Comparison of the indocyanine green dye method versus the combined method of indigo carmine blue dye with indocyanine green fluorescence imaging for sentinel lymph node biopsy in breast conservative therapy for stage ≤IIA breast cancer

Patients with early breast cancer who underwent Breast Conservative Therapy (BCT) between December 2003 and February 2017 were reviewed retrospectively. Indication for BCT was based on the guidelines of BCT issued by the Japanese Breast Cancer Society [24]. Patients with primary tumor size > 3 cm, multicentric breast cancer, and tumors fixed to the skin or muscle were not subject to BCT. Our analysis was restricted to those who had an indication for SLN biopsy (i.e., ≤T2 breast cancer with clinically negative axillary lymph nodes) based on preoperative examinations (mammography, ultrasound, enhanced multi-detector computed tomography, and enhanced magnetic resonance imaging, etc). Furthermore, in this study, only cases that underwent BCT with separate incisions (incision on the breast for tumor resection and a small axillar incision for SLN biopsy) were reviewed. Patients who had neoadjuvant chemotherapy, history of allergy to iodine, or history of prior breast and axillary surgery were excluded.

The main method for identification of SLNs for SLN biopsy was the ICG-D method until November 2011, while the CMF method was introduced in December 2011 after approval by the ethics committee of East Saitama General Hospital. There were no overlaps in the period in which ICG-D and CMF were performed. All resected SLNs were pathologically examined by hematoxylin and eosin stain. Pathological examinations were conducted by pathologists with more than 20 years’ experience. Until April 2008, all pathological diagnoses of SLN biopsy specimens were performed on permanent sections. In positive SLN cases, an additional axillary lymph node dissection was performed at a later date. After April 2008, intra-operative examination of SLN biopsy specimens became available, and was performed with one or more frozen sections. Positive SLNs in the frozen section resulted in additional axillary lymph node dissection on-spot. Additional axillary dissection of the lymph nodes was performed on a later date if the frozen section was negative but the permanent section was positive. This was not conducted if the diagnosis according to the permanent section was micrometastasis. For both ICG-D and CMF, axillary lymph node dissection was added when the SLN could not be identified. Post-operatively, a histopathologic examination was also performed on the resected specimens to confirm the diagnosis. Further examinations such as estrogen and progesterone receptor and herceptine score (HER2) were also conducted. If the herceptine score was + 2, the Fish method was added, and a final HER2-positive diagnosis was made when confirmed to be FISH positive. With the exception of 4 ICG-D cases, all operations were performed by the same operator (NT) who already had more than 14 years of experience in breast surgery when the first ICG-D in this study was performed (2004). In the 4 exceptional cases, the main surgeon had more than 10 years of experience while NT assisted the operation.

All patients in the study underwent tangential field irradiation of the whole breast (50Gy), with a dose per fraction of 2Gy in 5 weeks post-operatively. If margins were close to the tumor edge (≤5 mm), a boost of 10Gy was added. Indication for post-operative chemotherapy and hormone therapy was based on the St. Gallen International Expert Consensus at the time of the operation. Detailed information on all surgical procedures were provided to all patients and family members preoperatively and written informed consent was obtained.

The patient was placed in a supine position on the operating table and a small pillow was placed on the back of the operating side of the patient. The arm on the surgical side was extended perpendicular to the body axis. After the induction of general anesthesia and sterilization of the operation site, ICG 10 mg, diluted with 2 ml of distilled water, was injected into the sub-areolar region after induction of general anesthesia. The sub-areolar region was compressed to assess lymphatic drainage and a 3-4 cm incision was made at the axillar region and the green-stained SLNs were resected under shadowless surgical light conditions. After SLN biopsy, partial resection of the mammary gland was performed through a separate incision in succession.

After the induction of general anesthesia and sterilization of the operation site, a combination of 1 ml ICG (5 mg) and 1-2 ml indigo carmine (4-8 mg) was injected into the sub-areolar region (Fig. 1a). Fluorescence images were obtained using PDE, and subcutaneous lymphatic channels were detected over the skin usually within one or 2 minutes (Fig. 1b). If the lymphatic flow to the SLNs was not visible, the sub-areolar region was compressed to assess lymphatic drainage. The subcutaneous lymphatic channels were marked with a marker over the skin based on the PDE image (Fig. 1c). A 1.5-2 cm incision was made near the point of disappearance of the fluorescence in the axilla (Fig. 1d). Under shadowless surgical light conditions, the subcutaneous fat tissues were dissected, and the stained lymphatic channels were identified and carefully dissected, and traced until the first-drained lymph node was identified. Usually, one or more stained lymph nodes were identified (Fig. 1e) and resected by direct visual inspection. If the blue stained SLNs could not be identified, the operation field was directly inspected using PDE and the fluorescent lymph nodes were identified. Figure 1f illustrates a case where the blue stained lymph node could not be identified by direct visual inspection. However, using the PDE, we were able to easily identify the fluorescence of SLNs on the monitor. Fluorescence imaging was used for post-operative SLN confirmation as well. After SLN biopsy, partial resection of the mammary gland was performed through a separate incision.

Information on patient age, body mass index (BMI) (kg/m²), tumor characteristics such as location, T factor, N factor, stage, presence of recurrence, and pathological type, as well as time to detection of SLNs, presence or absence of SLN identification, and number of resected SLNs were also collected. Tumor location was defined as: A,upper inner quadrant; B, lower inner quadrant; C, upper outer quadrant; D, lower outer quadrant. If the primarily involved site could not be determined due to equal extent of involvement in two or more areas, tumor location was expressed as the predominant area that occupied the resection area [25].

The statistical analysis was done with StatMate IV (ATMS Co., Ltd., Tokyo, Japan) and Stata 14.1 (Stata Corp LP, College station, Texas, USA). We compared the baseline characteristics of the sample (age, BMI, and tumor characteristics), and intra- or post-operative findings (time to detection of SLNs, number of resected SLNs, pathological diagnosis of breast cancer, presence of estrogen and progesterone receptors, herceptine score) between patients who underwent ICG-D and CMF. The detection rates of SLNs in the two methods, including that of a subsample of SLN positive individuals based on intra-operative diagnosis, were also compared. We also examined whether the identification rate between the two methods differ by age and BMI by conducting stratified analyses. Student’s t-tests were used for the comparison of continuous variables. The proportion between the two groups were compared with Chi-squared tests unless it included cells ≤5, in which case Fisher’s exact tests were used. P < 0.05 was considered statistically significant.

Ethical approval to conduct this study was obtained from the ethics committee at the East Saitama General Hospital.

CMF was associated with a significantly higher SLN detection rate compared to ICG-D (96.4% vs. 83.7%; p = 0.003) (Table 3). There were only six CMF cases (5.5%) which required intra-operative PDE to detect SLNs. The SLN detection rate in CMF when considering these six individuals to not have had their SLNs identified was 90.9% (100/110). Among the 31 patients with confirmed SLN metastasis, ICG-D was only able to detect SLNs in 93.3% of the cases whereas CMF had a 100% detection rate (data shown only in text). For both ICG-D and CMF, axillary lymph node dissection was added when the SLN could not be identified. There were only 19 such cases (ICG-D 15 cases, CMF 4 cases). In terms of the pathological findings, age, and BMI, there were no significant difference between these 19 individuals and the rest of the sample (data shown only in text).

In patients aged ≥60 years and in those with BMI ≥ 25 kg/m², the SLN detection rate was particularly high in CMF as compared to ICG-D (Table 4). Specifically, in patients < 60 years of age, the SLN detection rate of CMF (vs. ICG-D) was only 1.04 times higher (94.2% vs. 90.6%; p = 0.716) while among the older age group, it was 1.32 times higher (98.3% vs. 74.3%; p < 0.001). In the case of BMI, the SLN detection rate was only 1.06 times higher (98.7% vs. 92.7%; p = 0.096) among those with BMI < 25 kg/m² while this figure rose to 1.55 (90.3% vs. 58.3%; p = 0.009) among those with BMI ≥ 25 kg/m².