Search

Purpose We aim to identify the features that allow the discrimination between healthy controls and diabetic patients with optical coherence tomography (OCT) data from the human retina. Methods Prior studies from our research group showed the association between the status of the blood‐retinal barrier and the consequent changes on the optical characteristics of the retina. Following the same rationale, we have demonstrated the possibility to discriminate healthy controls and diabetic patients by age group and healthy controls from diabetic subjects even though no visible changes could be identified, using a spectral domain OCT, Cirrus HD‐OCT (Carl Zeiss Meditec, Dublin, CA, USA), to gather the refractive index data in between the inner limiting membrane and the retinal pigment epithelium, and the support vector machine (SVM) for the automatic classification. In the work herewith presented, we applied a backward elimination process aiming to reduce the number of features used while simultaneously increase the classification performance. Results In this work we were able to reduce the number of features used in the SVM algorithm from the initial 46 features down to 10, while simultaneously increase the accuracy up to 97%. Moreover, the different feature set used for the different classifications suggest the presence of specific signatures within OCT data distribution. Conclusion These findings also suggest the possibility to use this noninvasive imaging modality to assess the neuronal condition and blood‐retinal barrier status. [ABSTRACT FROM AUTHOR]

Purpose To compute the vascular network of the human macula from spectral‐domain optical coherence tomography (OCT) to an extent similar to that of color fundus photography (CFP). Methods Macular cube protocol scans of 512x128x1024 and 200x200x1024 voxels of 20 eyes from 13 type 2 diabetic patients and 10 eyes from 10 healthy volunteers were collected from the Cirrus HD‐OCT (Carl Zeiss Meditec, Dublin, CA, USA) database. Additionally, CFPs and fluorescein angiograms (FAs) from the patients' eyes were gathered from the imaging database. Three distinct fundus references were computed from the OCT volumes after proper preprocessing. An additional OCT fundus image (OCTref) is computed as the principal component of these 3 OCT fundus references. The visible vascular network was manually segmented on CFP, FA and OCTref for comparison. Finally, a support vector machine (SVM) pattern classification algorithm was used to classify each pixel of the OCTref image into the vessel or non‐vessel classes from a set of 14 features computed from the OCT fundus references. Results Over 67% (67.8(7.2)%, average(SD)) of the vascular network manually segmented from the FA was manually segmented from the CFP, while this percentage raises to 69.2(8.9)% for OCT. In this way, the OCTref allows to compute an extended vascular network as compared to the CFP (102.8(14.4)%). When comparing the automatic versus the manual vascular segmentation, a specificity of 99.4(0.2)% and a sensitivity of 83.9(4.0)% were obtained. Overall, the accuracy of the automatic classification is of 98.0(0.4)%. Conclusion The proposed algorithm allows for the segmentation of the vascular network from OCT scans of the ocular fundus to a level similar to that of color fundus photography. [ABSTRACT FROM AUTHOR]

Purpose To compute the 3D retinal vascular network, from standard high‐definition optical coherence tomography (HD‐OCT) data. Methods Abnormal retinal vascular patterns were shown to be related to retinal and cardiac diseases. Studies have been performed using 2D fundus images. However, the obtained measurements may suffer from missing depth information for improved quantitative accuracy. In this work we resorted to Cirrus HD‐OCT data (Carl Zeiss Meditec, Inc., Dublin, CA, USA) to scan the human macula. Our approach to obtain vessel positioning (depth‐wise) takes advantage of the 2D automatic vessel segmentation of an OCT fundus image. To locate vessels in depth, we compare A‐scans where in the fundus image no vessels were detected to neighbor A‐scans containing vessels. Hence, we are able to identify the location of particular landmarks that characterize the presence of a vessel, such as, the hyper‐reflectivity and the shadowing effect (due to the light absorption by blood). Results Our algorithm is able to locate both above‐mentioned vessel markers (hyper‐reflective region and shadow). Preliminary data shows promising results. Both markers present good robustness and coherence, as demonstrated by the smoothness of vessels across different B‐scans and the possibility to discriminate between different depth paths of crossing vessels. Conclusion The findings suggest the possibility to compute the 3D vascular network, noninvasively, using a standard high‐definition OCT. Additionally, accounting for vessel depth will lead to improved measurements of retinal vascular network properties and therefore to possible better correlations between its shape, location, and disease status. [ABSTRACT FROM AUTHOR]

Purpose To reduce the speckle noise for improved structure segmentation on high‐definition spectral domain optical coherence tomography (OCT). Methods Eyes of 30 healthy volunteers and eyes of patients with age‐related macular degeneration (20), diabetic retinopathy (23), cystoid macular edema (2) and choroidal neo‐vascularization (13) underwent high‐definition spectral domain Cirrus OCT (Carl Zeiss Meditec, Dublin, CA, USA) using both the 200x200x1024 and the 512x128x1024 Macular Cube protocols. Results In the work herewith presented we have modified the traditional diffusion coefficient function, of the nonlinear complex diffusion type, to achieve an improved OCT data despeckling. The new filter was tuned in order to preserve edges, as in the transition from the vitreous to the ILM, to preserve the information within the retinal tissue and to be more effective while filtering the vitreous. A total of 88 volumetric scans gathered with the Cirrus OCT were filtered using the new nonlinear complex diffusion filter (NCDF) as well as with implementation current available filters, showing that the new NCDF filter allows for a better despeckling while preserving edges between retinal layers for eyes of healthy volunteers and patients with different pathologies. Conclusion These findings suggest that the speckle noise present in the Cirrus OCT can be successfully reduced while preserving important features using the newly developed NCDF filter. [ABSTRACT FROM AUTHOR]