Corneal Topography: Why a Map Answers a Different Question From a K Reading
What keratometry measures, and where it stops
A keratometer reduces the front of the eye to two numbers: the curvature of the steepest and flattest meridians across the central few millimetres of the cornea. That is enough to choose a soft lens base curve and to notice gross astigmatism. It is not enough to describe a surface.
The instrument works by assuming. It assumes the central cornea is a regular toric ellipsoid, that curvature is smooth from the centre outward, and that the meridians are 90 degrees apart. Where those assumptions hold, the two readings are accurate. Where they do not, the keratometer still returns two confident numbers, which is the dangerous part: a single K reading cannot tell you it was the wrong model.
How a topographer builds a map
The commonest design projects a set of concentric rings, the Placido pattern, onto the tear film and photographs the reflection. Each ring is a known circle in object space; its image is distorted by the cornea in a way that encodes local curvature at hundreds or thousands of points. Software then reconstructs a surface that would produce that reflection.
A second family of instruments uses a rotating Scheimpflug camera or a slit beam to measure the front and back surfaces directly, plus corneal thickness. That matters because the back surface contributes to total corneal power and cannot be seen in a reflection at all.
Millimetres into dioptres: the keratometric index
Corneal curvature is measured in millimetres of radius and reported in dioptres, and the conversion uses an assumed refractive index rather than the real one. The standard keratometric index is 1.3375, giving the relation F = (n - 1) / r. A radius of 7.8 mm therefore reads as about 43.3 dioptres, and 7.5 mm reads as roughly 45.0 dioptres.
The index of 1.3375 is not the refractive index of corneal tissue, which is closer to 1.376. It is a calibration constant chosen so that the instrument's output matches total corneal power as measured elsewhere, folding the back surface into a single front-surface number. This is worth knowing because it explains why a topography map's power scale and a refraction are not directly comparable.
Curvature is not the same as elevation
A curvature map answers: how tightly is the surface bent here. An elevation map answers: how far is this surface from a chosen reference sphere. The two are related but not interchangeable, and the distinction becomes visible when a cornea has a localised change.
Elevation is the more sensitive of the two for irregularity, because the reference is a smooth ideal and any deviation shows up directly, without being smoothed by a derivative. A region of abnormal elevation will also produce abnormal curvature, but the curvature response tends to be broader and less localised, which is why elevation maps are used where irregularity is suspected.
Within curvature maps there is a second distinction. Axial or sagittal curvature averages over the surface and is the familiar display; tangential curvature is more local and shows tight rings more sharply. A tight pattern that appears on tangential but not axial is a warning sign rather than a diagnosis.
What a map shows that two numbers cannot
Irregular astigmatism. If the two principal meridians are not perpendicular, or if curvature changes sharply within the pupil, a single cylinder and axis cannot correct it. Spectacles will leave a residual blur that changes with gaze direction. The map is the only common way to see this.
Contact lens warpage. A rigid lens that has been worn for years can mould the corneal shape. Comparing a map taken before lens removal with one taken after a period out of lenses shows whether the pattern is the cornea or the lens. Without that comparison, a K reading can be attributed to the wrong cause.
Post-surgical and post-trauma change. A scar or a suture produces a localised change that a two-meridian model cannot represent. The map localises it.
Keratoconus patterns. Inferior steepening, asymmetric bow-tie forms and displaced apex are recognisable map patterns. The map flags; the clinician diagnoses. A pattern is a reason for assessment, not a result.
Three misconceptions
- "My K readings are normal, so my cornea is normal." Normal central K readings are compatible with irregularity outside the sampled zone, or with a back-surface change the keratometer never sees.
- "A topography map is just a prettier keratometry." It answers a different question. One gives the curvature of an assumed shape; the other measures deviation from an ideal one.
- "The red areas mean damage." Colour scales are relative to the map's own range and are not comparable between instruments. Read the numbers and the pattern, not the colour.
When it is worth asking for one
Topography earns its place when a refraction does not behave: vision that is correctable to 6/6 in one direction and not another, a cylinder that keeps changing between visits, poor spectacle tolerance with a normal-looking prescription, or a history that makes an irregular surface plausible. It is also routine before orthokeratology and for rigid lens fitting.
None of this replaces an examination, and none of it should be read as a diagnosis. If you are given a map you do not understand, the fair questions are which map type it is, whether elevation is shown, and what the clinician was looking for.
Why two maps of the same eye differ
A Placido-based map is a photograph of the tear film, so anything that disturbs the tear film disturbs the map. A short break-up time, a dry patch, a lens that has just been removed, or an incomplete blink can all create local irregularities that disappear on a clean surface. Fixation matters for the same reason: if the eye is not centred on the instrument axis, the reconstruction is being built from an off-axis view.
That is why a single map is a snapshot rather than a trend. Where a pattern matters, the useful step is a repeat measurement after a blink and, where relevant, after a period out of contact lenses. Repeatability is part of the measurement, not an optional extra.