Palpation of the neck for metastases, by itself, may be associated with significant error rates. Therefore, CT is used extensively to image both the primary site and neck, and may increase the accuracy of neck staging from 70% to 93% . Computed tomography is especially useful in detecting occult nodal disease, determining the presence or absence of extranodal disease, demonstrating the resectability of extensive nodal disease, and evaluating inaccessible areas, such as the parapharyngeal space, retropharyngeal space, and tracheoesophageal groove (Figure 4).
Although MRI can detect nodal disease as well , we currently limit its use to selected primary sites, such as the nasopharynx, or to those cases in which a specific issue, such as perineural infiltration, needs to be addressed. Ultrasound is used more extensively in Europe than in the United States to evaluate nodal disease. Magnetic resonance imaging and ultrasound may be helpful in determining nodal fixation to the carotid artery, but no imaging method is definitive in this regard.
The choice between irradiation or dissection for treatment of the neck is based on the treatment modality planned for the primary site and the clinical status of the neck. More advanced neck disease is usually managed with combined therapy.
Management of the Clinically Negative Neck
Traditionally, elective neck treatment has been utilized to manage tumors with a high risk (> 20% to 30%) of occult metastatic disease because of the improved neck control associated with this strategy . The alternative is to adopt a "wait and see" approach, which may be associated with an increased rate of distant metastases and a worse prognosis, even if the neck failure is successfully salvaged [8,9]. However, it has been recently suggested that elective neck treatment may no longer be justified, since the use of CT and MRI to assess the neck should decrease the risk of occult disease to approximately 12% .
Because the morbidity of a modified radical neck dissection or elective neck irradiation is minimal, follow-up is simplified, and the cost of delivering the treatment in one instance is lower than that of an additional treatment period for a neck failure, we continue to advocate elective neck treatment in patients with a high risk of occult cervical metastases. An interesting use of decision analysis tools by Weiss et al further supported the elective management of the neck with either dissection or irradiation for those tumors at high risk for occult metastases .
Radiation and surgery can achieve similar regional control rates for N0 neck (Table 1) disease , and it is desirable to use only one therapeutic modality when possible to save the other for either recurrence or a second primary tumor. Therefore, when the primary tumor is to be treated with surgery, an elective neck dissection is indicated for primary sites with a greater than 20% risk of cervical metastasis. If, on the other hand, the primary site is to be treated with radiation, elective neck irradiation is indicated when the risk of cervical metastasis exceeds 20%.
Management of the Clinically Positive Neck
If the primary tumor is to be managed surgically, dissection is indicated for clinically positive necks. Preoperative irradiation is rarely used, except when a node is fixed to the carotid artery or deep cervical fascia. The node will usually become resectable, and in many instances, no viable tumor is found in the neck dissection pathologic specimen. When the primary site is to be irradiated, radiation alone is typically adequate for small N1 (1 to 2 cm) disease, especially for nasopharyngeal tumors. More advanced neck disease, particularly when regression is incomplete, requires a neck dissection 4 to 6 weeks after the completion of radiation therapy in order to improve regional control .
The terminology for describing neck dissections has become quite confusing since the initial description of the "classic" radical neck dissection in 1906 . An understanding of the classification of neck dissections is essential for selecting the appropriate type of neck dissection and for comparing treatment results. The description of neck dissections that follows is useful for such purposes .
Radical neck dissection entails the removal of lymph node levels I to V in continuity with the spinal accessory nerve, sternocleidomastoid muscle, and internal jugular vein (Figure 5). A modified radical neck dissection involves the removal of the same lymph node groups with preservation of one or more of the nonlymphatic structures (Figure 6). A selective neck dissection describes a cervical lymphadenectomy, in which one or more of the lymph node groups is not removed. Supraomohyoid and lateral neck dissections are selective neck dissections that remove nodal levels I to III and II to IV, respectively (Figures 7 and 8). Finally, an extended neck dissection includes the removal of lymph node groups or nonlymphatic structures not encompassed by a radical neck dissection.
The choice of a neck dissection is based on the primary site as well as the number, size, and location of positive lymph nodes. In addition, the results and morbidity associated with each type of neck dissection must be considered.
Radical vs Modified Radical Dissection
We prefer a modified radical neck dissection over a radical neck dissection in any situation in which equal regional control rates may be obtained. The removal of the spinal accessory nerve, sternocleidomastoid muscle, and internal jugular vein in a radical neck dissection is associated with trapezius muscle dysfunction, alteration of the neck contour, and venous engorgement. Furthermore, bilateral modified neck dissection may be performed simultaneously rather than in a staged fashion, as with a radical neck dissection.
Multiple authors advocate a modified radical neck dissection for N1 disease [14-16]. Since no lymph nodes are contained within the fascial wrappings of the sternocleidomastoid muscle or carotid sheath, all lymph nodes may be removed with a modified neck dissection, unless extensive extranodal spread and/or fixation are present. These are, in fact, our only indications for a radical neck dissection, as suggested by Bocca and Pignaturo .
Selective Neck Dissection
The advantage of a selective neck dissection is that it provides the same therapeutic benefit as a comprehensive neck dissection, with potentially less morbidity and decreased operative time. Cases must be carefully evaluated to ensure that they are appropriate for a selective dissection, however.
Use of selective neck dissection is based on the concept that metastases from a specific primary site occur in a predictable fashion. It is indicated for N0 or N1 neck disease associated with the primary sites described below. In the setting of more advanced neck disease or after previous treatment, lymphatic drainage is no longer predictable, and a comprehensive neck dissection is indicated.
The most commonly employed selective neck dissection is the supraomohyoid dissection for clinically negative necks associated with oral cavity primaries. Lateral neck dissection is best suited for laryngeal, oropharyngeal, and hypopharyngeal primaries . If significant unexpected lymph node metastases are encountered during a selective neck dissection, the dissection should be converted to a comprehensive neck dissection due to the unpredictable pattern of metastases in that event.
Extended Neck Dissection
Extended neck dissection is indicated for removal of the following specific nodal groups associated with selected primary sites: (1) anterior compartment nodes (level VI) with thyroid gland, esophageal, and subglottic laryngeal primaries; (2) facial nodes for lip and mid-face skin primaries; (3) periparotid and preauricular nodes for skin lesions of the temple, cheek, and anterior auricle; and (4) retroauricular and suboccipital nodes for posterior scalp and posterior auricular lesions.
A variation of an extended neck dissection is the inclusion of a retropharyngeal node dissection for pharyngeal tumors that have a particularly high incidence of nodal metastases to that area. The survival benefit of this dissection, however, has not been proven [19,20].