Optimal image sample size for corneal nerve morphometry

Optom Vis Sci. 2012 May;89(5):812-7. doi: 10.1097/OPX.0b013e31824ee8c9.


Purpose: Arbitrary numbers of corneal confocal microscopy images have been used for analysis of corneal subbasal nerve parameters under the implicit assumption that these are a representative sample of the central corneal nerve plexus. The purpose of this study is to present a technique for quantifying the number of random central corneal images required to achieve an acceptable level of accuracy in the measurement of corneal nerve fiber length and branch density.

Methods: Every possible combination of 2 to 16 images (where 16 was deemed the true mean) of the central corneal subbasal nerve plexus, not overlapping by more than 20%, were assessed for nerve fiber length and branch density in 20 subjects with type 2 diabetes and varying degrees of functional nerve deficit. Mean ratios were calculated to allow comparisons between and within subjects.

Results: In assessing nerve branch density, eight randomly chosen images not overlapping by more than 20% produced an average that was within 30% of the true mean 95% of the time. A similar sampling strategy of five images was 13% within the true mean 80% of the time for corneal nerve fiber length.

Conclusions: The "sample combination analysis" presented here can be used to determine the sample size required for a desired level of accuracy of quantification of corneal subbasal nerve parameters. This technique may have applications in other biological sampling studies.

Publication types

  • Comparative Study
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Cornea / innervation*
  • Diabetes Mellitus, Type 2 / complications
  • Diabetes Mellitus, Type 2 / diagnosis
  • Diabetic Retinopathy / etiology
  • Diabetic Retinopathy / pathology*
  • Humans
  • Image Enhancement / standards*
  • Microscopy, Confocal / methods
  • Microscopy, Confocal / standards*
  • Nerve Endings / pathology
  • Nerve Fibers / pathology*
  • Ophthalmic Nerve / pathology*
  • Reproducibility of Results
  • Severity of Illness Index