Patient selection
This retrospective observational case series included treatment- naive patients with type 2 MNV, diagnosed in our tertiary retina center (Medical Retina Unit, Department of Ophthalmology; Rudolf Foundation Hospital Vienna; Karl Landsteiner Institute for Retinal Research and Imaging) between 2008 and December 2017. The study adhered to the tenets of the Declaration of Helsinki. Ethics committee declared no requirement of approval due to the study’s retrospective data analysis. Consenting patients were initially diagnosed and classified into anatomic subtypes of neovascular lesions based on fundus examination by slit-lamp biomicroscopy (Haag- Streit AG, Bern, Switzerland) and multimodal imaging. Exclusion criteria were neovascularizations of different entities (p.e. idiopathic, myopic, posttraumatic, uveitic, dystrophic or secondary to pachychoroid diseases) as well as a subfoveal atrophy or a fibrosis at the time of recruitment. The fellow eye was examined for additional information on the drusen distribution and also excluded in case of a neovascularization, subfoveal atrophy or fibrosis.
Image Interpretation
The data records of patients with type 2 MNV were independently analyzed by two medical retina specialists based on near-infrared fundus reflectance (IR) images, blue-peak fundus autofluorescence (BAF) images, spectral-domain (SD) optical coherence tomography (OCT) scans, high resolution fluorescein angiography (FA) and indocyanine green angiography (ICGA) by a confocal scanning laser ophthalmoscope (Spectralis HRA+OCT; Heidelberg engineering, Heidelberg, Germany). Type 2 MNV was diagnosed in case of an early leakage and the absence of speckled hyperfluoresence in FA. ICGA differentiated the neovascular lesion as a well-defined network with circumscribed halo in early phases from other types like polypoidal lesions or hot spots mandatory for type 3 MNV. Moreover, ICGA was helpful to distinguish between neovascularization related to AMD vs. a secondary neovascularization masquerading AMD. Multiple SD-OCT scans through the lesion area were conducted and investigated to rule out a sub-RPE neovascularization and hence a mixed type component.
Characteristics
Greyscale variations in IR images on high resolution angiography and OCT scans indicated the existence of soft drusen (Fig.1a). SDD commonly appeared as hyporeflective dots (Fig.1e) or ribbons but could also occur as hyperreflective spots surrounding the perifoveal region, termed midperipheral SDD [15].
In BAF images, soft, hard and cuticular drusen appeared hypoautofluorescent in the center with an annulus of increased fundus autofluorescence [16, 17]. A reticular hypoautofluorescent pattern was typical for SDD [18]. Different staining levels in FA depend on the binding of dye to polar lipids with a higher proportion of fibronectin in contrast to neutral lipids with little adherence [19]. Soft drusen showed mild hyperfluorescence on FA (Fig.1b, 1f) but hypofluorescence on ICGA (Fig.1c), while hard drusen were often hyperfluorescent in both dye applications. SDD were hypofluorescent or not visible on FA and ICGA (Fig.1b, 1c, 1f) [6, 7, 20–22]. The hyperfluorescent “starry sky” appearance on FA was considered as typical for cuticular drusen (Fig.2a, 2c) [23]. OCT images were investigated to locate the layer and measure the size of retinal debris. Soft drusen (Fig.1d), hard drusen or cuticular drusen (Fig.2b, 2d) were identified as deposits below or within the RPE, whereas SDD were described as subretinal accumulation of material overlying the RPE zone, forming sharp, broad or rounded elevations [6]. A horizontal diameter of less than 63 mm was defined as threshold to distinguish small from larger drusen. The Spectralis software allowed for a correlation of the topography in SD-OCT scans and 30° en-face images. Drusen phenotypes, SDD, fibrosis or atrophy were identified and listed separately if co-existence occurred. Another senior clinical advisor without affiliation to the study concept was consulted in the case of grading disagreement.
SFCT was measured manually within SD-OCT scans as the greatest vertical distance between the RPE or the Bruch’s membrane and the sclerochoroidal interface.
Statistics
A univariate logistic regression model was performed for each factor (gender, age, SFCT) potentially influencing the occurrence of drusen or SDD on the affected eye. The same analyses were performed to investigate the fellow eye. All calculations were executed using R, release 3.3.3. Diagrams were developed by Microsoft Excel (Microsoft Corporation, Redmond, WA) and figures composed by Photoshop CC 14.0 (Adobe Systems Incorporated, San Jose, CA).