Noninvasive, simultaneous probing of 3-D cellular-resolution tissue morphology and depth-resolved function could significantly improve early medical diagnosis.
By Wolfgang Drexler
Optical coherence tomography (OCT) is an emerging diagnostic-imaging modality that enables in vivo cross-sectional tomographic visualization of internal microstructures in biological systems (see “How it works,” p. 29). Since its invention in the early 1990s, the motivating factor behind the development of OCT tools and techniques has been to enable noninvasive optical biopsies—specifically, in situ imaging of tissue microstructures with a resolution approaching that of histology but without the need for tissue excision and post-processing.1-7
To achieve this challenging goal, recent trends in OCT technology development have focused on improving image resolution, data-acquisition speed, tissue penetration, and contrast enhancement. New state-of-the-art delivery systems have facilitated the application of OCT in a variety of fields outside of ophthalmology (notably gastroenterology and cardiology), enhanced OCT tissue penetration, and provided access to internal organs for in vivo imaging. Additional extensions of OCT enable noninvasive depth-resolved functional imaging, providing polarization-sensitive spectroscopic blood-flow or physiologic tissue information. These advances should not only improve image contrast but enable the differentiation of pathologies via localized metabolic properties or functional (physiologic) state.
As a consequence, there has been a tremendous increase in publications (90 in 1998 vs. 900 in 2006), patents (9 granted patents in 1998 vs. more than 90 in 2006), and companies involved in the field of OCT (3 in 1998 vs. more than 20 in 2006). In 1998, OCT publications were cited approximately 700 times; in 2006, more than 10,000 citations referred to OCT research. In addition, sources have estimated that the market for OCT equipment will grow at a compound annual rate of more than 30% over the next four years, reaching €200 million (US$295 million) by 2011.8
The development of ultrabroad bandwidth light sources—particularly solid-state lasers, ASE light sources, and superluminescent diodes (SLDs)—helped to establish ultrahigh-resolution OCT as a viable clinical tool in the late 1990s.9-15 In fact, these technologies were a critical first step in enabling OCT to function as a type of optical biopsy, providing information on pathology in situ and in real time, with resolutions approaching that of excisional biopsy and histopathology (see Fig. 1). Especially in ophthalmology, where retinal biopsies are not possible, OCT can provide unprecedented morphologic and functional information for improved diagnosis and monitoring of the diseases that are the leading causes of blindness worldwide, as well as the therapies developed to combat these diseases (see “Retinal diseases push spectral-domain OCT into the short-wave infrared,” p. 28).16-21
FIGURE 1. Foveal portion of corresponding semithin histological section (A) and in vitro OCT image (B) of a perfusion fixed monkey retina(gc ax: ganglion cell axon layer, gc: ganglion cells, ipl: inner plexiform layer, inl: inner nuclear layer, Hf: fibers of Henle, onl: Outer nuclear layer, cis/cos: cone inner/outer segments, pe: Pigment epithelial layer, ch cap: choriocapillaris, ch: choroid, asterisk: darker faults in foveal floor indicative of foveal strain, d: epi-retinal debris). In vivo optical biopsy using ultrahigh resolution ophthalmic OCT of a patient with macular hole (D) compared with histology (C) of a different post mortem eye with comparable stage of macular hole. (Histology (C) provided by R. Brancato, Italy). Endoscopic ultra-high-resolution OCT tomogram (E, F) versus stained histological cross-section (C) of in vivo mouse colon with distally integrated beamsplitter enables visualization of colonic mucosa (CM), muscular mucosa (MM), submucosa (SM), muscularis externa (ME), and serosa (S) layers. Contrast enhanced portion, using local histogram equalization shows a surface layer of apical crypt cells (AC) as well as vertical structures in the mucosa that may correspond to crypt boundaries (C).
Again, using publications as a benchmark, it is clear that the most well-established clinical applications for OCT are in ophthalmology; about 50% of all OCT publications to date have been published in ophthalmic journals. Another 25% have been published in optical journals, reflecting the numerous technical advances that have been accomplished. The clinical impact of OCT, especially in ophthalmology, is also demonstrated by the fact that a fourth generation of commercial instruments was recently introduced and that more than a half-dozen companies are commercializing this technology for diagnostic applications in ophthalmology.
The concept of “the more you see, the better the diagnosis”—enabled by ultrahigh-resolution OCT—has also been successfully demonstrated in gastroenterology, cardiology, and many other fields.22-25 However, there are many technical challenges to achieving consistent ultrahigh-resolution OCT in turbid media. Until recently, resolution improvements were only possible with limited data-acquisition speeds because the dominant approach was time-domain OCT. With this technique, there is a fundamental tradeoff between sensitivity and data rate, so detection sensitivity is lower for faster scanning speeds and/or broader optical bandwidths used in ultrahigh resolution.
Recent advances in detection techniques are enabling dramatic increases in OCT imaging speeds. In particular, two alternatives to time-domain OCT that are gaining favor are Fourier-domain and frequency-domain OCT.26-29 These techniques are somewhat analogous to Fourier-transform spectroscopy and have significant sensitivity and speed advantages because they either measure the entire optical echo signal or detect the A-scan simultaneously rather than sequentially. Using Fourier-domain detection to obtain the interferometric information needed for the depth-resolved reflectivity profile can be accomplished by using either a spectrometer and high-speed CCD camera (also referred to as “spectral/Fourier domain OCT” in the literature) or by using a frequency-swept light source and detecting interferometric information with a photodetector (also referred to as “swept-source OCT or “optical-frequency-domain imaging” in the literature).
The acquisition speed for both approaches is governed by the read-out rate of the CCD camera or the sweep speed of the light source, respectively. Because both approaches detect all of the echoes of light simultaneously, they offer a significant efficiency/sensitivity advantage and a dramatic increase in line rate (A-scan rate) in comparison to time-domain OCT. For OCT imaging, in theory approximately 400 times faster imaging speed should be possible without sacrificing sensitivity. However, in practice, a factor of 50 to 100 faster imaging speed is achieved because of limitations in the CCD cameras. It is worth noting that, although the basic principle of Fourier-domain detection has been known since 1995, limitations in CCD technology and lack of recognition of the performance advantages delayed the use of this technique for nearly a decade. Three groups, working independently, simultaneously described the sensitivity advantage of this technique earlier in this decade.30-32 The first demonstration of retinal imaging using spectral/Fourier-domain OCT was performed in 2002, and high-speed imaging using line-scan CCDs was demonstrated in 2003.33 Spectral/Fourier-domain OCT was rapidly adopted and demonstrated to offer high-speed, ultrahigh-resolution, functional Doppler imaging.34-40
FIGURE 2. Three-dimensional OCT enables unprecedented volumetric representation of pathologic retina at different angled views (A) including from below (B). Virtual biopsy/surgery using OCT (C-F) allows the user to excise and remove any given layer or part of the retinal volume in order to visualize intraretinal morphology. In vivo ultrahigh-resolution OCT of human skin (G): rendering and virtual C-mode scans (a-e) of a pigmented mole (left). The epidermis is much thinner than at the fingertip. Penetration is lower, but highly absorbing structures of about 10 to 30 µm at the basal layer (b, c) and a wide vascular mesh work can be appreciated in the deeper dermal layers (d, e). Toward cellular-resolution retinal imaging: three-dimensional, adaptive optics, ultrahigh-resolution OCT of a small volume of the intraretinal tissue (H, I). The width of single cones (cf. arrows in I) estimated from the images is approximately 4 µm, in agreement with literature.
These improvements in OCT resolution and data-acquisition speed have enabled a generation leap in OCT technology and performance. Improving OCT resolution to the submicrometer level raises the possibility that subcellular resolution imaging is possible—with the proviso that the optical contrast of cellular features is sufficient at the imaging wavelength region. Significantly faster OCT data-acquisition speed of up to 400,000 measurements per second41-47 facilitates isotropically (equidistantly) sampled, nearly motion-artifact-free, volumetric imaging (see Fig. 2A-G), screening of larger (square-centimeter area) organs,48-49 and high-definition two-dimensional imaging. The latter especially improves the visualization of stratified organs such as the retina or colon by using high-transverse oversampling. Combining isotropic ultrahigh resolution and high data-acquisition speeds enables in vivo cellular-resolution imaging (see Fig. 2H-I).
Extending OCT: Optophysiology
In addition to techniques for improving OCT image resolution and data-acquisition speeds, many functional extensions of OCT technologies have been investigated. Of these, Doppler OCT (measuring blood-flow velocity) and polarization-sensitive OCT (imaging depth-resolved tissue birefringence) have been the most developed and successful.50-55 These techniques measure tissue parameters that provide information on the functional state of the tissue. They also help enhance image contrast.
In the stricter definition of functional measurements, electrophysiology is the “gold standard” for detecting physiologic/functional tissue, mainly retinal changes. This method requires contact, is time-intensive, does not enable depth resolution, and has limited transverse resolution. Noncontact, optical measurement of retinal response to visual stimulation with 10 µm spatial resolution has recently been achieved using functional ultrahigh-resolution OCT. This method takes advantage of the fact that physiological changes in dark-adapted retinas caused by light stimulation can result in local variations of the tissue reflectivity. This functional extension of OCT can be considered an optical analogue to electrophysiology and is therefore called “optophysiology,” capable of depth-resolved detection of retina physiology.56
To determine the sensitivity of optophysiology for the detection of changes in retinal reflectivity triggered by light stimulation, a dark-adapted living in vitro rabbit retina was exposed to a single flash of white light and optophysiology data was acquired synchronously with electroretinogram recordings. Throughout the functional experiments the isolated retinas were stimulated with single, 200 ms white-light flashes. A morphological B-scan was first taken from the measurement location (see Fig. 3A). Multiple ultrahigh-resolution OCT depth-reflectivity profiles (A-scans) were then acquired at one transverse location in the retina synchronously with ERG recordings (see Fig. 3B and 3C). The ultrahigh-resolution OCT A-scans were combined to form two-dimensional raw-data M-tomograms, presenting the retina reflectivity profile as a function of time. The optical data was processed using a cross-correlation algorithm to account for any movement of the retina caused by the solution flow, calculation of the optical background (average over the prestimulation A-scans of each M-tomogram), and generation of differential M-tomograms of the raw-data M-tomograms (see Fig. 3D).
FIGURE 3. Optophysiology using ultrahigh-resolution OCT enables the optical probing of depth-resolved retinal physiology. A morphological B-scan from the measurement location (A); multiple depth-reflectivity profiles (A-scans; B) are acquired at one transverse location in the retina synchronously with ERG recordings (C). Generation of differential M-tomograms of the raw-data M-tomograms (D); different optophysiological signals can be extracted from various retinal layers, enabling depth-resolved optical back-scattering changes that are due to physiological processes induced by the optical stimulus (E).
Optophysiological signals could be extracted from various retinal layers so that depth-resolved optical back-scattering changes that resulted from physiological processes induced by the optical stimulus could be detected (see Fig. 3E). The exact origin of the detected optophysiologic signals is unclear but might be related to the dipole reorientation (and therefore refractive-index changes) at the photoreceptor membrane. Alternatively, they could arise from light-induced isomerization of Rhodopsin in the outer photoreceptor segment or metabolic changes in the mitochondria of the inner photoreceptor segments. Noninvasive in vivo functional optical imaging of the intact rat retina was recently demonstrated in dark-adapted Long-Evans rat retinas.57
Due to these and other significant technology improvements in recent years, there might be some kind of saturation in future efforts to further improve the key technological OCT parameters: resolution, scanning/data-acquisition speed, sensitivity, and penetration. Therefore, the success of OCT in terms of visualizing morphology might lie in its ability to localize tissue function. This approach would put OCT in a unique position to perform noninvasive visualization of microstructural tissue morphology. In addition, because of OCT’s inherent depth-resolving nature, it would also enable unprecedented functional tissue information—ideally performed with a single measurement and dedicated post-processing.
Three-dimensional ultrahigh-isotropic-resolution OCT, in combination with ultrafast scanning/data acquisition, enabled a quantum leap in OCT performance in recent years. OCT can now be considered an optical analogue to CT or MRI, but with microscopic resolution. In addition to functional extensions of OCT, this technique appears to be on the verge of revolutionizing medical diagnosis in the very near future.
Editor’s Note: A detailed analysis of the emerging market for OCT equipment, components, techniques, and applications, Optical Coherence Tomography, Technology, Markets, and Applications, 2008-2012, will be published by BioOptics World in February 2008. For more information, go to www.bioopticsworld.com.
- A.F. Fercher and E. Roth, Proc SPIE 1986, 658, p. 48.
- J.G. Fujimoto et al., Optics Lett. 1986, 11, p. 150.
- R.C. Youngquist et al., Optics Lett. 1987, 12, p. 158.
- A.F. Fercher et a., Optics Lett., 1988. 13, p. 186.
- D. Huang et al., Science 1991, 254(5035) p. 1178.
- J.G. Fujimoto et al., Nature Medicine 1995, 1(9) p. 970.
- A.F. Fercher, J. Biomedical Optics 1996, 1, p. 157.
- Optometry Today 2007, 47(13) p. 22.
- A.M. Kowalevicz et al., Optics Lett. 2002, 27, p. 2037.
- A. Unterhuber et al., Optics Lett., 2003, 28(11) p. 905.
- T.H. Ko et al., Optics Exp. 2004, 12(10) p. 2112.
- D.S. Adler et al., Quantum Elect. 2004, 34(10) p. 915.
- W. Drexler et al., Optics Lett. 1999, 24(17) p. 1221.
- W. Drexler et al., Nature Medicine 2001, 7(4) p. 502.
- W. Drexler et al., Archives of Ophthal.2003, 121(5) p. 695.
- W. Drexler, J. Biomedical Optics 2004, 9(1) p. 47.
- E. Ergun et al., Investigative Ophthal. & Visual Sci., 2005, 46(1) p. 310.
- M.G. Wirtitsch et al., Am. J. Ophthalmology 2005, 140(6) p. 976.
- C. Scholda et al., Retina, 2006, 26(9) p. 1034.
- T.H. Ko et al., Ophthalmology 2005, 112(11) p. 1922.
- T.H. Ko et al., Arch. Ophthalmol. 2006, 124(6) p. 827.
- A.R. Tumlinson et al., Optics Exp., 2006, 14(5) p. 1878.
- A.R. Tumlinson et al., J. Biomedial Optics 2006, 11, p. 064003.
- P.R. Herz et al., Optics Exp. 2004, 12(15) p. 3532.
- Y. Chen et al., Endoscopy 2007, 39(7) p. 599.
- A.F. Fercher et al., Optics Communications 1995, 117 p. 43.
- B. Golubovic et al., Opt Lett. 1997 22(22), p. 1704.
- F. Lexer et al., Applied Optics 1997, 36(25) p. 6548.
- G. Häusler et al., J. Biomedical Optics 1998, 3, p. 21.
- R. Leitgeb, C.K. Hitzenberger, and A.F. Fercher, Optics Exp. 2003, 11(8) p. 889.
- M.A. Choma et al., Optics Exp. 2003, 11(18) p. 2183.
- J.F. de Boer et al., Optics Lett. 2003, 28(21) p. 2067.
- M. Wojtkowski et al., J. Biomedical Optics 2002, 7(3) p. 457.
- S.H. Yun et at., Optics Exp. 2003, 11(22) p. 2953.
- N. Nassif et al., Optics Lett. 2004, 29(5) p. 480.
- R.A. Leitgeb et al., Optics Exp. 2004, 12(10) p. 2156.
- M. Wojtkowski et al., Optics Exp. 2004, 12(11) p. 2404.
- B. Cense et al., Optics Exp. 2004, 12(11) p. 2435.
- R.A. Leitgeb et al., Optics Exp. 2003, 11(23) . 3116.
- B.R. White et al., Optics Exp. 2003, 11(25) p. 3490.
- S.H. Yun et al., Optics Lett. 2003, 28(20) p. 1981.
- W.Y. Oh et al., Optics Lett. 2005, 30(23) p. 3159.
- R. Huber et al., Optics Exp. 2005, 13(9) p. 3513.
- R. Huber et al., Optics Exp. 2006, 14(8) p. 3225.
- R. Huber et al., Optics Lett. 2006, 31(20) p. 2975.
- D.C. Adler et al., Optics Lett. 2007, 32(6) p. 626.
- R. Huber et al., Optics Lett. 2007, 32(14).
- S.H. Yun et al., Nature Medicine 2006, 12(12) p. 1429.
- B.J. Vakoc et al., Gastrointest. Endosc. 2007, 65(6) p. 898.
- X.J. Wang et al,, Optics Lett. 1995, 20(11) p. 1337.
- Z.P. Chen et al., Optics Lett. 1997, 22(14) p. 1119.
- J.A. Izatt et al., Optics Lett. 1997, 22(18) p. 1439.
- M.R. Hee et al., J. Opt Soc Am B 1992, 9(6) p. 903.
- J.F. deBoer et al., Optics Lett. 1997, 22(12) p. 934.
- J.F. de Boer et al., Optics Lett. 1999. 24(5) p. 300.
- K. Bizheva et al., Proc. Nat’l. Academy of Sciences USA 2006, 103(13) p. 5066.
- V.J. Srinivasan et al., Optics Lett. 2006, 31(15) p. 2308.
Wolfgang Drexler is full professor of biomedical imaging and head of the Biomedical Imaging Group at the School of Optometry and Vision Sciences, Cardiff University, Wales, U.K. He can be reached at DrexlerW@cardiff.ac.uk.
Retinal disease pushes spectral-domain OCT into the short-wave infrared
Optical-coherence tomography (OCT) is used to image hidden structures in normally opaque materials by detecting the backscatter of light from the structures. The signal is very weak and buried in the general light-scattering reflections from many points in the target material. By combining in an interferometer the sample signal with the signal from a reference arm, the general reflections are cancelled out and the remaining back-reflection signal is improved by a factor of 10. This permits recovery of the image using fast Fourier transforms of the interference signal as a function of wavelength.
However, the depth of the recovered signals is limited by many factors, not the least of which is the distance that light can travel in a material or media before the backscattered light is lost. The light interacts with molecules in various media in different ways, depending on the ratio of the size of the molecule and the wavelength of the light, or whether the media is air, paint, biological tissue, or even bones. When the mean size of molecules is larger than the wavelength, the scattering efficiency is large and fairly constant as a function of wavelength, according to the Mie solution for scattering. Using longer wavelengths in biomedical applications generally increases the depth of the tissue that can be probed.
Water absorption adds another limitation for imaging in tissue, as the light absorption increases strongly with wavelength, greatly attenuating signals the longer the path length. Meanwhile, melanin absorbs shorter wavelengths in skin and even in the retina where it is present to reduce stray light. The combination of scattering, melanin and water absorption with the available light sources that have the needed power, bandwidth and center wavelength to serve a particular application, determines the wavelengths used with OCT.
For imaging the retina, 840 nm is typically used because it easily passes through the aqueous humor as water absorption is relatively low and also through the melanin layer. While 840 nm works well for the center of the cornea, for the thicker parts at the edges, the near-infrared wavelength experiences too much scattering to form a useful image. This prevents the OCT system’s use for diagnosing a type of glaucoma by imaging the angle the cornea makes with the iris of the eye.
As a result, the short-wave infrared wavelength with reduced water absorbance at 1040 nm is now being used for retinal OCT. It is also being implemented for mapping the blood flow below the retina and for imaging the retina in patients with cataracts clouding the lens of the eye. The SWIR wavelength offers an advantage of greater tissue penetration depth due to lower scattering, while still being able to traverse the aqueous humor without too much absorbance (see figure).
While 800 nm spectral-domain OCT works well for imaging the center of the cornea, for the thicker parts at the edges the near-infrared wavelengths experiences too much scattering to form a useful image. Using a longer wavelength (1050 nm) enables greater tissue penetration depth due to lower scattering, while still being able to traverse the aqueous humor without too much absorbance. (Courtesy of Wolfgang Drexler/Cardiff University, Wales, U.K.)
For imaging several millimeters into organ tissues with endoscopy, examining plaque inside arterial walls, or imaging the first millimeters of skin, 1300 nm is frequently used. Although the absolute water absorption has increased at 1040 and 1300 nm over 840 nm, these tissues exhibit significantly greater light scattering. In addition, higher-power light sources are now permitted to be used under eye-safety regulations, due to the increased absorbance of these wavelengths in the humor. The increased power overcomes the water absorbance, enabling a net increase in imaging depth.
Acquisition speed is very important to doctors who may need to quickly get complete images before the patient moves. Imaging in the short-wave infrared using diode-array spectrometers for spectral-domain OCT has previously been limited to rates of less than 7700 A-lines/s, or swept-source OCT at 16,000 or 26,000 A-lines/s. Recent developments of a new series of InGaAs cameras have increased spectral-domain OCT system acquisition rate to more than 46,000 A-lines/s, enabling the generation of OCT images of 1000 A-lines at 46 frames/s. This imaging rate further enables additional applications, such as measuring blood flows and paves the way for more biomedical imaging applications to follow.
Douglas Malchow is business development manager for commercial products at Sensors Unlimited, part of Goodrich (Princeton, NJ). Contact him at email@example.com.
OCT: How it works
Optical coherence tomography (OCT) obtains high-resolution images of tissue by focusing near-infrared light into tissue and measuring the intensity and position of the resulting reflections (see figure). OCT offers resolution high enough to distinguish tissue microstructures; thus, it approaches the level of standard histopathology and can image tissue in vivo.
In a basic OCT system, a focused beam of white light is sent through a beamsplitter. The light is divided into two beams, with one beam directed to a reference mirror and the other to the sample being studied. Both are reflected back to the beamsplitter, which sends them to a detector where they interfere with each other. The detector signal is collected by a computer, which then translates the information into pictures for interpretation.
In a typical single-point OCT setup, scanning the light beam on the sample enables noninvasive cross-sectional imaging up to 3 mm in depth with micrometer resolution.
The high bandwidth of near-infrared light (800-1300 nm) enables OCT to achieve image resolutions up to 25 times greater than other imaging modalities; for example, 10 to 12 µm resolution vs. 100 µm for ultrasound. As a result, clinicians can examine anatomical structures in vivo on a microscopic scale to diagnose disease and better direct therapies. Another advantage of OCT is that it can be integrated with an optical fiber to carry the sampling light to the tissue, and several groups have integrated the fiber with standard endoscopes to image deep inside the body.
The key components in an OCT system are the optics (lenses and mirrors), laser/light sources (including superluminescent diode lasers, ASEs, and swept-laser sources), beamsplitters, filters, photodetectors, detector arrays, digital signal processors, CCD and CMOS cameras, and fiber optics (for endoscopic applications).