Room: Exhibit Hall | Forum 7
Purpose: A knowledge-based planning (KBP) system was previously created to improve treatment plan creation for whole breast irradiation using two tangential beams. Conversion of DVH estimations to deliverable plans requires adjustment of the optimization objectives. This work iterated beyond the previous work toward a better set of optimization parameters to improve the efficacy of achieving the treatment plan objectives.
Methods: New knowledge-based DVH estimation models were created in RapidPlanTM (Varian Medical Systems) using 60 and 49 previously treated plans for right and left breast patients, respectively. The models were used to create plans for 20 patients and compared with the clinically treated plans in terms of dose maximum (DM), mean tumor bed dose (TB), right lung mean dose (RL), left lung mean dose (LL) and heart mean dose (H).
Results: On average, for left breast plans DM decreased by 1.62%, TB decreased by 1.50%, LL decreased by 0.06% and H increased by 0.15% when using the KBP routine. On average, for right breast plans DM decreased by 0.43%, TB decreased by 0.85% and RL decreased by 0.86%. The isodose coverage and the DVHs for the treatment plans created using the KBP achieved clinical quality. On average, the treatment plans produced using the KBP routine were created in 5 minutes, which includes the time for auto-contouring of the breast tissue and skin flash structures necessary for optimization.
Conclusion: Improvements of the optimization parameters compared to the previous KBP routine demonstrates that this iterative approach yielded dosimetric gains. The KBP workflow has been optimized to help improve plan creation techniques for whole breast irradiation using tangential-beams. Currently, the KBP models are able to produce treatment plans of clinical quality.
Not Applicable / None Entered.