The role of postradiotherapy neck dissection in supraglottic carcinoma

Purpose: To evaluate our policy of performing neck dissection based on regional response after definitive radiotherapy in patients with supraglottic carcinoma and to identify the prognostic factors in this group of patients. Methods and Materials: Between 1970 and 1995, 121 patients with node-positive squamous cell carcinoma of the supraglottic larynx were treated with definitive radiotherapy. Sixty-nine percent of patients presented with 1997 AJCC Stage IV disease. The N-stage distribution was N1, 49; N2, 62; and N3, 10. The median size of the lymph nodes was 3 cm (range, 0.5-8 cm). Forty-five patients received once-a-day treatment with a median total dose of 65 Gy (range, 58.0-70.8 Gy) in 1.8-2.0 Gy per fraction over 48 days, and 76 patients received split-course accelerated hyperfractionation with a median total dose of 67.2 Gy (range, 63.2-73.6 Gy) in 1.6 Gy twice a day over 43 days. Patients whose lymph nodes were not clinically detectable at 4-6 weeks after the completion of radiotherapy (complete response) were followed without any neck dissection. Patients with persistent neck adenopathy (partial response) underwent neck dissection whenever possible. Mean follow-up of the living patients was 6.5 years. Results: Regional response was related to the size of lymph nodes at presentation. Eighty-seven percent of patients with nodal size of 3 cm or less had a complete response, whereas 43% of patients with nodal size greater than 3 cm had a partial response. The rate of regional control at 3 years for all patients in the study was 66%. The 3-year ultimate regional control rate after salvage neck dissection was 75%. A relapse in both the primary and regional sites was the most common pattern of relapse, accounting for 39% of all the failures. Local failure was associated with subsequent regional relapse with a relative risk of 4.3. For patients with complete response in whom postradiotherapy neck dissection was withheld, the regional control rates were 75% and 86% for N1 and N2, respectively. The rate of isolated regional relapse in this group of patients was 7.5%. In multivariate analysis, significant favorable factors predictive for regional control were female gender, accelerated hyperfractionation, and complete response; whereas factors predictive for overall survival were Karnofsky Performance Scale score and regional response. The rate of Radiation Therapy Oncology Group (RTOG) Grade 2 or 3 neck fibrosis was 17% and 23% for patients with and without postradiotherapy neck dissection, respectively. Conclusion: Isolated regional relapse is not common in patients with supraglottic carcinoma when a complete response is achieved at 4-6 weeks after definitive radiotherapy and postradiotherapy neck dissection is not performed. Female gender, accelerated hyperfractionation, and complete response are favorable predictors of regional control. (C) 2001 Elsevier Science Inc.