With accurate calculation algorithms in inverse planning for beamlet-based intensity modulated radiotherapy (IMRT)

it takes time to calculate the dose matrix

which represents the dose distribution of each beamlet element to each voxel for unit fluence. To reduce the calculation time

coarse or approximate algorithms are often a choice

but this results in a final dose distribution that cannot reflect the real value. In addition

it is necessary to test if a coarse algorithm is capable of calculating the dose matrix of beamlets. In this work

simulated dynamics optimization algorithm was applied to optimize the segment weight to minish the dose error from the dose matrix calculation. After calculating the dose matrix by ray-tracing algorithm which takes into account just the primary component of absorbed dose

the original beam profile intensity distribution was optimized by using the simulated dynamics algorithm. Before segmentation

the even-spaced algorithm and genetic algorithm were applied in clustering. The dose distribution of every segment was calculated accurately by using convolution-superposition algorithm

and the weight of any voxel was in inverse proportion to the voxel number of the PTV or OAR it belonged. The segment weight was optimized by using the simulated dynamics algorithm. By comparing the dose distributions before and after optimization of segment weight

one finds that the dose distribution is improved obviously.

It was also found that some segments which contained fewer beamlets or lower weight value could be omitted because the decay of dose distribution could be improved by re-optimizing the segment weight

and that some segments could be omitted by re-optimizing the segment weight during the process of beamlet-based IMRT.