Prolongation of overall treatment time as a cause of treatment failure in early breast cancer: An analysis of the UK START (Standardisation of Breast Radiotherapy) trials of radiotherapy fractionation.

Abstract

BACKGROUND: Tests of tumour treatment time effect in patients prescribed post-operative radiotherapy for early breast cancer have focussed on time to start of radiotherapy rather than overall treatment time. The START randomised trials of radiotherapy fractionation provide an opportunity to directly estimate the effect of treatment acceleration. METHODS: Between 1986 and 2002, a total of 5861 women with early breast cancer were recruited into the UK START pilot (START-P), START-A and START-B randomised trials. START-P and START-A tested 13 fractions of 3.0-3.3Gy against 25 fractions of 2.0Gy with a fixed treatment duration of 5weeks for all schedules; START-B tested 15 fractions of 2.67Gy in 3weeks against 25 fractions of 2.0Gy over 5weeks. Estimates of the effect of length of treatment for local-regional relapse and for a measure of late normal tissue effects (change in photographic breast appearance, for patients following breast conserving surgery) were obtained from Cox proportional hazards regression analyses stratified according to trial. RESULTS: At a median follow-up of 10years, 444/5831 (7.6%) patients with data available had a local-regional relapse, and 1135/3185 (35.6%) had mild or marked change in photographic breast appearance by 5years. Adjusting for prognostic factors, the estimate of the overall treatment time effect for local-regional relapse was 0.60Gy/day (95%CI 0.10 to 1.18Gy/day, p=0.02), and 0.14Gy/day (95%CI -0.09 to 0.34Gy/day, p=0.29) for change in photographic breast appearance. CONCLUSIONS: Combined analysis of the START trials generates the hypothesis that overall treatment time is a significant determinant of local cancer control after adjuvant whole breast radiotherapy, with approximately 0.6Gy per day 'wasted' in compensating for tumour cell proliferation.