A linear quadratic analysis of gynaecological brachytherapy

Four hundred and fourteen patients were treated by radical radiotherapy alone for cervical carcinoma at the Western Infirmary and the Royal Beatson Memorial Hospital between April 1982 and December 1987. All patients received external beam radiotherapy in addition to brachytherapy, using either manually inserted caesium (n = 107) or the Selectron afterloading machine (n = 307). Three mean Selectron dose rates were used: 0.91 Gy/h, 1.195 Gy/h and 1.74 Gy/h. During this period of time, the cumulative radiation effect formula was used to calculate an overall brachytherapy dose reduction to compensate for the increase in point 'A' dose rate. We have compared the local control rates and the incidence of late effects seen in these patients with theoretical parameters calculated using the linear quadratic (LQ) model. This model predicts a small rise in late effects as the dose rate increases, which is also seen in clinical practice; it also predicts a reduction in local control, partially offset by the addition of external beam radiotherapy, which would be most marked for early stage disease. There was a small fall in local control associated with Selectron treatment, but of a smaller size than predicted by the LQ model.