DR RAPPAPORT: This is a 41-year-old woman who previously had a hysterectomy and bilateral oophorectomy for menometrorrhagia four years ago at the age of 37. She underwent a radical mastectomy for a 3.5-centimeter, poorly differentiated adenocarcinoma, metastatic to two lymph nodes.

She was treated with FAC times six, and currently is on tamoxifen, which she has taken for three and a half years. The tumor is both ER- and PR-positive and overexpresses HER2 by FISH.

DR LOVE: This case raises the issue as to the optimal long-term endocrine treatment strategy in postmenopausal women. A number of theoretical models have been proposed to estimate whether an aromatase inhibitor initially or tamoxifen followed by an aromatase inhibitor is the optimal schedule (5.1). Dr Chlebowski, could you comment on these models?

DR CHLEBOWSKI: The problem I have with the models is they all assume they know what’s going to happen if you randomly assigned patients at first versus in the middle. I don’t think you can take those numbers in the middle and tack them onto a model. The models are interesting but not very informative.

The side-effect profile is more favorable for aromatase inhibitors than tamoxifen for endometrial cancer, stroke and PE. Bone loss is preventable and treatable.

DR LOVE: Can you discuss the trial reported at ASCO 2005 evaluating anastrozole in patients who had completed five years of tamoxifen?

DR CHLEBOWSKI: The Austrian trial, ABCSG-6a, looked at patients who had received tamoxifen with or without aminoglutethimide for five years. The patients were then randomly assigned to anastrozole or no treatment for three years. This extended adjuvant trial also showed a benefit from anastrozole.

DR LOVE: How did those data compare to what was seen in the MA17 trial with letrozole?

DR CHLEBOWSKI: The data looked about the same. Of course, in the letrozole trial, we talk about five years of letrozole, but it was reported after two and a half years of treatment.

DR LOVE: What you were saying about the models makes sense in terms of the point of randomization and the fact that looking at how relapse rate is affected at two or five years does not take into account all the recurrences and adverse events that occurred before that time.

DR CHLEBOWSKI: To me, that’s a big black box, to make that assumption. The other part of it is, when you look at the assumptions, trying to compare Jack Cuzick’s and Hal Burstein’s model, they get completely different answers.

Burstein used the primary studies’ main endpoints and Cuzick used a little different endpoint, and you get a totally different result, based on what you put in the model.

DR LOVE: Dan, from a clinical point of view, the idea of starting tamoxifen and switching to an AI requires clinicians to tell patients, “Here’s my long-term strategy for you. You’re going to go on a therapy that, for the first two or three years, is going to expose you to a greater risk of relapse, but in the long run, you’ll have fewer relapses.”

DR HAYES: In my own practice, I believe that for patients who have a reasonably good prognosis, tamoxifen alone is a perfectly adequate start. I actually disagree a bit with Rowan in that I’m not sure we know about long-term follow-up of women who have complete estrogen depletion. It took us a long time to see some of the unanticipated effects of tamoxifen.

Tamoxifen is a good drug. It’s not like we should throw it out. In patients who start out with a good prognosis, the maximum potential added benefit of an AI over tamoxifen can’t be more than one or two percent absolute survival over 10 years or so. Given what we know about tamoxifen and the benefits of it, I have started with that.

On the other hand, in this patient with positive nodes, I would consider starting with an aromatase inhibitor initially. She has a high risk of relapse up front, so the potential added proportional reduction in recurrences and death by using an AI is amplified. In addition, if you want to bring Kent’s biology into it, she’s HER2-positive, and one would believe that estrogen depletion should be a more effective strategy than a SERM in such a patient.

DR LOVE: Would you describe the type of patient you would treat with tamoxifen up front and discuss the issue of excess toxicities, such as endometrial cancer and stroke?

DR HAYES: First of all, let’s be clear that we’re talking about postmenopausal women only, not premenopausal patients. A postmenopausal woman with ER-positive, node-negative disease, especially if her tumor is two or three centimeters or less, I believe has a very good prognosis, and that’s the type of patient in whom I would recommend tamoxifen. I tell my patients that there’s no question in my mind that the AIs are more effective than tamoxifen, although the amplification of that benefit in a very good-prognosis patient is very small in terms of the added proportional reduction over tamoxifen.

With an AI, bone fractures are a risk. Granted, the fracture curves look like they are coming together, but you have to recall that there are precious few patients out in that tail beyond five years (5.2). That’s a very unstable statistical tail. Most of those patients are to the left of that, between three and four years.

DR LOVE: Chuck, where are you on the debate of tamoxifen versus aromatase inhibitors?

DR VOGEL: I’ve listened to everybody, and I actually agree with everything that’s been said. We worry about the cost issues, and there’s no question that I treat some patients with tamoxifen solely on the basis of cost.

I think that looking at the overall results, the aromatase inhibitors have a better toxicity profile, except for two areas, and they probably have slightly more efficacy. On the other hand, they are more costly.

So, in practice, I discuss all of these things with my patients. I have no problem if a patient elects to receive tamoxifen because of cost or whatever. It’s a very good drug. But, cost notwithstanding — and with the proviso that I don’t like starting the osteoporotic patients on aromatase inhibitors — I start with an aromatase inhibitor. I switch from tamoxifen to an aromatase inhibitor at two years, and I switch to letrozole at five years.

DR RAVDIN: I absolutely agree with the ASCO Technology Assessment statement that basically, for postmenopausal women with ER-positive disease, an aromatase inhibitor should be part of their adjuvant hormonal therapy (Winer 2005). I think it’s quite clear that these agents reduce the risk of relapse. Every single one of the major trials, all five of them, show marked reduction in the risk of relapse with aromatase inhibitors.

DR CHLEBOWSKI: Looking at the toxicity profile and the fact that you’re getting more recurrences if you start with tamoxifen first, I would rather start with an aromatase inhibitor and have fewer relapses, because those are not recoverable in a major sense.

I think it’s possible that shorter durations of hormone therapies will be better. We had more signals on that at ASCO this year with the two- versus five-year tamoxifen randomized trial showing absolutely no difference in survival after 10 years of follow-up (Valentini 2005).

DR RAVDIN: The clinical data are what drive things. And the same way that nobody could anticipate what the combined arm of ATAC was going to show, we don’t know what will be revealed with these switching strategies.

DR CHLEBOWSKI: I agree that the empirical data will rule, but what happens is you’re going to be behind for sure in terms of recurrences if you start with tamoxifen first, and in many cases that’s a terminal disease that you can’t recover from. I think you’ll have time to see what the empirical data will be for the majority of your patients because we’ll get some kind of signal from BIG FEMTA. We’ll have that result in probably three years.

Select publications