Purpose: Traditionally, GK-SRS treatment planning is performed using Manual Forward Planning (MFP) which can be time consuming and highly dependent on the experience of the planner. The objective of this study is to assess the variability in planning metrics between treatment plans generated by users with different levels of experience, with and without the use of a novel Fast Inverse Planning (FIP) algorithm.
Methods: Five patients treated to brain metastasis resection cavities, with target volumes ranging from 26.5-57.7 cc were chosen for this study. Manual forward plans and inversely optimized plans were generated for each case using Fast Inverse Planning (FIP) on the GK planning system by three planners with differing levels of experience (<10 cases beginner, <50 cases intermediate, >200 cases expert). A prescription dose of 30 Gy in 5 fractions to PTV, generated with 2mm expansion around the resection cavity (GTV), was used. All planners were requested to achieve a PTV coverage of 1.0. For each treatment site, Gradient Index (GI), Paddick Conformity Index (PCI), and beam-on time (BOT) were compared among the planners for FIP and MFP plans.
Results: Maximum variation in PCI among planners and across all MFP/FIP plans were 0.07 and 0.02, respectively. Similarly, the maximum variation in GI among planners and across all MFP/FIP plans were 0.74 and 0.38, respectively. Average planning time to achieve comparable plans using MFP and FIP was 60 minutes and 20 minutes respectively.
Conclusion: Plan quality metrics, PCI and GI, showed a tighter distribution for plans generated using FIP compared to MFP and FIP plans were at least equivalent in plan quality to MFP plans generated by an expert and intermediate planner. The automated workflow of inverse planning significantly reduced planning time to generate plans with equivalent plan quality across planners with differing levels of GK planning experience.
Optimization, Radiosurgery, Gamma Knife