This work was presented at the American Head and Neck Society Annual Meeting, 2010 Combined Otolaryngology Spring Meeting; April 28-29, 2010; Las Vegas, Nevada.
Mucoepidermoid carcinoma (MEC) is the most common malignancy of the major salivary glands. Prior reports noted histological grade and tumor stage as consistently important prognostic factors. This study reviewed the experience of patients with MEC at the University of Texas MD Anderson Cancer Center to determine the impact of clinical and pathologic findings on disease outcomes.
A retrospective clinical review was performed of patients with salivary gland MEC treated at a tertiary cancer center from 1990 to 2007.
A total of 125 patients were included. The 5-year overall survival and disease-free survival of all patients were 79.3% and 76.5%, respectively. Patients with low- and intermediate-grade disease had significantly better overall survival and disease-free survival than patients with high-grade disease, but no difference in survival rates was found between low- and intermediate-grade disease. Pathologic results of positive lymph nodes, extracapsular lymph node spread, and perineural invasion were all found to be poor prognostic indicators. On multivariate analysis, advanced disease stage and perineural invasion were found to be the most significant prognostic factors.
Salivary gland neoplasms account for approximately 5% of head and neck neoplasms, but fewer than 1% of malignancies in all body sites.1 However, despite their rarity, these carcinomas represent a highly diverse range of histological subtypes. Of these, mucoepidermoid carcinoma (MEC) is the most common, representing 30% to 40% of all salivary gland malignancies1 and up to 50% of parotid malignancies.2
Mucoepidermoid carcinoma tumors consist of both mucinous and epidermoid cells, although several additional cell types are also found. Furthermore, there is considerable variation in the type, distribution, and growth pattern of MEC cells, leading to several histological variants.3 Advances in immunohistochemistry, cytogenetics, and molecular genetics have provided information important for accurately differentiating MECs from pathologically similar neoplasms and has likely rendered earlier studies less accurate in their inclusion criteria.4, 5
The recurrent and metastatic behavior of MEC has been linked to histological grade and tumor stage, which have been confirmed in many retrospective studies as consistently important prognostic factors.6-9 These factors are often considered in guiding treatment strategies. Treatment recommendations for the various histological grades and stages, however, remain controversial. Even though radiotherapy techniques have changed over the past 20 years, the impact of these improvements on salivary gland malignancies has not been adequately studied. Furthermore, trials of systemic therapy have revealed limited success.10
To address these issues, we reviewed our experience of patients with MEC to determine the impact of clinical and pathologic findings on disease outcomes.
MATERIALS AND METHODS
In our institutional database, we identified more than 1500 patients with salivary gland malignancies who were treated at The University of Texas MD Anderson Cancer Center (MDACC) from 1990 through 2007, and all patients diagnosed with MEC of the salivary glands were determined. Institutional Review Board approval was obtained for this study. Epidemiological, clinical, pathologic, treatment, and follow-up data were obtained from patient medical records. Rapid tumor growth was a subjective evaluation by the patient and typically included noticeable enlargement of the tumor in less than 6 months. Patients with less than 2 years of follow-up were excluded, as were those who were only seen for a single consultation.
All specimens were reviewed by an MDACC head and neck pathologist. The specimens were analyzed for the following criteria: tumor grade (low [LG], intermediate [IMG], or high [HG]), surgical margin status, the presence of perineural invasion, and nodal status. Records from both outside institutions and MDACC were used to characterize tumors. Mucoepidermoid carcinoma was graded based on combined structural and cellular parameters into LG, IMG, and HG categories, in which LG was assigned to largely cystic lesions with <10% solid areas and no anaplastic cytologic features, IG comprised tumors with <40% of cystic formation and solid component with minimal cellular anaplasia, and HG comprised solid epithelial formation with focal glandular and mucinous differentiation and marked cellular anaplasia. Disease was classified using the TNM staging system in accordance with the American Joint Committee on Cancer (AJCC) classification.11 Treatment data included outside treatment received before patient presentation as well as treatment at our institution.
Descriptive statistics for scaled values and frequencies of study patients within the categories for each of the parameters of interest were enumerated with the assistance of commercial statistical software. Correlations between parameters and endpoints were assessed by the Pearson chi-squared test or, where there were fewer than 10 subjects in any cell of a 2 × 2 grid, by the 2-tailed Fisher exact test. Curves describing overall survival (OS), disease-free survival (DFS), and disease-specific survival (DSS) were generated by the Kaplan-Meier product limit method. The statistical significance of differences between the actuarial curves was tested by the log-rank test. Follow-up time was the time from initial presentation at this institution for the tumor of interest until the date of last contact or death for OS and DFS measurements. For DSS, the follow-up time was from the conclusion of treatment for the primary tumor of concern until the date of first recurrence, death, or last contact. In addition to patient status at the last clinic visit, last contact information also included correspondence with patients or their outside physicians. Hazard ratios (HRs) and proportional HRs were calculated according to the Cox regression technique. These statistical tests were performed with the assistance of the Statistica (StatSoft, Inc., Tulsa, Okla) and SPSS (IBM SPSS, IBM Corporation, Somers, NY) statistical software applications. Statistical tests were considered significant when P < .05.
A total of 202 patients with a diagnosis of MEC of the salivary glands were evaluated at the MDACC during the specified study period, and 125 were eligible for inclusion. Clinical and histopathologic characteristics are presented in Table 1. Our series included 57 males and 68 females (ratio, 1:1.2). The mean age at presentation was 48.2 years. The median length of follow-up was 63.5 months (range, 24-372 months).
Table 1. Clinical and Histopathologic Characteristics of Mucoepidermoid Carcinomas
No. (%) of Patients
NR indicates not reported.
No evidence of disease
Living with disease
Dead of disease
The parotid gland was the most frequent primary tumor site (86.4%), and tumors most commonly occupied the superficial lobe (86.9%). Eighty-six patients (75.5%) had stage T1 or T2 disease, 18 patients (15.3%) had regional metastases, and 5 patients (4%) had distant metastases at the time of original staging. In patients with regional or distant metastases on initial staging, 91.7% underwent computed tomography imaging, 16.7% underwent magnetic resonance imaging, and 25% underwent multimodality imaging.
Thirty-five patients (28%) presented untreated, 65 patients (52%) presented with their treatment partially completed, 6 patients (4.8%) presented with their planned treatment completed and with no evidence of disease, and 18 patients (14.4%) presented with recurrence or persistence. Forty-two patients (33.6%) were treated with surgery alone, 74 patients (59.2%) were treated with surgery and radiation therapy (XRT), 2 patients (1.6%) were treated with XRT alone, and 7 patients (5.6%) were treated with a multimodality treatment that included chemotherapy in addition to surgery and/or XRT.
Of the patients who presented to our institution with their treatment partially completed, 29 patients (44.6%) underwent a surgical re-excision of the primary site and/or a lymph node dissection, whereas 31 patients (47.7%) underwent postoperative radiotherapy alone. Criteria for additional surgical intervention included an excisional biopsy without parotidectomy in 14 patients, incomplete resection with residual disease confirmed on physical exam and/or imaging in 7 patients, residual lymphadenopathy on imaging in 3 patients, and recurrent disease on physical exam and imaging in 5 patients. Patients who underwent postoperative radiotherapy alone did not possess evidence of gross residual or recurrent tumor, except for in 1 patient who was not a surgical candidate. Criteria for postoperative radiotherapy included positive or close surgical margins in 25 patients, perineural or lymphovascular invasion in 14 patients, and extraglandular extension in 2 patients.
Of the 54 patients who possessed IMG tumors, 34 (61.8%) underwent postoperative radiotherapy due to evidence of positive or close margins in 28 patients, positive perineural or lymphovascular invasion in 7 patients, extraglandular extension in 5 patients, and tumor spillage in 2 patients. Of the 30 patients who possessed LG tumors, 12 (41.4%) underwent postoperative radiotherapy due to evidence of positive or close margins in 9 patients and perineural invasion in 3 patients. A significantly higher proportion of patients with IMG tumors were irradiated than were patients with LG tumors (P = .0047).
Forty-four (35.2%) patients underwent a neck dissection for primary or recurrent disease. Location of operation was equally divided with 22 patients undergoing neck dissection at an outside institution and 22 patients undergoing neck dissection at the MDACC. Criteria for neck dissection included lymphadenopathy on imaging in 16 patients, palpable lymphadenopathy in 7 patients, preliminary diagnosis of squamous cell carcinoma in 4 patients, recurrence in 2 patients, aggressive features in 1 patient, and for unknown reasons in 10 patients.
Twenty-four patients (19.4%) had a recurrence after treatment for their original disease, with 22 of these within 10 years after treatment. Fourteen patients had local recurrence, 12 had regional recurrence, and 13 had distant metastases. Twenty patients (16%) died of their disease, 19 of these within 10 years of presentation.
Outcome data are presented in Table 2. The 5-year OS and DFS of all patients were 79.3% and 76.5%, respectively. When compared by sex, the OS for men and women was 71.2% and 86.2%, respectively (P < .05); DFS was 68.2% and 83.8%, respectively (P < .05). Smokers with a 40 pack-year smoking history had an OS of 31.2% and DFS of 40%, compared with 84.5% and 78.6%, respectively, for those with <40 pack-years smoking and nonsmokers (P < .05 for OS, P < .05 for DFS).
Table 2. Survival Analysis of Clinical and Histopathologic Variables for Mucoepidermoid Carcinomaa
Hazard Ratio (95% CI)
Hazard Ratio (95% CI)
CI indicates confidence interval.
Cumulative overall survival (OS) and disease-free survival (DFS) at 5 years for all 125 patients.
Facial nerve palsy
Smoking >40 pack y
Parotid vs submandibular
Low vs intermediate
Intermediate vs high
Low/Intermediate vs high
T1/T2 vs T3/T4
N0 vs N1/N2
M0 vs M1
I/II vs III/IV
Positive vs negative
Clinical features of facial nerve palsy, skin involvement, fixed tumor, rapid tumor growth, and a submandibular gland location were found to be poor prognostic factors. By contrast, there was no statistically significant difference in OS or DFS between patients with and without pain (P = .098 for OS, P = .477 for DFS). Patients with T1 and T2 tumors together had an OS and DFS of 93.5% and 88.3%, respectively, whereas those with T3 and T4 tumors indicated significantly worse OS and DFS of 41.9% and 43.1%, respectively (P < .05, P < .05). In addition, patients with AJCC stage III and IV disease indicated a significantly worse OS and DFS than those with stage I and II disease (P < .05 for both) (Fig. 1).
The OS and DFS were 92.8% and 88.3% for LG, 95.1% and 90.7% for IMG, and 51% and 43.1% for HG disease, respectively (Fig. 2). There was a statistically significant difference in both OS and DFS for LG and IMG compared with HG (P < .05), but no difference between LG and IMG (P = .29). Positive lymph nodes on pathologic examination, extracapsular lymph node spread, and perineural invasion (Fig. 3) were all found to be indicators of poor prognosis. There was no statistically significant difference in OS for patients with positive surgical margins (P = .09); however, a decreased DFS was seen in this patient population (P < .05).
The relative strengths of several important variables were evaluated by the log-rank P and Cox regression HRs. On this multivariate analysis, advanced AJCC stage and perineural invasion of large nerves were found to be significantly poor prognostic factors. In addition, grade (HR = 5.78) appears to have a more profound effect on survival than site (HR = 4.357), and T stage (HR = 5.99) has a greater effect than grade (Figs. 4 and 5).
Recurrence data are presented in Table 3. There was a significant increase in incidence of recurrence among patients with HG disease compared with the combined group of patients with LG and IMG disease (P < .05). Patients with positive surgical margins were also noted to have significantly increased rates of distant metastasis (P < .05). The lungs were the most common site of metastasis.
Table 3. Effects of Histopathologic Features on Rates of Metastases (N = 125)
In this study, we report our single institution's experience of patients with MEC of the salivary glands. Although the 5-year OS for our patients with MEC was found to be 79%, in accordance with values in the literature (ranging from 68%-75%),12-14 those with advanced stage and HG tumors had an unfavorable prognosis. On univariate analysis, pathologic features that correlated with worse survival outcomes were high histological grade, large tumor size, nodal disease, metastatic disease, perineural invasion, and positive surgical margins. On multivariate analysis, disease stage and perineural invasion were found to be independent prognostic factors.
Demographic factors were assessed for effect on prognosis, and sex had a significant impact on both OS and DFS, with men having a less favorable OS and DFS than women. Although the etiology for this could not be gleaned from our data set (ie, tumor size or time to diagnosis), this effect may be attributed to older age, smoking behavior, and/or increased comorbidities in our male patients. Further studies are needed to address these possible confounders.
Tumor staging and histological grade appear to have profound effects on survival. We observed a significant decrease in survival for patients with large T3 and T4 tumors, the presence of nodal disease, and metastatic disease. Past reports have also indicated a significant trend of worsening survival from LG to IMG to HG MEC.15 A report by Nance et al showed no statistically significant difference in OS or DFS between patients with LG and IMG MEC,16 which deviates from previous findings in the literature. Our data support their conclusions that LG and IMG tumors result in identical survival outcomes whereas HG tumors indicate a significantly worse outcome.
The similar rates of survival between patients with LG and IMG lesions may be attributed to earlier diagnosis and/or improved treatment strategies for IMG tumors. It has been suggested that adjuvant XRT is indicated for HG carcinomas, stage III and IV tumors, tumors with extensive perineural or vascular invasion or with extraglandular extension, tumors of the deep lobe of the parotid, and incompletely excised tumors.3, 17 It remains unclear whether patients with IMG tumors should receive XRT based on histologic findings alone, so other clinical and pathologic characteristics should be considered in this decision. The most common factors that led our patients with IMG tumors to undergo adjuvant XRT included positive or close surgical margins, perineural or lymphovascular invasion, and extraglandular extension. The addition of postoperative radiotherapy in tumors with these high-grade features may have contributed to the improvement in survival of intermediate-grade MEC. Because a larger percentage of patients with IMG tumors possess high-grade features, and patients with LG tumors have always maintained a favorable survival curve, the positive effects of radiotherapy may be more evident in IMG MEC survival. In addition, the increased acceptance of the need for a superficial parotidectomy, as opposed to an excisional biopsy alone, has likely improved treatment results as well. Larger cohorts of patients will be necessary to further characterize the effectiveness of radiation for treating IMG tumors.
A significant number of our patients had positive surgical margins in this study, and there are several explanations for this. A majority of patients in this study presented to the MDACC with partially treated disease, specifically, positive surgical margins. Although there was no statistically significant difference in OS for patients with positive surgical margins, a clear trend toward worsening survival was observed in this patient population. Furthermore, patients with positive surgical margins showed decreased DFS, as well as increased rates of recurrence and distant metastases. A complete oncological resection should therefore not be compromised in favor of facial nerve preservation in tumors clearly involving the nerve. Biopsy of a suspicious-appearing nerve, including a thickened nerve or circumferential tumor involvement, can assist the surgeon in this intraoperative decision for nerve resection. The likelihood of local control without resection of a clinically or pathologically involved nerve is low; however, proximity of the tumor to the nerve alone does not warrant resection. These findings also highlight the importance of presurgical imaging and histological evaluation with fine-needle aspiration, which may provide greater information to both the patient and the operating surgeon prior to the resection of a salivary gland tumor.18, 19
Although MEC is the most common type of salivary gland carcinoma, few clinical trial data have addressed the role of systemic therapy specific to this type of tumor. This may reflect the fact that a majority of patients present with low-stage disease involving LG or IMG tumors that respond well to surgery or surgery with postoperative XRT. Similar to findings in the literature,10, 20-23 responses to systemic therapy were limited in our study. More than half of our patients receiving chemotherapy had stage IV disease on initial presentation, and all but 1 had HG tumors. These factors, no doubt, contributed to the 65% mortality rate in this patient group. Because MEC is rare, a clinical trial that addresses the role of systemic therapy and is based at a single institution is difficult, and multi-institutional cooperation is greatly needed.
Postoperative concomitant chemoradiotherapy has been recently implemented in an attempt to improve treatment outcomes for high-risk head and neck squamous cell carcinoma. The Radiation Therapy Oncology Group (RTOG) 9501/Intergroup trial demonstrated increased DFS in the combined therapy group compared with patients who received postoperative radiotherapy alone. Because those researchers observed a similar incidence of distant metastases in both groups, they proposed that the predominant effect of chemotherapy was to potentiate the locoregional effects of radiotherapy.24 In addition to increased DFS, the European Organisation for Research and Treatment of Cancer, in a similar trial, also showed statistically significant improvement in OS among patients who received postoperative concurrent chemoradiotherapy.25 Considering the poor prognosis for advanced-stage MEC, a similar combined postoperative chemoradiation regimen may prove similarly beneficial. Only a few studies have addressed this question in salivary gland carcinomas. Because these studies involved extremely small cohorts, the results are difficult to interpret.26, 27 An ongoing RTOG-sponsored trial (RTOG 1008) evaluating postoperative chemoradiation versus radiation alone for high-risk salivary gland carcinomas will perhaps offer more solid evidence for this approach. Further studies such as this are needed to address the use of novel treatment strategies for MEC.
High histological grade, advanced stage, perineural invasion, positive surgical margins, and submandibular location all portend for unfavorable outcomes among patients with MEC. These factors should therefore be taken into consideration when establishing the aggressiveness of a treatment strategy. The merging of tumor behavior between LG and IMG lesions may be attributed to earlier diagnosis and/or more successful treatment regimens for IMG tumors. Further advances in therapy are now needed to improve the outcomes for patients with HG histology and advanced-stage disease. Postoperative XRT is currently accepted as an appropriate adjuvant treatment for disease with aggressive features. However, systemic therapy has yet to play a definitive role in the successful treatment of MEC, particularly for radiosensitization and for the management of patients at high risk for distant metastasis. With the advent of molecular targeting and intensification strategies, improved control of advanced disease may be on the horizon.