Abstract
Purpose: :
Standard automated perimetry (SAP) and retinal nerve fiber layer (RNFL) thickness measured by optical coherence tomography (OCT) should provide equivalent assessments of glaucomatous neuropathy, but a method for the quantitative correlation of the tests is needed. The present study used a structure-function model to translate SAP data into RNFL thickness for mapping SAP data onto the TSNIT curve of RNFL thickness.
Methods: :
The model is based on nonlinear translations of SAP sensitivity to neuronal values and linear transformations of RNFL thickness to axonal counts. The SAP and OCT data were divided into 8 corresponding sectors and, for each sector, the model-predictions of RNFL thicknesses from SAP measures were compared to mean OCT-measured RNFL thicknesses. The data for the study were clinical SAP and OCT data from 60 eyes (32 patients) from the University of Houston (UHdata) and 53 eyes (53 patients) from the Bascom Palmer Eye Institute (BPdata).
Results: :
The general agreement between predicted and measured OCT thickness was good with a mean residual deviation (MRD) averaged across sectors of -5.2 µm for the UHdata and -12.8 µm for the BPdata. Mean RNFL thickness derived from SAP data was within 10 µm of measured thickness for 54 of 112 eyes (48%) and within 20 µm for 84 (75%) of the eyes. The mean absolute deviation (MAD) of the 8 sectors was 20.2 µm for the UHdata and 23.1 µm for the BPdata. The accuracy and precision of the sector analyses were lower than analyses of superior and inferior hemifields (MRD = 1.4 µm, MAD = 15.1µm for the UHdata and MRD = -0.3 µm , MAD = 13.2 µm for the BPdata).
Conclusions: :
The regional analysis should be clinically useful for direct comparisons of objective and subjective assessments of glaucomatous neuropathy on a common scale. The overall accuracy of estimating RNFL thickness from SAP data was good, but with relatively low precision. The imprecision is likely the result of, 1) deriving RNFL thickness on a linear scale from SAP measures on a logarithmic scale and, 2) inter-subject variations in the maxima of the TSNIT curve leading to variability in mapping the visual field onto the optic nerve head.
Keywords: imaging/image analysis: clinical • perimetry • ganglion cells