Clinical data on head and neck tumours presented by Withers et al. (The hazard of accelerated tumour clonogen repopulation during radiotherapy. Acta Oncol. 27: 131-146, 1988) are reexamined. Recently it was argued that the relationship between TCD50 and treatment time might be influenced by the practice of prescription [3,4]. This paper investigates the role of local tumour control rate in the interpretation of the data. In the genuine data set local tumour control rates were replaced by either tumour failure rate or by random numbers. TCD50 values were calculated from the original and these manipulated data and related to treatment duration. TCD50 increases with about 0.48 Gy/day in all cases. This unexpected result is explained by the lack of a dose-response relationship which causes a highly significant correlation between TCD50 and prescribed dose. Consistently, multiple regression analysis reveals a significant impact of prescribed dose (p < 0.0001) on TCD50 but not of treatment duration (p > 0.05). The increase of TCD50 with treatment time reflects solely dose-time prescription habits. The data provide no significant evidence either for an impact of treatment duration or for repopulation during radiotherapy in head and neck tumours. From this lack of evidence it may not be concluded that treatment duration is an unimportant parameter in radiotherapy.
