Duration of adjuvant chemotherapy: How much is enough?

(I've been away this week at the Epic Users's Group meeting in WI as part of my IT role at Hopkins, so I haven't been able to contribute to my blog. Here's another post continuing my series on recently published literature about breast cancer that patients and physicians might find relevant.)

How many cycles of chemo do you need? As a breast cancer oncologist, this question comes up in clinic a lot, particularly since I see a large number of women with early-stage disease who are being referred for my recommendations about adjuvant chemotherapy following lumpectomy or mastectomy. Since many of these patients are going to be cured even without chemo (especially if the cancer is estrogen receptor positive, meaning they will be receiving tamoxifen or an AI drug), minimizing the number of cycles of a toxic treatment like chemotherapy is critically important to achieve the optimal balance between benefits of therapy and side effects and risks. Unfortunately, we rarely have solid scientific evidence about how much is "enough," and I can appreciate how patients must think oncologists are incredibly arbitrary and vague when we advise six cycles of treatment not four, 20 weeks of therapy not 18, etc. And why do the recommendations always seem to be expressed as even numbers? Why not ever 3 or 7 cycles? (Don't ask me to try to answer that one…)

A recent trial published as an Early Release article in the Journal of Clinical Oncology came out online several weeks ago. This was a report of CALGB study 40101, entitled "Six Cycles of Doxorubicin and Cyclophosphamide or Paclitaxel Are Not Superior to Four Cycles As Adjuvant Chemotherapy for Breast Cancer in Women With Zero to Three Positive Axillary Nodes." If you are a JCO subscriber you can read the full article online. Also, there is a good summary article about the trial in the ASCO Post last week. I won't try to describe the study in detail except to say that early stage breast cancer patients who were either node-negative or had three or fewer positive nodes were randomly assigned to receive chemo with four cycles or six cycles of AC (Adriamycin/Cytoxan, a very widely used adjuvant regimen) or four cycles or six cycles of Taxol. The bottom line is that four-year relapse-free survival and overall survival was virtually identical when comparing the patients who received four cycles of either AC or Taxol with those who received six cycles of the same. Whether Taxol was superior or inferior to the AC was not reported yet, as further follow-up will be required before an answer can be determined. Having said that, the AC combination is a widely-used standard, and in very few situations would Taxol be given by itself as adjutant treatment. But clearly 6 cycles is almost always going to be more toxic than 4 cycles; now we know that it's not better, so in this population of patients (i.e., not necessarily applicable to all breast cancer patients), six cycles should NOT be used routinely.

So why can't we answer the question about duration of chemotherapy for all types of breast cancer patients and in other cancers as well?  Why is it so hard to get this information? Well, it just is, and that is very disappointing for oncologists and patients alike. These studies are very difficult to do, since they require large numbers of patients willing to be randomized and many years of follow-up. Remember that no more then 3-4% of all adult patients in the U.S. participate in cancer clinical trials. This trial enrolled 3171 women over a 6-year period of time, and it was performed in multiple institutions. That means that many thousands more treated in those same institutions either were offered the trial and declined it or were not eligible in the first place. And it took 10 years to get the results. That's because there was only a 1% difference in survival at 4 years for women treated with 6 cycles of chemo vs. 4 cycles - not statistically significant and practically not enough of a difference to justify the misery of increasing the length of the chemo by an additional 4-6 weeks. In a disease like early-stage breast cancer, where most patients do well - overall survival was > 95% - you only see differences in treatments when there is an "event," like a relapse, a death due to breast cancer or another cause, or a side effect so severe that it requires the patient to drop out of the study. In other words, only when something bad happens to a patient in the study do you really learn whether there are differences in treatment. And I should also say that it is not intuitive that four cycles of breast cancer chemo of this type are equivalent to six. It could have gone either way, and there are many other examples in oncology where shorter durations of treatment are definitely inferior to longer. So it requires a clinical trial of this magnitude to answer a question of this type.

Or does it? Maybe I'll save this discussion for another blog post, but researchers have been realizing that the days of these mega-trials, with thousands of patients randomized to different treatments that themselves represent only tiny variations in therapy, are numbered. In this era where more and more of our cancer treatments are becoming more personalized to the individual patient, taking into account the genomic make-up of their cancer and other individual variables, designing more trials like CALGB 40101 will not serve our patients well or give us the important answers we need. Furthermore, the cost of doing such mega-trials is already prohibitive, and our clinical cancer research infrastructure can barely support trials that are much more modest in scope as it is. Unfortunately, we still have few "personalized" treatments for early stage breast cancer, so adjuvant chemotherapy will continue to be used, but stay tuned. I'm confident that in a few years, I will have options to offer my early stage breast cancer patients much more attractive than AC X 4. At least it's not 6.