Room: Exhibit Hall
Purpose: The Gradient Index (GI) measures dose fall-off around a target. Achievable dose fall-off depends on planning target volume (PTV). In this study, we expand on our earlier published plan quality data and with more data a new predicting GI function that better predicts achievable GI for small PTVs is found.
Methods: Cranial VMAT SRS treatment plans for 121 patients with 444 lesions (1-13, median 3) were analyzed. All plans were generated in Eclipse and treatments were delivered on a TrueBeam STx (Varian Medical Systems). Individual Y jaw positions were limited to deliver only with the 2.5 mm MLC resulting in up to 4 isocenters for some cases. The heterogeneity index (HI), conformity index (CI= V100%/PTV), and GI (GI= V50%/V100%) were determined for each lesion. By analyzing the GI for each PTV, a new GI prediction model for small PTVs (< 0.35 cc) was developed.
Results: The PTV ranged 0.05 â€“ 10.99 cc (median 0.80 cc), and 1-4 (median 2) isocenters were used for each case. The HI ranged from 1.18-1.47 (median 1.35), CI ranged from 1.02â€“ 3.46 (median 1.18), and GI from 2.43 â€“ 14.88 (median 4.4). With more data for small lesions a new model (3.5/PTV^0.3) is found to be a better guide for achievable GI for PTV < 0.35 cc.
Conclusion: A model to calculate predictive GI is a useful tool during the treatment planning process and provides the bast measure if the desired dose fall-off has been achieved around each PTV. Our previously published data is based on 40 plans and by analyzing 121 more plans we found that the GI analysis matches well with previous developed GI prediction model for PTV > 0.35 cc (GI>0.35cc=4/PTV^0.2). A new prediction model for small lesions, GI<0.35cc=3.5/PTV^0.3, is developed to provide a realistic plan quality guide for the planners.