Abstract
Purpose :
To develop equations to convert automated central subfield thickness (CST) values among eyes with diabetic macular edema (DME) between Cirrus (Carl Zeiss Meditec) and Spectralis (Heidelberg Engineering) instruments.
Methods :
A random sample of 70% of the data from participants in a cross-sectional observational study (DRCR Retina Network Protocol O) who underwent two replicate scans on both Cirrus and Spectralis instruments were used to develop the conversion equations. The remaining 30% of the data and data from an independent study (CADME) were used to validate the equations. Concordance correlation coefficients and Bland-Altman 95% limits of agreement (LOA) were used to evaluate conversion equation performance.
Results :
A total of 262 CST scan pairs from 131 eyes of 75 participants in Protocol O were used to develop the equations. The validation samples included 112 scan pairs from 56 eyes of 32 participants in Protocol O and 150 scan pairs from 37 eyes of 37 participants in CADME. The proposed conversion equations with units of μm are (1) Spectralis = 40.78 + 0.95 × Cirrus and (2) Cirrus = 1.82 + 0.94 × Spectralis. The concordance correlation coefficients between observed and converted measurements ranged from 0.96 to 0.99. Converting CST between Spectralis and Cirrus, 95% of the converted measurements are expected to be within ±10% or ±32 µm of the observed measurements on the other instrument.
Conclusions :
The conversion equations developed in this study are suitable for the transformation of values from individual scans taken with Cirrus or Spectralis machines as well as averaged values from cohorts of eyes with DME. Variability of observed versus converted measurements is similar to that of test retest. Although previous work enabled the conversion of Spectralis and Cirrus measurements to time domain OCT equivalents, time domain OCT is now obsolete. Thus, the ability to convert CST values from one SDOCT instrument to another will benefit both clinical research and standard care efforts.
This is a 2021 ARVO Annual Meeting abstract.