The mandibular gingiva is the second most common site of oral cavity squamous cell carcinoma. This retrospective study was designed to determine the clinicopathological features of squamous cell carcinoma of the mandibular gingiva (MGSCC) and to establish a new risk model to predict overall survival. The study included 207 patients with primary MGSCC from January 2000 to September 2009. The medical charts were reviewed and data related to clinical characteristics, treatment provided, histopathological analysis, and follow-up were recorded. All patients underwent surgery as the first-line therapy; follow-up ranged from 1 to 171 months (median 63 months). Clinical characteristics and pathological outcomes were analyzed with respect to the 5-year overall survival rate. A survival risk model was established, and patients were classified into low-, moderate-, and high-risk groups based on the prognostic index designed in this study. The 5-year overall survival rates for the low-, moderate-, and high-risk groups were 92.3%, 76.9%, and 34.2%, respectively. Pathological node metastasis, perineural invasion, and extracapsular spread were the most significant predictive factors for 5-year overall survival. MGSCC is not aggressive, and the survival outcomes of MGSCC are better than those of squamous cell carcinoma (SCC) at other sites. It is suggested that patients with T2–T4 tumours undergo elective neck dissection and those with T1 tumours be followed up without addressing the neck.
In the population of northern China, the mandibular gingival mucosa is the second most common site of oral cancer, followed by the buccal mucosa (BSCC) and the floor of the mouth; the tongue is the most commonly affected site. The mandibular gingival mucosa is also the second most common site in the Japanese population, following tongue and floor of the mouth, and the third most common site in the USA. In South Africa, the mandibular gingiva is the most common site, followed by tongue and floor of the mouth. Thus there are geographical differences in the tumour locations. The clinical and pathological characteristics may differ in different regions of the world.
Squamous cell carcinoma of the mandibular gingival mucosa (MGSCC) is more common in elderly patients, and mandibular bone is more likely to be involved. It may be misdiagnosed as peri-apical or periodontal disease. Patients usually present with complaints of pain, swelling, tooth loosening, numbness of the lower lip, etc. The mandibular gingival mucosa site is thought to be rare, and the outcomes of treatment have been deemed to be poor. However, in the present authors’ experience, the survival outcomes of patients with MGSCC are better than those of patients with squamous cell carcinoma (SCC) at other sites of the oral cavity. Many institutions worldwide have investigated prognostic factors in MGSCC patients, but this research has been limited by the numbers of patients and prognostic factors investigated.
The hospital at which the present study was performed is one of the major medical institutions in the north of China, which has a population of more than 600 million. The aims of this hospital-based retrospective study were (1) to investigate the clinicopathological features and patterns of neck nodal metastasis of MGSCC in the population of northern China; (2) to compare the oncologic behaviour of MGSCC in this homogeneous population with that found in studies performed in other areas, such as America, Europe, and South Africa; and (3) to establish a new risk model to predict the survival of MGSCC patients.
Materials and methods
This research project was approved by the institutional review board of the study hospital in Beijing, China. Two hundred and seven patients with primary MGSCC treated in the department of oral and maxillofacial surgery of this hospital were identified from January 2000 to September 2009 and were included in the study. All 207 patients had primary cancer and had not undergone previous treatment. Patients who had not received previous treatment and who had pathologically proven SCC were included in the study; those with tumours arising primarily in the mandibular bone or retromolar trigone were excluded.
All of the patients underwent radiographic examinations, including panoramic radiography, computed tomography, magnetic resonance imaging, and ultrasonography. A baseline chest X-ray, complete blood count, and blood chemistries were also obtained. Clinical staging was based on the clinical and imaging findings according to the 2010 Union for International Cancer Control (UICC)/American Joint Committee on Cancer (AJCC), staging criteria, 7th edition. T4 tumours were defined as those invading the cortical bone. The mandibular gingiva refers to the mucosa overlying the alveolar process of the mandible. This lies between the line of attachment of the mucosa at the lower gingivobuccal sulcus and the line of free mucosa of the floor of the mouth, and extends posteriorly to the ascending ramus of the mandible. Clinical information, including tumour location, sex, age, clinical stage, smoking history, alcohol use, and treatment characteristics, were collected from the medical records.
All patients underwent surgery as the first-line therapy. Local excision of the primary tumour was performed with a margin of at least 15 mm. Frozen biopsies of the margins were obtained and if they were positive, additional tissue was resected and cryosectioned to ensure that the revised margin was free of tumour. Patients were treated with neck dissections if the nodes contained suspected metastatic lesions. A flow chart outlining the treatment of the 207 patients is presented in Fig. 1 . Marginal or segmental mandibulectomies were performed according to the extent of the bone invasion. Reconstruction plates with forearm flaps or vascularized fibula grafts were used to restore the defects. Postoperative radiotherapy was advised for patients with positive lymph nodes, pT4 tumours, or close margins (<4 mm).
As routinely practiced in the study hospital, patients were advised to return regularly at 1-month intervals for the first year, at 2-month intervals for the second year, 3-month intervals for the third year, and at intervals of 3–6 months for the fourth and fifth years.
Tumours were graded into well-differentiated, moderately differentiated, or poorly differentiated SCC. Perineural invasion was defined as carcinoma specifically tracking along or within a nerve. The detection of minor degrees of extracapsular spread (ECS) is aided by harvesting lymph nodes with their immediate pericapsular adipose tissue in position. Histological differentiation of regional lymph node metastases was obtained retrospectively from the pathology reports. One experienced pathologist (JYZ), who was blinded to the patient outcomes, reviewed all the available slices and recorded perineural invasion, vascular emboli, diffuse infiltration, and ECS features.
The clinical and pathological characteristics were analyzed using the Kaplan–Meier method, and factors significantly influencing the outcome were determined with the log-rank test. Univariate and multivariate analysis using a Cox proportional hazards model was applied to determine the covariates that best predicted survival rates. Statistical calculations were performed using commercially available software (IBM SPSS Statistics for Mac, version 20.0; IBM Corp., Armonk, NY, USA).
Two hundred and seven patients satisfying the inclusion criteria were included in the study. Of these 207 patients, 121 (58.5%) were male and 86 (41.5%) were female. They ranged in age from 15 to 86 years (median 64 years). Eighty patients (38.6%) had a history of smoking and 49 (23.7%) consumed alcohol. An exophytic lesion was the most common presentation, seen in 104 patients (50.2%), followed by ulcerative in 60 patients (29.0%) and infiltrative in 35 patients (16.9%). The clinical TNM staging was recorded for each patient and is presented in Table 1 .
|Characteristics||No. of patients (%)|
|Median (range)||64 (15–86)|
|Clinical N stage|
The pathological nodal status, degree of differentiation, perineural invasion, vascular emboli, diffuse infiltration, and ECS were recorded. The actual numbers of these pathological outcomes are presented in Table 2 .
|T stage||Differentiation grade||Perineural invasion||Vascular emboli||Diffuse infiltration||ECS|
n = 37
|25 (67.6%)||11 (29.7%)||1 (2.7%)||1 (2.7%)||36 (97.3%)||4 (10.8%)||33 (89.2%)||23 (63.9%)||13 (36.1%)||3 (9.1%)||30 (90.9%)|
n = 81
|52 (64.2%)||25 (30.9%)||4 (4.9%)||7 (9.0%)||71 (91.0%)||11 (14.1%)||67 (85.9%)||58 (74.4%)||20 (25.6%)||5 (7.1%)||65 (92.9%)|
n = 22
|8 (36.4%)||14 (63.6%)||0 (0)||3 (15%)||17 (85%)||4 (20%)||16 (80%)||14 (70%)||6 (30%)||2 (10%)||18 (90%)|
n = 62
|31 (50%)||25 (40.3%)||6 (9.7%)||11 (18.3%)||49 (81.7%)||15 (25%)||45 (75%)||53 (88.3%)||7 (11.7%)||10 (17.2%)||48 (82.8%)|
n = 202
Of the cN0 patients, 32.6% (45/138) had moderate-to-poor differentiation of the tumour, whereas 64.1% (41/64) of cN+ patients had moderate-to-poor differentiation. In the cN0 patients, perineural invasion, vascular emboli, ECS, and diffuse infiltration occurred in 6.0%, 10.5%, 4.2%, and 69.7%, respectively. Of the cN+ patients, perineural invasion, vascular emboli, ECS, and diffuse infiltration occurred in 22.6%, 32.3%, 24.6%, and 90.3% of the patients, respectively (see details in Table 3 ).
|Number of patients||pN+||Vascular emboli||Diffuse infiltration||ECS||Perineural invasion||Differentiation (moderate–poor)|
|T1||31||3/27 (11.1%)||4/31 (12.9%)||18/30 (60%)||1/27 (3.7%)||0/31 (0)||8/31 (25.8%)|
|T2||59||13/53 (24.5%)||5/57 (8.8%)||39/57 (68.4%)||1/50 (2%)||2/57 (3.5%)||17/59 (28.8%)|
|T3||15||4/14 (28.6%)||2/14 (14.3%)||9/14 (64.3%)||1/13 (7.7%)||1/14 (7.1%)||8/15 (53.3%)|
|T4||33||7/33 (21.2%)||3/31 (9.7%)||26/31 (83.9%)||2/30 (6.7%)||5/31 (16.1%)||12/33 (36.4%)|
Occult metastasis in cN0 patients
Overall, 138 patients were diagnosed as having a clinically negative neck. The distribution according to tumour size and the nodal status are presented in Table 3 .
Of the 127 cN0 patients who had a neck dissection, 27 had positive neck metastasis. Of these, 13 patients were pathologically N1, 11 were N2b, and three were N2c. The percentage of total occult neck metastasis was 21.3%. The rates of occult neck metastasis were 11.1%, 24.5%, 28.6%, and 21.2% for T1, T2, T3, and T4, respectively.
Among the total 207 patients, 196 (94.7%) had a neck dissection. Of these, 93.9% (184/196) had an ipsilateral neck dissection, while 6.1% (12/196) had a bilateral neck dissection. Of the 184 patients who had an ipsilateral neck dissection, 93 had an elective I–III neck dissection, 15 had an elective I–IV neck dissection, 21 had an elective I–V neck dissection, and 55 had a therapeutic neck dissection. Of the 12 patients who had a bilateral neck dissection, nine had a contralateral elective I–III neck dissection, one had a contralateral elective I–IV neck dissection, one had a contralateral elective I–V neck dissection, and one had a contralateral therapeutic I–V neck dissection.
Furthermore, 37.2% (77/207) of the patients received a marginal mandibulectomy and 62.8% (130/207) received a segmental mandibulectomy. Reconstruction required 107 free flaps and two pedicled flaps, including 87 fibular flaps, 17 forearm flaps, two iliac flaps, one rectus muscle flap, and two pectoralis major myocutaneous flaps. Eighteen patients underwent reconstruction with titanium plates and 58 patients had primary closure.
The follow-up period ranged from 1 to 171 months (median 63 months). A recurrence or metastasis occurred in 49.3% (102/207) of patients. Table 4 describes the types of recurrence, treatment provided, and the true survival rates after salvage treatment. Four patients who were free of recurrence or metastasis died of other causes. One had paralysis after radiation therapy, another died of a stroke at 14 months following surgery, and two patients died of heart disease without disease recurrence after 77 months of follow-up. Eight patients were lost to follow-up.
|Recurrence||Patients||Treatment||Success rate of operative salvage||Death|
|Local||33||OP ( n = 17), OP + RT ( n = 3), RT ( n = 2), quit ( n = 8), missing ( n = 3)||21.2%, 7/33||78.8% (26/33)|
|Regional||17||OP ( n = 5), OP + RT ( n = 6), quit ( n = 6)||11.8%, 2/17||88.2% (15/17)|
|Local-regional||15||OP ( n = 3), OP + RT ( n = 1), CCRT ( n = 2), quit ( n = 9)||20%, 3/15||80% (12/15)|
|Distant||9||Quit ( n = 8), CCRT ( n = 1)||–||100% (9/9)|
|Second primary malignancy||28||OP ( n = 18), OP + RT ( n = 4), quit ( n = 3), CCRT ( n = 3)||39.3%, 11/28
The relationship between the overall survival rate and the survival time is shown in Fig. 2 . The total 5-year overall survival rate was 71.8%. The actuarial overall survival rates of the patients according to the various clinicopathological factors are shown in Table 5 . An advanced T stage was found to adversely affect the survival rate. There was a significant difference between the T2 and T4 stages ( P = 0.020). The 5-year overall survival rates for the different TNM stage groups were 88.1% for stage I, 92.4% for stage II, 61.4% for stage III, and 53.5% for stage IV (stage I vs. stage II, P = 0.846; stage I vs. stage III, P = 0.034; stage I vs. stage IV, P < 0.001; stage II vs. stage III, P = 0.017; stage II vs. stage IV, P < 0.001; stage III vs. stage IV, P = 0.201). The 5-year overall survival rates in cases of perineural invasion, vascular emboli, and ECS were 35.6%, 54.3%, and 15.4%, respectively, but they were significantly increased in patients with negative pathological characteristics. However, the growth pattern and the patients’ smoking history, alcohol consumption, and sex had no effect on the survival rate.
|5-Year overall survival rate||Log rank P -value|
|T1||77.6%||T1 vs. T2, P = 0.017|
|T2||76.7%||T2 vs. T4, P = 0.020|
|N stage a|
|N0||85.6%||N0 vs. N1, P = 0.002|
|N1||58.6%||N0 vs. N2, P < 0.001|
|Well||81.8%||Well vs. moderately, P = 0.001|
|Moderately||56.7%||Well vs. poorly, P = 0.006|
|Neural invasion||P < 0.001|
|Vascular emboli||P = 0.012|
|ECS||P < 0.001|