Abstract
To demonstrate the merits of mass-based optimization in comparison with volume-based optimization using a simple test phantom.
Dose-volume-histogram-based (DVH-based) quadratic objective functions are converted into dose-mass-histogram-based (DMH-based) objective functions by multiplying per-voxel volumes by per-voxel densities within the objective. A digital phantom with a 1.0 g/cm
target is constructed for irradiation with two beams: one beam path contains a 0.2 g/cm
volume-of-interest (VOI0.2) while the orthogonal beam-path contains an equal-volume 0.8 g/cm
VOI (VOI0.8). Monitor-units are computed to achieve a 100 cGy average target dose for each individual beam, and for two-beam DVH-based and DMH-based optimizations.
For single-beam irradiation through VOI0.2, the average dose to VOI0.2 is 20.5 cGy. For single-beam irradiation through VOI0.8, the average dose to VOI0.8 is 25.2 cGy. Traversing the low density volume results in ∼23% lower dose. When DVH- and DMH-based optimizations are performed such that target dose-volume-histograms of the optimizations match, for the DVH optimization 60% vs. 40% of the dose is delivered through VOI0.2 vs. VOI0.8. For DMH-optimization, the split between dose delivered through VOI0.2 vs. VOI0.8 is 70% vs. 30%.
When density is constant, there is no difference between DVH- and DMH-based optimizations. However, in heterogeneous media, DMH and DVH solutions differ when low and high density materials have the same dose objectives. Delivering target dose through lower density VOIs facilitates target dose deposition due to a decrease in attenuating material, and the decreased attenuation lowers dose to the low density VOI. From mathematical and physical points of view dose-mass optimization is more general than dose-volume optimization.
