MENU

Click here to

×

Are you sure ?

Yes, do it No, cancel
  1. Home
  2. Meeting Program
  3. All
  4. Warm-Start Optimization in Auto-Planning

Warm-Start Optimization in Auto-Planning

X Huang1*, H Quan1 , B Zhao2 ,C Chen3, Y Chen3 , (1) Wuhan University, Wuhan(2) Peking University First Hospital, Beijing(3) Elekta, Beijing

Presentations

(Sunday, 7/14/2019)  

Room: ePoster Forums

Purpose: To develop a template based auto-planning method starting from an existing plan; and to evaluate the efficiency and plan quality by this method in comparison with auto-planning with updated optimization parameters but starting from beginning in each trial.

Methods: "Warm-start" optimization (WSO) is available In Monaco (Elekta, St. Louis, US) for online adaptive planning with the Unity MR-Linac (Elekta, Crawley, UK). Template based auto-planning has been developed with Monaco to facilitate unmanned planning trials with updated optimization parameters from a “cold-start�. To start from the beginning and modifying optimization parameters in small steps could ensure a steady convergence to clinically acceptable plans as evaluated by a set of pre-defined acceptance criteria, but the process could be inefficient. In this work, we developed an auto-planning system based on fluence map optimization (FMO) that retained the dose distribution and beam weights of the previous plan. For validation and evaluation of the new method, 10 prostate cases were planned and the results were compared with that by a “cold-start� auto-planning. The same scoring metric was used in plan quality evaluation.

Results: For both methods, an acceptable plan could be achieved with no more than 30 trials. However the plan quality with WSO was consistently better than “cold-start� auto-planning, especially in rectal dose sparing and in target dose uniformity. Most importantly, WSO auto-planning improved the efficiency tremendously. The average time for a planning trial was reduced from 2.1 minute to 0.5 minute.

Conclusion: WSO in auto-planning is advantageous than starting from beginning, as both methods using machine decision making to modify the optimization parameters. The WSO method will increase the productivities in using machine-learning for fine-tuning the techniques and eventually in clinical use of auto-planning.

Keywords

Not Applicable / None Entered.

Taxonomy

Not Applicable / None Entered.

Contact Email

Additional Links