FACIAL NODE INVOLVEMENT IN HEAD AND NECK CANCER

FACIAL NODE INVOLVEMENT IN HEAD AND NECK CANCER Patrick Sheahan, MB, AFRSCI, 1 Michael Colreavy, MB, FRCS (ORL), 1 Mary Toner, MB, FRCPath, 2 Conrad V. I. Timon, MD, FRCS (ORL) 1 1 Department of Otolaryngology Head and Neck Surgery, St. James Hospital, 17 Hampton Square, Navan Road, Dublin 7, Ireland. E-mail: sheahanp@eircom.net 2 Department of Histopathology, Trinity College, Dublin, Ireland Accepted 5 December 2003 Published online 14 April 2004 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/hed.20005 Abstract: Background. Facial node involvement in head and neck cancer is rarely documented. Furthermore, facial node removal may increase the risk of damage to the marginal mandibular nerve. Thus, although they may receive afferent lymphatic drainage from the oral cavity, facial nodes are rarely removed during neck dissection. Methods. We retrospectively reviewed the records of 29 patients with oral cavity or oropharyngeal carcinomas who underwent facial node sampling in 32 heminecks during neck dissection for oral cavity or oropharyngeal cancer. Results. Facial node metastases were present in seven patients. Facial node involvement was much more common among patients with palpable cervical lymphadenopathy. Positive facial nodes were associated with an increased risk of treatment failure and a poorer survival. Conclusions. In patients with oral cavity/oropharyngeal primary tumors and palpable cervical lymphadenopathy, consideration should be given to removal of facial nodes during neck dissection; however, further data are awaited before any benefits can be quantified. A 2004 Wiley Periodicals, Inc. Head Neck 26: 531 536, 2004 Keywords: lymph nodes, facial; lymph nodes, mandibular; neck dissection, squamous cell carcinoma; oral cavity Correspondence to: P. Sheahan B 2004 Wiley Periodicals, Inc. The lymph nodes of the face are found along the course of the branches of the facial vessels in the subcutaneous tissues of the face. They have been classified by Rouviere into four groups: malar (superficial to the malar eminence just lateral to the eye); infraorbital (in the canine fossa or nasolabial fold); buccinator (overlying the buccinator muscle); and mandibular (along the outer surface of the mandible, adjacent to the facial artery). 1 As well as receiving afferent drainage from the skin and subcutaneous tissues of the face, these nodes also receive some drainage from the gingival and buccal mucosae. 2 The efferent drainage is primarily to submandibular lymph nodes. The facial nodes may occasionally be involved with cancer of the head and neck; however, this is rarely documented. 3,4 The facial lymph node group of most interest to the head and neck surgeon is the mandibular group, also known as the supramandibular group (Figure 1). These nodes are located adjacent to the facial artery, at the anterior border of the masseter muscle, overlying the horizontal ramus of the mandible. The node of Stahr is described in Gray s Anatomy as a constant node that is found just where the facial artery crosses the horizontal ramus of the mandible. 5 At the lower border of the Facial Node Involvement HEAD & NECK June 2004 531

FIGURE 1. Location of supramandibular and buccinator nodes. body of the mandible, the mandibular nodes become contiguous with the prevascular and retrovascular nodes of the submandibular triangle. The prevascular and retrovascular nodes have often been termed facial nodes; however, this term should be reserved for those nodes located above the inferior border of the mandible only. 6 From a surgical point of view, the mandibular group of facial nodes are important because of their proximity to the marginal mandibular branch of the facial nerve. Surgical techniques used to preserve the marginal mandibular nerve, such as ligating the facial vein low in the neck and retracting it superiorly to elevate the nerve, may render the removal of these nodes during neck dissection more difficult. 7 However, the importance of removing these nodes during neck dissection has been disputed. 8 The appropriate management of facial nodes during neck dissection is of obvious importance. Despite this, there is a scarcity of information in the literature regarding involvement of facial nodes with metastatic carcinoma of the head and neck. With this in mind, this study was undertaken to examine the incidence of involvement of facial nodes with metastatic disease, the risk factors for such involvement, and the implications of facial node involvement on outcome. PATIENTS AND METHODS Over the past years, it has been the general policy of the senior author (CT) to sample the supramandibular facial nodes in patients undergoing neck dissection for oral cavity or oropharyngeal carcinomas whenever cervical metastases were sus- 532 Facial Node Involvement HEAD & NECK June 2004

pected. This study is a retrospective review of these cases. The cases were identified from the pathology records at St. James Hospital. From these records, the details of all patients who had undergone neck dissection and facial node sampling over a 9-year period between 1994 and 2002 inclusive were obtained. Patients were included in the study if they had a primary carcinoma in the oral cavity or oropharynx, had undergone neck dissection as part of their treatment, and had facial node sampling performed at the time of neck dissection. All neck dissections had been performed or supervised by the senior author (CT). Nodes were considered to be facial nodes only if they were located above the level of the lower border of the mandible in the soft tissues of the face. In all cases, the facial nodes were sent for pathologic examination separate from the main specimen, and in most cases, an intraoperative report had been issued at the time of surgery. In keeping with institutional policy, all neck dissections were sent to the laboratory while fresh and dissected before fixation in formalin. The patients medical records were also reviewed. Information obtained from the medical notes included clinical stage of the primary tumor, any previous history of cancer or radiotherapy, type of surgery performed and any adjuvant treatment administered, and outcome. RESULTS From a total of 76 patients who had undergone neck dissection for oral cavity or oropharyngeal primaries, 29 were identified as having undergone dissection and removal of facial nodes during neck dissection. These patients underwent sampling of facial nodes unilaterally (26 cases) or bilaterally (three cases), giving a total of 32 heminecks in which facial nodes had been removed. In all 32 heminecks, the supramandibular nodes had been sampled. Buccal nodes were sampled in two heminecks. In six heminecks, multiple supramandibular nodes were removed. In 21 cases, the facial nodes were sent to the laboratory for intraoperative consultation, and a report was issued at the time of surgery. In the other cases, the facial nodes were sent to the laboratory in formalin, separate from the main neck resection. Twenty-seven of the patients had squamous cell carcinoma. One patient had a hyalinizing clear cell carcinoma of the tongue base, and one other had a mucoepidermoid carcinoma of the tongue base. Twenty patients had a primary tumor within the oral cavity proper (eight floor-ofmouth, seven tongue, four retromolar trigone, one palate). One had a lower lip primary tumor, and eight had an oropharyngeal primary (five tonsil, three tongue base). Facial Nodes. Metastatic disease was present in the facial nodes in seven patients (24%). The details of these patients are shown in Table 1. In all cases, the facial nodes involved were the supramandibular nodes. One patient had two involved supramandibular nodes. Another patient also had metastatic involvement of the buccal node. In the other five, only one facial node had been dissected, and it was positive. The mean maximum size of the positive supramandibular nodes was 0.7 cm (range, 0.5 cm 0.9 cm) compared with a mean maximum size of 1.0 cm in negative supramandibular nodes (range, 0.6 cm 1.8 cm). The sites of primary disease in patients with facial node metastases were retromolar trigone Table 1. Patients with involvement of facial nodes with metastatic disease. Patient no. Primary site New/ recurrent Palpable neck disease Clinical classification Neck histology* Radiotherapy Outcome 1 Tonsil 2nd primary Yes T3 N2 13/37 Yes AWD 15 mo 2 RMT New Palpable T2 N1 0/15 Yes DOD 9 mo facial node 3 RMT New No T2 N0 1/24 Yes NED 26 mo 4 Tongue basey New Yes T2 N1 7/28 Yes AWD 8 mo 5 Floor of mouth New Yes T4 N2 0/30 ipsilateral Yes AWD 12 mo 3/47 contralateral 6 Tonsil New Yes T3 N1 7/34 Yes AWD 5 mo 7 Floor of mouth New Yes T4 N2 4/10 Yes DOD 7 mo Abbreviations: AWD, alive with disease; DOD, died of disease; NED, alive with no evidence of disease; RMT, retromolar trigone; NA, information not available. *Neck histology given as number of positive neck nodes over total number of dissected neck nodes, excluding facial nodes. yhistology = hyalinizing clear cell carcinoma. Facial Node Involvement HEAD & NECK June 2004 533

(two cases), floor of mouth (two cases), tonsil (two cases), and tongue base (one case). Apart from the patient with the tongue base primary tumor, who had a hyalinizing clear cell carcinoma of salivary origin, all the tumors were primary squamous cell carcinoma. One patient with positive facial nodes had previously undergone a total laryngectomy and right-sided modified radical neck dissection (without radiotherapy) for piriform fossa cancer. None of the others had any history of head and neck cancer or of skin cancer. There was no significant difference in the T classification of the primary tumor between patients with positive and negative facial nodes (Table 2). However, palpable lymphadenopathy was present in all but one patient with positive facial nodes and in only seven of the 22 patients with negative facial nodes. One patient with positive facial nodes had palpable facial adenopathy. Palpable facial adenopathy was absent in the rest of the patients. Two patients had positive facial nodes with no other histologic nodal involvement in the ipsilateral hemineck. One of these had palpable facial lymphadenopathy. The other had multiple metastatic nodes in the contralateral hemineck. Interestingly, the facial node on the side with cervical metastases was negative. Table 2. T stage, N stage, and outcome of patients with positive/ negative facial nodes. + Facial nodes Facial nodes T classification Tx 0 1 T1 0 2 T2 3 9 T3 2 7 T4 2 3 N classification N0 1 13 N1 3 2y N2 3 6 N3 0 1 Outcome No evidence 1 15 disease Recurrence 6 3 Death from 0 3 other causes No information 0 1 2-year determinate disease-free survival* 1/7 5/8 *2-year disease-free survival excluding surviving patients not censored for at least 2 years and excluding patients dying from other causes. yone patient with no palpable cervical lymphadenopathy had N1 disease on the basis of radiologic findings. In addition to surgery, 20 patients also underwent postoperative radiotherapy. This included all seven patients with positive facial nodes. The general indications for radiotherapy included the presence of positive margins in the neck or primary site, advanced primary tumor (pt3/4), the presence of two or more metastatic nodes, the presence of extracapsular spread, or involvement of a node larger than 3 cm. The details of patients undergoing radiotherapy and sites of failure are shown in Table 3. Only one patient with positive facial nodes had no evidence of disease after 2 years; regional recurrence developed in all the others. In contrast, only three patients with negative facial nodes had a recurrence. Seven of 10 patients with negative facial nodes censored at 2 years were alive, five with no evidence of disease, and two with recurrent disease. All three dead patients had died of other causes in the early postoperative period. One other patient had regional recurrence within 2 years. Regional recurrence developed in all five patients with palpable cervical adenopathy and positive facial nodes. On the other hand, among the eight patients with palpable cervical adenopathy and negative facial nodes, recurrence developed in two, one died from other causes, one was lost to follow-up, and the other four had no evidence of disease 8 months, 1 year, 14 months, and 5 years postoperatively. In addition, one further patient who had no palpable cervical adenopathy, but who had N1 disease on the basis of radiologic findings, was well with no evidence of disease at 4 years. Intraoperative Reports. Intraoperative histology reports had been issued for five of the cases of positive facial nodes. Of these, three had been reported as positive and two as negative. Intraoperative reports were negative for tumor in all 16 cases of negative facial nodes. DISCUSSION Metastases to facial nodes from mucosal cancer of the head and neck is rarely documented 3,4 ; because of this, these nodes have received scant attention in the literature. Dissection of facial nodes is not routinely performed during neck dissection; in fact, the upper limit of a radical neck dissection is generally given as the lower border of the mandible. 7 However, it has long been recognized that the supramandibular nodes may receive some of the lymphatic drainage from oral cavity mucosa. 1,2,5 Contiguous spread to facial nodes 534 Facial Node Involvement HEAD & NECK June 2004

Table 3. Number of patients with positive and negative facial nodes undergoing radiotherapy and sites of failure in patients with recurrent disease. + Facial nodes (n =7) Facial nodes (n = 22) Radiotherapy No radiotherapy Radiotherapy No radiotherapy No. of patients 7 0 13 9 Recurrent disease 6 0 3 0 Site of failure 6 neck 2 neck, 1 local + neck from submandibular nodes is another possibility, given that the presence of nodal metastases may interfere with normal lymphatic drainage patterns. 2 However, it is not known whether facial node dissection as a routine part of neck dissection is justified, particularly given that such a maneuver may increase the risk of damage to the marginal mandibular branch of the facial nerve. This study is, to our knowledge, the first to examine the incidence and significance of facial node involvement in head and neck cancer. The incidence of histologically positive facial nodes among patients with oral cavity or oropharyngeal primary tumors in our series, at 24%, was unexpectedly high. It is notable that facial node metastases occurred in the setting of palpable facial adenopathy in only one case; however, in most of the remaining cases, palpable cervical adenopathy was present. In general, facial node involvement was much more common among patients with palpable neck disease. Facial node metastasis was present in only one patient with no palpable lymphadenopathy, giving an incidence of facial node involvement in the N0 neck of 6.25%. On the other hand, six of the 16 heminecks with N+ disease had histologically positive facial nodes, for an incidence of facial node involvement of 37.5% in this group. Furthermore, most patients with positive facial nodes had multiple positive cervical nodes (Table 1). These findings would suggest that facial node metastases generally occur only in the setting of advanced disease. Facial node metastases may occur as a result of the facial nodes being the first echelon of lymphatic drainage from a primary site involved with carcinoma or as a result of disturbances to the normal pattern of lymphatic drainage by established cervical metastases. The possibility of facial node involvement as a result of first-echelon drainage is suggested in our series by the patient with the T2 retromolar trigone primary tumor, who had facial node involvement in the absence of any other cervical metastases. In another patient with a floor-of-mouth primary tumor, facial node involvement occurred in the absence of cervical metastases to the ipsilateral hemineck; however, it is possible that nodal disease in the contralateral hemineck may have led to alterations in the normal drainage patterns. One striking finding is the singularly poor outcome in patients with histologically positive facial nodes. The determinate 2-year disease-free survival among patients with positive facial nodes was only 14.3% compared with 62.5% for patients with negative facial nodes (Table 2). However, four of the patients with positive facial nodes and regional recurrence had close or positive margins at the site of resection of the primary tumor. Two others had regional recurrence despite fully controlled primary tumors and postoperative radiotherapy. One patient with a retromolar trigone primary had no recurrence. This patient had postoperative radiotherapy on the basis of two positive nodes, one facial node and one level 2 node. One may speculate as to whether the outcome may have been less fortunate had facial nodes not been dissected, particularly given the likelihood that radiotherapy might not have been administered if only one positive node had been found. In patients with oral cavity or oropharyngeal carcinomas and an N0 neck, the likelihood of facial node metastasis would seem to be low. Furthermore, it is questionable whether any survival benefit is likely to accrue as a result of routine facial node dissection, particularly in view of the data suggesting only a marginal survival benefit for elective neck dissection in the N0 neck over a policy of watchful waiting. 9,10 In the case of the N+ neck, on the other hand, the likelihood of facial node metastases would seem to be much higher. Furthermore, most of these patients would seem to have a poor prognosis, despite removal of the involved nodes and postoperative radiotherapy. Recurrent disease in facial nodes after neck dissection is rarely documented 3,4 ; however, it is not known whether residual disease in facial Facial Node Involvement HEAD & NECK June 2004 535

nodes may be a source of recurrent neck disease after apparently adequate control of disease in the primary site and neck. Although the number of patients who had undergone facial node sampling in this study was small, our findings would nonetheless suggest that consideration should be given to removal of the facial nodes whenever neck dissection is being undertaken for carcinoma of the oral cavity or oropharynx; however, clearly, further data are required before any benefits accruing from this procedure can be quantified. REFERENCES 1. Rouviere H. Anatomie des lymphatiques de l homme. Paris: Maison et Cie; 1932. 2. Sessions RB, Picken CA. Malignant cervical adenopathy: In: Cummings CW, Fredrickson JM, Harker LA, Krause CJ, Schuller DE, Richardson MA, editors. Otolaryngology head & neck surgery, 3rd ed. St. Louis: Mosby; 1998. p 1737 1757. 3. Tart RP, Mukherji SK, Avino AJ, Stringer SP, Mancuso AA. Facial lymph nodes: normal and abnormal CT appearance. Radiology 1993;188:695 700. 4. Chong VFH, Fan YF. Facial lymphadenopathy in nasopharyngeal carcinoma. Clin Radiol 2000;55:363 367. 5. Gray H. Anatomy of the human body, 26th ed. Philadelphia: Lea and Febiger; 1954. 6. DiNardo LJ. Lymphatics of the submandibular space: an anatomic, clinical, and pathologic study with applications to floor-of-mouth carcinoma. Laryngoscope 1998;108: 206 213. 7. Medina JE. Neck dissection. In: Bailey BJ, editor. Head and neck surgery. Philadelphia: Lippincott-Raven; 1998. p 1563 1593. 8. Maruyama S. Marginal branch of the facial nerve in submandibular dissection for T2 lingual carcinomas. Nippon Jibiinkoka Gakkai Kaiho 1998;101:1436 1441. 9. Fakih AR, Rao RS, Patel AR. Prophylactic neck dissection in squamous cell carcinoma of the oral tongue: a prospective randomized study. Semin Surg Oncol 1989;5: 327 330. 10. Vandenbrouck C, Sancho-Garnier H, Chassagne D, Saravane D, Cachin Y, Micheau C. Elective versus therapeutic radical neck dissection in epidermoid carcinoma of the oral cavity. Cancer 1980;46:386 390. 536 Facial Node Involvement HEAD & NECK June 2004