Linear-quadratic analysis of tumour response to fractionated radiotherapy: a study on human squamous cell carcinoma xenografts.

PURPOSE: To compare values for the alpha/beta ratio in experimental tumours irradiated either under conditions of clamping and short overall time or under more 'clinically realistic' conditions.
MATERIALS AND METHODS: Human squamous cell carcinomas, FaDu and GL, were grown in nude mice. Alpha/beta values were determined from local tumour control data after treatment with single doses and 2, 4, and 8 fractions under clamp hypoxia in 3.5 days, using maximum likelihood analysis. Effective alpha/beta values (alpha/beta(eff)) were determined from treatment with 12, 30, and 60 fractions under ambient conditions in a constant overall treatment time of 6 weeks.
RESULTS: After correction for an oxygen enhancement ratio of 2.7 the alpha/beta values were 15 Gy (95% CI 9; 24) for FaDu and 49 Gy (26; 122) for GL. In FaDu the TCD50 values after 12 to 60 fractions were not significantly different, the alpha/beta(eff) value was infinite (52; inf.). Unexpected from the high alpha/beta value, the TCD50 values of GL tumours increased from 37 Gy (28; 47) after 12 fractions to 59 Gy (52; 67) after 60 fractions: the alpha/beta(eff) value was 3 Gy (0.6; 12 Gy).
CONCLUSIONS: The results support the view that mechanisms other than recovery from sublethal radiation damage and repopulation of clonogenic tumour cells may importantly impact on treatment outcome when the number of fractions is changed in clinical radiotherapy.