A Clear Need for Randomized Trials
A Clear Need for Randomized Trials
Dr. Khan has provided a superb review of a recent series of papers on resection of intact breast primaries, some of which question the impact of this procedure on survival, and others that attempt to answer that question. The author is well qualified to discuss the issue, as she and Dr. Morrow authored the first in this recent series of articles that raise the question of improved survival in breast cancer.
The comparison of series of patients presenting with primary breast cancer and metastases treated by systemic therapy alone with those who also undergo resection of the primary tumor is not sufficient to assess therapies. It compares different therapies applied to different populations of breast cancer patients. That some women would be judged likely to survive long and benefit from local therapy, whereas others would be judged to have disease too advanced to trouble them with local therapy reflects good clinical judgment, not unintentional bias. Should the group undergoing local therapy fail to outlive those who do not receive such treatment, the clinical judgment of the treating physicians would be indicted rather than suggesting a lack of effect of local therapy. Some authors of the cited publications attempted to correct for this bias by multivariate analysis. That the difference in outcome persists can simply mean that given similar categories and parameters, the clinicians recommending treatment brought more critical insight to the decision than is reflected in our broad categories of staging.
More suggestive of therapeutic effect is the finding that among the population of women treated by surgery for the primary, outcomes were quite different depending on whether or not the tumor was resected to clear margins. Here, the bias of the treating clinician is eliminated, and surely, we can see a true benefit of treatment. Alas, those in whom clear margins could not be achieved must be considered to have had worse biology or inferior surgical skill applied to their treatment. Again, no scientific basis for a survival benefit associated with surgical treatment of the primary can be found.
I completely concur with Dr. Khan in her call for a randomized trial to address this increasingly important question and made a similar case in a presentation at the St. Gallen breast meeting last year.
The principle of not attempting to control the primary when metastasis was already established arose in a different era. The surgery not performed in M1 presentations was radical mastectomy, not excision and breast irradiation. The metastasis presented symptomatically or was found on physical examination or relatively crude blood tests and radiographs, not on positron-emission tomography/computed tomography or other techniques now used to detect clinically occult metastasis. The systemic therapies were often far less effective than those available today. These factors all made a general prohibition against surgery of the primary in M1 disease appropriate. The reasons for the prohibition no longer obtain.
Whenever surgery is performed in the presence of nodal metastases, we are creating an artificial partition among metastatic sites that is only partially valid. When axillary nodes are extensively involved, occult distant metastasis is also likely present. In that setting, the addition of irradiation to improve local control has been demonstrated to improve survival. This may be seen as a situation that is analogous to minimal M1 disease.
The flip side of the argument against local therapy was not directed at survival but at avoiding unnecessary treatment during the remaining year or years of life. When the metastases under discussion were symptomatic or clinically apparent, and when available systemic therapies were of limited effectiveness, death commonly came before complications of the primary tumor required intervention. Now, a long interval of remission is common and the primary tumor can progress to cause complications that require intervention. This may occur at a time when the patient has become more frail than at presentation. Careful clinical evaluation of the likely course of both primary disease and metastatic burden is required to provide best counsel regarding just the question of best managing the primary tumor, independent of a putative benefit on survival.