This article discusses the rationale for imaging cervical lymph nodes and reviews nodal anatomy and common drainage patterns, imaging features of pathologic lymph nodes, and the advantages of various imaging modalities available for evaluation and diagnosis of the lymph nodes.
Key Points
- •
In many cases of cervical lymphadenopathy, the clinical history, physical examination, and routine laboratory investigations can establish a presumptive diagnosis of reactively enlarged nodes related to viral or bacterial infection. Imaging these patients is generally unnecessary, unless atypical symptoms or findings concerning for aggressive infection are noted.
- •
Imaging, particularly computed tomography (CT), is valuable in cases concerning for suppurative lymphadenitis or deep neck infection by identifying abscesses requiring surgical drainage.
- •
Imaging findings such as stranding of adjacent fat on CT may help distinguish suppurative nodes from metastatic disease; however, common imaging findings such as nodal enlargement, enhancement, and hypodensity seen in the setting of lymphadenitis caused by typical or atypical pathogens can also be seen with nodal metastasis.
- •
In asymptomatic patients and patients with nonspecific symptoms, imaging may be indicated to assist in evaluating a palpable neck mass.
Introduction
Cervical lymphadenopathy is a common clinical finding, which can be related to reactive nodal hypertrophy, granulomatous processes, lymphoma and head and neck cancers or mimicked by nonnodal neck masses. Cross-sectional imaging has the ability to quickly assess the extent of cervical lymphadenopathy with greater accuracy than the physical examination and provide information regarding the relationship with adjacent vital structures. This information can have significant management implications in cases of acute infection, but is perhaps most important when applied to staging of head and neck malignancies. Head and neck squamous cell carcinomas (HNSCC) are the most common malignancy of the upper aerodigestive tract, and lymphatic spread is the most important mechanism of metastasis in these patients. Accurately staging nodal disease in HNSCC has both important prognostic and management implications. This article discusses the rationale for imaging cervical lymph nodes and reviews nodal anatomy and common drainage patters, imaging features of pathologic lymph nodes, and the advantages of various imaging modalities available.
Rationale for imaging
In many cases of cervical lymphadenopathy, the clinical history, physical examination, and routine laboratory investigations can establish a presumptive diagnosis of reactively enlarged nodes related to viral or bacterial infection. Imaging these patients is generally unnecessary, unless atypical symptoms or findings concerning for aggressive infection are noted. Imaging, particularly computed tomography (CT), has proved valuable in cases concerning for suppurative lymphadenitis or deep neck infection by identifying abscesses requiring surgical drainage. Imaging findings such as stranding of adjacent fat on CT may help distinguish suppurative nodes from metastatic disease; however, common imaging findings such as nodal enlargement, enhancement, and hypodensity seen in the setting of lymphadenitis caused by typical or atypical (eg, mycobacterial or bacillary) pathogens can also be seen with nodal metastasis ( Fig. 1 ).
In asymptomatic patients and patients with nonspecific symptoms, imaging may be indicated to assist in evaluating a palpable neck mass. In the case of a solitary palpable neck mass, imaging with ultrasonography (US), CT, or magnetic resonance imaging (MRI) may identify congenital cysts such as thyroglossal and branchial cleft cysts. Imaging findings in many of these patients are nonspecific, and recent studies have questioned the ability of US to make a confident preoperative diagnosis. However, imaging in these types of patients should be able to differentiate between nodal disease and nonnodal neoplasms in most cases. Nonnodal findings on imaging can also affect the differential diagnosis. For example, multiple cystic lesions within the parotid glands seen in combination with cervical lymphadenopathy can suggest the diagnosis of human immunodeficiency virus infection.
The prognostic value of accurate staging in the case of head and neck malignancies has produced a well-defined role for the use of imaging. Squamous cell carcinomas (SCC), which make up 80% to 90% of all upper aerodigestive tract malignancies, have shown a high predilection for metastasis to regional lymph nodes, with nodal metastases present in 20% to 30% of patients at the time of diagnosis depending on the site. The presence or absence of lymph node metastasis in these patients at the time of diagnosis has a profound impact on 5-year disease-specific survival rates. In 1 series evaluating HNSCC from multiple sites, the 5-year survival was reduced from 67.9% to 39.9% based solely on the presence of nodal metastases. Nodal stage in this series also affected survival, with advanced N2 or N3 patients, according to the American Joint Committee on Cancer (AJCC) staging system, showing significantly worse outcomes. Staging examinations by palpation alone have shown limited sensitivity of approximately 64%, specificity of 85%, and overall accuracy of approximately 75%. Also, clinical examination has proved inaccurate at evaluating size of cervical nodes regardless of clinical experience. The ability of imaging studies to evaluate nodal groups occult to palpation (such as retropharyngeal nodes) and provide accurate nodal measurements can increase sensitivity and specificity up to 81% and 96%, respectively, with imaging identifying palpably occult nodal metastases in up to 27% of patients.
Patients who are deemed N0 by clinical staging typically undergo elective treatment of cervical lymph nodes such as neck dissection if there is a 20% risk of occult metastases. Imaging studies may increase sensitivity and specificity over palpation; however, identification of micrometastases (metastatic deposits less than 3 mm in size) remains a significant challenge for imaging. Up to 25% of neck dissection specimens may harbor only micrometastases and use of highly sensitive molecular pathology techniques can show evidence for micrometastases in even more cases than traditional light microscopy. Treatment of these N0 patients remains controversial, and some investigators have suggested that the rate of occult metastases in patients undergoing thorough staging examinations including CT or MRI is low (less than 10%), making observation a viable option with equivalent outcomes to elective neck dissection. Other investigators have advocated reducing the number of staging evaluations used in patients with small primary tumors ; however, they continue to advocate the use of cross-sectional imaging as an accurate and expedient way to evaluate both the primary tumor and potential nodal metastases.
Nodal classification
The complex anatomy of cervical lymph nodes has led to numerous attempts at organization. The most widely used schemes historically have been those, such as that of Rouvière, describing nodal groups based on proximity to adjacent structures. This type of organization resulted in the familiar large groupings of cervical nodes such as the jugular or spinal accessory chains, which were then subdivided into superficial or deep, upper or lower, and medial or lateral subgroups. With the development of more advanced surgical approaches to head and neck cancer, including techniques such as selective neck dissections, a less cumbersome scheme was needed that could accurately reflect patterns of lymphatic drainage and standardize communication between clinicians. Building on the work of earlier investigators such as Lindberg and Shah and colleagues, the American Head and Neck Society (AHNS) and American Academy of Otolaryngology–Head and Neck Surgery developed a 6-level system of nodal classification. Some of the surgical boundaries initially defined proved difficult to define by imaging; therefore, surrogate landmarks were identified and an imaging-based classification of cervical lymph nodes was developed. The most recent AHNS consensus statement on nodal classification has been updated to reflect this system and the role of imaging in evaluating cervical lymph nodes in the setting of head and neck cancer. The result has been a widely accepted division of cervical lymph nodes into 7 numbered nodal levels with well-defined anatomic boundaries that can be easily identified by both surgeons and radiologists ( Fig. 2 ).
On cross-sectional imaging, level I lymph nodes are found inferior to the myelohyoid muscle, anterior to a line drawn at the posterior margin of the submandibular glands and within the boundaries of the mandible. This region can be subdivided into levels IA and IB by the anterior bellies of the digastric muscles, with level IA lymph nodes medial and level IB nodes lateral to the anterior belly of the digastric muscle.
Level II lymph nodes are found posterior to the line drawn at the posterior margin of the submandibular glands, anterior to the posterior margin of the sternocleidomastoid muscle and between the lower margin of the hyoid bone and the skull base ( Fig. 3 ). Level II nodes can also be subdivided into levels IIA and IIB, with level IIA nodes along the internal jugular vein and level IIB nodes posterior to the internal jugular vein and separated by a fat plane.
Level III lymph nodes are found lateral to the carotid arteries, anterior to the posterior margin of the sternocleidomastoid muscle and between the lower margin of the hyoid bone and lower margin of the cricoid cartilage (see Fig. 3 ).
Level IV lymph nodes are found lateral to the carotid arteries, anterior to the line connecting the posterior margins of the sternocleidomastoid and anterior scalene muscles and between the lower margin of the cricoid cartilage and the clavicle (see Fig. 3 ).
Level V lymph nodes are found posterior to the posterior margin of the sternocleidomastoid muscle from the skull base to the clavicle. Level V nodes can be subdivided in to levels VA and VB, with level VA nodes located between the skull base and lower margin of the cricoid cartilage and level VB nodes located between the lower margin of the cricoid and the clavicle anterior to the anterior margin of the trapezius muscle (see Fig. 3 ).
Levels VI and VII are outside the traditional boundaries of the radical neck dissection and, although they have been defined in the literature, may be less widely applied in clinical practice. Level VI lymph nodes encompass those found in the anterior central or visceral compartment medial to the carotid arteries and between the lower margin of the hyoid bone and the upper margin of the manubrium. Level VII lymph nodes encompass those found in a paratracheal location medial to the carotid arteries within the superior mediastinum between the upper margin of the manubrium and the innominate artery (see Fig. 3 ).
Some nodal sites, such as the nodal groups of the superior pericervical region including intraparotid, buccinator, suboccipital, and retroauricular lymph nodes, lie outside these defined levels and retain their traditional anatomic descriptors. Nodes found medial to the carotid arteries superior to the hyoid bone are referred to as retropharyngeal nodes (Rouvière), a similar example of an anatomically named nodal distribution.
Patterns of nodal drainage
The 7-level nodal classification system was designed to reflect common patterns of lymphatic drainage, and most malignancies of the head and neck have a predictable pattern of metastasis along these routes. Although the detailed anatomy of the lymphatic system of the head and neck and upper aerodigestive tract is beyond the scope of this review, recent works have mapped lymphatic drainage patterns and these anatomic pathways have been found to mirror clinical experience in patterns of metastatic disease in a variety of head and neck malignancies. The presence of predictable drainage patterns, particularly in the case of HNSCC, have led to the clinical practice of selective neck dissection involving nodal groups at highest risk rather than more comprehensive, and morbid, radical neck dissection. Therefore, a basic understanding of these patterns is essential when evaluating cervical lymph nodes in the setting of malignancy either clinically or by imaging. However, these patterns are not immutable, and skip metastasis to more distal nodal groups without involvement of usual proximal drainage pathways can occur in up to 15% of cases.
Oral Cavity Lymphatic Drainage
Lymphatic drainage of the oral cavity may extend from level I to level IV, depending on the location of the primary tumor. Anterior regions including the anterior floor of mouth, anterior mandibular gingival, and medial lower lip drain preferentially to level IA nodes, with the upper lip, lateral lower lip, buccal mucosal, and contiguous mucosa of the anterior palate draining preferentially to level IB nodes. The posterior portions of the oral cavity drain preferentially to level II.
Tongue Lymphatic Drainage
Drainage of the tongue is complex, with the ventral tongue following drainage patterns of the floor of mouth to level IA anteromedially and IB laterally ( Fig. 4 ). The dorsal oral tongue may drain either to level I nodes medially or level II nodes laterally. The tongue base shows drainage patterns more akin to pharyngeal drainage, primarily involving levels II and III ( Fig. 5 ). Level IV nodes may be involved in advanced disease, or via alternative drainage pathways. Midline lesions may have bilateral drainage patterns.
Nasopharynx Lymphatic Drainage
The nasopharynx is richly supplied with lymphatics and shows one of the highest rates of nodal metastasis when involved with SCC. The nasopharynx also serves as the catchment basin for lymphatic drainage of most of the nasal cavity and paranasal sinuses. The predominant lymphatic drainage pattern for these regions is to lateral retropharyngeal nodes and level II nodes, with drainage to more distal level III and IV nodal groups in more advanced disease and through alternative drainage patterns ( Fig. 6 ). Unlike other areas of the pharyngeal mucosa, nasopharyngeal lesions may drain posteriorly to level V. Midline lesions of the nasopharynx may result in bilateral drainage patterns.
Oropharyngeal Lymphatic Drainage
Oropharyngeal structures, including the soft palate, palatine and lingual tonsils, and posterior pharyngeal wall show a similar drainage pattern to that previously described for the tongue base. Drainage is primarily to level II and III nodal groups, with less common involvement of retropharyngeal and level IV nodes.
Hypopharynx Lymphatic Drainage
The hypopharynx is surpassed only by the nasopharynx in density of lymphatic drainage, with particularly dense confluence of lymphatic drainage in the piriform sinuses. Drainage of the hypopharynx is similar to that of the oropharynx, with predominant involvement of levels II to IV and occasional drainage to retropharyngeal nodes from the posterior hypopharyngeal wall. Anterior hypopharyngeal lymphatic drainage channels may directly communicate with laryngeal lymphatics, with eventual involvement of level VI nodes.
Laryngeal Lymphatic Drainage
Laryngeal drainage patterns can be divided into supraglottic, glottic, and subglottic patterns. Supraglottic structures, including the epiglottis and aryepiglottic folds, drain primarily to level III, with occasional drainage cephalad to level II. Prominent horizontal drainage patterns lead to an increased risk for bilateral metastasis. Glottic structures such as the true vocal fold and vocal ligaments have poor lymphatic drainage, making nodal metastasis from small glottic lesions less likely. More advanced glottic lesions, with involvement of the muscles of phonation, may have involvement of nodal levels II to IV and VI ( Fig. 7 ). Subglottic processes have primarily lateral drainage to levels III and IV.
Paratracheal Lymphatic Drainage
Processes involving the cervical esophagus, thyroid and parathyroid glands, and some subglottic locations show primarily paratracheal drainage to levels VI and VII. Other less common processes, including malignant melanoma and other cutaneous malignancies, have less predictable patterns of metastases. Melanoma may initially involve superficial nodes such as postauricular or suboccipital nodal groups, with additional involvement of posterior level V nodes depending on the location of the primary lesion.
Imaging characteristics of pathologic lymph nodes
Size
Size criteria are widely used in evaluating lymph nodes on cross-sectional imaging and are integral to the AJCC nodal staging system. Criteria for nodal enlargement have been studied by multiple groups advocating different techniques and cutoff values, with variability in the resulting sensitivity and specificity for detecting malignancy. Nodes may be measured in either greatest diameter (long axis) or minimal diameter (short axis). Nodes measuring greater than 1.5 cm in greatest diameter in levels I or II and greater than 1.0 cm in all other levels are considered abnormal and harbor metastatic disease in approximately 80% of cases. Other investigators claim minimal diameter measurements are more accurate and reproducible on both CT and US, with diameters exceeding 11 mm in levels I and II or 10 mm elsewhere considered abnormal. Nodes meeting these minimal diameter criteria harbor metastatic disease in approximately 75% of cases. Retropharyngeal nodes have been considered separately by some investigators, with those exceeding 6 to 8 mm in greatest diameter concerning for metastasis, particularly in the setting of nasopharyngeal tumors. Some investigators have suggested alternative measurement criteria for sonography, considering minimum diameter measurements greater than 6 to 7 mm in levels I and II or 5 mm in levels III and IV abnormal.
Shape
Size criteria alone may misinterpret reactive nodes as malignant and overlook small metastases. Therefore, additional morphologic criteria have been proposed. Normal lymph nodes tend to be flat or kidney-bean–shaped, with a fat-containing hilum. Malignant nodes typically show a rounded shape, with loss of the normal fatty hilum or focal cortical expansion ( Fig. 8 ). It has been suggested that a maximal diameter/minimal diameter ratio greater than 2 favors benign nodes and a ratio less than 2 favors malignancy. Nodal grouping can also help identify pathologic nodes. Grouping refers to 3 or more nodes in continuity with each other, each measuring at least 8 to 10 mm in diameter. Although not individually enlarged by these criteria, groupings of nodes such as this along the expected drainage pathway for a known malignancy are concerning and increase sensitivity for metastases.
