Significance of Size of Persistent/Recurrent Central Nodal Disease on Surgical Morbidity and Response to Therapy in Reoperative Neck Dissection for Papillary Thyroid Carcinoma

Abstract

Background:

To balance the risk of disease progression, morbidity, and efficacy of reoperative central neck dissection (RCND) in papillary thyroid carcinoma, the latest clinical guidelines recommend early surgery over surveillance when the largest diseased node is >8 mm in its smallest dimension. However, the evidence remains scarce. To determine an appropriate size for first-time RCND, the relationship between size of largest diseased central node, morbidity, and response-to-therapy following RCND was examined.

Methods:

A total of 130 patients who underwent RCND following initial surgery for persistent/recurrent nodal disease were reviewed. Patients with largest diseased central node measured preoperatively by ultrasonography were included. Eligible patients were categorized into three groups: largest central node <10 mm (group I), 10–15 mm (group II), and >15 mm (group III). Surgical morbidity and response to therapy at one year after RCND were compared between groups. To evaluate biochemical response, patients with structural incompleteness were excluded.

Results:

Group III not only had significantly more high-risk tumors (by American Thyroid Association risk stratification) at initial therapy (64.5% vs. 44.4%, respectively; p = 0.038), but this group also a higher risk of extranodal extension (35.5% vs. 16.0%; p = 0.055), recurrent laryngeal nerve involvement (19.4% vs. 0.0%; p < 0.001), incomplete surgical resection (48.4% vs. 7.4%; p < 0.001), new-onset vocal cord paresis (16.7% vs. 2.5%; p = 0.017), overall surgical morbidity (22.6% vs. 7.4%; p = 0.021), and biochemical incompleteness (80.6% vs. 67.9%; p = 0.004) than groups I and II combined did. However, overall morbidity did not differ between groups I and II (5.7% vs. 8.7%; p = 0.694). After adjusting for American Thyroid Association risk stratification, only the size of the largest diseased central node ≥15 mm (odds ratio = 7.256 [confidence interval 1.302–40.434], p = 0.001) was an independent risk factor for biochemical incompleteness following RCND.

Conclusions:

Patients with larger diseased central node(s) had a significantly higher risk of local invasion, surgical morbidity, and biochemical incompleteness. Relative to nodal size <10 mm, size >15 mm in the largest disease central node was an independent risk factor for incomplete biochemical response, while nodal size 10–15 mm was not. These findings imply that the recommended threshold of 8 mm might be too stringent and could be raised to 15 mm without increasing the surgical morbidity from RCND.

Introduction

Papillary thyroid carcinoma (PTC) is the most common type of thyroid cancer, and its age-adjusted incidence has doubled in the last 30 years (1). However, despite the excellent prognosis, with a disease-specific survival >90%, persistent/recurrent disease remains common (2). Depending on the initial tumor risk and response to initial therapy, up to 30–40% of PTC patients may have persistent/recurrent disease in the central (level VI) and/or lateral (levels II–V) compartments of the neck (3).

With greater use of high-resolution ultrasonography (US) and ultrasensitive thyroglobulin (Tg) assays, an increasing number of small-volume, subclinical persistent/recurrent central nodal disease is being diagnosed during routine follow-up surveillance. However, because the majority of these lesions do not progress and do not cause locoregional complications, active surveillance (i.e., no immediate surgery) and observing for growth have generally been advocated (3 –6). However, in order to balance the risk of locoregional complications (as a result of no immediate therapy) and the morbidity and efficacy of a therapy such as a reoperative central neck dissection (RCND), the latest guidelines have recommended the decision for RCND should be based on the absolute size of the largest diseased central node (in its smallest dimension) (3,6). Under the current recommendation, if the diseased central nodal node is >8 mm in the smallest dimension on preoperative imaging, early RCND is preferred over active surveillance, while for a smaller diseased node (i.e., ≤8 mm), a reasonable initial approach would be to adopt active surveillance after considering a number of factors such as proximity to adjacent vital structures (i.e., trachea, esophagus, and recurrent laryngeal nerve [RLN]), functional status of vocal cords (VCs), patient comorbidities, high-grade histology, Tg doubling time, and radioactive iodine (RAI) avidity (3,6). Although 8 mm appears a reasonable threshold given the small confined space within the central compartment, the evidence is anecdotal. To the authors' knowledge, no studies have specifically evaluated the impact of size of central nodal disease on the morbidity and oncological outcome (such as the response to therapy) following RCND. The current study hypothesized that both morbidity and incomplete response to therapy would increase as the size of the central nodal disease increases. To determine a size threshold that may increase surgical morbidity and incomplete response to RCND, the relationship between size of largest diseased central node and surgical morbidity and response to therapy following RCND was examined.

Patients and Methods

This retrospective analysis was approved by the local Institutional Review Board. All consecutive patients who underwent RCND at the authors' institution for persistent/recurrent central nodal disease from 1996 to 2011 were reviewed. RCND was defined as a unilateral or bilateral CND in a patient previously operated on in the central neck. Patients were included who had already undergone a total thyroidectomy with or without CND and/or lateral neck dissection and RAI ablation. Since prophylactic CND was not routinely performed until 2009, the majority of the CNDs at initial surgery were therapeutic in intent. To be eligible, patients had to have histologically confirmed PTC harbored in the central compartment and undergone a formal preoperative US examination of the neck by an experienced ultrasonographer at the authors' institution. During the examination, the size and location of the largest diseased central node (relative to the trachea/esophagus; i.e. pretracheal or paratracheal/esophageal) and the presence of concomitant lateral nodal involvement were recorded. The diagnosis of nodal metastasis was generally made by fine-needle aspiration cytology (FNAC), although occasionally it was based on a combination of clinical suspicion and suspicious sonographic characteristics. The size of the largest diseased central node was measured to the nearest 0.1 mm in three orthogonal dimensions, and the smallest dimension (in mm) was taken as the size of the nodal disease. To examine the association between size of largest diseased central node (measured by preoperative US) and subsequent RCND-related morbidity and response to therapy following RCND, patients were categorized into three groups: those with the size of the largest diseased central node <10 mm (group I), 10–15 mm (group II), and >15 mm (group III). To evaluate the biochemical response one year following RCND, patients with structural incompleteness after RCND were excluded. Structural incompleteness was defined as persistent or newly identified locoregional or distant disease seen on US, computed tomography/magnetic resonance imaging, and/or fluorodeoxyglucose positron emission tomography scan and confirmed by FNAC or histology. Biochemical incompleteness was defined as a stimulated thyroglobulin (sTg) ≥2 ng/mL.

For analysis, only the first RCND (i.e., the first reoperation involving the central compartment following initial operation) was analyzed. Patients with remnant recurrence or missing data on their initial operation, histology, perioperative VC and calcium status, and postoperative sTg and thyroid autoantibodies were excluded.

Decision for RCND

The decision was generally conducted as a joint decision-making process between the patient and the multidisciplinary team. All patients having central nodal disease on routine US were discussed in the multidisciplinary meeting. Patient preference and comorbidities, concomitant disease in the lateral compartment or distant sites, rate of disease growth, location of disease relative to the trachea, esophagus, and RLN were considered before a decision was made. There was no clear nodal size threshold for RCND during the study period.

Surgical treatment

Details of the surgical technique have been previously described (7,8). In general, only the central compartment with clinically identifiable disease was dissected, and the operation involved removal of all nodes and fibro-fatty tissue extending vertically from the RLN insertion and laterally from the medial border of the common carotid artery to the midline of the trachea. Whenever possible, the RLN was identified first and traced along its course before proceeding to dissection. In a densely scarred field, an intraoperative nerve stimulator probe was sometimes used for mapping out the course of the RLN (7). For those with RLN involvement by nodal disease, the involved RLN was meticulously dissected out by a combination of blunt and sharp dissection. The use of diathermy near the nerve was kept to a minimum. In case of RLN encasement, to achieve complete resection, the involved nerve was generally transected with the nodal disease, unless there was pre-existing contralateral VC paresis (VCP). After excision, the largest central node (that corresponded to the preoperative US) was marked with a stitch before submission to histopathology.

Histopathology

To ensure consistency, all excised specimens from eligible patients were re-examined by one pathologist (T.W.S.). Upon receiving the excised specimen in the laboratory, the outer surface of the nodes was inked and then cut into uniform slices each around 5 mm in thickness. A margin was deemed positive if tumor cells were found at the inked border (i.e., R1 resection). Histologic findings including number of central lymph nodes (CLN) retrieved, number of positive CLNs, CLN ratio [(positive CLNs/CLNs retrieved) × 100], presence of extranodal extension, and size of the largest metastatic focus were recorded. In the latter, the largest metastatic focus was measured to the nearest 0.1 mm by a built-in computerized measurement program on the microscope (Nikon Eclipse E600), and the smallest dimension of the largest metastatic focus was taken as the size of the diseased node. For patients having multiple metastatic foci, only the largest microscopic focus was measured.

Assessment of RCND-related morbidity

Both VCs were examined within one day before and one week after surgery with flexible laryngoscopy. Any reduced VC movement was recorded as VCP. Those with VCP were examined every eight weeks by otolaryngologists in the first six months. The presence of VCP lasting six months or more was regarded permanent. For hypocalcemia, those taking calcium ± calcitriol supplements were slowly weaned off their supplements. By definition, those who discontinued all supplements in the presence of normocalcemia less than six months after surgery were regarded temporary, while those who continued for six months or more were classed as permanent.

Adjuvant treatment, response to therapy, and surveillance

After RCND, all patients received an adjunctive RAI dose of 3 GBq or 80 mCi. External beam radiation therapy (EBRT) was considered in patients with incomplete surgical resection. To assess response to therapy, a post-RAI sTg level was taken one year after RAI. Tg autoantibodies were measured at the same time. Thyrotropin suppression to 0.5–2.0 mIU/L was recommended for high- and intermediate-risk patients. A follow-up visit was conducted every four to six months in the first two years, every six to eight months in the subsequent three years, and annually thereafter. Clinical examination, neck US, and non-stimulated Tg level were done during follow-up visits.

Laboratory methods

All postoperative Tg levels were measured in the same laboratory. However, the Tg assays were changed over time. Before August 2004, a radioimmunometric assay was used. After that, the assay was changed to an immunometric assay (Immulite 2000; Siemens Healthcare Diagnostics Products Ltd, Gwynedd, United Kingdom). This was calibrated against the CRM-457 standard. The reference range was <55 ng/mL, and analytical sensitivity was <0.2 ng/mL.

Statistical analysis

Statistical analysis was performed by the chi-square or Fisher's exact test to compare categorical variables, and the Mann–Whitney U-test was used to compare continuous variables between groups. For correlation between two continuous variables, Pearson's correlation test was performed. Variables that were significant in the univariate analysis were entered into multivariate analysis. Binary logistic regression analysis with a variable entrance criterion of ≤0.05 was conducted to identify independent factors associated with biochemical incompleteness after RCND. All statistical analyses were performed using SPSS Statistics for Windows v18.0 (SPSS, Inc., Chicago, IL).

Results

A total of 130 patients underwent first-time RCND. All procedures were done by the same group of surgeons. All patients had histologically confirmed PTC, and their largest diseased central node was preoperatively measured by US. After excluding those with thyroid remnant recurrence (n = 5) and insufficient data (n = 13), 112 patients were eligible. As initial therapy, all patients had a total thyroidectomy with or without neck dissection followed by 80 mCi RAI ablation postoperatively. Fifty (44.6%) patients had concomitant therapeutic CND at their initial surgery. On preoperative US, the mean size of largest diseased central node measured was 14.5 ± 10.39 mm (range 3–45 mm), and when categorized into groups I, II, and III, their mean sizes were 5.46 ± 2.50 mm, 12.09 ± 1.08 mm, and 28.54 ± 8.84 mm, respectively (p < 0.001). The majority (84.8%) of the largest diseased central nodes was located in the paratracheal/paraesophageal space. Following RCND, 18 (16.1%) had a positive microscopic margin on histology, and three (2.6%) patients had small amount of gross residual nodal disease (R2) in the central compartment. The rest had either a R0 or R1 resection. One of the three patients with a R2 resection received EBRT. One year following RCND, seven (6.3%) patients had structural incompleteness (five with locoregional and distant diseases, and two with distant disease only), and 78 (69.6%) patients had biochemical incompleteness. After a mean follow-up of 94.86 ± 33.28 months (range 50.18–216.30 months) after RCND, four of the seven patients died of metastatic PTC, while three were alive with stable structural disease. On analysis, of the 78 patients with biochemical incompleteness at one year, none of them died, but 14 (17.9%) patients developed new structural disease during follow-up requiring additional reoperations and/or RAI.

No significant differences in age, sex, time from initial operation to RCND, type of initial operation, and response to therapy one year after initial therapy were found between groups (see Table 1). However, group III had significantly more high-risk tumors at the time of initial operation than groups I and II combined (64.5% vs. 44.4%, respectively; p = 0.038). In terms of location of the largest diseased node, there was no significant difference between the groups (p > 0.05). Interestingly, although this did not reach statistical significance, group III had slightly less concomitant lateral nodal disease (p = 0.868) but more distant metastasis (p = 0.129) than the other two groups combined.

Table 1.

A Comparison of Baseline Patient and Tumor Characteristics Among Those with Largest Diseased Central Node <10 mm (Group I), 10–15 mm (Group II), and >15 mm in its Smallest Dimension (Group III) on Preoperative Ultrasonography

	Group I (n = 35)	Group II (n = 46)	Group III (n = 31)	p-Value ^a	p-Value ^b
Age at RCND (years)	54.52 ± 13.85	51.49 ± 16.67	49.23 ± 19.36	0.475	0.561
Sex (male/female)	5 (14.3)/30 (85.7)	14 (30.4)/32 (69.6)	9 (29.0)/22 (71.0)	0.115	0.542
Time from initial operation to first RCND (months)	38.26 ± 23.36	49.32 ± 32.31	30.63 ± 23.73	0.747	0.362
Type of initial operation				0.243	0.145
TT	18 (51.4)	20 (43.5)	17 (54.8)
TT + CND	11 (31.4)	20 (43.5)	0 (0.0)
TT + CND + SND	1 (2.9)	4 (8.7)	14 (45.2)
TT + SND	5 (14.3)	2 (4.3)	0 (0.0)
ATA risk stratification^c				0.295	0.038
Low	5 (14.3)	0 (0.0)	0 (0.0)
Intermediate	14 (40.0)	26 (56.5)	11 (35.5)
High	16 (45.7)	20 (43.5)	20 (64.5)
Response to therapy after initial therapy^d
Structural incompleteness	2 (5.7)	3 (6.5)	3 (9.7)	1.000	0.683
Biochemical response				0.765	0.593
sTg <2 ng/mL	6 (17.1)	9 (19.6)	7 (22.6)
sTg ≥2 ng/mL	27 (77.1)	34 (73.9)	21 (67.7)
Central nodal disease
Mean size (mm)	5.46 ± 2.50	12.09 ± 1.08	28.54 ± 8.84	<0.001	<0.001
Location				0.221	0.420
Pretracheal	6 (17.1)	5 (10.9)	6 (19.4)
Paratracheal/esophageal	29 (82.9)	41 (89.1)	25 (80.6)
Concomitant nodal disease in the lateral compartment	18 (51.4)	20 (43.5)	12 (38.7)	0.478	0.868
Concomitant distant metastasis	0 (0.0)	2 (4.3)	3 (9.7)	0.503	0.129

Statistically significant values are shown in bold.

Between groups I and II.

Between groups I and II combined and group III.

At initial operation based on the 2015 ATA risk stratification (3).

One year after initial operation and radioiodine ablation; those with structural incompleteness were excluded from evaluation of biochemical response.

RCND, reoperative central neck dissection; CND, central neck dissection; SND, lateral selective neck dissection; sTg, stimulated thyroglobulin; USG, ultrasonography; ATA, American Thyroid Association.

In terms of histologic and operative findings, there were no significant differences in mean number of CLNs retrieved (p > 0.05), number of positive CLNs (p > 0.05), CLN ratio (p > 0.05), extent (unilateral/bilateral; p > 0.05), or type of RCND (p > 0.05) between groups (see Table 2). However, group III had more frequent extranodal extension (35.5% vs. 16.0%, respectively; p = 0.055) than groups I and II combined. The chance of resection completeness was progressively lowered as nodal size increased (p < 0.05). Three patients had a R2 resection in group III, but none did in the other two groups. On microscopy, the mean size of the largest metastatic focus was highest in group III (p < 0.001), and there was a significant correlation between preoperative nodal size by US and size of metastatic focus (r = 0.961; p < 0.001). Most (72.3%) microscopic measurements corresponded to the preoperative US measurements within 3 mm in dimension. In terms of other operative findings, six patients in group III had RLN involvement, but none did in the other two groups (19.4% vs. 0.0%; p < 0.001). One of the six patients was suspected to have RLN involvement preoperatively on US imaging because extranodal spread was seen close to the tracheoesphageal groove. However, that patient did not have preoperative VCP on laryngoscopy. Intraoperatively, this patient was found to have nerve encasement by the nodal disease and required nerve transection. The other five patients were not suspected to have RLN involvement preoperatively. The proportion of inadvertent removal of parathyroid gland was similar between groups (p > 0.05).

Table 2.

A Comparison of Histologic and Operative Findings Among Those with Largest Diseased Central Bode <10 mm (Group I), 10–15 mm (Group II), and >15 mm in its Shortest Dimension (Group III) on Preoperative Ultrasonography

	Group I (n = 35)	Group II (n = 46)	Group III (n = 31)	p-Value ^a	p-Value ^b
Histological findings
Number of CLNs retrieved	8.23 ± 5.17	6.92 ± 4.72	9.69 ± 8.61	0.245	0.486
Number of positive CLNs	4.15 ± 3.60	3.25 ± 4.04	2.30 ± 2.59	0.178	0.303
CLN ratio (%)	50.96 ± 31.93	44.93 ± 37.22	35.21 ± 33.79	0.489	0.510
Extranodal extension	4 (11.4)	9 (19.6)	11 (35.5)	0.375	0.055
Largest metastatic focus (mm)	5.40 ± 2.37	12.61 ± 2.71	26.44 ± 9.84	<0.001	<0.001
Completeness of resection				0.034	<0.001
None (R0)	35 (100.0)	40 (87.0)	16 (51.6)
Microscopic positive (R1)	0 (0.0)	6 (13.0)	12 (38.7)
Gross disease left behind (R2)	0 (0.0)	0 (0.0)	3 (9.7)
Operative findings
RLN involvement	0 (0.0)	0 (0.0)	6 (19.4)	-	<0.001
Invasion/encasement	0 (0.0)	0 (0.0)	1 (3.2)	-	0.594
Adherent to RLN	0 (0.0)	0 (0.0)	5 (16.1)	-	0.001
Inadvertent PG removal	2 (5.7)	3 (6.5)	2 (6.5)	1.000	0.957
Extent of RCND				0.630	0.393
Unilateral	34 (97.1)	43 (93.5)	28 (90.3)
Bilateral	1 (2.9)	3 (6.5)	3 (9.7)
Type of RCND				0.270	0.583
RCND only	17 (48.6)	26 (56.5)	19 (61.3)
RCND + SND	18 (51.4)	20 (43.5)	12 (38.7)

Statistically significant values are shown in bold.

Between groups I and II.

Between groups I and II combined and group III.

RLN, recurrent laryngeal nerve; PG, parathyroid gland.

In terms of surgical morbidities (see Table 3), there were no significant differences in the rate of pre-existing VCP and hypocalcemia between groups (p > 0.05). There was one patient (in group III) who had a pre-existing permanent unilateral VCP that was contralateral to the side of the RCND. This was presumably related to the initial operation and not caused by central nodal disease. There were three patients in each group with pre-existing permanent hypocalcemia requiring calcium ± calcitriol. After excluding pre-existing morbidities, the overall rate of new-onset morbidities resulting from RCND was significantly higher in group III than it was in groups I and II combined (22.6% vs. 7.4%; p = 0.021). In group III, more than half (57.1%) of the total morbidities were temporary RLN injuries on the side of the RCND. Their time to recovery ranged from two to four months. One patient with permanent new-onset VCP was the patient with RLN encasement requiring transection. The incidence of new-onset VCP was significantly higher in group III than it was in groups I and II combined (16.1% vs. 2.5%, respectively; p = 0.017). However, the incidence of new-onset hypocalcemia and other complications were similar between groups. Two patients in group III suffered inadvertent esophageal injuries. One was recognized and repaired intraoperatively and recovered uneventfully, but the other was not recognized until two days postoperatively and required a second operation to repair the injury, but both had a R1 resection. New-onset VCPs and overall morbidity were not significantly different between groups I and II (p > 0.05).

Table 3.

A Comparison of Surgical Morbidity and Response to Therapy Following Reoperative Central Neck Dissection Among Those with the Largest Diseased Central Node <10 mm (Group I), 10–15 mm (Group II), and >15 mm in its Shortest Dimension (Group III) on Preoperative Ultrasonography

	Group I (n = 35)	Group II (n = 46)	Group III (n = 31)	p-Value ^a	p-Value ^b
Pre-existing VCP	0 (0.0)	0 (0.0)	1 (3.2)	—	0.268
New-onset VCP after RCND	1 (2.9)	1 (2.2)	5 (16.7)	1.000	0.017 ^c
Temporary (<6 months)	1 (2.9)	1 (2.2)	4 (13.3)	1.000	0.045 ^c
Permanent (≥6 months)	0 (0.0)	0 (0.0)	1 (3.3)	—	0.270^c
Pre-existing hypocalcemia	3 (8.6)	3 (6.5)	3 (9.7)	0.727	0.874
New-onset hypocalcemia after RCND	1 (3.1)	3 (7.0)	0 (0.0)	0.632	0.329^c
Temporary (<6 months)	0 (0.0)	0 (0.0)	0 (0.0)	—	—
Permanent (≥6 months)	1 (3.1)	3 (7.0)	0 (0.0)	0.632	0.329^c
Other complications after RCND	0 (0.0)	0 (0.0)	2 (6.5)	—	0.075
Overall RCND-related morbidity^**	2 (5.7)	4 (8.7)	7 (22.6)	0.694	0.021
Response to therapy one year after RCND
Structural incompleteness	1 (2.9)	2 (4.3)	4 (12.9)	0.889	0.191
Biochemical response
Mean sTg level (ng/mL)	11.15 ± 14.03	17.34 ± 15.11	25.21 ± 18.92	0.001 ^d	0.001 ^d
sTg <2 ng/mL	15 (42.9)	10 (21.7)	2 (6.5)	0.022 ^d	0.004 ^d
sTg ≥2 ng/mL	19 (54.3)	34 (73.9)	25 (80.6)

Statistically significant values are shown in bold.

Between groups I and II.

Between groups I and II combined and group III.

Excluding patients with either preexisting VCP or hypocalcemia.

Excluding patients with structural disease elsewhere

Sum of all new-onset VCP, hypocalcemia, and other complications.

VCP, vocal cord paresis; sTg, stimulated thyroglobulin.

In terms of response to therapy one year following RCND, although the incidence of structural incompleteness did not differ between groups (p > 0.05), biochemical incompleteness (sTg ≥2 ng/mL) became progressively more common as nodal size increased. The incidence rates of biochemical incompleteness for groups I, II, and III were 54.3%, 73.9%, and 80.6%, respectively. Mean sTg level was significantly higher in group III than it was in the other two groups combined (p = 0.001).

In multivariate analysis, after adjusting for American Thyroid Association risk stratification, the size of the largest diseased central node ≥15 mm by preoperative US (odds ratio [OR] = 7.256 [confidence interval (CI) 1.302–14.434], p = 0.001) was the only independent risk factor for biochemical incompleteness following RCND (see Table 4). When entered as a continuous variable, the size of central nodal disease (OR = 1.142 [CI 1.068–1.266], p = 0.012) was an independent risk factor for biochemical incompleteness following RCND.

Table 4.

A Multivariable Analysis for Achieving Biochemical Incomplete After Reoperative Cervical Neck Dissection ^a

Covariates	OR	CI	p-Value
ATA initial risk stratification
Low/intermediate	1
High	4.988	[0.912–27.294]	0.064
Size of the largest diseased central node on preoperative USG
<10 mm	1
10–15 mm	2.049	[0.657–6.391]	0.216
≥15 mm	7.256	[1.302–14.434]	0.024

Statistically significant values are shown in bold.

Size of persistent/recurrent central nodal disease remained significant (OR = 1.142 [CI 1.068–1.266], p = 0.012) when entered as a continuous variable into the model.

After excluding patients with structurally incomplete response following RCND.

OR, odds ratio; CI, confidence interval.

Discussion

To address the question of whether the cutoff of 8 mm by preoperative imaging is an appropriate nodal size threshold for recommending RCND over active surveillance in patients having persistent/recurrent central nodal disease (3,6), the present study examined the association between preoperative central nodal size by US and post-RCND morbidity and response to therapy at one year.

The data show that relative to those with smaller nodal disease (i.e., ≤15 mm; groups I and II combined), those with larger (>15 mm) diseased central nodes (i.e., group III) not only had a significantly higher incidence of RLN involvement (0.0% vs. 19.4%; p < 0.001), but also had a lesser chance of complete resection (i.e., R0 resection; 92.6% vs. 51.6%; p < 0.001). These findings are consistent with the view that bulkier central disease in the reoperative setting tends to be more locally invasive, and as a result, surgeons are less likely to be successful in complete resection.

In addition, the risks of new-onset VCP (16.7% vs. 2.5%, respectively; p = 0.017) and overall new-onset morbidity from RCND (22.6% vs. 7.4%; p = 0.021) were significantly higher in group III than they were in groups I and II combined. However, perhaps more relevant to the question of size threshold, the data show that all of these outcome parameters were not significantly different between those with central disease 10–15 mm (group II) and those with central disease <10 mm (group I). Therefore, in terms of the surgical morbidities following RCND, at least in the authors' experience, there were no differences for central nodal disease up to 15 mm in its smallest dimension. On the other hand, a size threshold of ≥15 mm more is associated with a more dramatic increase in morbidities, and is therefore perhaps a more clinically relevant cutoff than the more stringent threshold of 8 mm.

However, since four of the five new-onset VCPs in group III were only temporary in nature (i.e., full recovery within two to four months), one could argue that the actual impact of RCND on patients' quality of life in group III would have been limited, and their quality of life was significantly worse than either group I or group II. Nevertheless, it is worth pointing out that two patients in group III did have esophageal invasion, and both resulted in inadvertent esophageal perforation. Therefore, at times, an attempt at completely excising bulky nodal disease could result in a major surgical morbidity.

In terms of response to therapy one year after RCND, those having a larger (>15 mm) diseased central node (i.e., group III) had a significantly higher risk for biochemical incompleteness (i.e., sTg ≥2 ng/mL) compared with patients with a smaller (≤15 mm) diseased central node (80.6% vs. 67.9%; p = 0.004). Although the rate of biochemical incompleteness in group II was also significantly higher than that in group I (73.9% vs. 54.3%; p = 0.022), the mean sTg at one year in group III was highest among the three groups (p = 0.001). Furthermore, in the multivariate analysis, relative to a diseased central node <10 mm, only the size of a diseased node >15 mm was an independent factor for biochemical incompleteness (OR = 7.256 [CI 1.302–40.434], p = 0.024), while a nodal size of 10–15 mm was not (OR = 2.049 [CI 0.657–6.391], p = 0.216). This finding was consistent with a previous study, which showed larger central disease had less favorable outcomes in the reoperative setting (9). Therefore, similar to surgical morbidity, a size threshold of ≥15 mm o may represent a more significant cutoff value for response to therapy one year after RCND either <10 or 10–15 mm.

With regards to why patients with larger diseased central node were at a greater risk of biochemical incompleteness than those with smaller diseased central node, one reason is perhaps that the size of central nodal disease serves as a surrogate of the extent of subclinical local or distant disease (9), and therefore, despite successful local resection, some patients may still harbor residual disease that is only detectable biochemically. This assumption is certainly supported by the fact that at analysis, 14/78 (17.9%) patients with biochemical incompleteness following RCND developed new structural lesions, while none of the 27 patients with complete biochemical response following RCND developed new structural lesions. These findings are consistent with the experience of others that found postoperative sTg to be a useful marker for predicting persistent/recurrent disease following RCND (10 –12).

Despite the findings, several shortcomings must be acknowledged. First, although the results showed a significant association between the size of nodal disease and surgical morbidity and response to therapy after RCND, only a prospective randomized study could properly answer the question of whether a cutoff of 8 mm is appropriate. However, this would require a significant amount of time and effort involving multiple institutions. Also, although the results suggest that the cutoff of 15 mm is more clinically relevant than that of 8 mm, this may vary between institutions, depending on local surgical expertise and patient selection. Furthermore, although the findings imply that the current threshold of 8 mm could be increased to 15 mm without a significant increase in surgical morbidity, 8 mm remains a reasonable and safe cutoff. Furthermore, this was a single-institution retrospective analysis over a relatively long period and so selection biases might have had some influences on the results. Given that the decision for RCND was made by the patient and the multidisciplinary team, some patients with persistent/recurrent central disease might not have been operated and thereforenot analyzed in the study. Also, although the size of the largest diseased central node is a key parameter in deciding on RCND, other parameters such as patient comorbidities, interval growth rate, and US features (3,6,13) are also important, but they were not included in the present analysis. Lastly, for isolated, small-volume disease, in addition to surgery and active surveillance, other non-surgical therapeutic options such as ethanol injection, radiofrequency ablation, RAI therapy, and stereotactic EBRT are potential alternatives.

Conclusion

Patients with the largest diseased central node >15 mm were found to have a significantly higher risk of RLN involvement, incomplete surgical resection, new-onset VCP, overall surgical morbidity, and incomplete biochemical response in the reoperative setting than those with smaller (≤15 mm) central diseased nodes. Relative to nodal size <10 mm, a size >15 mm in the largest disease central node was an independent risk factor for incomplete biochemical response while size 10–15 mm was not. These findings imply that the current threshold of 8 mm might be too stringent and could be increased to 15 mm without a significant increase in surgical morbidity following RCND.

Footnotes

Author Disclosure Statement

The authors declare that they have no competing interests and nothing to disclose.

References

Hong Kong Cancer Registry. Cancer incidence and mortality in Hong Kong 1983–2013. Available at: www3.ha.org.hk/cancereg/e_stat.asp (accessed April 15, 2016 ).

Lang

, Wong

, Yu

, Lee

. 2016. Postoperative nomogram for predicting disease-specific death and recurrence in papillary thyroid carcinoma. Head Neck, 38:E1256–1263.

Haugen

, Alexander

, Bible

, Doherty

, Mandel

, Nikiforov

, Pacini

, Randolph

, Sawka

, Schlumberger

, Schuff

, Sherman

, Sosa

, Steward

, Tuttle

, Wartofsky

. 2016. 2015 American Thyroid Association management guidelines for adult patients with thyroid nodules and differentiated thyroid cancer: the American Thyroid Association Guidelines Task Force on Thyroid Nodules and Differentiated Thyroid Cancer. Thyroid, 26:1–133.

Perros

, Boelaert

, Colley

, Evans

, Gerrard Ba

, Gilbert

, Harrison

, Johnson

, Giles

, Moss

, Lewington

, Newbold

, Taylor

, Thakker

, Watkinson

, Williams

; British Thyroid Association. 2014. Guidelines for the management of thyroid cancer. Clin Endocrinol (Oxf), 81:1–122.

Clayman

, Steward

. 2016. Management of persistent or recurrent structural neck disease in differentiated thyroid carcinoma: point and counterpoint. JAMA Otolaryngol Head Neck Surg, 142:789–791.

Tufano

, Clayman

, Heller

, Inabnet

, Kebebew

, Shaha

, Steward

, Tuttle

; American Thyroid Association Surgical Affairs Committee Writing Task Force. 2015. Management of recurrent/persistent nodal disease in patients with differentiated thyroid cancer: a critical review of the risks and benefits of surgical intervention versus active surveillance. Thyroid, 25:15–27.

Lang

, Lee

, Ng

, Wong

, Wan

, Lo

. 2013. Evaluating the morbidity and efficacy of reoperative surgery in the central compartment for persistent/recurrent papillary thyroid carcinoma. World J Surg, 37:2853–2859.

Lang

, Wong

, Wan

. 2013. Postablation stimulated thyroglobulin level is an important predictor of biochemical complete remission after reoperative cervical neck dissection in persistent/recurrent papillary thyroid carcinoma. Ann Surg Oncol, 20:653–659.

Clayman

, Agarwal

, Edeiken

, Waguespack

, Roberts

, Sherman

. 2011. Long-term outcome of comprehensive central compartment dissection in patients with recurrent/persistent papillary thyroid carcinoma. Thyroid, 21:1309–1316.

10.

Al-Saif

, Farrar

, Bloomston

, Porter

, Ringel

, Kloos

. 2010. Long-term efficacy of lymph node reoperation for persistent papillary thyroid cancer. J Clin Endocrinol Metab, 95:2187–2194.

11.

Yim

, Kim

, Ryu

, Gong

, Hong

, Shong

. 2011. The outcomes of first reoperation for locoregionally recurrent/persistent papillary thyroid carcinoma in patients who initially underwent total thyroidectomy and remnant ablation. J Clin Endocrinol Metab, 96:2049–2056.

12.

Lee

, Roh

, Gong

, Cho

, Choi

, Nam

, Kim

. 2015. Risk factors for re-recurrence after first reoperative surgery for locoregional recurrent/persistent papillary thyroid carcinoma. World J Surg, 39:1943–1950.

13.

Lamartina

, Grani

, Biffoni

, Giacomelli

, Costante

, Lupo

, Maranghi

, Plasmati

, Sponziello

, Trulli

, Verrienti

, Filetti

, Durante

. 2016. Risk stratification of neck lesions detected sonographically during the follow-up of differentiated thyroid cancer. J Clin Endocrinol Metab, 101:3036–3044.