Author: Christopher Colyer, Leon F Dunn, Jonathan Dunning, Simon Goodall, Andy Schofield 👨🔬
Affiliation: GenesisCare 🌍
Purpose: Modern radiotherapy is characterized by complex treatment plans utilizing dynamic MLCs, gantry positions and recently, collimator rotations. The purpose of this work was to compare the plan complexity of 10,901 Monaco® (n = 17,664 beams) and 5,984 EclipseTM plans (n = 13,249 beams) treatment plans across 10 published metrics to derive control chart limits over a six-month period.
Methods: Approved treatment plans from Monaco and Eclipse were automatically processed using Sentinel (IsoAnalytics Pty. Ltd., Melbourne Australia) as they were exported for recalculation in parallel. Results for the 10 metrics were then compared across lung, breast, spine, prostate, brain, head and neck, abdomen, pelvis, rectum and limb cases. Control charts were derived for each treatment site per planning system.
Results: Across all plan types excluding the MI Total and Leaf Travel metrics. Monaco plans have a lower complexity on average when compared to Eclipse plans. However, both systems yield similar results overall, and a lower average complexity does not necessarily equate to a better plan. For MLC gap-based metrics; Q1 Gap, median MLC gap and SAS10, Monaco also produced larger segments on average. Control limits were derived, and are both treatment site and platform dependent.
Conclusion: Plan complexity metrics provide an additional parameter to determine plans that may be outliers with respect to plans within the same TPS and plan-class. Large datasets provide the statistical power to better highlight outlier plans with respect to the platform average. Plan complexity threshold values for deliverability need to be assessed against PSQA measurement outcomes to further solidify their potential value.