Introduction

Sentinel Lymph Node Biopsy (SLNB) has already become the standard of care for axillary staging in clinically node-negative breast cancer, decreasing postoperative arm morbidity without a negative effect on overall survival if Axillary Lymph Node Dissection (ALND) can be omitted [1-3]. Although SLNB is acceptable for clinically node-negative patients after Neoadjuvant Chemotherapy (NAC) [4], ALND has been performed for initially node-positive patients after NAC because of the high FalseNegative Rates (FNRs) (8.4% to 14.2%) of SLNB [5-7].

Among studies performed to evaluate the use of SLNB to stage the axilla after NAC in patients initially presenting with nodepositive disease, the American College of Surgeons Oncology Group (ACOSOG) Z1071 trial reported an FNR of 12.6%, which dropped to 7.2% when the positive node was marked with a clip and removed with the Sentinel Lymph Node (SLN) [8], although the difference was not statistically significant. Several techniques have been introduced to improve the accuracy of SLNB after NAC in node-positive patients. The marking axillary lymph nodes with Radioactive Iodine Seeds (MARI) procedure allowed removal of the marked metastatic lymph node without SLNB and showed an FNR of 7% [9]. The Targeted Axillary Dissection (TAD) procedure involves removing SLNs with the clipped node by localization with iodine-125 radioactive seeds and had an FNR of 2.0% [10]. Other labeling and localization techniques such as wire localization of clipped nodes [11-15], tattooing [16-18], and clipped node localization using Magseed or Radiofrequency Identification Tags (RFID) [19,20] are under investigation.

While an increasing number of reports have shown the usefulness of TAD for axillary staging, several studies have described the difficulty in identifying the clipped node [12,21,22]. These reports showed successful removal of the marked node if the clip was detected under US; however, if the clip was not visualized on US, it was removed as one of the SLNs, as a palpable node, as a part of ALND, or left behind in axilla. Because 23% to 75% of patients receiving NAC achieve Pathologic Complete Response (pCR) [23,24], accurate axillary assessment can result in these patients avoiding ALND. Thus, the purpose of this study was to investigate the accuracy of TAD using an Ultrasound (US)-visible clip for axillary staging of patients with node-positive breast cancer treated with NAC, with the secondary aim of assessing the US-identification rate of the clip and localization technique for the TAD.

Materials and Methods

Between August 2017 and September 2021, a prospective study was performed for patients with clinical T1-3, N1, 2, M0 breast cancer undergoing NAC followed by surgery approved by the Research Ethics Committee of Kameda Medical Center (approval number: 16-063, 16-071). Written informed consent was obtained from each patient before enrollment. Patients with internal mammary lymph node metastases, prior axillary surgery, or pregnancy were excluded.

Patient Evaluation and Clip Deployment

All patients underwent routine imaging, including mammography, US, and magnetic resonance imaging. ¹⁸F-fluorodeoxyglucose Positron-Emission Tomography (PET)/ Computed Tomography (CT) was performed for detection of metastases. Suspicious axillary nodes on any image were reevaluated by targeting US, and Fine-Needle Aspiration Biopsy (FNA) was performed for lymph nodes showing focal or diffuse cortical thickening (more than 3 mm thick), loss of fatty hilum, or an abnormally round or irregular shape [25]. If more than two suspicious LNs were identified, FNA was performed for two suspicious nodes or the most abnormal-appearing node. For a Targeted Lymph Node Biopsy (TLNB), a US-visible clip was placed in the biopsy-proven node under US guidance if metastases were identified by FNA.

NAC and Assessment of the Therapeutic Effect

NAC was administered using anthracycline- (four cycles of adriamycin with cyclophosphamide) and/or taxane-based regimens. In tumors showing Human Epidermal Growth Factor Receptor 2 (HER2) overexpression, anti-HER2 therapy with trastuzumab ± pertuzumab was added. US was performed one week after placement of the clip and then every 2 months until the surgery to assess the visibility and dislocation of the clip. After completion of NAC, the effect of NAC was evaluated using the Response Evaluation Criteria in Solid Tumors (RECIST) guideline [26].

Surgical Procedure

For a dual-technique SLNB, subareolar injection of radioisotope (technetium-99m sulfur colloid) was performed on the day before surgery. On the day of surgery, after patients were placed under general anesthesia, patent blue dye was injected into the subareolar region of the breast. Then, careful examination with axillary intraoperative US (IOUS) was performed to detect the clipped node. We used several localization techniques, including fluorescence techniques, wire localization, dye-injection method, or a combination of these techniques, under US guidance. For the fluorescence technique, after injection of 0.1-0.2 mL of indocyanine green dye (ICG) into the clipped node, ICG fluorescent imaging was performed using the Photodynamic Eye infrared camera (Hamamatsu Photonics k. k., Shizuoka, Japan). For the dyeinjection method, 0.1-0.2 mL of pyoktanin blue dye was injected into the clipped node. Pyoktanin blue is purple in color, which is distinguishable from the patent blue dye used in SLNB. When the fluorescence technique or dye injection was used in combination with wire localization, the dye was injected though an introducer needle of the hook-wire (Reverse-wire set: introducer needle, 21 gauge × 90 mm and reverse-wire, 0.25 mm × 230 mm; Breast Localization Needles, Hakko Co., Ltd. Nagano, Japan) without additional puncture. The presence of the clip in the removed node was confirmed using US. Specimen radiographs were obtained if the clip was undetectable on US.

During the surgery, the clipped node was removed first, followed by SLNB. Then, axillary evaluation was performed by palpation and palpable nodes were removed if present. All radioactive or blue nodes found in the axilla after removal of the clipped node were excised as SLNs. Palpable nodes were assessed separately. TAD consisted of TLNB, which indicates clipped node biopsy, SLNB, and palpable lymph node biopsy.

Pathologic Evaluation

Among the initial series of patients between August 2017 to January 2020, all patients underwent ALND after TAD (Group A). The lymph nodes removed with TAD were assessed as 2-mmthick slices by using formalin-fixed paraffin embedded tissue blocks stained with hematoxylin and eosin (HE). In the second series of patients between February 2020 to September 2021, the lymph nodes removed with TAD were sent for an Intraoperative Frozen Section (IOFS) analysis assessed by HE staining with 2-mm-thick slices (Group B), and ALND was not performed if the IOFS analysis showed no metastatic involvement. The lymph nodes removed by TAD were reexamined in permanent sections after surgery. Immunohistochemical (IHC) staining with pancytokeratin was performed in permanent sections for lymph nodes removed with TAD from all patients if negative pathologic results were obtained. Dissected axillary nodes were evaluated only by HE staining, with one section analyzed per lymph node. The clipped node, the SLN, the palpable node, and the dissected nodes were independently analyzed, and the pathologic report specified which lymph nodes showed metastatic involvement. Any metastatic foci, including isolated tumor cells (ITCs, ≤0.2 mm at the largest diameter) detected by HE or IHC staining, were considered node-positive, and pCR was defined by complete eradication of malignancy, including ITCs. Hormone receptor positivity was defined as an Allred score ≥ 3. Tumors with HER2 scores of 3+ were considered HER2-positive (HER2+). The positive result of fluorescence in-situ hybridization was considered HER2+ in tumors with HER2 scores of 2+. The US visibility of the clip, success rate of the TLNB, the localization technique of the clipped node, FNR of TLNB, TAD, IOFS analysis, and the rate of residual metastases in the dissected nodes after TAD in patients who underwent ALND were investigated.

Statistical Analysis

The FNR of TLNB or TAD, which was assessed in the patients who underwent ALND, was defined as the number of patients who showed negative TLNB or TAD results divided by the total number of patients with pathologically node-positive status. The FNR of IOFS (IOFS-FNR) was defined as the number of patients who showed a negative IOFS result during the surgery divided by the total number of the patients whose final pathological assessment showed metastatic carcinoma of any size. Fisher’s exact test was used for comparisons between the categorical variables of two groups, and the Chi-square test was used for comparisons among three or more groups. The Mann-Whitney U test was used for continuous variables to compare the two groups. A p value lower than 0.05 was considered statistically significant.

Results

A total of 26 patients (median age, 52 years; range, 3873 years) were enrolled in this study, including 13 patients in Group A and 13 patients in Group B. Overall, eight patients had T1 disease (30.8%), 14 had T2 disease (53.8%), and four had T3 disease (15.4%). Initial imaging revealed one, two, and three or more suspicious lymph nodes in eight (30.8%), nine (34.6%), and nine patients (34.6%), respectively. The predominant histologic tumor type was invasive ductal carcinoma (84.6%, 22/26), and the effect of NAC on the axilla was complete response in 16 patients (61.5%) and partial response in 10 patients (38.5%). After NAC, breast-conserving surgery (BCS) was performed in 10 patients (38.5%) and mastectomy was performed in 16 patients (61.5%). The pathological assessments showed persistence of lymph node disease in 17 patients (65.4%). ALND was performed for all 13 patients in Group A and for eight patients in Group B, including seven patients with residual nodal disease and one patient with nodal pCR. ALND was performed as a second surgery in one patient in Group B. The rate of second surgery was 7.69% (1/13). ALND was omitted in five patients in Group B, including four patients who achieved nodal pCR and one patient whose clipped node showed micrometastases in the permanent specimen. The clinicopathologic and treatment details of the patients are listed in Table 1. Nodal pCR was achieved in 34.6% (9/26) of the patients. Categorization of nodal pCR by subtypes was as follows: estrogen receptor-positive (ER+)/HER2-negative (HER2-) subtype, 0% (0/10); ER+/ HER2+ subtype, 28.6% (2/7); ER-/HER2+ subtype, 100.0% (3/3); and ER-/HER2- subtype, 66.7% (4/6). The effect of chemotherapy showed statistically significant differences among subtypes (p = 0.001).

Clip insertion and US visibility of the clip

As for the clips, Hydromark™ breast biopsy site markers (Hydromark, T3 shape; Devicor Medical Japan, Tokyo, Japan) were used in 13 patients of Group A and one patient of Group B. UltraCor™ Twirl™ breast marker (Twirl; C.R. Bard, Inc, NJ, USA) was used in 12 patients of Group B (p < 0.001). The bias in the type of clips used between the two groups was because Twirl had become available under the national insurance program during the period when patients in Group B were undergoing treatment whereas Hydromark still had not. One clip was used in 21 patients, whereas two clips were used in five patients. All 26 patients underwent successful clip insertion into the positive node. One week and 2 months after clip placement, each clip was located inside the node in all 26 patients. However, evaluation of clip displacement became difficult 4 months after placement in patients who remarkably responded to the NAC because only a clip was detectable on US (Figure 1).

Figure 1: US images of a Twirl and a Hydromark placed in axillary metastatic lymph nodes. The Twirl was visualized as an echogenic ring structure in the enlarged node after placement (a, arrow) and in the shrunken lymph node after NAC (b, arrow). The Hydromark was visualized as a hyperechoic echogenic clip with a hypoechoic gel in the abnormal LN one week after placement (c, arrow) and in the shrunken lymph node after NAC (d, arrow).

After completion of NAC, the US-identification rate of the clipped nodes was 96.2% (25/26), including 92.9% (13/14) of the Hydromark clips and 100% (12/12) in the Twirl clips (p = ns, Table 2). US identification of the Twirl clips was easier than that of the Hydromark clips. Although preoperative US failed to detect the Twirl clip in one patient, targeting IOUS successfully identified it after a mobile C-arm X-ray system (C-arm) detected the metallic clip in the operation room. On the other hand, the C-arm could not detect the missing Hydromark clip.

Intraoperative Localization and Excision of Clipped Nodes

The clipped node marked with the Twirl clip was successfully removed by wire-and-dye localization in 100% (12/12) of the patients, whereas the successful removal rate for the clipped node marked with the Hydromark clips was 78.6% (11/14), including 50% (1/2), 100% (2/2), and 80% (8/10) with the fluorescence technique, wire-and-fluorescence technique, and wire-and-dye localization, respectively (p = ns, Table 2).

Among the three cases showing failure of removal, the Hydromark clip was not clearly visualized in one patient. For this patient, a combination of wire-and-dye localization was attempted to the lymph node that appeared to be the clipped node on US, but it did not succeed. In another patient, although the Hydromark clip was clearly visualized on US and a fluorescence technique was used for excision, the clipped node was not included in the removed tissue. Further search was difficult since the ICG dye had spread over the surrounding tissues. In the third patient, the wire was fractured and missed the clipped node. In this study, although the wire was fractured or displaced intraoperatively in three patients (12.5%, 3/24), the clipped node was successfully removed in two patients by identifying the pyoctanin-stained lymph node or by detecting the Twirl clip under repeat IOUS. Since all three failed cases of TLNB were included in Group A, the clipped node was retrieved from the dissected axillary lymph nodes.

Pathologic Evaluation of Cases Involving TAD

TLNB removed a median of one clipped node (range, 1-4). Additional nodes were removed together along with the clipped node in six patients. The clipped node with SLN tracer uptake was observed in 61.5% of the patients (16/26, Table 3). In the five patients who required two clips, SLN tracer uptake was observed in both nodes in two patients and in one node in three patients. Metastases remained in the clipped node in a total of 16 patients. One false-negative event was noted in the TLNB cases, resulting in an FNR of 6.3% (1/16). The median number of SLNs removed from the 26 patients was two (range, 0-7). SLN tracer uptake was observed in a median of two nodes (range, 0-5). Although SLNB identified metastases in five patients with median of one positive node (range, 1-3), it did not pick up the false-negative node of TLNB. Palpable nodes removed from six patients indicated one positive node in three patients whose initial imaging indicated three or more nodal metastases. Among these three patients, two patients with the ER+/HER2- subtype also showed metastases in the clipped node, and one patient with the ER+/HER2+ subtype showed metastases only in the palpable node, resulting in a false-negative TLNB result. The median number of lymph nodes removed with TAD among the 26 patients was four (range, 2-10). The FNR of TAD was 0% (0/16).