Here at the annual meeting of the American Society of Clinical Oncology in Chicago, we see results of eagerly awaited clinical trials that have involved hundreds or even thousands of patients, millions of dollars, and years of hard work and analysis. The results can be a success; they may fail; but quite often fall somewhere in between. And even when successful, the results may not be as clear cut as one would like.
Such is the case with an important trial whose results came were reported at this meeting and simultaneously published in the New England Journal of Medicine about the use of aromatase inhibitors (AI) in post-menopausal women with breast cancer. As is so often the case, the researchers are highly regarded, the study well designed—but even with all the best efforts possible, left open are some extremely difficult questions.
In the current study, the question at hand was whether extending AI use from five years—which is the current standard of care with these drugs–to ten years resulted in any additional benefit among post-menopausal breast cancer patients whose cancer had not spread beyond the lymph nodes.
The study has received considerable press, and the reviews from the experts were somewhat mixed. It remains to be seen whether and how this study will influence clinical practice.
Briefly, post-menopausal women diagnosed with primary breast cancer were divided randomly into two groups: in group one, the women had already received approximately five years of treatment with a type of drug called an aromatase inhibitor (AI), which depletes the body’s estrogen production in an effort to prevent the breast cancer from retuning. They were then randomly selected to receive another five years of the same type of treatment with an AI. In group two, the women had also received approximately five years of an AI drug were then randomly chosen to receive placebo pills for another five years.
The press release headline version: women who had received the additional five years of treatment (or a total of about 10 years’ treatment with an AI reduced their chance of the cancer coming back by one-third (what is called the “hazard ratio”). That’s a significant finding, but tempered somewhat by the fact the chances of recurrence were somewhat low in the first place. In fact, rates of what’s called “disease (or progression) free survival (essentially, the percentage of patients whose disease did not recur during the study) was 91% in the women who got a placebo, and rose to 95% in the women treated with 10 years of AI medication. So while the “one-third” drop is correct, it’s actually a 4% absolute reduction in the chances that the breast cancer would return
Importantly though, the chances of death 10 years after the start of the study were almost identical: 93% in the AI group and 94% in the placebo group. There was also a substantial reduction in the risk of breast cancer in the opposite breast at 10 years of follow-up for the entire group: 3% for the AI/placebo group; 1% for the AI group.
Essentially, what these results tell us is that if a post-menopausal woman takes 10 years of an AI after diagnosis of a hormone sensitive breast cancer she will delay recurrence of the breast cancer, but she will not delay a death from breast cancer. The hope is that over time, those survival benefits will show up. But remember: these women were 65 when they were enrolled in the study, so the potential years of life gained at this point will be somewhat limited.
When the investigators discussed the results of the study, they acknowledged that most of the benefit was from avoiding the diagnosis and treatment of a new breast cancer in the other breast. It did not come from substantially reducing more distant recurrence, such as spread to bone, liver, lung or brain.
There are additional points that need to be emphasized in understanding the complexity in interpreting these results, especially when a woman currently taking the initial five years of adjuvant AI treatment is trying to make the decision whether to continue for another 5 years.
68% of all the women in this study had also received about 5 years of tamoxifen treatment before starting on their first 5 year round of AI treatment which occurred before they were randomized to an additional 5 years of AI treatment or placebo.
Tamoxifen is a drug that was approved in the mid-1970s for the treatment of recurrent estrogen sensitive breast cancer, but then became a standard workhorse for the adjuvant treatment of breast cancer (treatment to prevent breast cancer after primary surgery and radiation. So, many women had taken 10 years of medicines before they entered the study and randomized to an additional 5 years of treatment or placebo. That’s a long time to be treated.
And even despite that length of time on preventive therapy, 5% of the women who got the AI relapsed either in the other breast or elsewhere in her body. So that observation points out the serious problem with breast cancer that many people ignore: Most women with localized breast cancer relapse more than 5 years after diagnosis and treatment. In addition, according to the researchers, 5% of women with even the earliest most favorable breast cancer will die from their disease—some many, many years after diagnosis—even with all of this treatment aimed at preventing recurrence and death from the disease.
The study reported that there were not significant differences in quality of life between the two groups. However, in the women who received the additional five years of AI treatment, there were more fractures than those who received the placebo pill. In fact, 14% of the women in the AI treatment group had a fracture compared to 9% of the women who had received the 5 years of placebo treatment. Even though most of these fractures may be in less significant locations, they can still be problematic for women on AI treatment.
So what does all of this really mean? Does it change the way we will treat post-menopausal women with primary breast cancer? Is more really better?
That’s where the fog comes in:
We know that breast cancer can recur many years after diagnosis. We have learned through other trials that 10 years of adjuvant tamoxifen decreases the risk of recurrent breast cancer compared to 5 years of treatment. But the standard of care has moved on from tamoxifen to AI drugs. We don’t use much tamoxifen any longer to prevent breast cancer from returning in post-menopausal women. And prior to this clinical trial we didn’t have any good evidence that additional treatment beyond the currently recommended 5 years of adjuvant therapy with AI drugs really improves outcomes.
This study suggests that if you take an additional 5 years of treatment (or 10 years total) with an AI you will delay recurrence, but you won’t prevent death from breast cancer. Some have suggested it may be too early to see that benefit, so maybe it is premature to say conclusively that the additional 5 years of treatment doesn’t extend life. In addition, AI treatment reduced the risk of a breast cancer in the other breast by 2% over 10 years. That’s a real difference in statistics but again it isn’t an overwhelming number—but may make a lot of difference to women who can tolerate the AI medicine and don’t want to risk more treatment for another primary breast cancer even if that cancer is unlikely to threaten her life.
So will this study change the way doctors treat women with primary breast cancer? If a woman is post-menopausal with hormone receptor positive breast cancer will the doctor now recommend 10 years of AI treatment instead of 5 years as is currently recommended?
The experts who discussed the study didn’t exactly say “absolutely.” They did suggest doctors and their patients have a discussion on the benefits and risks, explaining what we know and what we don’t know as a result of this research. The same experts put a strong value on preventing another breast cancer from occurring, while others in the audience questioned whether the treatment with its associated cost, inconvenience and risk of fractures was worth avoiding a relatively small number of new breast cancers with no proof yet that it extends life.
So when asked the question whether this report will change clinical practice, one has to be a bit sanguine, since the results are not as straightforward as one would like. Simply put, different doctors and their breast cancer patients are going to come to different decisions.
As I wrote in the beginning of this commentary, life is not always clear cut. Neither is clinical research. We all want a crisp, clean answer to our questions. As this study shows, that is frequently not what we get.
When one thinks about it, that’s why the art of medicine—helping our patients make the decisions and choices that are best for them—remains an important part of our medical heritage and our current medical practice. I suspect—and hope–it will always be so.