Microvascular and structural alterations in the optic nerve head of advanced primary open-angle glaucoma compared with atrophic non-arteritic anterior ischemic optic neuropathy

A Deep Learning System Using Optical Coherence Tomography Angiography to Detect Glaucoma and Anterior Ischemic Optic Neuropathy

INTRODUCTION

Glaucoma and non-arteritic anterior ischemic optic neuropathy (NAION) are optic neuropathies that can both lead to irreversible blindness. Several studies have compared optical coherence tomography angiography (OCTA) findings in glaucoma and NAION in the presence of similar functional and structural damages with contradictory results. The goal of this study was to use a deep learning system to differentiate OCTA in glaucoma and NAION.

MATERIAL AND METHODS

Sixty eyes with glaucoma (including primary open angle glaucoma, angle-closure glaucoma, normal tension glaucoma, pigmentary glaucoma, pseudoexfoliative glaucoma and juvenile glaucoma), thirty eyes with atrophic NAION and forty control eyes (NC) were included. All patients underwent OCTA imaging and automatic segmentation was used to analyze the macular superficial capillary plexus (SCP) and the radial peripapillary capillary (RPC) plexus. We used the classic convolutional neural network (CNN) architecture of ResNet50. Attribution maps were obtained using the "Integrated Gradients" method.

RESULTS

The best performances were obtained with the SCP + RPC model achieving a mean area under the receiver operating characteristics curve (ROC AUC) of 0.94 (95% CI 0.92-0.96) for glaucoma, 0.90 (95% CI 0.86-0.94) for NAION and 0.96 (95% CI 0.96-0.97) for NC.

CONCLUSION

This study shows that deep learning architecture can classify NAION, glaucoma and normal OCTA images with a good diagnostic performance and may outperform the specialist assessment.

Characteristics of macrophage-like cells in acute nonarteritic anterior ischemic optic neuropathy and the normal fellow eyes on en face optical coherence tomography

Purpose

To quantitatively analyze macrophage-like cells (MLCs) at the vitreoretinal interface (VRI) in acute nonarteritic anterior ischemic optic neuropathy (NAION) using en face swept-source optical coherence tomography (OCT).

Methods

Twenty-five acute NAION eyes and 25 normal fellow eyes from 25 patients were included in the study. MLCs were visualized in a 3 μm 6 mm×6 mm en face OCT slab above the VRI centered on the optic nerve head (ONH). After semiautomatic binarization and quantification, we evaluated the MLC density between the two groups. We also investigated the relationship between MLC density and other OCT parameters, including the increase in peripapillary retinal nerve fiber layer (RNFL) thickness and loss of vessel density (VD) in radical peripapillary capillary (RPC).

Results

The MLC density in the affected eye of the ONH was highly correlated with that in the fellow eye (r=0.612, p=0.001). The MLC density significantly increased in acute NAION eyes (NAION vs. normal, 11.97 ± 6.66 vs. 9.31 ± 6.10 cell/mm², p=0.028). In sectorial analysis, the increase in MLCs was mainly in the superior regions (4.13 ± 7.49 vs. 0.94 ± 5.21 cell/mm², p=0.001). The VD of RPC decreased significantly in the affected eyes (NAION vs. normal, 37.54 ± 5.25 vs. 40.56 ± 4.25, p=0.016), and the loss of RPC was predominantly in the superior sectors and the temporal sectors when compared to the inferior sectors and the nasal sectors, respectively (superior vs. inferior, -3.54 ± 6.71 vs. -0.37 ± 8.07, p=0.004; temporal vs. nasal, -2.69 ± 8.72 vs. -1.22 ± 6.06, p=0.005). The loss of RPC and the increase in MLC density were greater in affected sector corresponding to the visual field (VF) defect.

Conclision

MLCs located above the VRI increased significantly in acute NAION eyes, especially in subregions corresponding to VF defect, which provides clinical evidence supporting that the inflammatory response participates in the pathological process of NAION. The magnitude of the increase in MLCs corresponds to the RPC loss in the quadrant analysis.

Optical coherence tomography evaluation of retinal nerve fiber layer thickness in non-arteritic anterior ischemic optic neuropathy and primary open angle glaucoma: a systematic review and Meta-analysis

AIM

To assess the differences in average and sectoral peripapillary retinal nerve fiber layer (pRNFL) thickness using spectral domain optical coherence tomography (SD-OCT) in patients with non-arteritic anterior ischemic neuropathy (NAION) compared with those with primary open angle glaucoma (POAG).

METHODS

A comprehensive literature search of the PubMed, Cochrane Library, and Embase databases were performed prior to October, 2021. Studies that compared the pRNFL thickness in NAION eyes with that in POAG eyes with matched mean deviation of the visual fields were included. The weighted mean difference (WMD) with 95% confidence interval (CI) was used to pool continuous outcomes.

RESULTS

Ten cross-sectional studies (11 datasets) comprising a total of 625 eyes (278 NAION eyes, 347 POAG eyes) were included in the qualitative and quantitative analyses. The pooled results demonstrated that the superior pRNFL was significantly thinner in NAION eyes than in POAG eyes (WMD=-6.40, 95%CI: -12.22 to -0.58, P=0.031), whereas the inferior pRNFL was significant thinner in POAG eyes than in NAION eyes (WMD=11.10, 95%CI: 7.06 to 15.14, P≤0.001). No difference was noted concerning the average, nasal, and temporal pRNFL thickness (average: WMD=1.45, 95%CI: -0.75 to 3.66, P=0.196; nasal: WMD=-2.12, 95%CI: -4.43 to 0.19, P=0.072; temporal: WMD=-1.24, 95%CI: -3.96 to 1.47, P=0.370).

CONCLUSION

SD-OCT based evaluation of inferior and superior pRNFL thickness can be potentially utilized to differentiate NAION from POAG, and help to understand the different pathophysiological mechanisms between these two diseases. Further longitudinal studies and studies using eight-quadrant or clock-hour classification method are required to validate the obtained findings.