PRESENTATION Rowan T Chlebowski, MD, PhD Dr Chlebowski is a Professor of Medicine at UCLA School of Medicine and Chief of Medical Oncology at Harbor-UCLA Medical Center in Torrance, California.

Slide 1

I am going to talk a little about the WINS trial that was reported at ASCO in a plenary session. Women in this study were 48 to 79 years of age, mostly postmenopausal with early-stage breast cancer, Stage I to IIIA, and had received primary surgery, radiation therapy and conventional systemic therapy. Patients with receptorpositive tumors received tamoxifen, while those with receptor-negative tumors received chemotherapy — half anthracycline based and half nonanthracycline based. The women on tamoxifen could have elected to receive chemotherapy as well. The patients were randomized 60:40, with fewer to the dietary intervention group, so we would have more resources to allocate to the dietary intervention group or control.

Slide 2

Randomization took place about 220 days after initial surgery, so patients were entered after they completed their primary therapy, while they were receiving hormonal therapy. The dietary intervention was done at 37 centers around the United States.

The diet group was given a dietary fat gram goal by centrally trained, registered dieticians, implementing a predefined low-fat eating plan. The dieticians were trained in behavioral intervention techniques. Patients received eight biweekly individual counseling sessions then one session every three months. There was no counseling towards weight reduction.

This was more of a switching trial than a weight-reduction trial. There were monthly group sessions, and patients self-monitored their fat intake. We captured information about their diet with telephone re-calls. The control group saw the dieticians every three months and talked about nutritional adequacy.

Fat gram intake for this group went from about 56 to 33 fat grams per day — about a 40 percent reduction in daily fat gram intake — which was sustained by most of the individuals.

There was a four-pound weight loss, which wasn’t much but still was three standard deviations different. The weight loss at least signals that some dietary change had occurred.

Slide 3

Here is our primary study endpoint of WINS relapse-free survival. This is an interim result, as the follow-up is continuing. As you can see, the hazard ratio was 0.76 with a p-value of 0.034 for Cox proportional hazard, and there was a three percent difference at five years.

We did a subgroup analysis by receptor status. The hazard ratio for relapse-free survival for patients with estrogen receptor-positive tumors was 0.85 and not significant. In the 478 patients with ER-negative disease, there was a hazard ratio of 0.58, with a 42 percent reduction in risk and eight percent absolute difference at five years. This is hypothesis generating but very intriguing to us, as the curves just break apart initially.

Slide 4

These are data from the Early Breast Cancer Trialists’ Collaborative Group, which is now published and available online. You can see that there was a 14 or 15 percent reduction in recurrence risk with tamoxifen. To the right, this is mortality risk going forward. That is why tamoxifen has been the standard therapy for all these years.

Slide 5

We also know that there are life-threatening toxicities for tamoxifen — endometrial cancer, pulmonary embolus, stroke. It took us 20 years to figure out that tamoxifen had an endometrial cancer risk. But we also should say that if we did 32 trials of tamoxifen versus nothing, it still took us 15 years to figure out that tamoxifen worked. So I think that if a trial has 3,000 or 4,000 patients in each arm, we’ll get to a toxicity issue sooner.

Slide 6

You’ve seen this before; the only point I want to make about the ATAC data is that they’re at 68 months. That means that almost a year of follow-up has occurred for people who have stopped their aromatase inhibitors.

So when we look at the toxicity of the ATAC trial, this is what it will be. There aren’t any studies for more than five years of aromatase inhibitors. People say, “I’m worried about long-term use of aromatase inhibitors.” Well, this should be the toxicity that you’re going to see, because it’s five years of toxicity.

Slide 7

These are the updated data on side effects. At the 68-month follow-up, half the patients have been off therapy for nearly a year, and we saw the expected reduction in all three life-threatening toxicities of endometrial cancer, stroke and venous thromboembolic disease. You can see all of them were significant.

Slide 8

There are 3,000 patients in each arm. Cardiac deaths were 49 versus 46 — so there wasn’t much of a signal for myocardial infarction issues with anastrozole — at least with the data from the 68- month follow-up from the ATAC trial.

Slide 9

The other point I want to make is that many oncologists have a lot of concern regarding bones. I think it’s going to be not only a preventable, treatable situation but also something that is likely to go away completely in the near future. There is no difference in hip fractures after 68 months with anastrozole and tamoxifen. You can see that hip fracture is one percent — a tenth of a percent per year. This is for a group of patients who had no prescreening when they entered the study and no ongoing protocol-defined follow-up for bone. These are high numbers. If you’re going to actually do any screening, any treating, you’re going to have lower numbers than that. The hip fractures are the ones that are associated with a survival detriment.

Slide 10

Here are the fracture rates from the update of the ATAC data presented by Tony Howell. You can see that at the five-and-a-halfyear point, the fracture rates look like they’re converging. For hormonal therapy, we know that when you stop estrogen/progestin or estrogen alone, bone loss is accelerated. It’s two to three times the rate of loss in normal menopause. Will we get accelerated bone recovery when we add back the hormones? We’ll know that soon from the bone density studies that are ongoing. There may not be long-term consequences of five years of aromatase inhibition.

Slide 11

The other issue is cognition. A few years ago, we thought that preclinical observational studies suggested that exogenous estrogens and lowering of progestins would be important in maintaining cognition. There were trials started in patients with dementia. The Women’s Health Initiative studies that I’ve been involved with have totally changed that concept.

In two randomized, placebo-controlled clinical trials, 16,000 otherwise healthy women received estrogen plus progestin therapy for five years, and the dementia was only two and a half years, but there was a 40 percent increase in stroke and doubling of dementia risk in a subset of patients 65 years of age or older who received estrogen/ progestin versus placebo for just two and a half years. In the estrogen-alone trial, there was a 40 percent increase in stroke and a 50 percent increase in dementia. A preplanned combined analysis had a statistically significant 76 percent increase in dementia. So the old idea is that exogenous estrogens maintain cognition. The new concept is that anything that causes arterial vascular events — like estrogen, estrogen plus progestin, maybe tamoxifen — will be likely to increase not only stroke but also decrease brain function and increase the risk of dementia. We don’t know what aromatase inhibitors will do to cognition, but the testing has to be two tailed, because one could equally say that it might have favorable impacts on cognition.

Slide 12

The BIG 1-98 study is our only switching study. We probably won’t get results from it for several years, and they may not be definitive. These are big studies, with 8,000 patients, but the switching parts — tamoxifen/ letrozole, letrozole/ tamoxifen — are half that size. There are 1,250 patients in each of those arms, numbers that would be small for switching.

Slide 13

Here are data similar to the ATAC data — 13.6 percent rate of relapse with tamoxifen versus 10.2 percent with letrozole — about the same difference after this relatively short period of follow-up.

Slide 14

Here are the data for deaths, cardiac — 26 with letrozole versus 13 with tamoxifen. You didn’t see anything at all like this in the MA17 trial, comparing letrozole versus placebo after tamoxifen — there was absolutely no signal of cardiac issue there.

It was alluded to that the population is different, but it actually isn’t that different. It is 15 percent Eastern European. Almost all the rest, 85 percent, is Western European, but there is no US, and the UK is underrepresented compared to the other trials.

The population is mostly Western European, so there was no great explanation for that difference.

Slide 15

Now we’ll go to the switching trials. This is the IES trial, which is exemestane versus tamoxifen after two to three years of tamoxifen.

Slide 16

You can see with this recent update that there was a 27 percent reduction in recurrence risk. And there was a trend towards a survival benefit, which we didn’t see in the ATAC trial. We’ll get into that later in terms of what that would mean when you’re switching in the middle and randomizing in the middle versus randomizing at the start. So there was a 20 to 27 percent reduction in risk of recurrence and a trend toward survival — but there were more myocardial infarctions on exemestane.

Slide 17

These were small numbers — 20 with exemestane versus eight with tamoxifen. Also, there were more vascular deaths — 15 versus seven. Is this a signal? There is much interest in this area. As the bone health goes away, this surfaces— but you could say these are very small numbers, with these relatively small differences, when you’re in trials involving thousands of patients.

Slide 18

The Austrian Breast Cancer Study Group 8 and ARNO combined trials with women on two years of tamoxifen. These were all postmenopausal women with earlystage breast cancer, almost all receptor-positive, who were randomly assigned to tamoxifen or anastrozole.

Slide 19

With that switching in the middle, at four standard deviations, there was a 40 percent reduction in recurrence risk. This trial of switching in the middle also showed a trend towards a survival difference.

Slide 20

There was a 42 percent reduction in risk of recurrence in the MA17 trial, with switching at the end.

Slide 21

Look at the toxicity data. For cardiovascular events, it’s six percent with letrozole versus six percent with placebo — no difference. Hypercholesterolemia: 16 percent versus 16 percent — no difference. What we have at the end is this interesting dichotomy between the BIG FEMTA trial with Western European patients, with randomization at first, and the MA17 trial with Western European, US and UK patients having no difference. That is an issue that has not been completely addressed.

Slide 22

I think the bone health issue is going away. At 33 months, you can see numerically more fractures on ATAC, IES and MA17, but for all these trials there was no prescreening or ongoing therapy for bones. One can calculate that 70 percent of these fractures occurred in women who had preexisting osteoporosis when they entered the trial. You should never get these numbers if you’re doing screening and intervention.

Slide 23

Just to show you that bisphosphonates work — this is ABCSG 12. Interestingly, they take premenopausal women with receptorpositive disease, make them postmenopausal with goserelin acetate and then give them anastrozole versus tamoxifen. So these women are made postmenopausal very suddenly.

Slide 24

What happens when you give them either zoledronic acid or not? If you don’t give them zoledronic acid, you can see that tamoxifen wasn’t able to fully prevent this rapid loss of bone. There was more loss with anastrozole, which was completely abrogated by zoledronic acid. Four milligrams of zoledronic acid every six months completely abrogated bone loss. The Z-FAST trial had about the same results.

Slide 25

The ASCO bone health guideline, which I helped write, was published in 2003. We’re in the process of updating it now, saying that every woman should obtain a bone mineral density reading; that we should follow or treat osteopenia as an option, treat women with osteoporosis and repeat bone mineral density tests annually.

We think that probably only about 20 percent of women then would need a bisphosphonate while on aromatase inhibitors, but we’ll have to see if that’s the case or not.

Slide 26

We’re going to finish by looking at the difference in these study designs. With the ATAC and BIG FEMTA trials of tamoxifen versus aromatase inhibitors, note that the randomization is at the start. With the switching trials, randomization is after two or three years. That means the women who progressed in the first two years, who had life-threatening toxicity of any kind, who had a myocardial infarction or couldn’t take the pills aren’t here. So we don’t know what the results would be if they randomized at the start. The only true sequencing trial we have is BIG FEMTA, and its results are two or three years away. You can see that the extended adjuvant trial is very much like a switching trial, because you’re giving this and then the randomization occurs. Women who relapse on tamoxifen, et cetera, won’t be around for the second randomization.

Slide 27

This effect has been alluded to in models. The problem that I have with the models in terms of trying to figure out what’s going on is they all assume they know what’s going to happen if you would have done the randomization at first versus the randomization in the middle. I don’t think you can take those numbers in the middle and tack them onto a model, because you just don’t know what happened in the middle. The models are interesting, but not very informative.

Slide 28

In conclusion, these are the rationales for using AIs: Efficacy against early recurrence peaks, side effects defined in five years. If you’re worried about the long-term effects of an aromatase inhibitor, I think we’ve seen them, unless you want to come up with a hypothesis of why, after estrogen comes back, you expect to see further side effects.

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

These are the rationales for using tamoxifen: Long experience with its use, concern over side effects with bone and coronary heart disease with aromatase inhibitors, no survival difference in these studies and cost associated with recurrence.

In the model analyses, all the models say they know what happens, and I don’t think anybody does. Reasonable oncologists can disagree.

* Presented at Research To Practice Breast Cancer Update CME Forum, Los Angeles, California, May 21, 2005. The enclosed graphics from the presentation are included in the PowerPoint files on CD 3. Please see page 1 for additional instructions.

SELECT PUBLICATIONS