BACKGROUND

This study was conducted to assess the effectiveness of the 1997 American Joint Committee on Cancer (AJCC) staging system to predict survival and local control of patients with maxillary sinus carcinoma and to identify significant factors for overall survival, local control, and distant metastases in patients with these tumors.

METHODS

Ninety-seven patients with maxillary sinus carcinoma were treated with radiotherapy at Stanford University and the University of California, San Francisco between 1959-1996. The histologic type of carcinoma among the 97 patients were: 58 squamous cell carcinomas, 4 adenocarcinomas, 16 undifferentiated carcinomas, and 19 adenoid cystic carcinomas. All patients were restaged clinically according to the 1977 and 1997 AJCC staging systems. The T classification of the tumors of the patients was as follows: 8 with T2, 18 with T3, and 71 with T4 according to the 1977 system and 8 with T2, 36 with T3, and 53 with T4 according to the 1997 system. Eleven patients had lymph node involvement at diagnosis. Thirty-six patients were treated with radiotherapy alone and 61 received a combination of surgical and radiation treatments. The median follow-up for surviving patients was 78 months.

RESULTS

The 5-year and 10-year actuarial survival rates for all patients were 34% and 31%, respectively. The 5-year survival estimate by the 1977 AJCC system (P = 0.06) was 75% for Stage II, 19% for Stage III, and 34% for Stage IV and by the 1997 AJCC system (P = 0.006) was 75% for Stage II, 37% for Stage III, and 28% for Stage IV. Significant prognostic factors for survival by multivariate analysis included age (favoring younger age, P<0.001), 1997 T classification (favoring T2-3, P = 0. 001), lymph node involvement at diagnosis (favoring N0, P = 0.002), treatment modality of the primary tumor site (favoring surgery and radiotherapy, P = 0.009), and gender (favoring female patients, P = 0.04). The overall radiation time was of borderline significance (favoring shorter time, P = 0.06). The actuarial 5-year local control rate was 43%. By the 1977 AJCC system (P = 0.78) it was 62% with T2, 36% with T3, and 45% with T4 and using the 1997 AJCC system (P = 0.29) it was 62% with T2, 53% with T3, and 36% with T4. The only significant prognostic factor for local control for all patients by multivariate analysis was local therapy, favoring surgery and radiotherapy over radiotherapy alone (P< 0.001). For patients treated with surgery, pathologic margin status correlated with local control (P = 0.007) and for patients treated with radiation alone, higher tumor dose (P = 0.007) and shorter overall treatment time (P = 0.04) were associated with fewer local recurrences. The 5-year estimate of freedom from distant metastases was 66%. The 1997 T classification, N classification, and lymph node recurrence were adverse prognostic factors for distant metastases on multivariate analysis. There were 22 complications in 16 patients, representing a 30% actuarial risk of developing late complications at 10 years.

CONCLUSIONS

The 1997 AJCC staging system was found to be superior to the 1977 AJCC staging system in predicting both survival and local control in this patient population. Combined surgical and radiation treatment to the primary tumor yielded higher survival and local control than radiotherapy alone. Other significant prognostic factors for survival were patient age, gender, and lymph node (N) classification. Prolonged overall radiation time was associated with poorer survival and local control. Late severe toxicity from the treatment of these tumors was a significant problem in long term survivors. Improved radiotherapy techniques should lead to decreased injury to the surrounding normal tissues. (c) 1999 American Cancer Society.