TL;DR
This paper presents an efficient and accurate method for calculating planar dose distributions for Flattening Filter-Free (FFF) IMRT, crucial for quality assurance (QA). They developed a novel bivariate Gaussian source model that accurately captures beam characteristics for both FF and FFF beams, leading to a robust algorithm that significantly simplifies FFF IMRT QA.
What Problem Does This Paper Address?
Intensity-Modulated Radiation Therapy (IMRT) is a sophisticated cancer treatment that delivers precise radiation doses to tumors. However, ensuring the accuracy of these complex dose deliveries requires rigorous quality assurance (QA). Traditionally, this involves time-consuming measurements. While computer-based independent dose calculations can speed this up, they rely on accurate source models. Flattening Filter-Free (FFF) beams, which are increasingly used due to their higher dose rates and reduced out-of-field dose, present unique challenges for these models. The removal of the flattening filter, which used to help shape the beam, alters the beam's characteristics, particularly reducing head-scattered photons. This simplification, however, requires a re-evaluation and adaptation of existing source models and dose calculation algorithms to maintain accuracy for FFF IMRT QA. This paper aims to bridge this gap by developing and validating an efficient algorithm specifically for FFF IMRT QA.
Methods
The core of this work lies in developing an analytical source model and an independent planar dose calculation algorithm.
The source model is a key innovation. It represents the radiation source within the linear accelerator as a combination of a primary point source and N bivariate Gaussian functions. These Gaussian functions are crucial for modeling "extra-focal" radiation, which includes head scatter, monitor chamber backscatter, and collimator exchange effects. The model was optimized by minimizing the difference between calculated and measured in-air output factors (Sc). A significant aspect is the direct comparison of this model for both Flattening Filter (FF) and Flattening Filter-Free (FFF) beams.
For dose calculation, they employed a backprojection and integration method to determine the beam fluence from the source model. This fluence is then modified by an off-axis ratio (OAR), modeled by a fourth-degree polynomial, to account for the cone-shaped profiles of FFF beams. The dose deposition process itself is described by an analytical kernel, which is a sum of three 2D Gaussian functions. To account for the ionization chamber's volume averaging effect during commissioning, a convolution-based method was used.
Datasets
The study utilized experimental data for commissioning and validation:
- In-air output factors (Sc): Measured using an IC-10 ionization chamber in a mini-phantom for various square and rectangular field sizes. These were used to optimize the source model parameters.
- Cross-beam profiles: Measured using a scanning ionization chamber in a water tank at 10 cm depth. These were used to optimize the off-axis ratio, dose deposition kernel, and MLC leaf-end transmission parameters.
- 2D diode array (MapCHECK) measurements: Used for the final validation of the planar dose calculation algorithm. Five head-and-neck FFF IMRT plans were delivered and measured with this array.
No new benchmark datasets were introduced; rather, existing measurement techniques and equipment were used for validation.
Key Insights and Ideas
- FFF beams simplify source modeling: The removal of the flattening filter significantly reduces head-scattered photons, making the extra-focal component of the source model simpler to manage. This is a crucial insight for developing efficient QA tools.
- Bivariate Gaussians are effective for extra-focal radiation: The paper demonstrates that a sum of bivariate Gaussian functions can accurately model the complex extra-focal radiation components, including collimator exchange effects, which were not explicitly modeled this way before.
- Planar dose calculation is sufficient for FFF IMRT QA: The study shows that an efficient planar dose calculation algorithm, rather than a full 3D calculation, is adequate and highly valuable for FFF IMRT QA, balancing accuracy with speed.
- Convolution for VAE correction is critical: Accounting for the Volume Averaging Effect (VAE) of ionization chambers during commissioning, using a convolution-based approach, is essential for achieving accurate beam profiles and, consequently, accurate dose calculations.
Key Results
- Source Model Accuracy: Excellent agreement was achieved for in-air output factors (Sc) for both FF (<0.25%) and FFF (<0.10%) beams.
- Reduced Scattered Photons: The removal of the FF reduced head-scattered photons by 34.7% for 6 MV and 49.3% for 10 MV.
- Planar Dose Calculation Accuracy: Superior agreement was observed for FFF IMRT planar dose distributions. In gamma comparisons (2%/2 mm criterion), the average passing rates were 96.2 ± 1.9% for 6 MV FFF and 95.5 ± 2.6% for 10 MV FFF.
- Efficiency: The commissioning time was less than two minutes on a personal computer, and the dose calculation for a typical IMRT beam took less than eight seconds.
Limitations and Future Work
- Limited to Homogeneous Phantoms: The current algorithm is primarily designed for homogeneous water phantoms. While it can be extended to calculate dose at various depths, the benefit in homogeneous phantoms is limited.
- Out-of-Field Underestimation: A slight underestimation (~1.5%) was observed just outside the field edge for larger fields (30x30 cm²), though its impact is considered small due to typical dose thresholds used in QA.
- Dose Rate Dependence: While the MapCHECK diodes showed dose rate dependence, it was deemed negligible for the specific IMRT delivery conditions tested and was not explicitly accounted for in the comparison.
- Future Work: The authors suggest that the dose calculation model can be extended to calculate planar dose distributions at various depths. They also note that the algorithm can be easily extended for linacs with two sets of independent jaws. The paper emphasizes providing an efficient and accurate algorithm for FFF IMRT QA rather than proposing new acceptance criteria.
Summary updated