False positives on OCT deviation maps are the norm, not the exception

Design

retrospective cross-sectional study using spectral-domain OCT, eyes stratified into three axial length groups, logistic regression

Population

194 healthy eyes: axial length under 23 mm, 23 to under 25 mm, and 25 mm or more

Primary outcome

false-positive rates on ganglion cell analysis and retinal nerve fibre layer deviation maps, and the patterns they take

Effect

false positives 76.1% (GCA) and 62.2% (RNFL); GCA by axial length 45.2%, 51.2%, 87.5% (p<0.001); axial length OR 1.09-2.64, age OR 1.04-1.06

This cross-sectional study imaged 194 healthy eyes on spectral-domain OCT and counted how often the deviation maps flagged something. Colour-coded abnormal areas appeared in 76.1% of ganglion cell analyses and 62.2% of retinal nerve fibre layer maps — in eyes with no disease.

The rate tracked axial length closely. For ganglion cell deviation maps it was 45.2% in eyes under 23 mm, 51.2% between 23 and 25 mm, and 87.5% at 25 mm or longer (p < 0.001). Axial length was an independent predictor across all maps (odds ratios 1.09 to 2.64), as was older age (1.04 to 1.06). The patterns also differed by eye type, which is the part that is directly usable: hyperopic eyes most often showed a small doughnut pattern (67.9%), while highly myopic eyes showed diffuse thinning and an inferior crescent (both 38.1%).

What changes is the weight you give a red map. In a long eye, an abnormal ganglion cell deviation map is more likely than not to be a normal eye being compared with the wrong normative population — so it should not, on its own, start a patient on lifelong drops or trigger a monitoring pathway that will follow them for decades. Learning the two or three recognisable false-positive patterns lets you name what you are looking at instead of hedging. The limits are worth stating: 194 eyes, one centre, one OCT platform, and normative databases differ between manufacturers, so the exact rates will not transfer even though the direction certainly does.

• Treat an abnormal ganglion cell map in an eye of 25 mm or longer as more likely false than true
• Learn the small doughnut pattern in hyperopic eyes and the diffuse thinning and inferior crescent patterns in myopic ones
• Weight older age as a further reason for caution in interpretation
• Require corroboration from disc appearance, visual field or repeat imaging before initiating treatment
• Rates come from 194 eyes on one platform; normative databases differ by manufacturer, so audit your own

The statistics, in plain English

These are false-positive rates measured in eyes confirmed healthy, so they describe how often the instrument flags normality — and at 76.1% the flag carries very little information on its own. An odds ratio range of 1.09 to 2.64 across maps means axial length's effect differs by which map you read, so a single number would mislead. With 194 eyes split across three axial length bands, the subgroup rates are estimates with meaningful uncertainty even though the trend across bands is clear and highly significant.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.