Room: AAPM ePoster Library
To compare ITV volumes between 4DCT simulation images and 4DCBCT pre-treatment images using a respiratory phantom with imbedded spherical targets and two different simulated respiratory rates
4DCT images of a respiratory phantom were acquired on a CT Simulator (Siemens Somatom Definition, Siemens, Malvern, Pa) utilizing the Anzai respiratory gating system and AZ-733V phantom (Anzai Medical, Tokyo, Japan). The phantom contains three imbedded spheres of different materials (acrylic, wood, and rubber) that simulate target volumes and moves a distance of 2 cm longitudinally when the motion is activated. Images of the phantom in both the 10 and 15 RPM modes were acquired on the CT simulator and imported into the treatment planning system (Raysearch Laboratories, Stockholm, Sweden). The phantom was then placed on the treatment couch and aligned to the reference CT dataset using a 3DCBCT (Elekta XVI System, Stockholm, Sweden). 4DCBCT datasets were then acquired of the respiratory phantom using both 10 and 15 RPM settings and imported into the treatment planning system. For all of the 4DCT and 4DCBCT datasets, the three spheres were contoured on each respiratory phase and combined to create ITVs. The 4DCT and 4DCBCT volumes were then compared using dice similarity coefficients (DSC), mean distance to agreement (MDA), and Hausdorff distance (HD).
The DSC ranged from 0.87 - 0.9 (10 RPM) and 0.61 - 0.83 (15 RPM). MDA values ranged from 0.19 - 0.25 cm (10 RPM) and 0.28 - 0.586 (15RPM). The HD ranged from 0.62 – 0.70 cm (10RPM) and 1.22 - 2.1cm (15RPM). Only one ITV comparison (15 RPM rubber sphere) fell below the 0.7 DSC threshold.
With one exception, all ITV comparisons exceeded the DSC threshold. The agreement between 4DCT and 4DCBCT images was lower for the 15 RPM images compared to the 10 RPM images.