Epithelial-myoepithelial carcinoma is an uncommon biphasic neoplasm most commonly arising in salivary tissue. It represents approximately 1% of all salivary neoplasms [2, 3]. In their review of 246 cases of salivary gland epithelial-myoepithelial carcinoma from the 1973–2010 SEER database, Vazquez et al.  noted that 57.3% of patients were female. The present study noted a similar female predominance, with 62.2% of patients being female. Vazquez et al. noted a rate of distant metastases at diagnosis of 4.5%. The present cohort shows an even lower M1 rate of 2.6%. Overall disease-specific survival at 60 and 120 months in the Vazquez study was 91.3% and 90.2%. The present study showed a lower overall survival of 72.7% and 59.5% at 60 and 120 months. This may be due to the inclusion in the present study of all sites including unknown primary, lung, and gastrointestinal tumors. These sites were less prevalent than salivary primary sites such as parotid and submandibular gland but had lower mean survival times, likely leading to the lower overall survival for the entire cohort. Additionally, the updated 1973–2014 data included in the present study showed a lower overall survival for parotid epithelial-myoepithelial carcinomas than submandibular and minor salivary gland tumors. In the 2015 study  patients with low-grade histology and lesions < 4 cm had significantly better survival vs. patients with high-grade histology and tumors > 4 cm. In the present study univariate analysis showed that T stage, N stage/presence of nodal metastases, M stage/presence of distant metastases, surgical treatment vs. nonsurgical treatment, overall AJCC stage, race, and age at diagnosis all significantly affected survival. Specifically, patients with T2, T3, and T4 tumors showed significantly shorter survival than patients with T1 tumors, patients with M1 tumors showed significantly shorter survival than M0 patients, and patients with nodal metastases showed significantly shorter survival than patients who were N0. Chen et al.  reported a rare case of a patient with a primary base of tongue epithelial-myoepithelial carcinoma with neck metastases and multiple lung nodules (T4N2cM1). The patient refused treatment other than biopsy and expired at 18 months post-diagnosis. In the present cohort parotid and submandibular gland were the two most common primary sites, but a statistically significant difference in survival by primary site was not found in the present study. Patients treated with surgery showed significantly increased survival relative to patients in whom surgery was not recommended, recommended but not performed for an unknown reason or in patients for whom surgery was recommended but it was unknown whether surgery was actually performed. Additionally, black patients and patients age 80 or greater at time of diagnosis had significantly shorter survival than other patients. These findings remained significant on multivariate logistic regression analysis, with overall AJCC stage, T, N, and M stage, age > 80, black race, and nonsurgical treatment significantly affecting survival.
It is interesting to note that between 1973 and 2010 there were 246 salivary gland epithelial-myoepithelial carcinomas reported (an average of approximately 6.6 per year), while in the updated data in the present study a further 70 epithelial-myoepithelial carcinomas of the major salivary glands (parotid + submandibular gland) were recorded in the next 4 years from 2010 to 2014 (an average of approximately 17.5 per year). This relative increase in numbers from 2010 to 2014 may be due to several factors. One significant factor may be Hurricane Katrina. Hurricane Katrina struck the Gulf Coast of the United States from August 23 to August 31, 2005 and significantly affected central Florida to Texas. These areas contained several major hospitals, and the SEER database specifically notes in their data set that Hurricane Katrina significantly impacted the Louisiana Cancer Registry’s data reporting for the second half of 2005, and SEER excluded Louisiana cases diagnosed from July–December 2005. As this period is contained within the 1973–2010 period, this may have had some effect on the epithelial-myoepithelial carcinoma cases reported during this time period. Additionally, increased familiarity by pathologists with epithelial-myoepithelial carcinoma may have led to increased histopathological identification of this rare tumor, and decreased misdiagnosis as more common neoplasms such as mucoepidermoid carcinoma or adenoid cystic carcinoma. Finally, the increased availability and use of imaging modalities such as ultrasound, computed tomography, and magnetic resonance imaging even at smaller regional centers may have led to both increased diagnostic sensitivity for epithelial-myoepithelial carcinomas presenting as neck masses, etc., and may have also led to increased discovery of epithelial-myoepithelial carcinomas as “incidentalomas” discovered on diagnostic imaging ordered for other, unrelated reasons.
This study has several limitations. The retrospective data makes recall bias and selection bias a possibility, and the reporting limitations of the SEER database meant that a minority of the cohort had missing TNM/overall stage data. However, SEER data has been frequently used to evaluate factors affecting patient survival in rare tumors for which randomized prospective trials would be difficult to assemble [2, 4, 5]. The survival analysis by treatment may be confounded by co-morbidities in some patients, particularly patients in whom surgery was not recommended.