Background: Real-time adaptive radiotherapy workflows require fast spatial dose calculations with clinical accuracy. Modern physics-based dose calculation algorithms often compromise between speed and accuracy. In contrast, deep learning methods have shown to be effective at predicting spatial dose distributions with high accuracy in sub-second times. Very High Energy Electrons (VHEE) have shown potential as a treatment modality in recent years due to their penetrative ability and conformality. Creating a need for fast and accurate VHEE spatial dose calculations.
Purpose: This study presents a deep learning-based algorithm that utilizes convolutional layers and the self-attention mechanism to predict VHEE beam spatial dose distributions in sub-second times.
Methods: The presented Electron-Dose Transformer Algorithm (E-DoTA) maps the 3D patient geometry and beam characteristics (using a vector representing the beam energy and a 3D Gaussian with a specific Gaussian positional spread representing the beam shape) to
a 3D dose distribution. E-DoTA uses a series of 3D convolutional layers to extract features from the patient geometries and beam shape, followed by a transformer to route information between the extracted features and the added energy vector, which are then upsampled to
the 3D dose distribution. E-DoTA is trained on 60,000 combinations of patient geometries and beam characteristics, derived from 15,000 independent patient geometries based on 12 distinct patient CT scans from the abdomen region. The model’s accuracy and prediction
speed are assessed using 8,000 previously unseen patient geometries and beam characteristics.
Results: E-DoTA predicts dose distributions of VHEE beams with high accuracy, achieving a gamma pass rate of 97.05% ± 3% (3mm, 1%) and an average relative dose error of 0.254% ± 0.096% in approximately 131 ms.
Conclusions: The fast and high-accuracy dose predictions allow the speed-up of VHEE spatial dose distribution calculations, which is currently only provided by slow Monte Carlo algorithms. Further optimizations to E-DoTA could allow for more accurate and faster dose calculations, thereby potentially accelerating VHEE radiotherapy workflows.

In order to create radiotherapy treatment plans for cancer patients, dose calculations need to be done as quickly as possible to get accurate results. However, current dose calculation algorithms take too much time to be deployed effectively. The current in house algorithm of the Medical Physics and Technology Section at the TU Delft, attempts to solve this problem by utilising a deterministic algorithm that has a significant time advantage over Monte Carlo algorithms. However, this comes with the cost of inaccuracy, one of which is that it assumes
all dose is deposited locally along the beam path. This is inaccurate as secondary particles created from non-elastic nuclear interactions can deposit their dose far from the beam path due to retaining significant kinetic energy. This thesis attempts to reduce this inaccuracy by mapping and quantifying the secondary particles to assess their contribution in non-local dose deposition. And analysing the relevant particle’s energy and angle distributions to gain insight into the development of the particle's characteristics with depth. Thereafter the relevant
particle’s are then utilised as a source to emulate their production in a primary proton beam at different depths to obtain the relevant 3D dose distributions. The analysis concluded that secondary protons are the most relevant secondary particle as they contribute to 88% of the secondary dose and have a significant range to deposit their dose non locally. By utilising the secondary protons as a source, it was found that the relative error between the integrated depth dose (IDD) of the scored protons and the IDD obtained directly from Monte Carlo simulations is equal to 5.1% in the z-direction and 3.4% in the x and y-direction. The absolute difference was found to be 1.54 × 10−5 Gy which is equal to 0.096% of the total dose and 2.75% of the dose contributed by all secondary particles. The results show that the methodology can produce accurate 3D dose matrices for secondary protons at different depths, which can then be used to improve the accuracy of the in house algorithm by adding the precalculated 3D dose matrices to the algorithm