A MOrphing Technique That Projects 3D Surface Objects to a STandard Metric (MOST); Geometric Modeling of Cervix Cancer Patients for Adaptive Radiation Treatment
S Oh*, Y Cho, Princess Margaret Hospital, Toronto, ONTH-C-137-5 Thursday 10:30AM - 12:30PM Room: 137
Purpose: Adaptive radiation therapy (ART) had been proposed based in order to correct the dosimetric discount due to geometric variations by means of feedback control. Numerical modeling of inter- and intra-patients' geometric variations is one of the important parts in ART. To this end, a MOrphing technique that projects 3D surface objects to a STandard metric (MOST) was developed.
Methods: A total of 174 clinical target volumes (CTVs) were obtained from 32 patients and used to evaluate the MOST. The MOST consists of three main steps; 1) deforming 3D mesh of CTVs to a sphere by parametric active contour (PAC) model, 2) sampling the deformed mesh at evenly distributed girds to be the standard metric, and 3) projecting the 3D data into 2D plane for further analysis. The performance of the MOST was evaluated with respects to 1) iteration number, 2) computation time, and 3) residual deformation, the residual distance between a sphere surface and the deformed node.
Result: MOST successfully transformed complex 152 meshes from 28 patients to the standard metric. Convergence was achieved with average iteration of 65 and one standard deviation (STD) of 74 for 137 cases. The computation time was 17.8 hours and this corresponds to 51.3% of the time for all 152 cases. The average±STD of residual deformation was 0.9±0.7-mm and more than 98% nodes had less than 3-mm residual deformation.
Conclusions: MOST demonstrated its ability to transform complex 3D meshes of cervix cancer into a simple standard metric. Inter- and intra-patients' geometric variation could be directly compared and analyzed using the transformed results by virtue of the regularity of standard metric. Numerical modeling of tumor dynamics, implemented using this technique, may provide a systematic way for the estimation of tumor evolution and optimal correctional action for ART.
Funding Support, Disclosures, and Conflict of Interest: This study is supported by funding from RaySearch Laboratories and Ontario Consortium for Adaptive Interventions Radiation Oncology (OCAIRO).