SU-E-J-151: Evaluation of a Real Time Tumour Autocontouring Algorithm Using In-Vivo Lung MR Images with Various Contrast to Noise Ratios
Bibliographic record
Abstract
PURPOSE: To quantitatively evaluate a lung tumour autocontouring algorithm using in-vivo lung cancer patient MR images with varying contrast to noise ratios (CNR) simulating images acquired at various MR field strengths. METHODS: , TE = 1.1ms. TR = 2.2ms, 275ms per image) under free breathing for approximately 3 minutes (650 images). Gaussian random noise is added to the 3T images to approximately simulate the equivalent CNR in images acquired at 1.5T, 1.0T, 0.5T, 0.3T and 0.2T. The moving tumour in all 3T images is contoured by a physician for reference. The first 20 of these manual contours are used for the parameters optimization of auto-contouring algorithm. The automatic contours from the remaining images are quantitatively compared with the physician's contours using the centroid's displacement and the Dice's coefficient (DC). RESULTS: The oncologist's contours of the 3T images show a maximum S-I motion of 26mm. Compared to the oncologist's contours, automatic contours have an average centroid displacement of 1.37mm, and an average DC of 0.881. The autocontouring algorithm's performance with images in the range of 1.5T to 0.5T equivalent CNRs is similar to that of the 3T data. However, for the lowest CNR datasets (0.2, 0.3T) an increase in centroid displacement and decrease in DC is observed, with mean displacements of 1.56mm, 1.71mm and DCs of 0.870, 0.836 for the 0.3T and 0.2T dataset, respectivelyConclusions: With in-vivo MR images, the autocontouring algorithm generated lung tumour contours similar to ones drawn by a physician (DC 〉 0.83). In this patient, additional CNR from 〉0.5T MRIs does not provide statistically significant improvement in the accuracy of our autocontouring software. E.Yip is supported by the Canadian Institutes of Health Research as well as Alberta Innovates - Health Solutions.
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 | 0.002 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.001 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 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 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".