Ten-year follow-up results of a randomised controlled study comparing level-I vs level-III axillary lymph node dissection for primary breast cancer.

The most appropriate level of axillary dissection for breast cancer remains unclear. The present randomised study compared the treatment results of level-I vs level-III dissection in T1/2/3 and N0/1a/1b (1987 UICC classification) breast cancer without distant metastasis. Between 1995 and 1997, 522 patients were enrolled, and 514 were eligible. They were stratified into breast-conserving surgery or mastectomy, and then further stratified into level-III dissection (group-A, n=258) or level-I dissection (group-B, n=256). All patients were given oral 5-fluorouracil at 200 mg day-1 and tamoxifen at 20 mg day-1, daily for 2 years. Group-A resulted in a significantly longer operation time (77.0 vs 60.5 min, P<0.0001) and significantly larger blood loss (62.1 vs 48.1 ml, P<0.0001) than group-B, but in no significant differences in the frequencies of arm oedema and shoulder disturbance. Group-A resulted in a significantly larger number of dissected nodes than group-B (18.7 vs 14.8, P<0.0001), but in no differences in the number of involved nodes (1.54 vs 1.44). There were no significant differences in the 10-year overall and disease-free survival rates: 89.6 and 76.6% for group-A vs 87.8 and 74.1% for group-B, respectively. In conclusion, level-III dissection resulted in a longer operation time and greater blood loss than level-I, but did not improve the survival rate. Level-III dissection is not a recommended surgery for T1-3/N0-1b breast cancer.