Accuracy of selective sentinel lymphadenectomy after neoadjuvant chemotherapy: effect of clinical node status at presentation.
BACKGROUND: Both neoadjuvant chemotherapy and selective sentinel lymphadenectomy (SSL) are increasingly being used in treating primary breast cancer. It is important to determine whether SSL can be used after neoadjuvant chemotherapy and whether clinical node status at presentation affects accuracy of SSL. STUDY DESIGN: Between 1995 and 2003, 53 evaluable cases of invasive breast cancer were treated with neoadjuvant chemotherapy followed by SSL and completion axillary node dissection. The accuracy of SSL and the number of failed SSLs were assessed in the entire group and in the subset that were clinically node positive at presentation. RESULTS: The sensitivity of SSL was 96%, the negative predictive value was 96%, and the sentinel node identification rate was 94%. Of the 53 evaluable patients, 23 had clinically node-positive disease at presentation (43%) and the remainder were clinically node negative (57%). Of the successfully completed SSL, the status of the sentinel lymph node corresponded to that of overall axillary status in 49 of 50 patients (accuracy rate 98%). Two of the 23 patients with clinically node-positive disease at presentation had unsuccessful SSL. Of the remaining 21 patients with a clinically positive axilla before systemic therapy, a false-negative SSL result occurred in 1 patient (accuracy 95%, sensitivity 91%). CONCLUSIONS: Selective sentinel lymphadenectomy after neoadjuvant chemotherapy is both feasible and accurate. Although early reports found a lower performance of SSL after neoadjuvant chemotherapy, this study suggests reevaluation of the current practice of full axillary lymph node dissection in this setting, particularly in those patients who are clinically node negative at presentation.
Lang, JE; Esserman, LJ; Ewing, CA; Rugo, HS; Lane, KT; Leong, SP; Hwang, ES
Volume / Issue
Start / End Page
International Standard Serial Number (ISSN)
Digital Object Identifier (DOI)