Computed tomography versus magnetic resonance imaging for diagnosing cervical lymph node metastasis of head and neck cancer: a systematic review and meta-analysis

Abstract

Computed tomography (CT) and magnetic resonance imaging (MRI) are common imaging methods to detect cervical lymph node metastasis of head and neck cancer. We aimed to assess the diagnostic efficacy of CT and MRI in detecting cervical lymph node metastasis, and to establish unified diagnostic criteria via systematic review and meta-analysis. A systematic literature search in five databases until January 2014 was carried out. All retrieved studies were reviewed and eligible studies were qualitatively summarized. Besides pooling the sensitivity (SEN) and specificity (SPE) data of CT and MRI, summary receiver operating characteristic curves were generated. A total of 63 studies including 3,029 participants were involved. The pooled results of meta-analysis showed that CT had a higher SEN (0.77 [95% confidence interval {CI} 0.73–0.87]) than MRI (0.72 [95% CI 0.70–0.74]) when node was considered as unit of analysis (P<0.05); MRI had a higher SPE (0.81 [95% CI 0.80–0.82]) than CT (0.72 [95% CI 0.69–0.74]) when neck level was considered as unit of analysis (P<0.05) and MRI had a higher area under concentration-time curve than CT when the patient was considered as unit of analysis (P<0.05). With regards to diagnostic criteria, for MRI, the results showed that the minimal axial diameter of 10 mm could be considered as the best size criterion, compared to 12 mm for CT. Overall, MRI conferred significantly higher SPE while CT demonstrated higher SEN. The diagnostic criteria for MRI and CT on size of metastatic lymph nodes were suggested as 10 and 12 mm, respectively.

Keywords:

• computed tomography
• magnetic resonance imaging
• metastasis
• head and neck cancer
• meta-analysis

Introduction

The occurrence of cervical lymph node metastasis in patients with head and neck cancers are very common.¹ The presence of cervical lymph node metastasis may affect the optimal treatment choice as well as prognosis in patients.² Management of patients presenting with cervical lymph node metastasis includes selective or radical neck dissection, followed by radiotherapy and/or chemotherapy depending on the pathological findings of the nodes.^3–5 Besides, the detection of cervical lymph node metastasis is very important for predicting prognosis in patients with head and neck cancers.^6–8

Many imaging techniques exist for identifying cervical lymph node metastasis in patients with head and neck cancers.^9–12 Among them, computed tomography (CT) and magnetic resonance imaging (MRI) are the most widely used tools.¹³ Both of them have improved accuracy of nodal staging over clinical palpation and the nodes which are clinically occulted can be visualized through these techniques.¹⁴ Usually the cervical lymph nodes demonstrate similar density as muscle on pre-contrast images of CT examination, and they can be separated from adjacent vessels by their differential enhancement after contrast administration.¹⁵ On the other hand, MRI is considered to have similar accuracy for identifying the cervical lymph node metastasis of head and neck cancer.^16,17 Because of the intrinsic high soft-tissue discrimination, MRI has become the preferred method for evaluating the soft tissues of the head and neck recently.¹⁸ Under current health care settings, the routine practice for evaluating patients with head and neck cancer is to perform either CT or MRI, but not both.¹⁹ Thus, to determine whether one of the two techniques is superior to the other is critical for providing guidance for clinical practice. Besides, since relevant studies utilized very different diagnostic criteria, it is warranted to determine the unified criteria that are most appropriate. A systematic review to assess all available evidence is thus needed for providing a comprehensive evaluation for these aims.

The aim of this study was thus to compare CT and MRI for detecting cervical lymph node metastasis in patients with head and neck cancer and to establish the unified diagnostic criteria by performing a systematic review and meta-analysis.

Methods

Inclusion criteria

The inclusion criteria were as follows: a) types of study: diagnostic accuracy test studies designed as cohort studies; b) participants: patients with biopsy proven head and neck cancers who would undergo neck dissection; c) index tests: CT and/or MRI; d) target condition: cervical lymph node metastasis; e) reference standard: histopathology examination; f) outcome: rates of true positive, false positive, false negative, and true negative or related data that could be used to calculate them.

Literature search

With no language restriction, the following databases were searched for retrieving studies: MEDLINE (1948 to 25 January 2014), EMBASE (1980 to 25 January 2014), China National Knowledge Infrastructure (1994 to 25 January 2014), VIP Chinese Journal Database (1989 to 25 January 2014), and Chinainfo (1998 to 25 January 2014).

The search strategy was optimized for all consulted databases, taking into account the differences in the various controlled vocabularies as well as the differences of database-specific technical variations.²⁰ Once relevant articles were identified, their reference lists were searched for additional articles. Both Medical Subject Headings (MeSH) and free text words were used in the search strategy with the following MeSH terms: “head and neck neoplasm”, “neoplasm metastases”, “SEN and SPE”, “Tomography, Spiral Computed” and “Magnetic Resonance Imaging”.

Study selection

Two reviewers independently examined the titles and abstracts of each search record to remove obviously irrelevant ones, and then retrieved the full text articles for potentially eligible articles. The full-texts were further examined according to the inclusion criteria. Discrepancies were resolved by consensus.

Data extraction

A standardized data extraction form was used by two authors independently for data extraction from included studies. Discrepancies were resolved by discussion, with input from a third author. The contents of the form included: name of first author, publication year, country, participants’ age, sex, number of included patients, tumor location, unit, details of CT and/or MRI, study design (prospective or retrospective).

Quality assessment

The methodological quality of included studies was assessed by The Quality Assessment Diagnostic Accuracy Studies statement-2 (QUADAS-2),²¹ which included four domains: patient selection, index test, reference standard, and flow and timing. Each domain was assessed in terms of risk of bias and the first three were assessed in terms of concerns regarding applicability. Signaling questions were included to assist judgments on risk of bias. The signaling questions in the QUADAS-2 were presented as shown in Table 1. The result for each item was categorized as yes (Y), unclear (U), or no (N). The summary risk of bias for each study was categorized as low (A), unclear (B), or high (C).

Table 1 Signaling questions in the QUADAS-2

Domain	Patient selection	Index test	Reference standard	Flow and timing
Signaling questions (yes/no/unclear)	1 Was a consecutive or random sample of patients enrolled?	4 Were the index test results inter preted without knowledge of the results of the reference standard?	5 Is the reference standard likely to correctly classify the target condition?	7 Was there an appropriate interval between index test(s) and reference standard?
	2 Was a case-control design avoided?		6 Were the reference standard results interpreted without know ledge of the results of the index test?	8 Did all patients receive a reference standard?
	3 Did the study avoid inappropriate exclusions?			9 Were all patients included in the analysis?

Abbreviation: QUADAS-2, The Quality Assessment Diagnostic Accuracy Studies statement-2.

Meta-analysis

Measures of diagnostic efficacy of CT and/or MRI included sensitivity (SEN), specificity (SPE), positive likelihood ratio (+LR), negative likelihood ratio (−LR), accuracy (ACC), and diagnostic odds ratios (DOR) with 95% confidence intervals (CIs). Summary receiver operating characteristic (SROC) curves were then drawn. The area under the curve (AUC) and Q* (the point where SEN is equal to SPE on the SROC curve) were calculated.

To detect any differences for SEN, SPE, AUC, and Q* between CT and MRI, a Z-test was conducted (Z= (VAL1–VAL2)/SQRT (SE1²+SE2²). The test standard was set at α=0.05. VAL indicates the mean of SEN, SPE, AUC or Q* of the CT or MRI and SE indicates the standard error of the corresponding variable.

Heterogeneity analysis

Heterogeneity between studies was evaluated by I² statistic.^22,23 If I²≤50% and P≥0.10, the heterogeneity was considered not significant and in such case the fixed-effects model would be used in meta-analysis. Otherwise, the random-effects model would be used.^24,25

Meta-regression

Meta-regression was used to determine any potential source of heterogeneity that might influence the overall assessment. The test standard for meta-regression was set at α=0.10. Relevant variables which might cause heterogeneities were tested, and any suggested sources of heterogeneity were considered as proof for a subgroup analysis. Variables detected by meta-regression included publication year (0= published before 2000; 1= published in or after 2000), race (0= Mongolia; 1= Caucasian), study type (0= retrospective; 1= prospective), risk of bias (0= high; 1= unclear; 2= low), blinding of the radiologists (0= no or unclear; 1= yes) and blinding of the pathologists (0= no or unclear; 1= yes). Meta-disc 1.4 and STATA 11.0 (StataCorp LP, College Station, TX, USA) were used to perform the statistical analyses.^26,27

Results

Selection of literature

The computerized and manual search retrieved a total of 306 articles. After assessing the titles and abstracts, 144 articles were found to be potentially relevant. After the full text assessment, 63 studies met the inclusion criteria and were included in this meta-analysis (Figure 1).^28–90

Figure 1 Flow chart of the literature search and selection.

Abbreviations: CT, computed tomography; MRI, magnetic resonance imaging.

Study characteristics

Of the 63 included studies, 24 were retrospective and 39 were prospective. A total of 3,029 participants were involved in these studies. Among those patients, 1,044 underwent both CT and MRI examination, 2,395 underwent MRI examination, and 1,678 underwent CT examination. Three kinds of unit of analysis were used, including node, neck level (the neck was classified as five levels according to anatomical landmarks), and patients. When node was considered as the unit of analysis, available studies involved 22 with CT and 30 with MRI. When neck level was considered as the unit of analysis, eight studies with CT and 16 with MRI were available. When patient was considered as the unit of analysis, available studies included eight with CT and eleven with MRI. The tumor locations included floor of mouth, nasopharynx, retromolar trigonum, mandibule, maxilla, supra-glottic larynx, oropharynx, laryngopharynx, hypopharynx, parotid gland, submandibular gland, tonsil, thyroid gland, cervical esophageal, paranasal sinuses et al. The characteristics of included studies are listed in Table 2.

Table 2 Study characteristics and included data sets for CT and MRI of the included articles

Study ID	Country	Study type	Patients (M/F)	Age (yr), mean (range)	Tumor location	Imaging modality	Unit
Adams et al^Citation28 1998	Germany	P	60 (16/44)	58.3 (38–76)	Tongue, FOM, Palate, MAN, MAX	MRI, CT	node
Akoglu et al^Citation29 2005	Turkey	P	23 (19/4)	58.3 (40–78)	Head and neck	MRI, CT	node
Anzai et al^Citation30 1994	USA	P	12 (7/5)	39–78	EAC, MAN, BCC, RMT, Lip, Oral cavity, Larynx	MRI	node
Ao et al^Citation31 1998	Japan	R	42 (9/33)	60 (39–78)	Larynx	MRI, CT	node
Bondt et al^Citation32 2009	The Netherlands	P	16 (9/7)	40–77	Tongue, NP, RMT, SMG, Cheek, RMT, SP, Nose	MRI, CT	neck level
Braams et al^Citation33 1996	The Netherlands	P	11 (7/4)	62.3 (46–73)	FOM, RMT, Cheek, Gingiva	MRI, CT	node
Braams et al^Citation34 1995	The Netherlands	P	12 (8/4)	65.3 (48–85)	Tongue, Lip, Gingiva, RMT, FOM	MRI	node
Bruschini et al^Citation35 2003	Italy	P	22 (19/3)	62.3 (46–79)	Larynx, OP, Oral cavity, Skin	CT	node
Curtin et al^Citation36 1997	Canada	R	213 (150/63)	59.6 (18–84)	Oral cavity, OP, HP, Larynx	MRI, CT	neck level
Dammann et al^Citation37 2005	Germany	P	64 (43/21)	56 (26–83)	Oral cavity, OP	MRI, CT	neck level
Ding et al^Citation38 2005	People’s Republic of China	P	92 (58/34)	53 (24–81)	Tongue	MRI	neck level
Dirix et al^Citation39 2010	Sweden	P	22 (13/9)	60 (41–83)	Oral cavity, Larynx, HP	MRI	node
Eida et al^Citation40 2003	Japan	P	111 (74/37)		FOM, Tongue, Palate, Gingiva, Cheek	CT	node
Fan et al^Citation41 2006	People’s Republic of China	R	42 (37/5)	53.6 (45–70)	OP, HP, Cervical esophageal	CT	patient
Fukunari et al^Citation42 2010	Japan	R	20	58 (23–81)	Tongue, Gingiva, Buccal, MAN, FOM	MRI	node
Gross et al^Citation43 2001	USA	R	26 (8/18)	40 (10–80)	Thyroid	MRI	node
Gu et al^Citation44 2000	People’s Republic of China	P	62	58 (44–77)	Head and neck	MRI	node
Guenzel et al^Citation45 2013	Germany	P	120 (95/25)	41–85	OP, Larynx	MRI	node
Guo et al^Citation46 2006	People’s Republic of China	R	48 (28/20)	56 (21–66)	Tongue, Buccal, Gingiva, FOM, Palate	MRI	node
Hannah et al^Citation47 2002	Australia	P	48 (34/14)	61 (26–92)	Oral cavity, OP, SGL, HP	CT	neck level
Hao et al^Citation48 2000	People’s Republic of China	P	60		Tongue, Gingiva, FOM, Palate, RMT, Buccal, Larynx, HP	MRI	node
Hafidh et al^Citation49 2006	Ireland	R	48 (42/6)	56 (32–80)	Oral cavity, OP, HP, Paranasal sinuses, Ear(skin)	MRI, CT	node
Hlawitschka et al^Citation50 2002	Germany	P	38 (28/10)	59 (41–89)	Tongue, Buccal, Palate, MAX	MRI, CT	node
Hoffman et al^Citation51 2000	USA	P	9 (6/3)	43–76	Oral cavity, OP, Lip	MRI	node, neck level
Jeong et al^Citation52 2007	Greece	R	47 (41/6)	56.3	Oral cavity, Larynx, OP, HP, PG	CT	neck level
Kau et al^Citation53 1999	Germany	P	111 (95/16)	29–78	Larynx, OP, LP, Lip, Ear	MRI, CT	node, neck level
Kawai et al^Citation54 2005	Japan	P	29 (23/6)	60 (28–81)	Tongue, OP, NP, Larynx, Buccal, Palate, PG, Gingiva	MRI	neck level
Ke et al^Citation55 2006	People’s Republic of China	R	20 (15/5)	54.5 (31–69)	Tongue, Larynx, Thyroid gland	CT	node
Krabbe et al^Citation56 2008	The Netherlands	P	38 (21/17)	59 (53–680)	Tongue, Gingiva, FOM, Tonsillar fossa	MRI, CT	node
Laubenbacher et al^Citation57 1994	Germany	P	22 (20/2)	54.4 (38–70)	OP, HP	MRI	node, neck level
Lee et al^Citation58 2013	People’s Republic of China	P	22 (21/1)	49.8 (26–66)	Tongue, Buccal, OP, FOM, HP, Palate, RMT, epiglottis, Pyriform sinus	MRI	patient
Lu et al^Citation59 2007	People’s Republic of China	P	13 (11/2)	58 (47–71)	Oral cavity, HP, OP, Larynx	CT	node
Lwin et al^Citation60 2012	UK	R	102 (68/34)	59 (23–89)	Tongue, FOM, Palate, Buccal, RMT, Tonsil, Gingiva	MRI	patient
Mcguirt et al^Citation61 1995	UK	P	49		Oral cavity, OP, HP	CT	node
Nakamoto et al^Citation62 2009	Japan	R	65 (50/15)	62 (27–81)	Larynx, HP, MAX, Tongue, OP, PG, Gingiva, FOM, NP, Ethmoid, EAM, Thyoid	MRI	patient
Nishimura et al^Citation63 2006	Japan	P	16 (13/3)	65.8 (37–76)	Cervical Esophageal	MRI	node
Olmos et al^Citation64 1999	The Netherlands	P	12 (6/6)	61.8 (44–73)	OP, Larynx, HP, Tongue, MAX	MRI	neck level
Ou et al^Citation65 2007	People’s Republic of China	R	24 (19/5)	50 (23–80)	Tongue, OP, Palate, Cheek, Maxillary sinus, Branchial cleft	MRI	node
Paulus et al^Citation66 1998	Belgium	R	25 (21/4)	48–74	SGL, Tongue, Glottis, Palate, RMT, FOM, HP, Vocal cord, Vestibule, Pyriform sinus	CT	node
Perrone et al^Citation67 2011	Italy	R	17 (10/7)	63 (15–85)	Head and neck	MRI	patient
Peters et al^Citation68 2013	The Netherlands	R	149 (120/29)	62 (40–78)	SGL, Glottis, NP, Cervical Esophageal	MRI, CT	patient
Pohar et al^Citation69 2006	USA	R	25 (17/8)	63.4	Oral cavity, OP, HP, Larynx, Nasal cavity	CT	node, neck level
Ren et al^Citation70 2000	People’s Republic of China	P	20 (18/2)	45–68	SGL	CT	node
Schwartz et al^Citation71 2004	USA	P	20 (20/0)	61 (42–78)	Oral cavity, OP	CT	node
Semedo et al^Citation72 2006	Portugal	P	20 (20/0)	57.3 (36–78)	HP, Larynx, OP	MRI	node
Seitz et al^Citation73 2009	Germany	R	66 (39/27)	63 (25–89)	Oral cavity, OP	MRI	node, patient
Stokkel et al^Citation74 2000	The Netherlands	P	54 (31/23)	60 (34–81)	Tongue, FOM, Gingiva, RMT, OP	CT	node
Stuckensen et al^Citation75 2000	Germany	P	106 (89/17)	59.6 (33–87)	FOM, Tongue, RMT, MAN, MAX, Buccal	MRI, CT	neck level
Sumi et al^Citation76 2007	Japan	R	38 (32/6)	65	HP, Gingiva, OP, Tongue, Larynx, FOM	MRI, CT	node
Sumi et al^Citation77 2006	Japan	P	26		OP, Gingiva, Larynx, Tongue	MRI	node
Sumi et al^Citation78 2003	Japan	P	32	24–80	OP, Gingiva, FOM, Tongue, Buccal, EAC	MRI	node
Sun et al^Citation79 2013	People’s Republic of China	R	114 (60/54)	51.2 (34–70)	Thyroid gland, Larynx, NP, HP, Tongue, PG, Cervical Esophageal, Maxillary sinus, Ear	CT	node
Sun et al^Citation79 2013	People’s Republic of China	R	86 (45/41)	52.7 (35–75)	Thyroid gland, Larynx, NP, HP, Tongue, PG, Cervical Esophageal, Maxillary sinus, Ear	MRI	node
Tai et al^Citation80 2002	People’s Republic of China	P	40 (24/16)	25–65	NP	MRI	patient
Takashima et al^Citation81 1997	Japan	R	50 (13/37)	57 (24–81)	Thyroid	MRI	node
Tuli et al^Citation82 2008	India	P	20 (12/8)	54.75 (30–85)	Tongue	MRI, CT	patient
Van den Brekel et al^Citation83 1991	The Netherlands	P	100	63±12.8	Tongue, FOM, SP, Lip, Tonsil, Pharyngeal wall, Ear, Tonsil, PS, SGL, Gingiva	MRI	patient
Vandecaveye et al^Citation84 2008	Belgium	P	36	41–81	Nasal cavity, SGL, FOM, OP, Glottis, Tongue, HP	MRI	node, neck level, patient
Wang et al^Citation85 1999	Japan	P	14 (10/4)	46 (26–71)	Thyroid	MRI	node
WIDE et al^Citation86 1999	UK	R	58	58.1 (32–82)	Tongue, FOM, Buccal, RMT, OP, Gingiva	MRI	neck level
Wilson et al^Citation87 1994	UK	P	12		FOM, Tongue, Tonsillar, Skin, Pinna, PG, Thyroid	MRI	neck level
Wu et al^Citation88 2010	People’s Republic of China	R	24 (23/1)	53.6 (45–85)	Larynx, HP	CT	node
Yoon et al^Citation89 2008	Korea	R	67 (58/9)	60 (24–85)	Larynx, Pharynx, Tonsil, Tongue, Oral cavity, Skin, MAX	MRI, CT	neck level
Yuan et al^Citation90 2000	People’s Republic of China	R	19 (12/7)	42–66	Larynx	MRI	neck level

Abbreviations: M, male; F, female; R, Retrospective; P, Prospective; EAC, external auditory canal; BCC, branchial cleft cyst; PS, piriform sinus; SGL, supra-glottic larynx; TGL, trans-glottic larynx; CT, computed tomography; MRI, magnetic resonance imaging; FOM, floor of mouth; MAN, mandibule; MAX, maxilla; RMT, retro-molar trigonum; NP, nasopharynx; SMG, submandibular gland; OP, oropharynx; HP, hypopharynx; LP, laryngopharynx; PG, parotid gland; SP, supropharynx; yr, years.

Quality of included studies

All included studies had fairly good applicability. For the risk of bias assessment, only two studies had a low risk of bias, five had a high risk, and 56 had an unclear risk (Table 3).

Table 3 Risk of bias of included studies

Study ID	Patient selection			Index test	Reference standard			Flow and timing		Summary risk of bias	Applicability
	1	2	3	4	5	6	7	8	9
Adams et al^Citation28 1998	U	Y	Y	Y	Y	U	Y	Y	Y	B	H
Akoglu et al^Citation29 2005	Y	Y	Y	U	Y	U	U	Y	Y	B	H
Anzai et al^Citation30 1994	U	Y	Y	U	Y	U	Y	Y	Y	B	H
Ao et al^Citation31 1998	U	Y	Y	U	Y	U	U	Y	Y	B	H
Bondt et al^Citation32 2009	Y	Y	Y	Y	Y	U	U	Y	Y	B	H
Braams et al^Citation33 1996	U	Y	Y	Y	Y	U	Y	Y	Y	B	H
Braams et al^Citation34 1995	U	Y	Y	Y	Y	U	U	Y	Y	B	H
Bruschini et al^Citation35 2003	U	Y	Y	Y	Y	Y	U	Y	Y	B	H
Curtin et al^Citation36 1997	Y	Y	Y	U	Y	U	U	Y	Y	B	H
Dammann et al^Citation37 2005	U	Y	Y	Y	Y	U	Y	Y	Y	B	H
Ding et al^Citation38 2005	U	Y	Y	Y	Y	U	Y	Y	Y	B	H
Dirix et al^Citation39 2010	U	Y	Y	U	Y	U	Y	Y	Y	B	H
Eida et al^Citation40 2003	U	Y	Y	Y	Y	U	U	Y	Y	B	H
Fan et al^Citation41 2006	U	Y	Y	Y	Y	U	U	Y	N	A	H
Fukunari et al^Citation42 2010	U	Y	Y	U	Y	U	U	Y	Y	B	H
Gross et al^Citation43 2001	U	Y	Y	Y	Y	U	Y	Y	Y	B	H
Gu et al^Citation44 2000	U	Y	Y	Y	Y	U	U	Y	Y	B	H
Guenzel et al^Citation45 2013	U	Y	Y	U	Y	U	U	Y	Y	B	H
Guo et al^Citation46 2006	U	Y	Y	U	Y	U	U	Y	N	A	H
Hannah et al^Citation47 2002	U	Y	Y	U	Y	U	U	Y	Y	B	H
Hao et al^Citation48 2000	U	Y	Y	Y	Y	Y	U	Y	Y	B	H
Hafidh et al^Citation49 2006	U	Y	Y	Y	Y	U	U	Y	Y	B	H
Hlawitschka et al^Citation50 2002	Y	Y	Y	U	Y	U	U	Y	N	A	H
Hoffman et al^Citation51 2000	U	Y	Y	U	Y	U	U	Y	Y	B	H
Jeong et al^Citation52 2007	U	Y	Y	Y	Y	U	U	Y	Y	B	H
Kau et al^Citation53 1999	Y	Y	Y	Y	Y	U	Y	Y	Y	B	H
Kawai et al^Citation54 2005	Y	Y	Y	Y	Y	U	Y	Y	Y	B	H
Ke et al^Citation55 2006	Y	Y	Y	Y	Y	U	Y	Y	Y	B	H
Krabbe et al^Citation56 2008	U	Y	Y	U	Y	U	U	Y	Y	B	H
Laubenbacher et al^Citation57 1994	U	Y	Y	U	Y	U	U	Y	Y	B	H
Lee et al^Citation58 2013	Y	Y	Y	U	Y	U	Y	Y	Y	B	H
Lu et al^Citation59 2007	Y	Y	Y	Y	Y	U	U	Y	Y	B	H
Lwin et al^Citation60 2012	U	Y	Y	Y	Y	U	U	Y	Y	B	H
Mcguirt et al^Citation61 1995	Y	Y	Y	U	Y	Y	U	Y	Y	B	H
Nakamoto et al^Citation62 2009	U	Y	Y	U	Y	U	U	Y	Y	B	H
Nishimura et al^Citation63 2006	Y	Y	Y	U	Y	U	Y	Y	Y	B	H
Olmos et al^Citation64 1999	U	Y	Y	U	Y	U	Y	Y	N	A	H
Ou et al^Citation65 2007	U	Y	Y	U	Y	U	U	Y	Y	B	H
Paulus et al^Citation66 1998	U	Y	Y	U	Y	U	U	Y	Y	B	H
Perrone et al^Citation67 2011	U	Y	Y	U	Y	U	U	Y	Y	B	H
Peters et al^Citation68 2013	U	Y	Y	Y	Y	U	U	Y	Y	B	H
Pohar et al^Citation69 2006	Y	Y	Y	Y	Y	U	U	Y	Y	B	H
Ren et al^Citation70 2000	U	Y	Y	Y	Y	U	U	Y	Y	B	H
Schwartz et al^Citation71 2004	U	Y	Y	Y	Y	U	U	Y	Y	B	H
Semedo et al^Citation72 2006	Y	Y	Y	Y	Y	U	U	Y	Y	B	H
Seitz et al^Citation73 2009	Y	Y	Y	Y	Y	Y	Y	Y	Y	C	H
Stokkel et al^Citation74 2000	U	Y	Y	U	Y	U	Y	Y	Y	B	H
Stuckensen et al^Citation75 2000	Y	Y	Y	U	Y	U	Y	Y	Y	B	H
Sumi et al^Citation76 2007	U	Y	Y	U	Y	U	Y	Y	Y	B	H
Sumi et al^Citation77 2006	Y	Y	Y	Y	Y	U	Y	Y	Y	B	H
Sumi et al^Citation78 2003	Y	Y	Y	U	Y	U	U	Y	Y	B	H
Sun et al^Citation79 2013	Y	Y	Y	Y	Y	U	U	Y	Y	B	H
Tai et al^Citation80 2002	U	Y	Y	Y	Y	U	U	Y	N	A	H
Takashima et al^Citation81 1997	U	Y	Y	Y	Y	U	Y	Y	Y	B	H
Tuli et al^Citation82 2008	Y	Y	Y	Y	Y	U	Y	Y	Y	B	H
Van den Brekel et al^Citation83 1991	U	Y	Y	Y	Y	U	Y	Y	Y	B	H
Vandecaveye et al^Citation84 2008	Y	Y	Y	Y	Y	Y	U	Y	Y	B	H
Wang et al^Citation85 1999	U	Y	Y	Y	Y	Y	Y	Y	Y	C	H
WIDE et al^Citation86 1999	U	Y	Y	Y	Y	U	U	Y	Y	B	H
Wilson et al^Citation87 1994	Y	Y	Y	Y	Y	U	U	Y	Y	B	H
Wu et al^Citation88 2010	U	Y	Y	U	Y	U	U	Y	Y	B	H
Yoon et al^Citation89 2008	U	Y	Y	U	Y	U	Y	Y	Y	B	H
Yuan et al^Citation90 2000	U	Y	Y	U	Y	U	Y	Y	Y	B	H

Abbreviations: Y, yes; U, unclear; N, no; A, high risk of bias; B, unclear risk of bias; C, low risk of bias; H, high applicability.

Comparison of CT and MRI in detecting cervical lymph node metastasis with node as unit of analysis

For CT, meta-regression analysis showed that the diagnostic efficacy was not affected by any of the tested variables. These variables thus did not account for heterogeneity between studies. After pooling 22 studies, we detected that CT had a mean (CI) SEN of 0.77 (95% CI 0.73–0.80), SPE of 0.85 (0.84–0.87), +LR of 3.84 (2.51–5.87), −LR of 0.34 (0.24–0.27), ACC of 0.8357, and DOR of 13.57 (6.99–26.33). The SROC was demonstrated in Figure 2 and the AUC was 0.8429 and Q* was 0.7745. For MRI, meta-regression analysis also showed that the diagnostic efficacy was not affected by any of the tested variables. After pooling 30 studies, we identified that MRI had a mean (CI) SEN of 0.72 (0.70–0.74), SPE of 0.84 (0.83–0.85), +LR of 5.06 (3.72–6.88), −LR of 0.27 (0.21–0.34), ACC of 0.8126, and DOR of 25.21 (15.97–39.80). The SROC is shown in Figure 2 and the AUC was 0.9054 and Q* was 0.8371.

Figure 2 Summary receiver operator characteristic curves of CT and MRI (node as unit of analysis).

Abbreviations: CT, computed tomography; MRI, magnetic resonance imaging.

By comparing the diagnostic efficacy between CT and MRI when node was treated as the unit of analysis, the results indicated that CT had a higher SEN, although the SPE and summarized diagnostic efficacy were comparable. The details are listed in Table 4.

Table 4 Comparison of meta-analysis results on diagnostic efficacy between CT and MRI

Unit	Variable	Number detected	SEN (95% CI)	SPE (95% CI)	AUC (SE)	Q* (SE)
Node	CT	2,483	0.77 (0.73–0.87)	0.85 (0.84–0.87)	0.8429 (0.0341)	0.7745 (0.0318)
	MRI	7,100	0.72 (0.70–0.74)	0.84 (0.83–0.85)	0.9054 (0.0198)	0.8371 (0.0215)
	P		0.0176	0.2739	0.1098	0.1262
Neck level	CT	1,665	0.84 (0.75–0.84)	0.72 (0.69–0.74)	0.8787 (0.0268)	0.8091 (0.0270)
	MRI	4,022	0.80 (0.77–0.82)	0.81 (0.80–0.82)	0.8860 (0.0262)	0.8165 (0.0269)
	P		1.0000	0.0000	0.8689	0.8702
Patient	CT	230	0.67 (0.52–0.80)	0.74 (0.68–0.81)	0.6860 (0.0815)	0.6418 (0.0643)
	MRI	716	0.78 (0.70–0.81)	0.76 (0.72–0.80)	0.8631 (0.0437)	0.7937 (0.0424)
	P		0.1992	0.6161	0.0491	0.0683

Abbreviations: AUC, area under the curve; CI, confidence interval; SE, standard error; CT, computed tomography; MRI, magnetic resonance imaging; SEN, sensitivity; SPE, specificity.

Comparison of CT and MRI in detecting cervical lymph node metastasis with neck level as unit of analysis

For MRI, meta-regression analysis detected that none of the tested variables accounted for heterogeneity between studies. After pooling 16 studies, it was detected that MRI had a mean (CI) SEN of 0.80 (0.77–0.82), SPE of 0.81 (0.80–0.82), +LR of 5.34 (3.24–8.82), −LR of 0.27 (0.20–0.37), ACC of 0.5257, DOR of 24.61 (12.21–49.61) and the AUC was 0.8860 and Q* was 0.8165 (Figure 3). For CT, similarly none of the tested variables accounted for heterogeneity. The pooling of available studies identified that CT had a mean (CI) SEN of 0.80 (0.75–0.84), SPE of 0.72 (0.69–0.74), +LR of 5.60 (2.13–14.73), −LR of 0.26 (0.19–0.36), ACC of 0.6888, DOR of 23.76 (7.87–71.79) and the AUC was 0.8787 and Q* was 0.8091 (Figure 3).

Figure 3 Summary receiver operator characteristic curves of CT and MRI (neck level as unit of analysis).

Abbreviations: CT, computed tomography; MRI, magnetic resonance imaging.

The comparison between CT and MRI showed that MRI had significantly higher SPE than CT while the other variables were comparable between these two techniques (Table 4).

Comparison of CT and MRI in detecting cervical lymph node metastasis with patient as unit of analysis

For the two studies, the pooled results showed that CT had a mean (CI): SEN, 0.81 (0.65–0.92); SPE, 0.35 (0.24–0.42); +LR, 1.14 (0.87–1.50); −LR, 0.70 (0.32–1.52); DOR, 1.66 (0.57–4.82) (Figure S1). For MRI, which included ten studies, meta-regression analysis showed that study type significantly affected the assessment of diagnostic efficacy (P=0.04) (Table 5). Based on the subgroup analysis according to study types, for the four retrospective studies, the pooled results indicated that MRI had a mean (CI) SEN, 0.77 (0.69–0.85); SPE, 0.48 (0.42–0.55); +CR, 2.42 (0.99–5.91); −CR, 0.54 (0.27–1.06); DOR, 5.24 (0.96–28.55) (Figure S2). For the five prospective studies, the pooled results showed that MRI had a mean (CI) SEN, 0.80 (0.72–0.86); SPE, 0.35 (0.67–0.86); +LR, 2.79 (1.44–5.40); −LR, 0.25 (0.08–0.76); DOR, 14.63 (3.64–58.70) (Figure S3). Pooling of the overall nine studies indicated the mean (CI) values for the following parameters to be: SEN, 0.79 (0.73–0.84); SPE, 0.56 (0.51–0.62); +LR, 2.64 (1.30–5.34); −LR, 0.37(0.20–0.71); DOR, 8.87 (2.42–32.55); AUC (0.8158); Q* (0.7498) (Figure S4).

Table 5 Results of meta-regression (MRI patient)

Variable	Coefficient	SE	P-value	RDOR	95% CI
Cte	−0.511	2.5493	0.8539	–	–
S	−0.330	0.1896	0.1798	–	–
Publication year	0.881	1.5156	0.6020	2.41	(0.02–300.01)
Race	1.786	1.1884	0.2298	5.97	(0.14–262.04)
Study type	3.288	0.9742	0.0432	26.80	(1.21–595.04)
Blinding of radiologists	−0.774	1.1952	0.5636	0.46	(0.01–20.70)
Blinding of pathologists	−0.290	1.5278	0.8615	0.75	(0.01–96.74)
Risk of bias	−0.227	0.9225	0.8217	0.80	(0.04–15.02)

Abbreviations: MRI, magnetic resonance imaging; CI, confidence interval; SE, standard error; RDOR, relative diagnostic odds ratio.

The comparison between CT and MRI showed that MRI had significantly higher AUC than CT while the other variables demonstrated no statistical significance between them. The details are listed in Table 4.

Lymph node size criteria

The size of metastatic lymph nodes used as diagnostic criteria of MRI and CT varied considerably among studies and among different neck levels (Table S1). To determine the best diagnostic criteria, a meta-analysis was conducted for different neck levels with lymph node unit data. For each neck level, the SROC curve was drawn to show the diagnostic efficacy of MRI for different node sizes (Figure 4). The results revealed that the minimal axial diameter of 10 mm in lymph node-bearing regions could be considered as the best size criterion for assessing cervical lymph node metastasis in patients with head and neck cancer (Table S2). For CT, the suggested criterion was 12 mm (Table S3). Considering the limited number of studies for CT, SROC curves were not drawn.

Figure 4 Summary receiver operator characteristic curves of CT and MRI (lymph node size criteria).

Abbreviations: CT, computed tomography; MRI, magnetic resonance imaging; SROC, summary receiver operating characteristic.

Discussion

Head and neck cancer is a common malignant neoplasm worldwide.¹ One of the most important factors that influences treatment approaches and therapeutic outcomes for patients with head and neck cancer is the presence of metastatic cervical lymph node. The accurate detection of the cervical lymph node metastasis is thus very important.^91,92 Clinical palpation used to be the method to detect cervical nodal metastasis before the development of imaging technologies. However, studies have shown that both the SEN and the SPE of this technique were unsatisfactory, with a high false positive rate of 25%–51%. The improvements in imaging technologies may make it possible for cervical lymph nodes metastasis in head and neck cancer patients can be effectively diagnosed, especially with CT and MRI.^{11,12,93–96} However, under current health care settings usually only one imaging technique will be performed. Thus a systematic evaluation regarding whether one of the two imaging techniques (CT and MRI) can have a better efficacy than the other will be critical to better guide the clinical practice.

In our systematic review and meta-analysis, we comprehensively evaluated all available evidence from 63 studies for evaluating this question whether one of the two imaging techniques (CT and MRI) can have a better efficacy. Besides pooling results from available studies, we assessed potential sources of heterogeneities via meta-regression and conducted sub-group analyses for significant heterogeneity sources detected. Our meta-analyses suggested that CT had a higher SEN than MRI when node was used as unit of analysis; MRI had a higher SPE when neck level was used as unit of analysis; and MRI had a higher AUC when patient was used as unit of analysis. Our findings showed that CT and MRI are effective tools for detecting the cervical lymph node metastasis in patients with head and neck cancer. Since the diagnostic criteria presented in relevant studies varied significantly, we also summarized available evidence to reveal the most appropriate ones for these two techniques, respectively. Usually, the diagnosis of metastatic cervical lymph nodes consisted of two parts, namely, structural and size changes. The structural changes included central necrosis or cystic degeneration, spherical (rather than flat or bean) shape, or abnormal grouping of nodes (a cluster of three or more lymph nodes of borderline size). In different studies, the description of the structural changes differed only mildly. However, the criteria for sizes differed considerably. Most authors recommended using the minimal axial diameter to assess metastasis. The criterion for minimal axial diameter varied between 5 to 15 mm. Our meta-analysis showed that the minimal axial diameter of 10 mm in lymph node-bearing regions could be considered as the best criterion for assessing cervical lymph node metastasis in patients with head and neck cancer for MRI, compared to 12 mm for CT. Several limitations should be acknowledged for the interpretation of our findings. Firstly, although we conducted meta-regression analyses and showed that the assessed variables largely did not account for heterogeneities between studies, additional undetected variables may account for heterogeneities which warrants further research. Secondly, in some of our analyses, only a very limited number of studies were available. For example, when focusing on the 12 mm size criterion, there was only one study available for evaluating CT with node unit, and future studies for evaluating relevant topics are warranted. In conclusion, through this comprehensive systematic review and meta-analysis, we identified that CT and MRI had acceptable diagnostic efficacy in detecting cervical lymph node metastasis in patients with head and neck cancer. When node was used as unit of analysis, CT had a higher SEN. When neck level was used as unit of analysis, MRI had a higher SPE. Out findings suggest that MRI is superior to CT in the diagnosis of cervical lymph node metastasis, especially in diagnosis confirmation. While CT had a better efficacy in diagnosis exclusion. The diagnostic criteria for MRI and CT for size of metastatic lymph nodes were established. Further high-quality studies are warranted to confirm our findings.

Supplementary materials

Figure S1 Meta-analysis of CT for detecting cervical lymph node metastasis in head and neck cancer patients (patient as unit of analysis).

Abbreviations: CT, computed tomography; CI, confidence interval; LR, likelihood ratio; df, degrees of freedom; SROC, summary receiver operating characteristic; AUC, area under the curve; SE, standard error.

Figure S2 Meta-analysis of MRI for detecting cervical lymph node metastasis in head and neck cancer patients (patient as unit of analysis) (retrospective studies).

Abbreviations: MRI, magnetic resonance imaging; CI, confidence interval; df, degrees of freedom; LR, likelihood ratio; OR, odds ratio.

Figure S3 Meta-analysis of MRI for detecting cervical lymph node metastasis in head and neck cancer patients (patient as unit of analysis) (prospective studies).

Abbreviations: MRI, magnetic resonance imaging; CI, confidence interval; df, degrees of freedom; LR, likelihood ratio; OR, odds ratio.

Figure S4 Meta-analysis of MRI for detecting cervical lymph node metastasis in head and neck cancer patients (patient as unit of analysis).

Abbreviations: MRI, magnetic resonance imaging; CI, confidence interval; df, degrees of freedom; LR, likelihood ratio; OR, odds ratio; SROC, summary receiver operating characteristic; AUC, area under the curve; SE, standard error.

Table S1 Study characteristics of lymph node size per neck level

Study ID	Method	Unit	I	II	III	IV	Retro	Others	TP	FP	FN	TN
Adams et al^Citation1 1998	CT	node	12	12	12	12	12	12	96	175	21	992
Adams et al^Citation1 1998	MRI	node	12	12	12	12	12	12	94	250	23	917
Akoglu et al^Citation2 2005	CT	node	15	15	15	15	15	15	21	2	6	12
Akoglu et al^Citation2 2005	MRI	node	15	15	15	15	15	15	16	1	11	13
Anzai et al^Citation4 1994	MRI	node	10	10	10	10	10	10	38	7	2	34
Braams et al^Citation7 1995	CT	node	11	10	10	10	10	10	5	10	4	13
Braams et al^Citation7 1995	MRI	node	10	11	10	10	10	10	5	6	10	134
Braams et al^Citation7 1995	MRI	node	11	10	10	10	10	10	8	10	14	167
Curtin et al^Citation9 1997	CT	neck level	5	5	5	5	5	5	57	415	1	62
Curtin et al^Citation9 1997	CT	neck level	7	7	7	7	7	7	56	396	2	81
Curtin et al^Citation9 1997	CT	neck level	8	8	8	8	8	8	55	372	3	105
Curtin et al^Citation9 1997	CT	neck level	9	9	9	9	9	9	53	329	5	148
Curtin et al^Citation9 1997	CT	neck level	10	10	10	10	10	10	51	291	7	186
Curtin et al^Citation9 1997	CT	neck level	11	11	11	11	11	11	46	210	12	267
Curtin et al^Citation9 1997	CT	neck level	12	12	12	12	12	12	43	157	15	320
Curtin et al^Citation9 1997	CT	neck level	15	15	15	15	15	15	32	76	26	401
Curtin et al^Citation9 1997	MRI	neck level	5	5	5	5	5	5	53	382	5	95
Curtin et al^Citation9 1997	MRI	neck level	7	7	7	7	7	7	52	367	6	110
Curtin et al^Citation9 1997	MRI	neck level	8	8	8	8	8	8	50	329	8	148
Curtin et al^Citation9 1997	MRI	neck level	9	9	9	9	9	9	48	281	10	196
Curtin et al^Citation9 1997	MRI	neck level	10	10	10	10	10	10	47	248	11	229
Curtin et al^Citation9 1997	MRI	neck level	11	11	11	11	11	11	41	167	17	310
Curtin et al^Citation9 1997	MRI	neck level	12	12	12	12	12	12	38	134	20	343
Curtin et al^Citation9 1997	MRI	neck level	15	15	15	15	15	15	30	67	28	410
Dammann et al^Citation10 2005	CT	neck level	10	10	10	10	10	10	32	17	8	236
Dammann et al^Citation10 2005	MRI	neck level	10	10	10	10	10	10	37	14	3	239
Ding et al^Citation12 2005	MRI	neck level	8	8	8	8	8	8	132	27	34	255
Dirix et al^Citation12 2010	MR-DW	neck level	10	10	10	10	10	10	30	3	2	93
Dirix et al^Citation12 2010	MR-DW	node	10	10	10	10	10	10	40	4	5	149
Dirix et al^Citation12 2010	MR-DW	patient	10	10	10	10	10	10	13	2	0	6
Eida et al^Citation13 2003	CT	node	8	9	6	7			3	5	3	162
Fan et al^Citation14 2006	CT	patient	10	11	10	10	10	10	23	11	4	4
Fukunari et al^Citation15 2010	MR	node	10	10	10	10	10	10	19	13	0	66
Gross et al^Citation16 2001	MR	node	11	10	10	10	10	10	143	22	6	39
Gu et al^Citation17 2000	MRI	node	10	11	10	10	10	10	8	3	1	50
Guenzel et al^Citation18 2013	MR	node	10	10	10	10	10	10	23	26	2	8
Guenzel et al^Citation18 2013	MR	node	15	15	15	15	15	15	20	6	2	28
Guo et al^Citation19 2006	MRI	node	10	10	10	10	10	10	8	3	1	36
Hafidh et al^Citation22 2006	CT	node	10	10	10	10	10	10	8	10	12	2
Hafidh et al^Citation22 2006	MRI	node	10	10	10	10	10	10	11	10	9	2
Hao et al^Citation21 2000	MRI	node	15	15	10	10	10	10	30	2	11	38
Kau et al^Citation26 1999	CT	neck level	15	15	15	15	15	15	6	17	1	17
Kau et al^Citation26 1999	MRI	neck level	15	15	15	15	15	15	2	17	1	15
Kau et al^Citation26 1999	CT	node	15	15	15	15	15	15	13	20	7	18
Kau et al^Citation26 1999	MRI	node	15	15	15	15	15	15	23	22	3	15
Kawai et al^Citation27 2005	MRSPIR	neck level I	5						8	28	0	22
Kawai et al^Citation27 2005	MRSPIR	neck level I	6						8	18	0	32
Kawai et al^Citation27 2005	MRSPIR	neck level I	7						8	10	1	39
Kawai et al^Citation27 2005	MRSPIR	neck level I	8						8	5	1	44
Kawai et al^Citation27 2005	MRSPIR	neck level I	9						8	1	1	48
Kawai et al^Citation27 2005	MRSPIR	neck level I	10						5	0	2	51
Kawai et al^Citation27 2005	MRSTIR	neck level I	5						8	24	0	26
Kawai et al^Citation27 2005	MRSTIR	neck level I	6						8	16	0	34
Kawai et al^Citation27 2005	MRSTIR	neck level I	7						8	7	0	43
Kawai et al^Citation27 2005	MRSTIR	neck level I	8						8	6	0	44
Kawai et al^Citation27 2005	MRSTIR	neck level I	9						8	1	1	48
Kawai et al^Citation27 2005	MRSTIR	neck level I	10						6	0	4	48
Kawai et al^Citation27 2005	MRSPIR	neck level II		5					25	21	0	12
Kawai et al^Citation27 2005	MRSPIR	neck level II		6					25	19	0	14
Kawai et al^Citation27 2005	MRSPIR	neck level II		7					25	16	1	16
Kawai et al^Citation27 2005	MRSPIR	neck level II		8					25	10	2	21
Kawai et al^Citation27 2005	MRSPIR	neck level II		9					25	1	6	26
Kawai et al^Citation27 2005	MRSPIR	neck level II		10					24	0	6	28
Kawai et al^Citation27 2005	MRSTIR	neck level II		5					25	22	0	11
Kawai et al^Citation27 2005	MRSTIR	neck level II		6					25	19	1	13
Kawai et al^Citation27 2005	MRSTIR	neck level II		7					25	19	1	13
Kawai et al^Citation27 2005	MRSTIR	neck level II		8					25	11	1	21
Kawai et al^Citation27 2005	MRSTIR	neck level II		9					25	6	2	25
Kawai et al^Citation27 2005	MRSTIR	neck level II		10					25	4	2	27
Kawai et al^Citation27 2005	MRSPIR	neck level III			5			5	15	7	0	36
Kawai et al^Citation27 2005	MRSPIR	neck level III			6			6	15	4	2	37
Kawai et al^Citation27 2005	MRSPIR	neck level III			7			7	15	2	2	39
Kawai et al^Citation27 2005	MRSPIR	neck level III			8			8	15	2	2	39
Kawai et al^Citation27 2005	MRSPIR	neck level III			9			9	13	0	3	42
Kawai et al^Citation27 2005	MRSPIR	neck level III			10			10	12	0	3	43
Kawai et al^Citation27 2005	MRSTIR	neck level III			5			5	15	10	0	33
Kawai et al^Citation27 2005	MRSTIR	neck level III			6			6	15	8	1	34
Kawai et al^Citation27 2005	MRSTIR	neck level III			7			7	15	3	4	36
Kawai et al^Citation27 2005	MRSTIR	neck level III			8			8	15	2	4	37
Kawai et al^Citation27 2005	MRSTIR	neck level III			9			9	11	0	4	43
Kawai et al^Citation27 2005	MRSTIR	neck level III			10			10	8	0	7	43
Ke et al^Citation28 2006	CT	node	15	10	10	10	10	10	10	3	3	4
Laubenbacher et al^Citation30 1994	MRI	neck level	15	15	15	15	15	15	13	7	5	9
Laubenbacher et al^Citation30 1994	MRI	node	15	15	15	15	15	15	65	126	18	312
Lee et al^Citation31 2013	MR-DW	patient	2	2	2	2	2	2	7	3	1	11
Lee et al^Citation31 2013	MR-TSE	patient	2	2	2	2	2	2	7	6	1	8
Lu et al^Citation32 2007	CT	node	15	10	10	19	10	10	11	1	3	6
Lwin et al^Citation33 2012	MR	patient	10	15	10	10	5	10	63	82	15	24
Mcguirt et al^Citation34 1995	CT	node	15	15	10	10	10	10	18	3	1	19
Nakamoto et al^Citation35 2009	MRI	patient	10	10	10	10	10	10	16	2	4	30
Olmos et al^Citation37 1999	MRI	neck level	10	10	10	10	10	10	22	11	2	27
Paulus et al^Citation39 1998	CT	node	15	15	10	10	10	10	8	1	0	4
Peters et al^Citation41 2013	CT	patient	3	3	3	3	3	3	10	56	0	1
Peters et al^Citation41 2013	CT	patient	4	4	4	4	4	4	8	48	2	9
Peters et al^Citation41 2013	CT	patient	5	5	5	5	5	5	6	29	4	28
Peters et al^Citation41 2013	CT	patient	6	6	6	6	6	6	5	18	5	39
Peters et al^Citation41 2013	CT	patient	7	7	7	7	7	7	5	6	5	51
Peters et al^Citation41 2013	CT	patient	8	8	8	8	8	8	4	5	6	52
Peters et al^Citation41 2013	CT	patient	9	9	9	9	9	9	3	1	7	56
Peters et al^Citation41 2013	CT	patient	10	10	10	10	10	10	3	1	7	56
Ren et al^Citation43 2000	CT	node	5	5	5	5	5	5	36	9	2	11
Schwartz et al^Citation44 2004	CT	node	10	15	10	10	10	10	21	1	6	68
Semedo et al^Citation45 2006	MR	node	10	10	10	10	10	10	24	8	1	30
Seitz et al^Citation46 2009	MR	node	10	10	10	10	5	10	92	6	12	18
Tai et al^Citation53 2002	MRI	patient	11	10	10	10	10	10	3	1	10	2
Van den Brekel et al^Citation56 1991	MRI	neck level	10	10	10	10	10	10	87	13	42	415
Van den Brekel et al^Citation56 1991	MRI	patient	10	10	10	10	10	10	63	6	15	46
Vandecaveye et al^Citation57 2008	MR-TSE	neck level	10	10	10	10	10	10	27	10	20	208
Vandecaveye et al^Citation57 2008	MR-TSE	node	10	10	10	10	10	10	34	10	40	217
Vandecaveye et al^Citation57 2008	MR-TSE	patient	10	10	10	10	10	10	20	5	1	7
Wang et al^Citation58 1999	MRI	node	10	10	10	10	10	10	23	0	15	130
WIDE et al^Citation59 1999	MRI	neck level	10	15	10	10	10	10	18	11	9	34
Wilson et al^Citation60 1994	MRI	neck level	5	5	5	5	5	5	17	16	0	18
Wu et al^Citation61 2010	CT	node	8	8	8	8	8	8	10	1	2	11
Yoon et al^Citation62 2008	CT	neck level	15	15	10	10	10	10	57	2	17	326
Yoon et al^Citation62 2008	MRI	neck level	15	15	10	10		10	57	2	17	326
Yuan et al^Citation63 2000	MRI	neck level	12	12	10	10	10	10	12	1	2	9

Abbreviations: MRI, magnetic resonance imaging; CT, computed tomography; MR-TSE,; MR-DW,; MRSTIR,; MRSPIR,; TP, true positive; FP, false positive; TN, true negative.

Table S2 Meta-analysis results on diagnostic efficacy of MRI on size of metastatic lymph nodes

Unit	Node size (mm)	SEN (95% CI)	SPE (95% CI)	AUC (SE)	Q* (SE)
Level I	10	0.768 (0.725–0.808)	0.901 (0.880–0.919)	0.9159 (0.0348)	0.8487 (0.0394)
	11	0.883	0.866
	12	0.803	0.786
	15	0.774 (0.709–0.830)	0.721 (0.682–0.758)	0.8653 (0.0295)	0.7959 (0.0287)
Level II	10	0.812 (0.778–0.844)	0.883 (0.861–0.902)	0.9151 (0.0341)	0.8477 (0.0385)
	11	0.542	0.953
	12	0.803	0.786
	15	0.774 (0.709–0.830)	0.721 (0.682–0.758)	0.8653 (0.0295)	0.7959 (0.0287)
Level III	10	0.801 (0.767–0.833)	0.894 (0.875–0.911)	0.9121 (0.0314)	0.8444 (0.0350)
	12	0.803	0.786
	15	0.785 (0.712–0.846)	0.704 (0.662–0.742)	0.8385 (0.0274)	0.7705 (0.0253)
Level IV	10	0.801 (0.767–0.833)	0.894 (0.875–0.911)	0.9121 (0.0314)	0.8444 (0.0350)
	12	0.803	0.786
	15	0.785 (0.712–0.846)	0.704 (0.662–0.742)	0.8385 (0.0274)	0.7705 (0.0253)
Retro	5	0.885	0.750
	10	0.780 (0.742–0.814)	0.899 (0.880–0.915)	0.9138 (0.0315)	0.8464 (0.0354)
	12	0.803	0.786
	15	0.785 (0.712–0.846)	0.704 (0.662–0.742)	0.8385 (0.0274)	0.7705 (0.0253)
Others	10	0.801 (0.767–0.833)	0.894 (0.875–0.911)	0.9121 (0.0314)	0.8444 (0.0350)
	12	0.803	0.786
	15	0.785 (0.712–0.846)	0.704 (0.662–0.742)	0.8385 (0.0274)	0.7705 (0.0253)

Abbreviations: MRI, magnetic resonance imaging; SEN, sensitivity; CI, confidence interval; SPE, specificity; AUC, area under the curve; SE, standard error.

Table S3 Meta-analysis results on diagnostic efficacy of CT on size of metastatic lymph nodes

Unit	Node size (mm)	SEN (95% CI)	SPE (95% CI)	AUC (SE)	Q* (SE)
Level I	5	0.947	0.550
	8	0.722 (0.465–0.903)	0.966 (0.928–0.988)
	10	0.617 (0.464–0.755)	0.864 (0.770–0.930)
	11	0.556	0.565
	12	0.821	0.850
	15	0.802 (0.711–0.875)	0.677 (0.573–0.771)	0.8519 (0.0818)	0.7830 (0.0776)
Level II	5	0.947	0.550
	8	0.769	0.917
	9	0.500	0.970
	10	0.607 (0.468–0.735)	0.510 (0.363–0.656)	0.7272 (0.1426)	0.6747 (0.1157)
	11	0.556	0.565
	12	0.821	0.850
	15	0.802 (0.711–0.875)	0.818 (0.746–0.876)	0.9083 (0.0599)	0.8402 (0.0658)
Level III	5	0.947	0.550
	6	0.500	0.970
	8	0.500	0.970
	10	0.746 (0.659–0.820)	0.809 (0.739–0.867)	0.8499 (0.0783)	0.7811 (0.0740)
	12	0.821	0.850
	15	0.723 (0.574–0.844)	0.577 (0.432–0.713)
Level IV	5	0.947	0.550
	7	0.500	0.970
	8	0.500	0.970
	10	0.746 (0.659–0.820)	0.809 (0.739–0.867)	0.8499 (0.0783)	0.7811 (0.0740)
	12	0.821	0.850
	15	0.723 (0.574–0.844)	0.577 (0.432–0.713)
Retro	5	0.947	0.550
	8	0.500	0.970
	10	0.746 (0.659–0.820)	0.809 (0.739–0.867)	0.8499 (0.0783)	0.7811 (0.0740)
	12	0.821	0.850
	15	0.723 (0.574–0.844)	0.577 (0.432–0.713)
Others	5	0.947	0.550
	8	0.500	0.970
	10	0.746 (0.659–0.820)	0.809 (0.739–0.867)	0.8499 (0.0783)	0.7811 (0.0740)
	12	0.821	0.850
	15	0.723 (0.574–0.844)	0.577 (0.432–0.713)

Abbreviations: CT, computed tomography; SEN, sensitivity; CI, confidence interval; SPE, specificity; AUC, area under the curve; SE, standard error.

References

• JemalASiegelRXuJWardECancer statistics, 2010CA Cancer J Clin201160527730020610543
   PubMed Web of Science ®Google Scholar
• FooteRLOlsenKDDavisDLBase of tongue carcinoma: patterns of failure and predictors of recurrence after surgery aloneHead Neck19931543003078360051
   PubMed Web of Science ®Google Scholar
• FerlitoARinaldoARobbinsKTChanging concepts in the surgical management of the cervical node metastasisOral Oncol200339542943512747966
   PubMed Web of Science ®Google Scholar
• TankéréFCamprouxABarryBPrognostic value of lymph node involvement in oral cancers: a study of 137 casesLaryngoscope2000110122061206511129021
   PubMed Web of Science ®Google Scholar
• ShahJCervical lymph node metastases: diagnostic, therapeutic, and prognostic implicationsOncology199041061652149826
   PubMedGoogle Scholar
• GolderWALymph node diagnosis in oncologic imaging: a dilemma still waiting to be solvedOnkologie200427219419915138355
   PubMedGoogle Scholar
• O’BrienCJMcNeilEBMcMahonJDSignificance of clinical stage, extent of surgery, and pathologic findings in metastatic cutaneous squamous carcinoma of the parotid glandHead Neck200224541742212001070
   PubMed Web of Science ®Google Scholar
• KauRJAlexiouCStimmerHArnoldWDiagnostic procedures for detection of lymph node metastases in cancer of the larynxORL J Otorhinolaryngol Relat Spec200062419920310859520
   PubMed Web of Science ®Google Scholar
• CastelijnsJAVan den BrekelMWImaging of lymphadenopathy in the neckEur Radiol200212472773811960218
   PubMed Web of Science ®Google Scholar
• CastelijnsJAVan den BrekelMWDetection of lymph node metastases in the neck: radiologic criteriaAJNR Am J Neuroradiol20012213411158879
   PubMed Web of Science ®Google Scholar
• HaoSPNgSHMagnetic resonance imaging versus clinical palpation in evaluating cervical metastasis from head and neck cancerOtolaryngol Head Neck Surg2000123332432710964315
   PubMed Web of Science ®Google Scholar
• SternWBSilverCEZeiferBAPerskyMSHellerKSComputed tomography of the clinically negative neckHead Neck19901221091132312275
   PubMed Web of Science ®Google Scholar
• KitagawaYNishizawaSSanoKProspective comparison of 18F-FDG PET with conventional imaging modalities (MRI, CT, and 67Ga scintigraphy in assessment of combined intraarterial chemotherapy and radiotherapy for head and neck carcinomaJ Nucl Med200344219820612571209
   PubMed Web of Science ®Google Scholar
• SchöderHCarlsonDLKrausDH18F-FDG PET/CT for detecting nodal metastases in patients with oral cancer staged N0 by clinical examination and CT/MRIJ Nucl Med200647575576216644744
   PubMed Web of Science ®Google Scholar
• LiHChenTWLiZLTumour size of resectable oesophageal squamous cell carcinoma measured with multidetector computed tomography for predicting regional lymph node metastasis and N stageEur Radiol201222112487249322653286
   PubMed Web of Science ®Google Scholar
• RumboldtZGordonLGordonLBonsallRAckermannSImaging in head and neck cancerCurr Treat Options Oncol200671233416343366
   PubMedGoogle Scholar
• ChandawarkarRYKakegawaTFujitaHYamanaHHayabuthiNComparative analysis of imaging modalities in the preoperative assessment of nodal metastasis in esophageal cancerJ Surg Oncol19966132142178637210
   PubMed Web of Science ®Google Scholar
• EscottEJRaoVMKoWDGuitierrezJEComparison of dynamic contrast-enhanced gradient-echo and spin-echo sequences in MRI of head and neck neoplasmsAJNR Am J Neuroradiol1997188141114199296179
   PubMed Web of Science ®Google Scholar
• HamsbergerHRHandbook of Head and Neck ImagingChicagoMosby Year-Book1995283298
   Google Scholar
• LefebvreCManheimerEGlanvilleJSearching for studiesCochrane Handbook for Systematic Reviews of Diagnostic Test Accuracy Version 5.1.0EnglandThe Cochrane Collaboration20116.16.43
   Google Scholar
• WhitingPFRutjesAWWestwoodMEQUADAS-2: A revised tool for the quality assessment of diagnostic accuracy studiesAnn Intern Med2011155852953622007046
   PubMed Web of Science ®Google Scholar
• LauJIoannidisJPSchmidCHQuantitative synthesis in systematic reviewsAnn Intern Med199712798208269382404
   PubMed Web of Science ®Google Scholar
• HigginsJPThompsonSGDeeksJJAltmanDGMeasuring inconsistency in meta-analysesBMJ200332741455756012958120
   PubMed Web of Science ®Google Scholar
• HigginsJThompsonSDeeksJAltmanDStatistical heterogeneity in systematic reviews of clinical trials: a critical appraisal of guidelines and practiceJ Health Serv Res Policy200271516111822262
   PubMedGoogle Scholar
• SchmidaCHStarkaPCBerlinbJALandaisPLauJMeta-regression detected association between heterogeneous treatment effects and study-level, but not patient-level, factorsJ Clin Epidemiol200457768369715358396
   PubMed Web of Science ®Google Scholar
• ZamoraJAbrairaVMurielAKhanKCoomarasamyAMeta-DiSc: a software for meta-analysis of test accuracy dataBMC Med Res Methodol200663116836745
   PubMedGoogle Scholar
• DeeksJJHigginsJPAltmanDGAnalysing data and undertaking meta-analysesCochrane Handbook for Systematic Reviews of Diagnostic Test Accuracy Version 5.1.0EnglandThe Cochrane Collaboration20119.19.43
   Google Scholar
• AdamsSBaumRPStuckensenTBitterKHörGProspective comparison of 18F-FDG PET with conventional imaging modalities (CT, MRI, US) in lymph node staging of head and neck cancerEur J Nucl Med1998259125512609724374
   PubMedGoogle Scholar
• AkoğluEDutipekMBekişRAssessment of cervical lymph node metastasis with different imaging methods in patients with head and neck squamous cell carcinomaJ Otolaryngol200534638439416343398
   PubMedGoogle Scholar
• OkumuraKFujimotoYHasegawaY[Retropharyngeal node metastasis in cancer of the oropharynx and hypopharynx: analysis of retropharyngeal node dissection regarding preoperative radiographic diagnosis]Nihon Jibiinkoka Gakkai Kaiho19981015573577 Japanese9642997
   PubMedGoogle Scholar
• AnzaiYBlackwellKEHirschowitzSLInitial clinical experience with dextran-coated superparamagnetic iron oxide for detection of lymph node metastases in patients with head and neck cancerRadiology199419237097157520182
   PubMed Web of Science ®Google Scholar
• De BondtRBHoeberigsMCNelemansPJDiagnostic accuracy and additional value of diffusion-weighted imaging for discrimination of malignant cervical lymph nodes in head and neck squamous cell carcinomaNeuroradiology200951318319219137282
   PubMed Web of Science ®Google Scholar
• BraamsJWPruimJNikkelsPGNodal spread of squamous cell carcinoma of the oral cavity detected with PET-tyrosine, MRI and CTJ Nucl Med19963768979018683307
   PubMed Web of Science ®Google Scholar
• BraamsJWPruimJFrelingNJDetection of lymph node metastases of squamous-cell cancer of the head and neck with FDG-PET and MRIJ Nucl Med19953622112167830116
   PubMed Web of Science ®Google Scholar
• BruschiniPGiorgettiABruschiniLPositron emission tomography (PET) in the staging of head neck cancer: comparison between PET and CTActa Otorhinolaryngol Ital200323644645315198047
   PubMedGoogle Scholar
• CurtinHDIshwaranHMancusoAAComparison of CT and MRI imaging in staging of neck metastasesRadiology199820711231309530307
   PubMed Web of Science ®Google Scholar
• DammannFHorgerMMueller-BergMRational diagnosis of squamous cell carcinoma of the head and neck region: comparative evaluation of CT, MRI, and 18FDG PETAJNR Am J Roentgenol2005184413261331
   PubMed Web of Science ®Google Scholar
• DingZXLiangBLShenJ[Magnetic resonance imaging diagnosis of cervical lymph node metastasis from lingual squamous cell carcinoma]Ai Zheng2005242199203 Chinese15694033
   PubMedGoogle Scholar
• DirixPVandecaveyeVDe KeyzerFDiffusion-weighted MRI for nodal staging of head and neck squamous cell carcinoma: impact on radiotherapy planningInt J Radiat Oncol Biol Phys201076376176619540069
   PubMed Web of Science ®Google Scholar
• EidaSSumiMYonetsuKKimuraYNakamuraTCombination of helical CT and Doppler sonography in the follow-up of patients with clinical N0 stage neck disease and oral cancerAJNR Am J Neuroradiol200324331231812637273
   PubMed Web of Science ®Google Scholar
• FanWYSunJWEvaluation of enhanced CT on the cervical lymph node metastasis of head and neck neoplasmsChinese Journal of Clinical Healthcare20069236237 Chinese
   Google Scholar
• FukunariFOkamuraKZezeRCervical lymph nodes with or without metastases from oral squamous carcinoma: a correlation of MRI findings and histopathologic architectureOral Surg Oral Med Oral Pathol Oral Radiol Endod2010109689089920299250
   PubMed Web of Science ®Google Scholar
• GrossNDWeissmanJLTalbotJMMRI detection of cervical metastasis from differentiated thyroid carcinomaLaryngoscope200111111 Pt 11905190911801967
   PubMed Web of Science ®Google Scholar
• GuYFQiuWLLuoJCComparison on MRI and CT for Diagnosing Cervical Lymph Node MetastasisJournal of Shang Hai Tie Dao University2000213336 Chinese
   Google Scholar
• GuenzelTFranzenAWiegandSThe value of PET compared to MRI in malignant head and neck tumorsAnticancer Res20133331141114623482793
   PubMed Web of Science ®Google Scholar
• GuoBShuDLRanWA [Clinical Study of Early-stage-diagnosis in Cervical Lymph Node Metastasis of Oral Carcinoma Using MRI]International Medicine and Health Guidance News2002122526 Chinese
   Google Scholar
• HannahAScottAMTochon-DanguyHEvaluation of 18 F-fluorodeoxyglucose positron emission tomography and computed tomography with histopathologic correlation in the initial staging of head and neck cancerAnn Surg2002236220821712170026
   PubMed Web of Science ®Google Scholar
• HaoSPNgSHMagnetic resonance imaging versus clinical palpation in evaluating cervical metastasis from head and neck cancerOtolaryngol Head Neck Surg2000123332432710964315
   PubMed Web of Science ®Google Scholar
• HafidhMALacyPDHughesJPDuffyGTimonCVEvaluation of the impact of addition of PET to CT and MRI scanning in the staging of patients with head and neck carcinomasEur Arch Otorhinolaryngol2006263985385916724209
   PubMed Web of Science ®Google Scholar
• HlawitschkaMNeiseEBredowJFDG-PET in the pretherapeutic evaluation of primary squamous cell carcinoma of the oral cavity and the involvement of cervical lymph nodesMol Imaging Biol200241919814538052
   PubMedGoogle Scholar
• HoffmanHTQuetsJToshiakiTFunctional magnetic resonance imaging using iron oxide particles in characterizing head and neck adenopathyLaryngoscope200011091425143010983936
   PubMed Web of Science ®Google Scholar
• JeongHSBaekCHSonYIUse of integrated 18F-FDG PET/CT to improve the accuracy of initial cervical nodal evaluation in patients with head and neck squamous cell carcinomaHead Neck200729320321017111430
   PubMed Web of Science ®Google Scholar
• KauRJAlexiouCLaubenbacherCLymph node detection of head and neck squamous cell carcinomas by positron emission tomography with Fluorodeoxyglucose F 18 in a routine clinical settingArch Otolaryngol Head Neck Surg1999125121322132810604409
   PubMedGoogle Scholar
• KawaiYSumiMNakamuraTTurbo short tau inversion recovery imaging for metastatic node screening in patients with head and neck cancerAJNR Am J Neuroradiol20062761283128716775280
   PubMed Web of Science ®Google Scholar
• KeZLiuMLiuYDiagnostic value of 18F-FDG PET/CT in the detection of the cervical lymph nodes metastasisLin Chuang Er Bi Yan Hou Ke Za Zhi2006206243245 Chinese16739373
   PubMedGoogle Scholar
• KrabbeCADijkstraPUPruimJFDG PET in oral and oropharyngeal cancer. Value for confirmation of N0 neck and detection of occult metastasesOral Oncol2008441313617306603
   PubMed Web of Science ®Google Scholar
• LaubenbacherCSaumweberDWagner-ManslauCComparison of fluorine-18-fluorodeoxyglucose PET, MRI and endoscopy for staging head and neck squamous-cell carcinomasJ Nucl Med19953610174717577562038
   PubMed Web of Science ®Google Scholar
• LeeMCTsaiHYChuangKSLiuCKChenMKPrediction of nodal metastasis in head and neck cancer using a 3T MRI ADC mapAJNR Am J Neuroradiol201334486486922997167
   PubMed Web of Science ®Google Scholar
• LuHJRibere-BrugelLydiaEmmanuelITThe comparison of PET/CT with contrast enhanced CT in the assessment of cervical lymph nodes in head and neck cancerModern Oncology20071515551557 Chinese
   Google Scholar
• LwinCTHanlonRLoweDAccuracy of MRI in prediction of tumour thickness and nodal stage in oral squamous cell carcinomaOral Oncol201248214915422154129
   PubMed Web of Science ®Google Scholar
• McGuirtWWilliamsDW3rdKeyesJWJrA comparative diagnostic study of head and neck nodal metastases using positron emission tomographyLaryngoscope19951054 Pt 13733757715380
   PubMed Web of Science ®Google Scholar
• NakamotoYTamaiKSagaTClinical value of image fusion from MRI and PET in patients with head and neck cancerMol Imaging Biol2009111465318769975
   PubMed Web of Science ®Google Scholar
• NishimuraHTanigawaNHiramatsuMPreoperative esophageal cancer staging: magnetic resonance imaging of lymph node with ferumoxtran-10, an ultrasmall superparamagnetic iron oxideJ Am Coll Surg2006202460461116571430
   PubMed Web of Science ®Google Scholar
• Valdés OlmosRAKoopsWLoftusBMCorrelative 201Tl SPECT, MRI and ex vivo 201Tl uptake in detecting and characterizing cervical lymphadenopathy in head and neck squamous cell carcinomaJ Nucl Med19994091414141910492358
   PubMed Web of Science ®Google Scholar
• OuYQLinYMagnetic resonance imaging diagnosis of 24 cases of cervical lymph node metastasis from oral carcinomaFujian Med J2936 Chinese
   Google Scholar
• PaulusPSambonAVivegnisD18FDG-PET for the assessment of primary head and neck tumors: clinical, computed tomography, and histopathological correlation in 38 patientsLaryngoscope199810810157815839778305
   PubMed Web of Science ®Google Scholar
• PerroneAGuerrisiPIzzoLDiffusion-weighted MRI in cervical lymph nodes: differentiation between benign and malignant lesionsEur J Radiol201177228128619716671
   PubMed Web of Science ®Google Scholar
• PetersTTCastelijnsJALjumanovicRDiagnostic value of CT and MRI in the detection of paratracheal lymph node metastasisOral Oncol201248545045522226512
   PubMed Web of Science ®Google Scholar
• PoharSBrownRNewmanNWhat does PET imaging add to conventional staging of head and neck cancer patients?Int J Radiat Oncol Biol Phys200768238338717379443
   PubMed Web of Science ®Google Scholar
• RenKZhangJRMaSSCT-Pathologic Correlative Study on the Cervical Lymph Node Metastasis of Laryngeal CancerChinese J Med Imaging20008347351 Chinese
   Google Scholar
• SchwartzDLFordERajendranJFDG-PET/CT imaging for preradiotherapy staging of head-and-neck squamous cell carcinomaInt J Radiat Oncol Biol Phys200561112913615629603
   PubMed Web of Science ®Google Scholar
• Curvo-SemedoLDinizMMiguéisJUSPIO-enhanced magnetic resonance imaging for nodal staging in patients with head and neck cancerJ Magn Reson Imaging200624112313116755542
   PubMed Web of Science ®Google Scholar
• SeitzOChambron-PinhoNMiddendorpM18F-Fluorodeoxyglucose-PET/CT to evaluate tumor, nodal disease, and gross tumor volume of oropharyngeal and oral cavity cancer: comparison with MRI imaging and validation with surgical specimenNeuroradiology2009511067768619727695
   PubMed Web of Science ®Google Scholar
• StokkelMPten BroekFWHordijkGJKooleRvan RijkPPPreoperative evaluation of patients with primary head and neck cancer using dual-head 18fluorodeoxyglucose positron emission tomographyAnn Surg2000231222923410674615
   PubMed Web of Science ®Google Scholar
• StuckensenTKovácsAFAdamsSBaumRPStaging of the neck in patients with oral cavity squamous cell carcinomas: a prospective comparison of PET, ultrasound, CT and MRIJ Craniomaxillofac Surg200028631932411465137
   PubMed Web of Science ®Google Scholar
• SumiMKimuraYSumiTNakamuraTDiagnostic performance of MRI relative to CT for metastatic nodes of head and neck squamous cell carcinomasJ Magn Reson Imaging20072661626163317968896
   PubMed Web of Science ®Google Scholar
• SumiMVan CauterenMNakamuraTMRI microimaging of benign and malignant nodes in the neckAJR Am J Roentgenol2006186374975716498102
   PubMed Web of Science ®Google Scholar
• SumiMSakihamaNSumiTDiscrimination of metastatic cervical lymph nodes with diffusion-weighted MRI imaging in patients with head and neck cancerAJNR Am J Neuroradiol20032481627163413679283
   PubMed Web of Science ®Google Scholar
• SunJTZhangZXZhangWJDiagnosis of molecular imaging on head and neck carcinoma and cercical lymph node metastasisChinese Journal of Coal Industry Medicine20131610491052 Chinese
   Google Scholar
• TaiCJShiauYCTsaiMHDetection of cervical lymph node metastases in nasopharyngeal carcinomas: comparison between technetium-99m methoxyisobutylisonitrile single photon emission computed tomography and magnetic resonance imagingNeoplasma200249425125412382024
   PubMed Web of Science ®Google Scholar
• TakashimaSSoneSTakayamaFPapillary thyroid carcinoma: MRI diagnosis of lymph node metastasisAJNR Am J Neuroradiol19981935095139541309
   PubMed Web of Science ®Google Scholar
• TuliHSSinghBPrasadVDiagnostic accuracy of 99mTc-MIBI-SPECT in the detection of lymph node metastases in patients with carcinoma of the tongue: comparison with computed tomography and MRINucl Med Commun200829980380818677208
   PubMed Web of Science ®Google Scholar
• Van den BrekelMWCastelijnsJACrollGAMagnetic resonance imaging vs palpation of cervical lymph node metastasisArch Otolaryngol Head Neck Surg199111766636732036191
   PubMedGoogle Scholar
• VandecaveyeVDe KeyzerFVander PoortenVHead and neck squamous cell carcinoma: value of diffusion-weighted MRI imaging for nodal stagingRadiology2009251113414619251938
   PubMed Web of Science ®Google Scholar
• WangQTakashimaSFukudaHDetection of medullary thyroid carcinoma and regional lymph node metastases by magnetic resonance imagingArch Otolaryngol Head Neck Surg1999125884284810448729
   PubMedGoogle Scholar
• WideJMWhiteDWWoolgarJAMagnetic resonance imaging in the assessment of cervical nodal metastasis in oral squamous cell carcinomaClin Radiol1999542909410050735
   PubMed Web of Science ®Google Scholar
• WilsonGRMcLeanNRChippindaleAThe role of MRI scanning in the diagnosis of cervical lymphadenopathyBr J Plast Surg19944731751798193855
   PubMedGoogle Scholar
• WuYQFanXCDengYValue of Spiral CT Scan on Cervical Lymph node Metastasis of Laryngo and Hypolaryngo carcinomaHei Long Jiang Medical Journal34761763 Chinese
   Google Scholar
• YoonDYHwangHSChangSKCT, MRI, US, 18F-FDG PET/CT, and their combined use for the assessment of cervical lymph node metastases in squamous cell carcinoma of the head and neckEur Radiol200919363464218843493
   PubMed Web of Science ®Google Scholar
• YuanYGHanDMFanEZThe evaluation of cervical lymph node metastasis of laryngeal cancer using magnetic resonance imaging (MRI)Lin Chuang Er Bi Yan Hou Ke Za Zhi20001410449451 Chinese12563714
   PubMedGoogle Scholar
• ShumJWDierksEJEvaluation and Staging of the Neck in Patients with Malignant DiseaseOral Maxillofac Surg Clin North Am201426220922124656885
   PubMed Web of Science ®Google Scholar
• SunFLiYFLiuJHXiongYImpact of postoperative adjuvant therapy on prognosis of low-risk cervical cancer: analysis of 208 casesNan Fang Yi Ke Da Xue Xue Bao201434401405 Chinese24670458
   PubMedGoogle Scholar
• LindbergRDistribution of cervical lymph node metastases from squamous cell carcinoma of the upper respiratory and digestive tractsCancer1972296144614495031238
   PubMed Web of Science ®Google Scholar
• BoccaECalearoCde VincentiisIOccult metastases in cancer of the larynx and their relationship to clinical and histological aspects of the primary tumor: a four year multicentric researchLaryngoscope1984948108610906748835
   PubMed Web of Science ®Google Scholar
• FriedmanMRobertsNKirshenbaumGColomboJNodal size of metastatic squamous cell carcinoma of the neckLaryngoscope199310388548568361287
   PubMed Web of Science ®Google Scholar
• JohnsonJTA surgeon looks at cervical lymph nodesRadiology199017536076102188292
   PubMed Web of Science ®Google Scholar

Disclosure

The first and corresponding authors had full access to all of the data in the study and had final responsibility for the decision to submit for publication. The authors have no conflicts of interest in this work.