Aims: To describe a method for computerised alignment and averaging of sequences in optical coherence tomography (OCT) B-scans and to present selected clinical observations based on the resulting improvement in retinal imaging. Methods: A methodological study and retrospective investigation of selected cases. Five human subjects were included, one healthy subject, two patients with central serous chorioretinopathy, one patient with branch retinal vein occlusion, and one patient with cilioretinal artery pseudo-occlusion. Based on computerised alignment of sets of B-scans obtained at identical retinal locations, average OCT images were produced and displayed in false colour or grayscale. These enhanced tomograms were compared with other morphological and functional characteristics. Results: Improved retinal imaging enabled assignment of the OCT image to retinal anatomy particularly at the outer layer of the photoreceptors and the retinal pigment epithelium, both in the healthy eye and in pathology. Identification of both post-oedematous structural disorganisation as well as post-ischaemic attenuation of the inner retina was superior to standard OCT images. Conclusions: Averaging of multiple OCT B-scans enhances the quality of retinal imaging sufficiently to reveal new details of retinal pathophysiology. Using the technique on OCT3 scans enables visualisation of details comparable with the results obtained using ultra high resolution OCT. F irst generation clinical instrumentation for optical coherence tomography (OCT1 and OCT2) enables reliable demonstration of changes in overall retinal thickness, detection of fluid in and behind the neurosensory retina, and, to some extent, identification of the retinal nerve fibre and photoreceptor layers. The intervening layers of the neurosensory retina are only vaguely discernible and, in spite of the improved resolution of the OCT3, the relation between histology and OCT has not been fully elucidated.Further improvements are possible with ultra high resolution instruments based on titanium sapphire lasers; however, this instrument is yet not commercially available.

RNFLT as measured by Stratus OCT standard protocols was significantly affected by age and refractive status. The effect on global RNFLT of a difference in refractive error of 10 diopters corresponded to the effect of a difference in age of 60 years. Theoretically, the effect of refractive status may be explained by artefacts of RNFLT measurement circle placement. The results suggest that the diagnostic accuracy of Stratus OCT may be improved by considering refractive status in addition to age when RNFLT is measured. For this purpose spherical equivalent seems as effective as axial length.

ABSTRACT.Purpose: To compare the performance of two machine learning classifiers (MLCs), artificial neural networks (ANNs) and support vector machines (SVMs), with input based on retinal nerve fibre layer thickness (RNFLT) measurements by optical coherence tomography (OCT), on the diagnosis of glaucoma, and to assess the effects of different input parameters. Methods: We analysed Stratus OCT data from 90 healthy persons and 62 glaucoma patients. Performance of MLCs was compared using conventional OCT RNFLT parameters plus novel parameters such as minimum RNFLT values, 10th and 90th percentiles of measured RNFLT, and transformations of A-scan measurements. For each input parameter and MLC, the area under the receiver operating characteristic curve (AROC) was calculated. Results: There were no statistically significant differences between ANNs and SVMs. The best AROCs for both ANN (0.982, 95%CI: 0.966-0.999) and SVM (0.989, 95% CI: 0.979-1.0) were based on input of transformed A-scan measurements. Our SVM trained on this input performed better than ANNs or SVMs trained on any of the single RNFLT parameters (p £ 0.038). The performance of ANNs and SVMs trained on minimum thickness values and the 10th and 90th percentiles were at least as good as ANNs and SVMs with input based on the conventional RNFLT parameters. Conclusion: No differences between ANN and SVM were observed in this study. Both MLCs performed very well, with similar diagnostic performance. Input parameters have a larger impact on diagnostic performance than the type of machine classifier. Our results suggest that parameters based on transformed A-scan thickness measurements of the RNFL processed by machine classifiers can improve OCT-based glaucoma diagnosis.

In healthy young adults with normal blood pressure and blood glucose, variations in retinal blood vessel diameters and blood pressure were predominantly attributable to genetic effects. A genetic influence may have a role in individual susceptibility to hypertension and other vascular diseases. The results suggest that retinal vessel diameters and the possible associated variations in risk of vascular disease are primarily genetic characteristics.