In doing research after my breast cancer diagnosis in 1989, I came across a National Cancer Institute clinical alert sent to oncologists. It announced the results of several clinical trials showing that women with node negative breast cancer benefited from chemotherapy. Prior to 1989, chemotherapy had been reserved for node positive women. Unfortunately, the alert didnt provide a number needed to treat analysis. So here it is: one hundred women would have to undergo chemotherapy for 10 to benefit. Ninety women would risk toxicities but get none of the benefits. And the harms include not only those that end your life like heart failure and leukemia but also more common ones that can RUIN your life: not being able to have a child or more children, loss of cognitive function, loss of libido, severe arthritis, seven times the risk of bone fracture. Thus began the standard practice to administer chemotherapy to women with node negative breast cancer that persists to this day: Treat everyone to improve the survival chances of a small minority.

Recently as I was researching estrogen receptors I was struck by the very strong bias in the medical community towards chemotherapy over hormonal therapy. For example, Craig Jordan, the developer of tamoxifen, recounts how his colleagues told him in the 70s to give up on tamoxifen. You are ruining your career. Chemotherapy is going to cure breast cancer. The prevailing idea was that you had to fight fire with fire. Tamoxifen was too gentle to be as effective. Even now, Europeans are more inclined to treat the same patients with hormonal therapy that American oncologists would treat with chemo.

In the 90s we were subjected to even bigger guns, more firepower with high dose chemotherapy and stem cell transplant. This tragedy was a watershed event for many advocates. More is not always better! New and unproven is not always better.

Another example of over-treatment is evident in the results of the MA-17 trial reported just last year. On top of surgery, radiation, for some chemotherapy and five years of Tamoxifen, node negative women will get 5 years of Femara and a bisphosphonate to counteract bone lose caused by Femara. And now another trial is looking at adding 5 MORE years of Femara. Femara is a drug that cuts off 99% of a women's estrogen. What is the toll of THAT? And remember -- a majority of these node negative women were cured with just surgery or surgery and radiation.

It seems to many of us that drugs are always being added but never subtracted.. Why would a drug company try to find out who doesnt need their drug especially if it is already on the market? That is one factor. Another factor has been the lucrative reimbursement policies that oncologists get for using chemotherapy as opposed to hormonal treatment or even doing nothing. Fortunately that reimbursement policy is being changed.

NOW, FINALLY with the development of gene expression profiling the question of taking away drugs is being seriously addressed. This is an extremely important development. If it had been prioritized it could have been done sooner, but this new technology, hopefully, has the POTENTIAL to be much more accurate because it allows you to look at more than one gene at a time and to look at levels of gene expression..

However, advocates need to be skeptical and ask the hard questions. Our livesboth their quantity and quality--are at stake. WHAT IS THE CLINCIAL RELEVANCE OF GENE PROFILING? Here on the screen I have listed five issues that I will be addressing today.

Number One: Is there enough evidence that gene profiles are ready for use in the clinic? I will be talking about OncotypeDX and MammaPrint because they are on the market today and because they will be talked about by two of our panelists.

There are no definitive rules for validating a biomarker and much disagreement. I agree with those who look for a BODY of evidence. I dont think we are there yet with OncotypeDX or MammaPrint-Im not sure we have done adequate studies and Im not sure all of the evidence we HAVE is transparent. All validation studies must be published and be part of the product labeling. We need studies that directly pit these new gene profiles to the best of the more traditional biomakers. A study published in the May 19th 2004 issue of the European Journal of Cancer showed that MammaPrint did NOT do noticeably better than the Nottingham Prognostic Index which uses traditional biomarkers to make treatment decisions. In addition, one of the Mammaprint validation studies includes some of the same tumor samples from which the profile was originally developed. That defeats the whole purpose of validating a biomarker. which is to see if the profile that you have developed from one group of samples performs similarly in a totally independent group.

In the Oncotype DX profile, the largest number of genes in the OncotypeDX profile are associated with the estrogen receptor. Yet comparisons were made to samples whose ER status was determined by the Ligand Binding Assay despite the fact, that ER status as determined by the Immuno-histochemical assay, IF PROPERLY DONE, is more accurate. In the Kaiser Permanente study OncotypeDX was compared to tumor size and grade. What about lymphatic invasion, or HER 2?

OncotypeDXs gene profile was developed and validated with stored tumor tissue. The beauty of using stored tissue is that you dont have to wait 10 or 20 years to see how women fare. But there are drawbacks. Does gene expression in paraffin embedded tissue change over time? Also, although the validation studies were designed prospectively, sampling biases may exist when using old stored tissue. Consider the NSABP- B14 samples used in the development of OncotypeDX. Only 1/3 of the pool from which these samples were taken had been preserved. Could this be a biased sample? Maybe these samples were from women who tended to have a family history and were particularly interested in furthering research? David Ransohof, one of our panelists, will be addressing bias in biomarker validation more thoroughly, so I will leave further discussion to him.

Two trials that have both randomized and prospective components, are in the works for these two genetic profiles. The Oncotype trial is called Tailor x and the MammaPrint trial is called MINDACCT. Other panelists will be describing these trials. I think such a new technology needs further testing.. It is crucial to establishing a new biomarkers superiority to the best combination of traditional biomarkers. For example the MINDACCT trial will be comparing how women do when their treatment decisions are made by MammaPrint vs. AdjvuantOnline. AdjuvantOnline is perhaps the best application of traditional biomarkers for breast cancer treatment decision making. MINDACCT is a 5 year study. The Onocotype study will be randomizing women in the middle risk group to either chemo or no chemo to see how accurate their profile is for this group. They will further validate their low risk and high risk groups with prospective data. Their trial will take about 7 years. Of course, the problem with these kinds of studies is that they can be out of date by the time they are published. But they are an important proof of principle.

Number Two: The second issue concerns FDA regulation and insurance coverage. FDA regulatory power over biomarkers is very weak. I believe it should be strengthened. A biomarker test that is only done in one lab -- called a homebrew -- is free to enter the marketplace without any regulation from the FDA.... Even those biomarker that do require FDA approval dont have to prove that they are better then what is already on the on the market..

Since Oncotype DX and MammaPrint are already on the market, pressure mounts from patients and doctors for insurance companies to pay for them. In fact, some already are. I think these gene profile tests should be available to women only within the context of a study or clinical trial, It is important to prevent their premature widespread use. Once they are on the market it is hard to take them off, if it doesnt work as well as expected. There are quite a few unproven biomarkers in different cancers already on the market. Raising healthcare costs, not to speak of the harm some of them have done-like the PSA test is a result of a weak regulatory policy.

Number Three: The third issue is acceptability. I am glad that these profiles are being developed and hopeful about the prospects of a test that will help women with breast cancer avoid unnecessary treatment. At the same time, I have reservations about HOW these profiles will be used in the real world. Will women choose chemotherapy even when they are told they have only a small chance of a recurrence? The bias towards chemotherapy and its overuse still permeates our society and will affect how gene profiles are used. Research has shown that many women will opt for chemotherapy for a one or two percent benefit. But OncotypeDXs low risk group has a 6.8% chance of distant metastases in 10 years. MammaPrints low risk group has a 15% chance of recurrence (both local and distant) in 10 years. Remember that is a median so half in the low risk group will be lower and half will be higher. Will women and their doctors consider these risks low enough to forgo chemotherapy, or in the case of MammaPrint to forgo both chemotherapy and hormonal therapy?

What happened at a recent meeting of The American Society of Breast Disease is a perfect illustration of what I am talking about. When the audience of doctors was asked how many have used OncotypeDX to make treatment decisions, almost the whole room raised their hands. The next question was: if your patient had a large tumor but a low risk genetic profile would you recommend that she forgo chemotherapy? Only 2 people raised their hands. We have to consider whether these profiles are worth it if they are just used as an add on to all the other traditional biomarkers already in use and if they are, how much value does it add? Oncotoype Dx costs 3,685 dollars which is not an insignificant amount. Whether it will be able to fulfill its potential to SIGNIFICANTLY cutback on the toxicities and the cost of unneeded chemotherapy and hormonal therapy remains to be seen.

Cost is a very real concern given the fact that ours is a market-driven health care system that encourages the use of the most expensive test and treatment. THAT IS FOR THOSE WHO ARE WELL INSURED. FORGET about those who arent; they dont even get standard of care. Their death rate from breast cancer is 50% than women of higher socioeconomic status.

Number Four: The fourth issue distinguishes between prediction and prognosis. The real promise of gene profiling will come when it can predict response to specific treatment. This would be the most effective way to prevent over-treatment and minimize toxicity. There is no demand among ER- women for hormonal therapy, and no demand for Herceptin by women who are Her2 negative. Both MammaPrint and OncotypeDX have tantalizing data that chemotherapy is of no help and might even be harmful for certain low risk women, while chemotherapy and not tamoxifen might help high risk women. But the subsets are small and the data has not been published in a peer reviewed journal. HOWEVER, even the ability to PREDICT response is ONLY half way there. The biggest challenge will be to develop new therapies for women who dont respond to any of the existing treatments.

Number Five: The fifth issue concern business motivation. Will there be so many targets, so many different types of cancer that drug companies will not find it profitable to make drugs for these small markets? WHAT WILL the PRICE OF THESE DRUGS BE? One example: The number of women with ovarian cancer who are HER 2 positive is so small that Genentech, the maker of Herceptin, decided it was not worth doing a clinical trial. This decision was not only a financial one but also a practical one: Was this group too small to provide enough trial participants for a clinical trial in a timely fashion?

In conclusion, advocates are hopeful that gene profiling for prognosis and prediction will prove effective. They can help to radically change cancer treatment and its decade long reliance on chemotherapy. Still, there are many question that need to be addressed.

I would like to leave you with several quick additional points about things important to advocates that should be started right now. Oncologists have to change how they discuss the risks and benefits of chemotherapy with patients. Hearing that chemotherapy reduces your risk by 30% is misleading; indeed meaningless unless you know your risk to begin with. If your risk of recurrence is 15% you are only reducing it by 5%. Further, regardless of how risk reduction is presented, it does not reflect the harm done to those who didnt need the treatment. WHAT IS THE HARM? There are the toxicities that can end your life: leukemia and heart failure. But more common, are the ones that can RUIN your life, like inability to have a child or have MORE children; the loss of libido; loss of cognitive function; severe joint pain; and bone fractures. These harms are usually ignored or understated and THEY ARE ALSO UNDERSTUDIED and UNDERFUNDED. A study in the recent issue of the JNCI showed that both short term and long term cancer survivors were twice as likely as their cancer free twins to have cognitive dysfunction. Could this be due to treatment? For women to make the right decisions, even with these new genetic profiles, we need to know about the harms of chemotherapy, or for that matter for any therapy.

All your eggs in one basket