Layered mass geometry: a novel technique to overlay seeds and applicators onto patient geometry in Geant4 brachytherapy simulations
Bibliographic record
Abstract
A problem faced by all Monte Carlo (MC) particle transport codes is how to handle overlapping geometries. The Geant4 MC toolkit allows the user to create parallel geometries within a single application. In Geant4 the standard mass-containing geometry is defined in a simulation volume called the World Volume. Separate parallel geometries can be defined in parallel worlds, that is, alternate three dimensional simulation volumes that share the same coordinate system with the World Volume for geometrical event biasing, scoring of radiation interactions, and/or the creation of hits in detailed readout structures. Until recently, only one of those worlds could contain mass so these parallel worlds provided no solution to simplify a complex geometric overlay issue in brachytherapy, namely the overlap of radiation sources and applicators with a CT based patient geometry. The standard method to handle seed and applicator overlay in MC requires removing CT voxels whose boundaries would intersect sources, placing the sources into the resulting void and then backfilling the remaining space of the void with a relevant material. The backfilling process may degrade the accuracy of patient representation, and the geometrical complexity of the technique precludes using fast and memory-efficient coding techniques that have been developed for regular voxel geometries. The patient must be represented by the less memory and CPU-efficient Geant4 voxel placement technique, G4PVPlacement, rather than the more efficient G4NestedParameterization (G4NestedParam). We introduce for the first time a Geant4 feature developed to solve this issue: Layered Mass Geometry (LMG) whereby both the standard (CT based patient geometry) and the parallel world (seeds and applicators) may now have mass. For any area where mass is present in the parallel world, the parallel mass is used. Elsewhere, the mass of the standard world is used. With LMG the user no longer needs to remove patient CT voxels that would include for example seeds. The patient representation can be a regular voxel grid, conducive to G4NestedParam, and the patient CT derived materials remain exact, avoiding the inaccuracy of the backfilling technique. Post-implant dosimetry for one patient with (125)I permanent seed implant was performed using Geant4 version 9.5.p01 using three different geometrical techniques. The first technique was the standard described above (G4PVPlacement). The second technique placed patient voxels as before, but placed seeds with LMG (G4PVPlacement+LMG). The third technique placed patient voxels through G4NestedParam and seeds through LMG (G4NestedParam+LMG). All the scenarios were calculated with 3 different image compression factors to manipulate the number of voxels. Additionally, the dosimetric impact of the backfilling technique was investigated for the case of calcifications in close proximity of sources. LMG eliminated the need for backfilling and simplified geometry description. Of the two LMG techniques, G4PVPlacement+LMG had no benefit to calculation time or memory use, actually increasing calculation time, but G4NestedParam+LMG reduced both calculation time and memory. The benefits of G4NestedParam+LMG over standard G4PVPlacement increased with increasing voxel numbers. For the case of calcifications in close proximity to sources, LMG not only increased efficiency but also yielded more accurate dose calculation than G4PVPlacement. G4NestedParam in combination with LMG present a new, efficient approach to simulate radiation sources that overlap patient geometry. Cases with brachytherapy applicators would constitute a direct extension of the method.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".