Multicentre study of detection and false-negative rates in sentinel node biopsy for breast cancer

Abstract

Background Sentinel node biopsy has recently evolved as a means of staging the axilla in breast cancer with minimal surgical trauma. The aim of this prospective multicentre study was to identify factors that influencd the detection and false-negative rates during the learning phase. Methods Data on all 498 sentinel node biopsies performed between August 1997 and December 1999 in Sweden were collected. Results A sentinel node was found in 450 patients (90 per cent). Preoperative scintigraphy visualized 83 per cent of all sentinel nodes. The detection rate was higher with same-day injection of tracer than with injection the day before (96 versus 86 per cent; P < 0·01). Dye injected less than 5 min or more than 30 min before the start of the operation lowered the detection rate (less than 60 per cent versus more than 65 per cent; P = 0·02). The detection rate varied from 61 to 100 per cent between surgeons. The false-negative rate was 11 per cent. The presence of multiple tumour foci and a high S-phase fraction increased the risk of a false-negative sentinel node, whereas the number of operations performed by each surgeon was less important. Conclusion Training of the individual surgeon influenced the detection rate, as did timing of tracer and dye injection. The false-negative rate seemed to be related to biological factors.