Poor Fixation from Nystagmus: Protocol Choice and Centration
A 54-year-old woman with congenital nystagmus and type 2 diabetes is scheduled for cataract surgery. Preoperative OCT is requested to exclude diabetic macular edema. Acuity is 20/60 and she cannot hold fixation for more than about two seconds; the iris video shows a slow pendular movement with an amplitude of roughly 3 mm at the pupil. The first macular cube yields signal strength 3/10 and the fovea is displaced to the edge of the scan. A 200×200 fast macula scan is attempted and is also degraded, with visible horizontal displacement between consecutive B-scans. She is cooperative and wants to avoid a return visit.
1. Which acquisition strategy do you choose, and why?
Read the worked answer
The problem is motion during acquisition, so the goal is to shorten the time the retina must be still while keeping the fovea inside the scan.
- Prioritize short per-B-scan acquisition time over dense raster coverage. A single high-speed line scan (6 mm) through the fovea, repeated with frame averaging and tracking, gives interpretable data about fluid even without a full cube.
- Use tracking and averaging features if available; they reject frames with motion and combine the good ones.
- Reduce the field if necessary. A smaller, centered scan is more likely to be captured in a still interval than a wide cube.
- Use the fellow eye or a landmark approach to locate the fovea before capture: identify the disc, then offset temporally, and manually adjust centration.
- Consider a portable/handheld device or imaging in a position where the head is stabilized, and allow brief rest periods between captures.
- Do not simply repeat the same failing protocol and accept a 3/10 scan; that converts a technical failure into a potentially misleading clinical result.
If a valid scan cannot be obtained, document the limitation rather than reporting thickness values.
2. How do you verify that the scan is centered on the fovea and that the result is trustworthy?
Read the worked answer
Centration and quality are separate checks and both must pass.
Centration
- Inspect the en face/IR reflectance image: the fovea appears as a small avascular depression temporal to the disc, with the temporal arcades arching around it.
- Confirm the foveal depression is visible within the B-scan set and not at the edge; if it is at the edge, re-center manually and recapture.
- Check that the scan is not so far off that a parafoveal location is being reported as 'central.'
Trustworthiness
- Signal strength of at least 6/10 (or the device's equivalent threshold).
- No visible discontinuity or horizontal shift between consecutive B-scans.
- Segmentation lines follow the ILM and RPE smoothly across the scan.
- The foveal contour on the B-scan matches the en face location.
If any of these fail, the thickness value should be labeled invalid and the reason documented. A scan that is centered but motion-degraded, or clean but off-center, is equally unusable for a treatment decision.
3. The physician wants a definitive answer today about edema. How do you handle this?
Read the worked answer
Be explicit about what the study can and cannot show.
- Report the objective findings: 'Signal 3/10, motion artifact, fovea at scan margin; automated thickness not interpretable.'
- If a line scan of acceptable quality was obtained, describe the qualitative findings (e.g., 'no cystoid spaces identified on the interpretable line scan') while noting that the negative predictive value of a single degraded scan is limited.
- Recommend a practical path: dilate and repeat with a faster protocol, image at a time when nystagmus is least (some patients have a null point or are quieter at certain gaze angles - ask the patient), or defer until after cataract surgery if the surgical plan does not hinge on the result.
- Do not let a low-quality scan be used to justify cancelling or delaying surgery, and do not let it be used to justify anti-VEGF injection either.
Document the conversation and the quality limitation in the record.