ASCO 2011 Annual Meeting, 3—7 June 2011, Chicago

Exemestane significantly reduces risk of invasive breast cancer in high-risk, postmenopausal women

Dr Paul Goss – Dana-Farber/Harvard Cancer Centre, Massachusetts General Hospital, Boston, USA

I’m here to present the results of an international breast cancer prevention trial, which is different from cancer treatment of course, on behalf of a very distinguished group of investigators in Canada, the United States, France and Spain, that collaborated with us over a seven year period between 2004 and today to conduct a clinical trial of a pill once a day, called an aromatase inhibitor, which will be familiar to some of your viewers, versus a placebo. The background to this trial was that tamoxifen, made in the UK, was the proof of principle that antagonising oestrogen prevents the growth of breast cancer. Tamoxifen worked its way from advanced end-stage disease all the way to early stage breast cancer, into breast cancer prevention and finally into FDA approval as a preventative for breast cancer. That happened in the late ‘90s; some of the seminal work was also done in the United Kingdom and Western European countries.  This was an exciting observation because a very substantial reduction in the incidence of invasive breast cancer, which is the commonest cancer in all women worldwide, there are 1.5 million women a year getting breast cancer, about 500,000 dying. It was a very important finding and the FDA set a guideline for this, it was not accepted or approved in other countries, to my knowledge, including the UK, and there was resistance to it from the get go, but the FDA said, “Fine, in certain women you can prescribe it.” And yet even despite that approval, recent data says only 4% of women ever took it. There are multiple reasons in the literature for that, one of which is clearly the side effect profile of tamoxifen. It, as you know, causes rare, serious life-threatening toxicities – endometrial cancer, blood clots, strokes and other things. When selling a breast cancer prevention idea to women, when it actually causes another cancer potentially, is a tough sell and a difficult debate. As I said, that’s one of the reasons.

Now the aromatase inhibitors were invented, again, in the UK, had the ability to more profoundly suppress oestrogen, didn’t have this mixed effect of tamoxifen, and, a long story cut short, came out over a decade and a half to be superior to tamoxifen and got approved worldwide, including the UK, as the treatment of choice for early stage breast cancer patients that would benefit from this medication. These were all post-menopausal women because that’s how this pill works, not in pre-menopausal women. During the course of those trials we gained the same evidence that tamoxifen had – new cancers in the opposite breast were profoundly shut down, the incidence was dramatically lower and better than tamoxifen and the side effect profile looked safer, with the one exception that there were differences in certain key ways, and I’ll come to the bone metabolism one in a minute, that were really important. But no life-threatening toxicities, no causes of treatment related deaths ever imaginable from the data that was collected around the world. So it was very logical to launch a trial with one of these drugs. And we used a placebo because it’s the most powerful experimental design, smallest experiment, cleanest toxicity profile, absolute proof of principle, but we counselled every woman who joined the trial about tamoxifen’s benefits and raloxifene. The literature said at the time, and the guidelines in the United States, that placebo controlled trials should be conducted because tamoxifen had not shown the net benefit that people wanted and it wasn’t being used anyway, so we couldn’t consider it a standard of care, despite being approved.

Our experiment lasted for seven years and we enrolled 4,500 women. Our goal was to not do any interim analysis but to reach 38 invasive cancers, it sounds like a terribly small number, but 38 invasive cancers should tell us absolutely and statistically certainly whether we achieved our goal or not with certainty rather than by chance. At this meeting we present the fact that we did. So the first and foremost goal of this trial, and only major goal, was to show that this drug could do that and it did. It reduced invasive breast cancer by extraordinarily, exactly 65%, it’s almost as if we were clairvoyant.

What was more surprising to us is we looked, of course, at the efficacy of the drug in other ways in terms of breast events. So it also prevented pre-invasive duct carcinoma in situ and it prevented these lesions that women often experience with mammography and MRI growing as well, benign breast disease but that has a troublesome harbinger of future cancer like hyperplastic lesions, atypical hypoplasia of the duct, lobular hypoplasia, LCIS. So all kinds of breast lesions were diminished, as they were with tamoxifen and that’s a good thing because that predicts that as you go further in follow up, you’re likely to have fewer and fewer invasive cancers in the future, even though you’ve used the drug for a finite period of time. That indeed was shown for tamoxifen - during the five years it has a good reduction but in the five years after that it has an even better reduction. So the efficacy was an important and successful endpoint.

We then looked at symptoms that aren’t dangerous but troublesome and we looked at serious potential organ toxicities that could be dangerous, even if they were asymptomatic. We looked, importantly, at 4,500 women, we made them give back their self-reported quality of life in many, many ways, a very detailed assessment of 4,500 women. We did find the symptoms we predicted which are aggravation, in some women, of menopausal symptoms or the occurrence of menopause for the first time, but none of them were deal breakers. In fact, some women did stop the medication because of that and we think that was a good thing. But we did not find any alarming new symptoms that hadn’t been anticipated. On the flip side, this was the extraordinary part, we found no untoward end organ effects that we were worried about. There were no increases in osteoporosis, no increases of clinical fractures, no second malignancies, no cardiovascular events of any sort whatsoever, and no treatment related death. It looked very clean and the quality of health that women reported back to us, quality of life, was that while they did notice through those sensitive instruments, quality of life effects because of hot flushes and waking at night and sweating, etc, they were minimal and overall the quality of life report was excellent.

What were the criteria for patient inclusion?

We chose women that mimicked the inclusion criteria of the very first seminal trial in the United States, which was P1, the tamoxifen trial. Women had to be post-menopausal, so that was different; they could be 35 years of age or over, because some women are menopausal earlier than others; they had to have one more risk factor at least, or more. Age of 60 alone allowed them into the trial; precursor lesions that we’ve discussed; ductal carcinoma started treated with mastectomy and a Gale score of at least 1.66 or above, the FDA threshold for tamoxifen. Gale is an interesting score because it takes into account all the reproductive history and all the risk factors that we know about – previous benign disease, family history, all kinds of things. So it gives a score, it’s an on-line tool that women use and doctors, and we set that at 1.66 or above. So that was the patient population.

What’s possibly more important, though, is that the actual cancer event rates were similar in our trial to other trials, despite the slight variation in inclusion criteria. But at the end of the day, if you said to whom does this apply, the first experimental answer, the correct answer, for any experiment which is successful, is those best fitting this result are those that could have gone into this trial, because that’s exactly who this result benefitted.

Where do you see the potential for replacing tamoxifen with exemestane?

Not in pre-menopausal women; not in BRCA1 or 2 carriers, which we haven’t tested in this trial and perhaps not for some women who would chose that route anyway. They will rival tamoxifen in price, they’ll become generic and the cost will fall, but it’s hard to imagine that any woman who wants to have breast cancer prevention who is menopausal would select tamoxifen above this, but I don’t want to give the impression that we’ve compared ourselves to tamoxifen or that tamoxifen is a valueless thing. As a breast cancer doctor, I wish more people had used tamoxifen and if this result only increased the debate to make more doctors use tamoxifen I’d be happy, but I think there’s a better choice, certainly for menopausal women it’s a better choice.