Comparison of treatment techniques for conformal radiotherapy of the prostate using dose-volume histograms and normal tissue complication probabilities.

The aim of this study was to evaluate the relative merits of the coplanar field arrangements most frequently used for conformal radiotherapy of the prostate using dose-volume histograms and normal tissue complication probabilities (NTCPs). Twelve patients with early prostate cancer underwent a planning CT scan of the pelvis. Isocentric plans for each patient were devised using three, four, six and eight conformal fields and beam-weights optimised using fast simulated annealing to give a dose homogeneity across the planning target volume of +/- 5% or better while minimising irradiation of the relevant organs at risk. The plans were then evaluated using dose-volume histograms of the organs at risk (bladder, rectum and both femoral heads) and the Lyman model of normal tissue complication probability for the rectum. Analysis of dose-volume histogram data averaged over the 12 patients indicates an advantage for six (p = 0.002) and eight (p = 0.0001) fields with respect to the percentage volume of the femoral heads receiving > 50% of the prescribed dose compared with three fields. There was a similar advantage for six (p = 0.0007) and eight (p = 0.0001) fields compared with four fields. Ranking of the treatment techniques indicates that the four-field technique is the worst with respect to femoral head irradiation but the best with respect to reducing rectal irradiation. A higher dose can be prescribed to the isocentre with the four-field technique for a 5% rectal NTCP. The six-field technique led to sparing of the bladder when the different treatment techniques were ranked using bladder dose-volume histogram data. We conclude that none of the techniques studied consistently proved to be superior when applied to this sample of patients with prostate cancer with respect to sparing all the organs at risk. The absolute differences between techniques are small and would be very difficult to detect with respect to clinically relevant endpoints.