Survival Impact of Occult Metastases in NSABP B-32: Sentinel Lymph Node Biopsy Versus Axillary Dissection in Node-Negative Breast Cancer.
Donald L Weaver, Takamaru Ashikaga, David N Krag, Joan M Skelly, Stewart J Anderson, Seth P Harlow, Thomas B Julian, Eleftherios P Mamounas, Norman Wolmark. University of Vermont College of Medicine, Burlington; National Surgical Adjuvant Breast and Bowel Project, Pittsburg, PA; Allegheny General Hospital, Pittsburgh, PA; Aultman Caner Center, Canton, OH
Background: Retrospective and observational analyses suggest occult lymph node metastases are an important prognostic factor for recurrence or survival. Prospective data with clinical outcome in randomized sentinel node trials has not been available.
Design: Women with breast cancer were randomized to sentinel lymph node biopsy (SLNB) plus axillary dissection or SLNB alone. SLN paraffin tissue blocks from cases with pathologically negative SLNs were centrally evaluated for occult metastases deeper in the blocks. Routine and cytokeratin immunohistochemical (IHC) stains were used at two widely spaced additional levels 0.5 and 1.0 mm deeper; findings were blinded and not used for clinical treatment decisions. Initial evaluation at participating sites was designed to identify all macrometastases >2 mm.
Results: Occult metastases were detected in 15.9% (95% confidence interval [CI]: 14.7% – 17.1%) of the 3887 cases. Log-rank tests indicated a significant difference between occult metastasis-positive and -negative patients for overall survival (OS; p=0.03), disease-free survival (DFS; p=0.02), and distant disease-free interval (DDFI; p=0.04), respectively. Corresponding adjusted hazard ratios (HR) for OS, DFS, and DDFI are 1.40 (CI: 1.05 – 1.86), 1.31 (CI: 1.07 – 1.60), and 1.30 (CI: 1.02 – 1.66), respectively. Five year Kaplan-Meier estimates for OS for patients with and without occult metastases detected were 94.6% and 95.8%, respectively. In a subset analysis by occult metastasis categorical size, HRs for isolated tumor cell clusters (ITC) and micrometastases are 1.29 and 1.66 (OS), 1.19 and 1.41 (DFS), 1.19 and 1.42 (DDFI), and, for survival without breast cancer death, 1.38 (CI: 1.02 – 1.87) and 1.91 (CI: 1.41 – 2.59), compared to no metastases having been detected. Five year Kaplan-Meier estimates of survival without breast cancer death are 98.4%, 97.8%, and 96.0% when no metastases, ITCs, or micro/macrometastases are detected.
Conclusions: Occult metastases are an independent prognostic variable in sentinel nodes that are negative on initial examination; however, the outcome difference magnitude at five years is small (1.2%). ITCs exert less impact than micrometastases. A clinical benefit from additional evaluation, including IHC, of initially negative sentinel nodes in breast cancer is not supported by this data.
Tuesday, March 1, 2011 2:00 PM
Platform Session: Section B, Tuesday Afternoon