Abstract
Purpose: :
To determine whether standard OCT contrast enhancement and shadow removal techniques can be extended to EDI-OCT images of the ONH in glaucoma patients.
Methods: :
We previously developed a compensation algorithm that, when applied to OCT images of the ONH, can remove shadow artefacts, enhance contrast, and improve deep tissue structure visibility (IOVS 2011;52(10):7738-48). This algorithm was tested with non-EDI OCT images of ONHs from subjects without ocular abnormalities. Here, its application is extended to both non-EDI and EDI-OCT images of ONHs from 5 glaucoma patients (3normal-tension and 2 high-tension). For each patient, non-EDI and EDI volumes of the ONH were acquired consecutively (Spectralis, Heidelberg Engineering). Each scan was composed of 145 B-scans of 384 A-scans averaged 9 times for speckle noise reduction. The interlayer contrast (ranging from 0 to 1, with 1 indicating a highly-detectable boundary) was computed across the anterior lamina cribrosa (LC) boundary using 5 B-scans. Similarly, the intra-layer contrast (ranging from 0 to 1, with 0 indicating complete blood vessel shadow removal) was computed in the LC using the same images.
Results: :
Without compensation, the inter-layer contrast (mean ± SD) was 0.33 ± 0.07 (non-EDI) and 0.35 ± 0.11 (EDI); EDI vs non-EDI, p > 0.05. With compensation, the inter-layer contrast values were significantly increased (indicating better anterior LC boundary visibility) in both cases (p < 0.001) to 0.90 ± 0.06 (non EDI) and 0.95 ± 0.01 (EDI); EDI vs non-EDI, p < 0.01. Without compensation, the intra-layer contrast (mean ± SD) was 0.93 ± 0.01 (non-EDI) and 0.94 ± 0.03 (EDI); EDI vs non-EDI, p > 0.05. With compensation, the intra-layer contrast values were significantly decreased (indicating successful shadow removal) in both cases (p < 0.001) to 0.55 ± 0.07 (non-EDI) and 0.46 ± 0.1 (EDI); EDI vs non-EDI, p < 0.05.
Conclusions: :
Our compensation algorithm can be applied to EDI images of the ONH and benefit from the previously demonstrated shadow removal and contrast enhancement while offering the advantages of deeper tissue imaging from EDI. Moreover, EDI, when combined with compensation, provides the best intra- and interlayer contrasts and thus the best LC visualization. Obtaining quantitative information and better contrast is a key step to allow automated segmentation of the ONH, quantification of ONH morphometry and biomechanics in vivo, and identification of potential risk indicators for glaucoma.
Keywords: image processing • lamina cribrosa • imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound)