Introduction

The management of the N0 neck in supraglottic laryngeal carcinoma (SGLC) remains controversial. There is a strong debate on how to treat the neck in patients with SGLC and N0 neck. The supraglottic part of the larynx is characterized by a rich lymphatic network and carcinomas of this region have a strong predilection for cervical metastasis. The incidence of metastasis has been reported to vary from 12 to 52% approximately, 30% of them being occult [1]. Therefore, many institutions advocate elective treatment of the N0 neck in patients with SGLC. N0 disease can be treated effectively with radiotherapy or with surgery alone. An ideal and generally accepted approach is to treat the N0 neck with selected initial therapy for the primary lesion. An elective neck dissection has the advantage of staging the neck disease, and determines whether an adjuvant therapy is required.

If surgery is decided, then the question of whether elective neck dissection should routinely be directed on both sides of the neck arises. Some offer bilateral neck dissection as a routine in all SGLCs [2–5]. Others consider performing an ipsilateral neck dissection on the side at higher risk of metastases according to primary tumor site and extent [6–9]. All of these studies both proposing and opposing routine bilateral neck dissection have been based on retrospective studies.

In 1991, we designed a prospective study on the management of N0 neck in patients with early stage (T1–T2) unilateral supraglottic carcinoma. The results of the study showed that routine bilateral neck dissection might not be necessary in the surgical treatment of all supraglottic carcinomas [10]. This result led us to determine whether T2–T4 stage supraglottic carcinomas require bilateral neck dissection in the management of N0 necks.

Materials and methods

A total of 72 consecutive patients, all of whom underwent elective bilateral neck dissection for the treatment of supraglottic squamous cell carcinoma, at Erciyes University, Medical Faculty, Kayseri, Turkey, between 1998 and 2006, were analyzed prospectively. Patients with carcinoma extending beyond the larynx and who had previously undergone therapy such as surgery, irradiation and/or chemotherapy were not included.

All tumors were staged according to the 1992 AJCC criteria [11]. The patients were divided into three groups according to the site of tumor growth: Group I consisted of 21 patients with lateralized (clear lateral) lesion reaching but not crossing the midline. Group II comprised 25 patients with cancer largely involving one side and crossing to the midline. Group III included 26 patients with carcinoma equally involving both sides of the larynx or growth into the midline larynx. Assessment of the primary tumor was accomplished by telescopic endovisual photography, direct laryngoscopy, ultrasonography, computerized tomography, and macroscopic evaluation of the surgical specimen.

N0 disease was decided on the palpation, ultrasonography (US) and computerized tomography (CT). All patients underwent both US and CT. If a patient had no palpable nodes clinically, but the US or CT scan indicated abnormal lymph node, the patients were reexamined with palpation. If palpation revealed no abnormal lymph nodes, then the patient was still marked N0. The same head and neck surgeons operated all patients on. The surgical procedure began with bilateral lateral neck dissection (LND) and proceeded to various types of laryngectomies selected according to the status of the primary. The lymphatics of zones II, III, and IV were removed. The principles of the surgery included removal of the primary tumor with negative margins in continuity with the preepiglottic space and with the lymphatic channels through the thyrohyoid membrane. A total of 144 specimens were obtained from 72 patients. Sides and different levels were delineated with the help of landmark stitches placed just after the removal of the specimens. The classification of the American Academy of Otolaryngology-Head and Neck Surgery was used to determine the level of nodal involvement [12]. All surgical specimens were evaluated by the same group of pathologists who are uninformed about the study. The face of the block of the bivalved lymph node served as the first section was obtained (5 μm). The next ten sections from the block discarded. The second of consecutive sections was obtained. This process was repeated for three times. All of three sections were stained hematoxylin and eosin and examined under light microscopy. Immunohistochemical staining was not performed. The term “ipsilateral” was used to define the most prominent side of tumor involvement.

Age, site of primary tumor, T stage, location and number of nodal metastasis were noted. Postoperative irradiation was given in cases with histopathologically positive lymph nodes with capsule invasion and/or multiple positive lymph nodes in the specimen and T4 lesion. The regular follow-up program was scheduled in all patients.

Results

All patients were male and their ages ranged between 30 and 80 years (mean 54.2 years). Of the 72 primary lesions, 12 were classified as T2, 44 were classified as T3 and 16 as T4. Seventy-two patients were divided as follows: 21 in group I, 25 in group II, and 26 in group III (Table 1).

Table 1 Tumor stage distribution according to groups
Full size table

The average number of lymph nodes examined histopathologically was 26.6 for each dissected neck side. Of the 72 patients, 16 were found to have occult regional metastases in pathologic examination (22.2%) (9 pN1, 4 pN2b, 3 pN2c). Extracapsular spread was observed in two lymph node metastases of two patients. The prevalence of occult metastases proportionally increased with T stage from 8.3 to 22.7 and 31.2%, respectively, for T2, T3 and T4 (Table 2). The rates of metastatic nodes were 23.8, 28.0, and 15.3% for group I, group II and group III, respectively, (Table 3). Bilateral neck metastases were found in 2 of 26 patients (7.7%) with central lesions (One with T4, the other with T3 primaries). There was only one patient (4%) with T3 primary who had both ipsilateral and contralateral lymph node metastasis in group II. None of the 21 patients with lateral lesion had contralateral neck metastasis (0%) (Table 3).

Table 2 Incidence of occult lymph node metastases due to tumor stages (n = 72)
Full size table
Table 3 Distribution of occult neck metastasis
Full size table

Level II was the most involved zone (19 necks of 16 patients) followed by level III (7 necks of the 16 patients). Level II involvement was associated with level III positivity in 7 patients. Level II through IV involvement was determined in only 1 patient. Isolated level III and IV positivity was not established.

The follow-up period ranged from 6 to 124 months (median 40.5 months). Two patients developed local recurrences (one from group II, the other from group III) and two patients developed neck recurrences (both from group III and with pT4N1M0, pT4N2C diseases). None of the patients with neck recurrences had local disease.

Discussion

The presence of lymph node metastasis from laryngeal cancer significantly reduces the probability of survival. Therefore, neck dissection is an essential part of the surgical treatment of SGLC.

There is a general agreement in that the surgical treatment of the primary tumor should be performed in conjunction with bilateral neck dissection in patients with ipsilateral or bilateral N+ neck(s). The indication to treat both sides of the necks electively in patients with clinically N0 necks is less clear. Bocca et al. [13] first proposed bilateral neck dissection in all patients with SGLC, since patients who required a delayed dissection had a smaller chance of surviving than patients who had undergone bilateral dissection as part of their initial treatment. An important argument in favor of the bilateral procedure is the better survival rate for patients in whom occult cancer is removed compared to that of those patients who are only treated if clinically manifest metastases appear later [3, 14, 15]. In addition, these authors believe that embryologically, the supraglottis does not form by fusion of two lateral masses; therefore, it may be thought of as a midline structure.

Critics opposed to doing bilateral neck dissection routinely in N0 neck patients appropriately point out that in the majority of cases an intact lymph node system is destroyed in the absence of cancer, taking away a barrier against the cancer for which it was removed. In 1990, DeSanto et al. [16] published a very attentive paper concerning the contralateral neck in the management of SGLC. They analyzed 247 patients, 222 of whom had undergone unilateral or simultaneous bilateral neck dissection. Patients were monitored either up to the time of their death or for a minimum of 3 years after treatment. Of the patients who were shown to be pathologically negative at the time of unilateral neck dissection, only one (1/98) developed contralateral neck disease later. This study was followed by a number of others which all showed that the prevalence of contralateral metastases in early lateral supraglottic carcinomas is less than 10%. All of these studies suggested that contralateral spread without ipsilateral spread is uncommon and the routine use of bilateral neck dissection may not be required [8, 10, 17]. Bilateral simultaneous neck dissection was not applied for all patients in these studies except the retrospective study of de Zinis et al., who included advanced primaries [8].

A thorough knowledge of the lymphatic pathways has paramount importance in estimating the pattern of metastasis. The first detailed description of the lymphatic system of the neck dates back the work of Rouviere [18]. There are two distinct endolaryngeal lymphatic systems: superficial submucosal and deep. The superficial system consists of many vessels and anastomoses also crossing the midline. Deep network is highly specialized with few anastomoses and a medial to lateral flow direction.

Only a few studies on the lymphatics of the larynx, which do not report important differences from the Rouivere’s original description have been published since then. Recently, Werner et al. [19] investigated the architecture and drainage patterns of the lymphatic system of the upper aerodigestive tract in 850 organ specimens with combined microscopic techniques with in vivo and in vitro lymphographic studies. They found that the lymph fluid of the supraglottic space drains through the lateral part of the thyrohyoid membrane in the mediolateral direction over three to six lymphatic channels. Two years later, Werner et al. [20] reported another interesting report on intraoperative lymphatic mapping of midline squamous cell carcinoma. They analyzed 11 patients with carcinomas originating from the epiglottis by intraoperative peritumoral radioisotope injection. They found that a tumor-free neck status on the side of the main lymphatic flow (sentinel node) correlated with no cancer spread in the contralateral neck. Furthermore, those patients who showed bilateral lymphatic drainage with a tumor-free ipsilateral neck side had no metastases on the contralateral side. The results of the study emphasized that the principle stated by De Santo el al. [16] in 1990 is justified.

In the present study, the rates of metastatic nodes were 23.8, 28.0 and 15.3% for group I, group II and group III, respectively. None of the 21 patients with lateral lesion had contralateral neck metastasis (0%). None of these patients developed neck recurrence within the median follow-up of 52.4 (24–124) months. Aforementioned literature and our findings suggest that the flow pattern of the supraglottis is preferentially unilateral, indeed less bilateral than suggested than by many authors. Thus, occult metastases are, like the palpable ones, usually ipsilateral. Interestingly, Biller et al. [21] reported that 12.7% of laryngeal and laryngopharyngeal cancer patients with delayed contralateral neck failure had ipsilateral histologically negative nodes at initial neck dissection. Delayed contralateral neck disease in such cases may be due to altered drainage pattern [22]. We consider that ipsilateral selective neck dissection seems to be acceptable in all lateral SGLCs. In patients with subclinical disease found pathologically in the dissected side, contralateral neck treatment should be attempted. There are several options in such a situation, (1) to perform contralateral neck dissection based on frozen-section analysis, (2) elective neck irradiation, and (3) planned second side neck dissection 2–3 weeks after the primary surgery. It is often not practical to do frozen section analysis of the full ipsilateral neck dissection specimen to decide whether a simultaneous contralateral neck dissection should be done. If the permanent section analysis determines metastatic lymph nodes, then a planned contralateral neck dissection or elective neck irradiation to the other side could be given. In the cases of patients with positive neck histopathology and who have had conservative larynx surgery, surgery might be the first choice. Otherwise, in cases of total laryngectomy and/or requiring postoperative radiation therapy (PORT), elective irradiation may be an appropriate alternative (Fig. 1). It is considered that morbidity of radiation (in case of elective treatment of the contralateral neck) is probably higher than selective neck dissection of level II–IV without IIb. Sentinel lymph node procedure is promising since it can make pre and per-operative diagnosis of the neck more reliable and may therefore be suitable to avoid an unnecessary contralateral neck dissection (20).

Fig. 1
figure 1

An algorithm for the management of SGLC with N0 neck

Full size image

Our study showed that bilateral neck involvement was more frequent in the group of patients with supraglottic primaries growing in the midline larynx (7.7%) than in those with cancer largely involving one side and crossing to the midline (4%) (Table 3). Whereas, 14.3% (1 of 7) of patients with ipsilateral neck metastasis had bilateral neck involvement in group II, 50% (2 of the 4 patients) had bilateral neck involvement in group III.

There was no isolated contralateral neck involvement in group II patients. Even in such lesions, ipsilateral metastasis was more frequent than bilateral ones in accordance with related literature [1]. Thus routine bilateral neck dissection in the management of N0 neck in patients with cancer largely involving one side but crossing to the midline is not required. These patients can also be treated as those who have well-lateralized SGLCs.

In medially localized tumors bilateral metastases are possible. Neck dissection should be conducted on both sides of the neck in patients with central primaries. When patients present cancers in the midline, it is difficult to decide which side of the neck to dissect and bilateral dissections are done in almost all of these cases. Werner et al. [20], thought that intraoperative sentinel node detection might facilitate identification of which side of the neck to initially dissect.

In conclusion, the data presented confirms that occult lymph node metastases are present in a high percentage of patients with supraglottic carcinoma and clinically N0 necks (22.2%), emphasizing the need for an elective treatment of the neck. On the other hand, this study does not support the use of routine bilateral neck dissection in the treatment of all lateral supraglottic carcinomas with clinically N0 neck. The rare involvement of the contralateral side of the neck in lateral lesions suggests that bilateral neck dissection may be better reserved for cases in which the probability of occult metastases is likely; either central or bilateral tumors and lateral tumors with clinically positive nodes in the ipsilateral side of the neck.