Room: Exhibit Hall | Forum 7
Purpose: The goal of this work is to develop a fast-direct aperture optimization (DAO) algorithm for IMRT treatment planning and adaptation.
Methods: A previously proposed fluence map optimization algorithm called Fast Inverse Dose Optimization (FIDO) was extended for DAO. FIDOâ€™s objective function was reformulated to optimize the aperture shapes and weights of IMRT plans. A key feature of this reformulated objective function and its derivatives is that, for each iteration of the optimization, these equations can be computed without having to recalculate the 3D dose distribution. Instead, these equations use a pre-calculated and efficient mapping that describes the effect of aperture change on the objective function and its gradient vector. A proof-of-concept algorithm was developed in MATLAB. Aperture shapes and weights were optimized using the aperture-based FIDO equations described above, implemented in a quasi-newton (two-loop recursion L-BFGS) optimization algorithm. This fast-inverse direct aperture optimization (FIDAO) algorithm was used to optimize 5-10 IMRT beams on the AAPM TG-119 c-shaped phantom, and a clinical prostate, liver, and head-and-neck case. The execution time and resulting plan quality were compared with the results obtained with an interior-point DAO algorithm implemented in matRad (an open-source treatment planning toolkit in MATLAB).
Results: In all four cases, the prototype FIDAO algorithm provided comparable or superior plan quality. Optimization times (number of iterations) were also significantly reduced with this prototype algorithm: 1.1s (50) vs. 15.4s (30); 0.3s (9) vs 23.8s (42); 0.5s (13) vs 30.7s (29); and 54.7s (50) vs. 562.0s (482) in the TG-119, prostate, liver, and head-and-neck cases, respectively.
Conclusion: A new direct aperture optimization algorithm based on FIDO was developed. This algorithm demonstrated promising dosimetric results and speed enhancement in a few IMRT test cases.
Funding Support, Disclosures, and Conflict of Interest: This work was funded by the Canadian Institutes of Health Research as well as through a tri-partisan research agreement between the London Health Research Institute, Philips Healthcare, and the Government of Ontario.