(Tuesday, 7/16/2019) 1:45 PM - 3:45 PM
Room: Stars at Night Ballroom 1
Purpose: The aim of the study is to develop a k-nearest-neighbors (kNN)-based plan quality quantification method that takes into account shapes and volumes variations of multiple structures.
Methods: We propose to perform plan quality quantification based on the statistical inference of reference plans that are nearest to the evaluated plan. They are selected by a novel plan similarity metric applicable to multiple PTVs. In this study, two plan quality metrics—dosimetric result probability (DRP) and dose deviation index (DDI)—are proposed to quantify plan quality amongst reference plans. To evaluate the performance of the proposed method, we studied 927 clinical approved head-and-neck treatment plans with two planning targets and analyzed eight organs-at-risk (OARs). Twelve sub-optimal plans identified by DRP were re-planned to validate the capability of the proposed methods in identifying inferior plans. A stand-alone application has been developed to perform automatic database construction, plan referencing, and plan quality evaluation and visualization.
Results: The kNN referencing method reduced the prediction confidence interval by 26.5% and DDIs by 5.5% averaged for all the OARs compared with referencing all historical plans. Additionally, the average root-mean-square of similar plans DVHs to the target plan is significantly smaller than the value of all historical plans. After replanning, median doses to left parotid, right parotid, larynx, pharynx, and oral cavity are reduced by 31.7%, 18.2%, 9.1%, 16.4%, and 15.1% respectively. Maximum doses to brainstem and spinal cord are reduced by 5.2% and 4.7%.
Conclusion: The proposed plan referencing and analysis methodology has been shown to be predictive of the current plan quality and can effectively identify plans with suboptimal quality. Therefore, the proposed workflow can potentially be applied in the clinics as a real-time plan quality assurance tool. Additionally, the method can potentially be directly applied to conduct intra- and inter-institution plan quality analytics.
Funding Support, Disclosures, and Conflict of Interest: This work is supported in part by a grant from NIH/NCI under grant number R01CA201212 and a master research grant from Varian Medical Systems.
Keywords
Quality Assurance, Statistical Analysis, Treatment Planning
Taxonomy
TH- External beam- photons: treatment planning/virtual clinical studies
Contact Email
