Surg Endosc (2016) 30:4785–4792
DOI 10.1007/s00464-016-4808-y
Oncologic outcomes of robotic thyroidectomy: 5-year experience
with propensity score matching
Kyung Tae1 • Chang Myeon Song1 • Yong Bae Ji1 • Eui Suk Sung1
Jin Hyeok Jeong1 • Dong Sun Kim2
Received: 21 September 2015 / Accepted: 3 February 2016 / Published online: 19 February 2016
Ó Springer Science+Business Media New York 2016
Abstract
Background The oncologic outcome of robotic thy-
roidectomy is not yet well established. The aim of this
study was to evaluate the recurrence rate after robotic
thyroidectomy in comparison with conventional thy-
roidectomy for papillary thyroid carcinoma (PTC).
Methods We analyzed 896 patients with PTC who either
underwent robotic (212 patients using a gasless unilateral
axillary or an axillo-breast approach) or conventional cer-
vical thyroidectomy (684 patients) with/without central
neck dissection between October 2008 and February 2014.
We excluded patients who underwent concomitant lateral
neck dissection or completion thyroidectomy, and cases
with T4 tumor, tumor lager than 4 cm, other types of
thyroid cancer, recurrent cancer, and distant metastasis. A
propensity score matching analysis was done with ten
covariates including age, gender, body mass index, tumor
size, multifocality, bilaterality, extrathyroidal extension,
type of thyroidectomy, extent of central neck dissection,
and RAI ablation to reduce selection bias.
Results In baseline data, the male ratio and the mean age
were lower in the robotic group. Stage, multifocality, and
bilaterality were higher in the conventional group. The rate
of total thyroidectomy was higher in the conventional
group. The two matched groups of each 185 patients were
produced and well balanced by propensity score matching.
& Kyung Tae
kytae@hanyang.ac.kr
1
Department of Otolaryngology-Head and Neck Surgery,
College of Medicine, Hanyang University, 222
Wangsimni-ro, Seongdong-gu, Seoul 133-792, Korea
2
Department of Internal Medicine, College of Medicine,
Hanyang University, Seoul, Korea
and Other Interventional Techniques
•
In the comparison of propensity score matched groups,
operative time was longer in the robotic group (P \ 0.001),
and postoperative complications did not differ between the
two groups, except for transient hypoparathyroidism and
formation of seroma. The recurrence rate did not differ
between the two groups after a mean follow-up of
43.6 months (0.5 and 1.1 % in the robotic and conventional
groups, respectively, P = 0.375).
Conclusion The oncologic outcome of robotic thy-
roidectomy in 5-year experience is comparable to that of
conventional thyroidectomy in selected patients with PTC.
Keywords Papillary thyroid carcinoma
Robotic
thyroidectomy
Oncologic outcomes
Gasless unilateral
axillary approach
Recurrence
Robotic thyroidectomy using a transaxillary approach has
gained in popularity and has been performed in many
institutes, although some surgeons still have reservations
about it. Since 2008, we have been performing robotic
thyroidectomy using a gasless unilateral axillary (GUA) or
a gasless unilateral axillo-breast (GUAB) approach to
avoid a visible neck scar and improve the cosmetic out-
comes [1–4].
In terms of technical feasibility and safety, we and other
authors have reported that early surgical outcomes and
complication rates of robotic thyroidectomy via the
transaxillary approach were comparable to those of con-
ventional transcervical thyroidectomy [1–14]. In addition,
with regard to the postoperative cosmetic satisfaction and
functional outcomes, we previously reported that swal-
lowing function, neck discomfort and paresthesia, and
overall quality of life (QOL) after robotic thyroidectomy
were similar to those after conventional cervical
123
4786
thyroidectomy and that postoperative cosmesis and voice
outcomes were better in the robotic group [15–20]. Also,
we and other authors demonstrated that the surgical com-
pleteness of robotic total thyroidectomy by transaxillary
approach was comparable to that of conventional thy-
roidectomy if the procedure was performed in properly
selected patients by expert robotic thyroid surgeons [21–
23].
In the assessment of a new surgical technique for
patients with thyroid cancer, oncologic safety and outcome
are the most important issues, in addition to technical
safety, functional or cosmetic outcomes, and surgical
completeness. The oncologic safety of robotic thyroidec-
tomy should not be overlooked or neglected in favor of
cosmetic or functional outcomes. However, the oncologic
outcome has not yet been well established for patients with
papillary thyroid carcinoma (PTC). The aim of the present
study was to compare the oncologic outcomes of robotic
thyroidectomy and conventional transcervical thyroidec-
tomy using propensity score matching to reduce patient
selection bias.
Patients and methods
Patients
We retrospectively analyzed the chart reviews of 896 con-
secutive patients with PTC who underwent thyroidectomy
with or without central neck dissection (CND) using a
conventional cervical approach or a robotic approach
between October 2008 and February 2014. Of the 896
patients, 684 underwent conventional cervical thyroidec-
tomy and 212 underwent robotic thyroidectomy using the da
Vinci Surgical System (Intuitive Surgical, Inc., Sunnyvale,
CA, USA). In the robotic group, the GUA approach was used
in 107 patients and the GUAB approach in 105 patients. In
the present analysis, we excluded patients who underwent
concomitant lateral neck dissection for confirmed or suspi-
cious lymph node metastasis in the lateral compartment or
completion thyroidectomy, and cases with T4 tumor, tumor
lager than 4 cm, other types of thyroid cancer, recurrent
cancer, and distant metastasis. We also excluded cases that
were not followed up to the time of writing.
The patients were not randomized for robotic versus
conventional thyroidectomy. The operative method was
decided according to the extent of disease, patient’s pref-
erence, and financial reasons. Robotic thyroidectomy is not
covered by the Korean National Health Insurance System.
The cost of robotic thyroidectomy is 3–4 times higher than
that of conventional thyroidectomy. So some patients
choose conventional procedures because of high cost of
robotic thyroidectomy. All robotic procedures were
123
Surg Endosc (2016) 30:4785–4792
performed by the same surgeon (K.T.). Conventional
transcervical thyroidectomy was performed by three expert
thyroid surgeons (K.T., C.M.S., and Y.B.J). All patients
provided informed consent, and for robotic thyroidectomy
the patients were informed of the possibility of conversion
to conventional thyroidectomy. The study was approved by
the Institutional Review Board of Hanyang University
Hospital.
Eligibility for robotic thyroidectomy using a GUA or
GUAB approach for PTC in our institute included tumors
of less than 4 cm with or without minimal extrathyroidal
extension (ETE), or metastatic lymph nodes less than 4 cm
in the central or lateral compartment on preoperative
ultrasonography (US) and/or computed tomography (CT).
Exclusion criteria for robotic thyroidectomy included PTC
with gross maximal ETE, multiple conglomerated lymph
node metastases with extensive invasion of surrounding
structures in the central or lateral compartment, and distant
metastasis. Patients with a history of neck or thyroid sur-
gery or irradiation were also excluded from the robotic
procedure.
Preoperative diagnosis of thyroid cancer and evaluation
of tumor extent were determined using US, CT, and fine-
needle aspiration cytology (FNAC) in all patients, and
FNAC was also performed to confirm lymph node metas-
tasis in patients with suspicious cervical lymph node
metastasis.
Although the extent of thyroidectomy was not same
among the surgeons in our institute, thyroid lobectomy was
usually recommended for patients with PTC who have an
intrathyroidal lesion that was less than 2–3 cm in diameter
without clinical cervical lymph node metastasis preopera-
tively. Therapeutic CND was performed in all patients with
suspicious lymph node metastasis on imaging study, and
prophylactic CND was routinely recommended in most
patients with clinically negative neck.
The operative procedures of robotic thyroidectomy via
GUA and GUAB approaches have been described previ-
ously [1, 3, 4]. A 5- to 6-cm low-collar incision was used
for conventional transcervical thyroidectomy. Intraopera-
tive neuromonitoring was used selectively in some patients
of both groups to identify the RLN.
Postoperative radioactive iodine (RAI) ablation at a dose
of 30–150 mCi was performed 2–4 months after operation.
In our institute, RAI ablation was performed in all patients
with gross ETE regardless of tumor size, tumor size larger
than 4 cm, or distant metastases. RAI ablation was also
recommended for selected patients with minimal ETE or
cervical lymph node metastasis, and tumor size 1–4 cm
and/or higher risk histologic features according to Ameri-
can Thyroid Association (ATA) guidelines [24].
Vocal cord paralysis was routinely determined by pre-
operative and postoperative flexible laryngoscopy in all
Surg Endosc (2016) 30:4785–4792
patients. Hypoparathyroidism was defined as any drop in
blood parathyroid hormone level below the normal limit,
regardless of hypocalcemic symptoms. Permanent recur-
rent laryngeal nerve (RLN) palsy or hypoparathyroidism
was defined as non-recovery within 6 months.
Recurrence was defined as the development of new
abnormal structural lesions on imaging study, such as neck
US, CT, or whole-body iodine scan, and was confirmed
pathologically using FNAC. Physical examination, neck
US, and serum TSH-stimulated or suppressed thyroglobu-
lin (Tg) measurements were used to detect recurrence after
surgery at 6- to 12-month intervals in all patients. Whole-
body iodine scan was performed selectively in some
patients who underwent total thyroidectomy and postop-
erative RAI ablation. Structural recurrence within
12 months after the initial surgery was considered as per-
sistent disease rather than recurrence.
Propensity score matching analysis
Propensity score matching was used to overcome patient
selection bias and confounding differences. For propensity
score matching, ten factors of baseline clinicopathologic
characteristics including age, gender, body mass index
(BMI), tumor size, multifocality, bilaterality, ETE, extent
of thyroidectomy, extent of CND, and RAI ablation were
selected as covariates. Propensity scores of individuals
were calculated using logistic regression analysis (SPSS
version 18.0, Chicago, IL), and the two matched groups
were produced and were compared for surgical outcomes,
complications, and recurrence.
Statistical analysis
Differences in continuous variables between the robotic
and conventional groups were compared using the Mann–
Whitney U test and differences in categorical variables by
the Chi-square test and Fisher’s exact test for small cell
variables. After propensity score matching, we assessed the
balance of baseline covariates of the two groups and
compared the perioperative outcomes and recurrence with
Wilcoxon signed rank test for continuous variables and
McNemar test for categorical variables. A P value less than
0.05 was considered statistically significant.
Results
Baseline clinicopathological characteristics
and perioperative outcomes
The baseline clinicopathological characteristics are listed
in Table 1. In comparison with the group undergoing
4787
conventional surgery, the female ratio was higher
(P \ 0.001) and the mean age was lower (P \ 0.001) in
the robotic surgery group. The body mass index did not
differ between the two groups. There was no difference in
tumor size, minimal ETE, and T and N classification
between the two groups. However, multifocality, bilater-
ality, and stage were significantly higher in the conven-
tional surgery group.
Robotic thyroidectomy was completed successfully in
all patients, and no case required conversion to transcer-
vical thyroidectomy. The proportion of total thyroidec-
tomies was significantly higher in the conventional surgery
group (P \ 0.001). The frequency of CND did not differ
between the two groups; however, the proportion of bilat-
eral CND was higher in the conventional surgery group.
Among patients who underwent total thyroidectomy,
postoperative RAI ablation was performed in 432 (74.2 %)
in the conventional group and 79 (75.9 %) in the robotic
group (P = 0.807). The mean of dose of RAI did not differ
between the two groups.
Comparison of perioperative outcomes
and recurrence outcomes after propensity score
matching
After propensity score matching, 185 patients were mat-
ched and the two matched groups were well balanced in the
ten covariates (Table 2). The significant differences
between the two groups seen on the baseline analysis in
age, gender, multifocality, bilaterality, stage, extent of
thyroidectomy and CND disappeared.
The comparison of perioperative outcomes, complications,
and recurrence between the conventional and robotic thy-
roidectomy groups after propensity score matching is pre-
sented in Table 3. The mean number of lymph nodes
retrieved did not differ between the two groups. The rate of
central compartment node metastasis and the mean number of
metastatic lymph nodes did not differ between the two groups.
The ratio of the number of positive lymph nodes to retrieved
lymph nodes also did not differ between the two groups. The
mean operative times of lobectomy and total thyroidectomy
were significantly longer in the robotic surgery group
(P \ 0.001 in both procedures). The volume of drainage was
higher in the robotic surgery group (P \ 0.001).
With regard to perioperative complications, there was
no significant difference between the two groups except for
transient hypoparathyroidism and seroma. Transient
hypoparathyroidism after total thyroidectomy was signifi-
cantly higher in the conventional surgery group (46.8 %)
than in the robotic surgery group (30.1 %, P = 0.017).
Postoperative seroma formation was higher in the robotic
surgery group (13 %) than in the conventional surgery
group (5.9 %, P = 0.043). Seroma was resolved by
123
4788 Surg Endosc (2016) 30:4785–4792

Table 1 Comparison of
Characteristics Conventional group (n = 684) Robotic group (n = 212) P value
baseline demographic and
clinicopathologic characteristics Age (years) 52.1 ± 11.8 40.3 ± 9.7 \0.001
of the robotic and conventional
cervical thyroidectomy groups Sex \0.001
Male 150 (21.9 %) 19 (8.96 %)
Female 534 (78.1 %) 193 (91.04 %)
Body mass index (kg/m2) 24.9 ± 3.5 (range 15.4–45.4) 24.1 ± 4.2 (range 16.8–45.6) 0.064
Tumor size (mm) 10.9 ± 8.3 9.2 ± 9.5 0.116
Tumor multifocality 205 (29.9 %) 40 (18.9 %) 0.001
Tumor bilaterality 147 (21.5 %) 21 (9.9 %) \0.001
Extrathyroidal extension 241 (35.2 %) 77 (36.3 %) 0.805
T classification 0.769
T1 423 (61.9 %) 127 (59.9 %)
T2 20 (2.9 %) 8 (3.8 %)
T3 241 (35.2 %) 77 (36.3 %)
N classification 0.432
Nx 74 (10.8 %) 21 (9.9 %)
N0 422 (61.7 %) 123 (58.0 %)
N1a 188 (27.5 %) 68 (32.1 %)
TNM stage 0.002
I 451 (65.9 %) 167 (78.8 %)
II 9 (1.3 %) 2 (0.9 %)
III 224 (32.8 %) 43 (20.3 %)
Extent of thyroidectomy \0.001
Total thyroidectomy 582 (85.1 %) 104 (49.1 %)
Lobectomy 102 (14.9 %) 108 (50.9 %)
Central neck dissection 610 (89.2 %) 191 (90.1 %) 0.799
Extent of CND \0.001
Unilateral CND 289 (47.4 %) 142 (74.3 %)
Bilateral CND 321 (52.6 %) 49 (25.7 %)
RAI ablation 432/582 (74.2 %) 79/104 (75.9 %) 0.807
Dose of RAI (mCi) 108.5 ± 44.5 109.1 ± 50.9 0.923
CND central neck dissection, RAI radioactive iodine

repeated aspiration. Transient RLN palsy occurred in 7
(3.8 %) and 3 (1.6 %) patients in the conventional and
robotic groups, respectively (P = 0.172). There was one
case of permanent RLN palsy in the conventional group.
Postoperative hematoma occurred in two cases (1.1 %) of
conventional thyroidectomy and three cases (1.6 %) of
robotic thyroidectomy. Re-exploration using the same
incision was performed in all five patients to manage
postoperative hemorrhage. There were no tracheal or bra-
chial plexus injuries or skin flap complications in the
robotic surgery group.
The RAI uptake rate at the first RAI ablation did not
differ between the two groups. Also, after RAI ablation the
mean level of serum TSH-stimulated Tg was similar
between the two groups (P = 0.344).
The mean follow-up times of the robotic and conven-
tional groups were 42.3 ± 21.7 months and
44.9 ± 19.2 months, respectively (P = 0.150). The recur-
rence rate did not differ between the two groups
(P = 0.375). Recurrence developed in one patient (0.5 %)
in the robotic group and in two (1.1 %) in the conventional
group. In the robotic group, one patient who underwent
lobectomy showed recurrence in a central compartment
lymph node at 38 months postoperatively. In the conven-
tional group, the sites of recurrence were the lateral com-
partment lymph nodes in two patients at postoperative 30
and 60 months in each. Of three patients with recurrence,
two patients of the conventional group underwent reoper-
ation for recurrent lateral lymph nodes and one patient who
had undergone robotic lobectomy wanted close follow-up
of recurrence in a 7-mm-sized lymph node in the central
compartment without reoperation. There was no distant
metastasis after surgery in either groups, and no patient
died for any reason.

123
Surg Endosc (2016) 30:4785–4792 4789

Table 2 Demographic and

Characteristics Conventional group (n = 185) Robotic group (n = 185) P value
clinicopathologic characteristics
of the robotic and conventional Age (years) 42.6 ± 8.9 41.8 ± 9.1 0.355
cervical thyroidectomy groups
after propensity score matching Sex 0.500
using ten covariates Male 20 (10.8 %) 19 (10.3 %)
Female 165 (89.2 %) 166 (89.7 %)
Body mass index (kg/m2) 24.6 ± 4.5 (range 15.4–44.2) 23.9 ± 5.0 (range 16.8–45.6) 0.501
Tumor size (mm) 7.9 ± 5.4 8.2 ± 5.3 0.200
Tumor multifocality 39 (21.1 %) 37 (20.0 %) 0.864
Tumor bilaterality 25 (13.5 %) 20 (10.8 %) 0.405
Extrathyroidal extension 66 (35.7 %) 77 (41.6 %) 0.177
T classification 0.331
T1 113 (61.1 %) 103 (55.7 %)
T2 3 (1.6 %) 7 (3.8 %)
T3 69 (37.3 %) 75 (40.5 %)
N classification 0.889
Nx 22 (11.9 %) 19 (10.3 %)
N0 98 (53.0 %) 98 (53.0 %)
N1a 65 (35.1 %) 68 (36.8 %)
TNM stage 0.615
I 146 (78.9 %) 142 (76.8 %)
II 1 (0.5 %) 0
III 38 (20.5 %) 43 (23.2 %)
Extent of thyroidectomy 0.302
Total thyroidectomy 111 (60.0 %) 103 (55.7 %)
Lobectomy 74 (40.0 %) 82 (44.3 %)
Central neck dissection 163 (88.1 %) 166 (89.7 %) 0.720
Extent of CND 0.839
Unilateral CND 112 (68.7 %) 117 (70.5 %)
Bilateral CND 51 (31.3 %) 49 (29.5 %)
RAI ablation 88/111 (79.2 %) 78/103 (75.7 %) 0.314
Dose of RAI (mCi) 104.1 ± 47.2 108.6 ± 35.0 0.482
CND central neck dissection, RAI radioactive iodine

Discussion matching to minimize patient selection bias. The propen-

Robotic thyroidectomy using a GUA or GUAB approach
has advantages, the most significant of which is the
excellent cosmetic outcome without scars on the anterior
neck [18]. Identification of the RLN and parathyroid glands
is subjectively easier than in conventional thyroidectomy,
and delicate dissection of the thyroid gland is possible with
the 3-dimensional magnified view.
However, the robotic thyroidectomy procedure is tech-
nically challenging and some surgeons still have concerns
about potential complications of the procedure and are
therefore reluctant to adopt a robotic technique for surgery
of thyroid cancer [25, 26].
In this study, we compared the oncologic outcomes of
robotic thyroidectomy with conventional transcervical
thyroidectomy in 5-year experience using propensity score
sity score defined as the conditional probability of assign-
ment to a particular treatment versus control given the
covariates can be used to balance the covariates in the two
groups [27, 28]. Exact matchings made using the propen-
sity score will remove all of the bias in the background
covariates. Propensity score matching has been proven to
reduce selection bias and increase precision in non-ran-
domized retrospective study and allows comparison
between different surgical procedures.
In propensity score matching analysis, the complication
rate was similar in the two groups of patients, except for
transient hypoparathyroidism and seroma. The incidence of
transient hypoparathyroidism was lower in the robotic
surgery group than in the conventional group and may be
related to the capsular dissection and 3-dimensional mag-
nification of the operative view, resulting in better

123
4790 Surg Endosc (2016) 30:4785–4792

Table 3 Comparison of surgical outcomes, complications and recurrence in the conventional and robotic thyroidectomy groups after propensity
score matching using ten covariates
Factors Conventional group (n = 185) Robotic group (n = 185) P value

Central lymph node metastasis 65/163 (39.9 %) 68/166 (41.0 %) 0.815

Number of lymph nodes retrieved
Unilateral CND 5.78 ± 3.84 5.50 ± 3.48 0.579
Bilateral CND 9.45 ± 4.78 8.98 ± 4.91 0.631
Number of positive lymph nodes
Unilateral CND 0.71 ± 1.63 0.95 ± 2.22 0.367
Bilateral CND 2.25 ± 2.68 2.09 ± 2.98 0.768
Ratio of positive/retrieved lymph nodes (%)
Unilateral CND 12.0 ± 23.5 21.0 ± 33.1 0.105
Bilateral CND 21.1 ± 25.9 23.7 ± 27.6 0.680
Operative time (min)
Total thyroidectomy 121.2 ± 47.2 157.9 ± 36.5 \0.001
No CND 137.0 ± 17.5 172.5 ± 24.7 0.078
Unilateral CND 128.6 ± 50.1 150.9 ± 36.5 0.024
Bilateral CND 112.3 ± 46.1 165.7 ± 38.8 \0.001
Lobectomy 98.9 ± 25.2 130.2 ± 38.7 \0.001
No CND 89.6 ± 29.9 157.7 ± 61.1 0.001
Unilateral CND 102.0 ± 39.1 124.9 ± 39.8 0.003
Volume of drainage (mL)
Total thyroidectomy 144.8 ± 48.6 229.5 ± 87.3 \0.001
Lobectomy 122.3 ± 60.9 207.7 ± 87.1 \0.001
RAI uptake rate at first RAI ablation 83/88 (94.3 %) 76/78 (97.4 %) 0.375
Stimulated Tg after RAI ablation (ng/ml) 1.05 ± 3.61 0.85 ± 2.18 0.344
Complications
Transient RLN palsy 7 (3.8 %) 3 (1.6 %) 0.172
Permanent RLN palsy 1 (0.5 %) 0 0.500
Transient hypoparathyroidism
Total thyroidectomy 52/111 (46.8 %) 31/103 (30.1 %) 0.017
Lobectomy 7/74 (9.5 %) 9/82 (11.0 %) 0.797
Permanent hypoparathyroidism
Total thyroidectomy 2/111 (1.8 %) 2/103 (1.9 %) 0.606
Lobectomy 0/74 0/82 NA
Hematoma 2 (1.1 %) 3 (1.6 %) 0.312
Seroma 11 (5.9 %) 24 (13.0 %) 0.043
Brachial plexus injury 0 0 NA
Follow-up (month) 42.3 ± 21.7 44.9 ± 19.2 0.150
Recurrence rate 2 (1.1 %) 1 (0.5 %) 0.375
Mean time to recurrence (month) 45.0 ± 21.2 38 0.667
Survival 185 (100 %) 185 (100 %) NA
CND central neck dissection, RAI radioactive iodine, Tg thyroglobulin, RLN recurrent laryngeal nerve, NA not applicable

preservation of the parathyroid glands in the robotic pro-
cedure. The formation of seroma and volume of drainage
were higher in the robotic surgery group than in the con-
ventional group because of the wider dissection area
including anterior chest wall to reach the thyroid gland.
However, seroma was resolved with the conservative
method of repeat aspiration. These results indicate that the
technical feasibility and safety of robotic thyroidectomy
are similar to those in the conventional approach if the
robotic procedure is performed by surgeons who are
familiar with the technique.
Even assuming that robotic thyroidectomy is technically
safe, oncologic safety has not yet been confirmed well for
its use in the surgical treatment of thyroid cancer although

123
Surg Endosc (2016) 30:4785–4792
one article recently published reported the comparable
oncologic outcome of trans-axillary robotic thyroidectomy
compared to conventional thyroidectomy [29].
In this study, the primary measurement of oncologic
outcome was structural recurrence. In order to evaluate
only the robotic thyroidectomy procedure avoiding the
influence of robotic lateral compartment neck dissection,
we excluded cases with clinically positive lateral com-
partment lymph nodes where concomitant lateral com-
partment neck dissection with thyroidectomy was needed
[30]. We also excluded cases of contraindication for
robotic thyroidectomy such as T4 tumor with maximal
ETE, tumors larger than 4 cm or recurrent cancer to ana-
lyze the recurrent rate in the groups with similar eligibility
for either robotic or conventional thyroidectomy. The
recurrence rate did not differ between the robotic (0.5 %)
and conventional (1.1 %) surgical groups of patients after
the mean follow-up of 43.6 ± 20.6 months in this study.
The usual average recurrence rate is 0–9 % with an average
of 4 % in PTC patients who are clinically node negative
[31]. The recurrence rate in pathologically node-positive
PTC is 3–4 % in patients with a small number of micro-
scopic lymph node metastases and up to 32 % in patients
with bulky locoregional metastases [31]. The low recur-
rence rate in the present study might be due to the rela-
tively short follow-up period and the inclusion criterion of
low-risk PTC patients.
Before final conclusions can be reached with regard to
the oncologic outcomes of robotic thyroidectomy, it is
essential to study the follow-up over a long-term period,
because PTC can recur even 20–30 years after thyroidec-
tomy. Nevertheless, two-thirds of recurrence usually
develops within the first decade after initial surgery, par-
ticularly in the first 5 years, the period of highest risk [32].
In the present study, there was no persistent structural
disease after initial surgery and all recurrences developed
at least 24 months after surgery, indicating indirectly that
the initial diagnosis was appropriate and the surgical
treatment would not leave residual structural disease. We
therefore suggest, with caution, that the results of this study
show that robotic thyroidectomy is as effective as con-
ventional cervical thyroidectomy, at least for the treatment
of low-risk PTC patients without lateral compartment
lymph node metastasis or maximal ETE.
The major limitations of our study are the heterogeneous
case–control groups and the retrospective non-randomized
study design although propensity score matching was used
to overcome these limitations. There might be also some
inter-operator variation in conventional thyroidectomy
although all conventional thyroidectomy was performed by
three expert thyroid surgeons. The short follow-up period is
also another major drawback. Further studies including
larger numbers of patients with the prospective matching
4791
control group and longer follow-up for up to 20 or 30 years
are necessary to overcome the limitations of this study.
In conclusion, when performed by surgeons experienced
in robotic techniques for thyroid surgery, the oncologic
outcome in 5-year experience of robotic thyroidectomy via
GUA or GUAB approaches is comparable to that of con-
ventional cervical thyroidectomy in selected patients with
PTC. Further studies with long-term follow-up in large
patient samples are necessary to determine the ultimate
long-term oncologic outcomes of robotic thyroidectomy.
Compliance with ethical standards
Disclosures Drs. Kyung Tae, Chang Myeon Song, Yong Bae Ji, Eui
Suk Sung, Jin Hyeok Jeong, and Dong Sun Kim have no conflicts of
interest or financial ties to disclose.
References
1. Tae K, Ji YB, Jeong JH, Lee SH, Jeong MA, Park CW (2011)
Robotic thyroidectomy by a gasless unilateral axillo-breast or
axillary approach: our early experiences. Surg Endosc
25:221–228
2. Tae K, Ji YB, Cho SH, Kim KR, Kim DW, Kim DS (2011) Initial
experience with a gasless unilateral axillo-breast or axillary
approach endoscopic thyroidectomy for papillary thyroid micro-
carcinoma: comparison with conventional open thyroidectomy.
Surg Laparosc Endosc Percutan Tech 21:162–169
3. Tae K, Ji YB, Cho SH, Lee SH, Kim DS, Kim TW (2012) Early
surgical outcomes of robotic thyroidectomy by a gasless unilat-
eral axillo-breast or axillary approach for papillary thyroid car-
cinoma: 2 years’ experience. Head Neck 34:617–625
4. Tae K, Ji YB, Jeong JH, Kim KR, Choi WH, Ahn YH (2013)
Comparative study of robotic versus endoscopic thyroidectomy
by a gasless unilateral axillo-breast or axillary approach. Head
Neck 35:477–484
5. Kang SW, Lee SC, Lee SH, Lee KY, Jeong JJ, Lee YS, Nam KH,
Chang HS, Chung WY, Park CS (2009) Robotic thyroid surgery
using a gasless, transaxillary approach and the da Vinci S system:
the operative outcomes of 338 consecutive patients. Surgery
146:1048–1055
6. Lee J, Kang SW, Jung JJ, Choi UJ, Yun JH, Nam KH, Soh EY,
Chung WY (2011) Multicenter study of robotic thyroidectomy:
short-term postoperative outcomes and surgeon ergonomic con-
siderations. Ann Surg Oncol 18:2538–2547
7. Kuppersmith RB, Holsinger FC (2011) Robotic thyroid surgery:
an initial experience with North American patients. Laryngo-
scope 121:521–526
8. Lee J, Nah KY, Kim RM, Ahn YH, Soh EY, Chung WY (2010)
Differences in postoperative outcomes, function, and cosmesis:
open versus robotic thyroidectomy. Surg Endosc 24:3186–3194
9. Noureldine SI, Jackson NR, Tufano RP, Kandil E (2013) A
comparative North American experience of robotic thyroidec-
tomy in a thyroid cancer population. Langenbecks Arch Surg
398:1069–1074
10. Yi O, Yoon JH, Lee YM, Sung TY, Chung KW, Kim TY, Kim
WB, Shong YK, Ryu JS, Hong SJ (2013) Technical and onco-
logic safety of robotic thyroid surgery. Ann Surg Oncol
20:1927–1933
11. Adam MA, Speicher P, Pura J, Dinan MA, Reed SD, Roman SA,
Sosa JA (2014) Robotic thyroidectomy for cancer in the US:
123
4792
patterns of use and short-term outcomes. Ann Surg Oncol
21:3859–3864
12. Lang BH, Wong CK, Tsang JS, Wong KP, Wan KY (2014) A
systematic review and meta-analysis comparing surgically-re-
lated complications between robotic-assisted thyroidectomy and
conventional open thyroidectomy. Ann Surg Oncol 21:850–861
13. Sun GH, Peress L, Pynnonen MA (2014) Systematic review and
meta-analysis of robotic vs conventional thyroidectomy approa-
ches for thyroid disease. Otolaryngol Head Neck Surg
150:520–532
14. Wang YC, Liu K, Xiong JJ, Zhu JQ (2015) Robotic thyroidec-
tomy versus conventional open thyroidectomy for differentiated
thyroid cancer: meta-analysis. J Laryngol Otol 129:558–567
15. Tae K, Kim KY, Yun BR, Ji YB, Park CW, Kim DS, Kim TW
(2012) Functional voice and swallowing outcomes after robotic
thyroidectomy by a gasless unilateral axillo-breast approach:
comparison with open thyroidectomy. Surg Endosc
26:1871–1877
16. Song CM, Cho YH, Ji YB, Jeong JH, Kim DS, Tae K (2013)
Comparison of a gasless unilateral axillo-breast and axillary
approach in robotic thyroidectomy. Surg Endosc 27:3769–3775
17. Song CM, Ji YB, Bang HS, Park CW, Kim H, Tae K (2014)
Long-term sensory disturbance and discomfort after robotic
thyroidectomy. World J Surg 38:1743–1748
18. Ji YB, Song CM, Bang HS, Lee SH, Park YS, Tae K (2014)
Long-term cosmetic outcomes after robotic/endoscopic thy-
roidectomy by a gasless unilateral axillo-breast or axillary
approach. J Laparoendosc Adv Surg Tech A 24:248–253
19. Song CM, Ji YB, Bang HS, Park CW, Kim DS, Tae K (2014)
Quality of life after robotic thyroidectomy by a gasless unilateral
axillary approach. Ann Surg Oncol 21:4188–4194
20. Song CM, Yun BR, Ji YB, Sung ES, Kim KR, Tae K (2016)
Long-term voice outcomes after robotic thyroidectomy. World J
Surg 40:110–116
21. Tae K, Song CM, Ji YB, Kim KR, Kim JY, Choi YY (2014)
Comparison of surgical completeness between robotic total thy-
roidectomy versus open thyroidectomy. Laryngoscope
124:1042–1047
22. Lee S, Lee CR, Lee SC, Park S, Kim HY, Son H, Kang SW,
Jeong JJ, Nam KH, Chung WY, Park CS, Cho A (2014) Surgical
completeness of robotic thyroidectomy: a prospective comparison
123
Surg Endosc (2016) 30:4785–4792
with conventional open thyroidectomy in papillary thyroid car-
cinoma patients. Surg Endosc 28:1068–1075
23. Lang BH, Wong CK, Tsang JS, Wong KP, Wan KY (2015) A
systematic review and meta-analysis evaluating completeness and
outcomes of robotic thyroidectomy. Laryngoscope 125:509–518
24. Cooper DS, Doherty GM, Haugen BR, Kloos RT, Lee SL,
Mandel SJ, Mazzaferri EL, McIver B, Pacini F, Schlumberger M,
Sherman SI, Steward DL, Tuttle RM (2009) Revised American
Thyroid Association management guidelines for patients with
thyroid nodules and differentiated thyroid cancer. Thyroid
19:1167–1214
25. Perrier ND (2012) Why I have abandoned robot-assisted
transaxillary thyroid surgery. Surgery 152:1025–1026
26. Inabnet WB 3rd (2012) Robotic thyroidectomy: must we drive a
luxury sedan to arrive at our destination safely? Thyroid
22:988–990
27. Rubin DB, Thomas N (1996) Matching using estimated propen-
sity scores: relating theory to practice. Biometrics 52:249–264
28. D’Agostino RB Jr (1998) Propensity score methods for bias
reduction in the comparison of a treatment to a non-randomized
control group. Stat Med 17:2265–2281
29. Lee SG, Lee J, Kim MJ, Choi JB, Kim TH, Ban EJ, Lee CR,
Kang SW, Jeong JJ, Nam KH, Jo YS, Chung WY (2015) Long-
term oncologic outcome of robotic versus open total thyroidec-
tomy in PTC: a case-matched retrospective study. Surg Endosc.
doi:10.1007/s00464-015-4632-9
30. Tae K, Ji YB, Song CM, Min HJ, Lee SH, Kim DS (2014)
Robotic lateral neck dissection by a gasless unilateral axillobreast
approach for differentiated thyroid carcinoma: our early experi-
ence. Surg Laparosc Endosc Percutan Tech 24:e128–e132
31. Randolph GW, Duh QY, Heller KS, LiVolsi VA, Mandel SJ,
Steward DL, Tufano RP, Tuttle RM (2012) The prognostic sig-
nificance of nodal metastases from papillary thyroid carcinoma
can be stratified based on the size and number of metastatic
lymph nodes, as well as the presence of extranodal extension.
Thyroid 22:1144–1152
32. Mazzaferri EL, Kloos RT (2001) Clinical review 128: current
approaches to primary therapy for papillary and follicular thyroid
cancer. J Clin Endocrinol Metab 86:1447–1463