Optical coherence tomography (OCT) has advanced considerably since it was first applied to the eye.1–7
It is an extension of a technique called low-coherence interferometry, which was initially applied to ophthalmology for in vivo measurements of eye axial length.1
At the time of introduction, it was used to obtain in vivo optical cross sections of the anterior segment,6
as well as retinal diseases, such as macular detachment, macular hole, epiretinal membrane, macular edema, and idiopathic central serous chorioretinopathy.4
OCT cross sections were also used to evaluate the optic disc and retinal layers5,8
such as the retinal nerve fiber layer (RNFL).9
Scan patterns that enabled reproducible measurements were developed,10
and these eventually became incorporated into a commercial system, which had an axial resolution of ~10 μm.
The first clinical system was limited to a scanning speed of 400 axial scans (A-scans)/s because of a physical constraint: a moving reference mirror. OCT uses low coherence interferometry to obtain A-scan intensity profiles, and the process requires light to be split and sent to both a reference arm with a mirror and to the sample. Provided the path length to the reference mirror and tissue match to within the coherence length of the light source, when the reflected beams recombine, interference occurs. Intensity information, in the form of a reflectivity profile in depth, can be extracted from the interference profile. Changing the location of the reference mirror allows backscattered tissue intensity levels to be detected from different depths in the tissue sample. This approach is referred to as time-domain (TD)-OCT because time-encoded signals are obtained directly. Several improvements in OCT hardware have been introduced since the first commercial TD-OCT system became available. Better axial resolution11–13
and increased scanning speed14–23
are the two main advancements that have recently become incorporated into commercial systems.
The implementation of broadband light sources into OCT systems11
improved the axial resolution from ~10 μm to as high as 2 μm in tissue.24
Acquisition speed has improved considerably by detecting backscattering signals in the frequency domain,14–23
which means backscattered depth information at a given location can be collected without the movement of a reference mirror. Frequency information is acquired with a broad-bandwidth light source, charge-coupled device (CCD) camera, and a spectrometer14,17,18,20
or by sweeping a narrow-bandwidth source through a broad range of frequencies with a photodetector.16,21–23
The approach that incorporates a broadband light source is often referred to as spectral-domain (SD)-OCT, whereas the latter is termed swept-source (SS)-OCT. In both approaches, intensity profiles (A-scans) are obtained using a Fourier-transform of the detected frequencies, and this facilitates rapid A-scan collection. In addition to improved scanning speed, frequency-domain-OCT also offers the advantage of higher detection sensitivity—that is, it exhibits higher signal-to-noise, given a perfect reflector.23,25
A summary of OCT detection techniques can be seen in .
Comparison of TD-, SD-, and SS-OCT Devices
With these speed and sensitivity improvements, it is now feasible to collect volumetric (three-dimensional; 3D) scans of tissue, whereas in the past, the amount of time required to do this would have been prohibitive. Broadband volumetric retinal imaging with SD-OCT at speeds of up to 312,500 A-scans/s26
and SS-OCT at 249,000 A-scans/s27
have been demonstrated. To date, most clinical systems operate at an acquisition rate of ~27,000 A-scans/s and an axial resolution of 5 to 6 μm. A summary of the current commercially available retinal imaging systems is presented in .
Description of Commercially Available SD-OCT Systems