It is being hailed as the biggest break-through in the treatment of breast cancer in years. An anti-estrogen drug called Femara (letrozole) has been found to decrease the risk of recurrences in post-menopausal women who have been treated for early-stage hormone-sensitive breast cancer. Leaders of an international clinical trial called off the trial when analysts found a big difference in responses between the two 'arms' of the study. The women who were already on Femara were advised to continue taking the drug, while those who had been given an inert placebo were offered the chance to start on Femara.
In the wake of this finding, most leaders of the cancer field have agreed that all post-menopausal women with a history of treated early-stage hormone-sensitive breast cancer should take this new drug after completing five years of treatment with the standard anti-hormonal drug, tamoxifen. "If you fit the profile it's very, very clear," said James Doroshow, MD, chairman of the division of medical oncology and therapeutics at City of Hope Cancer Center in Duarte, California (Gardner 2003).
The decision could affect 100,000 women per year in the US alone, and hundreds of thousands more around the world. The study was greeted with a barrage of enthusiasm the likes of which has not been seen in a quite some time. "This is up with the biggies, it is up with the major majors," enthused Larry Norton, MD, chief of medical oncology at Memorial Sloan-Kettering Cancer Center in New York and one of the study authors (Kolata 2003).
"I think it's exciting," said Dr. C. Kent Osborne, director of the breast center at Baylor College of Medicine and the Methodist Hospital in Houston (ibid.).
Dr. Harold J. Burstein of the Dana-Farber Cancer Institute, Boston, described the findings as an "unexpected and robust difference in the rate of events" (ibid.)
"Everyone thought letrozole would have a modest effect," said Dr. James N. Ingle, the principal investigator for the United States in the study. "Everyone was surprised" (ibid.)
The US government was quick to embrace the study. It is a "very important advance in breast cancer treatment," which will "improve the outlook for many thousands of women," according to Andrew von Eschenbach, MD, director of the National Cancer Institute, which led the study in the United States. The study was featured prominently at the government's website, www.cancer.gov
One always welcomes useful new treatments, regardless of their origin. But amidst the outpouring of celebratory news stories, I wish to interject a few words of caution. Let us see if we can figure out what exactly Femara did--and did not do--for the women who were enrolled in this important clinical trial over a four-year period.
Femara Reduces Recurrences
As stated, the 5,187 women who took part in this Canadian-led trial were all post-menopausal, and had been treated for early-stage breast cancer. All had already taken the standard drug tamoxifen for a period of 4.5 to 6 years following their initial treatment. It has been found that the protective effect of taking tamoxifen wanes over time, levelling off after about five years. Thus, the essential purpose of the present study was to find out if a new drug, Femara, could be given as a long-term therapy in place of tamoxifen after the latter had begun to lose its effectiveness.
And, make no mistake, the study did indeed show that there was a significant reduction in the risk of breast cancer recurrences in women who received Femara compared to those who received only an inert placebo pill. In total, during the four-year period of the study, 132 women taking the placebo experienced recurrences compared to just 75 who were taking Femara. Understandably, press reports have focused on the fact that Femara significantly reduced the occurrence of new tumors, which is true. However, we need to be quite clear on our terms of reference here. Femara reduces the risk of recurrence by 43%, but this is not the same thing as reducing deaths by 43%.
Put another way, Femara was shown to increase the period of "disease-free survival," or the length of time before the disease recurs, during which period the patient shows no evidence of disease (NED). In real terms the study demonstrated that after an average of four years 13% of women on the placebo, but only 7% of those on Femara, had suffered a recurrence.
Impact on Deaths
"Our study ushers in a new era of hope by cutting these ongoing recurrences and deaths from breast cancer after tamoxifen by almost one half," said the lead author, Paul E. Goss, MD. The reader can see that this statement is technically true, but its significance is easily inflated.
In this study, there were 2,594 women in the control group and 2,593 women in the Femara group. Over the four-year period in question, 17 women taking the placebo pills died of breast cancer compared to 9 taking Femara, for a total of 26. Thus, there was an overall reduction of 8 deaths from breast cancer. (For an excellent discussion of the different types of statistics that are used in evaluating the value of a treatment: http://www.annieappleseedproject.org/relrisverabr.html)
I read these figures differently from most commentators. What the figures say to me is that, for postmenopausal women, the risk of dying from early-stage breast cancer after undergoing conventional therapy is very small to start with. Even in the placebo group the death rate was under one percent over the duration of the study. Stated this way the results seem somewhat less spectacular than suggested by Dr. Goss's dramatic claim that the death rate had been cut in half.
That the overall death rate should be so low may initially seem surprising: we are accustomed, after all, to thinking of breast cancer as a relentlessly deadly disease. However, early-stage, postmenopausal, estrogen-receptor-positive breast cancer is generally not as aggressive a disease as many people fear. Naturally, after a first bout one needs to remain vigilant for any sign of a recurrence. But if a tumor does recur, either in the same or in the opposite breast, it can usually be adequately treated with conventional methods, such as further excisional surgery (as apparently happened to most of the women who experienced recurrences during this study).
According to surgeon Richard Evans, MD, in his thought-provoking book The Cancer Breakthrough You've Never Heard Of, this watch-and-wait strategy, followed whenever necessary by conservative surgery, is generally effective at removing the tumor before it spreads to internal organs (provided that the recurrent tumor is found before it becomes bigger than the original growth). (Go to the website of Texas Cancer Center for a fuller explanation of Dr. Evans' views: http://www.texascancercenter.com)
In the Femara study, there was an excess of 29 cases of distant metastases in the placebo group vs. those receiving the drug over the four years of the study. Unfortunately, the authors do not state whether this figure reached statistical significance, a crucial point. A true picture of the impact of treatment on the incidence of distant metastases is one of several important things that were lost through early termination of the study.
Most importantly, we should be absolutely clear on the fact that in this study Femara was NOT shown to increase overall survival in a statistically significant way. A total of 42 women in the placebo group died of all causes while taking part in the trial compared to 31 in the Femara group. While the trend was in Femara's favor, this difference did not reach statistical significance (technically, it had a "p" value of =0.25). In other words, even the relatively small survival advantage conferred by Femara in this regard may simply have been due to chance.
Some oncologists see the drug's positive impact on "disease-free survival" as a sure-fire indicator that the drug will eventually be shown to increase actual survival. For instance, Dr. Larry Norton, MD said he was confident Femara would save lives. "Disease-free survival predicts survival," he asserted. "It always has. And there is a very big decrease in cancer incidence so there will be a survival difference," he predicted in the New York Times.
However, not everyone sees it that way. For instance, in a leading cancer textbook, Principles and Practice of Radiation Oncology (3rd edition), Dr. Thomas Pajak, past Director of Statistics of the American College of Radiology, argued that "there is no general agreement about what constitutes a treatment failure or how NED [no evidence of disease, ed] survival is to be estimated." In his view, there is no substitute for proof of effectiveness based upon the actual survival of everyone in the trial.
"Absolute survival," he wrote, "is usually the most objective measure of treatment because a patient is either alive or dead. It calls for no interpretation" (p. 233). This is the view I ascribe to. Disease-free survival is a somewhat arbitrary measure, referring to a decrease in deaths due specifically to a particular kind of cancer over a limited period of time. It refers only to deaths that are directly attributable to the extension and progression of the disease. But deciding on a cause of death is somewhat subjective. It also does not take into account the possibility that the treatment itself may shorten the survival of some patients through its own adverse effects on the body. As we shall see, Femara is not without risks and side effects, the impact of which may only become apparent after patients have been followed for a considerably longer time than they were in the present truncated study.
It has become a common practice to call a halt to a clinical trial if one arm of the treatment appears in an interim analysis to be yielding significantly better results than a competing arm. In this case, the trial was terminated about halfway through its projected duration, after independent analysts noted that the Femara patients were having fewer recurrences of their breast cancer. Although I understand and respect the ethical concerns that prompted this decision, I think it was wrong to halt the study at this time. Because of this decision, the long-term effects of the treatment remain uncertain. The public may never discover through a rigorous trial whether Femara treatment actually increases overall survival, or what its possible long-term side effects may be. Meanwhile, hundreds of thousands of women will receive a treatment whose full effects remain obscure. Those women will, in effect, become unwitting participants in an open-ended experiment. To me, that, in its own way is just as unethical as withholding a promising treatment.
A similar point of view was expressed by John Bryant, PhD, and Norman Wolmark, MD, in an accompanying editorial in the New England Journal of Medicine. "[T]he primary aim of the study was not fully achieved--namely,--"to determine the disease-free survival and overall survival for women who have previously received ~5 years of adjuvant tamoxifen randomized to receive either letrozole 2.5 mg daily or placebo daily for five years," they wrote. "Thus, although the results demonstrate a meaningful biologic effect of letrozole therapy after tamoxifen therapy, they do not demonstrate a significant survival benefit, nor do they convey information about the optimal duration of treatment beyond two to three years" (Bryant 2003).
Early closure of a clinical trial can also "freeze" a desirable result in time, before it has a chance to trend downward. In other words, trials may be abruptly halted at the point that the treatment in question is still showing superiority (in this case over the placebo). A crude way of putting this is "quitting while you're ahead." A full-term trial might have yielded less dramatic results or even negative results concerning overall survival.
[c]2004 Ralph W. Moss, PhD. All Rights Reserved
800-980-1234 * www.cancerdecisions.com
by Ralph W. Moss, PhD, Director, The Moss Reports
COPYRIGHT 2004 The Townsend Letter Group
COPYRIGHT 2004 Gale Group