Abstract
BACKGROUND: Appropriate management of the clinically negative (N0) neck in supraglottic laryngeal cancer continues to be an area of controversy in head and neck surgery. Our treatment policy has been aggressive surgical management even in the clinically N0 neck.
METHODS: Between 1971 and 1991, 104 patients had the primary diagnosis of supraglottic laryngeal cancer. Ninety of these patients received their treatment at Roswell Park Cancer Institute and are the subject of this retrospective review.
RESULTS: All neoplasms included in this study were squamous cell cancers. The most common subsite involved with tumor in our series was the epiglottis, followed by the aryepiglottic folds and false cords. Supraglottic laryngectomy was performed of 29% of the cases; the remainder received total laryngectomy. Thirty-six percent of the patients had pathologic stage I/II disease, and 64% had stage III/IV. The 5-year survival rates were 100%, 81%, 73%, and 63% for stages I through IV, respectively. Fifty-seven patients had clinically N0 disease at presentation; of these 34 underwent elective neck dissection, and the remaining 23 patients were observed. Of those patients receiving neck dissection, 30% (n = 10) were found to have histologically positive disease, and of the 23 patients observed, 30% (n = 7) had histologically positive regional (neck) disease. Of the 17 clinically N0 and pathologically N+ patients, 82% (14 of 17) had involvement of level I (submandibular triangle), and 100% had involvement of level II. The incidence of bilateral disease in the clinically N0 patient was 44%. There were no local failures.
CONCLUSIONS: There is a high incidence of occult regional disease even in early-stage supraglottic squamous cell carcinoma of the larynx. In the surgical management the clinically N0 neck, we presently recommend bilateral neck dissection of levels I through IV to adequately address those regions at highest risk for occult disease. (Otolaryngol Head Neck Surg 1999;121:57-61.)
The surgical management of the clinically negative (N0) neck in supraglottic laryngeal carcinoma remains controversial. There have been divergent opinions as to the extent of dissection necessary to effect an oncologically satisfactory neck dissection.
The role of organ-sparing protocols and primary radiation therapy (RT) in the treatment of supraglottic laryngeal cancer is well recognized. However, in patients who are not candidates for or who refuse organ-sparing protocols and in patients in whom organ-preserving protocols fail, surgery remains a viable treatment option.
The pattern of nodal spread from laryngeal carcinoma has been described in several series. 1 – 5 These reports have provided the basis for the use of selective neck dissection in many centers.
From 1971 to 1991 at Roswell Park Cancer Institute (RPCI) there was a bias toward surgery as the sole modality of treatment for supraglottic laryngeal cancer. To assess the pattern of nodal metastasis in patients with clinically N0 necks, we reviewed our experience with 104 patients treated with primary surgical treatment. Our objective was to evaluate those areas (levels) of the neck most commonly involved with occult or recurrent disease.
METHODS AND MATERIAL
Between 1971 and 1991, 104 patients had the primary diagnosis of supraglottic laryngeal cancer; 90 of these patients received their primary therapy at RPCI and are the subject of this report. Approval for this study was obtained from the institutional review board at RPCI. All patients were operated on by the attending staff of RPCI's Department of Head and Neck Surgery. Age, site of primary, T stage, and location and number of nodal metastasis were obtained from hospital records, attending office notes, and the tumor registry. Survival curves were obtained with the Kaplan-Meier method. The log-rank test and Fisher's exact test were used to determine statistical significance.
Clinical stage distribution supraglottic laryngeal cancer
70125-7-table1.png)
Pathologic stage distribution supraglottic laryngeal cancer
70125-7-table2.png)
Survival rate
70125-7-table3.png)
A diagnosis of supraglottic laryngeal cancer was made in 63 men (70%) and 27 women (30%). The median age was 59 years. None of these patients received treatment for their neoplasm, aside from endoscopy and biopsy, outside of RPCI.
For the purposes of this review all tumors were staged according to the 1992 American Joint Committee on Cancer criteria. The clinical and histologic staging results are presented in Tables 1 and 2. On the basis of the pathologic findings, the stage distribution for this review included stages I (9 patients), II (23), III (29), and IV (29). The epiglottis was the most common primary site in our series (53%), followed by the aryepiglottic folds and ventricular bands (39%). The epicenter of the tumor could not be classified in the remaining (8%) cases. All tumors were squamous cell cancers.
Supraglottic laryngectomy was performed in 29% (n = 26) and total laryngectomy in 70% (n = 63) of our patients. One patient underwent epiglottectomy. Surgical margins of resection at the primary site were negative in 69 of 90 patients (77%), close (≤5 mm) in 15 (17%), and positive in 5 (6%). No patients received RT as the primary modality of treatment. Fourteen patients (16%) had planned postoperative adjuvant RT. Adjuvant radiotherapeutic treatment consisted of a wide-field technique treating both sides of the neck and the primary site. A minimum dose of 5000 rad was delivered during a 6-week period.
The classification system developed by the American Society of Head and Neck Surgery and the Committee for Head and Neck Surgery and Oncology within the American Academy of Otolaryngology-Head and Neck Surgery 6 was used to determine the level of nodal involvement. Most patients had nodal mapping based on lymph node clearing of the neck specimens, enabling us to accurately assess the location of regional metastatic disease within the neck specimens.
RESULTS
The 5-year overall survival rate for these patients was 100% for stage I, 73% stage II, 81% stage III, and 63% stage IV (Fig 1). The range of follow-up was 12 to 240 months. Patients with close or positive surgical margins were followed up for a minimum of 36 months; none of these patients had local failure.
There were no local and 21 regional failures in our series. Of the 21 patients with regional failures, 2 had pulmonary metastases at the time of presentation. All 19 remaining patients were salvaged by surgery. Nine of these failures were in patients whose tumors were originally staged as clinically N0.
In the group of patients treated with surgery alone for stage III/IV disease, there was a 29% regional recurrence rate, as compared with those patients receiving surgery with planned postoperative RT, in whom regional recurrence was only 15% (Table 3). Eighty-five percent of the patients with planned postoperative RT had stage III/IV disease at presentation. The 5-year survival rate in those patients receiving surgery alone was 64%, as compared with patients undergoing surgery with planned postoperative RT, in whom the 5-year survival rate was 82% (Fig 2). The occult regional metastatic rate was 26% for early-stage (I/II) and 33% for late-stage (III/IV) disease.
Fifty-seven patients within our series had clinically N0 necks. Thirty-four (60%) of these patients received elective neck dissection, and 23 (40%) were observed. Of the patients who received elective neck dissections, 30% (n = 10) were found to have pathologically involved cervical lymph nodes. Of the patients with clinically N0 necks who were observed, regional disease developed in 7 (30%). Overall, 17 patients (30%) with N0 disease were found to have, or subsequently develop, regional nodal disease.
70125-7-fig1.png)
Supraglottic laryngeal cancer survival by stage.
70125-7-fig2.png)
Survival rate: Surgery versus surgery plus RT in stage III and IV tumors.
As detailed in Fig 3 metastatic nodes were found in all levels of the neck. Of the 17 patients who had clinically N0 necks but pathologically positive nodes, 14 (82%) had involvement of the submandibular triangle (level I as defined by Robbins et al 6 ). This included 23% of the patients (n = 4) with nodal disease adjacent to the anterior belly of the digastric muscle and 59% (n = 10) with disease confined to the posterior belly of the digastric muscle. All patients (100%) with pathologically positive nodes had level II involvement. In the patients with clinically N0 necks, levels III, IV, and V were pathologically positive in 35%, 24%, and 12%, respectively. The incidence of bilateral disease in patients with N0 necks was 44%. Contralateral nodal involvement mirrored the ipsilateral dissected neck, with levels I through III found to be pathologically positive. Seven percent of the patients with N0 necks in our series had isolated N1 disease.
Regional failure in N0 necks occurred in 9 patients. Four patients had recurrences in the ipsilateral neck and 5 in the contralateral neck. The contralateral neck recurrences were in levels I, II, and III.
DISCUSSION
There were no local failures in our series, although 17% of our patients had close margins of resection (≤5 mm) and 6% had positive margins. None of the patients with close or positive margins received postoperative RT. We first attributed our successful local control rate to the fact that many of our patients received total laryngectomy. However, closer analysis of this cohort of patients demonstrated that all of the close or positive margins were at either the tongue base or pharyngeal mucosal margin. This suggests that our good local control was not a result of the type of surgery performed on the primary lesion but may be a result of the pathophysiology of the disease process.
70125-7-fig3.png)
Patterns of cervical metastasis of clinically NO patients. (Modified from Robbins KT, Medina JE, Wolf GT, et al. Standardizing neck dissection terminology. Arch Otolaryngol Head Neck Surg 1991;117:601-5.)
There is a high incidence of occult metastatic neck disease in supraglottic laryngeal cancer independent of the T stage of the primary lesion. Thirty-four patients with clinically N0 disease, 30% of whom had pathologically positive necks, received elective neck dissections. Sixty percent of these patients with positive necks had early (T1/T2) disease. There was a trend toward increased 5-year survival advantage in those patients with N0 necks who received prophylactic neck dissections (64% vs 50%, P = not significant). These findings corroborate the conclusions of other authors 1 , 7 who state the potential survival benefit in those patients who undergo elective treatment of regional disease in supra-glottic laryngeal cancer. Although our institution has had a surgical bias in the treatment of N0 regional disease, we consider RT an effective modality in the prophylactic treatment of the N0 neck.
In patients in whom surgery is chosen for treatment of the primary and regional disease, the anatomic levels to be included in the neck dissection remain controversial. Between 24% and 49% of the patients with supraglottic laryngeal cancer have either occult or clinically apparent cervical nodal metastasis. 2 , 8 – 11 Recent trends in head and neck surgery have focused on selective neck dissections to reduce the morbidity associated with radical neck dissections, while still maintaining oncologically sound principles. This is achieved only when the nodal groups at risk are addressed at the time of selective lymphadenectomy. 3 , 4 Inclusion of the submandibular triangle (level I) in selective lymphadenectomy for supra-glottic cancer, as defined by the anterior and posterior bellies of the digastric muscle, is controversial. Wenig and Applebaum 4 stated that 70% of nodal metastases in laryngeal cancer were limited to levels II and III. Similarly, Candela et al 5 found level I was involved in only 5% of the N0 necks and 8% of the N+ necks. Pietrantoni and Fior 12 stated, on the basis of their clinical review, that the submandibular triangle should not be dissected in the treatment of laryngeal cancer.
In our review of 57 patients with N0 supraglottic laryngeal cancer, level I was involved with regional disease in 82% of the patients with cervical nodal metastases. This finding represents a significant departure from previously reported series. There are several potential explanations for this difference. First, the extent of level I dissection may vary between institutions. Second, the pathologic processing of the dissected nodal basins may vary. At RPCI we have traditionally asked our pathologists to use lymph node-clearing techniques, which may increase the number of nodes identified. Each specimen was mapped to accurately represent the location of the nodes to vascular and muscular structures. Third, the definition of neck levels has changed from that of earlier publications. To avoid any inconsistencies in our data analysis, we have used the classification accepted by the American Society of Head and Neck Surgery. 1
Many centers use selective neck dissections as staging procedures in the treatment of the N0 neck, 4 , 5 , 13 saving RT for the treatment of patients with pathologically positive disease. There are several potential problems with this approach if all at-risk nodal levels are not dissected. First, although the incidence of N1 disease in our series is relatively low, there are patients who would receive postoperative RT unnecessarily, because of incomplete pathologic information from the clinician from the selective neck dissection, which includes only levels II, III, and IV Second, we believe it technically challenging to accurately define the separation of lymphatics between posterior level I and superior level II. Unlike some previous authors, we have found a high incidence of level I disease in our series of patients and a significant pathologic and clinical overlap between posterior level I and level II.
Surgical management of the N0 neck may be indicated in patients who refuse, do not qualify for, or experience failure of organ-sparing protocols. The surgical approach to the N0 neck may also be indicated in patients in whom primary RT or previous nonsurgical management fail. In this cohort of patients we presently recommend a selective bilateral neck dissection of levels I, II, III, and IV. We reserve postoperative RT for patients with multiple positive nodal levels and/or extracapsular spread. 14 , 15 This approach provides the most thorough prognostic information, provides excellent local control, and ensures surgical assessment of level I of the neck, a site that has been neglected in the treatment of supraglottic laryngeal carcinoma.
CONCLUSIONS
Because of the high incidence of occult metastatic disease in supraglottic laryngeal cancer even in early-stage disease, we recommend aggressive management of regional lymphatics. The successful treatment of regional disease in the N0 setting can be accomplished by primary surgical or RT. Because of the high incidence of disease in level I of our series, we presently recommend surgical management of regional lymphatics and dissections of levels I through IV.