ABSTRACT
Online tuning of particle accelerators is a complex optimisation problem that continues to require manual intervention by experienced human operators. Autonomous tuning is a rapidly expanding field of research, where learning-based methods like Bayesian optimisation (BO) hold great promise in improving plant performance and reducing tuning times. At the same time, reinforcement learning (RL) is a capable method of learning intelligent controllers, and recent work shows that RL can also be used to train domain-specialised optimisers in so-called reinforcement learning-trained optimisation (RLO). In parallel efforts, both algorithms have found successful adoption in particle accelerator tuning. Here we present a comparative case study, assessing the performance of both algorithms while providing a nuanced analysis of the merits and the practical challenges involved in deploying them to real-world facilities. Our results will help practitioners choose a suitable learning-based tuning algorithm for their tuning tasks, accelerating the adoption of autonomous tuning algorithms, ultimately improving the availability of particle accelerators and pushing their operational limits.
ABSTRACT
Very high energy electrons (VHEE) are a potential candidate for radiotherapy applications. This includes tumours in inhomogeneous regions such as lung and prostate cancers, due to the insensitivity of VHEE to inhomogeneities. This study explores how electrons in the VHEE range can be used to perform successful in vitro radiobiological studies. The ARES (accelerator research experiment at SINBAD) facility at DESY, Hamburg, Germany was used to deliver 154 MeV electrons to both prostate (PC3) and lung (A549) cancer cells in suspension. Dose was delivered to samples with repeatability and uniformity, quantified with Gafchromic film. Cell survival in response to VHEE was measured using the clonogenic assay to determine the biological effectiveness of VHEE in cancer cells for the first time using this method. Equivalent experiments were performed using 300 kVp X-rays, to enable VHEE irradiated cells to be compared with conventional photons. VHEE irradiated cancer cell survival was fitted to the linear quadratic (LQ) model (R2 = 0.96-0.97). The damage from VHEE and X-ray irradiated cells at doses between 1.41 and 6.33 Gy are comparable, suggesting similar relative biological effectiveness (RBE) between the two modalities. This suggests VHEE is as damaging as photon radiotherapy and therefore could be used to successfully damage cancer cells during radiotherapy. The RBE of VHEE was quantified as the relative doses required for 50% (D0.5) and 10% (D0.1) cell survival. Using these values, VHEE RBE was measured as 0.93 (D0.5) and 0.99 (D0.1) for A549 and 0.74 (D0.5) and 0.93 (D0.1) for PC3 cell lines respectively. For the first time, this study has shown that 154 MeV electrons can be used to effectively kill lung and prostate cancer cells, suggesting that VHEE would be a viable radiotherapy modality. Several studies have shown that VHEE has characteristics that would offer significant improvements over conventional photon radiotherapy for example, electrons are relatively easy to steer and can be used to deliver dose rapidly and with high efficiency. Studies have shown improved dose distribution with VHEE in treatment plans, in comparison to VMAT, indicating that VHEE can offer improved and safer treatment plans with reduced side effects. The biological response of cancer cells to VHEE has not been sufficiently studied as of yet, however this initial study provides some initial insights into cell damage. VHEE offers significant benefits over photon radiotherapy and therefore more studies are required to fully understand the biological effectiveness of VHEE.