Room: ePoster Forums
Purpose: To evaluate the efficiency and dosimetric impact of implementing a knowledge-based planning model for rectum site-specific plans.
Methods: A rectum specific model was trained using RapidPlanâ„¢, Varianâ€™s implementation of knowledge-based planning (KBP). 89 plans were used to train the model, consisting of both static IMRT and VMAT modalities. Each training plan was clinically accepted and included patient setups in both the supine and prone position. The model was used for 10 rectum patients, not included in the model, and a plan was generated through a single iteration of optimization without intervention. The model generated plan was compared against the treated plan for a variety of DVH parameters for the PTV, genitals, perineum, small bowel, bladder, and femoral heads.
Results: The model generated a clinically acceptable plan for all patients after a single iteration of optimization. Target coverage was improved with the average PTV D95% increasing by 1.25Gy with no change in PTV D2%. D0.03cc changed insignificantly for the genitals and small bowel but decreased by 2-3Gy for the femoral heads. Bladder dose increased by about 1Gy, for D50%, D35%, and D5%. For specific patients, dose sparing increased for all OARs by 1-7Gy with minimal change in target coverage. Overall, plan quality remained consistent with the clinical plans and VMAT planning time was reduced to 17 minutes with a single iteration, compared to 45-60 minutes for multiple iterations by hand.
Conclusion: The KBP implementation in Varianâ€™s RapidPlan is capable of rapidly generating clinically acceptable and, for some patients, superior plans without intervention when compared to manual planning. Planner workload is also reduced by about 50%, by achieving an acceptable plan within one iteration. Further refinement can also be achieved within a second iteration, creating the opportunity for planners to improve overall plan quality as well as increasing plan throughput.
Not Applicable / None Entered.