A B S T R A C TBackground and purpose: Inter-institutional studies highlighted correlation between consistent radiotherapy quality and improved overall patient survival. In treatment planning automation has the potential to address differences due to user-experience and training, promoting standardisation. The aim of this study was to evaluate implementation and clinical effect of a multicentre collaboratively-developed automated planning model for Intensity-Modulated Radiation Therapy/ Volumetric-Modulated Arc Therapy of prostate. The model was built using a variety of public institutions' clinical plans, incorporating different contouring and dose protocols, aiming at minimising their variation. Methods and materials: A model using 110 clinically approved and treated prostate plans provided by different radiotherapy centres was built with RapidPlan (RP), for use on intact and post-prostatectomy prostate cases. The model was validated, distributed and introduced into clinical practice in all institutions. To investigate its impact a total of 126 patients, originally manually inverse planned (OP), were replanned using RP without additional planner manual intervention. Target and organat-risk (OAR) metrics were statistically compared between original and automated plans. Results: For all centres combined and individually, RP provided plans comparable or superior to OP for all dose metrics. Statistically significant reductions with RP were found in bladder (V40Gy) and rectal (V50Gy) low doses (within 2.3% and 3.4% for combined and 4% and 10% individually). No clinically significant changes were seen for the PTV, independently of seminal vesicle inclusion. Conclusion: This project showed it is feasible to develop, share and implement RP models created with plans from different institutions treated with a variety of techniques and dose protocols, with the potential of improving treatment planning results and/or efficiency despite the original variability.

No abstract

A new method to measure the effect of the backscatter into the beam monitor chambers in linear accelerators is introduced from first principles. The technique, applicable to high-energy photon beams, is similar to the well-known telescopic method although here the heavy blocks are replaced by a very small, centred block on the shadow tray, thus the name 'ecliptic method'. This effect, caused mainly by backscattering from the secondary collimators, is known to be an output factor constituent and must be accounted for when detailed calculations involving the machine's head are required. Since its magnitude is generally small, experimental errors might obscure the behaviour of the phenomenon. Consequently, the procedure introduced goes along with an uncertainty assessment. Our theory was confirmed via measurements in cobalt-60 beams, where the studied effect does not contribute to the output factor. Measurements were also performed on our Saturne 41 linear accelerator and the results were qualitatively similar to those described elsewhere. The collimation systems were studied separately by varying one jaw setting while keeping the other at its maximum value. In the light of these results, we deduced an algorithm that can correlate the former data with the effect of backscattering to the beam monitor chambers for any rectangular field within 0.5%, which is of the order of the experimental uncertainty (0.6%). As we show, the experimental procedure is safe, simple, not invasive for the linac and requires only basic dosimetry equipment.

The formalism based on phantom and collimator scatter factors for high energy photon beams is deduced using a phase space description. The phantom scatter factors (Sp) depend on the field size and shape at the level of the phantom and are generally considered independent of the collimation details used to form the desired field provided the effect of contaminant electrons can be neglected. As demonstrated in this work, this behaviour leads to the applicability of the Clarkson method in irregular fields. However, for a given field formed with a tertiary collimator it is not a priori clear that the variations of extrafocal radiation due to secondary collimator setting do not affect the phantom scatter correction factors. In fact, the extrafocal radiation has a lower mean energy than that of unscattered photons, and this radiation can reach points well outside the radiation field increasing the irradiated phantom volume. Besides, transmission through the blocks contributes to phantom scatter. Therefore, for a given block-defined field, the associated phantom scatter dose, per unit of fluence in air on the central axis, should in principle increase when enlarging the secondary collimator field. To confirm this, isocentric Sp data for 6 MV photons were measured at 10 cm depth in water, reducing with cerrobend blocks several fields formed with the secondary collimators. In particular, when a 30 x 30 cm2 collimator field is reduced with blocks to a 7 x 7 cm2 field, the dose per unit of fluence in air is 1.4% higher than that of the square collimator field equating the given block field. Our calculations indicate that in this case the block transmission accounts for only 0.2% of this increment, showing that the remaining effect is due to extrafocal radiation. As a concluding remark, this work contributes to a better understanding of the classical Clarkson method for irregular fields giving, additionally, a formal interpretation of the commonly used quantities.

No abstract