Program Information
Development of Physician-Driven, Clinically-Relevant Pass/fail Criteria for a DVH-Based IMRT Patient Specific QA Program
A Price*, M Lawrence , S Das , L Rankine , University North Carolina, Chapel Hill, NC
Presentations
SU-F-T-291 (Sunday, July 31, 2016) 3:00 PM - 6:00 PM Room: Exhibit Hall
Purpose: To derive and implement physician driven metrics for dose volume histogram (DVH)-based patient specific quality assurance (QA) for prostate radiation therapy based on clinically relevant deviations.
Methods: An observership study presented 3 radiation oncology physicians from our institution with a series of artificial DVH-based IMRT QA results, each generated by introducing iteratively larger perturbations from a planned DVH (from our pool of 20 real patient plans) to portions of the bladder, rectum, and planning target volume (PTV) DVHs. Physician responses were quantified by recording acceptable deviations from both in-house defined clinical goals (e.g. V70Gy < 10cc for rectum, mean dose for bladder) and other accepted dosimetric indices from the literature . The newly defined DVH criteria were used to evaluate 14 clinical IMRT plans measured using ArcCHECK and processed using 3DVH (Sun Nuclear Corporation, Melbourne, FL), all of which passed the standard 3%/3mm 2D-Γ comparison at >90% (absolute dose). In addition, 4 intentionally perturbed plans were measured, all with close to 90% 2D-Γ passing rates (relative dose). We compared physician derived constraints against 2D-Γ results to establish the relationship with an accepted IMRT QA metric.
Results: Of the 14 clinical patients measured, 7 were flagged for physician review (i.e. did not pass physician-derived DVH constraints). The rectum and bladder ΔD0.1cc constraint was the root cause of 5/7 of these QA failures. Of the 4 perturbed plans, 3/4 patients failed multiple DVH criteria in the bladder and rectum. Observership results from one physician-derived metric (PTV ΔD0.1cc) were found to be too strict and excluded from analysis.
Conclusion: We derived clinically relevant physician driven criteria for DVH-based patient specific QA at our institution. In N=14 patient plans, implementation of these criteria reduced the need for physician review of patient QA results by 50%. Further refinement of PTV D0.1cc constraint is needed.
Funding Support, Disclosures, and Conflict of Interest: Partially funded by Sun Nuclear Corporation research agreement.
Contact Email: