Since the U.S. Food and Drug Administration approved the application of a robotic surgical system in adult and pediatric surgery in 2000, robotic surgery has made great strides in several surgical treatments, including cancer surgery [1]. In the field of oncological surgery, robotic surgical systems are currently most commonly used in the urology area, where there is narrow retroperitoneal space access. Robotic surgical systems are also commonly applied in general, gynecological, thoracic, and head and neck surgeries. Since the first use of robotic surgical systems for breast surgery by Toesca et al. in 2015, robot-assisted nipple-sparing mastectomy (RNSM) with immediate implant reconstruction has been recognized in the field of breast oncologic surgery [2, 3].

A robotic surgical system generally requires one camera port and two or three robotic instrument arms. The camera shows a three-dimensional field of view. The robotic instrument arm is an ergonomic device that can be bent flexibly, like a human joint, to gain access over a wide range in multiple directions. Due to these characteristics, the learning curve for robotic surgery is shorter than that for other surgical methods using laparoscopy or endoscopy. Moreover, minimally invasive surgery is possible through a small incision [1, 4, 5]. In addition, robotic surgery results in less intraoperative bleeding compared to conventional surgeries [6,7,8].

The classification of cancer stage is an important factor in evaluating the clinical patterns, predicting the prognosis, and determining the treatment of cancer [9]. For operable patients without systemic metastasis at initial diagnosis, the size of cancer (T stage) and stage of lymph node metastasis (N stage) are considered when planning the treatment. In breast cancer, surgery for node staging is performed in the ipsilateral axilla, since most lymphatic drainage related to breast cancer typically runs directly to the axilla [10]. Patients who are clinically negative for axillary lymph node metastasis are managed first with sentinel lymph node biopsy (SLNB), and patients with axillary lymph node metastasis consider undergoing axillary lymph node dissection (ALND). Although several prospective studies are investigating the possibility of omitting ALND and complementary adjuvant treatment [11, 12], ALND is still needed for breast cancer with axillary lymph node metastasis in some patients undergoing total mastectomy, who are also the main subjects of breast surgery involving robotic surgical systems [13].

A robotic surgical system can be used for lymph node dissection in various cancer surgeries for the evaluation of lymph node stage. Several studies have reported on surgery using robotic surgical systems for retroperitoneal lymph node dissection [14, 15], intra-abdominal lymph node dissection [16], and lymph node dissection in extra-abdominal areas, such as the neck or inguinal area [17, 18]. According to these reports, it was possible to distinguish microstructures, such as vessels and nerves around lymph nodes, better with a zoomed robotic system operation field than in open surgery. Moreover, the flexibility of the robotic instrument arms enables precise surgery, ensuring technical safety and minimally invasive surgery. However, there has been no report of robotic axillary lymph node dissection (RALND) in RNSM with immediate reconstruction using a single incision, except for our previous report [19].

During the conventional axillary lymph node dissection (CALND) procedure, in fact, surgeons and assistants need to pull on the retractors with full force to lift up the pectoralis major muscle, which increases the likelihood of damaging many microstructures and usually requires extension of the incision to provide sufficient visibility. To decrease these disadvantages, RALND could be a good solution.

In the current study, we compared the perioperative outcomes, including feasibility and safety, of RALND and CALND.

Methods

Patient selection

Thirty-two Korean patients underwent ALND in their robot-assisted nipple-sparing mastectomy (RNSM) by two surgeons from December 2017 to January 2021 at Severance Hospital, Seoul, Korea. Patients’ medical records were retrospectively reviewed. There is an international expert consensus that recommends robotic surgery for patients with early breast cancer with small tumors or for patients undergoing prophylactic mastectomy [20], and the recommendation was followed in this study.

Clinicopathological features, including age, body mass index (BMI), SLNB surgery, multiplicity (multifocality or multicentricity), estrogen receptor status, progesterone receptor status, HER2/SISH status, Ki-67 index, histological grade, tumor size, TNM stage, implementation of neoadjuvant chemotherapy, and adjuvant treatments, such as chemotherapy or radiation treatment, were reviewed. Surgical outcomes, including the number of retrieved and positive sentinel lymph nodes, retrieved and positive axillary lymph nodes, intraoperative bleeding amount, final incision size, and postoperative axillary complications, including postoperative axillary bleeding, lymphedema, and shoulder joint pain or disabilities, were also analyzed. Postoperative axillary bleeding was defined as a case requiring reoperation due to insufficient axillary vascular ligation or axillary vascular damage during the primary surgery. Lymphedema of the upper extremity after breast cancer surgery was defined as a clinical difference ≥ 20 mm in the left and right diameters in any part of the arm [21, 22]. Shoulder joint pain or disabilities were defined as cases requiring physical therapy or bursa injection in the department of rehabilitation medicine due to causes such as shoulder pain, limited range of motion, and adhesions after breast cancer surgery.

A multi-port robotic surgical system, the Da Vinci Xi system, or a single-port robotic surgical system, the Da Vinci SP system (Intuitive Surgical, Sunnyvale, CA), were used for RNSM. Only the Da Vinci SP system was used for RALND, as the Da Vinci Xi system experienced many collisions between the robot arms during surgery in a narrow space. Since December 2019, ALND using the SP system were performed according to the results of SLNB using intraoperative frozen sections [19].

This study was approved by the Institutional Review Board (IRB) of Yonsei University College of Medicine (IRB No. 4-2021-0425). Informed consent was waived due to the retrospective nature of the study.

Surgical technique

The surgical technique used was described in detail in a previous study [19]. In brief, RNSM was first performed through a linear incision at the level of the nipple on the mid-axillary line. Three patients who were previously treated with neoadjuvant chemotherapy with definite axillary lymph node metastasis underwent ALND immediately. The remaining patients were administered radioisotope injection for SLNB, and sequential ALND was performed according to the results of frozen biopsy for SLNB.

To obtain the surgical field of view, conventional retractors were used in CALND, while CO2 gas inflation was used in RALND. In CALND, most of the ALND procedures were performed with extension of the initial incision for better visualization. However, using the robotic surgical system, ALND was performed without additional extension of the incision. After finishing mastectomy and removing the breast specimen, the direction of the robotic instrument arms was adjusted to a vertical position, toward the patient’s head (Fig. 1). When the camera is inserted, the axillary basin and tissue should be seen at the center of the monitor. The edge of pectoralis major was located vertically at the affected right or left side. Then, axillary node dissection was performed by three robotic instrument arms, each of which had different functions: dissection, traction, and counter-traction.

Fig. 1
figure 1

Axillary structures and surgical access using a single-port robotic surgical system

Full size image

SLNB or ALND are performed through open incision before docking the robot to the breast [4]. If the SLNB is positive for axillary lymph node metastasis, CALND can be manually performed using the existing axillary or lateral incision, with or without extension, after robot-assisted mastectomy is completed.

The dissection area of ALND was the same in both CALND and RALND. Dissection of axillary level I and II lymph nodes underneath the pectoralis minor muscle was performed with careful attention to preserve the axillary vein, thoracodorsal nerve, long thoracic nerve, medial pectoralis nerve, and intercostobrachial nerves as much as possible.

Pathologic evaluations

TNM staging of breast cancer was determined according to the 8th American Joint Committee on Cancer Staging system. The expression status of estrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor 2 (HER2), and Ki-67 levels was examined by immunohistochemistry (IHC) staining. The definition of ER and PR positivity was 1% or more tumor nuclei staining by IHC, according to the American Society of Clinical Oncology/College of American Pathologists Guideline (ASCO/CAP) recommendation [23]. HER2 status was assessed by the Hercep test (Polyclonal, dilution 1:1500; DAKO, Produktionsvej, Glostrup, Denmark), and additional silver in situ hybridization (SISH; INFORM HER2 Dual ISH DNA Probe Cocktail Assay, Ventana Medical Systems, Oro Valley, AZ, USA) was applied when HER2 expression was equivocal in IHC. HER2 status depending on IHC and SISH was defined according to the College of American Pathologists Guidelines. IHC 0 or IHC 1 + of HER2 expression was considered as HER2-negative, and IHC 3 + or IHC 2 + /ISH + of HER2 expression was defined as HER2-positive [24]. Ki-67 indicates the proliferative index and Ki-67 < 14% was considered as low staining in IHC. Cancer subtypes by IHC staining were classified according to a previous study [25].

Statistical analysis

Continuous variables were analyzed using Mann‒Whitney test and reported as the median value with the minimum and maximum range. The comparison of categorical variables was evaluated using Fisher’s exact test. All statistical analyses were performed using SPSS (version 25.0, IBM Software, IBM, Armonk, NY, USA). A p value < 0.05 was considered statistically significant.

Results

Among 32 patients who underwent RNSM with ALND since December 2017, 11 patients underwent RALND (Fig. 2). Twenty-one CALND procedures were performed by two experienced surgeons, and the 11 RALND cases were performed by one of the two aforementioned surgeons.

Fig. 2
figure 2

Flowchart of patient selection

Full size image

The median age and body mass index of the total patients were 43.1 years (20‒59 years) and 22.3 kg/m2 (17.6‒28.2 kg/m2), respectively. Clinicopathologic characteristics of patients in the two groups are shown in Table 1. There were no statistically significant differences between the two groups. The median age of patients were 43 years in both the CALND (range 20‒57 years) and RALND (range 31‒59 years) groups. Three patients in the CALND group who did not undergo SLNB had previously been shown to have axillary lymph node metastasis on fine needle aspiration biopsy before neoadjuvant chemotherapy, and were preoperatively planned to undergo ALND regardless of the result of SLNB. Eighteen patients in the CALND group and all patients in the RALND group underwent ALND due to positive intraoperative frozen SLNB results. Half of all patients had luminal A-like type; among them, 37.5% had luminal B-like, 6.3% were HER2-positive, and 6.3% had triple-negative breast cancer.

Table 1 Clinicopathologic characteristics of CALND and RALND
Full size table

Two CALND patients and one RALND patient with TNM stage 0 had pathologic complete remission after neoadjuvant chemotherapy. Final pathologic results of the three patients were ductal carcinoma in situ with total tumor regression of invasive lesions after neoadjuvant chemotherapy. Prior to neoadjuvant chemotherapy, mastectomy and axillary lymph node surgery were planned for these patients. The proportion of patients diagnosed with stage III was higher in the CALND group than in the RALND group (23.8% in CALND vs. 9.1% in RALND). In addition, a high stage in breast cancer is generally due to advanced axillary lymph node metastasis, rather than tumor size. Six patients with stage III cancer all had T2 disease (≤ 5 cm), but two of them were N2 and four of them were N3 disease. The CALND group received more adjuvant chemotherapy compared to the RALND group (47.6% in CALND vs. 36.4% in RALND, p = 0.712).

Surgical outcomes, including the number of retrieved axillary lymph nodes, are shown in Table 2. A median of two and three sentinel lymph nodes was retrieved in the CALND and RALND groups, respectively, without statistically significant difference (p = 0.223). There was no significant difference in the median number of total resected axillary lymph nodes between the CALND and RALND groups (13 in CALND vs. 15 in RALND, p = 0.696). Median time to ALND was longer in the RALND group than in the CALND group, but was not statistically significant (35 min in CALND vs. 47 min in RALND, p = 0.126).

Table 2 Surgical outcomes of CALND and RALND
Full size table

The final incision size in RALND tended to be smaller than that in CALND (median 40 mm, range 30‒60 mm in RALND group vs. median 50 mm, range 35–60 mm in CALND group, p = 0.081) in Fig. 3. A higher proportion of patients in the RALND group had a final incision size ≤ 40 mm compared to patients in the CALND group (36.4% in CALND vs. 63.6% in RALND, p = 0.020), as shown in Table 2. There were no statistically significant differences in postoperative axillary complications. There was no case of postoperative axillary bleeding. There were a total of three cases of lymphedema, with no statistical difference in either the CALND group or RALND group (p = 0.968). The median follow-up period of lymphedema was 19 months. Shoulder joint pain or disabilities occurred in five cases in the CALND group and three cases in the RALND group, and there was no statistical difference between the two groups (p = 0.830).

Fig. 3
figure 3

Comparative analysis between CALND and RALND

Full size image

Discussion

The present study demonstrated the feasibility and safety of RALND compared to CALND during our early experience of RALND. The surgical effectiveness of RALND was not significantly different from that of CALND. RALND allows the retrieval of a similar number of axillary lymph nodes as CALND. RALND uses CO2 gas inflation rather than a mechanical retractor to secure axillary visibility, which was sufficient to expose the lymph nodes at axillary level I and II. The operating space was expanded, followed by tissue resection without retractors, as the CO2 gas was continuously inflated and maintained at a constant pressure in the axillary space during surgery. There was no need for retraction by assistants; thus, it may help to reduce surgical fatigue for surgeons and assistants. In addition, RNSM with RALND could be completed with a small initial incision (≤ 4 cm) without additional extension.

There are advantages to robotic surgery in various types of lymph node dissections in a small narrow working space. Kang et al. performed robot-assisted modified radical neck dissection for 33 patients with low-risk, well-differentiated thyroid cancer [17]. Robot surgery had several benefits in neck surgery, with shorter incision scar related to good cosmetic effect, fewer muscle deformities in the neck, and more stability in motions due to tremor-free robot facility compared to conventional open surgery. Mattei et al. referred to the monoblock technique by robotic surgery [26]. Pelvic lymph node dissection (PLND) is essential for node stating in high-risk prostate cancer, and extended PLND is related to a more precise N stage. A super-extended PLND is possible using a robot surgical system with minimal invasive surgery, and has a better overview of the surgical field than achieved with open surgery.

Several complications were previously reported in surgery for lymph node dissection of the extra-abdominal area. Postoperative infection, lymphocele, and prolonged lymphorrhea have been reported in inguinal lymph node dissection of melanoma or vulvar cancer [27, 28]. Minor complications, such as chyle leakage, vocal cord palsy, and minor lymphatic injuries, have been reported in neck lymph node dissection [17, 29, 30]. Careless detachment could be a reason for these complications, but they could also be caused by the lack of a tactile sense of the robot. In one case in this study, there was minor tearing of the superficial branch of axillary vein due to this lack of tactile sensation. Since it was difficult to maintain the surgical field, we opted for a conversion to open surgery to control bleeding by further extending the skin incision. When uncontrolled bleeding occurs, open conversion can be easily used to control the bleeding. More careful handling is required to avoid vessel injuries while using the robotic surgical system.

Two common incisions are used for RNSM surgery. One is the anterior line of the axillary region, and the other is the mid-axillary line of the NAC level [4, 31, 32]. We performed an incision in the anterior line of the axillary region for cadaveric attempts and early surgeries. As the number of surgeries increases, the incision is mainly placed in the mid-axillary line at the NAC level. An incision in the anterior line of axillary region is convenient for CALND. However, when a woman’s bra is worn, a scar in the axillary region is easily revealed. In addition, collisions of the robotic arms frequently occur during RNSM. In our experience, performing RNSM with an incision in the mid-axillary line is good for the convenience of robotic surgery and the cosmetic benefits of scarring. However, it is very difficult to obtain the surgical field of view of CALND without extended incision in the situation of performing ALND using an incision in the mid-axillary line. Axillary surgery can be performed with RALND without extension of an incision in the mid-axillary line.

There are several advantages to RALND. We compared the perioperative results of CANLD with those of RALND, and confirmed that RLAND was not inferior to CALND in terms of perioperative outcomes, including the ability to obtain axillary lymph nodes. Axillary lymph node level II can be sufficiently dissected through a magnified view of the robotic surgical system, without the need for mechanical retraction. Essential structures that need to be preserved, such as the axillary vein and long thoracic or thoracodorsal nerve, can be easily identified through the robotic surgical system. This approach is excellent for improving cosmetic effects and quality of life for patients by minimizing the incision length. However, there are also some disadvantages to the RALND procedure. A sufficient amount of education or training is required to acquire the necessary surgical skill. Therefore, appropriate training programs should be developed. Also, the operation time of RALND is slightly longer than that of CALND. If bleeding control is difficult during early experience, open conversion may be considered. However, this can be overcome by accumulating experience.

This study had several limitations. Since the time spent on CALND could not be investigated, it was difficult to accurately compare the operation time between the two groups. For RALND, videos were recorded during robotic surgery, and an accurate operation time could be measured through video analysis. However, there was no protocol for separately recording the operation time for CALND after breast surgery even if other indicators, such as the time period between the docking out time from the breast and the start time of breast reconstruction surgery, recorded in the operating room were considered as a suboptimal method. Data on the operation time were missing more often for CALND than for RALND. Moreover, long-term results were lacking in two respects. First, the evaluation of lymphedema at long-term follow-up after surgery was limited. Second, additional studies on oncological survival outcomes are needed. Since the present study investigated our initial experience of RALND and focused only on the short-term results, further studies are required.

Conclusion

RALND is a technique that clearly has advantages compared to CALND. RALND may be a feasible and safe technique if performed by well-trained surgeons in a well-equipped hospital.