CERVICAL METASTATIC DISEASE AND THE “UNKNOWN” PRIMARY
KEY POINTS
- Imaging is critical to planning the treatment for cervical metastatic disease.
- The use of imaging to detect the risk of micrometastases based on size criteria is overrated compared to historical risk data.
- No imaging study can eliminate the risk of micrometastases compared to the historical risk profile data in combination with data about the primary tumor in a way that justifies altering traditional decision making based on imaging findings alone.
- The full extent of gross metastatic disease can be known.
- Imaging can find otherwise undetectable head and neck primary cancers.
- Surveillance imaging can modify treatment plans.
INTRODUCTION
Etiology
The vast majority cervical node metastases are from squamous cell carcinoma (SCCA), predominantly from head and neck cancers of mucosal and skin origin. Some cervical metastatic disease, especially when confined to the low neck and supraclavicular area, is from primary sites of origin below the clavicles or due to lymphoma. This discussion is primarily one of the natural history of neck node metastases due to SCCA, but the other patterns of “solid tumor” malignant adenopathy will be considered. This model of that natural history also explains the behavior of metastases from most other tumor types and origins elsewhere in the head and neck. Where natural history varies with different histology and specific primary sites, it is discussed in conjunction with those particular tumor types and sites of origin and in more general terms in Chapters 21 through 24. Primary lymphoreticular processes and their tendencies to show related adenopathy are also discussed as pathologic entities in Chapters 26, 27, 28, and 159. This should be contrasted with the pathophysiologic tendencies of reactive and infectious adenopathy presented in Chapters 13 and 158.
Primary site management in SCCA and other head and neck carcinomas is closely linked to management of cervical lymph node metastases. Failure to control a metastatic tumor in the neck should be an unusual outcome unless the neck disease is very advanced; if the primary is controlled; and surgery, radiation, and chemotherapy have been used optimally. The clear exceptions to a good outcome are advanced neck disease with extranodal spread and fixation to the floor of the neck and/or carotid artery encasement. The specific plan for management of the neck will vary with the primary site treatment plan; therefore, the origin of the neck disease, as well as its extent, becomes pivotal in clinical decision making.
Prevalence and Epidemiology
Factors Modifying the Risk of Lymph Node Metastasis
The likelihood of lymph node metastasis is generally increased in more poorly differentiated tumors, ones that penetrate more deeply at the primary site, and by the capillary lymphatic density at the primary site.1,2 Vascular space invasion at the primary site predicts a high rate of nodal metastasis. Lymphatic spread also increases with recurrent lesions.
All histologic malignancies can show lymphatic spread. The factors just mentioned are probably more important than the specific tumor type with regard to risk of regional disease. For example, adenocarcinoma of the maxillary sinus is less likely to metastasize to nodes than the same histology arising in the parotid gland, assuming the same degree of differentiation because of differences in capillary lymphatic density at these two sites.
The risk of subclinical disease in the lymph nodes in a patient with a clinically negative neck has been understood for many years.3,4 A sample of this classic experience is presented in Figure 157.1. These incidences are discussed in detail in chapters on specific primary sites. In general, the risk of regional disease increases with the T stage. Lymph node involvement to “first echelon” nodes and then contiguous groups is usually predictable based on the primary site.3,5 Skips occur due to variations in capillary lymphatic anatomy. Substantial rerouting of lymphatic spread is caused by prior surgery and/or radiotherapy (RT).6
Midline primary lesions or lateralized primaries that grow to the midline may spread to both sides of the neck. Lateralized lesions will generally spread to ipsilateral nodes unless the primary site is one known for crossed lymphatic drainage. A large nodal mass may obstruct lymphatic vessels and reroute flow. Obstruction of the lymphatic pathways caused by surgery and/or radiation therapy may also shunt the lymphatic flow across the midline mainly through anastomoses in the submental area, as demonstrated in a brilliant and seminal work by Ugo Fisch.6
If a well-lateralized cancer metastasizes to contralateral nodes, level 2 is most commonly involved. A rare contralateral skip to level 3 is possible mainly in tongue cancers due in part to normal crossed drainage of capillary lymphatics.5 Whenever lymph node metastases appear in an unexpected distribution, a second primary must be suspected.
Extracapsular Spread and Other Factors Important to Diagnosis, Prognosis, and Clinical Decision Making
Snow et al.2 reviewed 326 radical neck dissection specimens. Extracapsular spread was documented in 23% of nodes <1 cm in diameter, 53% of nodes 2 to 3 cm, and 74% of nodes >3 cm in that work (Fig. 157.2). Such extracapsular spread has been associated with an approximately 50% decrease in survival in the past. This and other work1,2 has suggested that the number of involved lymph nodes do not correlate with the risk of extracapsular spread.
Lower neck regional disease is likely associated with an increased incidence of distant metastasis. Extracapsular spread, many positive nodes, and low neck disease in aggregate might, therefore, constitute a profile that justifies the use of fluorine-18 2-fluoro-2-deoxy-D-glucose positron emission tomography (FDG-PET) to look for distant disease.
Clinical Presentation
General
Metastatic disease most frequently is discovered as a neck mass, either as the presenting sign of a cancer or, more commonly, by physical examination and/or imaging studies done to evaluate a known head and neck region primary cancer. If the origin of the neck mass is uncertain after initial imaging, fine needle aspiration (FNA) biopsy of the palpable mass frequently is the next step to confirm that it is a nodal mass due to cancer.
The neck examination can be a very reliable source of clinical evaluation if learned correctly. However, even a well-trained examiner sometimes has difficulty identifying or excluding disease in the neck. This is especially true in the posttreatment neck. Imaging is particularly useful in a patient with a short, thick, and/or obese neck and the posttreatment neck. Once a metastatic node is suspected or confirmed, imaging is indispensable for assessing nodes not accessible on physical examination, such as retropharyngeal nodes, highest level 2 (retrostyloid space) nodes, and the deeper level 6 nodes (Fig. 157.3).
Normal structures that may be confused with a positive level 2 lymph node include transverse process of C1 or C2; the tail of the parotid gland; and a prominent, calcified carotid bifurcation. In level 1, it may be impossible to distinguish abnormal nodes from an abnormal submandibular gland. These normal sources of potential confusion are typically sorted out with imaging.
The Unknown Primary
Primary SCCA of the pharynx often presents with metastatic cancer in neck lymph nodes (Fig. 157.4). The primary is easily identified by physical examination in about 97% to 98% of patients.7 About 2% to 5% of the time, the primary cannot be found even by complete endoscopy by an experienced examiner.
Spontaneous remission of the primary may occur since some patients treated solely by neck dissection are cured without the primary site of origin established even by long-term follow-up. Spontaneous regression of other primary tumors has been reported. The concept of a cancer arising in branchial cleft remnants of the lateral neck should be abandoned even though it may rarely be the explanation (Figs. 153.18 and 157.4). Assuming a branchial origin can impede the search for a primary and lead to inadequate treatment and follow-up.
PATHOPHYSIOLOGY
Anatomy
Capillary Lymphatics
Capillary lymphatic distribution is the major explanation for the patterns and incidence of metastases to regional lymph nodes in head and neck cancer that will be presented in conjunction with those primary sites. The lymphatic metastatic process begins when cancer cells penetrate the lamina propria of the involved epithelial surface to reach the deeper capillary lymphatics.5 There are a few capillary lymphatics in the periosteum and perichondrium and none in bone and cartilage. There are no capillary lymphatics in the globe and few in the orbit. Muscle and fat also have few capillary lymphatics. This explains the very low rate of lymphatic node metastases from tumors arising in these structures and tissues.
Once tumors have penetrated deep enough in a capillary lymphatic–rich area, the risk of lymph node metastases will be related to the depth of cancer invasion.8,9 This trend has been proven and is used in clinical decision making for malignant melanoma of the skin and SCCA of the oral tongue. The nasopharynx and pyriform sinus have the most capillary lymphatic density.5 The paranasal sinuses, middle ear, and vocal cord have few, if any, capillary lymphatic channels.5 This capillary density distribution correlates very well with the high rates of nodal metastases in the pharynx as well as with the near zero rates of lymph node metastases in tumors confined to the paranasal sinuses and true vocal cord.
Lymphatic Trunks
Lymphatic collecting trunks form from the capillary lymphatics. Some of the lymphatic vessels have valves, and some do not. This sometimes has grave prognostic significance. In the deeper layers of the skin, the valves direct lymphatic flow in a predictable manner. However, superficial skin lymphatics are valveless, thus a cancer spread may be helter-skelter, resulting in unpredictable remote spread and satellite lesions (Fig. 157.5).
Lymphatic vessels draining the mucous membranes have valves, making their drainage patterns predictable in the previously untreated patient based on the known anatomic range of variation in the vessels.5 The lymphatic trunks empty into a large vein, near the junction of the jugular and subclavian veins at the surgical venous angle (Fig. 149.11). Lymphatic vessels may drain directly into the internal jugular or subclavian veins. There are also direct lymphatic to venous connections between lymph vessels and veins within the lymph node. The lymphatic and venous connections in nodes likely open up mainly after occlusion or obstruction of a lymphatic trunk. The direction of lymphatic flow with known common variations and less common aberrant pathways are well established.5 Variations that affect treatment decisions will be discussed for specific primary sites. Obstruction of lymphatic flow can be caused by tumor, postradiation fibrosis, surgery, or other traumatic interruption.6 The more complete the obstruction, the greater the change in lymph flow and resulting change in the expected pathway and pattern of lymph node metastasis. Postradiation alterations in redirected lymph flow are to some extent RT dose dependent (Figs. 157.5 and 157.6).
Lymph Nodes
About 150 to 350 lymph nodes lie above the clavicles.5 This is about one third of the lymph nodes in the body. This text will use the current standardized numbering systems for the cervical lymph node groups (Fig. 149.10). This system, agreed upon by surgeons, radiation oncologists, and diagnostic radiologists, has eliminated the previously confusing and less precise naming and renaming of these groups. Other important named groups include the retropharyngeal, facial, parotid, mastoid, suboccipital, and posterior neck nodes.
Inconstant, 1- to 2-mm lymph nodules and aggregates called intercalated nodes may lie along the course of a lymph vessel between its origin and its entrance to a larger node. Intercalated nodules and nodes can mimic satellite tumor nodules when involved with metastatic lymphatic spread.
Nodal Groups
One must be thoroughly familiar with the anatomic boundaries of the cervical lymph nodes as currently classified in levels 1 through 6 as well as the retropharyngeal, parotid, facial, lingual, mastoid, and posterior neck groups in order to properly assess metastatic disease from head and neck and other primary sites. The detailed anatomy, including anatomic variants, is discussed in conjunction with the normal anatomy of the neck (Chapter 149). The primary drainage fields of these nodes must also be understood. These include the following:
Level 1A: The lip, skin of the chin, more anterior buccal mucosa, gingiva, floor of the mouth, and tongue drain to these nodes. The efferent vessels drain to levels 1B and 2A.
Level 1B: The lips, floor of the mouth, buccal mucosa, gingiva, nasal vestibule, skin of the anterior face, oral tongue, palate, and submandibular and sublingual salivary glands drain to these nodes. The efferent vessels drain mainly to level 2A.
Level 2: The entire pharynx and larynx and other neck viscera drain to this group of nodes. Level 2 receives connections from levels 1 and 5; thus, they are a final common pathway for much lymphatic drainage.
Levels 3 and 4: The visceral compartment, mainly the larynx, hypopharynx, thyroid, and cervical esophagus, drain to these nodes.
Level 5: These nodes drain predominantly to the pharynx and larynx. However, they have important connections to parotid nodes and skin lymphatics, which make them very important in the treatment of skin cancers.
Level 6: These nodes drain their adjacent viscera and are contiguous with mediastinal lymphatics.
Retropharyngeal nodes drain the nasal cavity, nasopharynx, pharyngeal walls, pyriform sinus, hard and soft palate, middle ear, and eustachian tube.
Parotid nodes drain the skin of the frontal scalp, temple, nose, eyelids, pinna, external auditory canal, and malar area. They also drain the eustachian tube.
Mastoid or retroauricular nodes, occipital nodes, and posterior neck nodes drain the scalp and posterior periauricular soft tissues of the ear.
Pathophysiology and Patterns of Disease
The basic pathophysiology and related pathologic anatomic correlates of cervical lymph node metastatic disease as seen on imaging is due to tumor emboli that are transported by lymphatic channels to become trapped in the subcapsular sinus of a lymph node where the deposits may grow. This focal metastasis can then itself further embolize to a contiguous lymph node. The nodes are prepared in advance by the primary tumor to be areas both receptive to the deposits and are induced to prepare an environment that is favorable to metastatic tumor growth.
The deposits grow—at first possibly mainly peripherally because the afferent capillary lymphatics enter the capsule of the node—within the lymph node, and the metastatic deposits gradually replace its normal architecture (Fig. 157.7). Alternatively, the SCCA cells, especially if they lose the adhesive tendencies more prevalent in the better differentiated epidermoid cancers, might percolate through the lymphatic sinusoids, producing broad zones of altered architecture up to totally replacing the normal nodal architecture (Fig. 157.7J).
Eventually, areas of necrosis develop. The node may become completely replaced by tumor and become largely necrotic and cystic. This gross necrosis is often a late finding and the one most emphasized in the imaging literature as a criterion for metastases. With modern imaging, the metastatic deposits can be recognized much earlier in their natural history as smaller focal defects often as peripheral focal defects in the node (Fig. 157.7). Tumor cells within a node can also enter the venous circulation, possibly leading to pulmonary metastases since the lung capillary bed is a first-order filter of the cervical venous drainage.
The enlarging focal metastasis will continue this local growth and eventually invade the capsule of the lymph node and spread to perinodal tissue (Fig. 157.8). Extracapsular spread may be seen both in lymph nodes only focally involved as well as nodes nearly totally replaced by tumor (Figs. 157.7J and 157.8). Continued extracapsular growth may invade and become fixed to and/or encase surrounding structures.
The risk of cervical lymph node metastases is in large part dependent on the density of the capillary lymphatics at any given primary site. Tumors with less invasive margins are less likely to develop lymphatic metastases. For instance, verrucous carcinoma has indolent tumor margins and almost never invades the lymphatic system. The depth of invasion of the primary cancer also correlates with the risk of lymph node metastasis in mucosal cancer, as it does with malignant melanoma. For instance, in oral tongue cancers, the risk of lymph node metastasis increases rapidly with primary depth of invasion of 2 to 4 mm.8–11 Sarcomas usually originate in tissues with few or no lymphatics, such as muscle, bone, or cartilage, and generally have relatively low rates of lymph node involvement. These and other cancers normally at low risk for nodal metastases may spread by way of the lymphatic system once they involve areas with higher capillary lymphatic density.
Imaging of the Normal Lymph Node Size and Morphology: The Baseline
A node’s shape on axial sections depends on its orientation to the transverse plane of the body. This should be taken into account, as largest short axis dimensions are typically used even though they are relatively useless as a criterion for metastatic disease (Figs. 149.7–149.10).
Variations due to intra- and perinodal fatty foci and nodal folding should not be mistaken for focal defects produced by metastases (Figs. 149.7 and 149.8).
Computed tomography (CT) and magnetic resonance imaging (MRI) performed with and without intravenous contrast injection show normal nodes as homogeneous internally. Parenchymal changes seen within nodes on contrast-enhanced magnetic resonance (CEMR) and contrast-enhanced computed tomography (CECT) are very important in recognizing early metastases in nodes. Such changes can be seen with CECT in nodes with maximum short axis measurements as small as 3 to 5 mm (Fig. 157.7). Contrast is also required on CT to distinguish nodes from vessels. At ultrasound examination, normal nodes have a definable interface with surrounding tissues and homogeneous texture (Fig. 4.6). The texture may be interrupted by hilar structures including small nodal vessels; these normal structures can also be seen on good-quality CECT and CEMR studies. Blood flow patterns in the nodal vessels can be studied with color flow Doppler techniques (Figs. 4.1, 4.6, and 4.7 and Chapter 4). Resistive indices can also be calculated. Both resistive indices and flow patterns differ in normal and pathologic nodes. These gross flow features are a reflection of abnormal morphology seen on CT, magnetic resonance (MR), and ultrasound and are also related to perfusion data available on CT and MR studies.
The nodes of the head and neck vary in size as seen on axial CT and MR, normally from 1 to 15 mm in maximum short axis dimension12 (Table 149.2). With ultrasound, and on coronal and sagittal images, the long axis is seen and nodes are often 15 to 20 mm or more (especially in children) in length. Size criteria are very limited with regard to predicting the risk of lymph node micrometastases in otherwise normal nodes, thus considerably limiting the value of size criteria in clinical decision making to almost zero.
Imaging Criteria of Metastatic Disease in Lymph Nodes
For anatomic imaging studies, the diagnosis of metastatic SCCA in lymph nodes depends on size and the presence of focal parenchymal defects, including focal areas of low density or signal intensity, focal enhancement and foci of impacted keratin debris, and/or dystrophic calcification (Figs. 157.3–157.7). Irregularity of node margins, rim enhancement, and more obvious extranodal infiltration of bordering soft tissues help to identify tumor penetrating the node capsule (Figs. 157.2, 157.4, 157.5, 157.8, and 157.9). No criterion is specific for metastatic cancer. Lack of specificity is usually unimportant since the patients have known cancer. The lack of specificity is more of a problem in patients with a neck mass of uncertain etiology.