A comparison of accuracy of computer treatment planning systems in brachytherapy

An intercomparison was made between dose-rate point calculations with five commercial computer treatment planning systems used in brachytherapy. Sixteen Italian radiotherapy medical physics institutions performed a twofold test, respectively for point and linear sources, reproducing actual clinical implants and the results were compared to references values. The test was designed in order to determine the errors introduced separately by computation and by implant reconstruction via orthogonal films. The results show that for point sources a better accuracy can be achieved than for linear sources. It is also shown that a large variation of results exists between computation systems and the variability is larger for linear sources. The digitizing procedure adds a large amount of error. In the whole set of calculation points, the overall percent difference between computed and reference dose-rate values is larger than +/- 5% in 18% cases for point and in 52% cases for linear sources, with data input by digitizer. It seems that relevant errors do occur when computing actual dose rates, resulting in clinically relevant inaccuracies in the calculated absorbed dose and in its relative effectiveness.