CONTOURING A NEW PATH: ONTARIO’S COLLABORATIVE APPROACH TO AI IN RADIOTHERAPY
Bibliographic record
Abstract
The Radiation Therapy Community of Practice (RThCoP) embarked on a collaborative initiative to explore the implementation of artificial intelligence (AI) auto-contouring tools in radiation therapy planning. This initiative aims to understand the benefits and opportunities of auto-contouring solutions, their impact on program efficiency, and maintaining quality and safety while addressing the diverse needs of radiation therapy programs across multiple centres. The current state evaluation included representation from all participating regional cancer centres (RCCs) (n=7) using AI contouring tools, ensuring inclusivity and a broad range of perspectives. A series of structured meetings were conducted to gather insights and share experiences regarding AI contouring adoption. Key areas explored included integration into clinical workflows, addressing the learning curve associated with new technology, and measuring efficiency improvements. Feedback from these discussions will be used to develop actionable guidance, supplemented by evidence-based recommendations and consensus-driven best practices. Currently, three of seven RCCs utilize proprietary auto-contouring solutions embedded in their primary treatment planning system (TPS). Four of seven RCCs use third-party auto-contouring solutions, two of which plan to migrate to their primary TPS’s tools in the near future. Regardless of the tools in use, the initiative highlighted several benefits of adopting AI contouring, including reduced planning time, improved consistency in contouring, and enhanced resource allocation. However, opportunities for improvement include addressing variability in vendor solutions, training approaches to enhance confidence in AI-assisted workflows, and mitigating the perceived impact on the scope of practice for radiation therapists. Collaboration among centres allowed for the sharing of strategies to address these challenges, fostering a sense of community and shared learning. The current state evaluation has provided centres with an initial understanding of the benefits and opportunities for improvement when integrating AI contouring into clinical practice. Early AI adopters reported measurable improvements in workflow efficiency and reductions in inter-clinician variability. The initiative also underscored the importance of fostering a culture of continuous learning and adaptability in adopting emerging technologies. The work of the RThCoP aims to establish a foundation for scaling AI contouring practices across the broader radiation therapy community, with the potential to improve patient outcomes and optimize resource utilization on a larger scale. Most importantly, the initiative underscores the need for the RThCoP in fostering collaboration across Ontario, creating a platform for centres to share resources, align on best practices, and collectively address challenges in adopting AI contouring and beyond.
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 | 0.044 | 0.046 |
| Meta-epidemiology (narrow) | 0.000 | 0.001 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.003 | 0.006 |
| Science and technology studies | 0.025 | 0.013 |
| Scholarly communication | 0.013 | 0.006 |
| Open science | 0.006 | 0.018 |
| Research integrity | 0.003 | 0.004 |
| Insufficient payload (model declined to judge) | 0.007 | 0.001 |
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".