A fast method for rescaling voxel S values for arbitrary voxel sizes in targeted radionuclide therapy from a single Monte Carlo calculation

• Published in: Medical physics (Lancaster) Vol. 40; no. 8; pp. 082502 - n/a
• Main Authors: Fernández, María, Hänscheid, Heribert, Mauxion, Thibault, Bardiès, Manuel, Kletting, Peter, Glatting, Gerhard, Lassmann, Michael
• Format: Journal Article
• Language: English
• Published: United States American Association of Physicists in Medicine 01-08-2013
• Subjects: Algorithms; Anatomy; Cancer; Computed tomography; Dose‐volume analysis; dosimetry; Electrodeposition; internal dosimetry; interpolation; Interpolation; curve fitting; mathematical analysis; Medical imaging; Monte Carlo Method; Monte Carlo methods; Monte‐Carlo simulation; phantoms; Photons; Polynomials; radiation therapy; radioisotopes; Radiotherapy, Computer-Assisted - methods; targeted radionuclide therapy (TRT); Time Factors; Voxel S values (VSVs); internal dosimetry; Monte-Carlo simulation; targeted radionuclide therapy (TRT); Voxel S values (VSVs)
• ISSN: 0094-2405; 2473-4209
• DOI: 10.1118/1.4812684

Purpose:
In targeted radionuclide therapy, patient-specific dosimetry based on voxel S values (VSVs) is preferable to dosimetry based on mathematical phantoms. Monte-Carlo (MC) simulations are necessary to deduce VSVs for those voxel sizes required by quantitative imaging. The aim of this study is, starting from a single set of high-resolution VSVs obtained by MC simulations for a small voxel size along one single axis perpendicular to the source voxel, to present a suitable method to accurately calculate VSVs for larger voxel sizes.
Methods:
Accurate sets of VSVs for target voxel to source voxel distances up to 10 cm were obtained for high-resolution voxel sizes (0.5 mm for electrons and 1.0 mm for photons) from MC simulations for Y-90, Lu-177, and I-131 using the radiation transport code MCNPX v.2.7a. To make these values suitable to any larger voxel size, different analytical methods (based on resamplings, interpolations, and fits) were tested and compared to values obtained by direct MC simulations. As a result, an optimal calculation procedure is proposed. This procedure consisted of: (1) MC simulation for obtaining of a starting set of VSVs along a single line of voxels for a small voxel size for each radionuclide and type of radiation; (2) interpolation within the values obtained in point (1) for obtaining the VSVs for voxels within a spherical volume; (3) resampling of the data obtained in (1) and (2) for obtaining VSVs for voxels sizes larger than the one used for the MC calculation for integer voxel ratios (voxel ratio = new voxel size/voxel size MC simulation); (4) interpolation on within the data obtained in (3) for integer voxel ratios. The results were also compared to results from other authors.
Results:
The results obtained with the method proposed in this work show deviations relative to the source voxel below 1% for I-131 and Lu-177 and below 1.5% for Y-90 as compared with values obtained by direct MC simulations for voxel sizes ranging between 1.0 and 10.0 cm. The results obtained in this work show differences between the scored deposited energy and the emitted energy lower than 2% for electron radiation. Higher differences, attributable to the short considered radius of 10 cm in comparison with their penetration, can be found for photons. The authors’ results agree well with previously published data obtained by other authors using different methods.
Conclusions:
A reliable and fast approach for obtaining accurate VSVs for voxel sizes larger than the voxel size used for the MC calculation of the starting set of high-resolution VSVs was developed and successfully tested for three different radionuclides of interest for targeted radiotherapy: one pure beta (Y-90) and 2 beta-gamma emitters (Lu-177 und I-131). Applying the method of this work allows any interested reader to repeat the calculations for arbitrary radionuclides of interest and/or smaller high-resolution voxel sizes, provided the means for running MC simulations are available.