Abstract 7 — Deep Learning Differentiation of Inflammatory Lesions in Sacroiliac Joint MRI Based on Spondyloarthritis Research Consortium of Canada (SPARCC) System
Bibliographic record
Abstract
Objective To develop a deep learning algorithm for grading sacroiliitis based on SPARCC in magnetic resonance imaging (MRI). Method A total of 996 images with inflammatory lesions from 210 participants with MRI sacroiliitis were used for training and validation. The testing cohort consisted of 18 participants with and 19 without MRI sacroiliitis. One hundred and fifty four images from the testing cohort had inflammatory lesions identified by a pre-trained algorithm from our previous study[1]. The ground truth was defined by manually outlined regions of interests (ROIs) consisting of bone marrow edema (BME) at the sacroiliac joint. The performance of the deep learning pipeline in predicting the SPARCC score was compared to manual interpretation by two experienced readers. Result The intra-observer reliability and the Pearson coefficient between the SPARCC scores from two experienced readers and the deep learning pipeline were 0.83 and 0.86, respectively. The sensitivities in identifying all inflammatory lesions, deep lesions, and intense lesions were 0.83, 0.79 and 0.81, respectively. The Dice coefficients of the sacrum and ilium segmentation were 0.82 and 0.80, respectively. The accuracies of identifying the SI joint and reference vessel were 0.90 and 0.88, respectively. Conclusion The performance of AI algorithms in SPARCC scoring was compatible with manual scoring by experienced readers. This proposed deep learning pipeline could be the first demonstration of a complete and SPARCC-informed deep-learning approach in scoring STIR images in SpA.
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 | 0.002 | 0.003 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| 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".