Kaplan-Meier survival distribution for patients with maxillary sinus cancer according to histologic type. Abbreviations: ACC, adenoid cystic carcinoma; Aden, adenocarcinoma; MEC, mucoepidermoid carcinoma; Mel, melanoma; Sarc, sarcoma; SCC, squamous cell carcinoma; Undiff, undifferentiated carcinoma.
Bhattacharyya N. Survival and Staging Characteristics for Non–Squamous Cell Malignancies of the Maxillary Sinus. Arch Otolaryngol Head Neck Surg. 2003;129(3):334-337. doi:10.1001/archotol.129.3.334
To characterize the clinical behavior of non–squamous cell cancer of the maxillary sinus.
Cases of non–squamous cell maxillary sinus malignancy during 1988 through 1998 were extracted from the Surveillance, Epidemiology, and End Results database. Data for histologic type of tumor, tumor stage, and survival were analyzed using the Kaplan-Meier method to determine mean, median, and 5-year survival statistics for the overall cohort and for individual histologic types of tumor. The effect of tumor stage on overall survival was assessed. To determine relative survival, the non–squamous cell group was compared with a second group of patients with squamous cell carcinoma, matched according to age at diagnosis, year of diagnosis, and T stage.
One hundred eighty-eight cases of non–squamous cell malignancy were identified. The mean patient age was 57.8 years, and 143 patients (76%) presented with T3 or T4 tumors. There were 31 adenocarcinomas, 64 adenoid cystic carcinomas, 15 mucoepidermoid carcinomas, 22 melanomas, 45 sarcomas, and 11 undifferentiated carcinomas. The overall mean survival was 63.4 months, and 5-year survival was 45.6%. Adenoid cystic carcinoma exhibited the best mean survival (79 months), whereas melanoma and undifferentiated carcinoma exhibited poor mean survivals (30.3 and 12.8 months, respectively). T stage did not statistically affect overall survival (P = .86). Survival for patients with non–squamous cell cancer was substantially better than survival for the matched group with squamous cell carcinomas (mean survival, 41.5 months; 5-year survival, 27.4%).
Patients with non–squamous cell carcinoma of the maxillary sinus typically present with advanced T stage but have significantly better survival than patients with similar-stage squamous cell carcinoma of the maxillary sinus. T stage has less effect than histologic type of tumor on prognosis in non–squamous cell carcinoma of the maxillary sinus.
MALIGNANT TUMORS of the nasal cavity and paranasal sinuses constitute fewer than 1% of all malignancies and 3% of upper aerodigestive tract malignancies.1,2 Because of the overall rarity of these tumors, development of an understanding of their clinical behavior, response to therapy, and overall prognosis may be difficult. Malignancies of the maxillary sinus constitute the largest fraction of these paranasal sinus and nasal cavity malignancies. Although squamous cell carcinoma is the most common malignancy of the maxillary sinus, several other histologic types of tumor may be encountered as well. These include minor salivary gland malignancies, sarcomas, melanomas, and undifferentiated carcinomas. However, survival statistics for these rare histologic types within the maxillary sinus are limited. To appropriately counsel patients who are newly diagnosed as having non–squamous cell carcinoma of the maxillary sinus, survival data based on tumor characteristics, reinforced with adequate follow-up, are required.
To better delineate current clinical trends and prognostic variables in non–squamous cell malignancies of the maxillary sinus, we examined the Surveillance, Epidemiology, and End Results (SEER) database.3 Maintained by the National Cancer Institute, Bethesda, Md, and rigorously updated, this database is regarded as a gold-standard database for cancer surveillance in the United States. The SEER database samples several US geographic areas, representing an estimated 10% of the US population, and tracks several million cancers. Therefore, the SEER registry serves as an excellent source for the study of rare tumors, such as maxillary sinus malignancies. This database has been previously used to study cancer of the nasal cavity, major salivary gland malignancy, and other head and neck tumors.4- 10
The SEER database for 1988 through 1998 was examined. All cases of maxillary sinus malignancy (International Classification of Diseases for Oncology11 tumor code C31.0) were extracted from this database. Fields extracted from the database included standard demographic information (age, sex, and year of diagnosis) as well as tumor-dependent variables, including histologic type of tumor, tumor grade, and extent of disease. Survival time and vital status were similarly extracted.
All cases of non–squamous cell cancer of the maxillary sinus were then selected for analysis. Tumor analysis was restricted to those histologic types for which more than 10 cases were identified. Histologic types were decoded according to the International Classification of Diseases for Oncology. From the SEER extent-of-disease variable for the primary site, T stage was determined according to 1997 American Joint Committee on Cancer12 criteria for the staging of maxillary sinus malignancy. The extent-of-disease variable is a 2-digit field contained within the SEER database that is uniquely coded according to each primary site. The method for coding of this variable into T stage for the maxillary sinus is depicted in Table 1. Patients with indeterminate T stage or with distant metastatic disease were excluded from subsequent analysis.
A second reference group of patients with squamous cell carcinoma was also assembled from the SEER database. For each patient with non–squamous cell malignancy of the maxillary sinus, a patient with squamous cell carcinoma was randomly selected, matching for age (±5 years), year of diagnosis (±5 years), and T stage.
The data were then imported into the Statistical Package for the Social Sciences (version 10.0; SPSS Inc, Chicago, Ill) for subsequent processing. Standard descriptive statistics were computed for the variables under consideration. Kaplan-Meier survival analysis was conducted for the entire cohort and for each histologic type of tumor. Univariate comparisons among T stage and histologic types of tumor were conducted using the log-rank test.
To determine if patients with non–squamous cell malignancies of the maxillary sinus differed in survival relative to those with squamous cell carcinoma, comparative Kaplan-Meier survival analysis was conducted for the matched groups of patients with non–squamous cell and squamous cell carcinoma. Evaluation of survival differences was conducted using the log-rank test.
From the SEER database for 1988 through 1998, 188 non–squamous cell malignancies were identified. The mean patient age was 57.8 years, with 100 male and 88 female patients. The distribution of the histologic types of tumor is delineated in Table 2, along with T stage by histologic type. Most patients presented with either T3 or T4 disease. Patients with adenoid cystic carcinomas and sarcomas were particularly likely to present with advanced T-stage disease.
The mean Kaplan-Meier survival estimate for the overall cohort was 63.4 months (95% confidence interval, 55.1-71.8 months), with a median survival of 50.0 months.
The mean follow-up for patients alive at the end of the period was 47.0 months (range, 1-125 months), and no patients were unavailable for follow-up during the period. The 5-year survival for the overall cohort was 45.6%. Figure 1 depicts the Kaplan-Meier survival curves according to histologic type for patients with squamous and non–squamous cell tumors, and Table 3 lists mean, median, and 5-year survivals for each histologic type. Kaplan-Meier estimates of survival according to the extent (T stage) of the primary site across all histologic types are presented in Table 4. T stage was not predictive of survival in patients with non–squamous cell malignancy of the maxillary sinus (P = .86, log-rank statistic). Among 94 patients for whom nodal status was available, 1 patient had N1 disease, 4 had N2 disease, and 1 had N3 disease; the remaining cases were N0.
Appropriate matching to patients with squamous cell carcinoma according to age at diagnosis, T stage, and year of diagnosis was achieved for 158 of the patients with non–squamous cell malignancies. For patients with non–squamous cell sinus malignancies, the mean survival time was 61 months (95% confidence interval, 52-70 months) and the median survival was 51 months. In contrast, the mean survival for those with squamous cell carcinoma of the maxillary sinus was 42 months (95% confidence interval, 33-50 months), with a median survival of 20 months. The more favorable survival for patients with non–squamous cell malignancies of the maxillary sinus was statistically significant (P = .005).
As a subset of maxillary sinus malignancies, non–squamous cell cancers of the maxillary sinus are rare entities. Therefore, limited data exist in the literature regarding survival for patients with these lesions. In addition, the bone and soft tissues of the maxillary sinus may give rise to several different histologic types of malignant tumors. These include sarcomas, melanomas, carcinomas, and minor salivary gland malignancies. Although staging and survival for patients with squamous cell carcinoma of the maxillary sinus have been well established, development of survival data for patients with non–squamous cell malignancies of the maxillary sinus is hampered by the rarity and variety of these malignancies. In many instances, small numbers of patients with non–squamous cell malignancies of the maxillary sinus have been aggregated with those with squamous cell cancers, rather than reported separately.13- 17 To provide survival data that are useful in counseling patients with respect to prognosis, separate samples for these varying non–squamous cell histologic types and adequate follow-up are required. A previous study18 of non–squamous cell malignancies of the paranasal sinuses has indicated a similar distribution of histologic types of tumors, with sarcomas, adenoid cystic carcinomas, and adenocarcinoma accounting for most cases. Many of these non–squamous cell histologic types have distinctive biological behaviors.
Patients with adenoid cystic carcinoma had the longest mean and median survivals. Limited data exist on the survival and outcome of patients with adenoid cystic carcinoma of the maxillary sinus. In a study of 22 patients, Kim et al19 demonstrated an overall survival at 10 years of 37.6% for patients with adenoid cystic carcinoma of the maxillary antrum. Treatment failures occurred after 5 years, with primary site failure (73%) and distant metastasis (32%) being common. However, with adenoid cystic carcinoma, patients may live lengthy periods with metastatic disease, and 10-year survival statistics are often more representative.20 Other investigations have revealed similar findings.21
Survival for patients with undifferentiated carcinoma of the maxillary sinus was poor. Previous reports22- 24 have confirmed the aggressive behavior of sinonasal undifferentiated carcinoma, with the mean survival for patients with sinonasal undifferentiated carcinoma ranging from 11 to 15 months, with a large fraction of patients exhibiting eventual intracranial extension. Sinonasal mucosal melanoma accounts for only 1% of all melanomas. In a recent meta-analysis25 of sinonasal melanoma, patients with these tumors had a poor prognosis, with only 30% of patients living to the 5-year actuarial mark, a figure that is confirmed by the present data. Mucoepidermoid carcinoma and sarcomas of the maxillary sinus have received little attention in the literature. Traditionally, sarcomas of the paranasal sinuses have been associated with a dismal prognosis.18 The present data provide baseline information regarding survival for these 2 histologic types, but further study is required to determine survival and response to therapy for patients with these rare tumors.
Although non–squamous cell malignancies of the maxillary sinus are much rarer than squamous cell carcinoma, their prognoses in patients matched for age and T stage are substantially better than those of patients with squamous cell carcinoma of the maxillary sinus. Overall survival for patients with squamous cell carcinoma of the maxillary sinus has been traditionally poor, with expected 5-year survival ranging from 34% to 49%.1,14,17,26 Although the patients on whom the present data are based were not matched for N stage, cervical metastases in squamous and non–squamous cell malignancies of the maxillary sinus occur in 4% to 16% of patients and would likely have a limited effect on survival comparisons between the 2 groups.2,15
For non–squamous cell malignancy of the maxillary sinus, T stage did not statistically affect overall survival. The present data suggest that T stage for non–squamous cell malignancy may not be that helpful in determining patient survival. In fact, when the Kaplan-Meier analysis of survival according to T stage was conducted separately for each histologic type of tumor, for no type was the T stage a statistically significant predictor of overall survival (all P>.05). This finding suggests that the individual histobiological type of tumor, rather than extent of disease, is the important factor in determining overall survival in patients with non–squamous cell malignancies. For example, adenoid cystic carcinoma may exhibit significant microscopic perineural spread with T1 or T2 disease. Similarly, mucosal melanoma may spread with distant metastasis from a small maxillary tumor site. Therefore, patients with non–squamous cell malignancies of the maxillary sinus should be approached more from the standpoint of their histologic type of tumor and tumor-specific biological behavior, rather than their T stage of disease.
Although the present analysis provides a large number of patients with non–squamous cell malignancies diagnosed in a contemporary period, it has several limitations. The SEER database is a well-maintained and up-to-date clinical database of many malignancies. One significant drawback is that it is not possible to determine the effects of type of surgery and surgical margin status on survival. The status of surgical margins at the time of definitive surgery has been shown to have a significant effect on local recurrence, but its effect on overall survival in patients with maxillary sinus malignancy has been debated.26,27 Using the SEER database, it is not possible to determine the effect of margin status on survival. Furthermore, there are no fields in the SEER database for recording chemotherapy as a treatment modality; therefore, we cannot assess the effect of chemotherapy on overall survival. However, the effect of chemotherapy on survival is likely to be limited, as chemotherapy is not widely used in the United States for maxillary sinus malignancies.
Adenoid cystic carcinoma is the most common non–squamous cell carcinoma of the maxillary sinus. Survival for patients with undifferentiated carcinoma and melanoma involving the maxillary sinus is poor. Traditional T stage may have less bearing on overall survival for patients with non–squamous cell carcinoma compared with patients with squamous cell carcinoma of the maxillary sinus.
Corresponding author: Neil Bhattacharyya, MD, Division of Otolaryngology, Department of Otology and Laryngology, Brigham and Women's Hospital, Harvard Medical School, 333 Longwood Ave, Boston, MA 02115 (e-mail: firstname.lastname@example.org).
Accepted for publication July 25, 2002.