EP24.30: Pilot study of intra‐ and interoperator agreement of a proposed image quality scoring system for transvaginal ultrasound images
Bibliographic record
Abstract
This study aimed to test the intra- and interoperator reliability of a proposed scoring system to quantify TVUS image quality, which could be used in the development of artificial intelligence(AI) tools for diagnosis of endometriosis. A proposed three-point scoring system was developed from review of existing literature and expert assessment. A core of 1-3 was assigned, where 1 = image quality is poor, 2 = image quality is suboptimal and 3= image quality is optimal. If the image was deemed to be inaccurate, it was assigned a score of 0 and rejected. Six raters (two sonologists, two sonographers and two radiologists) assigned 150 images (50 cases [50 uterus and 100 ovaries]) a score of 0-3. This was repeated after a period of at least one week. Interobserver reliability was calculated with an Intraclass correlation coefficient (ICC) and intraobserver reliability was calculated with a weighted kappa (k). Poor levels of interobserver agreement were obtained between all six raters for all 150 images (ICC = .480), and for uterus only (ICC = .359). Moderate levels of agreement were achieved for images of the ovaries (ICC = .531). Agreement between the paired sonologists and sonographers was poor for all images (ICC = .336 and .425) as well as uterus (ICC = .253 and .299) and ovaries (ICC= .400 and .469) only. Moderate levels of agreement were achieved between the paired radiologists overall (ICC = .600) and for the uterus (ICC = .538) and ovaries (ICC = .621). Intraobserver agreement was weak to moderate among each of the raters overall (range k = 0.533-0.718, ICC = 0.636-0.819) and for the ovaries (range k = 0.467-0.751, ICC = 0.596-0.862). Weak to strong agreement was seen for the uterus (range k = 0.568-0.808, ICC = 0.546-0.983). All measures were statistically significant (p < .001). The proposed scoring system produced poor-moderate interobserver reliability and mostly weak to moderate levels of intraobserver reliability. More refinement of the scoring system may be needed to improve reliability.
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 | 0.056 | 0.075 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 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".