Use of En Face Optical Coherence Tomography to Monitor Papilledema in Idiopathic Intracranial Hypertension: A Pilot Study

Original Contribution Section Editors: Clare Fraser, MD Susan Mollan, MD Use of En Face Optical Coherence Tomography to Monitor Papilledema in Idiopathic Intracranial Hypertension: A Pilot Study Andrew R. Carey, MD, Thomas M. Bosley, MD, Neil R. Miller, MD, Timothy J. McCulley, MD, Amanda D. Henderson, MD Background: En face optical coherence tomography (OCT) uses the data acquired during OCT of the optic disc, which typically is used to determine measurements of the peripapillary retinal nerve fiber layer (ppRNFL), to generate a coronal composite fundus image rather than a crosssectional image. En face OCT has been reported to identify retinal changes related to papilledema in idiopathic intracranial hypertension (IIH) but has not been evaluated for monitoring papilledema. This study aimed to assess the reliability and validity of en face OCT for monitoring papilledema. Methods: Using the Pearson correlation coefficient (R), these measurements were compared with ppRNFL thickness as well as average diameter and estimated area. Four fellowship-trained neuro-ophthalmologists were asked to qualitatively rank en face images by the area of optic disc edema while masked from all other clinical data. Rankings were compared with ppRNFL thickness as a gold standard and with en face OCT characteristics using the Pearson correlation coefficient (R). Results: Experts were able to correctly identify an increase in average ppRNFL thickness .10 mm with a mean (SD) of 91% (±7%) accuracy. A ranking error among experts corresponded to a mean (standard error) change in the ppRNFL thickness of 6 (±6) mm. The mean Pearson correlation coefficient by the area of disc edema among experts was 0.92 (±0.13). Conclusions: Multiple objective parameters of en face OCT of optic disc edema have an excellent correlation with ppRNFL thickness. These results suggest that en face OCT is a valid clinical tool for monitoring papilledema in IIH. Journal of Neuro-Ophthalmology 2021;41:212–216 doi: 10.1097/WNO.0000000000000940 © 2020 by North American Neuro-Ophthalmology Society Division of Neuro-Ophthalmology, Wilmer Eye Institute, Johns Hopkins Medicine, Baltimore, Maryland. The authors report no conflicts of interest. Address correspondence to Andrew R. Carey, MD, Division of Neuro-Ophthalmology, Wilmer Eye Institute, Johns Hopkins Medicine, 600 N. Wolfe Street, Woods 459A, Baltimore, MD 21287; E-mail: [e-mail] 212 I diopathic intracranial hypertension (IIH) is a visionthreatening disorder that results in papilledema. The Idiopathic Intracranial Hypertension Treatment Trial (IIHTT) demonstrated that the risk of visual field loss is proportional to the degree of optic disc edema (1). Data from the IIHTT also demonstrated that optical coherence tomography (OCT) to measure the average peripapillary retinal nerve fiber layer (ppRNFL) thickness is more useful for monitoring papilledema and response to treatment than color fundus photographs (2). However, there are limits to OCT of the ppRNFL thickness, including unreliable segmentation in high-grade disc edema that can result in incorrect thickness measurements. En face OCT compiles various B-scan OCTs into a composite image, which then can be viewed coronally to form a fundus image based on the fundus reflectivity. This en face has a number of advantages: Because it is a digital image, various parameters can be measured using commercially available calipers that are part of the Cirrus OCT software; because it is grayscale, it is easily transmitted through fax; and because it is not as a dependent on the segmentation software compared with ppRNFL segmentation, it can deliver an image of adequate quality for interpretation even in the setting of poor signal quality and significant optic disc edema. In addition, the image is easily obtained through an undilated pupil and does not require a skilled photographer. En face OCT has been used to evaluate ppRNFL bundle defects in glaucoma (3), peripapillary atrophy (4), optic disc pits (5), microcystic macular edema in optic atrophy (6), and structural changes in the scleral canal (7). The IIHTT used en face OCT to evaluate retinal and choroidal folds in papilledema (8), but en face OCT has not been evaluated previously for monitoring papilledema. METHODS Four sets (2 eyes each of 2 patients) of 5 Cirrus optic disc OCT scans were collected longitudinally from patients with Carey et al: J Neuro-Ophthalmol 2021; 41: 212-216 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution FIG. 1. A-E. Sample of randomly assorted of en face optical coherence tomography images of papilledema from the right eye of one patient over a 6–8 month period, as presented for ranking. Correct ranking order least to greatest edema: A, C, E, D, B. OD, right eye. IIH over a period of 6–8 months, representing 20 scans in total. Inclusion criteria were the presence of optic disc edema at baseline, OCT at baseline, and a minimum of 4 follow-up OCT scans over a period of at least 6 months. Individual scans were reviewed to ensure image quality including adequate segmentation, and scans were excluded from relevant analyses if the segmentation was not adequate for interpretation. The scans were reviewed using the en face tool of the Cirrus software (Carl Zeiss Meditec, Inc, Dublin, CA). The caliper tool within the Cirrus software was used to measure the horizontal and vertical dimensions of the area of the optic disc edema. Using these data, the mean of horizontal and vertical diameters was calculated as well as the area of disc edema based on the mean diameter (circle model) and the assumption of an ellipse. These values were compared with ppRNFL thickness in the TABLE 1. ppRNFL, average diameter of disc edema, estimated area of disc edema by the circle method, and average rank of disc edema by neuro-ophthalmologists for each data point ppRNFL (mm) 95 114 117 215 427 98 91* 103 140 231 93 104 117 136 148 82 93 107 111 126 Diameter of Disc Edema (mm) Area of Disc Edema (mm2) Rank by Neuro-Ophthalmologists (1 = Least to 5 = Greatest) 1.82 2.07 2.10 3.24 3.81 1.41 1.68 2.04 2.67 3.18 2.07 2.21 2.27 2.61 2.68 2.04 2.19 2.28 2.42 2.50 2.6 3.4 3.5 8.3 11.4 1.6 2.2 3.3 5.6 7.9 3.4 3.8 4.1 5.4 5.6 3.3 3.8 4.1 4.6 4.9 1 2 3 4 5 2 2 2 4 5 1 2 3 4 5 1 2.75 3.75 2.5 5 *Segmentation errors caused artificially low ppRNFL measurement. ppRNFL, peripapillary retinal nerve fiber layer. Carey et al: J Neuro-Ophthalmol 2021; 41: 212-216 213 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution FIG. 2. Peripapillary RNFL OCT and en face optical coherence tomography from data point with segmentation errors. The arrow represents the anatomic ILM-RNFL interface, the red line represents ILM-RNFL segmentation, the arrowhead represents the anatomic RNFL-GCL interface, and the purple line represents RNFL-GCL. GCL, ganglion cell layer; ILM, internal limiting membrane; OCT, optical coherence tomography; RNFL, retinal nerve fiber layer. corresponding OCT using the Pearson correlation coefficient. The en face images were then extracted, and 4 fellowship-trained neuro-ophthalmologists were presented with a series of images from each eye, in random order (a sample is shown in Figure. 1), and then were asked to rank the images by the area of disc edema. Expert rankings were compared with the ppRNFL thickness and with objective measurements of the en face images using the Pearson correlation coefficient. RESULTS Average ppRNFL thickness measurements ranged from 82 to 427 mm with good segmentation. En face OCT measurements of the diameter of disc edema ranged from 1.363 to 3.135 mm for horizontal measurements, 1.463–4.485 mm for vertical measurements, and 1.413–3.810 mm for mean diameters; vertical diameters were greater than horizontal diameters in all instances. The ellipse method for the estimated area of disc edema resulted in a smaller estimation than that of the circle method in all instances ranging from 0 to 0.36 mm2, with a mean of 0.10 and an SD of 0.08 mm2. Both methods had an excellent agreement with ppRNFL thickness; circle method, an R-squared value of 0.868 and ellipse method, 0.866. The estimated area of disc edema by the circle method ranged from 1.568 to 11.399 mm2. Table 1 shows composite data for each OCT image, including ppRNFL thickness, mean diameter of disc edema, area of disc edema (by the circle model), and average ranking by experts. One OCT was found to have a segmentation error in the ppRNFL; it was omitted from analyses involving the ppRNFL but was otherwise included. The ppRNFL image and en face OCT image are shown in Figure 2. The ratio of average ppRNFL thickness to en face OCT measurements is displayed in Table 2. Figures 3 and 4 demonstrate scatter plots of the diameter of disc edema and area of disc edema, respectively, compared with the ppRNFL along with trend lines and R-squared values. Average ppRNFL thickness had the closest correlation with estimated area of disc edema by average diameter derived by the circle method, as determined by lowest SD and variance among parameters, although there was still a significant range. Rankings of papilledema by en face OCT measurements of the area of optic disc edema compared with the ppRNFL thickness ranged 0.94–0.97 with SD ranging ±0.02–0.03, with the estimated area having the highest correlation. Expert grading of the area of disc edema by fellowshiptrained neuro-ophthalmologists compared with objective OCT measurements is displayed in Table 3 and had the highest correlation with ppRNFL thickness. TABLE 2. Ratio of ppRNFL thickness to en face OCT measurements of optic disc edema Mean SD Max Min Variance Pearson correlation coefficient ppRNFL:Horizontal Diameter (mm:mm) ppRNFL:Vertical Diameter (mm:mm) ppRNFL:Average Diameter (mm:mm2) ppRNFL:Estimated Area by Average Diameter (Circle Method) (mm:mm2) ppRNFL:Estimated Area by Average Diameter (Ellipse Method) (mm:mm2) 65 19.20 136 47 368 0.86 49 13.51 95 35 183 0.87 56 15.64 112 40 245 0.88 30 8.57 63 24 73 0.93 31 8.51 63 24 72 0.93 OCT, optical coherence tomography; ppRNFL, peripapillary retinal nerve fiber layer. 214 Carey et al: J Neuro-Ophthalmol 2021; 41: 212-216 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution FIG. 3. Scatter plot of peripapillary retinal nerve fiber layer vs en face OCT measurements. OCT, optical coherence tomography; ppRNFL, peripapillary retinal nerve fiber layer. FIG. 4. Scatter plot of peripapillary retinal nerve fiber layer vs area of optic disc edema. ppRNFL, peripapillary retinal nerve fiber layer. Carey et al: J Neuro-Ophthalmol 2021; 41: 212-216 215 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution TABLE 3. Expert ranking of severity of optic disc edema vs objective OCT parameters OCT Feature ppRNFL thickness Horizontal diameter Vertical diameter Average diameter Estimated area R Mean (SD) 0.959 0.80 0.72 0.72 0.77 (±0.003) (±0.07) (±0.06) (±0.06) (±0.03) OCT, optical coherence tomography; ppRNFL, peripapillary retinal nerve fiber layer. Using en face OCT images, experts correctly identified worsening of optic disc edema corresponding to an increase in ppRNFL thickness of $10 mm with a mean accuracy of 91% (±7%). A ranking error among expert assessments of en face OCT corresponded to a mean change in the ppRNFL thickness of only 6 (±6) mm. Expert ranking of en face OCT images had an agreement of 0.92 (±0.13) by the Pearson correlation coefficient. CONCLUSIONS All objective parameters of en face OCT of optic disc edema had a high correlation with ppRNFL thickness. Experts were able to use en face OCT images to rank qualitatively the severity of papilledema with high accuracy and reliability, as well as to detect small changes in optic disc edema and ppRNFL thickness. En face OCT images are already available in the standard data acquired using the Cirrus OCT optic disc protocol, and these images may be more reliable than ppRNFL thickness measurements in the setting of poor scan quality because of low signal strength and segmentation errors. Similar to any other OCT images, en face OCT is susceptible to blink and motion artifacts. Because of individual anatomic variations, en face OCT images cannot be compared across patients, and direct comparison of en face OCT of optic disc edema with ppRNFL thickness cannot be related across patients. This was a pilot study designed for proof of concept and has certain limitations. First, sample sizes of patients and of neuro-ophthalmologists were small; therefore, results may not be generalizable. However, the sample did include a wide range of papilledema severity and neuroophthalmologists with a wide range of experience. Future research will involve automated measurements of the diameter and area of optic disc edema, larger crosssectional data from a validated sample such as the IIHTT, comparison with color photographs, and comparison with 216 longitudinal data to examine the ability to predict the risk of vision loss. In summary, we report a novel application of en face OCT for monitoring patients with papilledema, which appears to provide additional objective data that are both reliable and useful. STATEMENT OF AUTHORSHIP Category 1: a. Conception and design: A. R. Carey; b. Acquisition of data: A. R. Carey; c. Analysis and interpretation of data: A. R. Carey and A. D. Henderson. Category 2: a. Drafting the manuscript: A. R. Carey, T. M. Bosley, N. R. Miller, T. J. McCulley, and A. D. Henderson; b. Revising it for intellectual content: A. R. Carey, T. M. Bosley, N. R. Miller, T. J. McCulley, and A. D. Henderson. Category 3: a. Final approval of the completed manuscript: A. R. Carey, T. M. Bosley, N. R. Miller, T. J. McCulley, and A. D. Henderson. REFERENCES 1. Wall M, Falardeau J, Fletcher WA, Granadier RJ, Lam BL, Longmuir RA, Patel AD, Bruce BB, He H, McDermott MP; NORDIC Idiopathic Intracranial Hypertension Study Group. Risk factors for poor visual outcome in patients with idiopathic intracranial hypertension. Neurology. 2015;85:799–805. 2. Sheils CR, Fischer WS, Hollar RA, Blanchard LM, Feldon SE; NORDIC Idiopathic Intracranial Hypertension Study Group. The relationship between optic disc volume, area, and Frisén score in patients with idiopathic intracranial hypertension. Am J Ophthalmol. 2018;195:101–109. 3. Hood DC, Fortune B, Mavrommatis MA, Reynaud J, Ramachandran R, Ritch R, Rosen RB, Dubra A, Chui TY. Details of glaucomatous damage are better seen on OCT en face images than on OCT retinal nerve fiber layer thickness maps. Invest Ophthalmol Vis Sci. 2015;56:6208–6216. 4. Miki A, Ikuno Y, Weinreb RN, Yokoyama J, Asai T, Usui S, Nishida K. Measurements of the parapapillary atrophy zones in en face optical coherence tomography images. PLoS One. 2017;12:e0175347. 5. Maertz J, Kolb JP, Klein T, Mohler KJ, Eibl M, Wieser W, Huber R, Priglinger S, Wolf A. Combined in-depth, 3D, en face imaging of the optic disc, optic disc pits and optic disc pit maculopathy using swept-source megahertz OCT at 1050 nm. Graefes Arch Clin Exp Ophthalmol. 2018;256:289–298. 6. Gocho K, Kikuchi S, Kabuto T, Kameya S, Shinoda K, Mizota A, Yamaki K, Takahashi H. High-resolution en face images of microcystic macular edema in patients with autosomal dominant optic atrophy. Biomed Res Int. 2013;2013:676803. 7. Srinivasan VJ, Adler DC, Chen Y, Gorczynski I, Huber R, Duker JS, Schuman JS, Fujimoto JG. Ultrahigh-speed optical coherence tomography for three-dimensional and en face imaging of the retina and optic nerve head. Invest Ophthalmol Vis Sci. 2008;49:5103–5110. 8. Kupersmith MJ, Sibony PA, Feldon SE, Wang JK, Garvin M, Kardon R; OCT Sub-Study Group for the NORDIC Idiopathic Intracranial Hypertension Treatment Trial. The effect of treatment of idiopathic intracranial hypertension on prevalence of retinal and choroidal folds. Am J Ophthalmol. 2017;176:77–86. Carey et al: J Neuro-Ophthalmol 2021; 41: 212-216 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited.