Purpose: The purpose of the study was to show that delayed axillary lymph node dissection (ALND) has higher rates of lymphedema compared with immediate ALND, using data from NSABP-B32 at Beaumont Hospital.
Method: NSABP B-32 at Beaumont had 207 patients with follow-up data on 199 patients, randomizing clinically negative axilla to sentinel lymph node biopsy (SLNB)+ALND (GrA N=98), and SLNB+cytology±ALND (GrB N=101). All patients had preoperative volumetric arm measurements and only node negatives had routine postoperative measurements assessing lymphedema for 36 months. We contacted node-positive patients for postoperative measurements for this study. Twenty-four and 15 cytology-positive patients had SLNB+ALND in GrA and GrB, respectively (SubGrA1 N=24; SubGrB1 N=15). Fourteen hematoxylin and eosin-positive patients had delayed ALND (SubGrB2a N=14).
Results: Lymphedema rate for node-positive SLNB+ALND was 10.3% [SubGrA1 (3/24)+SubGrB1 (1/15)=4/39] and node-negative SLNB+ALND was 6.8% (SubGrA2=5/74). Lymphedema was 14.3% for delayed ALND in SubGrB2a (2 of 14) and 0% for 72 SLNBs in SubGrB2b. Our study comparing immediate and delayed ALND lymphedema was not statistically significant (10.3% vs. 14.3%, P=0.65). Comparing node-negative ALND (SubGrA2= 5/74=6.8%) to node-positive ALND (A1+B1+B2a=6/53=11.3%) was not statistically significant (P=0.52). Comparing lymphedema for node-negative ALND (SubGrA2) to SLNB (SubGrB2b) only approached significance (6.8% vs. 0%, P=0.058).
Conclusions: The rate of lymphedema was higher in delayed ALND but not statistically significant. Comparison, however, is difficult, given the limited sample size. We urge the other centers of NSABP-B32 to validate this, by contacting the node-positive patients for measurements. The lymphedema rate for SLNB alone was 0% and approached statistical significance when compared with node-negative ALND.