The Influence of Volume and Anatomic Location of Optic Disc Drusen on the Sensitivity of Autofluorescence

Background: Optic disc drusen (ODD) are acellular deposits in the optic nerve head. ODD can be diagnosed using different imaging modalities, including enhanced depth imaging optical coherence tomography (EDI-OCT) and autofluorescence (AF). It is unknown which factors determine the sensitivity of AF. The aim of this study was to investigate the effect of volume and anatomic location of ODD on the sensitivity of AF. Methods: Cross-sectional study. Results: A total of 38 patients (75 eyes) with ODD were included. In 12 of 75 eyes (16%) and in 11 of 38 patients (29%), EDI-OCT detected ODD that were not detected by AF. In 24 distinctly solitary ODD, both increase in ODD volume (P = 0.0388) and a more superficial ODD location (P < 0.0001) increased the possibility of AF detection of ODD, when performing a multivariate analysis. Conclusions: EDI-OCT is superior to AF in the diagnosis of ODD. Volume and anatomic location of ODD have a significant impact on the sensitivity of AF.

Original Contribution The Influence of Volume and Anatomic Location of Optic Disc Drusen on the Sensitivity of Autofluorescence Frederik Cornelius Loft, MB, Lasse Malmqvist, MD, Anne-Sofie Wessel Lindberg, MSc, Steffen Hamann, MD, PhD Background: Optic disc drusen (ODD) are acellular deposits in the optic nerve head. ODD can be diagnosed using different imaging modalities, including enhanced depth imaging optical coherence tomography (EDI-OCT) and autofluorescence (AF). It is unknown which factors determine the sensitivity of AF. The aim of this study was to investigate the effect of volume and anatomic location of ODD on the sensitivity of AF. Methods: Cross-sectional study. Results: A total of 38 patients (75 eyes) with ODD were included. In 12 of 75 eyes (16%) and in 11 of 38 patients (29%), EDI-OCT detected ODD that were not detected by AF. In 24 distinctly solitary ODD, both increase in ODD volume (P = 0.0388) and a more superficial ODD location (P , 0.0001) increased the possibility of AF detection of ODD, when performing a multivariate analysis. Conclusions: EDI-OCT is superior to AF in the diagnosis of ODD. Volume and anatomic location of ODD have a significant impact on the sensitivity of AF. Journal of Neuro-Ophthalmology 2019;39:23-27 doi: 10.1097/WNO.0000000000000654 © 2018 by North American Neuro-Ophthalmology Society O ptic disc drusen (ODD) are acellular, often calcified deposits in the optic nerve head, located anteriorly to lamina cribrosa (1). The prevalence is between 0.3% and 2.0% (2,3). Up to 87% of patients with ODD have visual field defects (3). Often they do not identify these field changes, as they develop very slowly (4-8). Although rare, severe complications such as acute loss of vision are known to occur (9-11). ODD in children are mostly buried deep Department of Ophthalmology (FCL, LM, SH), Rigshospitalet, University of Copenhagen, Denmark; and Department of Applied Mathematics and Computer Science (A-SWL), Technical University of Denmark, Denmark. The authors report no conflicts of interest. Address correspondence to Frederik Cornelius Loft, MB, Department of Ophthalmology, Rigshospitalet, University of Copenhagen, Valdemar Hansens Vej 1-23, 2600 Glostrup, Denmark; E-mail: [e-mail] Loft et al: J Neuro-Ophthalmol 2019; 39: 23-27 in the optic nerve head, whereas they are primarily superficial and visible on ophthalmoscopy in adults (12). Establishing the diagnosis of ODD is critical, as they may mimic the appearance of papilledema, especially if they are buried (1,13,14). Currently, ODD can be diagnosed using different imaging modalities such as B-scan ultrasonography, autofluorescence (AF) and, more recently, enhanced depth imaging optical coherence tomography (EDI-OCT). While EDI-OCT appears to be more reliable than other diagnostic methods (15-17), factors affecting the sensitivity of AF are unknown. Different parameters, such as ODD volume, ODD location, amount of ODD calcification, and optic nerve head anatomy might be important factors. Because AF of ODD, in theory, is inversely proportional to the depth of the ODD (18), it is expected that AF is less reliable in the diagnosis of ODD in children because of their high frequency of being buried. The purpose of our study was to investigate the effect of volume and anatomic location on the sensitivity of detecting ODD with AF. METHODS Patients Patients had a complete ophthalmic examination including EDI-OCT and AF (Heidelberg Engineering, Heidelberg, Germany). Examinations were performed at the Neuroophthalmology Clinic, Department of Ophthalmology, Rigshospitalet, Denmark, according to a previously published study (19). The patient group was selected by reviewing the files of 106 patients, classified by the International Statistical Classification of Diseases, 10th Revision, diagnosis code H47.3 "Drusen of the optic disc: Pseudopapilledema" and seen at the department from January 2009 to October 2014. The inclusion criteria were 1) visible ODD by slitlamp biomicroscopy or B-scan ultrasonography, 2) 18 years 23 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution FIG. 1. Measurement of the anatomic ODD location. After manual segmentation of a solitary ODD, the distance is measured from the top point of the ODD to the ILM (green bar). Arrowheads indicate blood vessels. BM, Bruch membrane; ILM, inner limiting membrane; ODD, optic disc drusen. or older, and 3) signed informed consent. All procedures adhered to the tenets of the Declaration of Helsinki, and the Scientific Ethics Committee of the Capital Region, Denmark (H-4-2013-040), approved the study. Optical Coherence Tomography and Autofluorescence All distinctly solitary ODD were located by searching through all included EDI-OCT scans. The ODD were characterized as having a hyperreflective margin with a hyporeflective center as reported in previous studies (15,20-22). Blood vessels might mimic ODD in EDI-OCT scans but were distinguished by their "figure of 8 configuration," a characteristic drop shadow or by their long tubular appearance, when scrolling through the EDI-OCT scans, as described by the ODD Studies Consortium (23) (Fig. 1). Requirements of ODD to be classified as distinctly solitary were as follows: 1) no other ODD located in the anterior/posterior direction; 2) a clear separation from other ODD in EDI-OCT en face view. In case of AF positivity, the following criteria applied as well: 1) the hyperautofluorescent area of the solitary ODD should not interfere with other hyperautofluorescent areas; 2) no AF negative ODD should be located within the hyperautofluorescent area of the solitary ODD. The identified solitary ODDs were classified as either AF positive or negative and segmented manually twice using the medical image segmentation tool ITK-Snap (ITK-Snap version 3.2.0, www.itksnap.org; in the public domain), enabling 3-dimensional (3D) reconstruction from which the ODD volume was calculated (Fig. 2). The inner limiting membrane (ILM) was detected using an automatic graph-based segmentation performed in MATLAB (MathWorks, Natick, MA). The top point of each ODD was defined as the point closest to ILM in 3D view. The distance in depth between the ILM and the top point of the ODD thereby described the anatomic location of solitary ODD (Fig. 1). The peripapillary retinal nerve fiber layer (RNFL) thickness was measured with optical coherence tomography in all patients, and the results of the AF-positive and AF-negative ODD eyes were compared. Statistical Analysis The included eyes were divided into an AF-positive and an AF-negative groups (Table 1) and analyzed, accounting for correlating eyes, using a mixed model analysis, and Fisher exact test was used for categorical variables. Statistical analysis was also performed on the solitary ODD. A generalized linear mixed model analysis with adjustment for correlating eyes was performed, to assess ODD volume and anatomic location as predictor for AF detection of ODD. This model was chosen because the binary outcome, as either AF positive or negative, is not normally distributed. The detection and volume measurement of the solitary ODD was performed twice with 1 month between the measurements. Intragrader variability was evaluated using Bland-Altman plot. The plot showed one outlier but acceptable variability, no trend, and the mean difference between volume measurements was 9.37 · 1025 mm3 (3.35 ± 0.2% of the mean volume of solitary ODD). A P value of ,0.05 was considered statistically significant. Statistical analysis was performed using SAS for Windows (Version 9.1; SAS Institute, Cary, NC). RESULTS A total of 38 patients (75 eyes) were included in this study. The RNFL was significantly thinner in the AF-positive group (P , 0.0001), when compared to the AF-negative group. There was no difference between the AF-positive FIG. 2. Different imaging modalities of a patient with an AF-negative ODD. A. Fundus photograph. B. AF with no hyperautofluorescence. C. EDI-OCT showing solitary ODD. D. Three-dimensional reconstruction of the 3 detected ODD (arrows). AF, autofluorescence; EDI-OCT, enhanced depth imaging optical coherence tomography; ODD, optic disc drusen. 24 Loft et al: J Neuro-Ophthalmol 2019; 39: 23-27 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution TABLE 1. Comparison of AF-positive and AF-negative eyes Gender ratio F:M Age (yrs) Spherical equivalent refraction (diopter) BCVA (ETDRS, letters) RNFL (mm) AF Positive (n = 62) AF Negative (n = 13) P value 51:11 31 (29) 20.31 (1.8) 86.5 (6) 70 (33) 9:4 32 (29) 20.25 (1) 87 (5) 101 (9) 0.28* 0.61† 0.2† 0.26† 0.0001† *Fisher exact test. † Mixed model (median [interquartile range]) analysis adjusting for correlated eyes. AF, autofluorescence; BCVA, best-corrected visual acuity; ETDRS, early treatment diabetic retinopathy study; RNFL, retinal nerve fiber layer. and AF-negative groups for sex, age, refraction, and bestcorrected visual acuity (Table 1). Of the 75 included eyes, 62 eyes had ODD visible on AF, where 74 eyes had ODD visible by EDI-OCT. Thus, in 12 of 75 eyes (16%) and in 11 of 38 patients (29%), the EDI-OCT managed to detect ODD, which the AF did not (Fig. 2). At no point did the AF manage to identify an ODD that the EDI-OCT could not detect. In the 75 included eyes, 23 distinctly solitary ODD were detected, and their volume and anatomic location were calculated. These were distributed in 14 eyes of 12 patients (2 cases of bilateral ODD). Of these 23 solitary ODD, 7 were detectable by AF. Both increase in ODD volume (P = 0.0388) and more superficial ODD location (P , 0.0001) increased the probability of AF detection of ODD, when performing a multivariate analysis adjusting for correlated eyes. None of the solitary ODD located more than 0.5 mm posterior to ILM (n = 13) were visible on AF (Fig. 3). DISCUSSION In our study, AF was inferior in the diagnosis of ODD compared to EDI-OCT. The ODD that were detectable using AF were located more anteriorly than the undetectable ones. Therefore, patients with buried ODD, which include the majority of children with ODD, are at a higher risk of being misdiagnosed and, therefore, at higher risk of undergoing unnecessary tests for the evaluation of optic disc edema. Accordingly, when buried ODD are suspected, EDI-OCT is the preferred diagnostic modality over AF and ultrasonography (20). Our results suggest that both volume and especially anatomic location of ODD have a major impact on the FIG. 3. Graph of AF-positive and AF-negative solitary ODD according to distance in millimeters from ILM to upper limit of the ODD and the volume of the ODD in mm3 (cubic millimeters). AF, autofluorescence; ILM, inner limiting membrane; ODD, optic disc drusen. Loft et al: J Neuro-Ophthalmol 2019; 39: 23-27 25 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution sensitivity of AF. However, other parameters such as degree of ODD calcification also could be important. ODD develop over time (24), and as proposed by Merchant et al (20), the ODD need a certain age and size in order to be detected by ultrasonography. Our study demonstrated that this appears to be the case for AF detection as well. The degree of development was not investigated in this study because, to the best of our knowledge, there is no method for determining this. Our data showed a clear separation of the AF-positive and AF-negative groups, in regard to vertical anatomic position (Fig. 3). The only overlap between the 2 groups is the 5 ODD with a location 0.4-0.5 mm posterior to ILM. Here, volume proved to be the determining factor, with the AF-positive ODD having a volume z10 times greater than those that were AF negative (Fig. 3). AF was not sensitive to any ODD located .0.5 mm posteriorly to ILM, although the detection of only ODD ,0.05 mm3 at this depth, might contribute to the finding. We found that RNFL thickness was significantly reduced in the AF-positive group. RNFL thickness decreases naturally over time (25,26), but there was no significant difference of age between the AF-positive and AF-negative groups. It is possible that axonal damage from the superficial ODD results in a reduction of the RNFL thickness or, more likely, that ODD simply are most easily detected by AF if the RNFL is already reduced. Regardless, our results showed that AF is associated with RNFL thinning. Besides a higher detection rate, EDI-OCT gives a clear picture of size, number, and location of the ODD. Because ODD diameter and volume had shown a negative correlation to RNFL thickness (15,27), this imaging modality may aid in our understanding of ODD pathophysiology. The main limitation of our study was that only solitary ODD were investigated, which reduced the amount of data that was analyzed. Furthermore, due to the inclusion criteria, all patients had ODD visible by either slit lamp or ultrasonography, thus excluding patients with ODD undetectable by these modalities, most likely smaller and deeply buried ODD. Segmentation of ODD was performed manually, but the method used to determine volume and anatomic location is a novel technique with an acceptable variability (27). This method allows for future longitudinal studies with great detail for changes in volume and location during the development of ODD. In conclusion, we found AF to be insufficient in the diagnosis of ODD in 16% of the eyes studied. Both volume and anatomic location of ODD had a significant impact on the sensitivity of the AF. The diagnosis of ODD based solely on AF is not recommended, and EDI-OCT should be the preferred imaging technique when buried ODD are suspected. 26 STATEMENT OF AUTHORSHIP Category 1: a. Conception and design: F. C. Loft, L. Malmqvist, A.-S. W. Lindberg, and S. Hamann; b. Acquisition of data: F. C. Loft, L. Malmqvist, A.-S. W. Lindberg, and S. Hamann; c. Analysis and interpretation of data: F. C. Loft, L. Malmqvist, A.-S. W. Lindberg, and S. Hamann. Category 2: a. Drafting the manuscript: F. C. Loft, L. Malmqvist, A.-S. W. Lindberg, and S. Hamann; b. Revising it for intellectual content: F. C. Loft, L. Malmqvist, A.-S. W. Lindberg, and S. Hamann. Category 3: a. Final approval of the completed manuscript: F. C. Loft, L. Malmqvist, A.-S. W. Lindberg, and S. Hamann. REFERENCES 1. Lam BL, Morais CG Jr, Pasol J. Drusen of the optic disc. Curr Neurol Neurosci Rep. 2008;8:404-408. 2. Friedman AH, Gartner S, Modi SS. Drusen of the optic disc. A retrospective study in cadaver eyes. Br J Ophthalmol. 1975;59:413-421. 3. Lorentzen SE. Drusen of the optic disk. A clinical and genetic study. Acta Ophthalmol (Copenh). 1966;44(suppl 90):1-180. 4. Pietruschka G, Priess G. Clinical importance and prognosis of drusen of the disc [in German]. Klin Monbl Augenheilkd. 1973;162:331-441. 5. Savino PJ, Glaser JS, Rosenberg MA. A clinical analysis of pseudopapilledema. II. Visual field defects. Arch Ophthalmol. 1979;97:71-75. 6. Mustonen E. Pseudopapilloedema with and without verified optic disc drusen. A clinical analysis II: visual fields. Acta Ophthalmol (Copenh). 1983;61:1057-1066. 7. Chambers JW, Walsh FB. Hyaline bodies in the optic discs: report of ten cases exemplifying importance in neurological diagnosis. Brain. 1951;74:95-108. 8. Malmqvist L, Lund-Andersen H, Hamann S. Long-term evolution of superficial optic disc drusen. Acta Ophthalmol. 2017;95:352-356. 9. Tan DK, Tow SL. Acute visual loss in a patient with optic disc drusen. Clin Ophthalmol. 2013;7:795-799. 10. Moreno M, Vazquez AM, Dominguez R, Rosas M. Severe and acute loss of visual field in a young patient with optic disc drusen. Arch Soc Esp Oftalmol. 2014;89:324-328. 11. Sarkies NJ, Sanders MD. Optic disc drusen and episodic visual loss. Br J Ophthalmol. 1987;71:537-539. 12. Erkkila H. Clinical appearance of optic disc drusen in childhood. Albrecht Von Graefes Arch Klin Exp Ophthalmol. 1975;193:1-18. 13. Fong CY, Williams C, Pople IK, Jardine PE. Optic disc drusen masquerading as papilloedema. Arch Dis Child. 2010;95:629. 14. Hu K, Davis A, O'Sullivan E. Distinguishing optic disc drusen from papilloedema. BMJ. 2008;337:a2360. 15. Sato T, Mrejen S, Spaide RF. Multimodal imaging of optic disc drusen. Am J Ophthalmol. 2013;156:275-282.e1. 16. Naoum S, Bouacha I, Drumare I, Marks C, Defoort-Delemmes S. Optic disc drusen in children: advantages of various imaging modalities [in French]. J Fr Ophtalmol. 2016;39:341-345. 17. Mrejen S, Spaide RF. Optical coherence tomography: imaging of the choroid and beyond. Surv Ophthalmol. 2013;58:387-429. cu B. Imaging methods in the diagnosis of optic disc 18. Tug drusen. Turk J Ophthalmol. 2016;46:232-236. 19. Malmqvist L, Wegener M, Sander BA, Hamann S. Peripapillary retinal nerve fiber layer thickness corresponds to drusen location and extent of visual field defects in superficial and buried optic disc drusen. J Neuroophthalmol. 2016;36:41-45. 20. Merchant KY, Su D, Park SC, Qayum S, Banik R, Liebmann JM, Ritch R. Enhanced depth imaging optical coherence tomography of optic nerve head drusen. Ophthalmology. 2013;120:1409-1414. Loft et al: J Neuro-Ophthalmol 2019; 39: 23-27 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited. Original Contribution 21. Yi K, Mujat M, Sun W, Burnes D, Latina MA, Lin DT, Deschler DG, Rubin PA, Park BH, de Boer JF, Chen TC. Imaging of optic nerve head drusen: improvements with spectral domain optical coherence tomography. J Glaucoma. 2009;18:373-378. 22. Malmqvist L, Li XQ, Eckmann CL, Skovgaard AM, Olsen EM, Larsen M, Munch IC, Hamann S. Optic disc drusen in children: the Copenhagen Child Cohort 2000 Eye Study. J Neuroophthalmol. 2018;38:xx-xx. 23. Malmqvist L, Bursztyn L, Costello F, Digre K, Fraser JA, Fraser C, Katz B, Lawlor M, Petzold A, Sibony P, Waner J, Wegener M, Wong S, Hamann S. The optic disc drusen studies consortium recommendations for diagnosis of optic disc drusen using optical coherence tomography. J Neuroophthalmol. 2018;38:xx-xx. Loft et al: J Neuro-Ophthalmol 2019; 39: 23-27 24. Frisen L. Evolution of drusen of the optic nerve head over 23 years. Acta Ophthalmol. 2008;86:111-112. 25. Won JY, Kim SE, Park YH. Effect of age and sex on retinal layer thickness and volume in normal eyes. Medicine (Baltimore). 2016;95:e5441. 26. Mansoori T, Balakrishna N. Effect of aging on retinal nerve fiber layer thickness in normal Asian Indian eyes: a longitudinal study. Ophthalmic Epidemiol. 2016;24:24-28. 27. Malmqvist L, Lindberg AW, Dahl VA, Jorgensen TM, Hamann S. Quantitatively measured anatomic location and volume of optic disc drusen: an enhanced depth imaging optical coherence tomography study. Invest Ophthalmol Vis Sci. 2017;58:2491- 2497. 27 Copyright © North American Neuro-Ophthalmology Society. Unauthorized reproduction of this article is prohibited.