Program Information
An Analysis of Knowledge Based Planning for Stereotactic Body Radiation Therapy of the Spine
J Foy*, R Marsh , D Owen , M Matuszak , University of Michigan, Ann Arbor, MI
Presentations
SU-E-T-97 (Sunday, July 12, 2015) 3:00 PM - 6:00 PM Room: Exhibit Hall
Purpose: Creating high quality SBRT treatment plans for the spine is often tedious and time consuming. In addition, the quality of treatment plans can vary greatly between treatment facilities due to inconsistencies in planning methods. This study investigates the performance of knowledge-based planning (KBP) for spine SBRT.
Methods: Treatment plans were created for 28 spine SBRT patients. Each case was planned to meet strict dose objectives and guidelines. After physician and physicist approval, the plans were added to a custom model in a KBP system (RapidPlan, Varian Eclipse v13.5). The model was then trained to be able to predict estimated DVHs and provide starting objective functions for future patients based on both generated and manual objectives. To validate the model, ten additional spine SBRT cases were planned manually as well as using the model objectives. Plans were compared based on planning time and quality (ability to meet the plan objectives, including dose metrics and conformity).
Results: The average dose to the spinal cord and the cord PRV differed between the validation and control plans by <0.25% demonstrating iso-toxicity. Six out of 10 validation plans met all dose objectives without the need for modifications, and overall, target dose coverage was increased by about 4.8%. If the validation plans did not meet the dose requirements initially, only 1-2 iterations of modifying the planning parameters were required before an acceptable plan was achieved. While manually created plans usually required 30 minutes to 3 hours to create, KBP can be used to create similar quality plans in 15-20 minutes.
Conclusion: KBP for spinal tumors has shown to greatly decrease the amount of time required to achieve high quality treatment plans with minimal human intervention and could feasibly be used to standardize plan quality between institutions.
Funding Support, Disclosures, and Conflict of Interest: Supported by Varian Medical Systems
Contact Email: