| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
EDUCATION EXHIBIT |
Molecular Imaging Program, National Cancer Institute, Bethesda, Maryland
There was a time when a renal cancer was just a renal cancer: a solid enhancing mass in the kidney that required no further description and was removed with a radical nephrectomy. As radiologists, we staged the mass according to whether it invaded the renal capsule, whether enlarged lymph nodes were present, and whether the renal vein and inferior vena cava were involved—and then we went on to the next case, leaving histology to the pathologists.
Since then, advances in our understanding and the treatment of renal cancer have occurred that bring into question the validity of several aspects of this practice paradigm. One manifestation of the evolution of our knowledge of renal cancer is the discovery of an increasingly complex array of tumor subtypes, clearly demonstrated in the preceding article by Prasad et al (1). These tumor types range in frequency from the common to the almost unheard of. Beginning with clear cell carcinoma, the most prevalent, Prasad et al (1) describe and illustrate the growing list of renal cancer subtypes: multilocular cystic, papillary, chromophobe, collecting duct, medullary, and mucinous tubular and spindle cell. These groupings allow patients to be treated more individually, some more aggressively and others less aggressively than the "one size fits all" policy of the prior era, although there is, of course, still variability in the biology of tumors within each subtype. Prasad et al (1) also make us aware of interesting, if rare, associations of renal cancer in younger individuals with chromosomal translocation-fusion (Xp11.2 trans TFE3), neuroblastoma-associated renal cancers, and hereditary renal cancer syndromes.
This comprehensive review of the diversity of renal carcinoma is not only a welcome addition to the radiology literature but also raises several questions: Where has all this new information come from? How will it affect imagers? Will molecular imaging play any role? Given the increasing complexity of renal cancer classifications in the last 10 years, what will the next 10 years bring?
Histology, which is based on H-E staining, reflects tissue microstructure and has been the traditional standard of reference for tumor diagnosis. Recently, the ability to differentiate cell types among cancers has advanced dramatically with improvements in immunohistochemistry and genomics. These have added and continue to add an increasingly complex lexicon, if not understanding, to conventional pathology (2). These biomarkers are better indicators than histologic structure of the tumor’s biology and provide insight into the origins of cancer. Advances in the genetics and immunohistochemistry of clear cell carcinoma, papillary cell types, chromophobe carcinoma, oncocytoma, and medullary renal cancer have already had clinical impact. Such subtyping provides physicians with better prognostic information and allows them to make better treatment decisions. Moreover, the growing list of renal tumors caused by inherited syndromes means that fewer patients are classified as having sporadic tumors of unknown genetics, and we are seeing more genetic and radiologic screening of seemingly unaffected individuals.
The challenge for imaging is not merely to recapitulate the H-E diagnosis or even the more sophisticated diagnosis based on immunohistochemistry or genomics. Rather, it is to characterize the lesion by its biology, its aggressiveness, and its response to therapy. Some may call it the "tricorder fantasy" after the device used in the television series "Star Trek," but imaging reflects the tumor in its living state and it is this biology that radiologists ultimately want to characterize; in comparison, histology is merely a surrogate. Realistically, radiologists are most useful to clinicians not when speculating about specific histologic diagnoses (and trying to make a diagnostic "home run") but when providing anatomic and physiologic information about the tumor’s size, location, vascularity, distribution, and growth pattern, thereby offering plausible working histologic diagnoses that direct further diagnostic studies and therapy.
As Prasad et al (1) make clear, there are some basic imaging features that generally characterize the most common histologic subtypes. Clear cell carcinomas tend to be vascular and heterogeneous. Papillary renal cancers tend to be hypovascular and homogeneous. Both chromophobe carcinomas and oncocytomas tend to be vascular and homogeneous (3,4). The key word is tend, and the operative principle is that a number of plausible renal cancer subtypes can be found to exhibit any given imaging pattern. If classic features are present, it is reasonable to suggest a specific renal cancer cell type when reporting results of imaging studies. However, one should indicate this to be the most likely diagnosis, since cancer is a heterogeneous entity with overlap of features among subtypes.
A glimpse into a possible future for imaging is illustrated by considering the current approach to the hereditary renal cancers. In this setting, knowing the genetics of the disease provides insight into the probable histology of the cancer, its likely biologic behavior, and its response to therapy (5). For instance, the constellation of findings constituting the von Hippel–Lindau syndrome (cerebellar hemangioblastomas, retinal angiomas, pheochromocytomas, pancreatic and renal cystic and solid lesions) is strongly associated with clear cell carcinoma of the kidney, which typically has a low Fuhrman grade, grows slowly, and responds well to enucleation or local ablation. Similarly, hereditary papillary renal cancer syndrome causes a very low grade type 1 papillary tumor that grows slowly and does not require aggressive treatment. The renal lesions associated with Birt-Hogg-Dubé syndrome (skin fibrofolliculomas, lung cysts, and renal lesions) are often, although not exclusively, chromophobe carcinomas, which are also generally slow growing.
Despite extensive work in this area and considerable progress over the past 10 years, we are only at the beginning of our understanding of kidney cancer genetics. For instance, a recent observation is that a subset of patients with the syndrome called multifocal cutaneous and uterine leiomyomas, or Reed syndrome, develop a distinctive papillary-like tumor (6,7). This syndrome has been termed hereditary leiomyoma RCC, and the tumors associated with it have a high nuclear grade and demonstrate very aggressive clinical behavior. These tumors rapidly spread by perinephric and lymphatic invasion (as opposed to hematogenous spread of tumor, more commonly seen with clear cell carcinoma) and are highly lethal. Therefore, they require more vigorous and proactive treatment than is usual for renal cancers that are either hereditary or sporadic. Thus, knowledge of hereditary renal cancer syndromes can lead to specific diagnoses and help direct treatment and surveillance policies. New, minimally invasive therapies such as cryotherapy and radiofrequency ablation are ideally suited to some of these diseases but are less well suited to others. Moreover, because they arise from derangements in specific metabolic pathways, such tumors may one day be amenable to treatment with targeted molecular therapies.
What might the future hold? The classification of renal cancers will no doubt become more complex. Indeed, the current WHO classification presented in Table 1 of the preceding article can be thought of as a starting point for the future (1). We already know that most tumors contain more than one histologic element, and future classifications will no doubt incorporate this information into a more complete description of tumors. For instance, the majority of tumors currently classified as "clear cell carcinoma," representing more than 70% of renal tumors, contain clear cell and granular cell elements as well as a wide variety of genetic abnormalities. The clear cell carcinoma of the future will undoubtedly be subdivided into finer and finer subtypes.
Will imagers of the future be able to provide more accurate and specific diagnoses that predict biologic behavior? If we look at fluorodeoxy-glucose (FDG) positron emission tomography (PET), as our best current example of molecular imaging, we have a long way to go. FDG PET would certainly be attractive as a prognostic and diagnostic agent. Unfortunately, across many primary renal tumors, FDG has a fairly low sensitivity. Indeed, at least 25% of metastatic lesions fail to take up FDG (8). The major benefit of FDG in renal cancer is that when it is taken up by the tumor, it is highly specific for cancer, but histologic subtyping according to uptake is not yet possible.
Is it reasonable to expect the development of new molecular imaging methods that will be able to identify specific histologic subtypes of renal cancer? Certainly not in the near term. Even if it were currently feasible to differentiate renal cancer subtypes by using imaging, it probably would not happen soon. The economics of such agents is unfavorable. With only 35,000 new cases of renal cancer per year in the United States and perhaps 700,000 worldwide, it is hard to imagine sufficient numbers of cases to justify the required monetary investment to bring the agent to market. For comparison, over 20 million Americans alone take the cholesterol-lowering agent Lipitor (atorvastatin; Pfizer, New York, NY) every day, and annual worldwide sales are over $12 billion. Drug company executives will be using examples of such drugs to compare profits to be derived from molecular imaging diagnostic agents.
However, molecular imaging agents that identify targets amenable to specific molecular therapies hold more promise. These imaging agents, like their matched treatments, span more than one type of cancer and therefore offer a larger, more viable market for drug developers. Recently introduced therapies with angiogenic inhibitors (eg, Avastin [bevacizumab, Genentech, South San Francisco, Calif] and sorafenib [Nexavar, Onyx Pharmaceuticals, Emeryville, Calif]) have proved to be successful in some cases of advanced renal cancer (9). Therefore, agents that identify angiogenesis and the specific targets of angiogenesis, such as integrins or vascular endothelial growth factor (VEGF) receptors, may assume greater importance. Similarly, monoclonal antibody–based imaging could identify tumors with targets amenable to treatments such as Herceptin (trastuzumab; Genentech) (the monoclonal antibody against HER2/neu receptors), cetuximab (Erbitux; ImClone Systems, New York, NY) (the monoclonal antibody against HER1 receptors), and Avastin (the monoclonal antibody against VEGF). It is clear that molecular imaging agents linked to molecular therapeutics will open up new diagnostic opportunities for radiology.
Thus, Prasad et al (1) have performed a service by updating radiologists on advances in the field of renal cancer. Advances in one field, such as genomics and histologic biomarkers, do not always translate immediately into advances in another field, such as imaging, but it is likely that such information will be extracted from images, if it becomes clinically important to do so. The imager of the future will certainly need to have a broad-based expertise in traditional anatomy and physiology, buttressed by a working knowledge of genomics and cellular biomarkers, all no doubt supplemented by a powerful, personal handheld computer.
 |
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
 Top References References |
|---|
2 Department of Radiology, University of Texas Health Science Center, San Antonio, Texas
We thank Dr Choyke for the insightful comments on our article on the histologic subtypes of RCC. As Dr Choyke mentions, we are witnessing increasing application of cytogenetics and molecular biology in oncology. The burgeoning cytogenetic and proteomic data have to some extent helped clarify cell of origin and elucidate specific oncologic pathways, thus paving the way to "molecular therapeutics" aimed at cellular-subcellular targets (1,2). Improved understanding of the biologic behavior of tumors with prognostic and management implications may result in change in screening paradigms and posttreatment surveillance imaging protocols. Oncology is increasingly drifting toward a "individualized patient management " paradigm based on genophenotype correlation.
Detailed studies of the ever-increasing spectrum of hereditary RCC syndromes (currently making up < 5% of RCCs) have immensely contributed to our understanding of the pathogenesis of sporadic RCCs (3). For example, it is well known that von Hippel–Lindau syndrome is associated with development of clear cell RCCs. It is interesting to note that up to 80% of sporadic clear cell RCCs demonstrate VHL gene mutations (3,4). Several drugs that act on distinct tumor pathways (VHL-associated hypoxia-inducible factor pathway, c-MET receptor tyrosine kinase pathway) are in various stages of development for potential use in patients with multicentric RCCs or advanced disease (2). Thus, knowledge of the histology of RCCs may determine what specific drugs may benefit a particular patient. Some histologic subtypes of RCC have unique imaging findings, which may permit prediction of histology with its attendant implications for management and prognosis. Also, the tumor response to molecular therapeutics may be vastly different than the response to standard cytoreductive therapy. Cystic degeneration of gastrointestinal stromal tumors responsive to Gleevec (imatinib mesylate; Novartis Pharmaceuticals, East Hanover, NJ) is a good example.
Improved understanding of the molecular aspects of RCC will help radiologists design functional imaging techniques aimed at tumor physiology or biology that provide useful insight into the vascularity and invasiveness of a tumor. Our unique position in the "patient management chain" permits us to not only diagnose and stage RCCs but to treat them as well. Radiologists are already treating patients with RCC using innovative technologies such as radiofrequency ablation and cryoablation. We may also be positioned to provide targeted, tumor-specific therapies, especially in patients with multifocal tumors and advanced disease. Thus, it is important for radiologists to understand the complex world of "oncologic cytogenetics" to better manage cases of RCC.
|
References |
|---|
|
TopReferencesReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOGRAPHICS | RADIOLOGY | RSNA JOURNALS ONLINE |
