ESTRO 2024 - Abstract Book

S3611

Physics - Dose prediction, optimisation and applications of photon and electron planning

ESTRO 2024

Conclusion:

The preliminary dosimetric data coming from the first 15 patients treated in the mono-institution phase 2 PRO Speed Trial, confirm the feasibility hypothesis coming from the former in-silico planning study of CK-SBRT treatment for localized PCa with simultaneous focal boost to the DIL up to 40 Gy. Moreover, the comparison of the median values of PTV dosimetric coverage and OARs dose constraints, proved that only small significant differences exist. These encouraging results need to be confirmed with longer follow-up data and with a larger cohort.

Keywords: Prostate CyberKnife, ultra-hypofractionation, SBRT

1907

Proffered Paper

Beyond the margin recipe: probabilistic and robust optimisation for cranial radiotherapy

Dan Welsh 1 , Joseph Wood 1 , Marcel van Herk 2 , Eliana Vasquez Osorio 2

1 The Christie, Christie Medical Physics and Engineering, Manchester, United Kingdom. 2 The University of Manchester, Radiotherapy Related Research Group, Manchester, United Kingdom

Purpose/Objective:

The use of PTV and PRV margins is a long-established and accepted method to account for geometric uncertainty in radiotherapy. However, the use of PTV margins can result in irradiation of large volumes of healthy tissue1. In the case of PRVs overlapping PTV, there is potential for underdosage of the CTV due to sparing a margin around a critical OAR. The degree and pattern of dose compromise is not likely to be optimal as the probability of a voxel being either tumour or OAR is not considered. Alternatives to margin-based treatment planning include robust optimisation and probabilistic planning2. Robust optimisation is now commercially available but there is limited clinical implementation3, and very limited clinical studies of probabilistic planning. This study compares the performance of robust optimisation and probabilistic planning to conventional margin plans for cranial radiotherapy, for cases where the target volume overlaps critical OARs (brainstem or optic apparatus).

Material/Methods:

13 patients were selected where the PTV overlapped one or more PRV for a critical OAR (brainstem, chiasm or optic nerves), each prescribed 60Gy/30#. Clinical margin plans utilised a 4mm PTV margin and a 3mm PRV margin. All plans were produced in RayStation 11B (RaySearch Laboratories, Stockholm, Sweden). For each case, two additional plans were created: 1) a probabilistic plan, optimised using an in-house implementation in the research interface of RayStation4, and 2) a robust plan, optimised using RayStation’s commercially available solution, using isotropic 4mm uncertainty. The set-up uncertainty parameters for the probabilistic and robust optimisation were determined from an IGRT audit of 50 brain patients (imaging protocol first 3# then weekly), with other geometric uncertainties estimated from the literature.