Encrypted login | home

Program Information

A Robust and Efficient Knowledge Based Plan Quality Analysis Tool for Lung SBRT

no image available
L Yuan

L Yuan1*, J Cai2 , Y Ge3 , F Yin4 , Q Jackie Wu5 , (1) Duke University Medical Center, Durham, NC, (2) Duke University Medical Center, Durham, NC, (3) Department of Software and Information Systems, University of North Carolin, Charlotte, NC, (4) Duke University Medical Center, Durham, NC, (5) Duke University Medical Center, Durham, NC


SU-H1-GePD-J(B)-4 (Sunday, July 30, 2017) 3:00 PM - 3:30 PM Room: Joint Imaging-Therapy ePoster Lounge - B

Purpose: To develop a plan quality check tool based on knowledge-guided planning and evaluate its clinical efficacy for Lung SBRT.

Methods: We utilized the previously reported knowledge models to develop a robust plan quality check tool capable of taking into account inter-patient plan quality variation caused by anatomical changes. 37 lung SBRT plans, 23 3D-CRT and 14 IMRT, are used in this study. Plan dosimetric quality are evaluated based on PTV dose conformity, ratio of 50% isodose volume to that of PTV (gradient measure), and D2% and mean doses in the OARs (Chestwall, lung, esophagus, heart and spinal cord). Knowledge models for the dose metrics are first trained with 3D-CRTplans. The deviations of the modeled dose metrics from the clinical values are calculated and the quality variation of the plans is represented by a composite plan quality index (CPQI) which is defined as the non-negative Principal Component of the standardized deviations. The 3D models were then applied to the IMRT cases to evaluate whether the CPQI can detect the dosimetric difference between the 3D and IMRT plans. Positive CPQI values mean better quality as compared to the model.

Results: The CPQI for the IMRT plans were calculated within seconds. The mean (s.d.) of the CPQI for the 3D and IMRT plans are 0.0(1.8) and 4.8(4.4), respectively. These results demonstrated good quality consistency for 3D-CRT plans and improvements in dosimetric quality for IMRT plans as expected. The individual dose metrics which have most significant average improvements are: PTV conformity and gradient measure: 0.2 and 2.6 respectively; heart D2% and mean dose: 3.9% and 1.0% respectively.

Conclusion: We developed a novel plan quality check tool based on knowledge-guided planning. It is a promising clinical tool for robust and efficient plan quality check, especially for plans demanding high standards such as lung SBRT.

Contact Email: