Encrypted login | home

Program Information

Biological Evaluation of Universal Multi-Criteria Optimization VMAT Prostate Plans


M Guerrero

M Guerrero*, S Chen , University of Maryland School of Medicine, Baltimore, MD

Presentations

WE-AB-209-4 (Wednesday, August 3, 2016) 7:30 AM - 9:30 AM Room: 209


Purpose: In recent years Multi-Criteria Optimization (MCO) has become commercially available as a potential optimization method for intensity modulated and volumetric arc therapies (IMRT and VMAT). The purpose of our work is to develop a universal set of objectives and constraints to generate prostate cancer MCO plans with VMAT and evaluate them using biological metrics.
Methods: We randomly selected ten prostate plans (two-arc VMAT) for patients with intermediate risk prostate cancer prescribed 75.6-79.2Gy to the prostate and seminal vesicles. We compared the plan scores to our clinical plans generated with standard IMRT optimization techniques. The planning target volumes (PTV) were 5-10 mm set-up margin expansion of the gross target volume. Using the same two arcs we generated MCO plans with a standard set of constraints and objectives using Raystation v4.5.The dose-volume histograms (DVH) for the (PTV), rectum and bladder were exported and used to calculate tumor control probability (TCP) and the complication-free tumor control probability P+ using biological models from the literature. We classified the plans by their level of difficulty based on the rectum-PTV overlap and develop MCO plans with similar levels of conformality as the clinical plans.
Results: The P+ values of the MCO plans had an average of 86.4%±2.3% and were equivalent to the clinical plans (P+ average of 85.0%±2.1%, p-value=0.33). The MCO NTCP values for the rectum ranged from 0.7% to 20% and were highly correlated with rectum-PTV overlap. NTCP values for the bladder were less than 1% for all MCO plans.
Conclusion: MCO VMAT prostate plans obtained with standardized objectives and constraints generate biologically equivalent plans to the clinical plans generated with standard optimization techniques.




Contact Email: