Poster — Wed Eve—47: IMRT Beam Angle Optimization Using a Hot‐Scripted Learning Algorithm for a Commercial Planning System
Bibliographic record
Abstract
Traditionally, beam angle selection for intensity‐modulated radiation therapy (IMRT) plans has been left up to the experienced dosimetrist. However, the large number of potential angle orientations suggests that the human plan may not be ideal. A learning algorithm (a variation of a genetic algorithm) was written to select beam angles to produce plans with more desirable dose‐volume histograms (DVH). The algorithm is used in parallel with the commercial Pinnacle3 Radiation Therapy Planning System. Starting with a generation of randomly chosen beam angles, the algorithm uses Pinnacle's built‐in hot scripting to call on the P3 IMRT portion of the software to perform dose calculations on each individual. Each set of angles is then ranked using a dosimetric fitness function that uses the same constraints that are used during the IMRT calculations. A new generation of beam angles is then constructed using both biologically and non‐biologically relevant operators. Operator weights are also adjusted each generation based on the dosimetric fitness function as well. Once completed, the algorithm was tested on several IMRT prostate cases. In each case, the DVH for the algorithm‐selected plan fit the dose constraints better than the human‐designed IMRT plan. A major drawback was the long running time of the algorithm (up to 11 hours), which was constrained almost entirely by the speed of Pinnacle's IMRT calculations. The ideal use of the code would be for overnight applications, with the human planner then using the optimized beam angles to plan as normal. This research was supported by the Graduate Internship program of the MITACS Network of Centres of Excellence and the Local Investigator Research Fund of the Windsor Regional Cancer Centre.
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.064 | 0.012 |
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 source (direct Gemma or distilled Codex), 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".