Acute toxicity and cost analysis of a phase III randomized trial of accelerated and conventional radiotherapy for squamous carcinoma of the head and neck: a Trans-Tasman Radiation Oncology Group study.

The primary purpose of the present analysis was to assess the feasibility and acute toxicity of a pure accelerated fractionation regimen in a cooperative group setting. This analysis included the first 320 patients entered on to the Trans-Tasman Radiation Oncology Group (TROG) randomized controlled trial which compared accelerated radiotherapy (ART) with conventional radiotherapy (CRT) in stage III and IV squamous cell carcinoma (SCC) of the head and neck. Patients were randomized to either 59.4 Gy in 33 fractions over 24 days (ART) or to 70 Gy 35 fractions over 49 days (CRT) after being stratified for site and stage. Accrual began in 1991 and the trial was closed on 3 April 1998 with the targeted 350 patients. The 3-year survival for the whole group was 54%, and the 3-year disease-free survival was 41%. Toxicity data were available on 303 patients (148 ART; 155 CRT). Mucosal toxicity was worse in the accelerated arm, and it peaked approximately 3 weeks earlier than the conventional arm. Skin toxicity was equivalent but occurred approximately 7 days earlier in the accelerated arm. Acute effects in both arms healed completely. Hospitalization was more common in the ART arm (71 vs 52 patients; P = 0.01) but the total bed days in hospital was not greatly different (1707 bed days for ART and 1607 bed days for CRT). Patients were more likely to require nasogastric (NG) feeding in the ART arm (49 vs 33 patients; P = 0.02). There were 1157 NG feeding days for ART and 1154 NG feeding days for CRT. The average cost of radiation treatment per patient including hospitalization, NG feeding and accommodation was $11,750 in the ART arm and $11,587 in the CRT arm. The accelerated arm has been shown to be a tolerable, practical and cost-equivalent regimen. The assessment of the therapeutic ratio of this accelerated protocol (ART) will be determined when the analysis of late effects and loco-regional control is made when the data are more mature.