Session Item

Treatment planning in Varian’s Ethos system uses the Intelligent Optimization Engine, which is designed to require minimum input to its optimiser of clinical objectives only rather than multiple objectives placed along DVH curves. It also aims to reduce normal tissue dose even further than requested. This focus on only clinical objectives enables fast and clear traffic-light analysis of plan quality, which is essential when making the choice between a scheduled plan and a newly-created plan of the day during Ethos on-line adaptive radiotherapy. The latest version of Ethos (1.1) allows use of RapidPlan models - knowledge-based models created from a library of plans - which use line functions to act along the whole of a DVH curve. This work aimed to determine if RapidPlan models for rectal cases could be used effectively in Ethos.

A RapidPlan model in clinical use generated in Eclipse v15.1 for rectal cancer cases was based on data from 28 patients. In the model, targets are 52Gy (boost) and 45Gy (elective), OARs are bladder, bowel, femoral heads, and genitals, and either 2 or 3 VMAT arcs were used. 10 patients were planned in Ethos based on clinical objectives from UK National Rectal Cancer IMRT Guidance; each patient was planned with the RapidPlan model (IOE+RP) and without (IOE), and with 2 and 3 arcs. Clinical objectives and other plan parameters were compared.

All OAR clinical objectives achieved in the IOE plans were also achieved in IOE+RP plans; in one case, a clinical objective not achieved in the IOE plans was achieved in the IOE+RP plans. There was a small increase in 52Gy and 45Gy minimum target dose in the 2-arc IOE+RP plans, 0.69% and 2.66% respectively, and a small decrease in the 3-arc IOE+RP plans, 0.32% and 1.1% respectively. The differences in V95% and in global maximum for both targets for all plans was negligible.

The average mean body, bladder and bowel doses were all lower for IOE+RP plans than for IOE plans: body mean dose was 3.0% lower for 2-arcs and 2.6% lower for 3-arcs; bladder mean dose was 16.6% lower for 2-arcs and 18.0% lower for 3-arcs; bowel mean dose was 14.0% lower for 2-arcs and 14.1% lower for 3-arcs. 

Mean optimisation times for IOE+RP were slightly slower than for RP plans, rising from 479.8s to 490.2s for 2 arcs, and from 494.5s to 505.8s for 3 arcs. Dose calculation times for all plans were between 29.8s and 57.8s, with IOE+RP plans being calculated on average around 3s faster than IOE plans.

Mean MUs: for 2-arc IOE, 658.7, for 2-arc IOE+RP, 703.4 (increase of 6.8%); for 3-arc IOE, 626.0, for 2-arc IOE+RP, 683.2 (increase of 9.1%).

Use of our clinical Eclipse RapidPlan rectal model can be used effectively to reduce OAR dose in Ethos rectal plans with minimal effect on other plan parameters, without the needed for adding further optimisation objectives. Clarity of analysis in Ethos is therefore preserved, with minimal change in plan generation time, which are both important in an on-line adaptive process.