Purpose: The objective of this study was to examine the incidence of late effects of hyperfractionated radiotherapy for head and neck cancer as a function of the dose delivered, as well as the daily interfraction interval. In addition, we wished to examine the influence of other prognostic factors including age, gender, primary site, T- and N-stage, and overall stage on the late effects of hyperfractionated radiotherapy.
Methods and materials: Between 1983 and 1987, 479 patients with advanced head and neck cancer were entered on a Phase ILE/II dose escalation trial of hyperfractionated radiotherapy. They were randomly assigned to receive a dose of 67.2, 72.0, 76.8, or 81.6 Gy, delivered at 1.2 Gy/fraction, twice a day (BID), 5 days/week. Of the 451 analyzable patients, 399 patients who received > or = 64.8 Gy and had a follow-up > 90 days were eligible for this study. Acute and late effects were scored with the RTOG/EORTC late radiation morbidity scoring scheme. For this analysis, patients were subclassified by the actual doses delivered and by an average daily interfraction interval of < or = 4.5 h or > 4.5 h. The incidence of late effects was estimated using a cumulative incidence approach.
Results: Fifty-nine patients received 67.2 +/- 2.4 Gy, 119 received 72.0 +/- 2.4 Gy, 98 received 76.8 +/- 2.4 Gy, and 123 received 81.6 +/- 2.4 Gy. The proportion of patients treated with a daily interfraction interval of > 4.5 h was 32, 50, 43, and 71%, respectively. The four treatment groups were well balanced with respect to pretreatment characteristics. The median follow-up was 1.71 years (range: 0.24-9.6) for all evaluable patients and 6.12 years for 85 alive patients. There was no significant difference in the incidence of late effects between the different dose levels. At 5 years, the cumulative incidence of late effects was 17, 14, 20, and 13% for grade 3, and 7, 3, 7, and 5% for grade 4. However, the incidence of late effects differed significantly with respect to daily interfraction interval. The cumulative incidence of grade 4 late effects increased from 6.3% at 2 years to 7.5% at 3 years to 8.0% at 4 years and 8.6% at 5 years with an interval of < or = 4.5 h, while it remained at a constant of 2.0% with an interval of > 4.5 h during the same period (p = 0.0036). Multivariate analysis showed that among the prognostic factors examined, daily interfraction interval of < or = 4.5 h was the only significant independent prognostic factor for the development of grade 3+ or grade 4 late effects (p = 0.0167 and p = 0.0013, respectively).
Conclusion: Results of this randomized Phase ILE/II trial of hyperfractionated radiotherapy in head and neck cancer showed no apparent dose-response relationship for late effects within the range of 67.2-81.6 Gy. Daily interfraction interval was a significant independent factor for the development of late effects in a multivariate analysis.