Multiple imaging modalities for the detection of optic nerve head drusen: Is echography still mandatory?

Abstract

In this study, we included 28 patients with optic nerve head drusen (ONHD) seen at the University Eye Hospital of Cologne, Germany. The mean age was 40.9 ± 22.4 years. Sixteen patients were female, and twelve were male. The mean visual acuity was 0.18 ± 0.39 and 0.14 ± 0.23 LogMAR for the right and left eyes, respectively. A summary of baseline and morphological characteristics is given in Table 1. Three patients showed unilateral ONHD (patient no. 10, 26 and 28). The mean deviation in the perimetry was −6.74 ± 8.05 dB and −4.03 ± 5.04 dB for the right and the left eyes, respectively. A definite grid shown is used for the optical coherence tomography (OCT) quantification (Fig. 1). In this collective, optic head drusen are seen on the surface of the optic disc (superficial ONHD) in 45 eyes. This type of drusen is easily seen using funduscopy and fundus photography (Fig. 2). Drusen located on the surface of the optic disc also can be diagnosed using fundus autofluorescence, spectral domain OCT (SD-OCT), enhanced depth imaging OCT (EDI-OCT) and swept source OCT (SS-OCT). The fundus autofluorescence shows hyper-reflective materials, which are characteristics for drusen. In SD-OCT, hyporeflective materials can be easily seen in the depth of the optic nerve. Using EDI-OCT and SS-OCT, these drusen can be easily displayed. Drusen present in OCT as oval structures with hyporeflective filling and shadowing effect underneath the drusen. In all cases, hyper-reflective dots are seen around the drusen. In some cases, only some hyper-reflective dots without inner hyporeflective fillings are seen. Differentiating drusen to blood vessels can be challenging in some cases. Unlike drusen, vessels represent as oval structures with hyper- or hyporeflective fillings and shadowing effect underneath them. The outer linings of vessels are usually smooth without the aforementioned hyper-reflective dots (Fig. 3). In this study, the mean depth of ONHD was 205.45 ± 95.01 μm. Drusen and other retinal structure in the height of retinal nerve fibre layer (RNFL) through the retinal pigment epithelium (RPE) were best displayed using SD-OCT. Other structures underneath the RPE were best displayed using EDI-OCT. Swept source OCT (SS-OCT) displayed all structures from RNFL to the choroid sufficiently (Fig. 4). The outer boundaries of structures underneath optic nerve head drusen were however best shown using EDI-OCT. In eight eyes, optic disc drusen are buried in the depth of the optic disc. This type of drusen cannot easily be diagnosed using funduscopy. Also, in fundus autofluorescence, no hyper-reflective materials can be seen. Swept source optical coherence tomography, EDI-OCT and SS-OCT cannot penetrate into the depth to demonstrate the drusen either. In this case, only ultrasound is able to demonstrate the drusen in the depth of the optic nerve (Fig. 5). Optical imaging modalities showed similar sensitivity, that is 0.82, 0.77, 0.80 and 0.83 for SD-OCT, EDI-OCT, SS-OCT and FAF, respectively. The use of ultrasound showed the highest sensitivity of 1.0. The Youden's index for these imaging modalities was 0.82, 0.77, 0.80, 0.83 and 1.0 for SD-OCT, EDI-OCT, SS-OCT, FAF and ultrasound, respectively. The RNFL thickness around the optic nerve head was not significantly different comparing controls and ONHD patients (Table 2). Ganglion cell layer inner plexiform layer (GCLIPL) thickness was thicker in the superior and inferior part of the optic nerve head in ONHD patients (p = 0.002 and p = 0.019, respectively). Our findings encourage the use of different imaging modalities to diagnose and follow ONHD. We showed that non-buried ONHD are best displayed using optical imaging modalities, such as fundus photography, fundus autofluorescence and OCT. The non-buried ONHD can also be displayed using ultrasound as well; however, light-based imaging modalities offer mostly a more precise depiction of structures. This might be relevant, if one would like to follow and compare the structures in the future. We showed that SD-OCT depicts ONHD above the RPE best, while EDI-OCT the structure underneath the RPE level best. Swept source OCT (SS-OCT) displays all structures from RNFL to underneath the RPE level sufficiently. In conclusion, new imaging modalities based on OCT, that is EDI-OCT and SS-OCT, have their potential for diagnosing ONHD. They are however not suitable for imaging buried ONHD. We support the use of multiple imaging modalities in ONHD, as each of them has its pros and cons.