Prevalence and Distribution of Segmentation Errors in Macular Ganglion Cell Analysis of Healthy Eyes Using Cirrus HD-OCT
Bibliographic record
Abstract
PURPOSE: To determine the frequency of different types of spectral domain optical coherence tomography (SD-OCT) scan artifacts and errors in ganglion cell algorithm (GCA) in healthy eyes. METHODS: Infrared image, color-coded map and each of the 128 horizontal b-scans acquired in the macular ganglion cell-inner plexiform layer scans using the Cirrus HD-OCT (Carl Zeiss Meditec, Dublin, CA) macular cube 512 × 128 protocol in 30 healthy normal eyes were evaluated. The frequency and pattern of each artifact was determined. Deviation of the segmentation line was classified into mild (less than 10 microns), moderate (10-50 microns) and severe (more than 50 microns). Each deviation, if present, was noted as upward or downward deviation. Each artifact was further described as per location on the scan and zones in the total scan area. RESULTS: A total of 1029 (26.8%) out of total 3840 scans had scan errors. The most common scan error was segmentation error (100%), followed by degraded images (6.70%), blink artifacts (0.09%) and out of register artifacts (3.3%). Misidentification of the inner retinal layers was most frequent (62%). Upward Deviation of the segmentation line (47.91%) and severe deviation (40.3%) were more often noted. Artifacts were mostly located in the central scan area (16.8%). The average number of scans with artifacts per eye was 34.3% and was not related to signal strength on Spearman correlation (p = 0.36). CONCLUSIONS: This study reveals that image artifacts and scan errors in SD-OCT GCA analysis are common and frequently involve segmentation errors. These errors may affect inner retinal thickness measurements in a clinically significant manner. Careful review of scans for artifacts is important when using this feature of SD-OCT device.
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 | 0.000 | 0.003 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| 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".