Comparison of Long-Term Outcomes Between Ipsilateral and Bilateral Central Neck Dissection in Occult Contralateral Central Lymph Metastasis in Papillary Thyroid Carcinoma With Unilateral Lateral Neck Metastasis Using Propensity Score Matching

This study aimed to evaluate long-term prognosis of contralateral central neck dissection in papillary thyroid cancer (PTC) patients with ipsilateral lateral neck metastasis. We aimed to produce clinical evidence to help determine the extent of central neck dissection (CND) focusing on the separation between the ipsilateral and contralateral sides. A total of 708 PTC patients who underwent total thyroidectomy and concomitant ipsilateral or bilateral CND with ipsilateral lateral neck dissection (LND) were retrospectively included between January 1997 and December 2015 at a single institution. The median follow-up time was 89.7 months, the mean age was 44.7 years and the mean tumor size was 1.5 cm. Among the study population, 507 were female (71.5%) and 201 (28.5%) were male. Locoregional recurrence (LRR) was observed in 26 (7.9%) patients and 30 (7.9%) patients in the ipsilateral and bilateral CND groups, respectively. There were 6 (1.8%) contralateral recurrence cases in the ipsilateral CND group and 5 (1.3%) cases in the bilateral CND group. Male sex (adjusted HR = 1.857, p = 0.034), larger tumor size (adjusted HR = 4.298, p = 0.006), and more metastatic ipsilateral central lymph nodes (adjusted HR = 1.078, p = 0.014) signicantly increased the risk of LRR. Ipsilateral CND only did not signicantly increase the risk of LRR (adjusted HR = 1.110, p = 0.712). There were no signicant differences in recurrence according to contralateral central neck dissection status after propensity score matching (p = 0.424), either. The incidence of hypocalcemia (p = 0.007) was higher in the bilateral CND group compared to the ipsilateral CND group. Surgeons may consider performing contralateral CND only for therapeutic purposes to reduce unnecessary complications. CNM central neck metastasis, CCND contralateral central neck dissection, CCNM contralateral central neck metastasis.


Introduction
Cervical lymph node metastases in papillary thyroid carcinoma (PTC) are quite common 1,2 and are associated with increased locoregional recurrence (LRR) [3][4][5][6][7] . However, PTC has a relatively good prognosis compared to other types of thyroid cancers 8 . For this reason, local recurrence and postoperative quality of life remain important issues. The 2015 ATA guidelines for thyroid cancer management recommend that prophylactic CND (ipsilateral or bilateral) should be considered in clinically involved lateral neck nodes 9 , but there is no detailed guideline for the extent of CND with the separation of the ipsilateral or contralateral compartment, and the extent of prophylactic CND depends on the preference of the individual surgeon. To date, there are various studies about the central compartment, however, only a few are about the contralateral central compartment in PTC patients with ipsilateral lateral neck metastasis. Some studies have been conducted to examine contralateral central neck metastasis status and its associated factors but without long-term outcomes dealing with the extent of CND 6, 10 . It is clinically signi cant to clarify long-term clinical outcomes for contralateral central neck dissection in PTC patients with ipsilateral lateral neck metastasis because of its important role in determining the extent of CND. It is certain that the incidence of surgery-related morbidities, such as hypoparathyroidism or hypocalcemia or VCP, were observed more prominently in bilateral CND compared to ipsilateral CND [11][12][13][14][15] . The purpose of this study was to evaluate the recurrence rate and the risk factors associated with recurrence in accordance with prophylactic contralateral CND. This, in turn, could potentially decrease postoperative morbidities from unnecessary neck dissection.

Methods
Patient selection. This study was approved by the Institutional Review Board at Samsung Medical Center(SMC) and all process were done in accordance with the relevant guidelines and regulations. Informed consent form from all participated patients in this study was obtained. We conducted a retrospective study at a single institution between January 1997 and December 2015. A total of 1,026 patients underwent lateral neck dissection (LND) with concomitant total thyroidectomy at the Thyroid Cancer Center of SMC, a tertiary referral center in Korea. Patients with the following were excluded: preoperative contralateral cN1a patients, a history of previous thyroidectomy, an age younger than 20 years or older than 80 years, non-PTC carcinomas (follicular/medullary/anaplastic), PTC variants, mixed type PTC, LND without thyroidectomy due to lateral neck recurrence, completion thyroidectomy, bilateral LND, distant metastasis, and follow-up duration less than 6 months (residual tumor or suspicious lymph node detected in the bilateral CND group within 6 months after initial surgery or loss to follow-up within 6 months) 16 . Finally, a total of 708 PTC patients who underwent total thyroidectomy with either ipsilateral central neck dissection (CND) + ipsilateral LND or bilateral CND + ipsilateral LND were enrolled (Fig. 1).
All patients were examined by ultrasonography (US) and/or computed tomography (CT) preoperatively to evaluate primary tumor and the presence of suspected nodal metastases. Korean Thyroid Imaging Reporting and Data System(K-TIRADS) by The Korean Society of Thyroid Radiology (KSThR) was used as a guideline for US-based diagnosis. Suspicious lesions were diagnosed by ne needle aspiration (FNA). Suspicious lymph nodes in the lateral compartment were con rmed by FNA or thyroglobulin (Tg) washout measurement. Based on the American Thyroid Association (ATA) Guidelines 9 , clinically involved lymph node disease (cN1) was de ned as lymph node metastases in the central neck compartment on preoperative imaging 16 .
Surgical methods. Surgical strategies were chosen according to the ATA Guidelines 9 . All patients underwent total thyroidectomy with either ipsilateral CND + ipsilateral LND or bilateral CND + ipsilateral LND. Therapeutic contralateral CND was typically performed after central lymph node metastasis was detected during preoperative US or surgery. Patients without contralateral central lymph node metastasis by preoperative exams underwent prophylactic contralateral CND at the discretion of the surgeon. The operations were performed by three endocrine surgeons (A, B, C) in our hospital; surgeon A performed routine bilateral CND and surgeon B and C preferred to perform bilateral CND only if contralateral central lymph node metastasis was suspected using preoperative images or gross inspection during surgery regardless of pathological contralateral central neck metastasis 16 . Approximately 61% of the patients with bilateral CND were patients who underwent surgery by surgeon A. The rest 39% of the patients with bilateral CND were patients who underwent surgery by surgeon B and C. Thyroditis with CLN enlargement is common in Korean patients who live in an iodine-rich area. Ultrasound features can be diverse depending on the severity of disease in patients with thyroditis. Among the patients who underwent surgery by surgeon B and C, 51% presented with CLT. In addition, our analysis is based on the cohort without radiologically detectable contralateral central lymph node metastasis. The term "ipsilateral" was used to indicate the same side as the main tumor, and "contralateral" was de ned as the opposite side relative to the main tumor. In cases of bilateral tumors, the largest tumor was considered to be the main tumor. Ipsilateral central lymph nodes were de ned as ipsilateral paratracheal, pretracheal lymph nodes, and precricoid (Delphian) nodes. Lymph nodes in the paratracheal region that were contralateral to the main tumor or opposite to a lateral neck metastasis were de ned as contralateral central lymph nodes. LND refers to a lateral neck compartmental lymph node dissection. LND was de ned as a neck dissection of levels II-V, conserving the internal jugular vein, spinal accessory nerve, and sternocleidomastoid muscle. Therapeutic LND was performed only in patients with lateral neck metastasis con rmed by FNA preoperatively. LND specimens were separated by the surgeon according to the neck level at the time of surgery 16 .
Histopathological examination of surgical specimens. Surgical specimens were microscopically examined by two or more experienced pathologists. The following histopathologic factors were assessed: main tumor size (longest diameter of the largest lesion), cell type of main tumor, multifocality, tumor bilaterality, extrathyroidal extension (ETE; microscopic or gross), regional LN metastasis (central or lateral compartment), and underlying conditions of the thyroid, such as chronic lymphocytic thyroiditis (CLT). The staging of thyroid cancer was determined in accordance with the 7th edition of the AJCC Cancer Staging Manual and the Future of TNM 17 .
Surgery-related outcomes. Various surgery-related outcomes were assessed. The terms "transient" and "permanent" were de ned on a six-month basis. "Transient" was de ned as a symptom duration of less than 6 months and "permanent" was de ned as a symptom duration of more than 6 months. Hypocalcemia was de ned as a serum calcium < 8.0 mg/dl. Calcium level were the lowest value within or more than six months of surgery accordingly. Recurrent laryngeal nerve injury described as vocal cord palsy (VCP) was diagnosed by laryngoscopy or patient's symptoms. Postoperative serum thyroglobulin (Tg) level in Table 1 was de ned as the stimulated Tg after the rst ablation. Postoperative follow-up and management. After the initial surgery, all patients underwent regular followup at 6-to 12-month intervals with clinical evaluations including physical examinations, US, CT, Iodine-131 ( 131 I) scans, and serum unstimulated Tg level. Suspicious lesions for recurrence were con rmed by US-guided FNA biopsy with or without Tg washout level and/or CT or positron emission tomography (PET). Locoregional recurrence was de ned as the presence of tumors or metastatic lymph nodes on cytology from FNA. Radioactive iodine (RAI) therapy was performed with 131 I at 4-12 weeks after surgery according to the ATA guidelines 9 . RAI was administered after thyroid hormone withdrawal or after stimulation with recombinant thyroid-stimulating hormone. When RAI treatment was no longer required, patients resumed regular follow-up. The last date of follow-up was de ned as loss to follow-up, withdrawal, recurrence, or death. In this study, all patients received postoperative RAI therapy, and TSH suppression was performed according to the ATA guidelines 9 .
Statistical analysis. Statistical analyses were performed using SPSS version 22.0 software (IBM Corp, Armonk, NY, USA), and statistically signi cant differences were de ned as P-values less than 0.05. Continuous variables are presented as mean ± standard deviation (SD), and categorical variables are presented as the number with percentage (%) and odds ratio (OR). The chi-square test and linear-by-linear association were used for categorical variables and the Student's t-test for continuous variables. For recurrence analysis, Cox regression was used to determine if the clinicopathological characteristics were associated with contralateral CND in patients with PTC. Propensity score matching was performed to adjust clinicopathological variables which can effect on recurrence. Under the matched condition, Kaplan-Meier and the log-rank test were adopted to analyze time-dependent LRR.

Results
Clinicopathological characteristics of 708 PTC patients who underwent total thyroidectomy with either ipsilateral CND + ipsilateral LND or bilateral CND + ipsilateral LND. The mean patient age was 44.7 years (range, 20-77) and the mean tumor size was 1.5 cm (range, 0.  In the bilateral CND group, there were 30 (7.9%) patients with LRR. Recurrence occurred at the ipsilateral thyroidectomy site in 6 (1.6%) patients, contralateral thyroidectomy site in 1 (0.3%) patients, ipsilateral central neck in 2 (0.5%) patients, and ipsilateral lateral neck in 18 (4.8%) patients. Contralateral recurrence occurred in 4 patients overall; there were 1 (0.3%) case in the contralateral thyroidectomy site, 3 (0.8%) cases in the contralateral lateral neck. There was no contralateral central neck recurrence in either groups (  Prognostic impact of contralateral central node dissection status on locoregional recurrence in 706 PTC patients after propensity score matching. Propensity Score matching was performed to compare "contralateral central node dissection" with "no contralateral central node dissection". Before propensity score matching, tumor size (p = 0.012), multifocality (p = 0.004) were more frequently observed in contralateral central neck dissection group. The patients who underwent contralateral central neck dissection received relatively higher dose RAI (117.6 mCi vs 142.2 mCi, p < 0.001) ( Table 4). After propensity matching, there were 130 matched pairs. There was no signi cant difference in patients' characteristics between two groups and loco-regional recurrence was analyzed. The median follow-up time was 93.7 months (range, 6-212 months). Of 11 recurred patients, 8(6.2%) and 3(2.3%) patients were in CCND (-) group and CCND (+) group, respectively. Recurrence-free survival rates in either ipsilateral or bilateral CND groups were 97.8% versus 97.6% at 5 years, 75.5% versus 97.6% at 10 years, respectively (p = 0.424) (Fig. 2). Furthermore, the recurrence-free survival rates for overall contralateral recurrence, including thyroidectomy site, central or lateral neck, in the ipsilateral or bilateral CND groups were 100.0% versus 99.2% at 5 years, 86.8% versus 99.2% at 10 years, respectively (p = 0.609) (Fig. 3).
There were no statistically signi cant differences in RFS rates according to contralateral central neck dissection status.   (Table 5).

Discussion
Previous studies have reported that the rate of regional lymph node metastasis in patients with PTC is approximately 30% up to 80% 18, 19 and central lymph node metastasis is already present in 40-60% of patients at the time of diagnosis [20][21][22] . Other studies demonstrated that the rate of contralateral central neck metastasis in patients with ipsilateral lateral neck metastasis was 34.7% and 5-30% in the nal pathology, respectively 23,24 . Since these rates were relatively high, there has been a controversy as to whether or not surgeons should perform contralateral central neck dissection (CND) without suspicious contralateral central lymph nodes on preoperative imaging. In addition, there have not been many studies on this subject to date, therefore the long term effect of prophylactic contralateral CND in PTC patients with ipsilateral lateral neck metastasis are still unclear. To the best of our knowledge, this study is one of the few studies to investigate the long-term outcome of prophylactic contralateral central neck dissection in PTC patients with ipsilateral lateral neck metastasis.
In our study, male sex, larger tumor size, and greater number of metastatic ipsilateral CLNs signi cantly increased the risk of LRR, however, ipsilateral CND only did not increase the risk of recurrence (Tables 3  and 4, Figs. 2 and 3). In the ipsilateral CND group, the rate of recurrence in the ipsilateral central or lateral neck was higher than those in the contralateral central or lateral neck ( Postoperative complications such as hypocalcemia were higher in patients who underwent bilateral CND than those who underwent ipsilateral CND (Table 4). This is a critical issue in thyroid surgery because hypocalcemia is closely associated with the patients' quality of life, and all endocrine surgeons should not ignore it.
Given the results above, prophylactic contralateral CND may not be necessary in patients with ipsilateral cN1b as a routine procedure. These retrospective data suggest that PTC patients with ipsilateral N1b presenting with evident lymphadenopathy intraoperatively or on preoperative imaging in the contralateral central compartment should undergo therapeutic contralateral CND. Unless there is an evident contralateral metastatic lymph node preoperatively or a detectable contralateral cN1a intraoperatively, ipsilateral CND may be su cient surgical treatment of the PTC patients with ipsilateral N1b.
This study had several limitations. First, our study was a non-randomized, retrospective, cohort study. Confounding variables or unmeasured factors may not have been identi ed, and there was possible surgeon bias in the decision to undergo ipsilateral CND versus bilateral CND. That is, the whole cohort was not treated in a random fashion. Thus, further prospective, randomized trials are needed to resolve this selection bias issue in order to clearly compare recurrence rates between the ipsilateral CND and bilateral CND groups. Second, a longer follow-up period is required to assess recurrence and survival more completely. Third, inter-observer variation in the detection and interpretation of cervical lymph node metastasis and inconsistent surgical management were involved because of the long-term period of data collection. Fourth, our ndings may not be applicable to smaller centers without a skilled team such as high volume surgeons or radiologists since the experts' factor could contribute to the low incidence of recurrence and contralateral recurrence. Performing prophylactic contralateral CND may rather decrease recurrence and help postoperative treatment planning and follow-up in small institutions with less skill in preoperative node evaluation; even this may demonstrate higher microscopic nodal disease. Despite these limitations, this study still has the value of a retrospective study in examining the role of CND in PTC as a prospective randomized trial may not easily feasible 28 . By analyzing adequate sample size in a single institution, we tried to minimize inadequate statistical power and local institutional variability using speci c inclusion factors and follow-up.
In conclusion, This study suggests that bilateral central neck dissection increased surgery-related complications, especially hypocalcemia, and may not reduce the risk of recurrence in PTC patients with unilateral N1b; therefore, surgeons may perform contralateral CND only in the presence of clinically evident or suspected nodal disease.

Declarations Funding Support
This work had no speci c funding.

Author Disclosure Statement
The authors have no con icts of interest to declare.   Recurrence-free survival rate for patients with ipsilateral lateral neck metastasis according to the extent of central neck dissection after propensity score matching.

Figure 3
Contralateral recurrence-free survival rate for patients with ipsilateral lateral neck metastasis according to the extent of central neck dissection after propensity score matching.