Implementing and Evaluating Aperture Morphing for CBCT-Based Adaptive Planning
A Zhao*, A Godley, K Stephans, Cleveland Clinic, Cleveland, OHTH-C-137-3 Thursday 10:30AM - 12:30PM Room: 137
Purpose: To implement and test a novel application of an online adaptive re-planning technique, segment aperture morphing (SAM), to kV cone-beam CT (CBCT).
Methods: Five IMRT prostate patients with kV CBCT image guidance were chosen. The patients were re-planned with Panther planning system (Prowess), 5 mm margin, prescribed PTV D95 of 76 Gy, prostate D100 of 78 Gy. Daily contours were generated on nine CBCTs each using ABAS (Elekta). SAM, as implemented in Panther, then executes to morph apertures based on changes in the beams eye view target shape thereby correcting for both deformation and translation of the target through MLC adjustments. The 45 SAM plans were compared to the current standard of IGRT based shifts. The Panther re-plan, SAM and shifted plan calculations were done without heterogeneity corrections to avoid uncertainties in CBCT electron density. The DVH parameters used in treatment planning were compared between SAM and shifted.
Results: The SAM method provided significantly better target coverage, P<0.0002 for all target DVH parameters. SAM maintained target coverage with average prostate D100 of 78Gy and PTV D95 76Gy, while repositioning left the target under-dosed at 72Gy. SAM also decreased rectal dose, in particular reducing V60 from 15.4 to 14.0% (P=0.0185). On average bladder dose is lower for repositioning but this at the cost of considerably reduced target coverage, (shifted bladder dose correlates to target dose with an r2~0.2.) In all but one fraction, the prostate is best covered by SAM. The contour generation takes ~20-60 seconds, SAM is instantaneous and the dose calculation ~20 seconds.
Conclusion: Aperture morphing based on target shape change determined by auto-segmentation consistently and significantly improves target coverage and reduces rectal dose compared to rigid IGRT shifts. It is a viable method for CBCT-based adaptive planning.
Funding Support, Disclosures, and Conflict of Interest: A research version of ABAS was provided by Elekta AB, Stockholm, Sweden, and a research version of Panther by Prowess Inc, Concord, CA.