How to Read an OCT in the Exam: Pattern Recognition in 10 Scans

In an OCT station, say what is present before naming the disease. Confirm quality and orientation, inspect the vitreoretinal interface, locate the abnormality by retinal compartment, then give a context-dependent differential. A dark space is not automatically oedema; a broken line is not automatically photoreceptor loss; a coloured thickness map is not a diagnosis.

This guide is for postgraduate ophthalmology examination revision. OCT findings must be integrated with history, examination and other imaging in clinical care; this is not patient-specific diagnostic or treatment advice.

The six-pass method: the same order on every scan

The most useful habit is refusing to jump from a familiar silhouette to a label. The International Nomenclature for Optical Coherence Tomography (IN•OCT) consensus provides standard language for describing morphology, while the vitreomacular-traction classification gives specific terms for the interface disorders below.¹² Use terminology consistently, but let the scan earn the term.

That sequence prevents two expensive errors in an exam. First, it separates an observation from an inference. Second, it makes you disclose uncertainty when an image alone cannot decide the diagnosis. A strong viva answer may end with “the leading differential in this context is __; I would correlate with __,” rather than treating structural OCT as a complete clinical examination.

Pass zero: quality and artefact are part of interpretation

Before “reading layers,” decide whether there is a reliable image. Look for poor fixation, low signal, blink dropout, media-opacity shadowing, decentration and motion. Check whether the B-scan corresponds to the en-face fundus registration. If a scan line is not through the fovea, do not call an absent foveal pit pathological. If a segmentation line runs through a PED or fluid pocket, do not report the automated thickness as though a human had measured it.

Shadowing gives useful information only when you name its source. Dense blood, exudate, a vessel or media opacity can attenuate the beam and obscure deeper structures. Conversely, relative hypertransmission beneath an RPE defect may be a clue, but it is not a standalone disease diagnosis. This is why a station answer should include “scan quality is adequate/limited” near the start. It shows that you understand an OCT is an acquisition, not a drawing of the retina.

The language of fluid: use the compartment precisely

Fluid terminology is one of the commonest marks lost in image stations. Your first task is spatial.

Say “intraretinal cystic spaces” rather than “CME” unless the context supports cystoid macular oedema as the appropriate clinical term. Say “subretinal hyperreflective material” before naming MNV if the stem has not supplied angiographic or clinical support. This is not pedantry. It keeps the description true even when the image has been cropped to remove the key clue.

Ten patterns: recognise them by what they distort

The following cards are a retrieval set, not a substitute for a retinal examination. For each one, practise a description, a mechanism direction, and one trap.

1. Vitreomacular adhesion (VMA)

VMA means the posterior hyaloid remains attached at the macula without the associated anatomical distortion required for traction. The international classification distinguishes it from VMT by the presence or absence of retinal architectural change.² Describe the attachment and preserved contour. Trap: calling every attached posterior hyaloid “traction.”

2. Vitreomacular traction (VMT)

VMT is attachment with distortion of the foveal surface, intraretinal cystic change, foveal elevation or other structural alteration attributable to traction. State focal or broad attachment only if measurable data are provided. Trap: seeing cysts and naming diabetic macular oedema when the interface is the dominant abnormality.

3. Epiretinal membrane (ERM)

An ERM appears as a hyperreflective line at the inner retinal surface, commonly with surface wrinkling, altered foveal contour or tractional thickening. Describe the membrane and the distortion separately. Trap: labelling any bright inner border as ERM when it may be normal internal limiting membrane reflectivity or scan artefact.

4. Full-thickness macular hole (FTMH)

An FTMH has a full-thickness foveal defect from inner surface to outer retina, often with elevated edges and an interface context. Avoid declaring stage from a single unfamiliar scan unless the exam has specified the classification system. Trap: calling a lamellar defect a full-thickness hole.

5. Lamellar macular hole and pseudohole

For a lamellar defect, describe partial-thickness tissue loss/irregularity and preserved outer retinal elements as visible. A pseudohole can occur with ERM-related steepening of foveal contour without a full-thickness defect. Trap: using “macular hole” without qualifying full versus partial thickness.

6. Intraretinal cystic spaces

Describe location, extent and retinal thickness. Differential direction includes diabetic macular oedema, retinal vein occlusion, inflammatory causes, traction and postoperative states, but no OCT image alone gives every cause. The AAO diabetic-retinopathy guidance places OCT within the broader assessment of retinal disease rather than as an isolated verdict.³ Trap: interpreting every cavity as active leakage.

7. Subretinal fluid

SRF separates neurosensory retina from RPE. In a stem, central serous chorioretinopathy, neovascular disease and inflammatory pathology can enter the differential; the correct choice depends on the RPE, choroid, clinical history and accompanying imaging. Trap: calling all SRF “CSCR.”

8. Pigment epithelial detachment (PED)

A PED is elevation of the RPE away from Bruch’s membrane. Describe whether the contents appear optically empty, hyperreflective or heterogeneous, then consider serous, drusenoid and fibrovascular directions in context. Do not diagnose vascular activity from elevation alone. The AMD Preferred Practice Pattern is useful for revising how OCT fits into AMD assessment.⁴

9. Drusen and outer-retinal/RPE change

Drusen can produce undulations or elevations of the RPE with material beneath it. Review the RPE contour, Bruch’s complex and overlying outer retinal bands. If the ellipsoid zone is irregular or disrupted, state that visible structural finding; do not casually translate it into a fixed visual prognosis. Trap: treating every RPE bump as a PED requiring a disease label.

10. RNFL and ganglion-cell maps

Peripapillary RNFL and macular ganglion-cell analyses need a different discipline: check signal, disc centring, segmentation and normative database flags, then correlate with disc examination and visual field. A red sector on a printout is a prompt for correlation, not proof of glaucoma. Trap: diagnosing glaucoma from a colour code without quality control or clinical correlation.

A compact pattern table for timed papers

A viva template that stays honest

Use the brackets rather than filling them with assumptions:

“This is a [right/left] [macular/peripapillary] OCT. Image quality is [adequate/limited] and I would verify [centration/segmentation] before interpreting the map. At the vitreoretinal interface there is [finding]. The foveal contour is [finding]. There is [no/IRF/SRF/sub-RPE] fluid, located [location]. The RPE and outer retinal bands show [finding]. In the supplied clinical context, this pattern is most consistent with [conditional differential]; I would correlate with [fundus, angiography/OCTA, visual field or clinical examination as relevant].”

The phrase “in the supplied context” is valuable. OCT cannot reliably distinguish all mimics, and OCTA, fluorescein angiography, fundus photography and clinical examination answer different questions. Do not say OCTA “shows leakage”: OCTA depicts flow signal, whereas FA characterises dye fluorescence behaviour.

When a comparison question gives you two scans

Many better questions do not ask for a diagnosis from one image. They show a baseline and follow-up, a horizontal and vertical raster, or OCT beside fundus/OCTA. Read both images through the same six passes before comparing them. State what changed and then decide whether the change is likely anatomical, acquisition-related or both.

For example, if follow-up OCT appears thinner, check whether the scan is similarly centred and whether segmentation remains correct before calling genuine resolution of fluid. If a cystic space has reduced but the ellipsoid-zone line remains disrupted, report the two observations separately. If a PED looks different between frames, compare its height, internal reflectivity and the RPE/Bruch’s boundaries rather than judging by colour alone. These are the small distinctions that stop a comparison station becoming a vague “better/worse” answer.

Use modality complementarity accurately. Structural OCT demonstrates cross-sectional reflectivity and morphology. Fundus photography records the visible surface appearance. FA describes dye transit and fluorescence behaviour. OCTA maps flow-signal information and has its own projection, motion and segmentation artefacts. In a written answer, “I would correlate structural OCT fluid with FA leakage where the question is activity and vascular permeability” is more precise than treating the tests as duplicates.

A layer-localisation drill

Once a week, cover the diagnostic label beneath a teaching scan and trace from vitreous to choroid with a finger or cursor. Name: posterior hyaloid; inner retinal surface; foveal contour; outer nuclear/photoreceptor region; RPE; Bruch’s complex; visible choroid. Then localise one abnormality using a relationship, not only a name: “above RPE,” “within retina,” “at inner surface,” or “below RPE.”

This drill is deliberately basic. Under examination pressure, candidates often remember an image pattern but lose the compartment. Compartment language transfers across diseases and devices, whereas a memorised colour scheme does not. If layer naming becomes uncertain on a low-quality image, say that the boundary is not reliably resolved. That is better examination practice than inventing precision.

Five mistakes examiners can see immediately

1. Starting with a disease label. Give morphology first.
2. Ignoring scan quality. An uncentred or segmented scan can create a false abnormality.
3. Using fluid terms interchangeably. IRF, SRF and sub-RPE fluid are different compartments.
4. Equating thickness with activity. Thickness is a measurement; disease activity needs context.
5. Turning a printout into a diagnosis. This is especially dangerous with RNFL maps and automated flags.

A practical revision drill

Take ten unlabelled scans from a trusted teaching set. Give yourself 30 seconds each. Write exactly four lines: quality; interface/contour; compartment/layers; conditional differential. Only then reveal the caption. On the next pass, pair each morphology with one competing explanation: VMT versus DMO, SRF in CSCR versus MNV context, PED versus drusenoid elevation. That contrast makes the pattern durable under pressure.

Use the verified retina study guide for the wider clinical framework. For report- and test-reading revision, see Read Ophthalmology Reports & Tests. You can also use high-yield free MCQs for short retrieval practice and the on-site ophthalmology glossary to standardise terminology. These are education resources, not clinical decision support.

Sources

All sources were checked on 18 August 2026. Re-check current society guidance before publication.

Footnotes

1. International Nomenclature for Optical Coherence Tomography (IN•OCT) consensus ↩

2. International vitreomacular traction study group classification ↩ ↩²

3. American Academy of Ophthalmology: Diabetic Retinopathy Preferred Practice Pattern ↩

4. American Academy of Ophthalmology: Age-Related Macular Degeneration Preferred Practice Pattern ↩