Analysis of thirty years of clinical trial data

Share :
Published: 23 Mar 2012
Views: 5653
Prof Richard Peto - Clinical Trial Service Unit, Oxford University, UK

Prof Richard Peto discusses the analysis of thirty years of treatment data, compiled by the University of Oxford.


Every five years trial data from all over the world was analysed to give a more definite answer on what treatments have the most substantial effects.  The discussion includes analysis of the effects of tamoxifen, chemotherapy and the expense of treatment and screening.

8th European Breast Cancer Conference - March 21-24, Vienna


Analysis of thirty years of clinical trial data


Professor Richard Peto – Clinical Trial Service Unit, Oxford University, UK

Richard Peto, thank you for giving us five minutes ahead of your plenary lecture this afternoon at the European Breast Cancer Conference. You’re going to go back thirty years in this massive work that you’ve undertaken from Oxford, amassing treatment outcome data from all over the world. What are you going to say this afternoon?


Well, for the past thirty years, every five years we’ve asked all the trialists in the world to send us their data then we analyse it. First we showed that simple treatments, hormonal treatment, chemotherapy, had some effect on five year survival then that it affected ten year survival, now that it affects fifteen year survival. And we can be now a lot more definite as to which treatments have the greatest effect and actually some old, out of patent treatments have substantial effects. You do need to test for ER status before using endocrine treatment but this is now possible at low cost. So, for example, our Indian colleagues tell me that it costs about £5 to get a reliable ER assay from an Indian sample. If the sample is ER positive, which in about, say, half of all Indian patients it would be, in about a third of all Chinese patients it would be, then the tumour will respond really excellently to things like Tamoxifen or aromatase inhibitors, both of which are now out of patent. We know that five years of Tamoxifen is more effective than two years of Tamoxifen; if you give five years of Tamoxifen then you can reduce by about one third the probability of dying from your breast cancer.


It’s the first real evidence that targeted therapy works.




People are treating targeted therapy as if it were new, this is thirty years old.


Well Tamoxifen is a 21st century drug that happened to get discovered fifty years too early. It binds to the oestrogen receptor, some breast cancers express the oestrogen receptor, about 80% of those in Britain, about 50% of those in India, and for those that do, Tamoxifen or some other endocrine treatment will reduce by about one third the fifteen year risk of dying from the disease but you do need to keep on with treatment for about five years.


And then the next range of studies that you did?


Also we found that chemotherapy does have some effect on long-term survival but the side effects of chemotherapy are really unpleasant. The drugs used need not be very expensive, you can get quite intensive chemotherapy regimens using out of patent drugs, so CAF would be out of patent drugs. But if you give a standard six courses of CAF, this is really quite nasty therapy and you really do need supportive care in order to give it. You can give less intensive therapy with something like 4AC which in principle requires just four visits to the clinic, it’s less effective than 6CAF, it still needs some supportive care and can produce some quite serious side effects but, again, it does affect long-term survival. With, say, the more intensive treatments you can knock about one third off breast cancer mortality with, say, 6CAF, with 4AC you can knock about one fifth off. But if you’re giving endocrine therapy and chemotherapy then the two add to each other. For a woman with ER negative disease, chemotherapy works, endocrine therapy doesn’t work; for a woman with ER positive disease both work and a combination works better than either alone and the results are there, it’s an effect on fifteen year survival, not just on five year survival. So these things are considered in many countries, even countries with limited medical resources, to be worth considering.


So they should be, because, as you say, they’re affordable and they’re deliverable.


The problem now, increasingly, is the cost of tests that are said to be necessary. And to work out whether endocrine treatment will work or not, all you need is an ER assay, you can do more expensive tests on RNA expression but actually a good ER assay is all that is needed and that’s a low cost test. On chemotherapy it seems as though, say, a good chemotherapy regimen will reduce your risk of death by about a third, no matter what that risk is. And so the absolute gain is greater for the higher risk patients. We need to classify women as higher risk or lower risk. Now, to some extent you can do this by the old things that doctors have been doing for ages – nodal involvement, looking at the tumour down a microscope. There are now tests which might cost thousands of dollars per patient, gene expression tests, they can help predict who is at greater or lower risk, it’s not clear that they predict who will get a reduction of more or less than one-third in whatever risk they have. And the tests now may finish up as ten times the cost of the drug and so if you have limited resources it’s not clear that they’re best invested in doing such tests.


So post the oestrogen/progesterone era we came to the ERB2, the HER2 new test and there you’re beginning to see the same pattern. The test for HER2 is probably not affordable in many important countries in the world but they’re coming down. The cost of the drugs…


The immunohistochemistry test isn’t so expensive, it’s nothing like these RNA expression tests. If you’re going to test for HER2 sensitivity then immunohistochemistry is a lot less expensive and is of comparable sensitivity with gene expression tests. So if you are in circumstances with limited resources then immunohistochemistry tests for oestrogen receptor status and HER2 status and possibly also for one or two other things such as progesterone receptor or Ki-67, but the key ones are the ER, oestrogen receptor, and the HER2 over-expression because these determine sensitivity to the relevant drugs. These aren’t very expensive, they’re not expensive in comparison, say, with the cost of a mastectomy or with the cost of surgery.


Or radiation therapy.


Or radiation therapy, yes. So these would be reasonably cost-effective in resource-limited settings and they do guide whether or not you’d benefit from endocrine therapy, that’s the ER test, whether or not you’d benefit from Herceptin, that’s the HER2 test, but Herceptin is an expensive drug, it’s inherently, at the moment, a very expensive drug so it may well be not affordable but Tamoxifen or an aromatase inhibitor which is out of patent, these are low cost drugs with very low risks in comparison with the benefits for women with hormone sensitive disease. And that’s the conclusion from randomising something like over 200,000 women in trials of Tamoxifen, longer versus shorter Tamoxifen, five years of Tamoxifen versus nothing, one or two years versus nothing, and more than 100,000 women in trials of various chemotherapy regimens.


You must be mature enough to be looking at late effects of some of these interventions. Where do you think that’s going to play out?


Well obviously the concern is the late cardiac effects from anthracycline-based regimens. If you give anthracyclines intensively you can have acute cardiotoxicity but you can also have some later risk of leukaemia and some later risk of cardiotoxicity. At the moment with the regimens that have been tested in general, the absolute gains outweigh what we know of these late hazards but we do need to watch women, a woman in her forties should be watched throughout her fifties, sixties, seventies, and it’s too soon to have done that. But for the moment with something like fifteen year follow up, the gains are outweighing the hazards in terms of anthracycline toxicity.


And trastuzumab hasn’t really influenced that very much yet, although it may.


No, I think that the overview has not yet had anything useful to say about trastuzumab, we will be looking at the trastuzumab trials this time round in collaboration with the investigators. But for the old things, the radiation to conserve breast, if you’re going to do breast conserving surgery then radiation to conserve breast really is necessary otherwise you have a high risk of recurrence and that produces some increased risk of death. Test for oestrogen receptor particularly, if it’s there really encourage long-term compliance and then the pros and cons of chemotherapy, the clinicians know better than we do, but if you can give it then even simple chemotherapy will reduce the ten year risk of death by something like 20% and stronger regimens, which are difficult to give and have serious side effects and have to be managed very carefully, such as six courses of CAF, these will reduce the overall risk of death by about one-third. And if you reduce the risk of death by a third from five years of Tamoxifen, by a third from a strong chemotherapy regimen, then the overall risk is reduced by more than half.


I think there’s a very strong message coming through now that breast cancer is eminently curable, especially early breast cancer. And that’s not percolated through to the public yet.


Some cases of breast cancer are curable; in Britain we have breast cancer rates that were drifting up through the 1950s, 1960s, 1970s and they would have gone up to a risk of nearly 3% of the women in Britain dying from breast cancer before they were 70. Suddenly, because of the improvements in early diagnosis and treatment, that turned around and now the risk is below 1.5%, the probability that a woman will die from breast cancer before 70 is just under 1.5%. So it’s half what it would have been without these improvements in treatment but a lot of these improvements would carry over to at least parts of the low and middle income countries. When I say parts I mean those people who would actually be able to afford mastectomy. Circumstances where a mastectomy can be given, then these drugs should be given, but the very expensive tests are not necessary, they are advertised as being necessary, they are on the back of the programme for the Vienna meeting as being necessary, but a simple collection of four immunohistochemistry stains has been shown by Dansit and Cuzick is about as informative if you’re trying to make treatment choices, and it is a lot cheaper.