Automated Volumetric Segmentation of Retinal Fluid using Optical Coherence Tomography

Abstract

Purpose : To develop an automated volumetric segmentation method to quantify retinal fluid using optical coherence tomography (OCT).

Methods : Using a commercial spectral-domain 70kHz OCT (RTVue-XR, Optovue), 3×3×2mm volumetric macular scans were obtained in diabetic macular edema (DME) study participants. The scans were processed with the novel automated retinal fluid segmentation algorithm. First, the fuzzy C-means method detected the fluid regions with low reflectivity. Second, the level-set method identified the boundaries of fluid filled regions. Last, artefactual fluid regions were removed using information on morphological characteristics (smoothness and aspect ratio) and vascular shadowing (OCT angiography). All these processing operated on segmented retinal layers above Bruch’s membrane. Fluid regions above the photoreceptor inner /outer segment (IS/OS) junction was classified as intraretinal fluid (IRF) and those below as subretinal fluid (SRF).

Results : Sixteen DME participants were scanned in one eye. Ten of them had retinal fluid in the macular scans based on clinician grading. These were used to test the automated algorithm. The central macular thickness was 307±172 µm in these eyes. This algorithm detected and quantified IRF and SRF (Fig. 1 A and B) creating fluid thickness map (1C), 3-dimensional rendering of the fluid (1E), as well as integrated angiogram with fluid spaces (1D). Compared to manually corrected segmentation, the automated segmentation had an overlap error of 10.6±2.4% and relative difference of 10.2±2.4%, indicating high accuracy and reliability.

Conclusions : This novel algorithm can automatically detect and quantify retinal fluid space accurately, offering an alternative and possibly more meaningful way to evaluate diabetic macular edema than total retinal thickness and volume.

This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.

 

Figure1. An eye with diabetic macular edema showing segmentation boundaries (in red) shown on a cross-sectional structural OCT (A) and en face structural OCT (B), thickness map of total retinal fluid. (D), en face OCT angiogram overlaid with thickness map (E) 3-dimensional rendering of the fluid spaces. (the volumes of IRF and SRF are 0.23 mm³, 0.061mm³, respectively)