Optical coherence tomography

Optical coherence tomography (English optical coherence tomography OCT) is a test method in which light of low coherence length using an interferometer for measuring distance of scattering materials is used. The investigation object is scanned point by point. Their main area of ​​application is the medicine.

Advantages over competing methods are the relatively high penetration depth (1-3 mm) in scattering tissue and simultaneous high axial resolution ( 0.5-15 microns ) with high measurement speed ( 20-300 kvoxel / s). The corresponding acoustic method is the ultrasound.

• 4.1 sampling, line width and depth measurement
• 4.2 OCT measurement methods

Principle

OCT was designed analogous to B -scan ultrasonography: The OCT image is composed of laterally adjacent axial interferograms of the object depth.

Each axial interferograms based on white light or short-coherence interferometry. This method compares here the path of the reflections of an axially directed into the depth of the measurement beam with the object to that of a reference beam in an interferometer ( Michelson interferometer usually ). The interferograms (optical cross-correlation ) of both arms result in a linear pattern, the strength of the light-reflecting structures and their relative optical path length as an axial depth profile (also called "A- Scan", ie amplitude -mode scan) maps. In the one-dimensional grid method, the beam is then passed transversely in one or two directions, bringing a two-dimensional tomogram ( brightness -mode scan) or a three-dimensional volume ( c -mode scan) can absorb.

Unlike the conventional light microscopy, is decoupled from the resolution in the longitudinal OCT transversal. The transverse resolution is determined by the numerical aperture of the optical system used. The longitudinal spatial resolution in the depth of the material, however, depends on the spectral width of the light used.

Application

Applications are primarily in medicine: mainly used in ophthalmology as well as for the early diagnosis of cancer and for skin examination. Here reflections are measured at interfaces of materials with different refractive index and thus reconstructs a three-dimensional image. Such a reconstruction is called imaging.

The main application is currently at the examination of the fundus and the posterior segment of the eye, since competing techniques such as confocal microscope, the fine layer structure of about 250-300 microns thick retina due to the small pupil size and the large distance from the cornea to the retina only may reflect inadequate. Other methods, in turn, are not suitable due to their high exposure of the human eye or be severely impaired and the vitreous humor of the eye ( eg, high-resolution ultrasound). Right here contactless measuring is an invaluable advantage, since it risks of infection and psychological stress are largely avoided.

A new field of application of OCT is cardiovascular imaging. Intravascular optical coherence tomography ( OCT) is a new system based on infrared light technology, the arteries may present with a resolution of 10-20 microns. Several preclinical and clinical series showed that OCT enables secure identification of intramural and luminal morphologies, such as plaques, thrombi, dissections and information on Lumen and stent dimensions. Studies comparing IVUS and OCT showed that OCT provides additional morphological information which allows an improved plaque characterization.

Axial resolution and bandwidth

After initial experiments with light sources limited bandwidth (a few nanometers) were relatively broadband light sources with high spatial coherence available and used. In most cases, the systems based on superluminescent diodes with a few tens of nanometers bandwidth ( typically 30 nm, equivalent to more than 30 microns resolution. ). Only in the year 1997, the leap from that standard resolution is successfully ventured up to the " ultra high resolution " (> 100 nm, corresponding to less than 3 microns axial resolution), the tomograms allows almost comparable with histological sections.

The following formula (derived from the Fourier relationship between correlation width and spectral width, measured at full width at half maximum ) allows for a spectrum with Gaussian distribution to calculate the corresponding axial resolution:

The dispersion in human tissue and especially in the vitreous of the eye destroy the coherence of the two arms. Nifty but balancing the dispersion in both arms allows restitute the coherence. The precision of the ultra-high resolution OCT have led to a rethinking of Ophthalmology, because ophthalmologists can get information suddenly, they knew only from the textbook. This already allows small changes in the early stages to see what is difficult or even impossible with other methods.

Recent developments in nonlinear optics allow light sources for other wavelength regions and develop with greater bandwidth ( see picture).

Sampling

In the time domain, the interference signal is sampled on arbitrary intervals (English sampled ). However, the sampling rate has no effect on the resolution. The curve is therefore indeed more accurately measured, the smallest width of a single signal but not narrow. However, the sampling rate is less than twice the carrier frequency of the signal results in aliasing artifacts, according to the sampling theorem by Nyquist -Shannon.

Measurement methods

Due to the combination of the autocorrelation ( cross-correlation of a time signal with itself) with the frequency spectrum of a function of the Fourier transformation is considered in the optical domain, the analogous relationship between the optical spectrum and the interference signal. Therefore one speaks on the one hand from the signal in the time domain (English time domain, TD ) and the other part of signal in the frequency range (English frequency domain, FD). Simply put, this means that either the reference arm altered in length and continuously the intensity of the interference measures, without taking into consideration the spectrum (time domain ), or the interference of the individual spectral components detected (frequency domain). This process was made ​​possible by the availability of rapid, sensitive cameras and fast computers.

The advantage of the FD method is the simple and rapid simultaneous measurement. Here can be acquired simultaneously, the full information on the depth, without the need for a moving part. This increases the stability and speed. One can see the difference of the methods in the fact that TD -OCT has to take the overall performance of the reference and measurement arm in each measurement point, while but the interference component represents only a very small portion, whereby the noise of the overall signal compared to the useful component predominates. For receiving in the frequency domain (FD -OCT ) in each spectral channel, only the corresponding spectral power is measured as background. Thus any interference from the other spectral regions are lost. The necessary dynamics of the detector decreases with the total power per channel. Consequently, with the same sensitivity need ( = sensitivity for the measurement of very small reflectance) frequency-domain measurements only a fraction of the power. FD -OCT is usually even insensitive, but far more effective than TD -OCT. In principle, a simultaneous measurement in the time domain is possible, but it requires non-linear processes that operate only at relatively high light intensities. But this contradicts the highly sensitive measurement at measuring signal powers below the NW area.

However, the Fourier transform of the complex number in the working area, so both methods are equivalent only if the complex-valued functions are known. However, the final signal should reflect the temporal evolution of the reflectivity ( = absolute value of the intensity in time), which is why it comes in intensity recordings in the frequency domain and the absence of the complex information to ambiguity. The result is the " flipping of the image " in the conventional FD method. Since the imaginary part of a function but corresponds to a phase shift of 90 °, we shifted the reference arm can ( a quarter of the wavelength so ) make the complex-valued function and thus reconstruct the complete temporal function by additional measurements with a 90 ° in the term.

Sampling, line width and depth measurement

The sampling rate in the frequency domain is attached via the Fourier transform to the measurement depth. A higher sampling rate, or number of pixels of a detector within the same spectral range thus increasing the area in which several objects can be clearly distinguished from each other. But here again the same restriction applies as in the time domain, if the line width, so the lowest possible single spectral line is exceeded, there is no additional information at the oversampling more. ( The line width is either by the light source in the temporal encoding, or by the imaging geometry and scattering effects in the spectrometer during spatial encoding limited ). A larger line width than sampling density leads to the Fourier transform to a decrease in the intensity of the object against the edge of the local space. When subsampling in turn leads to the formation of multiple images even away from the zero order of the local area, ie the area in the center of the measuring arm and the reference arm are of equal length. When subsampling therefore objects are mirrored in out of range.

OCT measurement methods

In recent times, many different methods have been developed for signal detection - in the following, a systematic overview of all possible methods. The holographic methods are the spatial, transverse to the longitudinal counterpart, temporal frequency range of the optical transit time. There is therefore a Fourier relationship between the longitudinal duration and temporal frequency and between the transverse deflection and transverse spatial frequency. In principle, there are two sub-groups, in which on the one hand, the signal is temporally coded (time encoded), is thus recorded sequentially, or spatially coded ( Spatially encoded), so spatially split, but is recorded simultaneously. Often unsystematic names such as " Fourier domain OCT " or " Spectral OCT" are used, but mostly confusing ( confusion with spectroscopic OCT and inaccurate - the frequency is related to the time correlation, not the wavelength) or sometimes meaningless are (there are no Fourier domain ). They are still listed in the table below for guidance as alternative names.

The methods differ in their imaging quality and applicability, due to the use of different components. Specifically, the FD method have to waste the advantage of no light and have a much higher sensitivity. The goal is a high sensitivity, when used less as possible mobile components and thus a high speed, such as 3D teFD and holographic methods. On the other hand, the phase coherence is better for the potentially slower process. In addition, it depends on the orientation of the grid method and its grid density; so in layered biological tissues a high grid density is usually desired in the low cross-section, which is supplied only with difficulty from the quick, easy on -face methods.

Extensions

In addition to purely topographical information, additional data can be evaluated from the original signal. Thus, by measuring a plurality of successive tomograms at the same point, the local displacement are used for Doppler velocity measurements ( Doppler OCT). In addition, various material properties such as scattering, absorption, polarization change (English polarization- sensitive OCT) and dispersion can be determined and displayed. In addition, we try to highlight or just search selectively for specific molecules (english molecular contrast OCT) tissue.

Benefits

The great technological advantage of OCT is the decoupling of the low resolution of the transverse resolution. The purely based on optical reflection, and thus non-contact measurement allows the elimination of the technologies applied to microscopy, thin sections, making the process allows microscopic images in living tissue (in vivo).

Due to the high selectivity of the action principle very small signals can be detected (below NW ) and be assigned to a specific depth at low input powers. Thus, this method is also well suited to study light-sensitive tissue.

The use of OCT is limited by the wavelength-dependent penetration depth of electromagnetic radiation in the object under investigation as well as the bandwidth- dependent resolution. Sophisticated broadband lasers enable the development since 1996 of the CLOCK -OCT ( ultra - high resolution OCT), which has driven several micrometers to a fraction of a micrometer depth resolution. Subcellular structures in human cancer cells can be represented in this way.

Similar procedures

OCT is related to other profile interferometric imaging methods ( which, however, can only measure surfaces) such as holography and optical coherence radar, which is used for high-precision three-dimensional representation of surfaces in aircraft and automotive industries.

In addition, the digital holography overlaps with the region of the OCT Here, the physical frame is placed in the Fourier plane and enhances the interference pattern by means of mathematical recalculation of the entire volume. The advantage here is the independence of the focusing ( which is compensated numerically ), which causes only a decrease in intensity, but no blur. Numerical holography has the disadvantage that it is very sensitive to speckle that occur when scattering materials. In addition, the holography as the " full-field " OCT variants not advantageous for suppressing the confocal crosstalk benefit. There is overlap in the phase modulation method in which the primary phase is modulated in in interference. An alternative to OCT in medicine is the multi-photon tomography, which allows for higher resolutions, however, the signal is limited to a depth of several hundred microns.

View

OCT is a relatively new process ( initial development in the late 1980s ) and begins to establish itself currently in various fields. Also, are not yet exhausted all the technical possibilities. The low loading of the object, the high resolution and increasing speed make the method very attractive. New light sources, detectors and scanners will allow the future to carry out high-resolution three-dimensional microscopy on living tissue in video speed. The amount of data for such high quality recording would reach some Gigavoxel per second; Current high resolution OCT methods achieve up to 250 Megavoxel per second, with the level in the year 2000 was still below 100 Kilovoxel per second. Ultra high-speed OCT with lower sensitivity achieved by parallel detection already up to 60 Gigavoxel per second.