Subglottic carcinoma: Review of a series and characterization of its patterns of spread.

From the Department of Otolaryngology-Head and Neck Surgery, Mayo Clinic, Rochester, Minn. (Dr. Strome); the Department of Otolaryngology-Head and Neck Surgery, University of Michigan Medical Center, Ann Arbor (Drs. Robey, Krause, and Hogikyan); and the Department of Pathology, University of Michigan Medical Center, Ann Arbor (Dr. Devaney).

[When this article was written, Dr. Strome was with the Department of Otolaryngology-Head and Neck Surgery, University of Michigan Medical Center, Ann Arbor.]

Reprint requests: Scott E. Strome, MD, Department of Otolaryngology- Head and Neck Surgery, Mayo Clinic, 200 First St. SW, Rochester, MN 55905. Phone: [phone]; fax: [phone].

Abstract

The rarity of primary subglottic malignancies, along with the varied definitions of the anatomic confines of this region, have limited our understanding of the patterns of tumor spread within the sub glottis. We conducted a retrospective chart review to analyze clinical and pathologic data in patients with subglottic carcinoma. A pattern of disease progression was identified, which is defined by the cartilaginous laryngeal framework, with the fibroelastic barriers susceptible to tumor invasion.

We conclude that although cartilaginous laryngeal structures are preserved until late in the disease course, the ability of tumors to invade the fibroelastic membranes provides them with an insidious means of escape. Specifically, tumor progression occurs primarily within the paraglottic space and extralaryngeal compartments; the potential for mucosal spread is limited. The lack of mucosal disease in patients whose cartilaginous laryngeal structures are intact may present a facade of normality in patients with advanced disease, and perhaps delay the early diagnosis of subglottic malignancies by physical and radiologic examination.

Introduction

Primary subglottic malignancies are rare. They represent only 1 to 8% of all laryngeal cancers. [1] Our understanding of the mechanisms of tumor spread in this area is limited by the rarity of primary malignancies in this region and by the lack of a uniform definition of the subglottic space. The purpose of this paper is to further explore the patterns of subglottic tumor spread by describing the findings of our retrospective review of clinicopathologic data on patients with primary subglottic carcinoma who presented to the University of Michigan Medical Center between 1964 and 1994. A knowledge of the patterns of tumor progression enhances the potential for earlier diagnosis and allows for more accurate staging and more effective treatment of the primary disease process.

While the inferior boundary of the subglottis is uniformly defined as the bottom of the cricoid cartilage, the definition of the superior boundary is less clear. The literature contains three different definitions of the superior limit of the subglottic space: 1) the inferior border of the true vocal folds (TVFs), 2) 5 mm below the free edge of the TVFs, and 3) 1 cm below the apex of the ventricle. [2,3] While the third definition is the one currently accepted by the American Joint Committee on Cancer, various investigators have used the others in studying these lesions. This lack of uniform reporting has hindered the study of this disease process.

Work by Pressman [4] and Kirchner [5-8] indicates that the supraglottic and glottic larynx are compartmentalized by fibroelastic membranes. Although these structures do play a role in defining the patterns of tumor spread, the current study suggests that they do not form reliable barriers in the subglottis. Even in cases of limited mucosal disease and an intact cartilaginous laryngotracheal infrastructure, paraglottic-space extension and spread of disease beyond the laryngotracheal confines do occur. The relative sparing of the regional cartilage in combination with minimal mucosal abnormalities may hinder early diagnosis and result in the understaging of clinically advanced disease.

Materials and methods

We performed a retrospective chart review of the records of all patients who were treated for primary subglottic malignancy at the University of Michigan Hospital during the 30-year period from 1964 to 1994. Only those patients who had presented with primary squamous cell carcinoma in the subglottic space--which we defined as the area 1 cm below the apex of the ventricle superiorly to the lower border of the cricoid inferiorly--were included in this study. Patients whose carcinomas extended into the subglottic space from elsewhere in the larynx were also excluded. Ten patients met our criteria for inclusion. All had a minimal followup of 2 years.

We noted each patient's age at diagnosis, race, sex, presenting signs and symptoms, social history, suspected diagnosis at the time of presentation, and the method of definitive diagnosis. We reviewed clinic notes and operative reports to identify the exact location of the primary tumor in the subglottic space. Tumor staging was performed in accordance with standard TNM (tumor, nodes, and metastases) criteria. One patient whom we could not stage because of incomplete records was included in the study on the basis of available demographic data, symptoms at presentation, and followup. We reviewed pathology reports and recorded tumor histology and pathological stage. Additionally, we retrieved and reviewed surgical pathology from patients who had undergone total laryngectomy in an attempt to define the histologic patterns of disease progression. Histopathology slides were available for 2 of the 10 patients. Of the remaining 8 patients, 7 had been treated with primary radiotherapy, including 1 patient who had undergone surgical salvage. Archival specimens were unavailable for the patie nt who had undergone surgical salvage and for the remaining 1 patient. We noted any mention of cartilage, thyroid, carotid, or trachea invasion and lymph node metastasis.

We also analyzed the treatment of the primary tumor, including adjuvant therapy. We reviewed the incidence of local recurrences, synchronous/metachronous distant metastases, and second primary tumors and the treatment modalities that were employed. Finally, we noted the disease status of each patient at their last followup and the cause of death.

Results

Demographic information on the 10 patients included in the study are presented in table 1. The number of men and women was equal, and their history of alcohol and tobacco use was similar to that observed in other reports of head and neck malignancies.

Symptoms at presentation included voice change in 7 patients, shortness of breath in 4, and hemoptysis in 3 (table 2). Five of the 6 patients for whom tumor descriptions were available had experienced tumor extension into the inferior aspect of the TVF. Overall, 5 patients were noted to have had TVF paralysis on clinical examination--3 unilateral and 2 bilateral. These findings suggest either deep thyroarytenoid muscle or cricothyroid joint involvement.

Information on tumor staging at the time of the initial diagnosis is presented in table 3. All 3 patients who had presented with stage II disease had involvement of the inferior aspect of the TVF without evidence of motion impairment. Five patients ultimately required a tracheotomy during therapy, 4 of them on an emergency basis. Only 1 patient had presented with a lymph node metastasis at the time of initial diagnosis, and this patient had locally advanced disease.

We performed a pathologic review of the archived histologic specimens from 2 patients who had undergone total laryngectomy as their primary treatment. One of these patients had previously failed induction chemotherapy. Twenty-four slides were available for study in one case, and 35 slides in the other. In both specimens, the striking low-power feature was the maintenance of intact outlines of most of the cartilaginous rings; the tongues of invasive carcinoma largely spared the islands of cartilage, seeming to preferentially invade through the spaces separating these cartilaginous islands (figure 1). Although minor erosions of cartilage by invasive tumor were found on close scrutiny, the bulk of the chondroid tissue appeared to have been resistant to the advancing bulk of the tumor. In one of these patients, histology showed that the TVF was uninvolved by the tumor (figure 2), although the anterior soft tissues and skeletal muscle at the same level as the vocal fold were largely replaced by invasive carcinoma (figure 3).

Survival as related to the tumor stage and the type of treatment employed is presented in table 4. Although the number of patients we evaluated is too small for meaningful survival analysis, the fact that only 1 of the 5 patients who had stage II disease or less and who had been treated with primary radiation therapy died of their disease within 5 years suggests that irradiation may be a viable treatment option in early-stage disease.

Four patients ultimately developed recurrent disease--1 at the primary site, 1 at the primary site and the neck, 1 at the primary site, neck, and distantly, and 1 with persistent disease. It is interesting that the 2 patients who developed secondary nodal disease had had a synchronous recurrence at the primary site. Four patients ultimately developed a second primary tumor--l cervical squamous cell carcinoma, 1 meningioma, 1 ampullar carcinoma, and 1 synchronous primary tumor of the lung.

Discussion

The principal conclusion that we derive from this study is that squamous cell carcinoma of the subglottic region is a locally aggressive neoplasm, and that fibroelastic barriers within the subglottis are susceptible to local tumor spread. Pathologic analysis of archived histologic specimens under low-power magnification revealed intact outlines of most of the tracheal rings and tongues of invasive carcinoma preferentially invading through the spaces that separate these cartilaginous islands. Although we found some minor erosions of cartilage by invasive tumor on close scrutiny, the bulk of the chondroid tissue appeared to have been resistant to the advancing bulk of the tumor. One specimen revealed that the TVF was uninvolved by the tumor, although the paraglottic-space tissues at the same level as the vocal fold were largely replaced by invasive carcinoma.

Analysis of these data suggests a possible mechanism for the spread of subglottic squamous cell carcinoma. The tumor appears to respect the cartilaginous laryngeal barriers until late in the disease process. Subglottic neoplasms enter early into the paraglottic space and spread within the region bounded medially by the conus elasticus and laterally by the hyaline laryngeal cartilages. Because the conus elasticus, cricothyroid, cricotracheal, and intertracheal membranes do not provide a reliable barrier to tumor spread, squamous cell carcinoma has the potential to enter the prelaryngotracheal space, often without manifesting overt clinical signs.

Support for this model is provided by the relationship of our patients' signs and symptoms to the descriptions of their tumor appearance at presentation. Four of the 5 patients who had early-stage disease presented with hoarseness. Patients with advanced disease often present with airway distress. Tumor descriptions revealed an incidence of bilateral TVF paralysis in 2 of the 5 patients with stage III or IV disease who presented with shortness of breath. The combination of these signs and symptoms in the patients with advanced disease and limited mucosal involvement is further evidence of significant paraglotticspace involvement.

In his study of whole-organ sections of laryngeal carcinomas, Kirchner demonstrated that although these tumors can spread via vascular or perineural routes, they spread primarily via direct extension. [6,7] The current study reveals that while the intrinsic fibroelastic barriers (eg, the conus elasticus) limit the spread of tumor in its earliest stages, tumors can spread freely within the defined space once these barriers are violated. Whole-organ section would be required to verify this model. However, the limited number of patients who have this site-specific disease process, combined with the lack of tissue availability of patients treated with radiotherapy, make whole-organ section difficult to perform. In fact, in Kirchner's study of 100 whole-organ sections, only 3 were subglottic primaries, and 1 of these patients had received previous radiation therapy. [7] The strong theoretical basis for this model in light of the anatomic and symptomatic descriptions of subglottic primary tumors in patients with varying stages of disease renders this model a plausible and useful tool for the clinical study of subglottic malignancy.

This anatomic model for the spread of subglottic tumors has three points of clinical relevance. First, because mucosal disease does not appear to play a significant role in this disease process, the diagnosis of primary subglottic malignancy is often delayed. This point is emphasized by a study performed by Berger et al, in which whole-organ sections were performed on the larynges of patients who presented with presumed T1N0M0 squamous cell carcinoma of the glottis and who failed primary radiation therapy. [1] This analysis revealed that 9 patients who had a primary subglottic tumor with secondary TVF involvement were erroneously diagnosed initially as having a primary TVF tumor. Because of the difficulty in diagnosis, these patients have historically presented with advanced disease and its attendant poor prognosis. Recognizing the potential for limited mucosal involvement in this disease process, it is incumbent upon the otolaryngologist to maintain a high index of suspicion for primary subglottic neoplasms in patients who present with unexplained voice change, respiratory distress, or TVF paralysis.

Second, because subglottic carcinomas appear to respect the boundaries of the intrinsic laryngotracheal cartilaginous infrastructure, imaging techniques may reveal no evidence of cartilaginous invasion, even in patients with advanced disease. This observation is at odds with the findings of Saleh et al, who reported that cricoid cartilage destruction was detected by computed tomography (CT) in 9 of 14 patients who had primary subglottic carcinoma. [9] Histologic confirmation of the CT findings was not provided. Therefore, although imaging studies may detect cartilage invasion in some patients, it is imperative that the physician recognize the propensity for the direct extension of subglottic carcinoma into the neck without causing any underlying cartilage abnormality.

Third, only 1 of our 10 patients with squamous cell carcinoma had lymph node metastasis at the time of presentation, and this patient had locally advanced disease. Moreover, the 2 patients who subsequently developed regional nodal metastasis also had a synchronous recurrence at the primary site. Thus, while Harrison argues that nodal metastases are often missed because of their location in the paratracheal and mediastinal region, the presence of primary nodal disease in only 10% of patients suggests that nodal dissection is necessary only in patients who have clinical evidence of regional metastases and in those who present with advanced or recurrent local disease. [10]

While symptoms varied, 70% of our patients experienced voice changes and 40% were short of breath. Five of the 6 patients for whom accurate clinical tumor descriptions were available experienced tumor extension into the inferior aspect of the TVF. These data are in accordance with Stell and Tobin's series, in which there was a 50% incidence of TVF fixation at the time of presentation. [11]

Previous studies have found that patients who are treated with primary surgery have higher survival rates than those who are treated with radiation. However, of the 4 evaluable patients in our study who were treated with primary radiation, only 1 had evidence of recurrence within 5 years. While the small number of patients in our study clearly precludes statistical evaluation, radiotherapy should be considered as a potential treatment option for patients with early-stage disease, given the slow growth rate of these tumors and their defined patterns of spread.

In addition to helping define the patterns of spread of subglottic tumors, this study serves to emphasize the need for aggressive airway management in this patient population. Five of the 10 patients in our series required an emergency tracheotomy for progressive airway obstruction, and another patient required a planned tracheotomy at the time of laryngoscopy. Previous studies have identified CT as an important component of the preoperative airway management of these patients because this may be the only means to adequately evaluate inferior tumor extension and thus avoid tumor spillage during tracheotomy. [9] Shapshay et al provide an algorithm for airway management in patients with tumors in this region that includes consideration of awake fiberoptic intubation and awake intubation with a ventilating tracheoscope for airway control. [2] Regardless of the method of airway control, the high incidence of airway obstruction in this patient population mandates careful preoperative consideration of optimal airw ay management for each patient.

The demographics, tumor histology, treatment, and survival of the patients reported in our study did not significantly differ from those found in previous reports. It is interesting that 50% of the patients in our study were women, a gender trend noted by previous investigators. It is unclear why a greater percentage of women have subglottic tumors than other squamous cell malignancies of the upper aerodigestive tract.

In conclusion, this paper brings to light several important features of subglottic carcinoma. Although it is a retrospective review, the low overall incidence of subglottic carcinoma makes this series of particular value. Specifically, because mucosal disease does not appear to play an important role in the disease process, the diagnosis is often delayed. Second, because subglottic carcinoma seems to respect the cartilagenous boundaries of the larynx and escape through the intercartilagenous infrastructure, imaging studies may reveal no evidence of invasion, even in the face of advanced disease. Finally, the low incidence of nodal metastasis in the paratracheal and mediastinal regions implies that resection of these compartments is not as important as the literature suggests. Irrespective of the treatment modality, airway considerations remain paramount in this population, given the high incidence of airway compromise.

References

(1.) Berger G, Harwood AR, Bryce DP, van Nostrand AW. Primary subglottic carcinoma masquerading clinically as T1 glottic carcinoma: A report of nine cases. J Otolaryagol 1985;14:1-6.

(2.) Kennedy KS, Parker GS. Subglottic and tracheal malignancies. Ear Nose Throat J 1992;71:233-7.

(3.) Shaha AR, Shah JP. Carcinoma of the subglottic larynx. Am J Surg 1982;144:456-8.

(4.) Pressman JJ, Dowdy A, Lobby R, Fields M. Further studies upon the submucosal compartments and lymphatics of the larynx by the injection of dyes and radioisotopes. Ann Otol Rhinol Laryngol 1956;65:963.

(5.) Kirchner JA, Carter D. Intralaryngeal barriers to the spread of cancer. Acta Otolaryngol (Stockh) 1987;103:503-13.

(6.) Kirchner JA. Two hundred laryngeal cancers: Patterns of growth and spread as seen in serial section. Laryngoscope 1977;87:474-82.

(7.) Kirchner JA. One hundred laryngeal cancers studies by serial section. Ann Otol Rhinol Laryngol 1969;78:689-709.

(8.) Kirchner JA. Invasion of the framework by laryngeal cancer: Surgical and radiological implications. ActaOtolaryngol (Stockh) 1984;97:392-7.

(9.) Saleh EM, Mancuso AA, Alhussaini AA. Computed tomography of primary subglottic cancer: Clinical importance of typical spread pattern. Head Neck 1992;14:125-32.

(10.) Harrison DF. The pathology and management of subglottic cancer. Ann Otol Rhinol Laryngol 1971;80:6-12.

(11.) Stell PM, Tobin KE. The behavior of cancer affecting the subglottic space. Can J Otolaryngol 1975;4:612-7.

(12.) Shapshay SM, Ruah CB, Bohigian RK, Beamis JF Jr. Obstructing tumors of the subglottic larynx and cervical trachea: Airway management and treatment. Ann Otol Rhinol Laryngol 1958;97:487-92.

                   Demographics of the study population

Gender (n=10)     Men: 5   Women: 5
Race (n=9)        White: 8 Black: 1
Tobacco use (n=9) Yes: 8   No: 1
Alcohol use (n=8) Yes: 5   No: 3


n = No. evaluable patients
                  Presenting symptoms of 10 patients with
                       primary subglottic carcinoma

Symptom             No. of Patients
Voice change               7
Shortness of breath        4
Hemoptysis                 3
Cough                      1
No symptoms                1
         Clinical involvement of regional structures, categorized
              by stage of primary disease at presentation [*]

                Site of Tumor Invasion
                 Thyroid                               Thyroid
                Cartilage              Trachea Carotid  Gland  Nodes
Stage
I/in situ (n=2)     0                     0       0       0      0
II (n=3)            0                     0       0       0      0
III (n=0)           -                     -       -       -      -
IV (n=4)            3                     3       1       3      1

(*.)Patients with early-stage disease did not manifest clinical evidence of local desease spread, yet TVF involvement was common. (Only 9 patients are included in this analysis because information on the initial tumor stage was not available for 1 patient.)

               Survival analysis of 9 evaluable patients [*]

Stage at            Treatment of    Survival
Presentation        Primary Tumor   Analysis
I/in situ (n=2)     Irradiation     1 NED [+]
                                    1 DOD [ss] 28 years
II (n=3)Irradiation 1 NED 14 months
                                    1 DOC [n] 6 months
                                    1 DOD 18 months
III (n=O)           -               -
IV (n=4)            1 Irradiation   1 DOD 8 months
                    3 Surgery/      3 NED [greater than]5 years
                     irradiation

(*.)Only 1 of the 5 patients with stage I or II disease died during the first 5 years of follow-up; all had undergone irradiation as the treatment of choice, (Only 9 patients are included in this analysis because information on the initial tumor stage was not available for 1 patient.)

(+.)No evidence of disease.

(ss.)Dead of disease.

(n.)Dead of other causes.