Pattern of Spread to the Lateral Neck in Metastatic Well-Differentiated Thyroid Cancer: A Systematic Review and Meta-Analysis

Abstract

Background:

There remains controversy surrounding the extent of lateral neck dissection required in patients with papillary thyroid cancer (PTC) and suspicious or confirmed metastatic lateral neck lymphadenopathy. The evidence for this clinical dilemma has never been reviewed systematically nor has there been an attempt to meta-analyze the data by lymph node levels to better characterize the pattern of spread.

Methods:

This meta-analysis used MEDLINE and EMBASE including all cohort studies reporting the pattern of lateral neck disease in patients who underwent a neck dissection for clinically, radiographically, or cytologically suspicious or confirmed metastatic lymphadenopathy for PTC. Our main outcome was the number of patients with positive involvement at a given level as a percentage of the cohort of patients with positive lateral neck disease, each level being measured separately.

Results:

Eighteen studies with a total of 1145 patients and 1298 neck dissections were included in our meta-analysis. Levels IIa and IIb had disease in 53.1% [95% confidence interval (CI) 46.6–59.5%] and 15.5% [CI 8.2–27.2%], respectively. Studies that did not distinguish between level IIa and IIb or in which both were collapsed into one category showed a total level II involvement of 53.4% [CI 49.7–57.1%]. Level III and level IV were involved in 70.5% [CI 67.0–73.9%] and 66.3% [CI 61.4–70.9%] of specimens. Studies that did not distinguish between level Va and Vb or in which both were collapsed into one category showed a total level V involvement of 25.3% [CI 20.0–31.5%]. Levels Va and Vb had positivity in 7.9% [CI 2.8–20.0%] and 21.5% [CI 7.7–47.6%], respectively, but had only three studies that could be meta-analyzed.

Conclusions:

This systematic review of the literature and meta-analysis of the pattern of spread indicates significant rates of lymph node metastasis to all lateral neck levels in patients with PTC with regional involvement. This evidence leads us to recommend a comprehensive selective neck dissection of levels IIa, IIb, III, IV, and Vb in patients with lateral neck disease from PTC. The evidence for level Va is lacking, as most studies did not distinguish between levels Va and Vb, and the border between the two levels was inconsistent. Future studies will need to address these sublevels separately.

Introduction

D espite a growing body of evidence on the pattern of spread of papillary thyroid cancer (PTC) to regional lymph nodes (1), there remains controversy surrounding the extent of neck dissection in patients with clinically, radiographically, or cytologically suspicious or confirmed metastasis to lateral neck lymph nodes. Due to the increasing incidence of thyroid cancer over the past four decades, there may also be a rise in the number of patients presenting with lateral neck disease and requiring neck dissection (2,3). Results from large population-based administrative and cancer databases, as well as data from large-volume institutions, have demonstrated that lymph node metastasis in PTC is a significant predictor of disease-free survival, disease-specific survival, or both (4 –9). The American Thyroid Association (ATA) in their 2009 thyroid nodule management guidelines advocated that a “therapeutic lateral neck compartmental lymph node dissection should be performed for patients with biopsy-proven metastatic lateral cervical lymphadenopathy” (10). The Triological Society followed in 2010, advocating a comprehensive selective neck dissection (11). Due to the variability in extent and definition of neck dissections, the ATA has recently published a consensus review on the anatomy, terminology, and rationale for lateral neck dissection in this population (12). The anatomic neck regions described include the upper jugular (II), the mid jugular (III), the lower jugular (IV), and the posterior triangle (V). Level II can be divided into level IIa nodes, which surround the internal jugular vein, and level IIb nodes, which lie posterior to the spinal accessory. Level V can also be divided into level Va nodes, which lie between the skull base and the level of the lower margin of the cricoid cartilage arch behind the posterior edge of the sternocleidomastoid muscle, and level Vb nodes, which lie between the level of the lower margin of the cricoid cartilage and the level of the clavicle (12). The ATA's review recommended a comprehensive selective neck dissection of levels IIa, III, IV, and Vb (12). The recommendations were based on relatively small sample and retrospective case series from single institutions. The evidence for this clinical dilemma on the extent of lateral neck dissection required in PTC with any lateral neck metastasis has never been systematically reviewed nor has there been an attempt to meta-analyze data by lymph node level to better characterize the pattern of spread.

This systematic review and meta-analysis focused on a single outcome: pattern of disease spread in patients with metastatic PTC. The outcome was measured by analysis of the final pathology specimen in patients who underwent a selective lateral neck dissection for suspected or confirmed metastatic PTC. We hypothesize that a meta-analysis in this population will reveal a high incidence of metastasis to all lateral neck levels with the greatest incidence of spread to levels III (middle jugular) and IV (lower jugular), followed by II (upper jugular) and V (posterior triangle), in that order.

Methods

Our review protocol has not previously been published or registered. We were specifically focused on the population of patients with PTC who underwent lateral neck dissection for metastatic lymph node spread.

Search strategy

A literature search was performed using the MEDLINE and EMBASE databases. The search was restricted to human subjects and adults (age≥18 years). The following search terms were used: (neck dissection) and (lateral or cervical lymphadenopathy) and (PTC or thyroid neoplasms or papillary carcinoma) not (medullary or follicular carcinoma). Key terms were explored and exploded in each database separately, and non-MeSH words were included when they were not included in the databases. The date of the last search was November 4, 2011.

Before reviewing any article, a number of techniques were used to ensure that all the relevant references were included in our search. Citations were cross-checked (snowballing) from key publications; citations from existing reviews were assessed (11,12); the related articles to key publications in PubMed were reviewed; forward citations were used; and two experts in the field (J.F. and I.W.) were surveyed to make sure that we were not missing any key references. This systematic review was planned, conducted, and reported in adherence with all of the standards of quality for reporting meta-analyses (MOOSE) (13).

Review process

The search results were reviewed in three stages. All three stages included the same two independent reviewers. The first two stages included a title and abstract review. A kappa statistic for inter-rater reliability was calculated for each stage. The full articles were then acquired for abstracts that had either been chosen by one or both reviewers. These articles were then assessed for study inclusion/exclusion criteria. Another kappa statistic was calculated for this step. The last phase involved two abstractors for each article. Any discrepancies were resolved by consensus. Reasons for exclusions were recorded at every step.

Study inclusion/exclusion criteria

Studies considered for inclusion treated adults with PTC or well-differentiated thyroid cancer (including follicular thyroid cancer) with lateral neck disease treated by a comprehensive neck dissection (levels II–IV or II–V). The study must have described disease spread by level and specifically indicate how many neck dissections from their cohort had at least one positive level on final pathology. For a study to be included, they had to describe their indications for surgery. The search included prospective and retrospective cohort studies.

Review articles, commentaries, and technique articles were excluded, although these publications were used only in the snowballing process. Case reports with 10 patients or less were excluded. Articles that focused on pediatric patients, prophylactic neck dissection, central neck dissections, medullary or anaplastic carcinoma, and studies that did not have adequate data for use in our meta-analysis were also excluded.

Data extraction

Each study was abstracted by the same two independent reviewers who reached a consensus about any discrepancies in the collection process. Authors were contacted when there were any questions about their methodology or results.

Summary measures

The main outcome measure was the number of neck dissections with positive neck node involvement at a given level as a percentage of the cohort of neck dissections with positive lateral neck disease, each level being measured separately. Since studies use a variety of inclusion criteria (clinical evidence of disease, radiographic evidence of disease, and/or fine-needle aspiration-proven neck disease), the denominator used for all studies was the number of neck dissections with pathology-proven metastasis to at least one level. This provided a consistent denominator for studies with slightly differing inclusion criteria and standards of care. By utilizing the number of neck dissections assessed and the event rates at each level, confidence intervals were calculated using the Comprehensive Meta-Analysis program version 2 (Biostat, Inc., Englewood, NJ) utilizing binomial distributions. Findings are presented as a percent, from the cohort of neck dissections, having involved disease at a particular level with its appropriate confidence intervals.

Bias across studies and heterogeneity

Risk of bias across studies may be present, particularly with regard to publication bias. As the topic involves surgical procedures and outcomes, it is very likely that smaller studies or those with unfavorable outcomes may not be published in the literature. A funnel plot was created to assess publication bias, and the Duval and Tweedie's trim-and-fill method was used for the analysis of each level to assess where the missing studies would be plotted.

For many of the reasons mentioned above, the a priori hypothesis was that a large degree of heterogeneity existed among studies. Important areas of variation include % female patients, prospective or retrospective, continent of primary data (North America yes/no), PTC or well-differentiated thyroid cancer, standard of care for neck dissection (II–IV or II–V), year of initiation of study, single- or multisurgeon, and single- or multicenter. The prespecified statistical plan to account for effects of heterogeneity included subgroup and metaregression analyses.

Assessment of study quality

A validated tool to assess bias in observational studies without a control or comparison group does not yet exist and is quite population specific. For this reason and to assess key differences between studies, data were collected on a number of additional study features, which may be considered as surrogates of study quality. Data on study design (retrospective vs. prospective), mean number of lymph nodes per neck, mean number of lymph nodes per level, pathology (PTC vs. well-differentiated thyroid cancer), whether the neck dissection specimen was marked or not, multicenter or single center, multisurgeon or single surgeon, and the institution's standard of care in treating patients with PTC with suspected lateral neck disease were collected. Data on the percentage of female patients in the population and the year of initiation of the study were also collected.

Statistical analyses

Eight overall statistical analyses were performed—one for each neck level (IIa, IIb, II, III, IV, V, Va, and Vb). Due to some anticipated heterogeneity, a more conservative statistical approach applying the random-effects model was chosen for all analyses. This method takes into account within-study variation, between-study variation, and the heterogeneity inherent in comparing results from different centers and surgeons. For each forest plot, the random-effects model summary-effect measure, 95% confidence interval (CI), and the I² statistic were included. To assess publication bias, a standard funnel plot was used as well as random-effects Duval and Tweedie's trim-and-fill.

Results

Study selection

Search results were reviewed in three stages. The first stage included a title review, which resulted in a reduction from 569 to 217 articles. This yielded an inter-rater reliability kappa statistic of 0.603±0.04. The second stage included an abstract review, which resulted in a reduction from 217 to 94 articles. This yielded an inter-rater reliability kappa statistic of 0.636±0.08. The final stage included an assessment of the full article for preselected inclusion/exclusion criteria that yielded 18 full-text articles for which we could extract our metameters (14 –31). Any disagreements at this stage were resolved by consensus. Figure 1 displays reasons for exclusion and number excluded at each stage.

FIG. 1.

Selection of studies for inclusion in the meta-analysis.

Study characteristics

The characteristics of the studies included in the meta-analysis are summarized in Table 1. The 18 studies were published between 2002 and 2012. Six of these studies were from North America, three from Europe, two from Israel, and seven from Asia. Each study had between 22 and 185 patients and 22 and 248 comprehensive neck dissections (II–IV or II–V) involving at least one level with positive nodes. This provided a total of 1145 patients and 1298 neck dissections included in this meta-analysis. The chosen metameter was extractable for levels IIa, IIb, II, III, IV, and V from 9, 9, 15, 18, 18, and 14 studies, respectively.

Table 1.

Included Studies and Evidence Table

Reference	Country	Study years	Total patients	Total positive NDs ^a	Level II	Level IIa	Level IIb	Level III	Level IV	Level V	Level Va	Level Vb	Standard of practice
Ahn et al., 2008 (14)	Italy	2005	37	28	65%	58%	22%	80%	76%	15%	—	—	I–VI
Farrag et al., 2009 (16)	United States	2002–2007	53	60	65%	56%	9%	66%	50%	40%	0%	40%	II–V
Khafif et al., 2008 (17)	Israel	NA	37	37	46%	—	—	68%	70%	16%	—	—	II–IV
King et al., 2011 (18)	United States	2000–2007	32	39	—	49%	62%	72%	67%	—	8%	31%	II–V
Koo et al., 2009 (19)	Korea	2005–2008	76	76	53%	51%	12%	72%	68%	16%	—	—	II–V
Kupferman et al., 2004 (20)^b	United States	1997–2002	39	39	52%	—	—	57%	41%	21%	—	—	II–V
Kupferman et al., 2008 (21)^b	United States	1993–2004	70	75	57%	—	—	62%	62%	53%	—	—	NA
Lee et al., 2007 (15)	Korea	2001–2006	46	55	60%	—	—	82%	75%	20%	—	—	IIa, IIb, III–V
Lee et al., 2008 (23)	Korea	2005–2007	167	191	56%	56%	7%	81%	75%	17%	—	—	II–V
Lee et al., 2010 (22)	Japan	1998–2007	70	80	63%	—	—	74%	82%	—	—	—	II–V
Lim et al., 2010 (24)	Korea	2005–2008	70	71	49%	—	—	74%	69%	16%	—	—	II–V
Merdad et al., 2012 (25)	Canada	2004–2011	185	248	49%	—	—	67%	61%	29%	—	—	IIa, IIb, III–Vb
Pingpank et al., 2002 (26)	United States	1991–2001	44	51	—	43%	21%	76%	59%	28%	—	—	II–V
Roh et al., 2008 (27)	Korea	2003–2006	52	57	—	—	—	72%	76%	—	13%	4%	Ib–Vb
Sivanandan et al., 2001 (28)	Singapore	1988–1999	70	80	49%	72%	17%	65%	56%	29%	—	—	I–V
Vayisoglu et al., 2010 (29)	Turkey	2006–2009	22	22	27%	—	—	55%	82%	18%	—	—	II–V
Yanir and Doweck, 2008 (30)	Israel	2001–2006	27	28	54%	27%	5%	68%	57%	20%	—	—	II–IV
Yuce et al., 2010 (31)	Turkey	1999–2008	48	61	46%	50%	7%	69%	66%	34%	—	—	II–VI

Total number of neck dissections (NDs) with positive disease.

Same first author, but different patient populations described.

NA, not available.

Pattern of spread

The random-effects model meta-analysis summary for each of the neck levels is represented in Tables 2 –7. Level IIa was involved with metastasis in 53.1% [CI 46.6–59.5%] and level IIb in 15.5% [CI 8.2–27.2%]. Studies that did not distinguish between level IIa and IIb or in which both were collapsed into one category showed a total level II involvement of 53.4% [CI 49.7–57.1%]. Level III and level IV were involved in 70.5% [CI 67.0–73.9%] and 66.3% [CI 61.4–70.9%] of cases. Studies that did not distinguish between level Va and Vb or in which both were collapsed into one category showed a total level V involvement of 25.3% [CI 20.0–31.5%]. Levels Va and Vb had three studies each (not shown) and had positivity in 7.9% [CI 2.8–20.0%] and 21.5% [CI 7.7–47.6%], respectively.

Table 2.

Percent of Necks with Involvement at Level IIa

	Statistics for each study
Reference	Event rate	Lower limit	Upper limit	Event rate and CI
Farrag et al., 2009 (16)	0.560	0.433	0.679
King et al., 2011 (18)	0.490	0.339	0.643
Koo et al., 2009 (19)	0.510	0.399	0.620
Lee et al., 2007 (15)	0.580	0.447	0.702
Lee et al., 2008 (23)	0.560	0.489	0.629
Pingpank et al., 2002 (26)	0.430	0.302	0.568
Roh et al., 2008 (27)	0.720	0.591	0.821
Vayisoglu et al., 2010 (29)	0.270	0.126	0.487
Yanir and Doweck, 2008 (30)	0.500	0.323	0.677
	0.531	0.466	0.595

I²=52.65

CI, 95% confidence interval.

Table 3.

Percent of Necks with Involvement at Level IIb

	Statistics for each study
Reference	Event rate	Lower limit	Upper limit	Event rate and CI
Farrag et al., 2009 (16)	0.090	0.039	0.193
King et al., 2011 (18)	0.620	0.461	0.757
Koo et al., 2009 (19)	0.120	0.064	0.214
Lee et al., 2007 (15)	0.220	0.130	0.348
Lee et al., 2008 (23)	0.070	0.041	0.116
Pingpank et al., 2002 (26)	0.210	0.119	0.343
Roh et al., 2008 (27)	0.170	0.093	0.290
Vayisoglu et al., 2010 (29)	0.050	0.008	0.264
Yanir and Doweck, 2008 (30)	0.070	0.017	0.243
	0.155	0.082	0.272

I²=86.93

Table 4.

Percent of Necks with Involvement at Level II (Not Otherwise Specified)

	Statistics for each study
Reference	Event rate	Lower limit	Upper limit	Event rate and CI
Ahn et al., 2008 (14)	0.650	0.461	0.801
Farrag et al., 2009 (16)	0.650	0.522	0.759
Khafif et al., 2008 (17)	0.460	0.309	0.619
Koo et al., 2009 (19)	0.530	0.418	0.639
Kupferman et al., 2004 (20)	0.520	0.366	0.670
Kupferman et al., 2008 (21)	0.570	0.456	0.677
Lee et al., 2007 (15)	0.600	0.467	0.720
Lee et al., 2008 (23)	0.560	0.489	0.629
Lee et al., 2010 (22)	0.630	0.520	0.728
Lim et al., 2010 (24)	0.490	0.376	0.605
Merdad et al., 2012 (25)	0.490	0.428	0.552
Sivanandan et al., 2001 (28)	0.490	0.383	0.598
Vayisoglu et al., 2010 (29)	0.270	0.126	0.487
Yanir and Doweck, 2008 (30)	0.540	0.358	0.712
Yuce et al., 2010 (31)	0.460	0.340	0.585
	0.534	0.497	0.571

I²=30.90

Table 5.

Percent of Necks with Involvement at Level III

	Statistics for each study
Reference	Event rate	Lower limit	Upper limit	Event rate and CI
Ahn et al., 2008 (14)	0.800	0.613	0.910
Farrag et al., 2009 (16)	0.660	0.532	0.768
Khafif et al., 2008 (17)	0.680	0.516	0.809
King et al., 2011 (18)	0.720	0.561	0.838
Koo et al., 2009 (19)	0.720	0.609	0.809
Kupferman et al., 2004 (20)	0.570	0.413	0.714
Kupferman et al., 2008 (21)	0.620	0.506	0.722
Lee et al., 2007 (15)	0.820	0.696	0.901
Lee et al., 2008 (23)	0.810	0.748	0.860
Lee et al., 2010 (22)	0.740	0.633	0.824
Lim et al., 2010 (24)	0.740	0.626	0.829
Merdad et al., 2012 (25)	0.670	0.609	0.726
Pingpank et al., 2002 (26)	0.760	0.625	0.858
Roh et al., 2008 (27)	0.720	0.591	0.821
Sivanandan et al., 2001 (28)	0.650	0.540	0.746
Vayisoglu et al., 2010 (29)	0.550	0.345	0.739
Yanir and Doweck, 2008 (30)	0.680	0.490	0.825
Yuce et al., 2010 (31)	0.690	0.564	0.793
	0.705	0.670	0.739

I²=39.34

Table 6.

Percent of Necks with Involvement at Level IV

	Statistics for each study
Reference	Event rate	Lower limit	Upper limit	Event rate and CI
Ahn et al., 2008 (14)	0.760	0.571	0.883
Farrag et al., 2009 (16)	0.500	0.376	0.624
Khafif et al., 2008 (17)	0.700	0.536	0.825
King et al., 2011 (18)	0.670	0.510	0.798
Koo et al., 2009 (19)	0.680	0.568	0.775
Kupferman et al., 2004 (20)	0.410	0.269	0.568
Kupferman et al., 2008 (21)	0.620	0.506	0.722
Lee et al., 2007 (15)	0.750	0.620	0.847
Lee et al., 2008 (23)	0.750	0.684	0.806
Lee et al., 2010 (22)	0.820	0.720	0.890
Lim et al., 2010 (24)	0.690	0.574	0.786
Merdad et al., 2012 (25)	0.610	0.548	0.669
Pingpank et al., 2002 (26)	0.590	0.452	0.715
Roh et al., 2008 (27)	0.760	0.633	0.853
Sivanandan et al., 2001 (28)	0.560	0.450	0.664
Vayisoglu et al., 2010 (29)	0.820	0.606	0.931
Yanir and Doweck, 2008 (30)	0.570	0.385	0.737
Yuce et al., 2010 (31)	0.660	0.533	0.767
	0.663	0.614	0.709

I²=65.65

Table 7.

Percent of Necks with Involvement at Level V (Not Otherwise Specified)

	Statistics for each study
Reference	Event rate	Lower limit	Upper limit	Event rate and CI
Ahn et al., 2008 (14)	0.150	0.059	0.332
Farrag et al., 2009 (16)	0.400	0.285	0.528
Koo et al., 2009 (19)	0.160	0.094	0.260
Kupferman et al., 2004 (20)	0.210	0.110	0.365
Kupferman et al., 2008 (21)	0.530	0.417	0.640
Lee et al., 2007 (15)	0.200	0.114	0.326
Lee et al., 2008 (23)	0.170	0.123	0.230
Lim et al., 2010 (24)	0.160	0.092	0.264
Merdad et al., 2012 (25)	0.290	0.237	0.350
Pingpank et al., 2002 (26)	0.280	0.174	0.417
Sivanandan et al., 2001 (28)	0.290	0.201	0.398
Vayisoglu et al., 2010 (29)	0.180	0.069	0.394
Yanir and Doweck, 2008 (30)	0.200	0.090	0.387
Yuce et al., 2010 (31)	0.340	0.233	0.467
	0.253	0.200	0.315

I²=75.97

Study heterogeneity

The studies ranged in heterogeneity with I² statistics of 53, 87, 31, 39, 66, and 76 for levels IIa, IIb, II, III, IV, and V, respectively. This ranges from moderate to high range of heterogeneity between studies.

Assessment of study quality and bias

Important study characteristics that are surrogates for study quality are summarized in Table 8. None of the studies report on exact technique of neck dissection nor did any of the studies report on the number of pathologists that reviewed their specimens.

Table 8.

Study Characteristics and Assessment of Bias

Reference	Study design (retro vs. prospective)	Year of initiation	Female (%)	Mean LNs/neck	Mean LNs/level	Cell pathology included	Method of marking level specified (Y/N)	Multicenter vs. single center	Multi vs. single surgeon
Ahn et al., 2008 (14)	R	2005	81%	NA	NA	WDTC	Y (labeled individually)	NA	NA
Farrag et al., 2009 (16)	R	2002	59%	NA	NA	PTC	Y (in situ)	Single	Single
Khafif et al., 2008 (17)	P	NA	62%	NA	9.9	PTC	NA	Single	NA
King et al., 2011 (18)	R	2000	59%	43	NA	PTC	Y (cut levels and sent separately)	Single	Single
Koo et al., 2009 (19)	R	2005	83%	31	4.7	PTC	N (IIb sent alone)	Multi	Multi
Kupferman et al., 2004 (20)^a	R	1997	80%	44	NA	PTC	N (oriented by senior surgeon)	Single	Single
Kupferman et al., 2008 (21)^a	R	1993	54%	NA	NA	WDTC	NA	Single	NA
Lee et al., 2007 (15)	R	2001	83%	NA	NA	PTC	NA	Single	NA
Lee et al., 2008 (23)	P	2005	82%	32	8.5	PTC	Y (rubber band)	Single	Single
Lee et al., 2010 (22)	R	1998	66%	NA	NA	PTC	NA	Single	Single
Lim et al., 2010 (24)	R	2005	81%	NA	NA	PTC	NA	Single	NA
Merdad et al., 2012 (25)	R	2004	50%	66	9.8	PTC	Y (surgical board)	Single	Single
Pingpank et al., 2002 (26)	R	1991	52%	NA	NA	WDTC	NA	Single	Multi
Roh et al., 2008 (27)	P	2003	77%	43	NA	PTC	NA	NA	Single
Sivanandan et al., 2001 (28)	R	1988	NA	NA	NA	NA	Y (sutures)	Single	NA
Vayisoglu et al., 2010 (29)	P	2006	61%	27	2.9	PTC	NA (IIb sent separately)	NA	NA
Yanir and Doweck, 2008 (30)	R	2001	56%	NA	NA	WDTC	Y (divided intra op)	Single	Single
Yuce et al., 2010 (31)	R	1999	52%	42	7.0	PTC	Y (ribbon)	Single	Multi

LNs, lymph nodes; R, retrospective; P, prospective; PTC, papillary thyroid cancer; WDTC, well-differentiated thyroid cancer; op, operatively.

Subgroup analyses and metaregression

The heterogeneity could not be explained by subgroup analyses of the prospective or retrospective nature of the study, continent of primary data (North America yes/no), whether the study focused on PTC or well-differentiated thyroid cancer, whether it was a single- or multisurgeon study, and whether it was a single- or multicenter study. A metaregression of percent female patients and of the year of initiation of the study also did not reveal any significant variation.

Discussion

This study is the first systematic review and meta-analysis of the literature on patients undergoing comprehensive lateral neck dissection for metastatic PTC. We demonstrated high rates of metastasis to all lateral neck levels. Included in our analysis were 18 unique studies from 2002 to 2012 with 1145 patients who underwent 1298 neck dissections. Level IIa, IIb, II, III, IV, and V had metastatic involvement in 53%, 16%, 53%, 71%, 66%, and 25% of neck dissections, respectively. Level Va and Vb only had three studies in each meta-analysis and demonstrated 8% and 22% positivity, respectively. Assessment of bias, subgroup analyses, and metaregression could not explain the moderate-to-high levels of heterogeneity between studies detected in our analysis.

Our findings, particularly at levels II and V, demonstrate the need to perform a comprehensive neck dissection in PTC patients with metastasis to the lateral neck. These systematically reviewed findings support the guidelines put out by the ATA in 2009 and the Triological Society in 2010, which were largely based on expert opinion and small retrospective cohort studies (10,11).

The true impact of local lymph node metastasis on survival in PTC remains controversial. Nonetheless, there are ample data in the literature associating lymph node metastasis with decreased disease-specific survival and increased risk of local recurrence. Outcome data from large institutional cohorts from the United States, Canada, and Germany have shown a significant and independent negative impact on survival with lymph node metastasis (32 –34). Analysis of large cancer registry databases from the United States and Sweden has similarly shown poorer outcomes and increased mortality rate in patients with locoregional lymph node metastases (8,35). Additionally, the presence of local lymph node metastasis has been associated with an increased risk of future local recurrence (4). As a result, lateral neck metastasis is currently considered a significant prognosticator according to the latest thyroid cancer TNM classification (36). Lymph node metastasis is not, however, a predictor in several institutional risk stratification schemes, including AMES, AGES, and MACIS (37 –39).

Considering the available evidence, lateral neck dissection for excision of metastatic lymph nodes may improve survival and decrease regional recurrence. However, the potential survival benefit must be weighed against the morbidity associated with neck dissection surgery. The most significant and common of these complications is shoulder dysfunction resulting from manipulation and injury of the spinal accessory nerve with surgery to levels II and V (9). Dissection of the posterior triangle (level V) is associated with higher rates of shoulder dysfunction (40); nonetheless, transection of the nerve remains a rare complication in experienced hands (41). In addition, there are no differences in health-related quality of life secondary to shoulder dysfunction between modified radical neck dissection and selective neck dissection at 1 year postoperatively (42). Synthesis of this literature suggests that the increased risk of recurrence and poorer survival outweigh the risk of complications from neck dissection.

The limitations of the study need to be considered. Whenever performing a meta-analysis to estimate an outcome or effect from a group of similar studies, the measure of heterogeneity informs whether the combined estimate is meaningful. Our a priori hypothesis of significant heterogeneity in a meta-analysis of observational studies proved to be true based on our moderate-to-high I². It has been suggested that quantification of heterogeneity is only one aspect of variability assessment across studies (43). The studies included in our meta-analysis have some diversity in their clinical and methodological aspects (Tables 1 and 8), but subgroup analyses and meta-regression on these factors could not explain or decrease the observed heterogeneity estimates. In our opinion, the clinical implication of the observed degree of inconsistency across studies is not significant. The results of our quantitative analyses were consistent with the qualitative trends across the individual studies. This is also evidenced by the fairly narrow confidence intervals around our estimates.

In this meta-analysis, the distribution of lateral neck disease involvement demonstrates the need for comprehensive neck dissection in patients presenting with confirmed PTC metastasis to any lateral neck level. Our findings corroborate the guidelines put out by the ATA in 2009 and the Triological Society in 2010 (10,11).

Footnotes

Disclosure Statement

The authors declare that no competing financial interests exist.

References

Noguchi

, Noguchi

, Murakami

. 1970. Papillary carcinoma of the thyroid. I. Developing pattern of metastasis. Cancer, 26:1053–1060.

Davies

, Welch

. 2006. Increasing incidence of thyroid cancer in the United States, 1973–2002. JAMA, 295:2164–2167.

Davies

, Welch

. 2010. Thyroid cancer survival in the United States: observational data from 1973 to 2005. Arch Otolaryngol Head Neck Surg, 136:440–444.

Beasley

, Lee

, Eski

, Walfish

, Witterick

, Freeman

. 2002. Impact of nodal metastases on prognosis in patients with well-differentiated thyroid cancer. Arch Otolaryngol Head Neck Surg, 128:825–828.

Noguchi

, Murakami

, Yamashita

, Toda

, Kawamoto

. 1998. Papillary thyroid carcinoma: modified radical neck dissection improves prognosis. Arch Surg, 133:276–280.

Shah

, Hall

, Eski

, Witterick

, Walfish

, Freeman

. 2003. Clinical course of thyroid carcinoma after neck dissection. Laryngoscope, 113:2102–2107.

Tubiana

, Schlumberger

, Rougier

, Laplanche

, Benhamou

, Gardet

, Caillou

, Travagli

, Parmentier

. 1985. Long-term results and prognostic factors in patients with differentiated thyroid carcinoma. Cancer, 55:794–804.

Lundgren

, Hall

, Dickman

, Zedenius

. 2006. Clinically significant prognostic factors for differentiated thyroid carcinoma: a population-based, nested case-control study. Cancer, 106:524–531.

Podnos

, Smith

, Wagman

, Ellenhorn

. 2005. The implication of lymph node metastasis on survival in patients with well-differentiated thyroid cancer. Am Surg, 71:731–734.

10.

American Thyroid Association (ATA) Guidelines Taskforce on Thyroid Nodules and Differentiated Thyroid Cancer. Cooper

, Doherty

, Haugen

, Kloos

, Lee

, Mandel

, Mazzaferri

, McIver

, Pacini

, Schlumberger

, Sherman

, Steward

, Tuttle

. 2009. Revised American Thyroid Association management guidelines for patients with thyroid nodules and differentiated thyroid cancer. Thyroid, 19:1167–1214.

11.

Hasney

, Amedee

. 2010. What is the appropriate extent of lateral neck dissection in the treatment of metastatic well-differentiated thyroid carcinoma? Laryngoscope, 120:1716–1717.

12.

Stack

Jr , Ferris

, Goldenberg

, Haymart

, Shaha

, Sheth

, Sosa

, Tufano

. American Thyroid Association Surgical Affairs Committee 2012 American Thyroid Association consensus review and statement regarding the anatomy, terminology, and rationale for lateral neck dissection in differentiated thyroid cancer. Thyroid, 22:501–508.

13.

Stroup

, Berlin

, Morton

, Olkin

, Williamson

, Rennie

, Moher

, Becker

, Sipe

, Thacker

. 2000. Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA, 283:2008–2012.

14.

Ahn

, Lee

, Yi

, Shong

, Hong

, Lee

, Choi

, Kim

. 2008. Diagnostic accuracy of CT and ultrasonography for evaluating metastatic cervical lymph nodes in patients with thyroid cancer. World J Surg, 32:1552–1558.

15.

Lee

, Wang

, Lee

, Son

, Kim

. 2007. Level IIb lymph node metastasis in neck dissection for papillary thyroid carcinoma. Arch Otolaryngol Head Neck Surg, 133:1028–1030.

16.

Farrag

, Lin

, Brownlee

, Kim

, Sheth

, Tufano

. 2009. Is routine dissection of level II-B and V-A necessary in patients with papillary thyroid cancer undergoing lateral neck dissection for FNA-confirmed metastases in other levels. World J Surg, 33:1680–1683.

17.

Khafif

, Ben-Yosef

, Abergel

, Kesler

, Landsberg

, Fliss

. 2008. Elective paratracheal neck dissection for lateral metastases from papillary carcinoma of the thyroid: is it indicated? Head Neck, 30:306–310.

18.

King

, Corbitt

, Miller

. 2011. Management of lateral cervical metastases in papillary thyroid cancer: patterns of lymph node distribution. Ear Nose Throat J, 90:386–389.

19.

Koo

, Yoon

, Kim

, Choi

, Lim

. 2009. Predictive factors of level IIb lymph node metastasis in patients with papillary thyroid carcinoma. Ann Surg Oncol, 16:1344–1347.

20.

Kupferman

, Patterson

, Mandel

, LiVolsi

, Weber

. 2004. Patterns of lateral neck metastasis in papillary thyroid carcinoma. Arch Otolaryngol Head Neck Surg, 130:857–860.

21.

Kupferman

, Weinstock

, Santillan

, Mishra

, Roberts

, Clayman

, Weber

. 2008. Predictors of level V metastasis in well-differentiated thyroid cancer. Head Neck, 30:1469–1474.

22.

Lee

, Kawata

, Nishikawa

, Yoshimura

, Takenaka

. 2010. Diagnostic criteria of ultrasonographic examination for lateral node metastasis of papillary thyroid carcinoma. Acta Otolaryngol, 130:161–166.

23.

Lee

, Sung

, Nam

, Chung

, Soh

, Park

. 2008. Is level IIb lymph node dissection always necessary in N1b papillary thyroid carcinoma patients? World J Surg, 32:716–721.

24.

Lim

, Choi

, Yoon

, Koo

. 2010. Occult lymph node metastases in neck level V in papillary thyroid carcinoma. Surgery, 147:241–245.

25.

Merdad

, Eskander

, Kroeker

, Freeman

. 2012. Metastatic papillary thyroid cancer with lateral neck disease: pattern of spread by level. Head Neck[Epub ahead of print]10.1002/hed.23149.

26.

Pingpank

Jr , Sasson

, Hanlon

, Friedman

, Ridge

. 2002. Tumor above the spinal accessory nerve in papillary thyroid cancer that involves lateral neck nodes: a common occurrence. Arch Otolaryngol Head Neck Surg, 128:1275–1278.

27.

Roh

, Kim

, Park

. 2008. Lateral cervical lymph node metastases from papillary thyroid carcinoma: pattern of nodal metastases and optimal strategy for neck dissection. Ann Surg Oncol, 15:1177–1182.

28.

Sivanandan

, Soo

. 2001. Pattern of cervical lymph node metastases from papillary carcinoma of the thyroid. Br J Surg, 88:1241–1244.

29.

Vayisoglu

, Ozcan

, Turkmenoglu

, Gorur

, Unal

, Dag

, Ocal

. 2010. Level IIb lymph node metastasis in thyroid papillary carcinoma. Eur Arch Otorhinolaryngol, 267:1117–1121.

30.

Yanir

, Doweck

. 2008. Regional metastases in well-differentiated thyroid carcinoma: pattern of spread. Laryngoscope, 118:433–436.

31.

Yuce

, Cagli

, Bayram

, Karasu

, Guney

. 2010. Regional metastatic pattern of papillary thyroid carcinoma. Eur Arch Otorhinolaryngol, 267:437–441.

32.

Mazzaferri

, Jhiang

. 1994. Long-term impact of initial surgical and medical therapy on papillary and follicular thyroid cancer. Am J Med, 97:418–428.

33.

Simpson

, McKinney

, Carruthers

, Gospodarowicz

, Sutcliffe

, Panzarella

. 1987. Papillary and follicular thyroid cancer. Prognostic factors in 1,578 patients. Am J Med, 83:479–488.

34.

Scheumann

, Gimm

, Wegener

, Hundeshagen

, Dralle

. 1994. Prognostic significance and surgical management of locoregional lymph node metastases in papillary thyroid cancer. World J Surg, 18:559–567; discussion 567–568.

35.

Ries

, Young

, Keel

, Eisner

, Lin

, Horner

. 2007. SEER Survival Monograph Cancer Survival Among Adults. U.S., SEER Program, 1988–2001, Patient and Tumor Characteristics., NIH Pub. No. 07-6215 edition. National Cancer Institute, SEER Program: Bethesda, MD.

36.

Edge

, Byrd

, Compton

, Fritz

, Greene

, Trotti

. 2010. AJCC Cancer Staging Manual, 7th. Springer: New York.

37.

Cady

, Rossi

. 1988. An expanded view of risk-group definition in differentiated thyroid carcinoma. Surgery, 104:947–953.

38.

Hay

, Grant

, Taylor

, McConahey

. 1987. Ipsilateral lobectomy versus bilateral lobar resection in papillary thyroid carcinoma: a retrospective analysis of surgical outcome using a novel prognostic scoring system. Surgery, 102:1088–1095.

39.

Hay

, Bergstralh

, Goellner

, Ebersold

, Grant

. 1993. Predicting outcome in papillary thyroid carcinoma: development of a reliable prognostic scoring system in a cohort of 1779 patients surgically treated at one institution during 1940 through 1989. Surgery, 114:1050–1057; discussion 1057–1058.

40.

Cappiello

, Piazza

, Giudice

, De Maria

, Nicolai

. 2005. Shoulder disability after different selective neck dissections (levels II-IV versus levels II-V): a comparative study. Laryngoscope, 115:259–263.

41.

Prim

, De Diego

, Verdaguer

, Sastre

, Rabanal

. 2006. Neurological complications following functional neck dissection. Eur Arch Otorhinolaryngol, 263:473–476.

42.

Kuntz

, Weymuller

Jr . 1999. Impact of neck dissection on quality of life. Laryngoscope, 109:1334–1338.

43.

Higgins

, Thompson

, Deeks

, Altman

. 2003. Measuring inconsistency in meta-analyses. BMJ, 327:557–560.