Image Acquisition: Scan Protocols, Centration, Alignment and Instrument Operation
Choosing the protocol
Match the protocol to the question. A macular volume/raster (e.g. 512 × 128 or 200 × 200) answers most retinal questions; a line scan through a lesion gives the highest sampling; a peripapillary circle (typically 3.4 mm) measures RNFL for glaucoma; radial scans suit the optic nerve head and peripheral lesions; en-face/C-scan slabs help with surface pathology. Fast cubes trade sampling density for speed; high-density scans trade speed for detail.
Alignment and centration
Seat the patient with chin on the rest, forehead against the bar, and align the eye to the internal fixation target. Macular scans must be centred on the foveal pit; disc scans on the disc centre; circle scans on the disc. Decentration produces oblique cut-through, apparent thickening/thinning and segmentation drift. The focus control sharpens the image laterally and raises signal; the z-offset/polarisation control moves the image within the acquisition window - do not confuse the two.
Signal and averaging
Check signal strength (quality) index before interpreting; most platforms want roughly 7/10 or their stated threshold, and repeat below it. Use frame averaging (ART) to reduce noise when the patient can cooperate, but remember averaging cannot fix motion, blink or media opacity. Dilate when wide-field or peripheral imaging is required, or when the pupil is small; avoid dilation where angle closure risk is unaudited and never dilate without following local protocol.
Longitudinal scanning
For follow-up, use the same instrument, same protocol and same centration, and use registration/change-analysis software. Comparing a dense scan today with a fast cube last year manufactures apparent change.
Traps: low signal attributed to pathology when the cause is cataract, dry eye or poor focus; absence of tracking blamed on the machine rather than on fixation; comparing scans from different devices.
Rule of thumb: before you interpret anything, confirm three things - correct centration, signal above threshold, and an identical protocol to the previous visit.