Embryology

The embryologic development of the larynx determines the pattern of metastatic spread of laryngeal cancer. The supraglottic larynx is derived from the buccopharyngeal primordium, which develops from the third and fourth branchial arches. The glottis and subglottis are derived from the tracheobronchial primordium from the sixth branchial arch and are formed by the union of lateral furrows that develop on each side of the tracheobronchial primordium. Therefore the larynx has a dual blood supply and lymphatic drainage. The supraglottis is supplied by the superior laryngeal arteries, and its lymphatic drainage follows these vessels to the carotid sheath to drain into deep cervical chain nodes in levels II and III. The glottis and subglottis are supplied by the inferior laryngeal arteries, and similarly, lymphatic drainage from these two regions follows these arteries to drain into prelaryngeal and pretracheal nodes (Level VI), before reaching the deep cervical chain nodes in level IV.³

The glottic region is formed by paired structures that fuse in the midline. The lymphatics drain unilaterally. The vocal folds have sparse lymphatics. Therefore, glottic cancers must invade deeply before getting access to lymphatic channels. These factors explain the lower incidence of lymphatic metastasis in glottic SCC, as well as the propensity for unilateral metastases. Because the supraglottis is formed without a midline union, its lymphatics drain bilaterally and the increased likelihood of bilateral lymphatic metastases from supraglottic carcinoma is ascribed to this embryologic factor, as described by Frazer in 1909.⁴ Pressman and colleagues⁵ noted that the inferior extent of supraglottic injection was the inferior false vocal cord, with the ventricle defining an anatomic barrier and halting further inferior flow of the dye.

Anatomy

A complete description of the anatomy of the larynx is beyond the scope of this chapter, so the following discussion is limited to a description of the boundaries of the larynx, its compartments, and its lymphatic drainage. The laryngeal framework is formed by three unpaired cartilages (epiglottis, thyroid, cricoid) and the paired arytenoid cartilages. The superior boundary of the larynx consists of the superior surface of the epiglottis and the superior edge of the aryepiglottic folds. The anterosuperior limit is formed by the lingual surface of the suprahyoid epiglottis and the hyoepiglottic ligament, which forms the superior boundary of the pre-epiglottic space. The anterior boundary of the larynx is formed by the thyrohyoid membrane and thyroid cartilage in the supraglottis, thyroid cartilage in the glottis, and cricothyroid membrane and anterior arch of the cricoid cartilage in the subglottis. The inferior limit of the larynx is defined by the horizontal plane, which passes through the inferior edge of the cricoid cartilage.⁶ The posterior and lateral borders of the larynx are composed of the laryngeal surface of the aryepiglottic folds, arytenoid cartilages, interarytenoid space, and posterior surface of the subglottic space (which is defined as the mucosa covering the surface of the cricoid cartilage).

The larynx is divided into three regions or sites: the supraglottis, the glottis, and the subglottis (Fig. 107-1). This division reflects the embryologic structure of the larynx and the anatomic barriers to spread of laryngeal cancer outlined earlier. The TNM staging system further subdivides the supraglottis and glottis of the larynx into multiple subsites, used to define the T stage (Table 107-1).⁶

Figure 107-1. Classification of laryngeal lesions by the anatomic site involved.

(Adapted from Ogura JH, Biller HF. Partial and total laryngectomy and radical neck dissection. In: Maloney WH, ed. Otolaryngology. Vol 4. New York: Harper & Row; 1971.)

Table 107-1 Anatomic Sites and Subsites of the Larynx

From Greene F, Page D, Fleming I, et al. AJCC Cancer Staging Manual. 6th ed. New York: Springer; 2002.

The supraglottis is composed of the suprahyoid and infrahyoid epiglottis (both the lingual and the laryngeal surfaces), aryepiglottic folds (laryngeal surfaces only), arytenoids, and ventricular bands (false vocal cords). The boundary between the suprahyoid and infrahyoid epiglottis is a horizontal plane passing through the hyoid bone. The inferior limit of the supraglottis is a horizontal plane through the lateral margin of the ventricle at its junction with the superior surface of the true vocal cord. The glottis is composed of the true vocal cords (both the superior and inferior surfaces) and includes the anterior and posterior commissures. The inferior boundary of the glottis is a horizontal plane 1 cm inferior to the inferior limit of the supraglottis (defined as the lateral margin of the ventricle at its junction with the superior surface of the vocal cord). The subglottis extends from the inferior limit of the glottis to the inferior edge of the cricoid cartilage. It is not divided further into any subsites.⁶

The mucosal lining of the larynx differs in the three regions. The epithelium of the supraglottis is predominantly of the pseudostratified columnar type, except at the edges of the aryepiglottic folds and the lateral borders of the epiglottis, which are stratified squamous epithelium. The supraglottic mucosa has an abundance of mucus glands and lymphatic vessels. The true vocal cords have a unique structure: stratified squamous epithelium covers a three-layered lamina propria (composed of superficial, intermediate, and deep layers). The intermediate and deep layers of the lamina propria make up the vocal ligament, which forms the superior border of the conus elasticus and interdigitates with the vocalis muscle. The true vocal cords (and thus the glottis) have few lymphatic vessels. The subglottis is lined by pseudostratified columnar epithelium, which is in close approximation to the cricoid cartilage and cricothyroid membrane.⁷

The laryngeal cartilages, hyoepiglottic ligament, thyrohyoid membrane, quadrangular membrane, conus elasticus, anterior commissure, and cricothyroid membrane form natural barriers to the spread of tumor. Within the larynx, the pre-epiglottic space (PES) and the paraglottic space (PGS) provide pathways for spread of laryngeal tumors (Fig. 107-2). The boundaries of the PES are: anteriorly, the thyroid cartilage and thyrohyoid membrane; superiorly, the hyoid bone, hyoepiglottic ligament, and valleculae; posteriorly, the anterior surface of the epiglottic cartilage and the thyroepiglottic ligament; and laterally, the PES is open and continuous with each of the two PGSs. The PES contains fat and areolar tissue⁷ and is frequently invaded by tumors because the cartilage of the epiglottis has multiple small fenestrations through which cancers arising from the infrahyoid epiglottis may pass. Superiorly, the hyoepiglottic ligament provides a barrier to spread of tumor to the tongue base (Fig. 107-3). The lymphatics of the PES drain through the thyrohyoid membrane, spreading to lymph nodes on both sides of the neck, primarily in zones II and III (Fig. 107-4).⁵ Supraglottic tumors with PES involvement are staged as T3 lesions.

Figure 107-2. Posterior oblique view of larynx showing confluence of pre-epiglottic and paraglottic spaces.

(Reprinted with permission from Myers EN, Suen JY, Myers JN, Hanna EYN. Cancer of the Head and Neck. 4th ed. Philadelphia: WB Saunders; 2003, Fig. 15-21.)

Figure 107-3. A, Total laryngectomy specimen with a deeply invasive supraglotttic carcinoma arising from the false vocal cord. B, Sagittal section shows cancer filling the paraglottic space without penetrating the hyoepiglottic ligament (arrows).

Rights were not granted to include this figure in electronic media. Please refer to the printed book.

(From Zeitels SM, Kirchner JA. Hyoepiglottic ligament in supraglottic cancer. Ann Otol Rhinol Laryngol. 1995;104:770.)

Figure 107-4. Histologic sections. A, Carcinoma on the laryngeal surface of the epiglottis growing through a normal epiglottic fenestration (light arrows) into the pre-epiglottic space. A typical fibroelastic pseudocapsule has formed (heavy arrows) around advancing carcinoma (hematoxylin-eosin; original magnification ×40). B, Atypical example of squamous cell carcinoma of the epiglottis located below the hyoepiglottic ligament did not grow through the epiglottic foramina into the pre-epiglottic space (arrow) (hematoxylin-eosin; original magnification ×10). C, Squamous cell carcinoma arising on the laryngeal surface of the epiglottis (light arrow) demonstrating classic pattern of invasion through foramina in epiglottic cartilage with pre-epiglottic space extension (heavy arrow) (hematoxylin-eosin; original magnification ×1).

Rights were not granted to include this figure in electronic media. Please refer to the printed book.

(From Zeitels SM, Vaughan CW. Endoscopic management of early supraglottic cancer. Ann Otol Rhinol Laryngol. 1990;99:951.)

The PGS lies lateral to the true and false vocal folds and extends laterally to the thyroid cartilage (Fig. 107-5). The boundaries of each PGS are: medially (from superior to inferior), the quadrangular membrane, laryngeal ventricle, and conus elasticus; laterally, the thyroid cartilage (anteriorly) and the mucosa of the medial wall of the pyriform sinus (posteriorly); and inferolaterally, the cricothyroid membrane.⁸ Anteriorly, each PGS is continuous with the PES, and tumors may spread along this pathway (see Fig. 107-2). Paraglottic space involvement in either a glottic or supraglottic tumor is staged as T3 and is significant because the extent of the PGS means that tumors in this space may spread to involve any or all of the three regions of the larynx (Fig. 107-6).

Figure 107-5. Dimensions of the paraglottic space located between the mucosa of the larynx and its cartilaginous framework.

(Reprinted with permission from the American Academy of Otolaryngology—Head and Neck Surgery; from Myers EN, Alvi A. Management of carcinoma of the supraglottic larynx: evolution, current concepts, and future trends. Laryngoscope. 1996;106:561.)

Figure 107-6. A T3 supraglottic carcinoma extending to the glottis through the paraglottic space. Note the extension below the floor of the ventricle (arrow) with widening of the paraglottic space (hematoxylin-eosin, gross coronal section).

(From Weinstein GS, Laccourreye O, Brasnu D, Tucker J, Montone K. Reconsidering a paradigm: the spread of supraglottic carcinoma to the glottis. Laryngoscope. 1995;105:1131.)

Conservation laryngeal surgery is undertaken on the basis of the theory of compartmentalization of the larynx, which evolved from the work of Frazer,⁴ Pressman and colleagues,⁵ and Tucker and Smith.⁹ Pressman and colleagues found that this separate embryologic derivation explains why supraglottic tumors of substantial bulk do not spread across the laryngeal ventricle to the vocal cord. In experiments using submucosal vital dyes and radioisotopes, Pressman also noted that the inferior extent of supraglottic injection was the inferior false vocal cord. The ventricle was an anatomic barrier to the inferior flow of the dye and thus was confirmed as a barrier to tumor spread. Using animals, cadavers, and whole organ serial sections of human tumor specimens, Tucker and Smith confirmed that elastic tissue barriers within the larynx explained the findings of the dye studies. Although these studies confirm compartmentalization of the larynx and clinically supraglottic tumors are observed to invade the glottis (and vice versa) uncommonly, there is no true anatomic barrier separating the supraglottis from the glottis.

Treatment of Premalignant Lesions

The treatment of a premalignant lesion should aim to eradicate the lesion while preserving voice quality and laryngeal function. Accurate diagnosis is critical for proper decision making, and a number of factors related to the characteristics of the lesion, as well as patient factors, affect the treatment decision-making process (Table 107-5).

Table 107-5 Factors to Consider in Treatment of Premalignant Laryngeal Lesions

Risk Factors

Second Primary Tumors

The single greatest risk factor for head and neck SCC is prior head and neck SCC. The annual risk of a second primary tumor (SPT) following an index head and neck SCC is 1% to 7%, with the risk persisting for at least 10 years.^57,90–92 The cumulative risk of developing an SPT is at least 20% and is greater for those who continue to use tobacco and alcohol. An individual with stage I or II head and neck SCC is more likely to die from an SPT than from the index tumor. The majority of SPTs develop in the head and neck, but a significant proportion occur in the esophagus or lung. Second primary tumors may be synchronous (identified within 6 months of the index tumor) or metachronous (diagnosed more than 6 months after the index tumor). In a study of 875 patients with SCC of the head and neck, 207 had SPTs develop within 5 years of their index tumor.⁹³ Of these patients, 31% had a third primary malignancy develop and 10% had a fourth primary malignancy develop. For laryngeal cancer, the lung is a significant site for synchronous and metachronous SPTs. An isolated lung malignancy in a patient with larynx cancer is more likely to be an SPT than a metastasis from the laryngeal cancer. Therefore an isolated pulmonary nodule should be considered an SPT until proven otherwise.⁹² Prior RT is associated with SPTs in a small number of cases.⁷⁴ Slaughter and associates⁹⁴ examined clinically normal tissue adjacent to head and neck cancers and identified many histologic changes seen in the malignant cells in the adjacent normal appearing tissue, leading him to propose the concept of “field cancerization.” Improved understanding of head and neck carcinogenesis has provided a molecular explanation for Slaughter’s observations.^50,95 Bedi and colleagues examined X-chromosome inactivation and performed microsatellite analysis to evaluate allelic loss at chromosomes 3p and 9p in females with multiple primary head and neck cancers. Both the original cancer and the second malignancy arose from a single clone.⁹⁵ Califano and coworkers similarly observed that tissues adjacent to malignant and premalignant lesions shared common genetic changes.⁵⁰ Multiple tumors generally do not arise from multiple transforming events, but instead a single transforming event produces a cell with growth advantage that spreads throughout the mucosal surface. The tumor may accumulate further genetic damage, resulting ultimately in additional malignancies that are geographically distinct but genetically related to the original cancer.⁹⁵

Cervical Nodal Metastases

The incidence of cervical metastases from SCC of the larynx, as well as the nodal groups involved, vary according to the site of the primary tumor (Fig. 107-9). Due to its rich network of lymphatics, supraglottic SCC has the highest incidence of regional metastases—both clinically apparent and occult metastases. Cervical metastasis has been confirmed pathologically in 10% of T1 supraglottic lesions, 29% of T2, 38% of T3, and 57% of T4.¹⁰⁴ The incidence of occult metastases (cN0, pN+) in supraglottic SCC varies from 12% to 40% for all T stages.^105–108 The incidence of occult metastases is greater in tumors with a higher T stage: T1 0% to 14%, T2 20% to 21%, T3 28% to 35%, and T4 40% to 75%.^106,109 Supraglottic SCC usually metastasizes to levels II, III, and IV. Levels I and V are involved by metastases rarely and only when other nodal levels are also involved.^110,111 Bilateral metastases (palpable and occult) occur frequently in supraglottic SCC and are more common in midline or bilateral tumors.^104,112 Therefore, in cases of supraglottic SCC, surgical treatment of the N0 and N1 neck is usually with a bilateral selective neck dissection (levels II-IV). For N2 or N3 disease, a comprehensive neck dissection (levels I-V) is indicated.

Figure 107-9. The six nodal levels of the neck.

(Redrawn from Robbins KT, Clayman G, Levine PA et al. Neck dissection classification update: revisions proposed by the American Head and Neck Society and the American Academy of Otolaryngology–Head and Neck Surgery. Arch Otolaryngol-Head Neck Surg. 2002;128:751-758, Figure 1).

Glottic SCC has a low risk of cervical metastasis: in a series of 910 patients, the overall incidence of pathologically confirmed nodal metastasis was 5.9%, with an incidence of occult metastasis of 18%.¹¹³ Similar to supraglottic SCC, the incidence of regional metastasis correlates with T stage. In the same series, nodal metastases were found in 0.1% of T1 tumors (one patient only), 5% of T2, 18% of T3, and 32% of T4.¹¹³ The nodes at risk of metastasis from glottic SCC are those in levels II, II, and IV, as well as level VI (prelaryngeal, pretracheal, and paratracheal nodes). Bilateral or contralateral metastases are rare.

Primary subglottic carcinoma is rare, and descriptions of the clinical behavior of these tumors are based on a small number of patients. The paratracheal nodes (level VI) are most frequently involved with metastases, including contralateral or bilateral metastases. Metastases to levels III, IV and V are uncommon. Although these tumors are aggressive and have a poor prognosis, the incidence of cervical metastasis is generally reported to be low, with a range of 4% to 27%.^101,114 However, Harrison¹¹⁵ detected metastatic tumor in approximately 50% of serially sectioned paratracheal nodes. Metastases to mediastinal lymph nodes are common (up to 46%), but they are classified as distant metastases.¹¹⁶

Distant Metastases

Distant metastases from laryngeal SCC include not only hematogenous metastases to distant organs, but also lymphatic metastases to nodal groups outside the neck.⁵⁹ The most common site for distant hematogenous metastases is the lung. The liver and skeletal system (ribs, vertebrae, and skull) are affected less often. The mediastinum is the most common site for distant lymphatic metastases.¹¹⁷ Distant metastases are uncommon at initial presentation. Patients who develop distant metastases have almost always had regional metastases diagnosed at some stage in the course of their disease. The incidence of distant metastasis varies according to the site of the primary tumor: 3.1% to 8.8% in glottic SCC, and 3.7% to 15% in supraglottic SCC.^117–121 Supraglottic SCC usually has a higher incidence of distant metastases compared with glottic SCC. The frequency of distant metastases from subglottic SCC is less certain because primary tumors in this site are rare; however, Spector and colleagues¹¹⁷ observed that 14.3% of subglottic SCCs developed distant metastases. Clinical and pathologic factors associated with an increased risk of distant metastases include advanced stage of primary tumor (especially T4 stage); presence of cervical metastases (especially N2 and N3 disease); duration, level, and extracapsular spread of cervical metastases; and locoregional recurrence.^117,121,122 Lymphatic metastasis to the skin is also a sign of advanced disease and, similar to distant metastases, has a grave prognosis.¹²³

History and Physical Examination

A history of the presenting symptoms and any associated UADT symptoms should be elicited, with particular attention paid to symptoms suggesting impending airway obstruction such as stridor and dyspnea. Additional information that should be gathered includes exposure to risk factors for laryngeal cancer (primarily tobacco and alcohol), medications, and medical comorbidities such as cardiovascular, pulmonary, or renal disease that may affect the type of treatment that can be offered.

A complete head and neck examination is performed for suspected laryngeal cancer. Initial assessment should determine whether urgent intervention is required, for example, to secure an airway for a patient with a large obstructing lesion. Subsequent assessment is directed at assessing the location and extent of the primary tumor, palpating for neck nodes, and examining the rest of the upper aerodigestive tract for synchronous tumors.

Indirect laryngoscopy (IDL) using a mirror may provide an excellent three-dimensional view of the larynx and any lesions. However, images are not recorded, the hypopharynx is not adequately visualized, and in patients who do not tolerate IDL, the larynx cannot be evaluated completely. Flexible fiberoptic laryngoscopy permits photographic and video documentation, as well as examination during dynamic maneuvers (e.g., Valsalva’s maneuver, coughing, swallowing). Laryngeal videostroboscopy is useful for the documentation of small lesions on the vocal fold and assessment of the mucosal wave before and after treatment. However, videostroboscopy cannot reliably distinguish intraepithelial neoplasia from invasive carcinoma, nor is it reliable for determining the depth of invasion of vocal cord carcinoma.¹²⁴

In the patient with laryngeal cancer, specific examination findings that should be noted include the site of the primary tumor; the subsites involved; extension to adjacent structures; vocal cord mobility (normal vs. impaired vs. fixed); and patency of the airway. A biopsy may be taken in the office via a transoral approach or via the working channel of the flexible laryngoscope, although this is usually performed in the operating room under general anesthesia, unless there are concerns regarding a patient’s fitness for anesthesia. The base of tongue and the laryngeal skeleton are palpated to assess for extralaryngeal spread, and the neck is palpated to assess the nodal status.

Imaging

Radiologic imaging is a critical part of the evaluation of a patient with a suspected laryngeal malignancy.¹²⁵ When feasible, imaging is performed before operative endoscopy and biopsy to obtain images before potential edema and distortion from biopsy and manipulation of the larynx. Information obtained with imaging can be integrated with endoscopic findings and, in some cases, can guide the surgeon to focus attention on deeper areas not normally visible on endoscopy requiring a biopsy. Cross-sectional imaging, using either computed tomography (CT) scanning or magnetic resonance imaging (MRI), yields the most useful information in the evaluation of tumors of the larynx.

Computed tomography is the most commonly used radiologic investigation in the evaluation of laryngeal cancer, although MRI is superior in some situations. CT is superior to MRI for imaging bony structures (such as ossified cartilages) and calcifications.¹²⁵ MRI is superior for detecting cartilage invasion and has enhanced ability to discriminate soft tissues. The disadvantages of MRI of the larynx compared with CT include increased imaging time (resulting in motion artifact), decreased spatial resolution, inferior ability to detect bone invasion, and increased cost.¹²⁶ MRI is also contraindicated for patients with pacemakers or other ferromagnetic foreign bodies, and false-positive findings are common because MRI is often unable to distinguish clearly between reactive inflammatory changes and tumor.¹²⁷

Evaluation with CT includes a standard neck scan to evaluate extralaryngeal structures and lymph nodes, and finely cut images (≈1 mm) through the larynx to avoid partial volume averaging of the vocal folds, and allow multiplanar reconstruction.¹²⁸ The axial images should be aligned parallel to the true vocal cords to allow correct interpretation of tumor location and size.¹²⁵ Intravenous contrast is especially important because the primary tumor usually enhances with contrast, improving its visualization, and because the contrast will differentiate between nodal disease and blood vessels in the neck. Contrast (gadolinium)-enhanced, T1-weighted MRI images improve the definition of pathologic tissue, although some authors have found the use of contrast adds no information when assessing laryngeal tumors.^129,130 Special MRI series such as fat suppression are often helpful, especially when assessing the paraglottic and pre-epiglottic spaces.¹²⁵

Positron emission tomography (PET), using radiolabeled glucose (18 F-fluorodeoxyglucose, FDG), is able to distinguish between malignancy and other pathologic processes such as inflammation.¹³¹ PET is particularly useful in the head and neck when performed in conjunction with CT (PET/CT), giving superior anatomic localization of suspicious areas.¹³² PET has several potential uses: the staging of malignant tumors (especially for the detection of regional and distant metastases); the diagnosis of the unknown primary; the detection of synchronous lesions in addition to the index malignancy; and in the diagnosis of residual or recurrent cancer, which is particularly challenging in the irradiated larynx (see later).¹³³ PET is being used more commonly in the head and neck as our understanding of its abilities and applications improves.

Radiology plays an important role in the staging of laryngeal SCC because deep invasion by the primary tumor is assessed poorly by clinical and endoscopic examination. Several characteristics of the primary tumor used for staging cannot be determined without imaging including invasion of the PES, the PGS, the thyroid or cricoid cartilage, the extralaryngeal tissues, the prevertebral space, mediastinal structures, and encasement of the carotid.⁶

Invasion of the PES is diagnosed on CT by a loss of fat in the PES and replacement with a moderately enhancing mass.¹²⁸ On MRI, tumor has a low to intermediate signal on T1-weighted images, with a high signal on T2, and enhances with contrast on T1.¹²⁸ MRI is highly sensitive in the diagnosis of PES and base of tongue invasion and is superior to CT.¹³⁴ The radiologic appearance of the thyroid and cricoid cartilages depends on the degree of ossification. Although CT is superior to MRI for imaging bone and calcium, MRI is superior to CT in the detection of laryngeal cartilage invasion.^126,127,135 MRI is highly sensitive for cartilage invasion and has a high negative predictive value; however, it is not good at distinguishing between reactive inflammation and tumor, resulting in a lower specificity and a positive predictive value of approximately 70%.^126,127,135 CT is less sensitive than MRI in the diagnosis of cartilage invasion. Findings on CT such as extralaryngeal spread, sclerosis, erosion, and lysis suggest cartilage invasion.¹³⁶ MRI is also better than CT at detecting cricoid invasion in subglottic tumors.¹³⁷ The accurate diagnosis of cartilage invasion is important because it will upstage a carcinoma to T3 or T4a and is associated with a worse response to radiotherapy, poorer local control, and an increased risk of chondroradionecrosis.^126,138

Imaging is also important to stage the neck accurately and detect distant metastases. In one series, CT and MRI were equivalent in the detection of cervical metastases in larynx cancer and superior to clinical examination and ultrasound. However, ultrasound-guided FNA and PET were the most accurate.¹³⁹ PET has not been used extensively in the pretreatment evaluation of laryngeal cancer because standard PET lacks anatomic detail. However, fused PET/CT scanning addresses this issue and in all likelihood will become more widely used for this indication.¹⁴⁰

The lungs are the most common site of metastasis for lung cancer, and second primary tumors commonly occur in the lungs as well. A chest radiograph or a chest CT should be obtained to exclude pulmonary lesions. In patients with advanced primary tumors and low cervical metastases, PET/CT is useful for identifying distant metastases or SPTs that would alter patient management.¹⁴⁰

Operative Endoscopic Examination

Unless there are medical contraindications, all patients with suspected laryngeal cancer should undergo an endoscopic examination under general anesthesia. Direct laryngoscopy allows the clinician to examine the larynx in greater detail, palpate the larynx, and obtain a biopsy for histologic analysis. Angled endoscopes are useful to evaluate the laryngeal surface of the epiglottis, anterior commissure, ventricles, and undersurface of the true vocal cords. If vocal cord mobility is noted to be abnormal on preoperative examination, palpation of the vocal cord and arytenoid cartilage will differentiate arytenoid fixation from vocalis muscle invasion causing cord immobility. Invasion of adjacent sites in the oropharynx or hypopharynx is assessed with direct pharyngoscopy and palpation of the tongue base. The neck may be palpated under general anesthetic to assess nodal status. The complete examination should be documented with photos, video, and a diagram of the extent of the primary tumor. Esophagoscopy is performed to exclude an SPT in the esophagus. Bronchoscopy is not required routinely if the preoperative chest imaging is normal.¹⁴¹ Biopsies of the primary tumor and any suspicious areas are taken to confirm the histologic diagnosis and to determine the extent of tumor.

Caution should be exercised when performing direct laryngoscopy for patients presenting with laryngeal tumors obstructing the airway. Maintenance of a secure airway at all times is the paramount concern, and for severely compromised airways a tracheostomy under local anesthetic may be necessary. However, if possible, debulking of the obstructing tumor (using either cup forceps, a microdebrider, or the CO₂ laser) is the preferred technique to avoid a tracheostomy and to maintain the airway so that thorough evaluation and staging can be performed before selection of the definitive treatment.¹⁴² If a tracheostomy is necessary, a high tracheostomy is performed so that any subsequent resection allows preservation of as much of the trachea as is oncologically possible. Surgical treatment is usually required for patients requiring tracheostomy because significant destruction of the larynx is usually present. The traditional treatment of an obstructing laryngeal SCC requiring a tracheostomy was an emergency laryngectomy (after a biopsy and frozen section biopsy had confirmed the diagnosis). However, emergency laryngectomy does not offer a survival advantage over tracheostomy with delayed laryngectomy and does not allow the patient to be counseled and prepared both psychologically and nutritionally before such radical laryngeal surgery.^142,143 Tracheostomy has not been found to increase the risk of peristomal recurrence, which is more strongly associated with advanced local disease, especially subglottic tumor extension.^143–146

Treatment of the Early Primary Tumor in Glottic Squamous Cell Carcinoma

Early glottic SCC (defined as stage I or II disease, i.e., T1-2N0) may be treated with either RT or surgery, without the need for elective treatment of the neck.¹⁴⁷ Primary RT for T1 glottic SCC provides 5-year local control rates of 81% to 90% and laryngeal preservation in 90% to 98% of patients.^148–150 For T2 tumors with normal vocal cord mobility, RT achieves local control in 64% to 87%, with laryngeal preservation rates of 75% to 87%.^148,151,152 The effectiveness of RT may be overestimated due to complete excision of the tumor at the time of the initial biopsy. In 12 of 60 patients undergoing partial laryngectomy for early glottic SCC, Stutsman and McGavran found that there was no tumor in the pathologic specimen.^153,154

Surgical treatment of early glottic SCC also aims to preserve the larynx and is referred to as conservation laryngeal surgery or partial laryngectomy. Traditionally, these limited laryngeal resections were performed via an external approach: Cordectomy and vertical hemilaryngectomy (VHL) are the two classic open procedures for the treatment of early glottic carcinoma. Cordectomy is the removal of the diseased true vocal cord via a laryngofissure. Vertical hemilaryngectomy removes the ipsilateral true and false vocal cords, extending laterally to the perichondrium of the thyroid cartilage. The lamina of the thyroid may be removed to allow the soft tissues adjacent to the larynx to collapse medially to reconstitute the glottis for phonation, or it may be preserved, with transposition of soft tissue (such as a strap muscle) medial to the lamina to recreate the glottis. Variations of the VHL (extended VHL) have been described to include the resection of the anterior commissure, contralateral true vocal cord, arytenoid, and supraglottic or subglottic tumor extension. Open surgical treatment oncologic results are reported as having 90% to 98% local control rates and 93% to 98% larynx preservation rates.^155,156

In the past 30 years, endoscopic approaches analogous to the open procedures have been developed to accomplish the same resection without disruption of the supporting structures of the larynx. The oncologic results for transoral laser microsurgery (TLM) have been reviewed by Ambrosch.¹⁵⁷ For Tis-T2 tumors, local control rates are 80% to 94% with a greater than 94% laryngeal preservation rate.¹⁵⁸ In the hands of experienced surgeons, local control and laryngeal preservation rates following TLM equal those following open surgical techniques.¹⁵⁸ In comparison with open surgical techniques, TLM avoids a tracheostomy, the hospital stay is shorter, the cost is reduced, and there is a lower incidence of dysphagia postoperatively.¹⁵⁷

Lesions of the middle third of the true vocal cord have the best local control rates and may be treated by either TLM, open cordectomy or RT; local control approaches 100% after surgical excision, whereas RT achieves a 95% local control rate.¹⁵⁹ Radiation failures may be caused by unrecognized deep invasion. Following surgical excision, repeat surgery or RT may be used to treat residual or recurrent tumor. Although RT alone has excellent results, a second course of radiation for a recurrence or for a second tumor cannot be offered. Recurrent tumors may not be amenable to conservation surgery after previous RT.

T2 glottic lesions with impaired vocal cord mobility warrant special consideration. Although classified as T2 lesions by the TNM system, glottic tumors with impaired vocal cord mobility have a worse prognosis than tumors that are classified as T2 on the basis of supraglottic or infraglottic invasion. Impaired vocal cord mobility is usually secondary to either tumor bulk or deep invasion. Radiotherapy is less effective in controlling these lesions, and this is probably often due to tumor volume. Dickens and colleagues^148,160 noted that 4% of tumors less than 15 mm recurred after radiotherapy, whereas 26% of larger lesions of a similar stage recurred, even when only one true vocal cord was involved. T2 tumors managed by primary RT showed a 30% local failure rate, which, when surgically salvaged, improved to 94%.^19,161 Harwood and DeBoer¹⁹ observed that impaired vocal cord mobility resulted in lower control rates in T2 lesions, and they suggested the classification be divided into T2a and T2b on the basis of mobility. In this analysis, a 70% local control rate was noted for the former category versus 51% in the latter. McLaughlin and colleagues¹⁶² noted recurrence rates of 11% and 26% for T2a and T2b tumors, respectively.

Voice quality following either surgery or RT is influenced by the extent of tumor and depth of invasion. Small superficial tumors will have excellent voice quality with either surgery or RT. Deeper tumors with muscular invasion will have inferior voice outcomes with either treatment modality. Furthermore, the control rate with RT will be lower. Surgical treatment provides a better assessment of tumor extent and in some cases may result in upstaging of the tumor. For small superficial tumors, voice quality with surgery or RT is generally good with comparable voice results.¹⁶³

Early Glottic Squamous Cell Carcinoma and the Anterior Commissure

Anterior commissure involvement has been associated with decreased local control rates with surgery and RT.^164,165 The topic has been controversial, and several explanations for the poorer control rates have been offered. One hypothesis for the decreased effectiveness of RT has been underdosage with supervoltage RT at the tumor-air interface. Increased dose fractions (to >2 Gy) are believed to have solved this problem. The anterior commissure is also a difficult region to assess, and deep invasion may not be recognized, resulting in understaging and thus undertreatment. The lack of perichondrium at the insertion of the anterior commissure tendon was thought to increase the risk of cartilage invasion. Kirchner and Carter¹⁶⁶ examined laryngeal carcinoma with anterior commissure invasion using whole organ sections of the larynx and discovered that the anterior commissure tendon is a strong barrier to cancer spread. Deep invasion was seen only in cases where the tumor had invaded the supraglottis superiorly or the subglottis inferiorly. Spread of tumor across the anterior commissure did not increase the risk of deep invasion. Kirchner concluded that supraglottic spread provides access to the PES and that subglottic spread provides access to the thyroid cartilage and the cricothyroid membrane.

Frontolateral VHL obtains local control rates of 80% to 90% for T1 carcinomas involving the anterior commissure.^167,168 Supracricoid partial laryngectomy (SCPL) is a more extensive procedure that removes the anterior commissure and the anterior two thirds of the vocal folds. Laccourreye and colleagues have reported a 5-year local control rate of 98% for T1 and T2 glottic tumors with anterior commissure involvement.¹⁶⁹ Bron and colleagues¹⁷⁰ reported local control of 94.5% for 45 previously untreated laryngeal SCCs involving the anterior commissure treated with SCPL. Although oncologically effective, voice quality is significantly impaired following this technique.

Early reports of TLM for glottic carcinoma considered anterior commissure involvement to be a contraindication. Krespi and colleagues noted a high rate of failure at the anterior commissure.¹⁷¹ This region may be difficult to visualize at the time of surgical resection. With improved understanding of the anatomy, instrumentation, and technique, excellent control rates have been obtained.^172,173 Pearson and Salassa¹⁶⁵ reported their initial experience of 39 patients with anterior commissure involvement. They had no local failures among 17 pT1 and pT2a tumors. The majority (19/22) of advanced tumors with anterior commissure involvement (pT2b, pT3-4) tumors were controlled with endoscopic surgery. Steiner and colleagues¹⁵⁸ reported results on 263 patients with early glottic tumors treated over a 10-year period and observed a modest decrease in local control and laryngeal preservation rates, with equivalent 5-year survival. For T1a tumors, local control was 90% when the anterior commissure was not involved and 84% with anterior commissure involvement. The corresponding laryngeal preservation rates were 99% versus 93%. Similar findings were seen with T1b and T2a tumors.

Treatment of the Advanced Primary Tumor in Glottic Squamous Cell Carcinoma

Advanced glottic SCC (stage III or IV disease) is associated with vocal cord fixation, cartilage invasion, transglottic spread of tumor, subglottic extension, laryngeal framework invasion, extralaryngeal spread, lymph node metastases, and distant metastases. These features portend worse prognosis. The treatment of choice for T3 and T4 glottic tumors remains controversial because of the heterogeneity of the tumors and lack of reliable studies comparing surgery and radiotherapy for T3 and T4 carcinoma of the larynx.¹⁷⁴ T3 and T4 glottic tumors are discussed separately in this chapter. T3 glottic carcinomas are unusual because despite their advanced T stage, in general they have a low risk of nodal metastasis. Furthermore, the spectrum of disease with T3 lesions is variable, ranging from low-volume tumors invading the vocalis muscle to cause fixation to very large transglottic tumors. Tumor volume and transglottic spread of T3 tumors predict increased aggressiveness, increased rate of lymph node metastasis, and poorer response to treatment. Tumors larger than 1.5 cm, subglottic extension, and lymph node metastasis laterally or to paratracheal or anterior pretracheal nodes predict failure above the clavicle.^175,176

Traditionally, T3 tumors have been treated with total laryngectomy as single-modality therapy. Open VHL and more extensive partial laryngectomies are used in carefully selected cases. Kirchner and Som¹⁷⁷ reported a 2-year survival rate of 60% following open partial laryngectomy and noted that failures occurred when there was inferior tumor extension in the larynx. Biller and Lawson reported a 73% absolute 2-year, tumor-free control rate following partial laryngectomy (with resection extended to include a portion of the cricoid cartilage when there was >5 mm of subglottic extension).¹⁷⁸ Pearson and coworkers have published extensive experience of near-total laryngectomy (NTL) for patients with tumors not suitable for other conservation procedures.¹⁷⁹ This procedure preserves one arytenoid and a portion of the cricoid cartilage to create a diversionary voice shunt from the trachea. Patients remain tracheotomy dependent for breathing and use the shunt to produce speech.

Radiation therapy has a local control rate of approximately 50% for T3 tumors, which is lower than that of surgery.¹⁸⁰ For T3 tumors, the return of vocal cord mobility following RT predicts a good response. In a small series of 14 patients with T2b-3 laryngeal carcinomas, all five patients who did not have return of cord mobility and none of nine patients who had return of vocal cord mobility developed recurrent cancer.¹⁸¹ Tumor volume may predict the response to irradiation, with poor results in larger tumors.¹⁸² For small T3 tumors that would be amenable to partial laryngectomy, primary RT could be considered in those not wishing to pursue surgical option, although surgical resection generally has higher local control.^183,184 Intensified RT regimens including twice-daily treatments and the use of intensity-modulated radiotherapy (IMRT) may improve local control.¹⁸⁵ Further discussion on RT can be found in Chapter 112. Total laryngectomy or CRT is recommended for bulky T3 tumors or tumors not suitable for conservation laryngeal surgery.

In general, T4 glottic carcinoma is not considered amenable to conservative laryngeal resection. Options for T4 glottic carcinoma include total laryngectomy (usually with postoperative RT or CRT), near-total laryngectomy, or primary CRT in selected low-volume disease with limited cartilage destruction to preserve the larynx. Near-total laryngectomy may be considered in selected cases with limited subglottic extension and no interarytenoid involvement.¹⁸⁶ Recent encouraging results for TLM for T3-4 laryngeal cancer support its use in carefully selected cases by experienced surgeons for organ preservation.¹⁸⁷

Primary RT for T4 glottic carcinoma has poor local control rates. A historical review of primary RT for laryngeal SCC found that only 2 of 25 T4N0 glottic carcinomas treated with primary RT survived 5 years.¹⁸⁸ In patients unable or unwilling to undergo concurrent CRT and unwilling to have a total laryngectomy, primary RT may provide a chance of local control. In some patients, the addition of newer agents such as cetuximab may increase the effectiveness of RT with acceptable risk of increased toxicity.¹⁸⁹

Although attempts to avoid total laryngectomy are worthwhile, RT may result in significant local tissue destruction, scarring, and persistent edema. Thus the larynx may be preserved, but the patient may be left with a severely compromised organ with a restricted airway, a poor voice, dysphagia, and/or aspiration. If RT is selected for primary management, close follow-up is required as the success of this approach relies on the early detection of residual or recurrent disease, which may be challenging in a larynx scarred from irradiation. A total laryngectomy is usually required for salvage if recurrent disease is diagnosed. The successful salvage rate for recurrences after RT is approximately 60%. Persistent postradiation edema predicts persistent disease: 45% of patients with edema persisting for longer than 6 months after RT had a deep recurrence.¹⁹⁰ In these situations close follow-up with endoscopy and imaging is vital.¹⁹¹ Distinguishing between recurrence and chondroradionecrosis of the larynx can be challenging. Deep biopsies required to obtain an accurate diagnosis can induce or exacerbate chondroradionecrosis. Positron emission tomography (PET) scanning has been helpful to resolve this dilemma and assists surveillance for tumor recurrence.¹⁹²

Selected T4a tumors may be considered for nonsurgical organ preservation trials (see “Nonsurgical Organ Preservation” later). The success of organ-preservation protocols for bulky T4a tumors is lower than for T3 tumors because cartilage destruction predicts poor response to organ preservation protocols. These patients were excluded from the RTOG 91-11 trial comparing concurrent CRT, induction chemotherapy followed by RT, and RT alone.¹⁹³

When total laryngectomy is undertaken for T4 disease, hemithyroidectomy or subtotal thyroidectomy is recommended for cases of palpable abnormality, subglottic tumors, or glottic tumors with greater than 1 cm of subglottic extension.¹⁹⁴ Thyroid gland invasion may be predicted if a positive Delphian node or cartilage destruction is present. Cancer is found in 3% to 8% of thyroid specimens.

Treatment of the Neck in Glottic Squamous Cell Carcinoma

The incidence of nodal metastasis for glottic SCC is lower than either supraglottic or subglottic SCC. Because the true vocal cords are nearly devoid of lymphatics, tumors limited to the glottis rarely metastasize to regional nodes. When metastasis does occur, the nodes at risk are the prelaryngeal, pretracheal, and paratracheal nodes, as well as the upper, mid, and lower deep cervical chain nodes (levels II, III, and IV).

Occult metastases from T1-2 glottic SCCs are uncommon, and elective treatment of the N0 neck is not required.¹⁹⁵ T3 glottic SCC is more controversial. In general, occult nodal metastases are uncommon from T3 glottic carcinomas unless there is transglottic spread of the tumor, which has a higher rate of occult metastasis.¹⁹⁶ A national survey of otolaryngologists published in 2003 reported that 87% of respondents treated the neck of patients with T3N0 glottic SCC, and 90% treated the neck of those with T4N0 glottic SCC.¹⁹⁷ T4 glottic carcinomas have a higher risk of occult metastases, approximately 20%, and treatment of the neck is recommended. If the primary tumor is being treated surgically, an ipsilateral selective neck dissection is recommended. For glottic carcinoma, the paratracheal nodes and levels II-IV are dissected. Adjuvant treatment with RT or CRT is used depending on pathologic findings in the neck dissection specimen.^198,199 If RT is used to treat the primary tumor, the central compartment and ipsilateral lateral neck are included in the field.

For all T stages, clinically evident nodal disease (N+) warrants aggressive treatment, the choice of which depends on the management of the primary tumor. If surgery is used for the primary tumor, an ipsilateral comprehensive neck dissection is indicated for N+ disease. Adjuvant treatment may be indicated depending on pathologic findings.

Postoperative RT is recommended when multiple nodes, extracapsular spread, extralaryngeal invasion, and perineural or lymphovascular invasion is present. Recently, two large studies have demonstrated improved locoregional control with use of platinum-based chemotherapy given concurrently with RT.^200,201 Cooper and colleagues²⁰⁰ found a 10% increase in the 2-year locoregional control rate (82% vs. 72%), although there was an associated 43% increase in acute grade III or higher toxicity (34% vs. 77%). Similarly, Bernier and associates observed an 11% increase in 5-year progression-free survival, which was also associated with a higher rate of acute toxicity.²⁰¹ A comparative analysis of the combined results of these two studies found that extracapsular spread and microscopically positive surgical margins were the only risk factors for which adjuvant chemotherapy enhanced the efficacy of RT in both studies.²⁰² Other variables trended toward statistical significance, but patients who had two or more positive lymph nodes (without extracapsular spread) as their only risk factor did not seem to benefit from the addition of chemotherapy.

Improved locoregional control rate with primary concurrent CRT has led to the addition of chemotherapy to RT as an adjuvant treatment for patients with adverse pathologic findings in the surgical specimen. The decision to use chemotherapy in this setting requires careful consideration of the ability of the patient to tolerate the treatment. The success of this therapy is dependent on completion of the treatment regimen without significant breaks.^203–206 Encouragingly, a recent analysis of the tolerability of concurrent CRT in patients older than 70 years of age demonstrated high compliance with a regimen of concurrent carboplatin and RT.²⁰⁷

Treatment of the Early Primary Tumor in Supraglottic Squamous Cell Carcinoma

For early supraglottic cancer, the two most frequently used treatment options are supraglottic partial laryngectomy (SGL) or RT. SGL for early supraglottic SCC (as well as some T3 tumors) is achieved by an open supraglottic laryngectomy (OSGL) or by the more recently described TLM.

The traditional OSGL was first described by Alonso in 1947 and subsequently refined by Ogura and Som.^210–212 The oncologic validity of OSGL as a treatment for supraglottic SCC comes from the principle that the supraglottis is embryologically separate from the glottis and subglottis, and for this reason, supraglottic SCC remains localized to the supraglottis in most cases, despite the lack of an anatomic structure preventing invasion of the glottis.²¹³ OSGL removes all laryngeal structures superior to the floor of the ventricle, preserving both true vocal cords, both arytenoids, the base of tongue, and the hyoid bone (Fig. 107-10). Indications for OSGL are T1, T2, or selected T3 supraglottic tumors. T3 tumors with PES and no transglottic spread and/or vocal cord impairment are amenable to OSGL. Contraindications to OSGL include poor general physical condition (e.g., advanced age, lung disease, neurologic disease, preexisting dysphagia or aspiration), glottic involvement, impaired mobility or fixation of the vocal cord, thyroid or cricoid cartilage invasion, involvement of the base of tongue to within 1 cm of the circumvallate papilla, or involvement of the deep muscles of the tongue.²¹⁴ The functional morbidity of OSGL is significant, and almost all patients will experience some aspiration postoperatively. Therefore careful patient selection is crucial to the overall success of this procedure: Candidates for OSGL must have good pulmonary function to tolerate the expected aspiration. An OSGL is referred to as an extended OSGL when a more extensive resection is carried out for the treatment of SCC involving the lingual surface of the epiglottis, the base of tongue, or one of the arytenoids.²¹⁴

Figure 107-10. Outline of resection for open supraglottic laryngectomy. Note inclusion of entire pre-epiglottic space.

(Redrawn from Som ML. Conservation surgery for carcinoma of the supraglottis. J Laryngol Otol. 1970;84:657.)

The functional and oncologic outcomes of OSGL are well documented. Because OSGL disrupts the pharyngeal muscles, the strap muscles, and the sensory innervation of the pharynx and larynx, swallowing is markedly impaired, especially in the early postoperative period. A tracheostomy is required in all patients to provide a safe airway and to protect the lower airway from aspiration. Suarez and colleagues²¹⁵ reported that 10% of patients in their series required a completion total laryngectomy for chronic aspiration. Advanced age (older than 65 years) was the major risk factor for intractable aspiration. A further 24% of subjects required permanent tracheostomy. Bron and colleagues²¹⁶ noted approximately one quarter of their patients aspirated in the early postoperative period. The median duration until normal oral feeding was established was 16 days, and the median duration until decannulation was 17 days.

OSGL yields excellent local control in the treatment of early supraglottic cancer in the range of 80% to 100%.^217–222 In patients treated with OSGL for T1-3 supraglottic SCC, Bron and colleagues²¹⁶ reported 5-year local control, locoregional control, and overall survival rates of 92.5%, 90%, and 92%, respectively. Adjuvant RT was given in 30% for positive surgical margins or adverse pathologic findings in the lymph nodes. Local control and survival were poor when cartilage invasion was present or when there was extralaryngeal extension of disease (i.e., pT4). Suarez and colleagues²²² treated 193 patients with supraglottic SCC with OSGL and ND. Nearly half of the subjects received adjuvant RT. Local control was 98% for T1 tumors and 91% for T2 tumors. No improvement in overall survival was noted with the addition of RT. In Bocca’s series of 537 patients with supraglottic cancer (predominantly T2), local control was 94% for stage I and 82% for stage II. Overall 5-year survival was 78%.²¹⁸

Robbins and colleagues²²⁰ treated 139 T2 and T3 supraglottic SCCs with OSGL if technically possible, or with total laryngectomy or radiotherapy if not. Local control at 3 years was 100% for the OSGL group and 91% for the total laryngectomy group. Those subjects treated with primary RT had a 69% local control rate, which improved to 85% with salvage total laryngectomy. Five-year survival was 89% for those subjects treated with OSGL primarily, 78% with TL, and 70% with RT.

Transoral laser microsurgery for supraglottic carcinoma, first described by Vaughan in 1978, has become an accepted alternative to OSGL for supraglottic SCC.²²³ The indications for TLM are similar to those for OSGL (i.e., T1-2 and selected T3 tumors), although some institutions will use TLM for more advanced lesions such as T4.^224,225 Few contraindications to TLM exist: incomplete exposure of the tumor; tumor involving the great vessels of the neck; and tumor location and/or bulk requiring an extensive resection that would place the patient at high risk of aspiration (e.g., extensive tongue base involvement).²²⁵ The indications for neck dissection are not changed with TLM, and both sides of the neck may be dissected at the same time the primary is resected, or more commonly performed in a staged fashion several weeks after the initial TLM.²²⁶ The local control rates of TLM are similar to those of OSGL; however, the functional morbidity is much less following TLM because the extrinsic muscles of the larynx, pharyngeal muscles, cartilaginous framework of the larynx, and superior laryngeal nerves are left intact.²²⁷ The functional advantages of TLM include the possible avoidance of a temporary or permanent tracheostomy; less impairment of swallowing postoperatively including less aspiration; and a lower incidence of pharyngocutaneous fistulae.^224,228

Steiner’s group reported their experience with TLM for early supraglottic carcinoma in 1998.²²⁴ Only 4% of subjects were treated with adjuvant RT—96% underwent TLM only. No tracheostomies were performed at the time of the resection, although one patient required a temporary tracheostomy to manage aspiration. The average duration of nasogastric tube feeding was 6 days, and overall swallowing was excellent postoperatively. The oncologic results were equivalent to OSGL: 5-year local control rates were 100% for pT1 and 89% for pT2 tumors, which were salvaged with either a second TLM or with OSGL to give an ultimate local control rate of 97%. Overall 5-year survival was 76%.

Iro and colleagues²²⁹ reported the results of TLM in a series of 141 supraglottic cancers, which were predominantly T1-2. Half of the subjects also had neck dissections, and 45% received adjuvant RT. Tracheostomies were required in 13% of subjects during the course of treatment and were permanent in 9%. All but one of the subjects requiring a permanent tracheostomy were staged as pT4 and underwent extensive resection of the tongue base. Five-year local control rates were 86%, 75%, 75%, and 78% for stages I, II, III, and IV, respectively. Five-year, disease-free survival was reported as 66% (75% for stages I and II, 56% for stages III and IV).

Motta and colleagues²³⁰ performed TLM on 124 subjects with T1-3N0 supraglottic carcinoma and observed local control rates of 82%, 63%, and 77% for T1, T2, and T3 lesions. Laryngeal preservation rates were 89%, 85%, and 95%. Overall survival rates were 91%, 88%, and 81%. Advanced supraglottic tumors were also included in a series of TLM reported by Rudert and associates.²³¹ The majority of patients (24 of 34) received RT and nearly 20% required tracheostomies. Local control was 100%, 75%, 78%, and 37% for T1, T2, T3, and T4 tumors; overall survival was 71% for stages I and II and 50% for stages III and IV tumors.

The functional outcomes of patients who underwent either TLM or OSGL for predominantly T1-2 supraglottic SCC were compared by Peretti and colleagues.²²⁷ A temporary tracheostomy was required in 14% of TLM patients for an average of 4.5 days compared with 100% of OSGL patients who were decannulated after an average of 35 days. No permanent tracheostomies were required in either group. A feeding tube was used in 21% of the TLM patients for an average of 5 days, compared with 100% of OSGL patients, for an average of 19 days. No permanent feeding tubes were placed in either group. The average period of hospitalization was 11 days for the TLM group versus 26 days for the OSGL group. No significant difference in voice quality between the two groups occurred.

The role of adjuvant RT in TLM has been unclear, with up to 94% of subjects receiving adjuvant RT after TLM in some series.²³²

Zeitels and colleagues²³³ concluded that small T1-2 supraglottic lesions in subsites amenable to endoscopic resection (suprahyoid epiglottis, aryepiglottic fold, vestibular fold) could be treated successfully with TLM without RT but recommended that RT should be given after endoscopic resections of larger lesions (T2-3) in less favorable sites (e.g., infrahyoid epiglottis). In this series there was a high rate of local failure after RT in those patients in whom negative margins were not achieved at TLM, suggesting that RT was unable to control residual disease at the primary site. Davis and coworkers described the concept of TLM as a neoadjuvant treatment to remove all carcinoma at the primary site (with negative margins), followed by RT to treat any residual microscopic disease and treat the neck.²³⁴ Using this approach for T2 supraglottic SCC, TLM (and neck dissections for N+ subjects) was performed, with 83% of subjects receiving adjuvant RT. Over one third of subjects were staged as pT3 secondary to PES invasion. Local control was 100% in the surgery-only group and 97% in the combined-modality group. Regional control was 96% in the N0 subjects who received RT to the necks and 91% in the N+ subjects who received surgery and RT to the necks. Overall survival was low at 63% for the TLM/RT group and 50% for the TLM-only group due to a high rate of non–cancer-related deaths. A permanent feeding tube was required in 3%, but there were no permanent tracheostomies. The average time until normal swallowing was established was less than 14 days.

Because RT generally has an inferior local control rate compared with surgery, especially for bulky lesions, Agrawal and colleagues²³² proposed that a TLM procedure to debulk the primary tumor may improve the oncologic outcome. In a prospective trial of TLM with adjuvant RT for T1-2N0-1 supraglottic SCC, the 3-year local control rate was 97%, with one recurrence that was salvaged by TL. The overall survival rate was 88%. The functional results were not as good as in other TLM series, and this was probably due to the addition of the adjuvant RT: 21% of subjects experienced a prolonged period of dysphagia (3 to 10 months), and 9% required a permanent gastrostomy tube. Subject accrual for this trial became difficult as some coinvestigators were reluctant to use RT for all patients because they became increasingly comfortable in the use of surgery as single-modality treatment for these patients.

Radiation therapy plays an important role in the treatment of patients whose tumors are not amenable to partial laryngectomy, who are medically unfit for surgery, or who prefer to avoid surgery. In general, surgical excision has a higher local control rate for early supraglottic tumors than RT.²³⁵ Radiation therapy alone is also less effective than surgery (with adjuvant RT) or CRT in patients with advanced laryngeal carcinoma; however, it may be used to treat patients who are not eligible for CRT protocols but wish to attempt to preserve their larynx. However, RT is not without complications including dysphagia, aspiration, laryngeal edema, and chondronecrosis, which may require tracheostomy or total laryngectomy.²³⁵

Harwood reported initial local control rates with RT of 70% for T1N0 tumors, 68% for T2N0, and 54% for T3-4N0.²³⁶ Mendenhall and colleagues²³⁵ treated 209 supraglottic carcinomas with primary RT and obtained initial local control rates of 100%, 85%, 64%, and 36% for T1, T2, T3, and T4 lesions. Patients with recurrent disease were salvaged with either a total laryngectomy or, if possible, an OSGL, to give an ultimate local control rate of 100%, 88%, 81%, and 57%. The cause-specific survival for the series was 100%, 92%, 75%, 47%, and 32% for stages I, II, III, Iva, and IVb. Tumor volume (<6 cm³) and vocal cord mobility were significant predictors of local control. Hinerman and coworkers²³⁷ reported on 274 patients with supraglottic SCC treated with RT with or without neck dissection. The 5-year local control rate after RT was 100%, 86%, 62%, and 62% for T1, T2, T3, and T4 tumors. The 5-year cause-specific survival was 100%, 93%, 81%, 50%, and 49% for stages I, II, III, Iva, and IVb.

Primary RT was compared with partial laryngectomy (and neck dissection) by Spriano and colleagues²³⁸ for the treatment of 166 stage I and II supraglottic SCC. The surgical group had a superior 5-year, disease-free survival rate of 88% versus 76% for the RT group. Salvage surgery was effective in 66% of the partial laryngectomy failures and in 50% of the RT failures. Altered fractionation regimens have also been used to treat supraglottic SCC, and several reports have suggested improved local control rates with twice-daily hyperfractionated RT compared with once-daily RT.^239–241

Treatment of the Advanced Primary Tumor in Supraglottic Squamous Cell Carcinoma

Advanced primary supraglottic tumors (i.e., T3 or T4) have traditionally been treated with total laryngectomy, bilateral neck dissections, and postoperative RT. However, laryngeal preservation has become an important aim in the treatment of laryngeal cancer in an attempt to improve the quality of life of patients.²⁴² The therapeutic options for these advanced lesions currently include surgical excision (total laryngectomy or partial laryngectomy in selected cases) with bilateral neck dissections, followed by adjuvant RT (or concurrent CRT); primary RT with surgical salvage (usually total laryngectomy); or concurrent CRT with surgical salvage.²⁴³

The Department of Veterans Affairs (VA) Laryngeal Cancer Study, published in 1991, was a landmark trial in the development of nonsurgical organ preservation and chemotherapy for the treatment of advanced laryngeal cancer.²⁴⁴ Nearly two thirds of the subjects (208 of 332) had supraglottic SCC, with the majority of tumors classified as T3. The section on nonsurgical organ preservation later in this chapter describes the design and results of the important studies of nonsurgical treatment of advanced laryngeal cancer.

A second important trial investigating the nonsurgical treatment of advanced laryngeal SCC, the Head and Neck Intergroup Study (RTOG 91-11), was reported in 2003.¹⁹³ Over two thirds of the subjects (356 of 518) had supraglottic tumors. Large-volume T4 tumors that extended into the tongue base by more than 1 cm or penetrated through cartilage were excluded. The concurrent CRT group had statistically significant higher rates of locoregional control (78%) and laryngeal preservation (88%) compared with the induction chemotherapy– and radiation therapy–alone arms. The findings of this study have led to concurrent CRT becoming the standard nonsurgical organ preservation protocol for advanced laryngeal cancers with T2, T3, and low-volume T4 tumors (without gross cartilage destruction). Further information on nonsurgical preservation is discussed later.

The surgical excision of advanced primary tumors has traditionally been achieved by total laryngectomy, and although this remains the most commonly used procedure, conservative laryngeal resections such as OSGL, extended OSGL, supracricoid partial laryngectomy (SCPL), or TLM may be used for selected cases. Supracricoid partial laryngectomy, first described in 1959 by Majer and Rieder, is an organ-preserving surgical technique in which the aim is to preserve function without the need for a stoma of any kind.²⁴⁵ It may be used as a primary treatment or as a salvage procedure after RT.²⁴⁶ This procedure resects both true vocal cords, both false vocal cords, both paraglottic spaces, the entire thyroid cartilage, and the epiglottis. The hyoid bone and the cricoid cartilage are preserved. One of the arytenoids may also be included in the resection, but at least one intact cricoarytenoid unit must be preserved for postoperative function.²⁴⁵ Following the resection of a supraglottic cancer, the surgical defect is closed by tightly apposing the hyoid bone and base of the tongue to the cricoid cartilage, termed cricohyoidopexy (CHP). A SCPL (with CHP) is used for selected T2, T3, and T4 supraglottic and transglottic tumors: T2 tumors not amenable to OSGL because of involvement of the true vocal cords or anterior commissure, extension to the floor of the ventricle, and/or impaired motion of the true vocal cord; T3 transglottic and supraglottic tumors with true vocal cord fixation and/or pre-epiglottic space invasion; and T4 transglottic and supraglottic tumors with limited invasion of the thyroid cartilage without extension to the outer thyroid perichondrium, or extralaryngeal spread.²⁴⁵ Contraindications to SCPL include poor general health, poor pulmonary function, invasion of the cricoarytenoid joint, invasion of the cricoid cartilage, involvement of the posterior commissure, extension to the subglottis, invasion of the hyoid bone, extension of tumor to the outer perichondrium of the thyroid cartilage, or extralaryngeal spread.^214,247

The oncologic results of SCPL with CHP are excellent, with 5-year survival rates of 67% to 95% and local control rates of 88% to 95%.^245,248 Total laryngectomy is the recommended salvage procedure for recurrent disease. In general, laryngeal function is good after SCPL. A tracheostomy and feeding tube are required for all patients initially but are temporary in most.^246,249 A completion total laryngectomy may be required on occasion for intractable aspiration.^170,246 The major disadvantage of SCPL is poor voice quality.²⁴⁵

In summary, for T3 or T4a supraglottic tumors without extensive base of tongue involvement or cartilage destruction, either nonsurgical organ preservation (using concurrent CRT) or, in selected cases, conservative laryngeal surgery (with adjuvant RT) may be used primarily in an effort to preserve a functional larynx. Adjuvant RT or concurrent CRT may be required after primary surgical treatment, and total laryngectomy may be required as a salvage procedure after either nonsurgical or surgical organ preservation treatment. For extensive T4 tumors with gross cartilage destruction and/or extralaryngeal spread, total laryngectomy is the preferred initial treatment.

Treatment of the Neck in Supraglottic Squamous Cell Carcinoma

Supraglottic SCC has a high incidence of clinically apparent and occult regional metastases. Supraglottic SCC usually metastasizes to levels II, III, and IV. Bilateral metastases occur frequently in supraglottic SCC. Elective treatment of the clinically negative neck is indicated in all supraglottic SCCs except T1 lesions. The choice of either surgery or RT for the neck will vary according to how the primary tumor is to be treated. For early-stage tumors (T1-2), either surgery or RT may be used as a single modality to treat both the primary site and the neck. The indications for neck dissection do not alter when the primary tumor is treated with a partial laryngeal resection. For advanced primary lesions (T3-4), combined modality therapy is usually indicated. Often, nonsurgical organ preservation such as CRT or RT is used initially, with surgery to the primary and/or neck for salvage. Alpert and colleagues²⁵⁰ investigated the use of RT for the treatment of clinically negative necks in SCC: for patients staged N0, both necks were treated with RT; and for patients with ipsilateral nodal metastases, the ipsilateral neck was treated with surgery and the contralateral neck was treated with RT. RT was effective in this role because failure was observed in only 3% of node-negative necks.

The standard surgical treatment of the neck in N0 supraglottic SCC is bilateral selective neck dissection of levels II to IV. Several studies have observed that levels I and V are rarely involved by metastases when clinically apparent metastases are present in levels II-IV and are never involved when occult metastases are present in the lateral neck.^104,251,252 A selective neck dissection (levels II to IV) is as effective as a comprehensive neck dissection (levels I to V) for N0 (and N1) disease.²⁵³ Further studies of N0 supraglottic (and glottic) SCC have observed that metastases are not common in sublevel IIb and level IV, and therefore Ferlito and colleagues have suggested that a more targeted selective neck dissection of sublevel IIa and level III may be adequate for the elective surgical treatment of the neck in these cases.¹¹¹ An exception to these findings was a study that found occult metastases were present in level I in 82% of supraglottic SCC staged cN0.¹⁰⁸

Treatment of the N+ neck is less controversial. If surgery is the primary treatment modality, a selective neck dissection (levels II to IV) is as effective as a comprehensive neck dissection for N1 disease, but for more extensive nodal disease (N2-3), a comprehensive neck dissection should be undertaken.^253,254 Bilateral neck dissection has been found to decrease the regional failure rate of the surgical treatment of supraglottic SCC from 20% of cases (in which the contralateral undissected neck was the most common site for recurrence) to 8%.²⁵⁵ Adjuvant RT is indicated for neck dissections showing unfavorable pathologic findings including multiple involved nodes, extracapsular spread, or soft tissue extension of tumor.^209,218,256

In summary, for supraglottic SCC, surgical treatment of the N0 and N1 neck is usually with a bilateral selective neck dissection (levels II to IV). For N2 or N3 disease, a comprehensive neck dissection (levels I to V) is indicated for the N+ neck, with a contralateral selective neck dissection (II to IV) if the contralateral neck is N0. Primary nonsurgical treatment should include RT to both necks.

Nonsurgical Organ Preservation Treatment

A number of studies in the 1980s found that the addition of chemotherapy before RT for surgery for advanced head and neck cancer resulted in high rates of complete tumor regression associated with prolonged survival.^258–260 Response to chemotherapy predicted tumor radiosensitivity. However, this improved response did not translate into increased survival.²⁶¹

These early studies provided the rationale for the VA laryngeal cancer study reported in 1991.²⁴⁴ This was the first large prospective randomized study to demonstrate the effectiveness of the use of chemotherapy to define the more favorable patients and avoid laryngectomy in responders. This multi-institutional study reported on 332 patients with stage III or IV previously untreated laryngeal SCC. Subjects were randomized to receive either induction chemotherapy (IC) followed by RT, or surgery (total laryngectomy) and RT (Fig. 107-12). Two cycles of IC were given to the first group: nonresponders to the IC underwent immediate total laryngectomy, whereas responders received a third cycle of chemotherapy followed by RT (with salvage total laryngectomy for residual disease) (see Fig. 107-12). The laryngeal preservation rate was 64% for patients assigned to the IC/RT group, although only 39% had a fully functioning larynx. The 2-year survival for both groups was 68%. In a follow-up study, a further 10% decrease in the number of those with a functioning larynx was noted at 5 years, and there were no statistically significant differences in survival between the two groups.²⁶²

Figure 107-12. Schematic of the Department of Veterans Affairs Laryngeal Cancer Study.

In an effort to determine the relative contributions of chemotherapy and RT to laryngeal preservation, as well as to determine whether concurrent chemotherapy as a radiosensitizer would improve the likelihood of organ preservation, the Head and Neck Intergroup conducted a prospective randomized study that included 518 patients randomly assigned to three arms (Fig. 107-13).¹⁹³ Subjects received either IC followed by RT, concurrent CRT, or RT alone. Previously untreated stage III and IV glottic or supraglottic carcinomas requiring total laryngectomy for surgical treatment were eligible. This study excluded subjects with T1 tumors and bulky T4 tumors with laryngeal cartilage destruction, invasion through the laryngeal cartilages, or extensive tongue base invasion. At 2 years, the larynx was preserved in 88% of patients in the concurrent CRT group, compared with 75% and 70% in the IC/RT- and RT-only groups. The overall survival rate for all three groups was similar (5-year survival 54% to 56%). The high rate of laryngeal preservation in the concurrent CRT group was associated with an 82% incidence of high-grade toxicity compared with an incidence of 81% and 62% in the IC/RT and RT groups, respectively. In those who failed nonsurgical organ preservation therapy, total laryngectomy salvaged approximately three quarters of the patients.

Figure 107-13. Schematic of the National Cancer Institute Intergroup Laryngeal Cancer Study.

Adelstein and associates compared RT with concurrent CRT for stages III and IV laryngeal cancer and observed a significant difference in the recurrence-free interval (51% vs. 62%).²⁶³ The 5-year overall survival was similar for the two groups (48% vs. 50%), which the authors believed was due to the efficacy of salvage surgery and non–cancer-related deaths. Most other studies of concurrent CRT in the head and neck do not address the larynx specifically and do not have large numbers of laryngeal cancer patients.²⁴³ The utility of IC has been re-examined in a recent report by Urba and colleagues,²⁶⁴ in which a single cycle of chemotherapy was used to select patients with advanced laryngeal cancer for treatment with concurrent CRT. The 3-year, cause-specific survival and overall survival rates were 87% and 85%, respectively, with a laryngeal preservation rate of 70%.

Despite the encouraging results, nonsurgical organ preservation is not without disadvantages, especially with regard to its toxicity, which must be taken into consideration when deciding on treatment. In the RTOG 91-11 study, acute toxicity was more common and severe with concurrent CRT when compared with the other two groups and included chemotherapy-related side effects (e.g., neutropenia, nausea, vomiting) and increased rates of severe radiation induced mucositis.¹⁹³ The total rate of severe toxic effects, acute and late, was 82% in the concurrent CRT group. There was no difference between the three groups with regard to speech, with 3% to 8% reporting moderate or worse speech impairment at 2 years. However, the incidence of dysphagia was high: 26% of patients treated with concurrent CRT had problems swallowing at 1-year post treatment. At 2 years, there was no significant difference between the groups, with 15% of the concurrent CRT group reporting ongoing dysphagia. Previous RT and CRT may also pose problems for both the patient and the surgeon if surgery (usually total laryngectomy) is required after nonsurgical organ preservation for residual or recurrent disease, chondronecrosis, or intractable aspiration. In a subsequent report on the RTOG 91-11 trial, Weber and colleagues²⁶⁵ concluded that the morbidity of TL after organ preservation was acceptable but observed that up to 30% of patients developed a pharyngocutaneous fistula. The incidence was highest in the concurrent CRT group, but this difference was not statistically significant.

Therefore two important principles must be adhered to when nonsurgical larynx preservation treatment protocols are undertaken. First, if neoadjuvant chemotherapy or concurrent CRT is used, total laryngectomy must be performed for salvage in patients who fail to respond if survival is not to be compromised.²⁶⁶ Second, the success of CRT is dependent on the completion of the treatment protocol. The efficacy of RT is significantly diminished if treatment breaks are taken due to toxicity.²⁶⁷ Because nonsurgical larynx preservation therapy has significant toxicity (and mortality of 2% to 4%), patients with comorbidities and poor social support to help them through this challenging treatment regimen are less likely to complete the treatment and thus benefit from the addition of chemotherapy.

The success of nonsurgical larynx preservation treatment outside closely supervised clinical trials has also been questioned. Hoffman and colleagues⁶⁰ observed a concerning decrease in larynx cancer survival between 1985 and 2001, in patients with advanced glottic cancer, early supraglottic cancer, and T3N0M0 supraglottic cancer. During this same period there was a marked increase in use of nonsurgical larynx preservation treatment. Recently Chen and colleagues²⁶⁸ analyzed 7019 larynx cancer cases from the National Cancer Database and observed that total laryngectomy was associated with a statistically significant increase in survival compared with CRT for stage IV laryngeal cancer.

Stomal Recurrence

After total laryngectomy, recurrence of SCC around the stoma is a grave sign. Stomal recurrence is often insidious and only discovered after the disease has become extensive. Recurrent disease may be secondary to paratracheal node involvement, invasion of the thyroid gland, intraoperative tumor spill with implantation of cells in the stoma, or incomplete excision of tracheal invasion by subglottic spread. Subglottic SCC is most frequently associated with stomal recurrence.^297,298

The treatment of stomal recurrence is morbid and often unsuccessful, so prevention of stomal recurrence is paramount. When the risk of stomal recurrence is great (i.e., in primary subglottic SCC, glottic SCC with greater than 1 cm of subglottic extension, or T4 glottic tumors), the ipsilateral thyroid lobe is removed together with the larynx and bilateral paratracheal node dissections are performed. Adjuvant RT is also given, and the treatment field should include the upper mediastinum. In a large series reported by Rubin and colleagues,¹⁴⁶ the majority of stomal recurrences (12 of 15) were from primary subglottic tumors.

Dissection of the mediastinum and wide local resection for stomal recurrence was pioneered by Sisson, who proposed a four-category classification system (Fig. 107-14).²⁹⁹ Significant morbidity and mortality are associated with the surgical treatment of stomal recurrence including injury to the great vessels, mediastinitis, hypocalcemia, and fistula formation. Aggressive treatment of stage III and IV stomal recurrence is usually not indicated because the prognosis for these patients is poor.³⁰⁰ Sisson concluded that mediastinal failure was an important factor in advanced head and neck cancer and should be considered for T4 glottic and subglottic tumors in which mediastinoscopy was positive for nodes.²⁹⁹

Figure 107-14. Stomal recurrence. A, Stage I. B, Stage II. C, Stage III. D, Stage IV.

(From Sisson GA Sr. Ogura memorial lecture: mediastinal dissection. Laryngoscope. 1989;99:1264.)

Disease Factors

Patient Factors

Several factors independent of the stage or histologic characteristics of the tumor may affect the outcome of laryngeal SCC. Patient variables that have consistently been found to be important predictive factors include comorbidity (general health), age, and performance status. Comorbidity may be measured using one of several validated scales such as the Washington University Head and Neck Comorbidity Index (developed specifically for head and neck cancer patients)³¹⁹ and is an independent prognostic factor for patients with laryngeal cancer.^329,330 Advanced age (older than 65 years) is associated with a worse prognosis for SCC of the head and neck including laryngeal carcinoma.^302,319 Patients with a higher performance status, as measured by scales such as the Karnofsky Performance Status, have a more favorable outcome.^331,332 The significance of gender as a prognostic factor is controversial, with some reports finding no difference^319,331 and others reporting a worse prognosis for women with laryngeal cancer.^333,334 For patients undergoing RT for laryngeal SCC, a normal hemoglobin level is associated with increased overall survival.³³⁵

Molecular Markers of Prognosis

Clinicians and researchers alike are searching for molecular markers that will allow the identification of patients with laryngeal cancer who have adverse prognostic features so that more aggressive and successful treatment may be offered. The most commonly mutated gene in human cancer including laryngeal cancer is p53; however, its prognostic value in laryngeal cancer is uncertain.^336–338 Epidermal growth factor receptor (EGFR) overexpression has been found to be a marker for aggressiveness, nodal metastasis, and survival.^338,339 EGFR overexpression is also associated with radioresistance and poor local control of laryngeal cancer,³⁴⁰ and therefore a tumor’s EGFR status may be used to predict radiosensitivity and chemosensitivity.³³⁸ Vascular endothelial growth factor (VEGF) positivity is associated with increased overall survival in patients irradiated for laryngeal SCC.³⁴¹ Cyclin D1 and cyclin D3 overexpression are reported to be significant predictors of disease-free survival, and determining the cyclin D1/D3 coexpression status of laryngeal carcinomas has allowed stratification of patients into prognostic groups, with patients overexpressing both cyclins having the worst prognosis.³⁴² Markers of epithelial differentiation such as type 2 cyclo-oxygenase (Cox-2) and galectin-3, may also have predictive value. Low expression of Cox-2 in laryngeal SCC is associated with aggressive biologic behavior (including invasion), whereas galectin-3 expression is associated with improved overall survival.³³⁸ At present, much of the data regarding molecular markers are conflicting and no reliable prognostic molecular markers exist for clinical use.