Prophylactic Bilateral Central Neck Dissection Should Be Carefully Evaluated Based on Prospective Study of 581 Cases With cN0 Papillary Thyroid Carcinoma

Background: prophylactic central lymph node dissection (PCND) had been a basic consensus for patients with papillary thyroid carcinoma (PTC) in China. However, unilateral or bilateral central lymph node dissection (CND)was still controversial. This study aimed at investigating the safety and long-term benet for the patients with bilateral central lymph node dissection (BCCD). Methods: 581 patients were enrolled and divided randomly into the test and control groups according to a different range of CND. 285 patients were prospectively assigned to undergo lobe thyroidectomy plus BCND in the test group, in comparison 296 patients were assigned to undergo lobe thyroidectomy plus ipsilateral central lymph node dissection (ICND) in the control group. Results: We found that the numbers of total LN and N1a in the test group were higher than that of the control group (p = 0.002), but there was no difference in the number of metastasized lymph nodes (p = 0.857) and tumor recurrence (p = 0.308). Additionally, in the aspect of postoperative complication (1 day after surgery), the serum levels of parathyroid hormone in the BCND group were lower than that in the ICND group (P = 0.010), and the numbers of transient laryngeal nerve palsy were higher than that(p = 0.033). meanwhile we further found that tumors size larger than 1cm and tumor side lymph node metastasis were related to the contralateral lymph node metastasis. Conclusion: BCND resulted in more positive lymph nodes and similar surgical complications compared with ICND but did not change the patient's long-term prognosis. It may be an alternative for patients with tumor sizes larger than 1cm. Meanwhile, and it should be done in large medical centers, with rich experience in PG and RLN protection


Introduction
Central lymph node involvement was common in patients with papillary thyroid carcinoma (PTC) [1], which may affect tumor recurrence and patient's survival [2][3][4], but PCND had not been recommended by American Thyroid Association (ATA)guidelines for the reason that it could obviously increase surgical complication which were caused by injury of the parathyroid gland (PG) and recurrent laryngeal nerve (RLN). Nowadays with the development of RLN and PG protection, the safety of CND has been con rmed based on the existing study [5], and PCND was also recommended by our guideline in China [6] for the reason that although central lymph node metastasis did not affect the survival rate of patients, once occurred, it would increase outpatient number, cost, and the psychological pressure especially.
when con rmed with PTC, lobe thyroidectomy plus PCND were considered as the rst treatment option for the patient with unilateral (cN0) PTC in china [6], and occult lymph node metastases were also found frequently in the nal para n pathology, so there was no debate in the role of the PCND in China [7][8], except for the extent of the PCND. Basing on the safety of the ipsilateral PCND, bilateral PCND including removal of prelaryngeal, pretracheal, and bilateral paratracheal nodes was proposed in patients with unilateral PTC in recent years [9][10].
BCND could decrease the risk of overlooking contralateral metastases but add a higher risk of RLN palsy and hypoparathyroidism. Now with the development of RLN and PG protection, the risk mentioned above had decreased signi cantly. More comparative studies had suggested that BCND may be an effective alternative to ICND for cN0 PTC because of its long-term oncologic outcome and the similar complications [11].
Although more retrospective studies reported that there were a large number of occult metastasis lymph nodes in the contralateral region, there were still many controversies in the eld of tumor recurrence and metastasis for lacking a prospective randomized controlled study. It was hard to sure if the patients with CN0 PTC would bene t from BCND in the long-term oncologic outcome [12][13]. In the present study, prospective randomized controlled studies with standard BCND were rare, especially the speci c dissecting range and operative process. We aimed to prospectively compare the results of ICND and BCND and provide a clinical evidence for bilateral central lymph node dissection in patients with unilateral cN0 PTC.

Patients characteristics
A total of 581 consecutive patients were enrolled in this study from 2015.12 to 2017.12. The characteristics of patients were shown in table 1. All the patients were diagnosed as unilateral cN0 PTC.
They were randomly assigned to one of the following groups: (1) lobe thyroidectomy plus ICND (control group); or (2) lobe thyroidectomy plus BCND (test group). Exclusion criteria included previous neck surgery, any preoperative (i.e., clinical and ultrasound examination) evidence of bilateral PTC, macroscopically in ltrating tumors, isthmus cancer, age more than 55 years, or distant metastases. The randomization was done by using computer-generated random number tables. All the operations were performed by the same surgeon. The study had been registered with North America Clinical Practice (NCT02648399) and approved by the Ethics Committee of the Fujian Medical University Union Hospital (2015KY026), and all patients provided informed consent.

Study Endpoints
The primary endpoint was to compare the long-term (2 years) oncological outcome, while the secondary aim of the study was to compare postoperative complications, operative time, and the numbers of total and metastatic LN, T and N stage between the two groups.
2.3 De nitions PTC were de ned as clinically cN0 in the absence of any preoperative (i.e., clinical and ultrasound examination) evidence of multifocal disease or lymph node involvement, respectively. All surgical procedures were performed by an experienced endocrine surgeon. ICND included removal of prelaryngeal, pretracheal, and paratracheal lymph node on the tumor side. BCND included the removal of prelaryngeal, pretracheal, tumor side and contralateral paratracheal lymph node. Due to the presence of the thyroid lobe, there were some differences in the dissection of the paratracheal lymph node on the contralateral side. The upper bound of the dissection was generally located at the level of the inferior thyroid artery. First, we separated the outer side of the thyroid gland, pulled the lower pole of the thyroid lobe upward and inside, then continued separating downward to the thymus level in the caudal direction and protecting thymus vein (Fig 1). We could nd the inferior parathyroid gland, which was closely related to the thymus gland named thymus-related inferior parathyroid gland and attempt to preserve it in situ. We also nd the characteristics that the right RLN intersected with the common carotid artery on the right side, and the RLN was close to the surface of the esophagus on the left side [14]. Therefore, we could quickly nd the RLN in the lower part of the neck and separated along the surface of RLN to the level of the inferior thyroid artery (Fig 2). The extent of dissection under the inferior thyroid artery was consistent with that of conventional dissection of the tumor side (Fig 3) [15]. Pathological staging was de ned in accordance with the 2017 National Comprehensive Cancer Network pTNM staging system.

Follow-up and postoperative treatment of hypoparathyroidism
Laryngoscopy was performed both preoperatively and postoperatively in all patients to check vocal cord motility. Complications were considered de nitive if they lasted more than six months. Brie y, postoperative suppressive levothyroxine treatment was administered to all patients. All patients underwent thyroid (parathyroid) function (TSH, T3, T4, PTH, Ca) measurements under suppressive levothyroxine treatment after 2 months of surgery and ultrasound neck scan after 12 months of surgery. Follow-up data were obtained by outpatient consultations or telephone contact.

Data Collection
General characteristics, intraoperative factors, pathologic examination, the number of lymph nodes and metastatic lymph nodes in the resected specimens, and postoperative complications were collected retrospectively. The seventh edition of the American Joint Committee on Cancer (AJCC) staging was used for all the recruited patients. The primary endpoints were tumor recurrence or metastasis, the second endpoints were the operative time, the numbers of the total LN and the metastatic LN, the ratio of PG auto-transplantation, the serum levels of PTH, the postoperative complications, the ratio of IONM.

Statistical Analysis
Statistical analysis was performed by SPSS 26.0, Chicago, IL, USA. All values were presented as mean± standard deviation. A T-test or Chi-Square test was used to determine statistical signi cance, requiring p<0.05 was considered to be statistically signi cant.

Patents characteristics
There were 296 and 285 patients, respectively, in the ICND and BCND group. And no signi cant difference was found between the two groups concerning age, sex, tumor size, number of lesions, T stage, the ratio of IONM, PTH and Ca (pre-operation) ( Table 1).

Short-term Results
Patients in the BCND group experienced more operative time (p = 0.006) and more N1a. (p = 0.004). Also, with more PG numbers of auto-transplantation and removed lymph nodes in the BCND group than that of the ICND group. But there was no signi cant difference in the number of metastasized lymph nodes, T stage. Additionally, in the aspect of postoperative complication, the serum levels of PTH (1 day after surgery) in the BCND group (P = 0.010) were lower than that in the ICND group, and the numbers of transient laryngeal nerve palsy (p = 0.033) were higher than that. while in the aspect of post-operation hemorrhage, Ca (1 day after surgery), the serum levels of PTH (6 months after surgery), Ca (6 months after surgery), there were no signi cant difference. (Table 1).

Long-term Results
No signi cant difference was found in the number of tumor recurrence. Additionally, in the aspect of postoperative complication, no signi cant difference was found including de nitive hypoparathyroidism, de nitive laryngeal nerve palsy. There were 65 and 220 patients, respectively, with contralateral side LN (+) and contralateral side LN (-) in the BCND group. No signi cant difference was found between the two subgroups in the BCND group including age, sex, T stage, number of lesions, lesion located in the lower pole (Table 2), except the tumor side LN (+) (p = 0.000) and tumor size(p=0.000) ( Table 2).

Discussion
The importance of the ICND had been accepted by the most surgeons in China for the reason that occult or microscopic lymph node metastases of the tumor side were found in 50-80 % of patients with CN0 PTC [6,[16][17]. Unfortunately, ICND still carried the risk of contra central lymph node metastases being overlooked in approximately one-quarter of patients [18]. This could imply a potential risk of tumor recurrence and residual. The central lymph node as a whole contains prelaryngeal, pretracheal, bilateral paratracheal lymph node, so some scholars suggest that we should dissect bilateral central lymph node called BCND which may change the TNM stage and reduce the risks of tumor recurrence [19][20].
Conversely, one of the main arguments against BCND was the higher risk of complications including parathyroid and RLN injury, and that there was no long-term bene t for the patients [21][22].
To dispute the controversy, we designed this study to investigate the feasibility of BCND and its impact on complications and long-term bene ts in patients. Regarding the result, BCND was associated with prolonged operative time compared with the ICND, but it just took about 10 minutes longer at our institution. And we found there were 54.73% and 66.32% lymph node metastases in the group of ICND and BCND respectively, it was consistent with the existing research conclusions that occult lymph node metastases are frequent in cN0 PTC, especially in the BCND group(p = 0.004). But there was no difference in the number of tumor recurrence, metastatic LN, and T stage. Meanwhile the serum levels of PTH (1 day after surgery) in the BCND group (1 day after surgery) (P = 0.010) was lower than that in the ICND group, and the numbers of PG auto-transplantation (p = 0.025) and transient laryngeal nerve palsy (p = 0.033) were higher than that. Basing the result mentioned above, BCND didn't seem to gain more bene ts for the patients with cN0 PTC.
But the permanency complications rate, including the injury of permanent parathyroid and RLN, were not signi cantly different between the two groups. Based these data, we seem to see a reason that BCND could be carried out. What caused it? the reasons for this were as following. First, with the help of new ways of nding RLN [14], we can protect it as more as possible, even though there were only about 30% patients with IONM because of health policy. Meanwhile, there were more patients with transient laryngeal nerve palsy in the BCND group than that of the ICND group, but all the patients recovered after 2 months in the function of the vocal cords which was con rmed by the electronic laryngoscope. Moreover, with the improvement of PG recognition and auto-transplantation, the function of the PG (Postoperative 6 months) in the BCND group was the same with groups of ICND. We had to emphasize the importance of PG transplantation as soon as possible, especially in the patents that the PG was not in situ or misresection [23]. with the help of new ways of nding RLN, PG recognition and auto-transplantation [24][25][26], BCND may be safe in the patients with cN0 PTC in our study based on the long-term results because it may guarantee an adequate surgical resection in patients with occult nodal disease and reduced the risk of tumor recurrence theoretically. So BCND may be a feasible treatment option, but who was suitable for BCND will be a question to be discussed.
We further found that tumors size larger than 1cm and tumor side lymph node metastasis were related to the contralateral lymph node metastasis in patients with PTC. As the table shows, the total number of lymph nodes dissected was more than 10, so it was too di cult to make a rapid and accurate diagnosis in 30 minutes for the pathologist, especially in the condition of occult metastasis or large specimen, or in some primary hospitals. So, it's not feasible that we dissect the contralateral side LN after knowing the result of the frozen pathology of the tumor side LN during operation [27]. In addition, there was no difference in tumor recurrence between the two groups that may be related to the inert biological behavior of thyroid cancer, and its follow-up time should be longer. Therefore, we do not recommend routine BCND in all, except for patients with tumor size larger than 1 cm.
The major advantage of the present study was randomization. It was clear that this method of patient selection may decrease some concerns regarding recruitment bias; meanwhile, patients in the two groups were operated during the same period, and all the involved surgeons performed both surgical procedures. But we had to point out that our study had some limitations. Firstly, the major disadvantage was the lack of follow-up data for 5/10-years tumor recurrence, even if there was no signi cant difference in terms of 2 -years tumor recurrence between two groups. Secondly, due to the dissection of contralateral lymph node and the destruction of the original tissue structure, it was not conducive to the next operation, and contralateral central lymph node dissection could affect the PG, therefore the status of the PG should be recorded clearly on the medical documents which was useful for next operation. Thirdly, we did not detect tumor genetic mutation and use IONM routinely. After separating the outer side of the thyroid gland and pulling the lower pole of the thyroid lobe upward and inside, then we could quickly nd the RLN and protect the PG.

Figure 2
This is the anatomical display after bilateral central lymph node dissection (BCND). We can protect the recurrent laryngeal nerve and parathyroid gland very well.

Figure 3
This is the anatomical display after ipsilateral central lymph node dissection(ICND).