Background: The detailed summary of results from Bristol where 270 patients with carcinoma of the cervix were treated with either 75 cGy/h from manually loaded caesium or 150 cGy/h by remote afterloading (Newman, G. Increased morbidity following the introduction of remote afterloading, with increased dose rate, for cancer of the cervix. Radiother. Oncol. 39:97-103, 1996) can be analysed to study the radiobiological factors that could have contributed to the different outcomes reported. The increases in grade 3 late complications from 4 to 22%, and in grade 2 + 3 from 12 to 32%, in spite of a reduction of 20% in dose, imply a rather large difference in biological effect between the two systems, which might or might not be due to dose rate differences. Purpose: The possibility that a significant part of the effect might have been due to the dose rate is investigated, with particular attention to what values of radiobiological parameters might explain it, or can be excluded. A similar set of clinical data from Tilsburg (Rodrigus, P., de Winter, K., Venselaar, J. and Leers, W.H. Evaluation of late morbidity in patients with carcinoma of the uterine cervix following a dose rate change. Radiother. Oncol. 42:137-141, 1997), where dose rate was doubled from 54 to 107 cGy/h with a dose reduction of 20%, is also considered. Methods: Linear quadratic modelling is employed to calculate Biologically Effective Doses (or ERDs) corresponding to the clinical protocols used. Repair of sublethal radiation damage at a range of half-times is assumed, both for mono-and bi-exponential components. When the LDR is doubled it is called MDR in the present study. Results: The maximum ratios calculated for the BEDs of 16 Gy at MDR to 20 Gy at LDR were 1.06-1.15, assuming alpha/beta = 4-2 Gy, the latter being an unlikely extreme for rectal or urinary complications. These maxima occurred over a narrow range of t(1/2) values from 1.5 to 2.5 h. If t(1/2) were as long as 4-7 h or as short as 0.5-0.75 h, the biological effects would have been equal. The theoretically ideal dose reduction factors, calculated using the t(1/2) values derived from the clinical data, are in the range of 24-29% instead of 20%. Conclusions: Somewhat greater dose reduction factors for late complications were suggested by this analysis than the 20% that has been commonly used when the dose rate is increased, both from the Bristol and Tilsburg data. The Bristol data showed no loss of therapeutic ratio on changing from manual to remote afterloading. Due to the close-to-optimum choice of the dose reduction factor which was actually used, some values for the half-time of repair of late complications in gynaecological brachytherapy could be estimated and used to calculate the theoretical dose reduction factors. (C) 1997 Elsevier Science Ireland Ltd.
