Results from application of scripted based, algorithmic approach to multi-target SRS planning, evaluation and characterization of volume dependent metrics

Abstract

Purpose : Transitioning away from fixed beam toward VMAT approach for multi-target SRS, we developed a standardized algorithmic approach for treatment planning, and a script-based evaluation application characterizing high, intermediate and low dose regions proximal to targets and throughout the brain. The evaluation script was used to compare metrics for clinically treated fixed- and VMAT-based plans to quantify benchmark norms. Methods and Materials: Plans were examined for 79 patients (37 Fixed/47 VMAT) treating 179 (120 fixed/59 VMAT) targets. Dual purpose structures used for optimization and evaluation include 5 mm thick shells around the PTV (HDRing) and around the HDRing (MDRing) to control/measure dose fall off around the targets, and Brain – (PTV + 5 mm) to quantify for low dose regions. Effective gradients (GrEff) were calculated using V100% [cc] and V50% [cc] in HDRing and MDRings. Volume dependence of metric value distributions were characterized with quantile regression. Results: Conformity index (CI) decreased rapidly toward unity with increasing volume, plateauing near 0. 5 cc. Conformity index was significantly improved for VMAT plans (1. 19 ± 0. 17 vs 1. 40 ± 0. 46, p<0. 001) whereas effective gradients (%/cm) were reduced (117. 55 ± 17. 26 vs 137. 62 ± 26. 50, p<0. 001). Gradients decreased with increasing target volume (TV) converging near 4 cc for fixed field plans. Quantiles for volumes outside the PTVs receiving 12 Gy or more were smaller for VMAT than fixed beams, increasing as smaller powers of volume (e. g. 0. 45 vs 0. 51). Doses 5-10 mm from targets were similar. Volume of Brain – (PTV+05) receiving at least 5 Gy depended on cumulative PTV volumes and were less for fixed vs VMAT beams. Automation of metric collection improved evaluation of newly generated treatment plans and expedited the transition to multi-target VMAT-based SRS. Conclusions: Development of standardized algorithmic approach to optimization plus script based metrics calculation improved the SRS planning process and evaluation.