The Impact of Prophylactic Axillary Node Dissection on Breast Cancer Survival—A Bayesian Meta-Analysis

Richard K. Orr MD, MPH
Original Article
Volume 6, Issue 1 / January ,

View full article HTML | View full article PDF | Download Citation

Abstract

Background: Because of the general acceptance of the NSABP B-04 study, prophylactic axillary node dissection for women with clinically negative axillae is considered diagnostic, but not therapeutic, by many oncologists. Nevertheless, several authors have shown that B-04 did not include enough patients to exclude a small survival advantage.

Methods: A Bayesian meta-analysis of the available literature was performed comparing standard treatment to standard treatment without axillary node dissection. Six randomized controlled trials were identified, consisting of nearly 3000 patients and spanning four decades.

Results: All six trials showed that prophylactic axillary node dissection improved survival, ranging from 4% to 16%, corresponding to a risk reduction of 7%-46%. Combining the six trials showed an average survival benefit of 5.4% (95% CI = 2.7-8.0%, probability of survival benefit. > 99.5%). Adjusting for biases in the individual studies did not alter the conclusions, nor did subset analysis of Stage I patients.

Conclusions: Axillary node dissection improves survival in women with operable breast cancer. Nevertheless, two important limitations of this analysis are noteworthy. Few of the patients in the six trials had T1a tumors, so extrapolation of these results to this subset (and those with nonpalpable tumors) may be inappropriate. Essentially no patients in the six trials were treated with adjuvant therapy, as contrasted to current clinical practice. It is possible that the risk reduction seen in this meta-analysis may be diminished in patients receiving adjuvant chemotherapy. Despite these limitations, this study suggests that axillary dissection should be performed in most women with palpable tumors for diagnostic, as well as therapeutic, purposes.

Keywords

Go to Issue Contents