Purpose: To develop a versatile and flexible simulation method using an optical design software for designing imaging-based tomography systems and to validate the complete workflow for reconstructing test dose distributions, within the context of 3D scintillation dosimetry.
Methods: The simulation method was developed using Zemax OpticStudioÂ®, a software for designing virtual prototypes of optical systems. Exploiting OpticStudioâ€™s ray tracing features, three systematic key steps were implemented with the end goal of imaging-based 3D emission-computed tomography: (1) modeling different types of imaging systems and viewing positions; (2) computing their respective projection matrices, and (3) simulating photorealistic images of a fluorescent light pattern emitted within a plastic scintillator volume. To validate, 3D benchmark and clinical spinal SBRT dose distributions were reconstructed using simulated images as input projections within a 60x60x60-mmÂ³ volume discretized at a resolution of 1.5-mm. Without loss of generality, reconstructions were obtained using an inverse tomographic maximum likelihood-expectation maximization algorithm. The reconstruction quality was assessed using 2D and 3D correlation coefficients.
Results: The method enabled the simulation and tomographic modeling of three distinct imaging systems: a standard conventional camera and two focused plenoptic cameras, one with a single focal-length orthogonal microlens array and one with a triple focal-length hexagonal microlens array. The tomographic performance could be compared based on system type and number of projections. In particular, for all systems, an average 32% increase in 3D correlation was observed when using 2 vs 1 orthogonal projections, whereas an average 3% increase resulted from adding a third orthogonal projection. Moreover, the plenoptic cameras were not found to significantly outperform the conventional camera.
Conclusion: By taking advantageÂ of optical design tools, we've shown a streamlined and versatile workflow to conceive and compare scintillator-based dose detectors. Our work paves the way for theÂ rigorous optimization of a new generation ofÂ 3D dosimetryÂ systems.
Funding Support, Disclosures, and Conflict of Interest: Madison Rilling was supported by a Natural Sciences and Engineering Research Council of Canada (NSERC) Alexander-Graham-Bell doctoral scholarship. This research was supported by the NSERC Industrial Research Chair in Optical Design.