In the News

by Ellen Mahoney, M.D. FACS, The Times-Standard, 10/10/06.

The appearance of breast cells changes as a normal healthy cell becomes malignant. This was noted centuries ago, with the invention of the microscope, and for a long time, all we had was the appearance of a cell to tell us what its nature and potential was. The degree of deviation in appearance from a normal breast cell made the diagnosis of cancer, and treatment choices limited.

About 100 years ago the sorting of genetic characteristics based on appearance of the pea or the fly was noted, and the field of genetics was born. It was also noted that the material in the nucleus of the cell was copied and divided between the two daughter cells as the cells divided. But no one understood exactly what a gene was, or what it did, or how.

A mere 50 years ago the DNA molecule was described. Finally, we understood that genes are sections of the DNA molecule that serve as blueprints for making proteins in the cells. These proteins serve both the structure and the function of the cells. Every cell has all of the genes necessary to make all of the proteins in the body. It was around this time that the concept of mutation was born; something, be it a random mistake or a toxic foreign influence, changes the structure of the DNA while it is dividing. Most of these changes in the blueprint have no influence, and a few cause the cell to die. But some let the cell live, but cause changes in the DNA that ultimately can result in the gene making the wrong protein for that cell. One mutation doesn't make cancer, but once the DNA is changed, and passed down to a daughter cell, the chances become greater that more mistakes will be made in successive generations of that cell, and over many years the result can be a cell that not only grows and divides without regard to whether you need more of that cell type, but that also is able to digest its way through normal tissue, leave the organ where it originated, live in the circulation of the body, attach itself to tissue elsewhere in the body, and recruit a new blood supply to feed itself, all while subverting the immune system which is always on the lookout for cells that are not normal. Then it is cancer. But many mutations of different types can result in cells that LOOK the same. And the appearance of the cell doesn't really tell us what functional genes are in it. Now we know that cancer cells vary a great deal in their ability to move and grow, and we also have a large variety of treatments. How do we tell which cancers are more dangerous than others, and how do we match the most aggressive treatments to the most aggressive cancers? We do not want to risk under-treating or over-treating our patients, yet it is clear that the appearance of the cell is not enough to help us make these decisions.

The first major individualization based on the actual state of the DNA in a cancer cell was the determination of the estrogen receptor protein. This happened in the 1970s. Cells that had DNA that made this protein had a better prognosis, and also could be damaged by drugs that interact with this protein. This was the first time that we were able to match treatment with the actual state of the DNA, and it is still the most powerful. There have been some others, notably Her2/neu protein. This is an abnormal protein, so its manufacture is directed by an abnormal gene in the DNA. Tumors that have a lot of this type of damaged DNA can be effectively treated by an antibody to the protein, called Herceptin. These tests can help us select specific treatments based on specific genes, as determined by their products.

Progress in this area, though very exciting, has been very frustrating because it has been so slow. Meanwhile, every day, women and their doctors are trying to find ways to select treatment, and even someday to give better prognostic information. We have identified a lot of genes in breast cancer, but we don't really know what all of them do yet. Cell cultures to determine the sensitivity of an individual's cancer cells to individual drugs, analogous to culturing bacteria and testing antibiotics against them, have not been successful in breast cancer.

Scientists decided to take another route to getting this clinical information by looking back at a cumulative collection of about 400 breast cancers in women who had small tumors, and who did not have nodes involved (by our older and more inaccurate ways of determining that fact). Most, but unfortunately not all of them, had estrogen receptor proteins on the tumor cells and were treated with tamoxifen. This was a good choice, as the patients who have small ER+ tumors without nodes involved present the greatest dilemma in choice of therapy. Many of these patients will do just as well with or without chemotherapy if they are given hormone-blocking drugs. Is there a way to tell in advance which of these patients should get chemotherapy?

The 400 tumors were analyzed for genes that they had in common, and out of hundreds of genes, 16 genes were chosen, along with five reference genes. This panel was then compared, again retrospectively, to the tumors of 2600 women in several small studies (where many of the authors are employees or consultants for Genomic Health, the company that makes the test). The score is reported in three categories denoting the chance of recurrence: low, medium or high, and the results were intriguing. Intriguing enough to justify further study, using the score to select therapy, and then seeing if there is a difference in survival, but not intriguing enough to justify the widespread use of a very expensive ($3,700) test, not always covered by insurance, to make clinical decisions today. Furthermore, one of the independent attempts to validate this panel of genes and the recurrence score, done at MD Anderson Cancer Center, did not show a relationship between the genes selected and the recurrence score.

The test was validated by "predicting the past," looking at a set of genes that happened to occur frequently in a population and seeing how patients who had these genes did. We don't even really know if the 16 genes that were selected are the most important ones, since there was nothing but observation used to choose them. There is absolutely no proven value in using Oncotype DX to select particular drugs and, at best, the test has limited utility in giving us one more possible "tie-breaker" to add to other information we have about a tumor that has mixed messages in its prognostic features. Prospective trials, where treatments are chosen on the basis of a particular set of genes and a survival benefit demonstrated, are now underway, and if they show a benefit, they will be the gold standard. There is also a competing 70-gene product called Mammaprint which is also undergoing prospective trials, so Genomic Health is working very aggressively to try to make clinicians think that OncotypeDX is the final word!

What should you do now? The main question is whether your case already has enough traditional indicators making the treatment decision very clear without the test. Many will, and getting the Oncotype Dx test, if it is ultimately proven to be inaccurate, may actually lead to wrong decisions about your care. The next question is whether you fit the criteria for the test. Clinicians, faced with tough decisions, are already extending the use of the test beyond the categories of patients that Genomic Health has validated it for, even retrospectively. While this is tempting, and even I have fallen into this temptation, the results have to be viewed extremely skeptically, as there is not even retrospective validation for these patients. Third, if you and your doctor want to do the test, and have thought through what you will do differently depending on the results, be sure that you get insurance approval with a promise to pay before the test is ordered.

And stand by! Even though this test isn't all that we need now, it possibly will prove to be a next step in letting us know more about the biochemical capability and susceptibility of individual cancer cells based on their DNA, and that is the way the field is moving. The tests we really need to help us choose proper treatments and to help us know which tumors are particularly dangerous to your long term health are coming. They just aren't here yet.

For more information, call the Humboldt Community Breast Health Project at 825-8345.

Dr. Ellen Mahoney is an Arcata surgeon and president of the Humboldt-Del Norte County Medical Society.