6 Diagnostic Capability of Optical Coherence Tomography …
141
AL and thickness of the retinal layers are stronger in the outer regions, perhaps due
to the lower shear resistance of the thinner peripheral retina [65, 66].
The introduction of the latest SD-OCT devices led not only to a dramatic increase
in mapping speed and some increase in axial resolution, but the examination of the
choroid became also possible. In the past years, promising results have been obtained
by the manual segmentation and measurement of choroidal thickness on OCT images.
Li et al. and Sogawa et al. demonstrated strong negative correlation between AL and
choroidal thickness measured in the subfoveal area of young and healthy eyes (r
−0.624 and r −0.735, respectively) [67, 68]. Unfortunately, it is not possible to
obtain choroidal thickness from TD-OCT images due to the poor penetration and
thus low resolution beneath the RPE, which is one of the shortcomings of this study.
As the growth of the eyeball is stipulated to continue until the age of 20 years
[69], it is important to note that a longitudinal study spanning from adolescence to
early adulthood would be necessary to evaluate the effect of AL on the thickness of
intraretinal layers of the macula under and above the age of 20 years. Szigeti et al.
hypothesized that as the eyeball stops to grow the nuclear layers follow the shape of
an elongated globe and get thinner by lateral stretching, while the other layers are not
capable of this stretching [56]. It should, however, be taken into consideration with
such a study that longer AL decreases the magnification of fundus imaging, making
transverse dimensions appear smaller on the OCT scan, in inverse proportion to AL
[52]. Based on the Szigeti et al. study it is suggested that the effect of AL should be
taken into consideration when using OCT image segmentation techniques in future
clinical studies involving [56].
6.3 Capability of Optical Coherence Tomography Based
Quantitative Analysis for Various Eye Diseases
The diagnostic capabilities of OCT have renovated the ophthalmology practice and
provided demonstrable clinical benefits. Widespread clinical adoption of this technology has resulted in ophthalmic OCT images obtained each second by the medical
community, anywhere in the world. During the last decade, the upgrade of scanning
speed, resolution, and sensitivity has significantly increased the potential of OCT to
visualize more detailed retinal structures. However, the amount of data to be analyzed has also increased significantly. Automatic analysis algorithms or software are
therefore essential to the clinical applications because the huge amount of volumetric
data is no longer possible to be analyzed by visual identification or manual labeling.
As the retina is a multi-layered tissue, it is important to segment the various layers
or surfaces to fully explore the retinal structure and function. The development of
OCT segmentation software has progressed extensively during the last decade. It
was originally a proprietary software solution of individual manufacturers of OCT
but it has become a generic software solution of various research groups that have
developed algorithms to automatically detect retinal surfaces [18, 70–83]. A review
141
AL and thickness of the retinal layers are stronger in the outer regions, perhaps due
to the lower shear resistance of the thinner peripheral retina [65, 66].
The introduction of the latest SD-OCT devices led not only to a dramatic increase
in mapping speed and some increase in axial resolution, but the examination of the
choroid became also possible. In the past years, promising results have been obtained
by the manual segmentation and measurement of choroidal thickness on OCT images.
Li et al. and Sogawa et al. demonstrated strong negative correlation between AL and
choroidal thickness measured in the subfoveal area of young and healthy eyes (r
−0.624 and r −0.735, respectively) [67, 68]. Unfortunately, it is not possible to
obtain choroidal thickness from TD-OCT images due to the poor penetration and
thus low resolution beneath the RPE, which is one of the shortcomings of this study.
As the growth of the eyeball is stipulated to continue until the age of 20 years
[69], it is important to note that a longitudinal study spanning from adolescence to
early adulthood would be necessary to evaluate the effect of AL on the thickness of
intraretinal layers of the macula under and above the age of 20 years. Szigeti et al.
hypothesized that as the eyeball stops to grow the nuclear layers follow the shape of
an elongated globe and get thinner by lateral stretching, while the other layers are not
capable of this stretching [56]. It should, however, be taken into consideration with
such a study that longer AL decreases the magnification of fundus imaging, making
transverse dimensions appear smaller on the OCT scan, in inverse proportion to AL
[52]. Based on the Szigeti et al. study it is suggested that the effect of AL should be
taken into consideration when using OCT image segmentation techniques in future
clinical studies involving [56].
6.3 Capability of Optical Coherence Tomography Based
Quantitative Analysis for Various Eye Diseases
The diagnostic capabilities of OCT have renovated the ophthalmology practice and
provided demonstrable clinical benefits. Widespread clinical adoption of this technology has resulted in ophthalmic OCT images obtained each second by the medical
community, anywhere in the world. During the last decade, the upgrade of scanning
speed, resolution, and sensitivity has significantly increased the potential of OCT to
visualize more detailed retinal structures. However, the amount of data to be analyzed has also increased significantly. Automatic analysis algorithms or software are
therefore essential to the clinical applications because the huge amount of volumetric
data is no longer possible to be analyzed by visual identification or manual labeling.
As the retina is a multi-layered tissue, it is important to segment the various layers
or surfaces to fully explore the retinal structure and function. The development of
OCT segmentation software has progressed extensively during the last decade. It
was originally a proprietary software solution of individual manufacturers of OCT
but it has become a generic software solution of various research groups that have
developed algorithms to automatically detect retinal surfaces [18, 70–83]. A review