Sentinel lymph node biopsy (SLNB) has become a standard technique for axillary staging in early breast cancer. It has replaced complete axillary lymph node dissection (ALND) because of an accuracy rate greater than 90% in predicting axillary lymphatic involvement. A false-negative (FN) SLNB is defined as when the sentinel lymph nodes (SLNs) are reported to be negative in the histologic assessment, but other axillary nodes harbor cancer cells.

According to a meta-analysis of 183 articles involving 9306 patients, the FN rate (FNR) of SLNB ranged from 4.6% to 16.7%,1 with a mean of 7.5%.2 A prospective multicenter study showed that different factors, such as tumor size, location, and surgeon experience, can affect the FNR,3 and patients who undergo neoadjuvant chemotherapy (NAC) before SLNB may have a higher FNR.4 This may be due to several factors. First, NAC can cause the lymphatic pathways to be altered because of the formation of granulation tissue, fat necrosis, and fibrosis; therefore, the mapping agent might be delivered to a nonsentinel node in the axilla. Second, tumor emboli or cellular materials could block the lymphatic channels.5 Third, although NAC may cause regression of the tumor in some SLNs, it may not do so in other nodes. Therefore, a negative SLN could be found even though there are other involved nodes in the axilla.

Some studies have reported no difference in the FNR of SLNB in patients treated with NAC compared with those who had initial surgery.6 However, it should be kept in mind that chemotherapy can alter the lymphatic pathways by fibrosis or tumor emboli, and that finding a sentinel node without treatment effects after NAC in a patient with a previously pathologically positive node may imply an FN result.7

With the increasing use of NAC, it has become essential for every breast care unit to analyze its institutional accuracy and compare it with international standards. This study aimed to investigate the FNR of SLNB after NAC in patients who were initially node-positive but became node-negative after treatment.

The study included patients with breast cancer with histologically proven positive axillary lymph nodes who had undergone SLNB and an ALND after NAC between 2015–2019 and had been operated on by the first author (R.O.) at a private center. The collected data included the patient's age at the time of surgery, pathologic tumor size, Ki-67 level, breast involvement side, family history of breast or ovarian cancer, type of surgery, lymphovascular invasion (LVI), and the number of extracted and involved lymph nodes in SLNB and ALND. Using the convenience sampling method, only patients with available information about the timing of chemotherapy, SLNB, and ALND were included in the study. Patients who had positive lymph nodes in their ultrasound examination after NAC were excluded.

All HER2-positive patients received docetaxel, carboplatin, and trastuzumab (TCH), with or without pertuzumab, because of insurance and drug access limitations. Luminal subtype and triple-negative cases received a doxorubicin, cyclophosphamide, and paclitaxel (ACT) chemotherapy regimen. All patients underwent surgery 3–4 weeks after completing NAC. Patients with clipped positive lymph nodes before chemotherapy were excluded. A dual-tracer technique was used for all SLNBs. During surgery, the detected sentinel nodes were sent for frozen section examination. The SLNs were considered positive if isolated tumor cells, micrometastases, or macrometastases were detected in the frozen or permanent histologic examination.

ALND was performed in all patients with any residual tumor in an SLN or when only 1 or 2 negative nodes were found during the operation. All excised lymph nodes were assessed by hematoxylin-eosin staining in the permanent histology examination.

The final analysis was conducted on 368 patients whose mean (SD) age was 46.58 (10.91) years (range, 18–79 years). The median pathologic tumor size was 15 mm (IQR, 23.5 mm). The characteristics of these patients are shown in Table 1. Of the 368 patients who underwent histological analysis of their SLNs, 205 (55.7%) had a positive SLN in the histologic analysis after surgery, whereas 163 (44.3%) had negative results. The FNR of SLNB was 9.8% (16 patients), as shown in Figure 1.

Among the 16 patients with FN results, 5 had 1 involved node, whereas the rest had 2 involved nodes in ALND (Figure 1). Table 2 shows the sensitivity, specificity, and accuracy of SLNB after NAC. The results of crude and adjusted analyses of the pathologic factors associated with FNR are reported in Table 3.

According to the final analysis, a higher number of ALNs dissected was associated with a higher FN rate (odds ratio (OR), 1.21; 95% CI, 1.01–1.45), whereas a higher number of SLNs excised was associated with a lower FN rate (OR, 0.42; 95% CI, 0.18–0.97).

Our study revealed that 9.8% of patients with breast cancer with pathologically proven positive axillary nodes experienced an FN SLNB after NAC. Small-scale studies evaluating SLNB after NAC in patients with biopsy-proven clinical N1 (cN1) disease have shown various results.4,8

The GANEA study9 examined the safety of SLN biopsy in a large number of patients with early breast cancer after NAC. The study included 419 women, and the FNR was found to be 9.4% in node-negative and 15% in node-positive patients. The overall FNR was 11.5%. Also, the National Surgical Adjuvant Breast and Bowel Project (NSABP) B-27 trial10 included both cN0 and cN1 disease and reported an SLN FNR of 10.7% after chemotherapy.

Three large trials in Europe, the United States, and Canada evaluated the FNR of SLNB after NAC in women initially presenting with biopsy-proven node-positive breast cancer.11–13

The SENTINA trial in Europe11 was conducted in 103 different centers in Germany and Austria and evaluated 1737 patients who underwent NAC. In this study, SLNs were identified in 80.1% of the patients who converted from cN+ to clinical N0 after NAC (ycN0), and the FNR was 14.2%. The FNR was 24.3% when only 1 SLN was removed, and it dropped to 18.5% when 2 SLNs were found.

The ACOSOG Z1071 Alliance trial in the United States12 recruited 756 women from 136 institutions, of whom 649 had SLNB and ALND following NAC. The FNR was 12.6%, which means that positive nodes were found in ALND despite the presence of negative SLN(s) in 39 patients.

The SN FNAC trial in Canada13 recruited 153 patients, and the SLNs were assessed by immunohistochemistry (IHC). Any size of metastases, including isolated tumor cells, was considered positive, and the FNR decreased from 13.3% to 8.4% by using IHC.

When only 1 SLN is removed, the FNR tends to be high. Patients with only 1 SLN constituted 31% of patients in SENTINA, 20.4% in the Alliance trial, and 26.5% in the SN FNAC trial. In a subgroup of patients with at least 3 SLNs removed, the FNR decreased to 8.6% in SENTINA and 9.1% in the Alliance trial. In our study, all patients with only 1 or 2 SLN(s) underwent ALND because of the probability of a high FNR. Only patients with 3 negative SLNs were spared from ALND.

In line with the above trials, a meta-analysis by van Nijnatten et al.14 showed that the FNR was worse if only 1 SLN was removed compared with 2 or more SLNs (23.9% vs 10.4%). In addition to harvesting more than 1 SLN and IHC assessment of the SLNs, the use of a dual tracer during SLNB and clipping or marking the positive nodes before NAC could decrease the FNR.

In a meta-analysis by Tee et al.,15 1921 patients with breast cancer were included. The FNR of SLNB using a dual tracer was 11% compared with 19% with a single tracer. The FNR was 20% when only 1 SLN was removed and dropped to 12% and 4% when 2 and 3 SLNs were harvested, respectively. Similarly, a recent study showed that removing 3 or more SLNs after NAC in patients with breast cancer decreased the FNR from 19.1% to 8.7%.16

We used dual mapping in all our patients, but we excluded the patients who had undergone targeted axillary dissections because marking or clipping involved axillary lymph nodes is not done routinely in many centers before or during NAC.

All SLNs in our study were examined by hematoxylin-eosin, and IHC was used only in a minority with suspicious hematoxylin-eosin results. In comparison with other studies, our FNR was lower than expected, especially considering our inclusion criteria that enrolled only patients with 1 or 2 negative SLNs. The subgroup analysis of the 14 patients with FNR, considering the number of SLNs, had no significant results.

Our false negative rate is very close to the value reported in the study by Lazar et al.17, which found an identification rate of 93.13% with an FNR of 7.4% for SLNB after NAC in 102 patients with breast cancer in a tertiary single center.

We found a relation between a larger post-NAC residual tumor in the primary site and FNR. This association was also detected by Ozmen et al. in 2009 in a study on 77 patients with locally advanced breast cancer who had received NAC, where a higher FNR was found in patients with tumors larger than 2 cm.18 However, this finding must be considered with caution because of the small sample size, especially given the previous studies that showed an indirect association between tumor size and the FNR of SLNB.3,19 Future studies with larger sample sizes could clarify this association.

The strength of our study is that all operations were performed by a single experienced surgeon, which ensures the similarity of referred patients for NAC and could eliminate the impact of surgical skills on the FNR. This study is valuable because most previous studies were conducted in developed countries, and the few studies that were conducted in developing countries, including Iran, had a very small sample size.

Our study had some limitations, including the small sample size and its retrospective nature. The number of FN results (14 patients) is not sufficient for any subgroup analysis. The association between FNR and tumor subtype, receptor status, and other histologic factors was not statistically significant.

In conclusion, SLNB after NAC in previously node-positive patients is feasible and accurate with a low FNR. According to our study, when deciding to spare ALND in the presence of a negative SLNB in the neoadjuvant setting, the size of the residual tumor, i.e., the response of the primary tumor to systemic therapy, should be considered.

This cross-sectional study received approval from the ethics committee of Imam Khomeini Hospital of Tehran University of Medical Sciences (ethics number IR.TUMS.IKHC.REC.1402.311).