Robotic thyroidectomy using a bilateral axillo-breast approach (BABA RT), first introduced in 2008, is a popular method of thyroid surgery [1, 2]. BABA RT combines the da Vinci robot surgical system with the BABA surgical method, requiring four small incisions at the circumareolar and skin creases of the axilla [3]. Its advantages include an operative approach similar to that of conventional thyroidectomy, while not leaving scars on the neck [4, 5]. BABA RT has shown oncological safety and surgical completeness with cosmetic advantages in patients with papillary thyroid carcinoma (PTC) [6, 7]. To date, however, BABA RT has seldom been performed in PTC patients with extensive neck lymph node metastasis (LNM).

BABA RT is safe in patients with advanced stage thyroid disease [7]. Moreover, robot-assisted modified radical neck dissection (MRND) was shown to be feasible and safe in patients with PTC and lateral LNM [8], resulting in the performance of RT in selected patients with PTC and lateral LNM [9]. Although robot-assisted BABA MRND (robotic BABA MRND) was found to be as successful as open MRND, few PTC patients to date have been surgically treated by this method [10]. Therefore, there is not much evidence for MRND and procedures are not standardized.

In the present study, we review cases of robotic BABA MRND and report the indications, operation procedures, and surgical outcomes of the patients who underwent robotic BABA MRND with PTC and lateral LNM at a single institution.

Materials and methods

Patients

The medical records of patients with PTC and lateral LNM who underwent robotic BABA MRND at Seoul National University Hospital from March 2010 to July 2016 were retrospectively reviewed. Lateral LNM was confirmed preoperatively in all patients by fine needle aspiration with wash out thyroglobulin (Tg) test. Patients who had an extra-nodal extension, and vessel invasion and/or vagus nerve invasion of lateral neck node were contraindication of robotic BABA MRND. This study was approved by the Institutional Review Board of Seoul National University Hospital (H-1609-109-794).

Procedures of making a flap and performing BABA RT

The patient was placed in supine position with neck extension [11]. After aseptic draping, a dilute epinephrine solution (1:200,000) was injected into the subcutaneous layer of the anterior chest to lift the skin and subcutaneous layer, a procedure called hydro dissection. Incisions were made in the skin at the circumareolar and axilla, followed by dissection of the subcutaneous area with a vascular tunneler. The da Vinci robot was docked through an inserted trocar. BABA RT with bilateral central lymph node (LN) dissection was performed prior to robotic BABA MRND.

Procedures of robotic BABA MRND

The surgical procedure for robotic BABA MRND is shown in Fig. 1. After BABA total thyroidectomy, the robotic instrument was docked at the location of lateral LNM. PK dissecting forceps were docked in the ipsilateral axillary trocar, and Prograsp forceps were docked in the contralateral axillar trocar (Fig. 1A, B). The sternocleidomastoid (SCM) muscle was separated from the sternohyoid and sternothyroid muscles using a permanent cautery hook. The sternohyoid and sternothyroid muscles were retracted in the medial direction using Prograsp forceps, and the SCM muscle was retracted laterally using PK dissecting forceps. The medial border of the lateral LN (right level III) was dissected from the internal jugular vein and the lateral border of right level III was dissected from the SCM muscle using a harmonic scalpel or permanent cautery hook. Prograsp forceps were carefully passed through the SCM muscle from level IV to level V. The entire lateral border of the SCM muscle was dissected with a permanent cautery hook while lifting the SCM muscle with Prograsp forceps.

Fig. 1
figure 1

The procedure of robotic-assisted modified radical neck dissection with a bilateral axillo-breast approach (robotic BABA MRND). Fine needle aspiration of 37-year-old woman indicated papillary thyroid carcinoma in the right thyroid nodule and metastatic papillary thyroid carcinoma in right level III. A and B preparing for robotic BABA MRND. C and D Dissection of lymph node in level V while preserving the spinal accessory nerve. E and F Pull the sternocleidomastoid (SCM) muscle by polydioxanone (PDS) suture. G and H Dissection of level IV was started from the junction between the internal jugular and right subclavian vein while preserving the phrenic nerve I and J Dissection of level III K and L Dissection of level II up to right posterior belly of digastric muscle

Full size image

Level V LNs were dissected with medial traction of Prograsp forceps, while preserving the brachial plexus and spinal accessory nerve with a permanent cautery hook (Fig. 1C, D). Both robotic instruments were utilized for tenting and detailed observation of level V, followed by two laparoscopic clippings of level V. A 1 -0 polydioxanone (PDS) suture was passed through the left areola trocar, and sutures anchored by the PDS were used by the assistant to pull the SCM muscle (Fig. 1E, F). Lifting of the SCM muscle ensures a sufficient workspace for LN dissection, which proceeded from level IV to level II.

To dissect level IV, the omohyoid muscle was lifted with PK dissecting forceps. The bottom of level IV and the junction between the internal jugular and right subclavian veins could be clearly observed with proper motions of the PK dissecting forceps and the Prograsp forceps (Fig. 1G, H). To dissect level IV in an upward direction, the omohyoid muscle was pulled down by the Prograsp forceps. Careful dissection of level IV was required to prepare the phrenic nerve, with level IV dissection followed by continuous dissection of the medial borders of level III (Fig. 1I, J) and level II, the latter while preserving the right spinal accessory nerve. The right posterior belly of the digastric muscle was exposed in level II (Fig. 1K, L). Clipped level V LNs were moved to level IV through the bottom of the SCM. The lateral borders of levels III and II were dissected continuously. LNs from level IV to level II were carefully dissected while preserving the phrenic nerve. Dissected LNs were removed with lap bags, and sutures anchored by the PDS were released and removed through the areolar trocar. Midline closure with Vicryl sutures was used for midline closure, followed by placement of a Jackson-Pratt drain.

Postoperative management and outpatient follow-up

Patients who underwent robotic BABA MRND started food intake 6 h after surgery. The amount of drainage fluid was regularly checked during each patient’s hospital stay, and the drain was removed on the morning the patient was discharged, usually on postoperative day 3. The final pathology report was available 2 weeks after surgery. All patients were encouraged to undergo radioactive iodine (RAI) treatment, based on published guidelines for radioactive treatment of differentiated thyroid cancer [12]. RAI dose was based on pathologic aggressiveness. Whole body scintigraphy (WBS) and serum stimulated thyroglobulin (sTg) tests were performed to check remnant thyroid tissue or detect distant metastases after RAI treatment. Tumor recurrence was monitored regularly by ultrasonography and by measuring sTg concentrations.

Results

The 15 patients identified included 14 women (93.3%) and one man (6.7%), of mean age 37.1 ± 9.3 years and mean body mass index 22.1 ± 3.3 kg/m2 (Table 1). Ten (66.7%) and five (33.3%) patients had main tumors on the right and left sides, respectively, with a mean tumor size of 1.0 ± 0.6 cm. Eight (53.3%) and seven (46.7%) patients had right and left sided MRNDs, respectively, whereas five (33.3%), two (13.3%), and eight (53.3%) patients were classified as having T1, T2, and T3 tumors, respectively. Four patients (26.7%) had bilateral tumors.

Table 1 Clinicopathological characteristics of patients with papillary thyroid carcinoma and lateral lymph node metastasis
Full size table

Surgical outcomes are shown in Table 2. The mean operation time was 272.7 ± 33.8 min. The mean numbers of central retrieved and central metastatic LNs were 8.1 ± 6.8 and 2.9 ± 4.2, respectively, and the mean numbers of lateral retrieved and lateral metastatic LNs were 20.7 ± 7.2 and 5.3 ± 4.4, respectively. Mean postoperative hospital stay was 3.1 ± 0.4 days, and the mean total amount of drainage was 348.8 ± 163.7 mL. Mean follow-up period was 18.7 ± 19.1 months. The median sTg concentration after the first RAI treatment was 0.8 ng/mL (range 0.1–36.5 ng/mL), with 8 (53.3%) and 14 (93.3%) of these patients having stimulated Tg concentrations below 1 ng/mL and below 2 ng/mL, respectively.

Table 2 Surgical outcomes of patients with papillary thyroid carcinoma and lateral lymph node metastasis who underwent robotic-assisted modified radical neck dissection with bilateral axillo-breast approach
Full size table

Table 3 shows surgical outcomes after robotic BABA MRND. The rates of transient and permanent hypocalcemia were 46.7 and 0%, respectively, and the rates of transient and permanent vocal cord palsy were 6.7 and 0%, respectively. There were no reports of chyle leakage, postoperative bleeding, and wound infection. One patient experienced Horner’s syndrome, with the symptoms spontaneously regressing.

Table 3 Surgical complications in patients with papillary thyroid carcinoma and lateral lymph node metastasis who underwent bilateral axillo-breast approach robotic thyroidectomy with modified radical neck dissection
Full size table

Discussion

To date, the surgical outcomes of robotic BABA MRND in patients with PTC and lateral LNM have been unclear. This study reports our initial experience and surgical outcomes of robotic BABA MRND in 15 patients with PTC and lateral LNM.

The main advantage of BABA RT is the ability of the surgeon to see the operation field symmetrically, as in open surgery [13]. This advantage was equally applicable to robotic BABA MRND. Most surgeons stand at the right side of the patient during open MRND and look upward. The operative view during robotic BABA MRND is identical because the camera is inserted through the right areola of the patient. The similarities between open surgery and BABA MRND enable the surgeon to more easily adapt to the latter [11, 14].

As in open MRND, the first surgical procedure during lateral neck dissection in robotic BABA MRND is dissection of structures adjacent to the SCM muscle and exposure of LNs, thereby widening the surgical field. Compared with open MRND, in which the adjacent area is exposed by a large incision and an army navy retractor, enabling dissection of the sides of the SCM muscle, BABA MRND uses different instruments, including PK dissecting forceps and Prograsp forceps [10]. The movements of both these instruments are precise and sufficient for traction and counter-traction of the SCM muscle. The endowrist function of these instruments and the magnification by the camera enable the performance of complex tasks in lateral compartment areas [15]. In addition, the PDS is used to lift and anchor the dissected SCM muscle [10], securing a larger workspace. This surgical technique has the advantage of causing little or no injury to the skin, with operative visibility similar to that of open MRND.

The surgical procedure and the extent of the dissected area of robotic BABA MRND are similar to those of open MRND. Dissection of the lateral LN is easier when starting from level V, because the level V space is narrow, whereas the space between levels II and IV is relatively wide. The location of LNs was identified by clipping level V twice with laparoscopic clippers, which transferred level V LNs to level IV through the SCM muscle. During LNM dissection from level IV to level II, MRND can be performed safely, while preserving the SCM muscle, spinal accessory nerve, internal jugular vein, phrenic nerve, vagus nerve, and transverse cervical artery. The dissected LNs are removed using the lap bag after surgery and separated based on clipping sites.

Following robotic BABA MRND, patients underwent RAI and were monitored by measuring their concentrations of sTg. sTg is a good indicator of remnant thyroid tissues, thereby confirming the completeness of surgery [16]. A low sTg concentration indicates complete thyroid removal and oncological safety. A study of 1026 PTC patients who underwent BABA RT showed that the mean sTg level was 0.4 ng/mL, with 65.1% of these patients having sTg concentrations below 1 ng/mL [3]. In this study of patients with PTC and LNM, we found that 53.3 and 93.3% of patients had sTg concentrations below 1 and 2 ng/mL, respectively. These findings show that the oncological completeness of robotic BABA MRND in selected patients was acceptable (Fig. 2A, B). Furthermore, the cosmetic result of robotic BABA MRND was excellent (Fig. 2C).

Fig. 2
figure 2

Oncological and cosmetic results of robotic BABA MRND. A 29-year-old woman underwent robotic BABA MRND for papillary thyroid carcinoma with lymph node metastasis in lateral neck compartment. The patient underwent ultrasonography and whole body scintigraphy after 2 cycle of radioactive iodine treatments. A Whole body scintigraphy, B ultrasonography of lateral neck compartments, C cosmetic results 2 weeks after operation

Full size image

The rates of temporary hypocalcemia and temporary vocal cord palsy in the present study were 46.7 and 6.7%, respectively. This rate of temporary hypocalcemia was in the range reported for open MRND (1.6–50%) [17], and this rate of temporary vocal cord palsy was no higher than that of RT (up to 20%) [16]. One patient experienced postoperative Horner’s syndrome, with symptoms improving 6 months after surgery without treatment or intervention. Brachial plexus injuries, which were reported in studies of RT using a transaxillary approach, were not observed in the present study [18,19,20].

This study had several limitations, including its retrospective design, which may be associated with selection bias. Second, the number of included patients was small, which may have increased the incidence of complications. Prospective, large-scale observational studies are needed to determine the prevalence of complications, such as temporary hypocalcemia, temporary vocal cord palsy, and Horner’s syndrome, in patients undergoing BABA MRND for PTC and lateral LNM.

In conclusion, this study showed that the surgical procedures for robotic BABA MRND are similar to those for open MRND, with an acceptable complication rate, in selected patients with PTC and lateral LNM. Robotic BABA MRND may be safe and effective and an excellent surgical option in selected PTC patients with lateral LNM.