and for the BONSAI (Brain and Optic Nerve Study with Artificial Intelligence) Study Group Objective: To compare the diagnostic performance of an artificial intelligence deep learning system with that of expert neuro-ophthalmologists in classifying optic disc appearance. Methods: The deep learning system was previously trained and validated on 14,341 ocular fundus photographs from 19 international centers. The performance of the system was evaluated on 800 new fundus photographs (400 normal optic discs, 201 papilledema [disc edema from elevated intracranial pressure], 199 other optic disc abnormalities) and compared with that of 2 expert neuro-ophthalmologists who independently reviewed the same randomly presented images without clinical information. Area under the receiver operating characteristic curve, accuracy, sensitivity, and specificity were calculated. Results: The system correctly classified 678 of 800 (84.7%) photographs, compared with 675 of 800 (84.4%) for Expert 1 and 641 of 800 (80.1%) for Expert 2. The system yielded areas under the receiver operating characteristic curve of 0.97 (95% confidence interval [CI] = 0.96-0.98), 0.96 (95% CI = 0.94-0.97), and 0.89 (95% CI = 0.87-0.92) for the detection of normal discs, papilledema, and other disc abnormalities, respectively. The accuracy, sensitivity, and specificity of the system's classification of optic discs were similar to or better than the 2 experts. Intergrader agreement at the eye level was 0.71 (95% CI = 0.67-0.76) between Expert 1 and Expert 2, 0.72 (95% CI = 0.68-0.76) between the system and Expert 1, and 0.65 (95% CI = 0.61-0.70) between the system and Expert 2. Interpretation: The performance of this deep learning system at classifying optic disc abnormalities was at least as good as 2 expert neuro-ophthalmologists. Future prospective studies are needed to validate this system as a diagnostic aid in relevant clinical settings.

Assessment of the vasculature within the optic nerve, peripapillary superficial retina, macula, and peripapillary choroid can be determined in glaucoma using optical coherence tomography angiography (OCTA). Decreased perfusion within the pre-laminar layer of the optic nerve has been correlated with glaucoma severity. The peripapillary superficial retinal vessel density allows diagnosis and detection of glaucoma progression in a manner similar to the peripapillary retinal nerve fiber layer (RNFL) thickness.Furthermore, decreased peripapillary vessel density of the intact hemiretina or unaffected eye of glaucomatous eyes suggests that vascular changes can occur prior to detectable visual field damage. The accuracy for glaucoma detection of the macular ganglion cell (MGC) thickness compared to macular vessel density has differed among studies. Several studies have reported reduction of macular vessel density as well as its ganglion cell thickness. Results of studies evaluating the parapapillary choroid have shown a greater prevalence of choroidal microvasculature dropout in glaucomatous eyes with a parapapillary gamma zone,which is associated with central visual field defects or glaucoma progression. It remains unclear whether the reduced vessel density in glaucoma is a primary event or secondary to glaucomatous damage. Further studies are warranted to elucidate this question.

Background: Peripapillary and macular microvasculature alterations after nonarteritic ischemic optic neuropathy (NAION) have been investigated in several studies. We aimed to explore the vascular changes from acute NAION (aNAION) to chronic NAION (cNAION). Methods: This prospective observational study composed of 16 eyes with aNAION and 40 healthy agematched controls. Eyes with NAION were followed up for more than 6 months after acute event. Optical coherence tomography angiography (OCTA) was used to evaluate peripapillary and macular vessel densities (VDs). The customized software was used for calculating deep retinal VD to attenuate the large superficial vessel projection effect. Result: The mean age of patients with NAION and controls was 56.13 ± 13.2 and 54.46 ± 15.5 years, respectively (P = 0.195). Radial peripapillary capillary density was significantly lower in both eyes with aNAION and eyes with cNAION than healthy eyes. Peripapillary capillary density decreased significantly from the acute to the chronic phase of NAION with values of 41.77 ± 4.05% and 34.35 ± 7.30%, respectively (P , 0.001). The mean superficial macular VD was 46.83 ± 3.47% in aNAION and 44.49 ± 4.50% in cNAION eyes with no significant difference between them (P = 0.252), but both were lower than control eyes. Deep macular VD was not affected in aNAION and cNAION eyes compared with control eyes. Correlation analysis in eyes with cNAION revealed that there were significant correlations between peripapillary nerve fiber layer and the capillary density (r = 0.772, P , 0.001) and between ganglion cell complex thickness and corresponding superficial macular VD. Conclusions: Although a decrease in peripapillary capillary density in aNAION eyes with active disc edema progressed when evaluated in the cNAION state, progressive VD loss was not observed in the macular area, suggesting a nonprogressive nature of macular vessel involvement in NAION.

Purpose: To compare peripapillary perfused capillary density (PCD) on optical coherence tomography angiography among resolved acute angle-closure (AAC), primary open-angle glaucoma (POAG), and control eyes. Design: Prospective, cross-sectional, observational study. Methods: All patients with resolved AAC or POAG of varying severity and controls were enrolled. We obtained 4.5 × 4.5 mm2 optical coherence tomography angiography images of the optic nerve head. PCD was analyzed using customized software with major vessel removal. Continuous variables were assessed using the analysis of variance and Bonferroni correction test. A marginal model of generalized estimating equations was used to adjust for confounding factors and interocular correlation. Results: The study included 44 eyes with resolved AAC (mean duration of elevated intraocular pressure, 8.1 ± 10.9 days), 69 eyes with POAG, and 49 control eyes. PCD showed a similar decrease between AAC and POAG eyes (P > 0.99). After adjusting for age and sex, the mean difference in global PCD between each of the POAG stage groups and the AAC group was the highest in the severe POAG group (−3.43; 95% confidence interval [CI], −11.38 to 2.52; P = 0.211), followed by the mild POAG (0.68; 95% CI, −3.26 to 4.62; P = 0.735) and moderate POAG (0.20; 95% CI, −5.21 to 5.61; P = 0.942) groups. The duration of increased intraocular pressure did not affect PCD (P = 0.258 and 0.168 for global and annular PCDs, respectively). Conclusions: The degree of microvascular attenuation in AAC eyes was not different from that in POAG eyes.

No abstract

No abstract

No abstract

This work aims at determining the ability of a deep learning (DL) algorithm to measure retinal nerve fiber layer (RNFL) thickness from optical coherence tomography (OCT) scans in anterior ischemic optic neuropathy (NAION) and demyelinating optic neuritis (ON). The training/validation dataset included 750 RNFL OCT B-scans. Performance of our algorithm was evaluated on 194 OCT B-scans from 70 healthy eyes, 82 scans from 28 NAION eyes, and 84 scans of 29 ON eyes. Results were compared to manual segmentation as a ground-truth and to RNFL calculations from the built-in instrument software. The Dice coefficient for the test images was 0.87. The mean average RNFL thickness using our U-Net was not different from the manually segmented best estimate and OCT machine data in control and ON eyes. In NAION eyes, while the mean average RNFL thickness using our U-Net algorithm was not different from the manual segmented value, the OCT machine data were different from the manual segmented values. In NAION eyes, the MAE of the average RNFL thickness was 1.18 ± 0.69 μm and 6.65 ± 5.37 μm in the U-Net algorithm segmentation and the conventional OCT machine data, respectively (P = 0.0001).