Assessment of ischemia in acute central retinal vein occlusion from inner retinal reflectivity on spectral domain optical coherence tomography

Figures & data

Figure 1 Grading standard used for a CRVO with none to mild ischemia (grade 1).

Notes: Top left panel: the SD-OCT B-scan for the normal right eye is shown for comparison. Top right panel: color fundus photograph of the left eye with CRVO. The veins are dilated, and a few scattered dot hemorrhages are present in all four quadrants of the fundus with some in the macula. Best-corrected visual acuity was 20/40. There was no relative afferent pupillary defect. Bottom left panel: the SD-OCT B-scan of the left eye showing excellent delineation of the inner retinal layers (yellow arrows; each layer seen in the normal scan of the top left panel can be identified here). Mild cystoid macular edema is present. Bottom right panel: a mid-phase frame from the fluorescein angiogram of the left eye shows the absence of ischemic zones in the retina and the presence of mild fluorescein leakage.
Abbreviations: CRVO, central retinal vein occlusion; SD-OCT, spectral domain optical coherence tomography.

Figure 2 Grading standard used for a CRVO with moderate ischemia (grade 2).

Notes: Top left panel: the SD-OCT B-scan for the normal right eye is shown for comparison. Top right panel: color fundus photograph of the left eye with CRVO. The veins are more dilated than in the case with mild ischemia (Figure 1), and more dot hemorrhages are present in all four quadrants of the fundus. The best-corrected visual acuity was 20/300. A moderate relative afferent pupillary defect was present. Bottom left panel: the SD-OCT B-scan of the left eye showing partial loss of delineation of the inner retinal layers and increased reflectivity (yellow arrows; compare to top left panel). Worse cystoid macular edema is present than in the case with mild ischemia and subretinal fluid is present, but neither factor was included in grading for ischemia. Bottom right panel: a mid-phase frame from the fluorescein angiogram of the left eye shows the absence of ischemic zones in the retina and the absence of fluorescein leakage.
Abbreviations: CRVO, central retinal vein occlusion; SD-OCT, spectral domain optical coherence tomography; N, nasal; S, superior; T, temporal; I, inferior.

Figure 3 Grading standard used for a CRVO with severe ischemia (grade 3).

Notes: Top left panel: the SD-OCT B-scan for the normal right eye is shown for comparison. Top right panel: color fundus photograph of the left eye with CRVO. The veins are dilated and severe intraretinal hemorrhage is present in all four quadrants of the fundus. The red arrow indicates an area with nearly confluent intraretinal hemorrhage. The yellow arrow indicates an area relatively spared from intraretinal hemorrhage, which is the area for which the SD-OCT can be assessed without confounding from blood. The visual acuity was counting fingers at 2 feet. A marked relative afferent pupillary defect of the left eye was present. Bottom left panel: the SD-OCT B-scan of the left eye showing severe loss of delineation of the inner retinal layers and marked increase in reflectivity (yellow arrow; compares to top right panel). There is also increased reflectivity and loss of delineation of inner retinal layers at the red arrow, but in this case, the changes arise from a combination of the intraretinal blood and the ischemia. The region indicated by the yellow arrow is used for grading as it lacks the confounding inner retinal hemorrhage. Marked cystoid macular edema is present but this factor was not included in grading. Bottom right panel: a mid-phase frame from the fluorescein angiogram of the left eye shows zones of capillary nonperfusion and zones in which perfusion cannot be graded due to the presence of intraretinal hemorrhage.
Abbreviations: CRVO, central retinal vein occlusion; SD-OCT, spectral domain optical coherence tomography; N, nasal; S, superior; T, temporal; I, inferior.

Figure 4 Bland–Altman plots of intragrader repeatability and intergrader reproducibility in the assessment of SD-OCT images in acute central retinal vein occlusion.

Notes: (A) Bland–Altman plot for 2 gradings of the same set of 39 SD-OCT images performed by grader 1. (B) Bland–Altman plot for 2 gradings of the same set of 39 SD-OCT images performed by grader 2. (C) Bland–Altman plot for gradings of graders 1 and 2 of the same set of 39 SD-OCT images. The dotted lines represent the limits of agreement.
Abbreviations: SD, standard deviation; SD-OCT, spectral domain optical coherence tomography.

Table 1 Characteristics of patients and eyes

Characteristic	Overall sample	Grade of ischemia			P-value
		Mild (N=13)	Moderate (N=17)	Severe (N=9)
Age (mean, SD)	70, 13	68, 11	69, 15	76, 11	0.2163
Gender (F:M)	18:21	6:7	6:11	6:3	0.3076
Hypertension, N (%)	24 (62)	7 (54)	11 (65)	6 (67)	0.7815
Diabetes, N (%)	8 (21)	5 (38)	1 (6)	2 (22)	0.775
Primary open angle glaucoma, N (%)	5 (13)	1 (8)	3 (18)	1 (11)	0.7057
Other macular conditions (number without: list of other conditions, N)	32: DME (3), ERM (2), AMD (2)	10: DME (2), ERM (1)	14: ERM (1), AMD (2)	8: DME (1)	0.3378

Abbreviations: AMD, age-related macular degeneration; DME, diabetic macular edema; ERM, epiretinal membrane; F, female; M, male; N, number of patients; SD, standard deviation.

Table 2 Distribution of baseline characteristics and outcomes by baseline grade of ischemia

Characteristic	Baseline grade of ischemia			P-value
	Mild	Moderate	Severe
N	13	17	9
Baseline VA	0.480 (0.300, 0.651)	0.602 (0.300, 1.088)	0.903 (0.521, 1.626)	0.1246*
12-month VA	0.200 (0.100, 0.449)	0.200 (0.100, 1.000)	0.480 (0.300, 1.577)	0.1266*
12-month change in VA	−0.300 (−0.439, 0.049)	−0.100 (−0.410, 0.300)	0.000 (−0.400, 0.300)	0.5780*
Baseline CSMT	551 (407, 593)	659 (421, 708)	596 (414, 736)	0.1901*
12-month CSMT	332 (231, 429)	314 (251, 446)	495 (330, 694)	0.0931*
12-month change in CSMT	−179 (−95, −272)	−280 (−61, −345)	−124 (194, −321)	0.4822*
Prevalence of baseline RAPD (%)	15.4	23.5	44.4	0.4096**
Incidence of ASNV over 12 months (%)	15.4	17.6	8.9	0.9040**

Notes: Entries are medians (interquartile range). Grades of ischemia are 1 (mild), 2 (moderate), and 3 (severe).

* P-values reflect Kruskal–Wallis testing.

** P-values reflect likelihood ratio testing.

Abbreviations: ASNV, anterior segment neovascularization; CSMT, central subfield mean thickness; N, number of patients; RAPD, relative afferent pupillary defect; VA, visual acuity.

Figure 5 The time course of logMAR visual acuity graphed as a function of time in months and stratified by baseline ischemia grade as graded from SD-OCT images.

Note: The lines represent grade 1 ischemia subgroup (green line), the grade 2 ischemia subgroup (blue line), and grade 3 ischemia subgroup (red line).
Abbreviations: logMAR, logarithm of the minimum angle of resolution; SD-OCT, spectral domain optical coherence tomography.

Figure 6 Graphs of the relationships of three SD-OCT signs to each other.

Notes: In each panel, for each grade of ischemia, the green bar (y) indicates the proportion of eye that had the sign, and the red bar (n) indicates the proportion of eyes without the sign. (A) Graph of the proportion of eyes with the p-MLM sign as a function of the baseline grade of ischemia by SD-OCT grade. (B) Graph of the proportion of eyes with the PAMM sign as a function of the baseline grade of ischemia by SD-OCT grade. (C) Graph of the proportion of eyes with the PAMM sign as a function of the presence or absence of a p-MLM sign.
Abbreviations: PAMM, paracentral acute middle maculopathy; p-MLM, prominent middle-limiting membrane; SD-OCT, spectral domain optical coherence tomography.

Figure 7 Evolution of SD-OCT signs of ischemia over time.

Notes: (A) Near infrared reflectance imaging photograph of the right fundus of a 60-year-old woman with an acute central retinal vein occlusion of the right eye. (A) Red arrows indicate the hyporeflective lesion indicative of PAMM. Green arrow indicates the location of the scan in panel B. (B) Radial line scan from the SD-OCT showing the location of PAMM (purple arrow). The INL is hyperreflective compared to the adjacent normal region (green arrow). (C) Radial line scan from the SD-OCT 17 days after the image of B showing resolution of PAMM, development of cystoid macular edema, and new development of the p-MLM sign (turquoise arrow) at the outer border of the INL. In addition, a new sign, a hyperreflective line at the inner border of the INL similar to the p-MLM sign, is seen (yellow arrow).
Abbreviations: INL, inner nuclear layer; PAMM, paracentral acute middle maculopathy; p-MLM, prominent middle-limiting membrane; SD-OCT, spectral domain optical coherence tomography.

Table 3 Number of interventions over 12 months in acute central retinal vein occlusion

Outcome	Baseline ischemia status by SD-OCT grading			P-value
	1 (mild)	2 (moderate)	3 (severe)
Number of anti-VEGF injections	6 (5, 9), 11	6 (4, 9), 12	6 (3, 7), 12	0.8540
Number of PRPs	0 (0, 0), 0	0 (0, 0), 1	0 (0, 0), 1	0.5094
Number of grid laser treatments	0 (0, 0), 1	0 (0, 0), 0	0 (0, 0), 0	0.3679
Number of intravitreal steroid injections	0 (0, 0), 1	0 (0, 0), 1	0 (0, 0), 2	0.9238
Incisional glaucoma procedures (total cases: number of cases of specified ischemia grade)	1:13	1:17	0:9	0.5712
Laser cyclophotocoagulation (total cases: number of cases of specified ischemia grade)	1:13	0:17	1:9	0.2941

Notes: The number of anti-VEGF injections applied for 12 months following the diagnosis of central retinal vein occlusion. Entries are medians followed by interquartile ranges in parentheses, in turn, followed by the maximum number for the subgroup. The number of anti-VEGF injections is the sum of bevacizumab, ranibizumab, and aflibercept injections. The number of intravitreal steroid injections is the sum of intravitreal triamcinolone and dexamethasone sustained release device injections.

Abbreviations: PRPs, panretinal photocoagulations; SD-OCT, spectral domain optical coherence tomography; VEGF, vascular endothelial growth factor.