Program Information
Automated Complexity Analysis of Volumetric Modulated Arc Therapy Plans
K C Younge*, M M Matuszak , L A Janes , C J R Anderson , J M Moran , D A Roberts , University of Michigan, Ann Arbor, MI
Presentations
TH-AB-BRB-3 (Thursday, July 16, 2015) 7:30 AM - 9:30 AM Room: Ballroom B
Purpose: Volumetric modulated arc therapy (VMAT) plans can become excessively complex with little additional dosimetric benefit due to the degenerative nature of inverse planning. Our goal is to improve the efficiency and quality of the VMAT planning process through the use of an automated complexity metric script.
Methods: A metric that quantifies the complexity of VMAT plans by evaluating individual segment modulation was developed at our institution. This metric finds the ratio of the MLC leaf side length over aperture area for each control point, and calculates an MU-weighted average of all controls points in the plan. A plug-in script was created to enable use of the metric within our commercial treatment planning system. The script calculates the complexity metric and then compares it to metrics from previous clinical VMAT plans, all of which have passed patient specific pre-treatment quality assurance. To set a threshold complexity value, we used the complexity metric script to analyze 517 clinical VMAT plans created in 2013-14 that passed QA (composite analysis, 95% of points passing 4%/1mm) and 57 that failed QA. This threshold was optimized based on the resulting true and false positive rates.
Results: With the clinically set threshold, the complexity metric has a true positive rate of 44% for identifying plans that failed QA with a corresponding false positive rate of 6%. The false positives correspond to highly modulated plans, which may benefit from being reoptimized even though they initially passed patient specific QA.
Conclusion: Our in-house complexity metric script has the ability to streamline the VMAT planning and treatment process by alerting the planner when overly complex plans are generated. Those plans can be reoptimized before pre-treatment QA is performed to attempt to reduce the plan complexity and therefore improve the plan delivery accuracy.
Funding Support, Disclosures, and Conflict of Interest: This project was supported in part by P01CA059827.
Contact Email: