Factors affecting failed localisation and false-negative rates of sentinel node biopsy in breast cancer - results of the ALMANAC validation phase

Background. Despite the widespread application of sentinel lymph node biopsy (SLNB) for early stage breast cancer, there is a wide variation in reported test performance characteristics. A major aim of this prospective multicentre validation study was to quantify detection and false-negative rates of SLNB and evaluate factors influencing them. Methods. Eight-hundred and fourty-two patients with clinically node-negative breast cancer underwent SLNB according to a standardised protocol that used a combination of radiopharmaceutical 99mTc-albumin colloid and Patent Blue V dye. SLNB was followed by standard axillary treatment at the same operation in all patients. Results. Sentinel lymph nodes (SLNs) were identified in 803 (96.1%) of 836 evaluable cases. The median number of SLNs removed per patient was 2 (range 1-9). There were 19 false negatives, resulting in a sensitivity of 263/282 (93.3%) and accuracy 782/803 (97.6%). SLNs were successfully identified by blue dye in 698 (85.6%), by isotope in 698 (85.6%), and by the combination of blue dye and isotope in 782 (96.0%) of 815 patients. Among 276 node positive patients, one or more positive SLNs were identified by blue dye in 251 (90.9%), by isotope in 246 (89.1%) and by the combination of blue dye and gamma probe in 258 (93.5%). Obesity, tumor location other than upper outer quadrant and non-visualisation of SLNs on the pre-operative lymphoscintiscan were significantly associated with failed localisation (p < 0.001, p=0.008, p < 0.001, respectively). The false-negative rate in patients with grade 3 tumors was 9.6%, compared with 4.7% in those with grade 2 tumors (p=0.022). The false-negative rate in patients who had one SLN harvested was 10.1%, compared with 1.1% in those who had multiple SLNs (three or more) removed (p=0.010). Conclusion. SLNB can accurately determine whether axillary metastases are present in patients with early stage breast cancer with clinically negative axillary nodes. Both success and accuracy of SLNB are optimised by the combined use of blue dye and isotope. SLNB success decreases with increasing body mass, tumor location other than the upper outer quadrant and non-visualisation of hot nodes on the pre-operative lymphoscintiscan. This study demonstrates reduction in the predictive value of a negative SLNB in grade 3 tumors.