Inter-Instituion Application of Knowledge-Based IMRT Treatment Planning

Abstract

Intensity Modulated Radiation Therapy (IMRT) has allowed a large degree of healthy tissue sparing while delivering therapeutic dose to tumors. However, the treatment planning process for IMRT is iterative and time consuming and the resultant plan quality is dependent on the skill and experience of the planner.Following the work of Chanyavanich, a knowledge-based approach to IMRT treatment planning was used to generate high quality IMRT plans for patients from another hospital, using previously treated Duke plans as a reference library. An image-similarity metric was used to identify the patient from our database with the most similar anatomy to each new patient. Parameters from the Duke plan were then modified and applied to the new patient, resulting in quality dose distributions.In conclusion, the treatment planning time was reduced to approximately ten minutes for all cases, and the resultant plans were frequently of higher quality than the original, manually produced plans. The quality of the Duke treatment plans was preserved as the plans were adapted to new patient anatomy.