Association of Occult Metastases in Sentinel Lymph Nodes and Bone Marrow With Survival Among Women With Early-Stage Invasive Breast Cancer

CONTEXT Immunochemical staining of sentinel lymph nodes (SLNs) and bone marrow identifies breast cancer metastases not seen with routine pathological or clinical examination. OBJECTIVE To determine the association between survival and metastases detected by immunochemical staining of SLNs and bone marrow specimens from patients with early-stage breast cancer. DESIGN, SETTING, AND PATIENTS From May 1999 to May 2003, 126 sites in the American College of Surgeons Oncology Group Z0010 trial enrolled women with clinical T1 to T2N0M0 invasive breast carcinoma in a prospective observational study. INTERVENTIONS All 5210 patients underwent breast-conserving surgery and SLN dissection. Bone marrow aspiration at the time of operation was initially optional and subsequently mandatory (March 2001). Sentinel lymph node specimens (hematoxylin-eosin negative) and bone marrow specimens were sent to a central laboratory for immunochemical staining; treating clinicians were blinded to results. MAIN OUTCOME MEASURES Overall survival (primary end point) and disease-free survival (a secondary end point). RESULTS Of 5119 SLN specimens (98.3%), 3904 (76.3%) were tumor-negative by hematoxylin-eosin staining. Of 3326 SLN specimens examined by immunohistochemistry, 349 (10.5%) were positive for tumor. Of 3413 bone marrow specimens examined by immunocytochemistry, 104 (3.0%) were positive for tumors. At a median follow-up of 6.3 years (through April 2010), 435 patients had died and 376 had disease recurrence. Immunohistochemical evidence of SLN metastases was not significantly associated with overall survival (5-year rates: 95.7%; 95% confidence interval [CI], 95.0%-96.5% for immunohistochemical negative and 95.1%; 95% CI, 92.7%-97.5% for immunohistochemical positive disease; P = .64; unadjusted hazard ratio [HR], 0.90; 95% CI, 0.59-1.39; P = .64). Bone marrow metastases were associated with decreased overall survival (unadjusted HR for mortality, 1.94; 95% CI, 1.02-3.67; P = .04), but neither immunohistochemical evidence of tumor in SLNs (adjusted HR, 0.88; 95% CI, 0.45-1.71; P = .70) nor immunocytochemical evidence of tumor in bone marrow (adjusted HR, 1.83; 95% CI, 0.79-4.26; P = .15) was statistically significant on multivariable analysis. CONCLUSION Among women receiving breast-conserving therapy and SLN dissection, immunohistochemical evidence of SLN metastasis was not associated with overall survival over a median of 6.3 years, whereas occult bone marrow metastasis, although