Room: AAPM ePoster Library
Purpose: To reduce the maximum possible error (max-error) in a quality assurance (QA) program by using a novel machine-learning (ML) algorithm for predicting QA plan passing rate by minimizing max-error.
Methods: A total of 498 patient-IMRT QA-plans were measured using a commercial 2D diode array and delivered on five linear accelerators. Passing rates were based on a tolerance of 3%/3mm local dose and distance-to-agreement. Each treatment plan was characterized by 78 complexity features. Two ML models were used to predict passing rate: (1) an ordinary-least-squares (OLS) linear model that minimize the mean of the expected error; and (2) Chebyshev-minimax (MM) linear solution that minimizes max-error. An 80% training dataset was used for fitting and optimization, and a 20% testing dataset was used to compare predictions.
Results: Max-error of IMRT QA passing rate predictions was 7.6% for the OLS model and 3.0% for the MM model. The mean-square-error was, however, 1.4% and 2.8% for the OLS and MM models respectively highlighting the tradeoff between both methods. Following optimization and testing with hold-out sets (20% plans), the OLS and MM model max-errors were 7.0% and 3.8% respectively. To ensure all plans have at least a 90% passing rate, plans predicted to have 97.0% passing rate or lower (79 out of 497, 16.0%) would require QA using the OLS model due to its maximum 7.0% error. With the MM model, however, all plans predicted to be 93.8% or lower (24 out of 497, 4.8%) would require QA, resulting in more than a 3x reduction compared to the OLS model.
Conclusion: Efficient QA programs that ensure safe IMRT treatments require accurate identification of plans that are likely to fail QA-criteria. Using Chebyshev-minimax optimization, a 95% reduction of resources for IMRT QA can be achieved while meeting current passing criteria as suggested by TG 100.