Seventy‐First Annual Scientific Meeting of Canadian Organization of Medical Physicists, RBC Place London, London, Ontario, June 5–7, 2025
Bibliographic record
Abstract
Purpose: Head and neck cancer (HNC) patients undergoing radiotherapy often experience significant anatomical changes that necessitate replanning.However, replanning is resource-intensive and decided on short notice.Therefore, we aim to predict the progression of anatomical changes throughout radiotherapy so that replanning can be anticipated in advance.Methods: We trained a variational autoencoder (VAE) to learn condensed latent vectors of CBCT scans using our in-house dataset of 420 HNC patients and 5323 CBCT scans.Subsequently, we developed a model to predict changes in these latent-space vectors over time based on the initial CBCT and clinical features (eg.staging, chemotherapy, etc.).Points along the predicted latent trajectory were decoded by the VAE to reconstruct synthetic future CBCT images.We evaluated the model by calculating the Dice score between actual and predicted body masks for each fraction.Additionally, we assessed the percent change of the area between the first and subsequent fractions in true versus predicted images, a useful replanning metric for determining shrinkage.Results: Initially, we trained our models on binary masks of single image slices, rather than the full CBCT.Our model achieved an average test set Dice score of 0.94 and area error of 5.9% for all fractions.The area error decreases as more CBCTs are incorporated throughout treatment.We are expanding our model to predict 3D anatomical changes and implementing a replanning flagging system based on the expected changes. Conclusion:In a DIBH treatment, 3D CRT techniques is dosimetrically advantageous and more forgiving to tolerance over VMAT.
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 | 0.003 | 0.004 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.002 | 0.001 |
| Scholarly communication | 0.004 | 0.001 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.371 | 0.147 |
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".