Depth-Resolved Visualization of Perifoveal Retinal Vasculature in Preterm Infants Using Handheld Optical Coherence Tomography Angiography

Abstract

Purpose: To establish methods to visualize depth-resolved perifoveal retinal vasculature in preterm infants using handheld optical coherence tomography angiography (OCT-A).

Methods: In this exploratory study, eyes of preterm infants were imaged using an investigational noncontact, handheld swept-source OCT-A device as part of the prospective BabySTEPS infant retinal imaging study. We selected high-quality OCT-A volumes at two developmental stages for analysis. Customized MATLAB scripts were used to segment retinal layers, test offset parameters, and generate depth-resolved OCT-A slabs. The superficial (SCP), intermediate (ICP), and deep (DCP) capillary plexuses were visualized and qualitatively assessed by three image graders.

Results: Six eyes from six preterm infants were included in this analysis. A three-layered perifoveal retinal vasculature was successfully visualized in all three eyes (three infants) in the 40 weeks postmenstrual age (PMA) group (one of three eyes with treated type 1 retinopathy of prematurity [ROP]). No obvious ICP or DCP was found in good-quality scans of the three eyes (three infants) in the 35 weeks PMA group (three of three eyes developed type 1 ROP).

Conclusions: Custom segmentation parameters are useful to visualize perifoveal retinal vasculature in preterm infants. At term age, a three-layered capillary structure is visible in most eyes, while prior to detectable flow within the ICP and DCP, the perifoveal vasculature may be better visualized in two layers.

Translational relevance: Development of segmentation parameters for depth-resolved OCT-A of perifoveal retinal vasculature in preterm infants facilitates the study of human retinal vascular development and vascular pathologies of ROP.

Conflict of interest statement

Disclosure: P.R. Patel, None; R. Imperio, None; C. Viehland, (P), Theia Imaging, LLC (F); D. Tran-Viet, None; S.J. Chiu, US10366492B2 (P); V. Tai, None; J.A. Izatt, (P), US10366492B2 (P), Leica Microsystems (F); C.A. Toth, (P), US10366492B2 (P), Alcon (F), EMMES Inc. (C), Theia Imaging, LLC (F); X. Chen, None

Figures

Figure 1.

Comparison of using IPL/INL junction and OPL/ONL junction as the reference boundary for segmentation of the ICP and DCP. The reference boundary using the IPL/INL junction (ICP: from 8.74 µm above to 17.48 µm below the IPL/INL junction; DCP: 17.48 µm below the IPL/INL junction to the OPL/ONL junction) was compared to the reference boundary using the OPL/ONL junction (ICP: from 8.74 µm above the IPL/INL junction to 34.96 µm above the OPL/ONL junction; DCP: from 34.96 µm above to at the OPL/ONL junction). In an infant eye without macular edema, no obvious difference in ICP and DCP vascular patterns was observed between the two reference boundaries (A). In contrast, in the infant eye with macular edema, the ICP and DCP slabs showed different vascular patterns in the area of macular edema (B, arrowheads). In the cross-sectional B-scan with flow overlay, the optimal segmentation lines using the IPL/INL junction as reference (C, magenta) and the OPL/ONL junction as reference (C, yellow) appeared close and nearly parallel to each other, while the reference lines were further apart in areas of macular edema (left side of D, magenta and yellow). By using the OPL/ONL junction as the reference boundary, most of the anomalous INL flow signals were assigned to the ICP.

Figure 2.

Segmentation of three-layered perifoveal retinal vasculature in three preterm infants imaged at around 40 weeks postmenstrual age. The larger arterioles and venules were almost exclusively located at the SCP; the parafoveal ICP exhibited a closed, loop-like vascular pattern; and the parafoveal DCP exhibited singular dendritic-like processes that extend toward the fovea. The area of signal loss in B was caused by blockage from vitreous hemorrhage. In the infant with macular edema, anomalous perifoveal endbulb-like flow signals could be seen in the ICP (C). A colored composite for the three capillary plexuses was generated and shown in the right-most column (SCP: yellow; ICP: cyan; and DCP: magenta).

Figure 3.

Cross-sectional B-scans with flow overlay in an infant eye at 40 weeks PMA showed that vertical location of the SCP varied based on its proximity to the fovea. Distal to the fovea, the SCP was present throughout the GCL (A and right side of B, arrowheads). In contrast, closer to fovea, the SCP was located closer to the GCL/IPL junction (B, arrows). Of note, the DCP terminated further from the fovea compared to the other capillary plexuses. Cyan lines represented the segmentation boundaries of the SCP, ICP, and DCP.

Figure 4.

Segmentation of two-layered perifoveal retinal vasculature in three preterm infants imaged at around 35 weeks postmenstrual age. The SVC in these three infant eyes exhibited vascular dilation and tortuosity, while the DVC was largely avascular. Cross-sectional B-scans with flow overlay showed that the flow signals in all three eyes were largely above the IPL/INL junction. Scant flow signal along the IPL/INL junction may represent early deep vascular penetration (A, arrows). In one infant, neovascularization and inner retinal vessel protrusion were present, but the inner limiting membrane was not breached (A, arrowhead). Cyan lines represented the SVC and DVC boundaries.