To avoid complications such as wound breakdown, skin flap necrosis, and exposure of the carotid artery after neck dissection, the incisions must be placed correctly.

The skin of the anterior lateral aspect of the neck is supplied by descending branches of the facial, submental, and occipital arteries and by ascending branches of the transverse cervical and suprascapular artery (Fig. 117.1). These arterial branches anastomose, forming a superficial network of vessels that runs predominantly in a vertical direction. Although some or all the main arteries or their perforating branches may be ligated or divided during a neck dissection, the superficial, predominantly vertical, vascular plexus must remain intact to ensure adequate blood supply to the skin flaps. Thus, the incisions in the neck that are more likely to safeguard the blood supply to the skin flaps are the single transverse incision in the mid portion of the neck, the superiorly based apron-like incision from mastoid to mentum for combined neck dissection with intraoral procedures (Fig. 117.2A), and the apron-like incision used when a neck dissection is performed in conjunction with a laryngectomy (Fig. 117.2B).

The Y incision and the double-Y incision jeopardize the blood supply to the inferior and middle skin flaps, respectively, and suffer from placing a trifurcate incision over the carotid artery (Fig. 117.2C and D). The modification of the Schobinger incision creates a long anterior medial flap, the tip of which may necrose as a result of the limited ascending blood supply (Fig. 117.2E). The MacFee double transverse incision transects the ascending and descending blood supply to the central portion of the flap (Fig. 117.2F). This flap, however, usually fares well even in the previously irradiated patient.

Located in the anterolateral aspect of the neck, the platysma is a wide, quadrangular, sheet-like muscle that extends obliquely from the upper chest to the lower face. This muscle, a remnant of the panniculus carnosus, is located immediately deep to the subcutaneous tissue and thus provides an easily identifiable plane to raise skin flaps during neck surgery. In most neck dissections, flaps are elevated by dissecting in a plane immediately deep to the platysma; however, when the extent of the disease is such that the platysma must be left attached to the specimen, flaps can be elevated easily in a plane immediately superficial to this muscle. The beginning head and neck surgeon must remember that, because of its oblique direction, the platysma does not cover an inferiorly based triangle in the anterior aspect of the neck and most of the posterolateral aspect of the neck. Here, flaps must be elevated in a subcutaneous plane created by the surgeon. While making the incisions for a neck dissection and elevating the skin flaps, it is helpful to remember that the posterior border of the platysma is either over or slightly anterior to the external jugular vein and the greater auricular nerve.

Identifying the marginal mandibular branch is essential to perform an adequate excision of the lymph nodes in the submandibular triangle. The practice of ligating the anterior facial vein low in the submandibular triangle and
retracting it superiorly to “protect the marginal branch” can also result in elevation of the prevascular and retrovascular lymph nodes, thus precluding their appropriate removal. When indicated, it is preferable to identify the nerve and thoroughly remove these lymph nodes.

The nerve can be identified about 1 cm in front of and below the angle of the mandible by incising the superficial layer of the deep cervical fascia that envelops the submandibular gland, immediately above the gland, in a direction parallel to the direction of the nerve. The incised fascia then is gently pushed superiorly, exposing the nerve that lies deep to it but superficial to the adventitia of the anterior facial vein. The submandibular retrovascular lymph nodes are usually near the nerve and must be carefully dissected away from it. As this is done, the facial vessels are exposed and can be divided.

Below the jugular foramen, the external branch of the spinal accessory nerve is located medial to the digastric and stylohyoid muscles and lateral or immediately posterior to the internal jugular vein (IJV). Occasionally, the uppermost portion of the nerve is posteriomedial to the vein. From here, the nerve runs obliquely downward and backward to reach the medial surface of the sternocleidomastoid muscle (SCM) near the junction of its superior and middle thirds (two to three finger breadths below the tip of the mastoid). Although the nerve can continue its downward course entirely medial to the muscle (18%), more commonly, it traverses and appears in the posterior border of it (82%) (1). Here, the nerve is always located above the point where the greater auricular nerve turns around the posterior border of the SCM, also known as Erb point (2). The mean distance between Erb point and the spinal accessory nerve is 10.7 mm, SD ± 6.3. It then runs through the posterior triangle of the neck and crosses the anterior border of the trapezius muscle. The mean distance between this point and the clavicle is 51.3 mm, SD ± 17 (2). Two anatomic characteristics of this portion of the nerve are relevant to avoid injuring it in the course of a neck dissection. First, the spinal accessory nerve is located rather superficially as it courses through the middle and low posterior triangle of the neck, and it can be easily injured while elevating the
posterior skin flaps. Second, the nerve does not enter the trapezius muscle at the anterior border of it but courses along the deep surface of the muscle in close relationship with the transverse cervical vessels. Therefore, isolating the nerve to the level of the anterior border of the trapezius does not ensure its preservation during surgical dissection below this point, particularly in a bloody operative field.

The levator scapulae is a triangular muscle located deep in the lateral aspect of the neck, anterior and medial to the splenius capitis muscle. It extends from the transverse process of the atlas and the next three cervical vertebrae to the superior angle and the spine of the scapula. The action of levator scapulae is to raise the medial angle of the scapula and incline the neck to the corresponding side with rotation of the neck in the same direction. With the trapezius muscle, the levator scapulae makes a shrug possible.

The nerves to the levator scapulae, which vary in number from 1 to 3, branch off the fourth and fifth cervical nerves and travel posteriorly and inferiorly. They cross the anterior border of the levator scapulae and remain on the surface of the muscle for a short distance. The nerves to levator scapulae are under the fascia of this muscle; thus, in the course of any neck dissection, but especially in a radical neck dissection (RND) or a modified radical neck dissection (MRND), it is crucial to keep the plane of dissection superficial to the fascia of the levator in order to preserve these nerves. The dorsal scapular nerve is inconsistent in its anatomic relations in the posterior triangle of the neck and contributes to the innervation of the levator scapulae in a minority of cases (3). Because one of the functions of the levator is to draw the scapula and the shoulder upward and medially, inadvertent or unnecessary resection of the nerves to it during an RND may add to the resulting deformity and functional disability of the shoulder.

At the base of the neck, the thoracic duct is located medial to and behind the left common carotid artery and the
vagus nerve. From here, it arches upward, forward, and laterally, passing behind the IJV and in front of the anterior scalene muscle and the phrenic nerve. It then opens into the IJV, the subclavian vein, or the angle formed by the junction of these two vessels. The duct is anterior and medial to the thyrocervical trunk and the transverse cervical artery. Precise knowledge of these anatomic relationships is important to avoid injuring the duct during a neck dissection. It is even more important when the surgeon is called on to search for and repair a chyle leak during or after a neck dissection. To prevent a chyle leak, the surgeon also must remember that the thoracic duct may be multiple in its upper end and that at the base of the neck it usually receives a jugular, a subclavian, and usually other minor lymphatic trunks, which must be ligated or clipped individually.

The deep cervical fascia of the neck is divided into three layers: superficial, middle, and deep (Fig. 117.3). The superficial or investing layer surrounds the entire neck. It arises from the vertebral spinous processes and the ligamentum nuchae and encircles the entire neck to attach itself again to the spinous processes on the opposite side. This fascia divides to enclose the trapezius muscle. At the anterior border of this muscle, the two layers fuse into a single layer that crosses the posterior triangle of the neck. It divides again to surround the inferior belly of the omohyoid muscle and the sternomastoid muscle. At the lateral border of the strap muscles, it sends fibers between them before fusing in front of them as it extends onto the other side of the neck. This fascia also envelops the submandibular and parotid glands.

The middle layer of the deep cervical fascia, also called the visceral fascia, surrounds the visceral structures of the anterior portion of the neck. The deep layer of the deep cervical fascia or prevertebral fascia surrounds the deep muscles of the neck. Among them, it covers the splenius capitis, the levator scapulae, and the scalene muscles. It extends onto the other side of the neck, covering the prevertebral muscles.

The carotid sheath encircling the jugular vein, common carotid artery, and vagus nerve is formed by all layers of the deep cervical fascia. The carotid sheath originates superiorly at the jugular foramen, where it attaches to the skull base. It then follows the course of the vessels, traversing the anterior cervical triangle and extending inferiorly into the thoracic inlet.

The lymph node regions of the neck are shown in Figure 117.4. The six levels currently used encompass the complete topographic anatomy of the neck. The concept of sublevels has been introduced into the classification since certain zones have been identified within the six levels, which may have clinical significance.

Level I is divided in two sublevels: sublevel IA (submental), which includes the lymph nodes within the triangle bound by the anterior bellies of the digastric muscles and the mylohoid muscle, and sublevel IIB (submandibular), which includes the lymph nodes with the boundaries of the triangle formed by the anterior and posterior bellies of the digastrics and body of the mandible.

Level II (upper jugular) includes the lymph nodes located around the upper third of the IJV and adjacent spinal accessory nerve extending from the level of the
skull base to the level of the inferior border of the hyoid. The anterior/medial boundary is the stylohyoid muscle (the radiologic correlate is the vertical plane defined by the posterior surface of the submandibular gland), and the posterior (lateral) boundary is the posterior border of the SCM. Two sublevels are recognized in level II: sublevel IIA contains nodes in the region anterior to the spinal accessory nerve (SAN) and sublevel IIB contains nodes posterior to the SAN.

Level III (midjugular) includes the lymph nodes located around the middle third of the IJV extending from the inferior border of the hyoid bone to the inferior border of the cricoid cartilage. The anterior/medial boundary is the lateral border of the sternohyoid muscle, and the posterior/lateral boundary is the posterior border of the SCM.

Level IV (lower jugular) encompasses the lymph nodes located around the lower third of the IJV extending from the inferior border of the cricoid cartilage to the clavicle.

The anatomic boundary that separates the medial border of levels III and IV from the lateral border of level VI has traditionally been the lateral border of the sternohyoid muscle, a landmark that cannot be easily discerned on imaging studies. Therefore, the medial aspect of the common carotid artery has been suggested as an alternate boundary to separate these levels in an axial plane in imaging studies (4).

Level V (posterior triangle) comprises predominantly of the lymph nodes located along the lower half of the spinal accessory nerve and the transverse cervical artery. The supraclavicular nodes are also included in posterior triangle group. The superior boundary is the apex formed by convergence of the sternocleidomastoid and trapezius muscles, the inferior boundary is the clavicle, the anterior boundary is the posterior border of the SCM, and the posterior boundary is the anterior border of the trapezius muscle. A horizontal plane marking the inferior border of the anterior cricoid arch separates two sublevels. Sublevel VA, above this plane, includes the spinal accessory nodes. Sublevel VB, below this plane, includes the nodes that follow the transverse cervical vessels and the supraclavicular nodes with the exception of Virchow node, which is located in level IV.

Level VI (anterior compartment): Lymph nodes in this compartment include the pre- and paratracheal nodes, precricoid (delphian) node, and the perithyroidal nodes including the lymph nodes along the recurrent laryngeal nerves. The superior boundary is the hyoid bone, the inferior boundary is the suprasternal notch, and the lateral boundaries are the common carotid arteries.

Other lymph node groups: Lymph nodes involving regions not located within these levels should be referred to by the name of their specific nodal group; examples of these are the superior mediastinum, the retropharyngeal, the periparotid, the buccinator, the postauricular, and the suboccipital lymph nodes.

Lindberg first described the clinical pattern of cervical metastases for various primary sites in 1972 (5). Shah (6), in 1990, examined the distribution of nodal metastases in 1,119 RNDs performed both electively and therapeutically for squamous cell carcinoma (SCCA) of the oral cavity, oropharynx, hypopharynx, and larynx. He found a consistent distribution of cervical metastases (Table 117.1). The majority of metastases from cancer of the oral cavity were found in lymph nodes in levels I, II, and III. On the other hand, most metastases from carcinomas of the oropharynx, hypopharynx, and larynx were found in levels II, III, and IV. The lymph nodes in level V were not involved in the absence of metastases at other levels.

The significance of other lymphatic groups, such as the retropharyngeal lymph nodes (RPLNs) and paratracheal nodes, has also been characterized (7). Metastases to the RPLNs can occur with SCCA of the hypopharynx, tonsil, soft palate, posterior and lateral oropharynx, nasopharynx, and supraglottis.

Metastases from cancers of the larynx can occur to the paratracheal lymph nodes (PTLNs). In a study of 91 selected patients with carcinoma of the larynx who underwent PTLN dissection, Weber et al. (8) found that metastases to these lymph nodes occurred primarily in patients with subglottic and transglottic tumors, but they also occurred in patients with glottic tumors. In a prospective study of 45 patients with advanced glottic cancers, who underwent bilateral paratracheal node dissection as part of their treatment, Shenoy et al. (9) found metastases in the ipsilateral paratracheal nodes in 4% of the cases and contralateral metastases in 2%. Failure to address the paratracheal nodes for laryngeal cancer is a known cause of stomal recurrence.

Using the histologic demonstration of metastasis in a lymph node as the gold standard, several studies have shown that ultrasonography (US), computed tomography (CT), and magnetic resonance imaging (MRI) have a
higher sensitivity and specificity than clinical examination in the detection of metastases in lymph nodes. In a prospective study of 48 patients who were to undergo neck dissection, Haberal et al. (10) found that the sensitivity and specificity of palpation for the detection of metastases in the lymph nodes were 64% and 85%, respectively, while the corresponding values were 72% and 96% for US and 81% and 96% for CT. Adams et al. (11) reported similar results for US and CT; in addition, they reported a sensitivity of 80% and a specificity of 79% for MRI. What these observations indicate is that a “negative” US, CT, or MRI of the neck cannot be relied to withhold elective treatment of the cervical lymph nodes, since they cannot detect metastases in 19% to 28% of the patients staged clinically N0. Obviously the imaging criteria used to diagnose metastasis in a lymph node are not absolutely reliable. The imaging criterion used most often to consider a cervical node suspicious for metastasis is size larger than 1.5 cm in maximum diameter for nodes located in levels I and II and larger than 1 cm for nodes in other regions of the neck (12). Although a correlation exists between the size of a lymph node and the presence of histologic metastasis (Table 117.2), it is clear that not all enlarged lymph nodes contain metastatic deposits and that nodes smaller than 1 or 1.5 cm can indeed contain metastases. Thirty-three percent of all metastases from SCCAs of the head and neck are found in lymph nodes smaller than 1 cm, 10% of tumor-positive neck dissection specimens contain only metastases less than 3 mm in diameter, and, more importantly, 25% of all clinically occult lymph node metastases are too small to be detected by any of the currently available imaging techniques (13). DiNardo (14) demonstrated that 88% of metastases from the carcinomas of the floor of the mouth were found in nodes that were 1 cm or less in diameter.

The presence of a central area of lucency within a node shown on CT was at one time considered equivalent to the presence of necrotic tumor within a node; however, such a finding can be caused by an artery with plaque formation or a fatty inclusion in a lymph node. Another imaging criterion used to “diagnose” metastasis in a lymph node is the shape of the node, as determined by the ratio of its long (l) and short (s) axis diameters. In a study by Steinkamp et al. (15), 730 enlarged cervical lymph nodes in 285 patients were examined using ultrasound, and the l/s ratio was calculated. Histologic examination after neck dissection revealed that 95% of the enlarged cervical nodes, shown on ultrasound to have a l/s ratio of more than 2, were correctly diagnosed as benign. Nodes presenting with a more circular shape and a l/s ratio of less than 2 were diagnosed correctly as metastases with 95% accuracy.

In an attempt to overcome the lack of sensitivity of morphologic imaging criteria, US was combined with fine needle aspiration biopsy (US-FNAB). This technique appeared more promising for the preoperative evaluation of the N0 neck as it enables sampling of lymph nodes as small as 3 mm in diameter and adds the advantages of cytologic evaluation (16). However, the usefulness of this technique is strongly dependent on the skill and time of the ultrasonographer and on the experience of the cytopathologist. Furthermore, the outcomes of a wait-and-see policy after negative US-FNAB have been disappointing. In a study of 92 patients with tumors of the oral cavity, staged T1 and T2, who were observed after a negative US-fine-needle aspiration (US-FNA), metastases in neck nodes became apparent subsequently in 19 (21%) (17). In a more recent study, Wensing et al. (18) found that palpation and US with or without US-FNAB missed occult lymph node metastases in 22% of the patients with oral cavity SCCA. These figures are troubling because the incidence of lymph node metastases in patients with such tumors, who are observed without any intervention to the neck, is about 25%.

Prospective studies using 18-fluorodeoxyglucose (FDG) positron emission tomography (PET) to assess lymph node metastases from SCCAs of the oral cavity have shown a sensitivity and specificity higher than MRI, CT, and US. However, current FDG-PET techniques are also limited in the detection of tumor foci smaller than 1 cm (19,20). Kyzas et al. (21) performed a meta-analysis of 32 studies that assessed the diagnostic performance of PET scans in patients with SCCA of the head and neck. In patients staged clinically N0, the sensitivity of 18 PET scan was only 50% (95% CI: 37% to 63%), whereas specificity was 87% (95% CI: 76% to 93%). These authors also compared the performance of PET scan with that of “conventional diagnostic methods,” that is, CT, MRI, and ultrasound with fine-needle aspiration (FNA) by analyzing studies that had also used these diagnostic methods on the same patients. The sensitivity and specificity of PET scans were 80% and 86%, respectively, and of “conventional diagnostic tests” were 75% and 79%. Thus, at the present time, the role for PET scan in the evaluation of the N0 neck is limited as it will not detect metastases in 20% to 50% of the cases.

In another effort to overcome the limitations of current imaging techniques, particularly their inability to differentiate a tumor-infiltrated lymph node from a normal or a reactive one, de Bree et al. (22) have experimented with radioimmunoscintigraphy (RIS). With this technique, SCCA specific monoclonal antibodies labeled with ⁹⁹mTc were given intravenously to patients with SCCA who underwent neck dissection. Unfortunately, RIS was not superior to CT and MRI for the detection of lymph node metastases.

CT and MRI are valuable in the evaluation of lymph nodes that are not easily accessible to the hands of the examiner, such as the retropharyngeal, upper mediastinal, and, in some patients, the PTLNs. They are also valuable in the assessment of resectability of large metastatic deposits in the neck, because in most instances they can define the relationship of a metastatic tumor with critical structures, such as the common and the internal carotid artery, the cervical spine, the vertebral artery, and the brachial plexus. If tumor involvement of the common or the internal carotid artery is suspected, a systematic preoperative evaluation should include four-vessel cerebral angiography to determine the status of the contralateral carotid and to assess collateral intracerebral circulation. In addition, an attempt should be made during the angiography to measure carotid back pressure and to assess dynamically the collateral circulation by using balloon occlusion techniques while monitoring the patient for evidence of neurologic deficits under normotensive and hypotensive conditions.

FNA has become a valuable diagnostic tool in patients with a mass in the neck in whom metastatic carcinoma is suspected. According to numerous recent reports, the specificity of FNA ranges from 94% to 100% and the sensitivity ranges from 92% to 98%. The interobserver variability between cytopathologists in one series has been reported as 8%. FNA has been found to be most accurate in the diagnosis of epithelial malignancies, achieving nearly 100% accuracy.

FNA is indicated in the patient with a solid mass in the neck after a thorough examination of the mucosal surfaces and skin of the head and neck region fails to reveal a primary tumor. If the mass is located in the supraclavicular area, FNA is essential in the initial evaluation of the patient. The cytopathology findings then can guide the clinician’s search for a primary tumor below the clavicles.

Sentinel lymph node biopsy (SLNB) is feasible and useful as a staging procedure in patients with early carcinomas of the oral cavity and in particular for patients with cancers of the oral tongue. Proponents of this technique point out that it allows accurate histopathologic staging of the neck by examining the sentinel lymph node (SLN) with serial sectioning and immunohistochemistry, and it avoids unnecessary neck dissection and its possible complications.

The SLNB is based on the belief that cancers metastasize via lymphatics to regional lymph nodes in an orderly fashion, that this process is embolic in nature, and that the lymph node that first receives lymphatic drainage from the primary site can be identified and excised for histologic analysis (23).

Early studies in patients with SCCAs of the mucosal surfaces of the head and neck investigated the methodology and feasibility of SLNB (23,24,25,26). In general, the primary cancer should be accessible to infiltration with a radioisotope-labeled colloid to perform a lymphoscintigraphy and a blue dye to aid in intraoperative localization of the sentinel node. These localizing methods have been shown to be complementary. Ross et al. (26) in a multicenter study investigated SLNB in 134 patients with SCCA of the oral cavity and oropharynx, staged T1/T2 N0. Lymphoscintigraphy was performed preoperatively; blue dye and a gamma probe were used intraoperatively to aid in the identification of sentinel nodes. Sentinel nodes were identified in 93% of the cases. The number of sentinel nodes varies, but in a series of 48 patients studied by Ross et al. (24), the mean number of sentinel nodes harvested was 2.4.

Subsequent studies have examined the utility of SLNB in patients with oral cavity or oropharyngeal cancers staged T1/T2 N0. The sensitivity of the procedure is 90% when the histopathology of the sentinel node is compared with that of the neck dissection specimen (27). It results in histopathologic upstaging of the clinically N0 neck in 36% of the patients when the nodes are examined with routine hematoxylin-eosin staining; serial sectioning and immunohistochemistry upstage an additional 8% of the cases (24). The detection of micrometastases can be further enhanced by using highly specific tumor markers and molecular methods (28,29).

Recently, Civantos et al. (30) reported the results of a North American Multi-Institutional Prospective Study that evaluated the utility of SLNB in T1/T2 oral SCCAs. The study included 140 patients (68% oral tongue, 19% the floor of the mouth) from 25 institutions who underwent SLNB and selective neck dissection (SND) (levels I-IV). The negative predictive value (NPV) was 94% when the SLNs were examined with hematoxylin-eosin stains and 96% when they were examined with serial sectioning and immunocytochemistry (30). The NPV among experienced surgeons was 100% versus 95% for less experienced ones. The SLN was the only positive node in 51% of the cases with a positive SLN. The false-negative rate was 9.8% overall. Interestingly, however, the false-negative rate was 10% in patients with cancers of the oral tongue but was 25% in patients with cancers of the floor of the mouth. In the experience of the University of Miami (31), the NPV of SLNB was 88.5% in patients with cancers of the floor of mouth (FOM) and 95.8% when these patients were excluded. Similarly, Ross et al. (26) have reported that the identification of SLNB in patients with FOM cancers was lower (86%) than in patients with tumors in other locations (97%); the sensitivity of SLNB in FOM cancers was also lower (80%), compared with other tumor locations (100%). It appears that lymphoscintigraphy in cancers of the FOM is not as helpful in identifying the SLN due to the shine-through effect of the radioactivity at the primary site; it obscures the lymph nodes in level I, which are one of the primary echelons of lymphatic drainage for the FOM and lower gum. Obviously, this limits the utility of SLNB in patients with tumors in these locations.

The SNDs consist of the removal of only the lymph node groups at highest risk of containing metastases according to the location of the primary tumor, preserving the spinal accessory nerve, the IJV, and the SCM. There are four main types of SNDs: