SU-E-T-205: Comparison of Dose-Volume and Dose-Mass 4D Inverse Optimization in NSCLC




To compare dose-volume (Dvh) and dose-mass (Dmh) inverse treatment planning results, performed on time-resolved (4D) CT scans for non-small cell lung cancers (NSCLC), thereby accounting for the temporal variation in lung volume and mass.


Ten NSCLC cases were retrospectively studied. The estimated range of tumor motion was in excess of 0.5 cm (0.7 cm to 1.8 cm) in at least one principal direction. For each case five phases, uniformly distributed over the breathing cycle, were chosen for planning. For each of the phases two IMRT plans, with Dvh and with Dmh quadratic objective cost function applied only for the organs at risk (OARs), were generated. Pure dose objective functions were used for the targets. The final dose in each case for each optimization scheme was achieved through a deformable dose accumulation over the five breathing phases. Plan quality of Dvh and Dmh optimization was evaluated by several dose indices (DIs), represented by the dose to certain volume of an OAR. Tallied DIs were cord1%, heart33%, lungs20%, lungs30%, and esophagus50%, as well as lung volumes encompassed by 20 Gy and 30 Gy isodose lines. Statistical significance in the differences among those indices was quantified with two-tailed paired t-tests with a significance level of 5%.


On average Dmh optimized plans outperformed the 4D Dvh optimized plans. The statistically significant dose reduction in cord DI1, heart DI33, esophagus DI50, lung DI20 and DI30 were 27.5%, 9.9%, 21.3%, 5.6%, and 14.7% respectively. Isovolumes of 20 Gy and 30 Gy differed by 9.6% and 6.7% respectively.


The presented findings suggest that on average 4D Dmh optimized plans Result in better OAR sparing than 4D Dvh optimized plans. This is true for the majority of the tallied dosimetric endpoints, indicating that the potential sparing and its magnitude are patient specific.