Purchase this article with an account.
Sowmya Srinivas, Muneeswar gupta Nittala, Amir H Hariri, Maximilian Pfau, Julie Gasperini, Srinivas R Sadda; Quantification of intraretinal hard exudates with Enface Optic Coherence Tomography. Invest. Ophthalmol. Vis. Sci. 2016;57(12):4232.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To evaluate the accuracy and reproducibility of intraretinal hard exudate (HE) quantification from Spectral domain optical coherence tomography (SD-OCT) images of eyes with diabetic retinopathy (DR).
Twelve eyes of 12 subjects with diabetic retinopathy followed over time (total of 30 visits across the cohort) were included in this IRB approved cross-sectional study. All subjects underwent SD-OCT imaging on a single instrument (Cirrus OCT; Carl Zeiss Meditec, Inc.) at the South Coast Retinal Center. HE's were quantified using two different methods. First, manual segmentation and quantification of HEs was performed on all OCT B scans (512 X 128 macular cube protocol) using 3D-OCTOR software, yielding both volume and area of HE’s. Second, segmentation of HEs was performed on en face OCT image generated using commercial instrument software (Cirrus OCT Enface analysis 7.5). The en face image was based on a slab between the outer border of the ganglion cell layer (GCL) and the external limiting membrane (ELM). Lipid HE’s appear as sharply demarcated bright foci in these en face images. To quantify these HE’s, the en face image was manually segmented with computer software (Photoshop version 5.0 extended; Adobe). The total area of HE’s obtained from exhaustive B-scan segmentation was compared the area from the en face segmentation, using Pearson correlation. All grading was also repeated by a second masked grader, and agreement between graders was assessed using intra class correlation (ICC).
The mean (standard deviation) HE area was 1.78 (1.37) mm2 with B-scan based segmentation and 0.72(0.82) mm2 with the En face approach. The mean absolute difference was 1.01 (0.64) mm2. There was, however, an excellent correlation in the total HE area between the B-scan segmentation and En face approaches (r= 0.95, P<0.0001). Inter grader reproducibility showed a high level of agreement with an ICC of 0.92 (95% CI 0.894-0.988) for the en face approach, and an ICC of 0.99 (95% CI 0.977-0.997) for B-scan segmentation.
Quantification of HEs in eyes with diabetic retinopathy using en face analysis can be performed reliably and correlate well with HE measurements obtained by exhaustive segmentation of all B-scans in dense volume OCTs. The en face approach may be a valuable strategy for quantification of HEs or other intraretinal bright objects (e.g. migrated pigment) on OCT scans.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.
This PDF is available to Subscribers Only