Does low-tech disc assessment correlate with retinal nerve fibre layer loss, measured with high-tech, spectral-domain OCT?

Abstract

 

Purpose

 

To investigate, whether clinical optic disc assessment and grading correlates with retinal nerve fibre layer thickness (RNFLt) loss, measured with spectral domain (sd) OCT.

 

Methods

 

143 normal and 369 glaucoma patients from the Erlangen Glaucoma Registry (www.clinicaltrials.gov (NCT00494923) were examined with indirect ophthalmoscopy with a 78 dpt lens and graded according to the criteria of Jonas (1). Eyes were assessed by two experienced examiners (CM, RL) from 0 (normal) to IV (progressed) according to the appearance of the neuroretinal rim (NRR) (I=vertical elongation of cup, II=NRR notching one clock hour, III=NRR loss three clock hours, IV=temporal NRR loss). RNFL was measured under a circumpapillary ring mean 360° and in 45° sectors with SPECTRALIS OCT (Heidelberg Engineering). Standardized clinical examination included standard static white on white perimetry, 24-h IOP profile, stereographic optic disc slides and Scanning Laser Tomography (HRT III, Software Version 3.0; Heidelberg Engineering, Heidelberg, Germany). The effect of glaucoma stage and RNFL on the diagnostic accuracy was evaluated using receiver operating characteristic (ROC) curves and non-parametric tests with SPSS 20.

 

Results

 

183 eyes were classified stage I, 106 stage II, 52 stage III and 28 stage IV. There was no significant difference for refraction and disc area. Mean RNFL thickness was for 0 97.5±9.5µm, for I 81.0±14.0µm, for II 67.6±13.5µm, for III 59.0±12.6µm and for IV 49.6±9.3µm (Figure1). Mean RNFLt was significantly different from 0 in all stages I-IV (p<0.05%). Areas under the ROC curves (AUROC) for discrimination between healthy and glaucoma stages showed best discrimination in stage I with mean RNFLt (0.84), followed by nasally inferior (0.80), for stage II with mean RNFLt (0.96), followed by temporally inferior (0.95), for stage III with temporally inferior (0.99), followed by mean RNFLt (0.98) and for stage IV with both mean RNFLt and temporally inferior, (1.0), followed by temporally superior.

 

Conclusions

 

Clinical low-cost optic disc assessment of NRR according to Jonas’ criteria reflects glaucomatous RNFLt loss measured with SD-OCT very well, although stage I showed a wide overlap with stage 0. 1) Jonas JB et al. Survey of Ophthalmol 1999; 43: 293-320

 

 

Fig.1: Boxplot diagramme showing RNFL thickness with SD-OCT in morphologic disc stages.

 

Keywords: 550 imaging/image analysis: clinical • 627 optic disc • 610 nerve fiber layer