15th Congress of the European Society of Surgical Oncology (ESSO), 15–17 September 2010, Bordeaux, France

Interview with Professor Remy Salmon (Institut Curie, Paris, France)

Oncotype DX gene assay test

Remy, it’s great to see you here and you’ve been actually using this genomic classifier, the oncotype DX in practice. Could I take you back, first of all, to the St Gallen meeting, I think it was 2009, where they had the consensus which gave support to the idea of using gene testing. What happened at that time and do you think it marked a way forward?

Genetic testing is not  recent information; the first data on that, going back to the beginning of the 21st century with work done mainly in Holland with the seventy genes test and several other tests, it was the Amsterdam signature, Rotterdam signature. In parallel there were the developments in Pittsburgh from Bacon and colleagues of the oncotype design. The research is coming from each part of the ocean and, in certain ways, the war is not finished, it’s still working the United States against Europe and trying to do the best things for the patients.

So there are systems in the US and in Europe and both are representing a way of actually determining genetically what might happen to your patient with early breast cancer.

It’s a little more complicated, if I can, because the management of breast cancer is in constant improvement. Constant improvement means that there is never 50% of the management which is going to change one year because of the discovery of one test, of one technique and so on. Because the main improvement is early diagnosis, which means we have smaller and smaller cancer and we are dealing with patients now in which the hope of survival at ten or fifteen years is more than 80%. To make progress in the last 20% is becoming more and more difficult. On the other hand there are patients in whom you are expecting a very good result because they have good usual prognostic factors and they do badly, and we don’t understand why. So the genetic testing came in addition from the size, the nodes, grade, lymph node invasion, to add more information and in order in good prognosis patients to select the very good prognosis patients and to distinguish them from the bad prognosis patients. As usual there is, in the middle, the intermediate patients which represent a constant situation in which we are facing difficulties.

Clearly genetic testing could go into more widespread use, how do you recommend doctors to use it for their patients with early breast cancer?

At the present time in France we are involved in a European trial called the MINDACT trial related on the signature from the Amsterdam signature. In the United States it’s now widely used, the oncotype DX test using 21 genes, and to select in the small tumour with positive hormonal receptors and node negatives with a little extension with one, two, three nodes positive, to select the patient in which we can add some benefit from chemo, or in which chemotherapy will be of no benefit. There is an important chance because in the publication, at the present time, there is between the teams from 20-30% of the decision for chemotherapy which is changed to hormonal therapy, which obviously saves money but also in what we call the cognitive function of our patient, allows them to avoid chemotherapy with all the side effects of the chemotherapy which is, from my point of view, an enormous improvement. 

So what do you do with your patients now?

A patient is operated on, she has a biopsy and she is operated on and then we discuss on the tumour, the nodes, the chest X-rays, liver test and so on. And we meet in a local consensus conference, it’s a multi-disciplinary team conference, in French it’s called a réunion de concertation pluridisciplinaire, RCP. Then we discuss the best post-operative treatment according with the local referential. But it’s life that there are patients who don’t fit with the referential so we need additional information. Take an example, if you have a young patient, 35 years old, who has a highly aggressive cancer, 3cm diameter, highly proliferative cancer, negative receptors, positive nodes, nobody will discuss about chemotherapy. At the opposite, you take an old lady, 75 years old, small tumour, negative nodes, positive receptor, low proliferation, nobody will give chemotherapy to her. But when you deal with a perimenopausal patient, the females are separated into before 50 and after 50 – before 50 you are young, after 50 you are old, which means you are menopausal. It obviously doesn’t work in real life, but the consensus and the rules are very often decided on this cut off at 50 years old. So if you are 49½ years old and you have positive receptors but they are positive but not so strong and the oestrogens are positive, progesterones are negative, there is a proliferative index in the middle, you have a difficulty to decide what to give as medical treatment. So, as I said a little earlier and this morning, there are two solutions. Either there is a dominant male in the consensus, in the multi-disciplinary team, and the dominant male says, “OK, this patient will receive that. So I decide my patient will never receive chemo.” The other solution, which is a more objective solution, is to use a genetic test in which this intermediate patient can be separated either into a good group or a bad group.

So what you’re describing is authority based medicine in comparison with evidence based medicine. Do you think the genetic testing can now give the evidence base for making those decisions?

Yes. It relates to evidence but we are dealing with men, with females of course, but it will not solve everything but it will help us to improve the management of the patient. With the advantage of genetic testing is that they are objective, reproducible and it can be shared from the United States to France. While if you just look at the pathology on a slide, there is always some kind of subjectivity, what you called not evidence based but authority based.

So you have that now, if you have a patient with a small tumour, oestrogen receptor positive, that’s a good candidate, there would have been doubt, what about HER2, for instance, and what about nodal status?

With HER2 it’s simple. HER2 is simple, it’s solved, it has a targeted treatment with Herceptin. So if the patient has HER2, the patient will receive Herceptin and if the patient receives Herceptin, she will have chemo. That’s no discussion, nobody will discuss it any more unless she is 80 years old or if the tumour is less than 5mm. That’s not any more a question, HER2 patients.

What about nodal status, then, how does that come into it? You have a lot of nodes, does that mean you still don’t need the test?

The genetic testing is used at the present time for selecting in negative nodes or small node involvement. When there are a lot of nodes the genetic testing will probably be used in the near future to select which drug to give to the patient, so to predict the sensitivity to a certain drug which means there are two different ways to think - one is the prognostic way and the second is the predictive way. The prognostic way is to say, “Madam, your overall survival and disease free survival is at 75%, 85%, 95%.” That’s the prognostic. The predictive way is to say, “You have positive hormonal receptors so you can receive hormonal therapy. You have an HER2 tumour, so you have to receive a targeted therapy with Herceptin.” That’s true for hormonal treatment and targeted treatment, but for chemotherapy, at the present time, the only way to decide for chemotherapy until the genetic testing, was the proliferation. So the highest proliferation, the highest likelihood of being efficient with chemo, but that was not at all specific and, in addition, it was not specific of any kind of drug which means you can give gemcitabine, adriamycin, cyclophosphamide and there was no specific.

Has there been any clarification now since the era of genetic testing in the choice of cytotoxic chemotherapy?

It remains very complicated because of the heterogeneity of the tumour and because, in the clonal evolution you can have some part of the tumour which can express certain genes which can predict their sensibility to a certain drug and in another part of the tumour, some genes which are a demonstration that the drug will not be efficient. So we are really facing a difficulty of homogenisation of the tumour with the risk that if you select a certain clone of tumour, which is very sensitive to the drug, the next clone here, which is not sensitive to the drug, will take a selective advantage because they were living in equilibrium of if this one dies, the other clone will develop.

And, indeed, tumours evolve.

Yes, with eventually a cure. But it’s speculation, that’s still under research.

What about radiosensitivity, can you do any predicting in that situation?

Well that’s a difficult question, I have to remind you that, first of all, I am a surgeon but of course I have been working in the Curie Institut so I have some idea about radiation therapy. No, there are some works in genetic testing and with predicting the radiation therapy sensitivity and more of it to predict side effects of radiation therapy in order to avoid the definitive side effect of radiation therapy with fibrosis or other stuff. But it’s not yet used on an everyday basis.

So far then, we can get more refined predictive information for patients. What do you feel is the prospect of improving survival and reducing morbidity overall among breast cancer patients?

Well there are two parts to your question because you say how can we improve the survival and the quality of life?  Would you like to live longer but in a bad condition or a little shorter but in a better condition? That doesn’t relate any more to medicines, that’s getting into philosophy. Obviously we like the targeted therapy which reminds us of bombing in some place in the Middle East, about the targeted therapy without collateral damage. So we try, of course, to avoid collateral damage, which is not very easy and we discover from time to time with new drugs, collateral damage which we have not expected. Clearly, the best way to improve the treatment, as well for the treatment and for the aesthetical outcome, for instance in breast cancer, is to discover the cancer at a very early stage. The earlier the stage, the lighter the treatment and the better the result on a long-term basis because it’s easier, which it’s quite evident, that’s common sense; the bigger the tumour, if you come with a T4 tumour, it’s a pity but you are not necessarily obliged to cure the patient. But if you have a less than 1cm tumour, then you have to fight to try to improve and not to let the patient with strong side effects.

And because with refined detection techniques you can actually find these very small tumours, then there may be a case for saying that genetic testing might be your best shot for finding out the aggressivity or benignity, if I can use that word, of the tumour.

One of the first answers is the smaller the tumour, so the genetic testing requires a very small amount of tumour which is part of the question because you don’t have a lot of material to use to analyse and do the classical test, genetic testing. So the genetic testing requires a small amount of tumour and it’s easy and reproducible. It’s much easier in the American test, oncotype, because it’s done in parafin while the European tests are done on frozen sections and if you do a frozen section you have to take a sample, put that in the frozen section and put another part of the tumour aside to do the classical test in order to compare them. So if the tumour is too small you cannot do, for instance, MINDACT. The MINDACT trial is reserved for the patients with tumours of 1cm and over; when it’s less than 1cm, for instance in my institute, our pathologists just refuse to do the test.

So where do you see the application of oncotype DX principally then? In these very early tumours?

Yes, you can use it easily because it’s one of the slides of the pathologists which is sent to the lab and to the referential laboratory and is very easy to use.

What then are your recommendations to busy doctors with worried patients who need to know what’s the best way forward?

Do your best to get oncotype because it will improve the management of your patients, that’s clear. The genetic testing in the good selected cases, which is the middle cases I told you about, not the young, very aggressive cancer, not the older very slow growing cancer, but in the middle. The use of genetic testing, namely oncotype, will help to improve the management of them.

How much would it improve risk adapting decision making?

There’s between a 20 and 30 chance of decision for chemo. You will reduce the fee, all the patients with cancer accept to be operated on. You remove the tumour, that’s good, they feel better after the tumour is removed. But when they wake up after my surgery, they say, “Doctor, should I have received chemotherapy?” And receiving chemotherapy is considered as being an aggressive danger, a risk, something more dangerous, their life is more concerned. They know they will lose their hair, no more eyebrows, ovarian function will stop and that’s not good news. So if we can reduce the amount of patients who receive chemo, it’s clearly an improvement.

So potentially you feel that it’s possible to spare patients treatment which isn’t going to do them any good. Does the test using oncotype DX help you decide which of the endocrine therapies to use?

No. At the present time you have no test which can decide if tamoxifen or aromatase inhibitor is better. You usually use aromatase inhibitor because we have been reading international multi-centric trials deciding that an aromatase inhibitor is better than tamoxifen.

But as far as chemotherapy is concerned, it looks now, to you, that it could be quite a help?

Clearly,yes. If you can spare 30% of the patients indication of chemotherapy, we all think it’s of enormous importance.