To most patients with cancer—and even to the physicians caring for them—the word “genomics” has little or no meaning. Many are vaguely aware that genomics has something to do with genes and may somehow relate to a genetic risk for cancer within families. However, when told that genomics entails an attempt to “fingerprint” their cancer, they immediately grasp the basic concept. Patients and physicians know that hardly a month passes without the popular press announcing a major breakthrough in our understanding of cancer.

Most oncologists also understand genetic fingerprinting as it relates to cancer: that it will help differentiate tumors destined to grow quickly from those that will not, as well as predict where an individual tumor is likely to spread. Oncologists expect that through genomics, tests will be developed to help fulfill the long-standing hope of identifying which tumors are either sensitive or insensitive to current toxic therapies, thus sparing patients from certain treatment regimes that may be ineffective.

Remarkably fast progress

The last decade has seen amazing progress in understanding the basic biology of cancer and developing exciting new treatments—including the tyrosine kinase inhibitor STI 571 against chronic myelogenous leukemia, as well as the tumor-angiogenesis inhibitors. During this period almost every major hereditary form of cancer has been linked with specific mutant genes. Syndromes that were controversial clinical entities, such as the hereditary nonpolyposis colorectal cancer syndrome, have been proven to exist by elucidation of their molecular basis. Alongside the advances in molecular oncology have been the technological and scientific triumphs of the Human Genome Project and private genomics companies.

Such rapid progress could not have been realized without the development of automated, fluorescent DNA-sequencing approaches. DNA microarray technology makes possible large-scale genotyping and parallel gene-expression analysis. Most important, computational approaches have been refined to the point where these huge and complex data sets now take days rather than years to analyze.

Perhaps the advance most visible to the lay public is the identification of individual genes involved in cancer. Many of these genes can contain mutations that, when present in the germline state, result in highly penetrant hereditary cancer syndromes. The detection of such mutations can, through careful follow-up screening, significantly improve an affected patient’s prognosis.

Other genes, although not directly involved in cancer, have been demonstrated to be of prognostic significance and, potentially, to influence disease management. Several studies of p53, for example, indicate that presence of the gene portends a worst-case prognosis in node-negative breast cancer.¹ The vast majority of p53 mutations, in fact, are missense mutations resulting from single base-pair changes. DNA microarray technology can be an accurate and economical alternative to fluorescent sequencing in detecting these mutations.²

Breakthroughs spawn unique challenges

The excitement of the last decade must, however, be tempered by the scientific, economic, and ethical challenges that still confront us in applying these breakthroughs in the clinic.

For example, identifying a specific present mutant sequence is and will continue to be crucial due to the importance of genotype-phenotype relationships. Yet the process of confirming the mutation poses several problems.

First, because the focus on genotype-phenotype relationships is relatively new, the current studies are small, and securing a quicker enrollment of patients into clinical studies of genotype-phenotype relationships remains a challenge. Moreover, clinical databases often are simply too limited to allow us to understand the penetrance and phenotype of every mutation in any of the known cancer-related genes. This limitation might be overcome, however, by using the archived, paraffin-embedded material in hundreds of surgical pathology laboratories. To do that ethically, medical chart researchers and genetic counselors will need to obtain follow-up information and consent from patients. Economically, such an effort will require a major funding commitment.

Second, the full potential of automated, miniaturized, parallel genotyping technologies has not been realized. If automated “lab on a chip” technologies are to become widely available, private and public assistance is needed. Implementing such genotyping infrastructure requires hiring well-trained personnel to support the research. In addition, members of the research community must dispel—particularly for graduate students—the notion that high-throughput clinical genomics is merely a big “fishing expedition,” by supporting and encouraging genomic investigation.

Third, confirming a frame-shift/nonsense mutation from deletions and insertions (common in hereditary cancer genes) requires that sequence data be read by humans rather than machines, a process that can increase costs significantly. Indeed, the real bottleneck to progress today lies in developing integrated approaches to accurately collect and process the enormous amount of data collected from the trials. Funding research that is focused on the development of integrated systems to perform such studies should be a top priority. This funding effort should address concerns about how to mine data in standard written medical charts as well.

We would be remiss if we failed to mention that this exciting new area of clinical practice also raises ethical concerns. For example, is it ethical to use the large banks of tissue available in surgical-pathology tissue blocks around the world without the consent of the people from whom the tumors were resected? Conversely, if one of those people cannot be tracked down, is it ethical not to proceed with this research? What are the implications of telling someone that an inherited set of metabolic enzymes puts him or her at a slight but real increased risk of developing cancer?

Already, experts such as the secretary of the U.S. Health and Human Services Advisory Committee on Genetic Testing and representatives from other areas of genomic research are pondering and making recommendations about these moral and ethical dilemmas. It’s worth noting that some start-up clinical genomics companies have business plans that include ethical review boards made up of outside contractors. Simply put, we can’t afford not to do this research, and we must establish a way to carry it out openly and ethically.