Program Information
Development of a Clinical Knowledge-Based Model for High Dose Rate (HDR) Prostate Brachytherapy Planning
A Plypoo*, J Roeske , M Surucu , A Solanki , M Harkenrider , H Kang , Loyola Univ Medical Center, Maywood, IL
Presentations
SU-H4-GePD-T-1 (Sunday, July 30, 2017) 4:30 PM - 5:00 PM Room: Therapy ePoster Lounge
Purpose: To develop a comprehensive knowledge-based planning model from patient datasets to be used as a planning baseline for HDR prostate brachytherapy.
Methods: Source spatio-temporal loading patterns were determined from 22 HDR prostate treatments of 12 patients utilizing 17-18 catheters implanted through a freehand flexineedle template. Dwell times from all the plans at each dwell position (0.5cm step interval) were compiled and averaged to the nearest 0.5 second to construct spatio-temporal patterns for each catheter and implemented in a clinical knowledge-based model. To analyze the model’s competency in achieving dosimetric goals, the model was imported into 15 previous prostate HDR cases and was evaluated without further adjustments (i.e. baseline evaluation). The planning evaluation goals of our clinic were then assessed against plans generated with geometrical optimization (GO) also without adjustments.
Results: Each individual catheter in the model contains distinctive temporal patterns relative to their spatial positions within the template and in proximity to critical structures. Mean target coverage (V100), V150, and V200 were 90.4%±6.8%, 29.3%±12.0%, and 6.4%±3.3% for the model plans vs 96.7%±3.9%, 63.1%±18.8%, and 20.4%±15.8% for GO plans (p<0.005 for paired t-tests). The V75 goal for bladder (p=0.0046) and rectum (p<0.0001) were met for 60% and 40% of the model plans, and 27% and 0% of GO plans respectively. The urethra Dmax goal (p=0.0019) was rarely met for both methods, 13% of the model plans, but in none of GO plans.
Conclusion: A knowledge-based clinical plan model containing source spatio-temporal patterns for HDR prostate was developed with a straightforward method which can be easily adapted into any clinical workflow. The model provides a robust planning foundation towards a clinically acceptable plan and standardization of the planning process. Further model development will include more automation processes and incorporate more anatomical and spatial variables.
Contact Email: