You are here: Home: BCU 2|2003: Peter Ravdin, MD, PhD

	Peter Ravdin, MD, PhD
Clinical Associate Professor of Oncology, Department of Medicine, Division of Medical Oncology, University of Texas Health Science Center at San Antonio

Edited comments by Dr Ravdin

ATAC trial data update

After 47 months, the disease-free survival data from the ATAC trial look very good. The curves continue to diverge, and there is no sign that they are coming together. It’s like a primary election with 80 percent of the precincts reporting, and a pretty solid lead for a candidate. It’s very likely that in five years, the anastrozole arm will be superior in terms of relapse-free survival. Given its current lead, it’s virtually impossible that it won’t be.

Tamoxifen only reduces mortality by about 30 percent, so there’s a lot of room for improvement. At this time, no mortality data on anastrozole has been presented. It is possible that because of the kind of relapses anastrozole prevents that there won’t be as great an effect on survival as on disease-free survival. However, if it has a large effect on either, it would still be considered beneficial.

ATAC trial data update: Impact on clinical care

Until now, I had not changed my clinical practice based on the early ATAC results. I was waiting to see more data and whether or not the curves were coming together. However, at 47 months, the divergence of the curves shows a three-percent advantage for anastrozole. There will not be three-percent events in either arm over the next year; therefore, the anastrozole advantage will continue to be the same or greater in the next year.

I will now tell patients that there are two options. One option, tamoxifen, seems less efficacious in the short-term, but we know its short- and long-term toxicities. With anastrozole, the time to relapse is substantially improved at the four-year point, but we really don’t have any long-term safety or efficacy data. The FDA did, however, find adequate evidence to allow approval of the drug in the adjuvant setting. There is a risk with either therapy, and some patients will want the new therapy with the potential to be better.

I’m particularly tempted to use anastrozole in patients at increased risk for a thrombotic event. I include age as a risk factor, because, for example, a 65-yearold woman has a one- to two-percent chance of a major thrombotic event over the next five years. Anastrozole would not elevate that event rate; therefore, if the patient has good bone mass, placing her on anastrozole becomes a safety issue, regardless of whether there is improved efficacy.

Anastrozole and bone loss

The updated safety data still shows that anastrozole is better in terms of thrombotic events and endometrial cancer, but worse in terms of fractures. I would like to see more mature safety data.

There is an important carryover toxicity effect that might be expected after the drugs are stopped. For example, tamoxifen has a carryover effect in endometrial cancer risk for five years after therapy. The real question with anastrozole is whether the bone loss patients experience during therapy increases their risk for the rest of their lives.

I’m not half as worried about osteoporosis as I am about thrombotic events with tamoxifen. Osteoporosis is like watching a hurricane come in from Africa — you can prepare for it, predict where it’s going to hit and do something to make yourself less vulnerable to its effects. In patients I start on adjuvant anastrozole, I routinely check baseline bone mineral density to determine which patients need to be monitored soon or started on a bisphosphonate to protect bone mass.

Prevention of bone loss: Bisphosphonates

Bone loss can be managed. Dr Gnant presented a study at San Antonio looking at ovarian ablation with anastrozole versus tamoxifen and bisphosphonates. They saw protection from bone loss by adding zoledronate. In addition, the women who received tamoxifen and ovarian suppression without a bisphosphonate had a drop in bone loss, which was corrected when they got zoledronate.

The data presented by Dr Gnant is important with regard to anastrozole because without agents like zoledronate, osteoporosis would be a major issue. But this study showed that bisphosphonates have the potential to totally prevent the risk of bone loss.

Advantages to a well-powered trial

It is heartening that the number of women in the ATAC trial was greater than the number of postmenopausal, ER-positive women who received five years of tamoxifen in the overview. I applaud the ATAC trial because it is enormously overpowered to determine which drug is better. The beauty of a trial being overpowered is that you can begin to ask subset questions — that’s the power of the overview.

I also like overpowered trials because they are almost impossible to replicate. For example, ATAC may be the only chance to ask what predicts for benefit with tamoxifen or anastrozole, because if survival and toxicity advantages emerge for the anastrozole arm, it may be impossible to do such a trial again and ask the biological questions as corollaries to it. To do it all in one 9,000- patient study is immensely powerful. This trial will never be replicated.

Use of other aromatase inhibitors in the adjuvant setting

I do not use letrozole for adjuvant therapy in the nonprotocol setting. It’s probably equivalent to anastrozole, but I don’t see any significant advantages. If there was a problem with anastrozole, it would have shown up in this study of 9,000 patients, and I would be able to warn my patients or switch them if necessary. With letrozole, I have no way of knowing if there’s an issue.

I have been looking at whether exemestane might have some advantages compared to anastrozole. There will be trials to test this. Exemestane is a very different aromatase inhibitor — it’s irreversible and it has a steroidal structure. Early laboratory evidence suggests it will not be associated with bone loss.

The resistance mechanisms of exemestane might also be different, which could be both better and worse. Remember that tamoxifen can actually be read as an estrogen. I’m curious to see if a drug with a steroid backbone, such as exemestane, might also be interpreted in some systems as an estrogen. Perhaps the same resistance mechanisms that cause resistance to tamoxifen might also cause resistance to exemestane.

Calculating expected benefit from therapy

The informal way to calculate a patient’s expected benefit is to multiply her risk of recurrence by the relative risk reduction expected from a therapy. For example, if a patient has a 60 percent risk of recurrence and your therapy reduces that by 40 percent, multiplying those two percentages together gives a 24 percent benefit.

This is a rough approximation, and it doesn’t quite work that way. In reality, the 40 percent reduction is not a 40 percent composite reduction, but rather a 40 percent reduction that occurs in each year. It compounds like interest in a bank account and actually ends up giving less of an effect than you would expect.

The reason is mathematical, not biological. On a curved surface, you can’t lay your ruler down across the entire curve. Rather, the ruler is essentially touching the curve at any given point and telling you what the slope is at that point. That’s what the hazard is — it tells you what the actual ratios of the slopes would be if you put the ruler at four years on both curves.

I wrote a program on adjuvant therapy that shows how the numerical method for calculating benefit works. It’s accessible online at www.adjuvantonline.com.

Small reductions may offer big benefits

A two percent absolute difference sounds modest, but it can be important. The overview suggests that the proportional benefits hold up for a given therapy, irrespective of the baseline risk. If low-risk patients benefit 20 percent from a given therapy, high-risk patients receive a 20 percent relative benefit as well. So for the low-risk patient, a 20 percent benefit may be only one or two percent. But for the high-risk patient with a 50 percent risk, a 20 percent difference is a 10 percent risk reduction. This therapeutic index gets higher and higher with risk.

I would not expect any therapy to be effective in 100 percent of patients because breast cancer is a very heterogeneous disease. Patients can have ER-positive and -negative disease or HER2-positive and -negative disease. Different cancers express different genes and, therefore, have potentially different vulnerabilities to therapy. I would be amazed if any single therapy was effective in more than 50 percent of patients. Given that, when you see a 20 percent effect, an additional two percent represents perhaps 20 percent of that overall population, which is significant. And perhaps that particular agent benefits 20 percent of the patients who aren’t benefiting from other strategies.

I believe the conquest of breast cancer is not going to be one magic bullet, but rather identifying sets of patients whose cancers have specific vulnerabilities. Maybe none of those sets will be greater than 20 percent of the total, so even the greatest therapies may be very effective only for a small set of patients.

Impact of therapy on early versus late relapses

The divergence of curves with effective adjuvant therapy has not been adequately studied, and I think there is an enormous hidden story there. Some curves begin to diverge within the first year, continue to diverge for the first five years and then parallel each other. Curves like this tell me the therapy is killing the rapidly progressive, early relapsing clones.

The last overview showed that the proportional benefits for chemotherapy emerged entirely because of impact on relapses within the first five years. There was no impact at all on relapse from the average chemotherapy after five years — a fascinating result.

With chemotherapy, we are not yet touching the late, slowly proliferating population, which accounts for perhaps one-third of all relapses, particularly in ER-positive disease. This is where vaccines may be of particular benefit. In contrast, there was a curve for a particular therapy presented at San Antonio that showed no difference in the first five years, but the advantage accumulated in the second five years. The curve suggested the therapy showed no advantage against the rapidly progressive clones in the early relapsers, but that the advantage emerged in the late relapsers.

Hormone therapy is more balanced than chemotherapy in the impact on the second five years. In NSABP P-1 and B-14, the curves actually slightly diverge. The therapy is probably acting on the slower and stalled clones. This has not been adequately studied, and I think it’s worth some additional research.

CALGB 9741 results and the impact on clinical care

I think the results of CALBG 9741 will pressure physicians to introduce the dose-dense regimen. Many clinicians will want to obtain some experience with dose density in their very high-risk patients, who might have received very high-dose chemotherapy with stem-cell support in the past. For those patients, these new dose-dense regimens are extremely interesting.

The problem is that clinicians will want more information about the toxicity profile. There was evidence that the sequential, noncombination dose-dense regimen was less toxic. Many of us are leery about giving very large cumulative doses of doxorubicin, so this will take some soul searching. I don’t think physicians are going to suddenly adopt one of the dose-dense regimens tomorrow morning; rather, I think they will ease into it.

Palliation versus cure for metastatic disease

Physicians generally agree that the targets of treatment for patients with metastatic disease are long-term health maintenance and disease palliation. Today, we don’t talk about intensive therapy for cure. At San Antonio five years ago, there would have been discussions about high-dose chemotherapy with stem-cell support. Now we discuss the breadth of single agents available to support patients. We’re looking to maintain patients’ function for prolonged periods, rather than to induce cure.

Physician acceptance of capecitabine

Kathy Miller presented a case at a seminar recently in which the patient progressed shortly after receiving adjuvant ACT. When the members of the audience were asked what agent they would use next, the most common answer was capecitabine alone. I was a surprised by that because the acceptance of capecitabine was slow in the beginning.

I attribute this slow acceptance to two factors. First, capecitabine was approved at too high a dose; so many physicians had an unfavorable first experience using it. Second, capecitabine is the first drug I can think of that was approved before there were any publications in the literature.

Physician acceptance has grown as lower doses have been tried and patients’ tolerance has improved. In addition, articles have suggested a relatively high response rate with capecitabine as first-line therapy and in combination therapy, particularly with docetaxel.

Capecitabine/docetaxel combination in the adjuvant and metastatic settings

The capecitabine/docetaxel study was important because a very large number of patients consider combination adjuvant therapy. Therefore, the most valuable outcome of this trial in the metastatic setting was learning whether or not this combination might have promise in adjuvant therapy. Patients with metastatic disease in really desperate situations might have a high response rate with the combination, but that’s not the majority of patients. Most patients don’t relapse, and those who do don’t usually find themselves in a desperate situation early on.

If there had been a sequential arm of capecitabine followed by docetaxel, I don’t think we would have seen a great deal of difference between the two arms, as in many other crossover studies.

Select publications

Home · Search