is higher than extracellular pH measured using pH electrodes. This pH gradient is consistent with the tumor cells being able to buffer intracellular pH at a physiologic level9. Also, in human glioma patients, spectroscopy has enabled detection of metabolically active tumor deposits up to 2.0cm peripheral to regions of contrast enhancement, thereby emphasizing the poor spatial localization of tumors based only upon anatomic derangements10.
The next frontier for oncologic imaging will be characterization of cellular function. Assessing morphology using imaging has allowed for advancements in the treatment of cancer. However, further improvements in outcome will likely come from exploitation of functional or metabolic differences, or by imaging functional alterations associated with therapeutic processes. Many functional and/or metabolic imaging studies require the use of positron emission tomography (PET). PET scanning is complicated by the very short half-life of positron emitting radionuclides. The anatomic resolution of PET scanners is also relatively poor, but this is being overcome by design of hybrid CT/PET imaging devices. At this time, PET scanners are not widely available for veterinary use. However, who would have predicted 10 years ago the degree to which magnetic resonance imaging has become integrated into veterinary imaging. There is little doubt that PET imaging in veterinary medicine will increase.
Three aspects of metabolic imaging will be considered here: 1) imaging glucose metabolism using 18F-fluorodeoxyglucose; 2) imaging tumor hypoxia; and 3) imaging gene expression.
18F-FDG allows evaluation of glucose metabolism. FDG uptake is proportional to the glycolytic rate of viable tumor cells, indicating the increased tumor demand for glucose11. After phosphorylation, FDG is trapped within the cell because it is not a substrate for subsequent pathways and dephosphorylation is scarce or absent in tumor cells. Therefore, cellular concentration of FDG represents the glycolytic activity of viable tumor cells11. The sensitivity and specificity of FDG imaging has initiated the use of whole-body FDG scans for pretreatment metastasis screening. Changes in FDG uptake during or after therapy have prognostic significance in lymphoma and in breast, lung and colorectal carcinoma in humans. Evidence that FDG PET can predict response early during the course of therapy opens up new possibilities for optimizing therapy planning and prognostic evaluation11.
Most solid tumors have regions where oxygen tension is low. Hypoxia develops from diffusion limitations of oxygen (chronic hypoxia) and from transient alterations in perfusion (acute hypoxia). For decades it has been known that hypoxic cells are radioresistant and hypoxia was often blamed for the failure of irradiation to control solid tumors. Recently, tumor hypoxia has also been associated with an aggressive phenotype. Hypoxia induces genomic instability, prevents apoptosis, promotes proangiogenic factors and induces oncogenes and transcription factors. Thus, hypoxia's deleterious effects on a tumor and host are multifactorial and if hypoxia could be detected non-invasively, patients with aggressive tumors could be identified prior to therapy and directed into more specific approaches. One way to detect regions of low oxygenation is through the use of 2-nitroimidazoles. These drugs are reductively activated under conditions of low oxygenation and become irreversibly bound to cellular macromolecules. Immunohistochemical detection of these compounds in tissue biopsies is used to quantify tissue hypoxia12. However, biopsy-based methods are invasive and there is always the uncertainty of sampling error. Thus, it would be desirable to detect nitroimidazoles using imaging. This has become possible through labeling of nitroimidazoles with radionuclides and PET imaging of 18F radiolabeled nitroimidazole has been used with success13. Hypoxia imaging will become more commonplace as techniques are refined, and will likely be incorporated into radiation treatment planning technology to allow concentration of dose in regions where the inherent radiosensitivity of the tumor cells is low.
Gene therapy trials are underway worldwide with the goal of either reversing genetic deficiencies or using virus replication or gene products to kill cancer cells14. A major problem is determining the location and extent of transgene expression. This is of critical importance with regard to the specificity of treatment as well as avoidance of toxicity. In vitro determination of transgene expression is facilitated by inclusion of a reporter gene in the construct, with detection using methods such as fluorescence. In vitro, however, detection of gene expression is more difficult, and given the uncertainty of where translation is occurring a detectable reporter gene would be valuable in vivo as well. There is much interest in imaging gene expression. In theory, this could be accomplished by administering a radiolabeled substance that binds to a reporter gene product15. One approach undergoing investigation is use of the somatostatin receptor as a reporter gene with subsequent detection by quantification of intravenously administered radiolabeled ligand binding16. Commercial radiolabeled somatostatin receptor ligands are available, but synthetic ligands with higher binding efficiency are also being developed.
Since the discovery of x-rays in 1896, imaging has been used primarily to assess anatomy. Structural alterations of cancer patients enable treatment to be formulated and also provide prognostic information. As cancer treatment becomes more effective, there is a need to understand more about the biology of the tumor and host, and to monitor the outcome of therapy. Functional, or metabolic, imaging is the next imaging frontier. Exciting developments providing information at the molecular level are occurring at a rapid rate. It is incumbent upon veterinarians to become familiar with these new imaging techniques. Animal cancer patients hold a wealth of untapped information that will undoubtedly contribute to improved health of animals and humans. Veterinarians are in an ideal position to be key players in unlocking this novel and exciting information.
1. Spencer C, Ackerman N, Burt J: The canine lateral thoracic radiograph. Veterinary Radiology 22:262-266, 1981
2. James A, Osterman F, Bush R, et al: The use of compound B-mode ultrasound in abdominal disease of animals. J Am Vet Radiol Soc 17:106-112, 1976
3. Harvey C, Albrecht T: Ultrasound of focal liver lesions. Eur Radiol 11:1578-1593, 2001
4. Forrest L, Dodge R, Page R, et al: Relationship between quantitative tumor scintigraphy and time to metastasis in dogs with osteosarcoma. J Nucl Med 33:1542-1547, 1992
5. Campostrini F, Gregianin M, Rampin L, et al: How iliopelvic lymphoscintigraphy can affect the definition of planning target volume in radiation therapy of pelvic and testicular tumors. Int J Radiat Oncol Biol Phys 53:1303-1313, 2002
6. Van Camp S, Fisher P, Thrall D: Dynamic CT measurement of contrast medium wash in kinetics in canine nasal tumors. Vet Radiol & Ultrasound 41:403-408, 2000
7. Tsien C, Eisbruch A, McShan D, et al: Intensity-modulated radiation therapy (IMRT) for locally advanced paranasal sinus tumors: incorporating clinical decisions in the optimization process. Int J Radiat Oncol Biol Phys 55:776-784, 2003
8. Schoepf U, Becker C, Obuchowski N, et al: Multislice computed tomography as a screening tool for colon cancer, lung cancer, and coronary artery disease. Eur Radiol 11:1975-1985, 2001
9. Prescott D, Charles H, Poulson J, et al: The relationship between intracellular and extracellular pH in spontaneous canine tumors. Clin Cancer Res 6:2501-2505, 2000
10. Pirzkall A, Nelson S, McKnight T, et al: Metabolic imaging of low-grade gliomas with three-dimensional magnetic resonance spectroscopy. Int J Radiat Oncol Biol Phys 53:1254-1264, 2002
11. Kostakoglu L, Goldsmith S: 18F-FEG PET evaluation of the response to therapy for lymphoma and for breast, lung and colorectal carcinoma. J Nucl Med 44:224-239, 2003
12. Raleigh J, Dewhirst M, Thrall D: Measuring tumor hypoxia. Semin Radiat Oncol 6:37-45, 1996
13. Koch C, Evans S: Non-invasive PET and SPECT imaging of tissue hypoxia using isotopically labeled 2-nitroimidazoles. Adv Exp Med Biol 510:285-292, 2003
14. Voss S, Kruskal J: Gene Therapy: A Primer for Radiologists. Radiographics 18:1343-1372, 1998
15. Weissleder R, Mahmood U: Molecular imaging. Radiology 219:316-333, 2001
16. Zinn K, Chaudhuri T, Krasnykh V, et al: Gamma camera dual imaging with a somatostatin receptor and thymidine kinase after gene transfer with a bicistronic adenovirus in mice. Radiology 223:417-425, 2002