Commissioning of GPU-based multi-criteria optimizer combined with plan navigation tools for high-dose-rate...

Area of Science:

• Medical Physics
• Radiation Oncology
• Computational Biology

Background:

• High-dose-rate (HDR) brachytherapy planning requires efficient and accurate optimization algorithms.
• Existing workflows can be time-consuming and may not fully explore the Pareto-optimal solution space.

Purpose of the Study:

• To report on the clinical commissioning of a GPU-based multi-criteria optimization (gMCO) algorithm and its graphical user interface (gMCO-GUI).
• To validate the gMCO algorithm's performance against established clinical planning systems for HDR prostate brachytherapy.

Main Methods:

• The gMCO workflow was validated against Oncentra Prostate (OcP) and Oncentra Brachy (OcB) using 40 HDR prostate brachytherapy patient plans.
• 2,000 Pareto-optimal plans were generated per patient, with one plan selected via gMCO-GUI for export.
• DICOM RTPLAN files were used for exporting dwell positions/times to OcP/OcB for dosimetry calculation. Validation included TG43 implementation, volume calculations, DVH comparisons, and 3D dose distribution analysis (DSC, 95% HD, γ analysis).

Main Results:

• gMCO dose rates showed minimal differences (-0.4% to 0.3%) compared to flexisource consensus data.
• gMCO volumes agreed within ±2% for 3/5 analytical shapes.
• Key DVH indices differed by ±1.2%, with mean DSC > 0.98 and mean 95% HD < 0.4 mm for isodose lines. Over 99% of voxels passed γ analysis (1%/1 mm threshold).

Conclusions:

• The GPU-based MCO workflow was successfully integrated into clinical practice.
• The gMCO algorithm demonstrated high accuracy and reliability when validated against established brachytherapy planning systems.