I N T E R V I E W  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.

Select Excerpts from the Interview*
* Conducted on May 20, 2005

 Tracks 2-3

DR LOVE: Can you discuss the Women’s Intervention Nutrition Study (WINS) that you presented at ASCO?

DR CHLEBOWSKI: The issue of dietary fat intake and breast cancer has been around for about 25 years. So to address this issue, we conducted a randomized clinical trial. We entered 2,437 women aged 48 to 79 from 37 clinical centers in the United States. They all received standard breast cancer management, including surgery, radiation therapy if indicated, tamoxifen for five years if estrogen receptor-positive and a defined chemotherapy if estrogen receptornegative. The estrogen receptor-positive patients could also receive chemotherapy.

DR LOVE: What was the rationale for looking at nutrition and breast cancer recurrence?

DR CHLEBOWSKI: The whole issue probably began with the country-to-country differences in breast cancer incidence. Twenty-five years ago, the Japanese not only had fewer cancers but they had a much lower recurrence rate, and at that time, the difference in obesity between American and Japanese women wasn’t that great.

There were a number of studies that looked at cohorts and suggested that dietary fat intake would affect recurrence. More recently, the attention in the observational studies has shifted away from dietary fat and more towards obesity and physical activity.

DR LOVE: When you look at those components — dietary fat, obesity and physical activity — what would be the mechanism of action as to why it would affect breast cancer recurrence?

DR CHLEBOWSKI: The prevailing thought was that obesity or dietary fat could be related to estrogen levels, and they can show that association. Of course, that would make it much less interesting if that were the only mechanism because now we have aromatase inhibitors.

One always had the consideration that it is more like the metabolic syndrome where you end up having obesity, insulin resistance associated with coronary heart disease, dementia, diabetes and the cancers — colon and breast cancer — at least in some of those studies. That kind of mechanism, metabolic syndrome, insulin regulatory pathways, had also been under consideration but not the primary focus.

DR LOVE: What were the endpoints of the study?

DR CHLEBOWSKI: Our primary study endpoint was relapse-free survival, which included all breast cancer recurrence sites, including contralateral breast cancers. We found that the dietary group had a longer relapse-free survival than the control population. 12.4 percent of the control group had a relapse compared to 9.8 in the diet group, which was a 2.6 percent absolute difference at five years or a 24 percent reduction in risk of recurrence.

DR LOVE: Do you think this is now something that should be brought up to women with breast cancer?

DR CHLEBOWSKI: We’re not quite there yet in that we recognize the need for further follow-up, a peer-review publication and probably a confirmatory study. Having said that, the diet was associated with nutritional adequacy and can be recommended for other health reasons and would also have no appreciable side effects.

The only issue that I could see is this guilt factor. For instance, the American Cancer Society has recommendations for dietary change after cancer diagnosis. They’re all based on inference. Now we have a little signal that there might be a cancer benefit as well, but I don’t think we would tell every patient that they have to do it. If somebody wanted to do something, you’d say, “Here’s something that you could do. It may inf luence your breast cancer, but it also has other potential benefits.”

Track 4

DR LOVE: Can you update us on the Women’s Health Initiative (WHI) trials and summarize some of the most important findings that have come out in the last couple of years, particularly related to breast cancer?

DR CHLEBOWSKI: There are two key areas. First would be the WHI hormone trials, and the initial data reported was the comparison of estrogen plus progestin versus placebo in women who had a uterus (Rossouw 2002). Basically, the surprising finding was that coronary heart disease was increased by approximately 25 to 30 percent, as opposed to the prestudy estimates that anticipated it would be reduced with hormone use. Breast cancers were increased, as expected, but surprisingly, prognostic characteristics worsened. After one year of estrogen plus progestin use, abnormal mammograms increased by 74 percent.

These were some of the major findings. Basically, what happened with that was the FDA changed the labeling of estrogen plus progestin, and 33 million fewer prescriptions for menopausal hormone therapy were written in 2003 versus 2001.

Also, the trial of estrogen only versus placebo for otherwise healthy women who had a prior hysterectomy was stopped early. In that study, there was no effect on the global health index — the overall balance of risks and benefits (Anderson 2004). Actually, coronary heart disease didn’t increase. Rather, there was approximately a nine percent decrease, but it was not statistically significant.

The other very interesting finding was that estrogen-only therapy ended up trending towards a decrease in breast cancers. There were approximately 24 percent fewer breast cancers on the estrogen-only arm. We’re in the process of doing further analyses, and I believe this may well represent a real event. Short-term estrogen may well be associated with a reduction in breast cancer risk.

 Track 5

DR LOVE: Can you summarize where we are in terms of the major randomized trials of aromatase inhibitors?

DR CHLEBOWSKI: We have 68 months’ follow-up on the ATAC data, which means the majority of patients have completed their therapy and approximately a year of follow-up afterwards (Howell 2005). I see the toxicity data as being final, unless one proposes a new mechanism whereby returning estrogen back to the physiological levels causes different toxicities. The five-year toxicity data is very favorable for anastrozole compared to tamoxifen, because the three life-threatening toxicities — endometrial cancer, arterial and venous vascular events — were all significantly less with anastrozole.

DR LOVE: While endometrial cancer is certainly a disturbing event, I guess you could raise the question as to whether it is really life threatening.

DR CHLEBOWSKI: Endometrial cancer has a 15 percent mortality rate associated with it. When we look at the hip fractures, which are life threatening, the incidence was low and really not different at all between the patients on anastrozole versus tamoxifen. Tony Howell just reported on the cardiac deaths data, which were 49 versus 46 — really no difference at all after five years.

 Track 13

DR LOVE: Putting aside toxicities, there has been a lot of discussion about whether the long-term — 10-, 15-, 20-year — relapse rate in some patients would be lower starting with tamoxifen for some period of time, followed by an aromatase inhibitor. What are your thoughts on that hypothesis?

DR CHLEBOWSKI: If you start with tamoxifen, after two and a half, three or five years, more patients will have relapsed than on an aromatase inhibitor (1.1). A substantial number of those patients will be irretrievable — they have incurable disease — and so you’re banking on the fact that you’ll be able to capture more patients later, but we don’t have any data for that. That’s just speculation.

While I believe sequencing therapy may be better, ultimately, I still don’t see any reason not to start with the most effective therapy. An aromatase inhibitor followed by tamoxifen or a nonsteroidal aromatase inhibitor makes more sense to me. We have to wait to see the data from the BIG FEMTA trial, which includes an arm with letrozole as initial treatment followed by tamoxifen.

Track 16

DR LOVE: Can you discuss the issue of race, ethnicity and breast cancer subtypes?

DR CHLEBOWSKI: In the Women’s Health Initiative, we have a large population, so we examined risk factors for breast cancer by ethnicity (Chlebowski 2005). We took 160,000 women, in whom we have a racial ethnic distribution, and tracked them as a prospective cohort. As expected, we found the ageadjusted breast cancer risk for African-Americans, Hispanics and Asian Pacific Islanders, compared to Caucasians, were all substantially less — 30 percent or so less (1.2), which is about the same as the SEER data.

We corrected for the Gail risk model, and their hazard ratios moved towards unity but were still different. We looked at approximately 22 risk factors, so we really have extensive risk-factor correction, and what we found was very interesting. We could almost completely explain the reduced risk of breast cancer in Hispanics and Asian Pacific Islanders. Their hazard ratios, with our risk factors in the final models, were 0.98 and 0.94.

DR LOVE: What about the African-American population?

DR CHLEBOWSKI: Interestingly, for African-Americans, their hazard ratio compared to Caucasians was 0.75, which statistically was significantly lower. Even more interestingly, they had substantially lower risk for hormone receptor-positive cancers — about 50 percent of the Caucasians’ risk.

The most interesting finding was our equivalent of triple-negative cancers — ER-negative, PR-negative and high grade. African-Americans have nearly a fivefold-increased risk of having triple-negative cancers compared to Caucasians. In African-Americans, 31 percent of the cancers were triple negatives compared to 10 percent in Caucasians.

Other groups are looking at whether the triple negatives have genetic profiles similar to basaloid cancers, which are faster growing and more difficult to treat. We consider this kind of a unifying hypothesis in that, with these more common basaloid cancers, with the same mammographic screening or frequency, you’re going to find higher-stage cancers, and with the same therapy that was given in the cooperative groups, they will have a worse outcome.

It’s interesting that in several African-American populations, the triple negatives are close to 50 or 60 percent of the population, instead of a third. Studies are now underway looking at the genetic profiles of these mixed African-American populations compared to Caucasian populations.

SELECT PUBLICATIONS