The field of radiotherapy is constantly evolving, with a growing demand for next-generation online and real-time adaptive workflows. However, the accurate simulation of particle transport in sub-second times has remained a challenge. Current analytical pencil beam algorithms (PBA) and stochastic Monte Carlo (MC) methods are unable to provide the level of precision required for these advanced workflows. In this study, the authors propose a novel data-driven millisecond speed dose calculation algorithm called DoTA. This algorithm accurately predicts the dose deposited by mono-energetic proton pencil beams for arbitrary energies and patient geometries. Protons have a forward-scattering nature, so the authors frame 3D particle transport as modeling a sequence of 2D geometries in the beam's eye view. DoTA combines convolutional neural networks (CNNs) that extract spatial features such as tissue and density contrasts with a transformer self-attention backbone. This backbone facilitates information routing between the sequence of geometry slices and a vector representing the beam's energy. The algorithm is trained using low noise MC simulations of proton beamlets from 80,000 different head and neck, lung, and prostate geometries. The results demonstrate that DoTA can predict beamlet doses in just 5 milliseconds with an impressive gamma pass rate of 99.37% (1%, 3 mm) compared to ground truth MC calculations. This represents a significant improvement over analytical PBA algorithms in terms of both precision and speed. In fact, DoTA offers MC accuracy that is 100 times faster than PBAs for pencil beams. Moreover, DoTA is capable of calculating full treatment plan doses in just 10 to 15 seconds depending on the number of beamlets used. It achieves an outstanding gamma pass rate of 99.70% (2%, 2 mm) across nine test patients; surpassing all previous analytical pencil beam and deep learning-based approaches. The authors highlight that DoTA represents a new state-of-the-art in data-driven dose calculation; its speed being comparable to even commercial GPU MC approaches making it suitable for adaptive treatments that require sub-second calculations. Additionally, its straightforward implementation could potentially offer similar benefits to other steps of the radiotherapy workflow or other treatment modalities such as helium or carbon treatments.
The field of radiotherapy is changing and getting better with new ways to treat patients. It can be hard to accurately predict how particles will move in the body very quickly. The current methods we use are not always precise enough for advanced treatments. The authors of this study have come up with a new way to calculate how much radiation should be given called DoTA. DoTA can accurately predict how much radiation should be given using protons, which are a type of particle. It combines different types of computer programs to help make accurate predictions. They trained the program using simulations of different situations and found that it was very accurate and fast. This new method is much faster than the current ones we use and can give accurate results in just a few seconds. It represents a big improvement in how we calculate radiation doses for patients."
Definitions
- Radiotherapy: A treatment that uses high-energy rays or particles to kill cancer cells.
- Adaptive: Able to change or adjust based on new information or circumstances.
- Simulation: Creating a model or representation of something, often using a computer, to understand or predict its behavior.
- Algorithm: A set of instructions or rules followed by a computer program to solve a problem or complete a task.
- Dose: The amount of medicine or radiation given for treatment.
Radiotherapy: Achieving MC Accuracy with DoTA
Radiotherapy is a field that is constantly evolving, and there is an increasing demand for next-generation online and real-time adaptive workflows. However, accurately simulating particle transport in sub-second times has remained a challenge. Analytical pencil beam algorithms (PBA) and stochastic Monte Carlo (MC) methods are unable to provide the level of precision required for these advanced workflows. In response to this need, researchers have proposed a novel data-driven millisecond speed dose calculation algorithm called DoTA.
What Is DoTA?
DoTA accurately predicts the dose deposited by mono-energetic proton pencil beams for arbitrary energies and patient geometries. Protons have a forward-scattering nature, so the authors frame 3D particle transport as modeling a sequence of 2D geometries in the beam's eye view. To do this, they combine convolutional neural networks (CNNs) that extract spatial features such as tissue and density contrasts with a transformer self-attention backbone. This backbone facilitates information routing between the sequence of geometry slices and a vector representing the beam's energy. The algorithm was trained using low noise MC simulations of proton beamlets from 80,000 different head and neck, lung, and prostate geometries.
Results
The results demonstrate that DoTA can predict beamlet doses in just 5 milliseconds with an impressive gamma pass rate of 99.37% (1%, 3 mm) compared to ground truth MC calculations; this represents a significant improvement over analytical PBA algorithms in terms of both precision and speed – it offers MC accuracy that is 100 times faster than PBAs for pencil beams! Moreover, DoTA can calculate full treatment plan doses in just 10 to 15 seconds depending on the number of beamlets used; it achieves an outstanding gamma pass rate of 99.70% (2%, 2 mm) across nine test patients; surpassing all previous analytical pencil beam and deep learning-based approaches!
Conclusion
The authors conclude that DoTA represents a new state-of-the art in data driven dose calculation; its speed being comparable to even commercial GPU MC approaches making it suitable for adaptive treatments that require sub second calculations! Additionally its straightforward implementation could potentially offer similar benefits to other steps of radiotherapy workflow or other treatment modalities such as helium or carbon treatments!
Next generation online and real-time adaptive radiotherapy workflows require precise particle transport simulations in sub-second times, which is unfeasible with current analytical pencil beam algorithms (PBA) or stochastic Monte Carlo (MC) methods. We present a data-driven millisecond speed dose calculation algorithm (DoTA) accurately predicting the dose deposited by mono-energetic proton pencil beams for arbitrary energies and patient geometries. Given the forward-scattering nature of protons, we frame 3D particle transport as modeling a sequence of 2D geometries in the beam's eye view. DoTA combines convolutional neural networks extracting spatial features (e.g., tissue and density contrasts) with a transformer self-attention backbone that routes information between the sequence of geometry slices and a vector representing the beam's energy, and is trained to predict low noise MC simulations of proton beamlets using 80,000 different head and neck, lung, and prostate geometries. Predicting beamlet doses in 5 ms with a very high gamma pass rate of 99.37% (1%, 3 mm) compared to the ground truth MC calculations, DoTA significantly improves upon analytical pencil beam algorithms both in precision and speed. Offering MC accuracy 100 times faster than PBAs for pencil beams, our model calculates full treatment plan doses in 10 to 15 s depending on the number of beamlets, achieving a 99.70% (2%, 2 mm) gamma pass rate across 9 test patients. Outperforming all previous analytical pencil beam and deep learning based approaches, DoTA represents a new state of the art in data-driven dose calculation and can directly compete with the speed of even commercial GPU MC approaches. Providing the sub-second speed required for adaptive treatments, straightforward implementations could offer similar benefits to other steps of the radiotherapy workflow or other modalities such as helium or carbon treatments.
