Diagnostic ability of confocal near-infrared reflectance fundus imaging to detect retrograde microcystic maculopathy from chiasm compression. A comparative study with OCT findings

Abstract

Purpose To evaluate the ability of confocal near-infrared reflectance (NIR) to diagnose retrograde microcystic maculopathy (RMM) in eyes with temporal visual field (VF) loss and optic atrophy from chiasmal compression. To compare NIR findings with optical coherence tomography (OCT) findings in the same group of patients. Methods Thirty-four eyes (26 patients) with temporal VF loss from chiasmal compression and 41 healthy eyes (22 controls) underwent NIR fundus photography, and macular OCT scanning. VF loss was estimated and retinal layers thickness were measured on OCT. Two examiners blinded to the diagnosis randomly examined NIR images for the presence of hyporeflective abnormality (HA) and OCT scans for the presence of microcystic macular abnormalities (MMA). The total average and hemi-macular HA area and number of microcysts were determined. The groups were compared and the level of agreement was estimated. Results The OCT-measured macular retinal nerve fiber and ganglion cell layers were thinner and the inner nuclear layer was thicker in patients compared to controls. HA and MMA were detected in 22 and 12 patient eyes, respectively, and in 0 controls (p<0.001, both comparisons). HA was significantly more frequent than MMA in patients with optic atrophy, and agreement between HA and MMA (both total and hemi-macular) was fair (kappa range: 0.24–0.29). The mean HA area was significantly greater in the nasal than temporal hemi-macula. A re-analysis of the 14 eyes with discrepant findings allowed to confirm RMM in 20 eyes (20/34) indicating that OCT detected RMM in 12 and missed it in 8 eyes. On the other hand, NIR correctly detected 18 out of 20 eyes, overcalled 4 and missed 2. Conclusions RMM is a frequent finding in eyes with severe VF loss from long-standing chiasmal compression. NIR photography appears to be more sensitive than OCT for detecting RMM and may be useful as screening method for its presence.