Sir,

As subretinal drusenoid deposits, also known as reticular pseudodrusen, carry an increased odds ratio for the development of choroidal neovascularization (2.6),1 the recognition of this finding is warranted in clinical evaluations of non-neovascular age-related macular degeneration (AMD).

Imaging subretinal drusenoid deposits requires optical coherence tomography (OCT) resolutions adequate to determine the retinal pigment epithelium (RPE) position relative to drusen and OCT algorithms that include subretinal structures. As the low reflectance of retinal tissue limits OCT resolution, subretinal drusenoid deposits are more easily detected with high-resolution B-scan protocols that use line averaging to enhance detail and reduce speckle noise. Current OCT drusen detection algorithms typically use lower resolution single-line raster scans to shorten scan acquisition time and maximize the area scanned for segmentation. Segmentation protocols typically identify drusen only beneath the RPE, missing subretinal structures, such as subretinal drusenoid deposits.

In a representative series of six eyes with subretinal drusenoid deposits, macular cube scans (500 × 128 and 200 × 200 protocols) obtained with the Zeiss Cirrus HD-OCT (Carl Zeiss Meditec, Dublin, CA, USA, v4.5.1.11) failed to show subretinal drusenoid deposits with adequate resolution. However, subretinal drusenoid deposits visualization was possible with the Cirrus HD-OCT in all six of our cases if manual segmentation was performed on C-scan (advanced visualization, RPE algorithm, 37 μm slab elevated above the RPE; Figure 1).

Figure 1
figure 1

SD-OCT manual segmentation and subretinal drusenoid deposits (SDD). Six eyes (a–f) with SDD well demonstrated by line scan protocol (top row). SDD are absent on drusen maps (second row), and cube-scanning protocols exclude subretinal structures seen on individual B-scans (third row). However, SDD are well demonstrated through manual segmentation (arrows to representative structures, bottom row).

Full size image

In summary, SD-OCT macular cube scans for non-exudative AMD have a limited ability to show important subretinal structures, such as subretinal drusenoid deposits, because of inherently lower B-scan resolution and lack of analysis internal to the RPE. However, manually segmented en face curved C-scans on the Cirrus HD-OCT can display subretinal drusenoid deposits without changes to the protocol. With the advent of pharmacologic therapy for non-neovascular AMD, an assessment for subretinal drusenoid deposits should be included in automated macular analyses.