How to Read an OCT: Hard-Core Tips and Tricks

Last updated: October 2026

Most OCT reading errors happen before the reader looks at any number. This page is a set of hard-earned rules for reading OCT printouts the way experienced retina and glaucoma specialists do: quality first, raw B-scans always, and coloured maps last.

Step 0: Before you read anything

• Check the patient's name, eye, date and device. Many "progression" cases turn out to be a right eye compared against a left eye.
• Signal strength: below the manufacturer's recommended threshold (for example, under 6/10 on Cirrus), thickness values are unreliable. Weak signal usually thins RNFL artificially. Media opacity, dry eye and small pupils are common causes.
• Centration: the RNFL circle must be centred on the disc, and the macular cube on the fovea. A decentred circle shifts thickness between sectors.
• Segmentation lines: scroll through the B-scans and check that the coloured lines follow the true layer boundaries. When they don't, every number is wrong.
• Motion and blink: look for horizontal discontinuities (motion) and dark horizontal bands (blinks) on the en face image.

RNFL: avoid red disease and green disease

• "Red disease" is a false-positive abnormal result. Common causes are high myopia, tilted discs, large peripapillary atrophy, poor signal, and decentration. In myopes the superior and inferior bundles shift temporally, so the "normal" peak positions in the database don't match the patient.
• "Green disease" is a false-negative normal result. Disc oedema, an epiretinal membrane, vitreopapillary traction or a thick peripapillary vitreous can make a thin RNFL look thick.
• Normative databases usually do not include high myopes, children or every ethnic group. Treat their colour codes with caution in those patients.
• Look at the shape of the TSNIT curve (the classic "double hump") and compare the two eyes. Asymmetry between fellow eyes is often more telling than either eye alone.
• Floor effect: RNFL thickness stops falling at around 50 µm, because blood vessels and glial tissue remain. In advanced glaucoma, follow the macular GCIPL and the visual field instead of the RNFL.
• Test-retest noise: a change of only a few microns between visits is usually within measurement variability. Confirm a real change with repeat scans on the same device.

Macular ganglion cell analysis (GCIPL / GCC)

• Around half of the retinal ganglion cells sit in the macula, so the GCIPL often shows early glaucoma damage, especially in myopes whose RNFL is hard to interpret.
• Glaucomatous GCIPL loss tends to be arcuate and respects the horizontal raphe, often inferotemporally first.
• Neuro pearl: GCIPL loss that respects the vertical meridian (binasal loss in both eyes, i.e., the nasal half of each macula) points to a chiasmal lesion. It can appear before field loss. A homonymous pattern (nasal half of one macula plus temporal half of the other) suggests a retrochiasmal lesion, such as the optic tract.
• Any macular disease (ERM, DME, AMD, macular oedema) distorts GCIPL segmentation. Always check the macular B-scan before calling GCIPL loss glaucoma.

Macula: read the B-scan layer by layer

• Fluid location matters:
  • Intraretinal fluid (IRF): hyporeflective cysts within the retina (DME, CME, MNV activity)
  • Subretinal fluid (SRF): between the neurosensory retina and RPE (CSCR, nAMD, VKH, optic pit)
  • Sub-RPE: pigment epithelial detachment (PED)
• MNV clues: a shallow, irregular PED with a "double-layer sign" suggests type 1 (sub-RPE) MNV. Hyperreflective material above the RPE suggests type 2. Intraretinal hyperreflective lesions with fluid near a PED suggest type 3 (retinal angiomatous proliferation).
• Polypoidal choroidal vasculopathy: look for a sharply peaked or notched PED, often with a thumb-like protrusion, and a thick choroid. Confirm with ICGA.
• Drusen vs reticular pseudodrusen: drusen sit beneath the RPE; reticular pseudodrusen (subretinal drusenoid deposits) sit above it and carry a higher risk of progression.
• Outer retinal tubulation looks like fluid but is a tube of degenerating photoreceptors. It is not a sign of activity, and it does not need anti-VEGF.
• Prognosis markers: disruption of the ellipsoid zone (EZ) and external limiting membrane predicts poor visual recovery. In DME, disorganisation of the retinal inner layers (DRIL) also predicts worse vision.
• Hole vs pseudohole: a full-thickness macular hole goes through all layers down to the RPE. A lamellar hole has irregular inner tissue loss with intact outer retina. An ERM pseudohole has steep foveal walls without tissue loss.
• Hydroxychloroquine toxicity: look for parafoveal EZ loss (the "flying saucer" sign). In Asian patients, damage often starts in a pericentral pattern outside the standard central scan, so use wider scans.
• Pachychoroid: an enhanced-depth (EDI) scan showing a thick choroid with dilated outer vessels supports CSCR, PCV and pachychoroid neovasculopathy.

Artifacts you must recognise

• Mirror artifact: in highly myopic eyes or staphylomas, the image folds over at the top of the scan.
• Shadowing: vitreous floaters, haemorrhage or large vessels block the signal beneath them and mimic thinning.
• Segmentation failure: very common with ERM, VMT, high myopia, large PEDs and advanced disease.
• Out-of-range or cut-off scans: part of the retina falls outside the scan window and thickness is not measured.

OCT angiography (OCTA)

• Projection artifact: superficial vessels are "projected" into deeper slabs. Use projection-resolved software and check the B-scan with flow overlay before calling an outer-retinal MNV.
• Slab boundaries depend on segmentation. Wrong segmentation creates false flow voids or hides real lesions.
• Poor signal causes apparent loss of flow. Never interpret capillary dropout on a low-quality scan.
• OCTA shows flow, not leakage. It cannot tell you whether a lesion is active; combine it with structural OCT findings such as fluid.

Following progression

• Use the same device and scan protocol each time. Values from different devices are not interchangeable.
• Use device follow-up mode (registration to the baseline) whenever possible.
• Reset the baseline after cataract surgery, because better signal can change thickness readings.
• Combine structure with function. OCT tends to detect change earlier in early disease; the visual field becomes more useful in advanced disease.