Well-differentiated thyroid cancer has been the most prevalent cancer in Korea since 2009, and its incidence is increasing worldwide [1–3]. Most cases are papillary thyroid cancers, which most frequently occur in young woman. Because the prognosis and survival rate of thyroid cancer are excellent, cosmesis plays an important role in the surgical strategy of some thyroid cancer patients. Asian patients are at higher risk for hypertrophic scars [4, 5], and an incision at the anterior neck is considered cosmetically undesirable.

To address these concerns, in 2004, we developed the bilateral axillo-breast approach (BABA) for endoscopic thyroidectomy, which we combined with the da Vinci robotic system. The da Vinci system has several merits: EndoWrist, with a high degree of freedom of motion, allows the operator to perform more sophisticated and complex maneuvers, and high-definition three-dimensional (3D) imaging provides a good operative view with excellent depth perception [6]. We have reported the surgical clearance and safety of BABA robotic thyroidectomy (RoT) compared with open thyroidectomy [7, 8]. Over the past 5 years, we have performed more than 1000 cases of BABA RoT, and we have analyzed the surgical outcomes to assess the safety and effectiveness of this method.

Here we report the clinicopathologic characteristics, operation types and times, tumor, node, metastasis classification system (TNM) stage (based on the 7th edition of the American Joint Committee on Cancer/International Union against Cancer [AJCC/UICC]), results after radioactive ablation therapy, and disease recurrence rates of 1026 cases of thyroid disease treated by BABA RoT.

Patients and methods

Patients

Between February 2008 and February 2012, a total of 1026 individuals with thyroid disease underwent RoT by BABA at Seoul National University Hospital and affiliated hospitals. These operations were performed by five surgeons at three academic centers. Informed consent was obtained from all patients, and the protocols were approved by the institutional review boards of Seoul National University Hospital.

Indications for BABA RoT [9] included low-risk, well-differentiated thyroid carcinoma (<2 cm, no definite evidence of local invasion, cervical nodal metastasis by the preoperative pathologic and radiologic finding), Graves disease, or a benign thyroid nodule up to 8 cm in size. Preoperative diagnoses for thyroid nodules were made by ultrasound and ultrasound-guided fine-needle aspiration biopsy. All patients underwent computed tomography to evaluate tumor location and cervical lymph node metastasis. Patients with previous neck surgery, distant metastasis or multiple lateral neck node metastases, definite tumor invasion to an adjacent organ, or lesions located in the posterior part of the thyroid adjacent to the tracheoesophageal groove or Berry ligament were excluded. All patients also had a breast examination with mammography or ultrasound to detect breast lumps. Large body habitus, high body mass index, and thyroiditis were not considered factors in selection of patients for BABA RoT.

Patients with malignancy underwent total thyroidectomy. Thyroid lobectomy was considered in patients with small (<1 cm), low-risk, unifocal, intrathyroidal papillary carcinoma in the absence of radiologically or clinically involved cervical nodal metastases, and in patients who preferred to undergo thyroid lobectomy [10]. Most proven cases of malignant thyroid carcinoma underwent prophylactic ipsilateral central compartment lymph node dissection (LND). After lateral neck node metastases were excluded in preoperative fine needle aspiration, selective lateral LND was performed in occasional cases in which metastases were suspected in one or two lymph nodes on preoperative imaging.

Operative methods

The RoT procedure has been previously described [8, 9, 11]. Briefly, RoT involves four stages: creation of a working space, robot docking, console time (when the surgeon operates via the da Vinci system console), and closure.

The patient was given general anesthesia and placed in the supine position, extending the neck. After draping, the trajectory lines and working space were drawn on the chest and neck. Creation of a working space was started by hydrodissection. Epinephrine (1:200,000) in a 0.9 % NaCl solution was injected subcutaneously in the working space under the platysma in the neck.

Incisions were made bilaterally at the superomedial edge of the breast areolas and at the axillary folds. Bilateral axillary and circumareolar 8–12 mm ports were inserted after the working space was developed, by using a vascular tunneler through the port incision. The flap extended from the thyroid cartilage superiorly to 2 cm below the clavicle inferiorly and laterally from just beyond the medial border of the sternocleidomastoid muscle. The working space was made by low pressure (5–6 mm Hg) insufflations of CO2 gas.

The robot was docked, and a camera was inserted through the right breast. Monopolar electrocautery or ultrasonic shears were inserted through the left breast. Graspers (Prograsp and Maryland forceps, Intuitive Surgical Inc., Sunnyvale, CA) were inserted through the right and left axillary ports.

The midline of the strap muscle was identified and separated. After visualizing the cricothyroid membrane, isthmus, and trachea, the isthmus was divided with the ultrasonic shears, which facilitated dissection of the gland laterally and posteriorly, and allowed for optimal visualization of the superior thyroid pedicle. Thyroidectomy was performed while identifying and preserving the parathyroid glands and recurrent laryngeal nerve. The dissected thyroid lobe was extracted with an endoplastic bag. Lesion-side central compartment dissection was performed, with care taken to ensure that the recurrent laryngeal nerve was not damaged in patients with a concomitant malignancy or suspicious nodule. The contralateral lobe was dissected in the same manner.

After complete removal of the thyroid, the operative field was irrigated with saline, and meticulous hemostasis was achieved. The midline was closed by robotic endosuturing. In cases of selective lateral LND, a short segment of one double polydioxanone loop (PDS*II, Ethicon Inc., Cincinnati, OH) was used for traction of the sternocleidomastoid muscle.

Follow-up

All patients were followed-up at 2 weeks, 3 months, and 6 months postoperatively and then annually. In all patients, recurrent laryngeal nerve function was assessed from the vocal cord movement by using video laryngoscopy preoperatively, and at 2 postoperative weeks. Hypomotility, median fixation, or paramedian fixation of the vocal cords was recorded as vocal cord palsy. Laryngoscopy was repeated at each visit until vocal cord movement returned. In the case of transient vocal cord palsy, vocal cord movement returned within 6 postoperative months. In cases with a vocal cord abnormality that persisted at 6 postoperative months, permanent vocal cord palsy was diagnosed.

During hospitalization, all patients underwent tests for serum total calcium, ionized calcium, phosphorus, and parathyroid hormone (PTH) to evaluate hypoparathyroidism. A serum total calcium level of <8 mg/dL with a hypocalcemic symptom during hospitalization was defined as transient hypoparathyroidism. Permanent hypoparathyroidism was defined as a serum PTH level of <5 pg/mL with an ongoing requirement for oral calcium supplementation beyond 6 months.

Indications for radioactive iodine (RAI) ablation were as follows: all patients with stage III or IV disease; all patients aged <45 years with stage II disease; most patients aged ≥ 45 with stage II disease; and selected patients with stage I disease, especially those with multifocal disease, nodal metastases, extrathyroid or vascular invasion, or aggressive histologies [9].

Patients for whom RAI ablation was indicated underwent therapy at 8–12 weeks after thyroidectomy. Before the administration of 131I, blood was drawn to measure TSH-stimulated serum thyroglobulin level (stimulated Tg). An 131I whole-body scan was performed on the second day after RAI ablation. In all patients with differentiated carcinoma who did not undergo RAI ablation, the TSH-suppressed serum thyroglobulin level (suppressed Tg) was checked at 3 postoperative months. All patients were examined by ultrasound at 1 postoperative year and annually thereafter.

Results

This study included 1026 consecutive cases (108 males and 918 females) with a mean age of 39.7 ± 9.8 years (range 13–70 years). According to postoperative pathology, the diagnoses of the cases were as follows: papillary thyroid carcinoma (n = 955), follicular thyroid carcinoma (n = 11), medullary thyroid carcinoma (n = 2), adenomatous goiter (n = 34), and follicular or Hürthle cell adenoma (n = 9). Nine cases were Graves disease without tumor in the final pathology report. One case was prophylactic thyroidectomy in a patient with MEN2A, with no neoplastic changes within the thyroid gland.

Of the eight cases that underwent completion thyroidectomy, five had no malignancy in the completion thyroidectomy specimen, and three had papillary thyroid carcinoma. Two cases were patients with newly diagnosed papillary thyroid carcinoma who had a history of previous endoscopic thyroidectomy. The mean malignant tumor size was 0.8 ± 0.49 cm (range 0.1–5.0 cm), and the mean benign tumor size was 1.9 ± 1.48 cm (range 0.2–6.0 cm). The operation types are summarized in Table 1.

Table 1 Patient characteristics
Full size table

The operating time of BABA total RoT with central LND included a mean of 38 ± 13 min (range 20–90 min) of working space creation and 75 ± 26 min (range 25–175 min) of console time. The mean postoperative hospital stay was 3.3 ± 0.7 days (range 2–10 days) (Table 1). The mean number of lymph nodes retrieved by central LND was 4.9 ± 3.74 nodes (range 1–24 nodes), and the number retrieved by selective lateral LND of levels II–IV was 13.2 ± 5.07 nodes (range 7–20 nodes).

Intraoperative bleeding occurred in two cases. The right internal jugular vein was torn during selective lateral LND and was repaired with a small incision on the neck area. Postoperative bleeding occurred in two cases. One case occurred on postoperative day 3 and stopped spontaneously and the other underwent decompression on postoperative day 6. One case of pneumothorax was found at postoperative day 2. Among the 872 patients who underwent BABA total RoT with central LND, transient hypoparathyroidism occurred in 341 patients (39.1 %) and transient vocal cord palsy occurred in 124 patients (14.2 %). Permanent hypoparathyroidism occurred in 13 patients (1.5 %), and permanent vocal cord palsy occurred in two patients (0.2 %).

A total of 889 patients with papillary thyroid carcinoma underwent BABA total RoT. The mean age of these patients was 40.2 ± 9.5 years (range 13–66 years), and the male-to-female ratio was 1: 8.8 (91 male and 798 female subjects). The mean tumor size on histological analysis was 0.75 ± 0.39 cm (range 0.1–2.7 cm). In total, 81.4 % of patients had papillary thyroid microcarcinoma. Multifocality and bilaterality were observed in 286 (32.1 %) and 169 (19.0 %) cases, respectively. Extrathyroid extension occurred in 446 cases (50.2 %), and was microscopic in 417 cases (46.9 %) and gross in 29 cases (3.3 %). In terms of TNM stage according to the 7th edition of AJCC/UICC, 693 patients (77.9 %) had stage I disease, 191 patients (21.5 %) had stage III disease, and five patients (0.6 %) had stage IVa disease (Table 2).

Table 2 Pathologic characteristics of 889 patients with papillary thyroid carcinoma who underwent bilateral axillo-breast approach robotic total thyroidectomy
Full size table

In total, 500 patients (56.2 %) received RAI ablation. The median stimulated Tg level of patients at the first RAI ablation was 0.4 ng/mL (range <0.1–79.84 ng/mL), and 65.1 % of patients (295 of 453) had stimulated Tg levels <1.0 ng/mL. Forty-two patients were referred outside of the hospital after 2 postoperative weeks or delayed RAI ablation until after pregnancy and childbirth. In five cases, the Tg levels at the first RAI ablation were not measured.

For the RAI ablation-negative group, the median suppressed Tg level at 3 postoperative months was <0.1 ng/mL (range <0.1–7.0 ng/mL), and 99.4 % of patients (335 of 337) had a suppressed Tg level of <1.0 ng/mL. Fifty-two patients were referred outside of the hospital. One patient whose Tg level was 1.6 ng/mL at the first follow-up was recommended RAI ablation. This patient had no problem on ultrasound and he refused RAI ablation. His Tg level became <0.1 ng/mL at his second visit. Another patient with a Tg level of 7.0 ng/mL was checked by positron emission tomography and ultrasound, which revealed no evidence of disease.

After a median follow-up of 23 months, there was no locoregional recurrence in patients who underwent RoT. There was one case of lung metastasis, detected by the first 131I whole-body scan with stimulated Tg <1.0 ng/mL, and this patient underwent high-dose RAI ablation.

Discussion

Since Gagner [12] performed the first endoscopic neck surgery in 1996, various endoscopic thyroidectomy techniques have been proposed. Shimazu et al. [13] developed the axillo-bilateral breast approach to overcome the disadvantages of using one direction with lateral approaches and the disadvantages of instrument interference with the axillary approach. The axillo-bilateral breast approach provides excellent cosmetic results without neck scarring, but is limited by the difficulty in visualizing both lobes of the thyroid. Accordingly, in 2004, we developed BABA for optimal visualization during total thyroidectomy [14]. In our experience, BABA endoscopic thyroidectomy provides a good symmetrical surgical view of both thyroid lobes, a similar operative approach to conventional open thyroidectomy, and an excellent cosmetic outcome [14]. However, endoscopic operation has several limitations [15–18]. In particular, the use of a 2D visual representation and nonflexible endoscopic instruments may make surgical field visualization and instrument manipulation difficult. Also, endoscopy personnel must be experienced, and the learning curve is rather steep. To overcome some of the limitations of manual endoscopic operation, we combined the BABA technique with a robotic system.

Serum Tg levels may be used to detect recurrent or residual disease in most patients who undergo total thyroidectomy [10, 19–21]. Although the levels may be measured while TSH is suppressed, the test is more sensitive when TSH is stimulated [19, 20, 22, 23]. Stimulated Tg is directly correlated to the amount of residual thyroid tissue [20, 21, 24–26], and it may be used to predict persistent or recurrent disease by measurement of the thyroid bed 131I uptake after first RAI ablation [26–28]. A recent study of patients undergoing open thyroidectomy showed that 25.3 % of 438 patients with differentiated thyroid carcinoma showed a stimulated Tg <2.0 ng/mL after first RAI ablation [29]. In another report, 48.3 % of 729 patients with low-risk thyroid carcinoma showed a stimulated Tg level ≤1 ng/mL [24].

A group in Italy reported surgical completeness with minimally invasive video-assisted thyroidectomy (MIVAT) for papillary carcinoma [21]. MIVAT was carried out through a 15 mm incision between the sternal notch and the cricoid cartilage with needlescopic instruments and a 30 infinity 5 mm endoscope [21, 30]. The mean stimulated Tg after operation obtained with MIVAT (5.3 ± 5.8 ng/mL) was similar to that obtained with open thyroidectomy (7.6 ± 21.7 ng/mL). A separate report showed a mean stimulated Tg after video-assisted total thyroidectomy of 5.5 ± 6.2 ng/mL (range <0.1–25.3), with 21 % of 152 patients showing a stimulated Tg <0.1 ng/mL [30]. RoT using a transaxillary approach used a 5 to 6 cm vertical incision along the lateral border of the pectoralis major muscle in the axilla, created working space over the anterior surface of the pectoralis major muscle and between the sternal and the clavicular heads of the sternocleidomastoid muscle, and exposed thyroid under strap muscle with an external retractor [15]. Rather than reporting stimulated Tg levels, many studies of RoT using a transaxillary approach provide suppressed Tg levels at 4 postoperative months with or without RAI [15, 31–33]. Another group reported stimulated Tg levels at first RAI ablation after RoT by unilateral axillo-breast or axillary approach (mean: 12.70 ± 15.01 ng/mL, median: 8.8 ng/mL, range 0.1–62.6 ng/mL) or open thyroidectomy (mean: 4.90 ± 8.57 ng/mL, median: 1.9 ng/mL, range <0.1–65.8) [34].

In the case of BABA RoT, we previously reported the surgical outcomes of our initial experience of 109 patients [8]. The median stimulated Tg level of 57 patients who underwent RAI ablation was 0.2 ng/mL (range <0.1–36.4), and 74 % had a stimulated Tg level of <1.0 ng/mL [8]. We examined the mean stimulated Tg after first RAI ablation among patients who underwent BABA RoT compared to those who underwent open thyroidectomy, using the propensity score to show surgical completeness. The mean stimulated Tg levels and percentage with stimulated Tg level of <1.0 ng/mL were 1.4 ± 3.9 ng/mL (range <0.1–36.4) and 69.1 %, respectively, for the RoT group, compared to 1.2 ± 3.1 ng/mL (range <0.1–38.7) and 68.6 %, respectively, for the open thyroidectomy group. The surgical completeness of BABA RoT was not inferior to that of open thyroidectomy [7]. In the present study, the median stimulated Tg level of patients after first RAI ablation was 0.4 ng/mL (range <0.1–79.84), and 65.1 % (295 of 453) had a stimulated Tg level of <1.0 ng/mL. These results are similar to those of our previous study, and the stimulated Tg level achieved was lower than that obtained when using a different approach (Table 3).

Table 3 Stimulated thyroglobulin after first RAI ablation among patients with papillary thyroid carcinoma
Full size table

Lee et al. [8] addressed some limitations of robotic thyroid surgery, which were a greater operative time and a smaller number of retrieved lymph node in robotic group than in open group. The operative time of robot surgery includes working space creation time, docking time, console time and closure time. The console time was 123 ± 29 min and retrieved lymph node was 2.5 ± 3.2 in previous RoT group. The console time was 75 ± 26 min in the present group and improved than previous group as surgeon’s experience expanded. The number of retrieved lymph node was also improved as surgeon’s experience expanded. The mean postoperative hospital stay of 3.3 days for RoT group was similar with that of 3.2 days for open thyroid surgery, which was reported previously [8, 14, 35].

The incidence rates of postoperative hypoparathyroidism and vocal cord palsy vary depending on the assessment modality and extent of surgery [36–38]. In the present group, 39.1 % of patients experienced transient hypoparathyroidism and 14.2 % of patients experienced transient vocal cord palsy, compared with rates of 17.7 % and 2.5 %, respectively, for open thyroidectomy in our institution [35]. For open thyroidectomy, the rates of transient hypoparathyroidism were reported range from 14.0 % to 39.6 % [36–42]. In the present study, transient hypoparathyroidism rate is within acceptable rate compared with previous reports but higher than those of open thyroidectomy in our institution [35]. Transient vocal cord palsy rates are high. It is thought that any of vocal cord movement abnormality such as hypomotility without hoarseness is also considered as transient vocal cord palsy in our institution. After conventional open thyroidectomy, reported rates of permanent hypoparathyroidism range from 0 to 4.9 % and rates of permanent vocal cord palsy range from 0 to 3.6 % [36–42]. In the present study, the rates of permanent hypoparathyroidism (1.5 %) and permanent vocal cord palsy (0.2 %) were similar or better than those previously reported for conventional open thyroidectomy [35]. One of the two patients with permanent vocal cord palsy recovered from vocal cord hypomotility by 18 months after operation. We had experience of bleeding control after BABA endoscopic surgery [18], which took endoscopic approach using the same incision sites. We had not experienced immediate postoperative bleeding among the cases of BABA RoT. However, we assume that endoscopic decompression and bleeding control in BABA RoT could be done in the same way as in BABA endoscopic operation.

We previously reported that BABA was associated with a higher degree of patient cosmetic satisfaction and BABA induced some sensory impairment in the anterior chest area including on the breasts and nipples but it improved significantly 3 months after surgery [14, 43]. And the operative scars are almost invisible [8, 14]. This cosmetic outcome is the main drive for oncoplastic thyroid surgery. The results of our study show that RoT using BABA can result in an oncologically safe operation using oncoplastic principles [44, 45]. Additional research is needed to compare the recurrence rate between open thyroidectomy and BABA RoT because of relatively short follow-up periods.

The present data includes 174 patients of our initial BABA robotic total thyroidectomy which we previously reported [7, 8]. All of these 174 operations were performed by one expert operator with the same team. To assess the oncologically safety and completeness of BABA RoT, we investigated 1026 cases of BABA RoT performed by several surgeons with or without technical expertise of BABA in our institution and affiliated hospitals. Although the study involved five surgeons with varying RoT experience at three centers, the results were similar to those obtained in our previous study. In BABA, the docking of the robot was designed to optimize visualization of the midline of the strap muscle by using the right breast port for the camera. This symmetrical view allowed both thyroid lobes to be dissected with a similar view and with a medial approach rather than open surgery. With this advantage, the surgeon can perform total thyroidectomy both completely and safety.

Conclusions

BABA RoT is an oncologically safe and effective method that provides good surgical completeness as well as low rates of postoperative complications and recurrence for lower-risk patients with differentiated thyroid cancer.