A comparison of sentinel node biopsy before and after neoadjuvant chemotherapy: timing is important.

BACKGROUND: Because neoadjuvant chemotherapy is being used more frequently, the optimal timing of sentinel node biopsy (SNB) remains controversial. We previously evaluated the predictive value of SNB before neoadjuvant chemotherapy in clinically node-negative breast cancer. Our identification rate of the sentinel node among 52 patients before chemotherapy with a mean tumor size of 4 cm was 100%. In this study, we compared the identification rates of SNB before and after neoadjuvant chemotherapy and evaluated the false-negative rate of SNB after chemotherapy.

METHODS: A retrospective institutional database review identified 36 women who underwent SNB after neoadjuvant chemotherapy for breast cancer from 1999 to 2004. The initial clinical tumor size and lymph node status, SNB pathology, axillary lymph node dissection pathology, and residual pathologic tumor size were reviewed.

RESULTS: Sixteen of 36 patients had a clinically negative axilla before neoadjuvant therapy. SNB after neoadjuvant therapy was successful in 29 patients (80.6%), although 7 patients did not map (19.4%). Six of the 7 patients who failed to map had a clinically positive axilla initially. Axillary disease was found in 6 of 7 of these patients at dissection (85.7%). Of the 29 patients who mapped successfully, 13 (45%) were SNB negative, and 16 (55%) were SNB positive. Of the 13 SNB-negative patients, 2 had a positive axillary lymph node dissection, yielding a false-negative rate of 11%. Thirteen patients who mapped had a clinically positive axilla before therapy (45%). Of the 11 patients with true-negative SNBs, 7 (64%) were clinically node negative at presentation. The initial tumor sizes on examination ranged from 2 to 9 cm (mean, 5.0 cm), and residual pathologic tumor sizes ranged from 0 to 6 cm (mean, 1.8 cm). Failure to map correlated with a clinically positive axilla at presentation (100% vs 45%) but did not correlate with initial tumor size.

CONCLUSIONS: Sentinel node identification rates are significantly better when mapping is performed before neoadjuvant chemotherapy (100% vs 80.6%), with failure to map correlated with clinically positive nodal disease at presentation and residual disease at axillary lymph node dissection. Among patients who map successfully after chemotherapy, the false-negative rate is high (11%). Given these findings, we currently recommend SNB before neoadjuvant chemotherapy for clinically node-negative patients, and raise concerns about the use of SNB after neoadjuvant therapy in patients with an initially clinically positive axilla.