A simulation technique for computation of the dosimetric effects of setup, organ motion and delineation uncertainties in radiotherapy.

In this study, we introduce a novel simulation technique to incorporate delineation errors into radiotherapy treatment margins and combine them with organ motion and set-up errors to investigate the cumulative dosimetric effects in different tumour sites. The effects of applying patient realignment correction protocols for radical treatments of prostate, lung and brain tumours were also modelled. Simulations were based on data from measurements using image-guidance techniques, including the use of fiducial markers in prostate and breathing correction techniques for the lung. The use of different sizes of planning target volume (PTV) margins was also evaluated. The prostate clinical target volumes' V99% showed up to 3.2% improvement with reduction in treatment uncertainties. For the lung plans, the V99% increased by up to an average of 10% with increase in treatment margin size from 0.5 to 1.5 cm. This improvement was, however, at the detriment of the dose delivered to the critical organs where the maximum dose received by the spinal cord increased by up to 0.5 Gy per fraction. These results were used to deduce the possible margin reductions and dose escalation achievable with reduced uncertainties.

A fuzzy convolution model for radiobiologically optimized radiotherapy margins.

In this study we investigate the use of a new knowledge-based fuzzy logic technique to derive radiotherapy margins based on radiotherapy uncertainties and their radiobiological effects. The main radiotherapy uncertainties considered and used to build the model were delineation, set-up and organ motion-induced errors. The radiobiological effects of these combined errors, in terms of prostate tumour control probability and rectal normal tissue complication probability, were used to formulate the rule base and membership functions for a Sugeno type fuzzy system linking the error effect to the treatment margin. The defuzzified output was optimized by convolving it with a Gaussian convolution kernel to give a uniformly varying transfer function which was used to calculate the required treatment margins. The margin derived using the fuzzy technique showed good agreement compared to current prostate margins based on the commonly used margin formulation proposed by van Herk et al (2000 Int. J. Radiat. Oncol. Biol. Phys. 47 1121-35), and has nonlinear variation above combined errors of 5 mm standard deviation. The derived margin is on average 0.5 mm bigger than currently used margins in the region of small treatment uncertainties where margin reduction would be applicable. The new margin was applied in an intensity modulated radiotherapy prostate treatment planning example where margin reduction and a dose escalation regime were implemented, and by inducing equivalent treatment uncertainties, the resulting target and organs at risk doses were found to compare well to results obtained using currently recommended margins.