|
The The following presentation entitled "Optical Biomedical Imaging" describes another light based diagnostic technology developed by Mediscience Technology Corp. This technology is in addition to the Fluorescence Spectroscopy also developed by Mediscience.
To view the Mediscience PATENT PORTFOLIO for Optical Biomedical Imaging, click here.
Optical Biomedical Imaging By S. K. Gayen and R. R. Alfano Researchers in optical biomedical imaging are pursing a dream that someday soon they will develop noninvasive modalities enabling doctors too peer inside the human body by shining light. By analyzing how light interacts with tissue, these �optical body scanners� would help doctors diagnose lesions such as tumors in the breast, and would determine whether a tumor is malignant or benign. Although an impressive array of techniques-among them, x-ray computed tomography (CT), magnetic resonance imaging (MRI), Ultrasound, and radioisotope imaging�are already available for imaging parts of the body, there still are limitations: � X-rays or other established techniques may not accurately detect tumors in their earliest stage when they are small and most treatable. X-rays also are not suitable for imaging young, dense breasts and cannot distinguish between benign and malignant tumors. In addition, their high energy can ionize the tissue and may be potentially harmful if used to often for routine screening or at high intensity. � Radioisotope imaging has limited application and exposes the body to radiatioactivity that is potentially harmful. � Ultrasound lacks the resolution to detect objects with linear dimensions smaller than a few millimeters, and like X-rays, may not provide any information about the tissue chemistry. � MRI is a powerful technique with sub millimeter spatial resolution and the ability to detect some specific chemicals. However it cannot detect some important elements such as oxygen, and the cost of superconducting magnets needed for its operation makes it highly expensive. The need for an imaging modality that is noninvasive, safe, inexpensive, compact, and capable of monitoring tissue chemistry in viva preferably in real time, provides the impetus for the intense research activities in biomedical optical imaging. This article provides a brief introduction to the state of the art in these endeavors. Background Biomedical imaging techniques are sought to identify, locate and diagnose an �object� such as a tumor inside a biological tissue. The successful use of optical methods to solve this imaging problem has eluded researchers since the Victorian doctors futile attempt to detect breast tumors using light from candles. The interest in this technique, known as diaphamography, has increased throughout the century as brighter light sources and sophisticated image recording methods were invented and experimental data accrued to suggest that light can monitor tissue function. In its simplest form, the procedure involves illuminating the part of the body to be imaged with bright light and search for indications of pathology in the observed transillumination pattern. The physical basis of transillumination is the difference in the transmission of light though normal tissue and an object, such as a tumor, inside it. This differential transillumination in principle should lead to the formation of a �shadow� of the tumor in the transmitted light. However observation of the shadow image is frustrated since light scattered by the tissue severely degrades, and for thick tissues, completely buries the image in the background light. The key to successful development of optical imaging techniques is to deal with the problem of light scattering effectively. Understanding how light propagates through a highly scattering medium is an important step in resolution of the problem. Light is both scattered and absorbed by biological tissues, resulting in diffuse transmittance. Characteristics such as intensity, coherence, and polarization of the incident light change as it is absorbed and scattered inside the tissue. The extent of these changes depends on the wavelength of light, the type of tissue through which it propagates and the tissue thickness. Consequently, the light that emerges from the medium has very different characteristics than the incident light. Absorption results from electronic transitions within the energy levels of the atoms and molecules in the tissue. Scattering arises from variations of refraction in the microscopic and macroscopic structures of the tissue; it increases strongly as the wavelength of light becomes comparable to the size of the scattering elements. Light scattering is a statistical process, meaning that the probability that a photon will be scattered has certain value that depends on its wavelength and the type and thickness of the tissue. The transmitted light thus is comprised of unscattered (or coherently scattered), weakly scattered and multi-scattered photons. The coherently scattered or ballistic photons propagate in the direction of the incoming beam, transverse the shortest path, retain most of the characteristics of the incident photons and carry the maximum information about the interior structure of the scattering medium. The multiple scatter photons travel long distances within the medium, lose many of their initial characteristics, carry little information about the structures inside the medium and emerge later in all directions. They constitute what is known as the diffusive component of the transmitted light. The photons that scatter slightly in the forward direction retain significant initial properties and information on structures hidden in the scattering medium. These photons are called snake photons because their trajectories resemble a wriggling snake. Since the photons of an incident short pulse of light spend different times in transit through the intervening medium, the transmitted pulse becomes broadened with ballistic photons arriving first followed by the snake and diffusive photons.****Ballistic photons are the most effective in forming a shadow image. Snake photons also generate translumination images whose resolution depends on the position and width of the temporal slice used in imaging. The snake photons that form a well resolved shadow image arrive much earlier than most of the diffuse photons. The intensity of the forward-transmitted light is attenuated by absorption and scattering. The relative intensity of the three components in the transmitted beam vary depending on the color of light and characteristics of the sample. In a highly turbid medium, the ballistic component is extremely weak, while the diffuse component is most intense. To form images of the structure inside the turbid medium using the ballistic and snake photons the more numerous diffusive photons need be strongly discriminated against. Two alternative approaches are being pursued to alleviate this problem. The first involves filtering out the image bearing photons from the background of multi-scattered image-blurring photons using clever experimental techniques. The second approach sometimes referred to as the inverse problem is based on detection of multiple scattered photons at various positions around the object. The aim is to construct the image from these measured intensities, known experimental parameters, and tissue characteristics using models of light propagation and sophisticated computer algorithms. The impetus for solving the inverse problem stems from the realization that for tissues that are more than a few centimeters thick, the only transmitted light is the scattered light. Both frequency and time domain data may be used in this approach. The thrust is to develop appropriate algorithms to generate a map of the interior structure of a turbid medium in a reasonable time. The Monte Carlo method focuses on the particle nature of light and promises higher accuracy at the cost of higher computational complexity and time. Alternate methods based on the diffusion equation and the radiative transfer equation are being pursued by various groups. What light to use?Perhaps the most important consideration in optical imaging in tissues is the nature of the light to be used. Laser light- characterized by a high degree of spectral brightness, directionality, cohesion, monochromaticity, and optional short time duration- is used by most researchers these days. Several factors need to be considered to choose the optimal wavelength. First, shorter wavelengths provide better spatial resolution, but are strongly absorbed by the tissue. Second, to form a sharp image, one needs light that is scattered the least. Finally the wavelength range should be suitable to detect the key chemicals that can indicate how the body is functioning. Unlike ultrasound and X-rays only light offers this capability of spectroscopic monitoring of body chemistry. Near infrared light spanning the 700 to 1300 nm spectral range appears to be the judicious choice. Light in this range is not as strongly absorbed by tissue as the visible light, and so will have a higher transmission and less likelihood to cause burns. The availability of broadly wavelength-tunable solid state lasers such as Ti:sapphire and Cr:forsterite, to cover this spectral range is another advantage. The wavelength tenability offers the possibility of monitoring different chemical species by tuning to different wavelengths. Near-infrared light is also less scattered than viable light and is not known to cause any tissue damage even with prolonged exposure and intensity levels that may be needed for routine screening. However, the main problem of scattered light must still be confronted. Photon sorting techniquesSeveral schemes have evolved over the years to sort out the image bearing ballistic and snake photons from the multi-scattered diffuse photons. These schemes exploit one or more of the changes that scattering induces on the characteristics such as directionality, polarization, coherence, and temporal duration of the incident light. Since the image-bearing photons change the least, the idea is to devise a gate that will let the photons with a specific initial property through, but will block the others. One such scheme is a space gate that exploits the fact that the ballistic and snake photons come out of the tissue in the incident direction, while the multiple-scattered light emerges in all directions. So, a small aperture centered on the direction of incidence and placed after the sample will collect the ballistic and snake light and effectively discriminate against a significant fraction of the scattered light. Similarly, a time gate capitalizes on the fact that image-bearing photons emerge from the sample sooner than he diffuse photons. To realize a time gate in practice, one needs a shutter that will open for a short duration, typically a few picoseconds, to let early photons through and then close in time to leave out the delayed scattered photons. Similar gating schemes based on polarization and coherence of light have also been devised. Often a single gate may not be discriminating enough. For example, a space gate cannot filter out the photons that have first been scattered out of the incident direction and then back into it. However, a time gate employed after the space gate can cut out such �on-line� scattered photons, since scattering makes them travel longer distances and emerge out of the sample later than the ballistic photons that propagate straight through. The implementation of an optical imaging modality based on any of the gating methods mentioned above requires an appropriate laser and beam delivery optics to illuminate the sample, a detection scheme to monitor the transmitted light, and a signal processing unit to construct the image from the colleted data and known experimental parameters. The form in which data should be collected is a major consideration to the researchers, and three different approaches-time resolved, frequency domain and continuous wave are now used. Time-resolved methods commonly use ultrashort light pulses from an ultrafast laser to illuminate the sample. A time gate, often combined with a space gate, is then used to measure the time evolution of emergent light. Image information is extracted by analyzing the relevant temporal slice of the data. Frequency-domain techniques use modulated laser light for sample illumination and monitor the phase shifts associated with the average delays experienced by photons traveling through the sample for image construction. These two approaches provide theoretically equivalent data related by Fourier transform. However, in a frequency-domain technique, information is available only at certain discrete frequencies that are used in the experiment, while a time-domain method generates data at a large number of frequencies for image construction. As the name suggests, continuous wave laser light is used in continuous wave methods, where detection schemes are rather simple but the signal processing is more involved. Major techniques in these three approaches are highlighted in the following section. Overview of available techniquesSpace Gating Confocal imaging and Fourier spatial filtering are two commonly used space-gating techniques. In confocal imaging, an optical fiber collimator that faces the transmitter collects the light that passes straight through. Since the diameter of the collecting fiber is small, it only accepts forward propagating light and rejects off-axis scattered light. A two-dimensional image may be formed point by point by scanning the illuminator and collector fibers synchronously across the object while keeping them precisely aligned. In a Fourier gate, a collimated beam of light shines the object, and a lens of focal length F placed at a distance F focuses the photons that retain their collimation at a distance of 2F from the object. A small aperture placed at the focus of the lens transmits the image-bearing light and rejects most of the diffuse component. Polarization GatingThe polarization gate makes use of the fact that scattering
events depolarize an incident beam of linearly depolarized light so that ballistic and snake photons
retain much of the initial polarization while the multiple-scattered photons
are depolarized. In practice a polarization gate is implemented by shining the
object through a linear polarizer and collecting the emerging light through a
second linear polarizer. The degree of polarization [ defined as (Ip �
Is)/(Ip + Is) where Ip and Is are transmitted intensities
with the axis of the second polarizer parallel and perpendicular, respectively,
to that of the first] of the transmitted light is used to select the
image-bearing component, since it is expected ideally to be unity for ballistic
light and zero for completely depolarized diffuse light/ Time-resolved methodsThe time-resolved imaging methods developed to date may be classified broadly into incoherent and coherent techniques, depending on how the selection of ballistic and snake photons is achieved. In incoherent techniques, they are selected by opening a time gate for a short duration. A reference light pulse opens the gate and the light emerging from the sample can pass through for as long as the reference pulse is on. An optical delay line in the path of the reference pulse enables its temporal overlap with the desired part of the transmitted light. The ratio of the intensity of the light that passes trough when the reference pulse is present to that when it is off determines the contrast of the gate. These gates are often known as pulse gates, since the characteristics of the reference pulse play a major role in defining their properties. Examples of pulse gating techniques are optical Kerr gate (OKG), second-harmonic generation (SHG) gate, sum and difference frequency generation parametric gate, stimulated Raman scattering (SRS) gate, and time-correlated single photon counting. Imaging using a streak camera is an example of an incoherent electronic gate that operates by converting the temporal intensity profile of the transmitted light to a spatial one. It has a limited dynamic range of 104. The interference between image-bearing photons and a reference beam derived from the same initial pulse of light forms the mechanism for selective detection in coherent techniques. Photons that are coherent with the reference give rise to a gated signal, while defuse light does not. The coherence time of the reference pulse determines the duration of the gate. So, long-duration but broadband band light pulse may be used to generate ultra-fast gates. Optical coherent imaging (OCI), holographic methods, four wave mixing (FWM) gates, a coherent anti-Stroke Raman scattering (CARS) gate are coherent techniques. Frequency-domain MethodsThe frequency-domain methods are based on generation and measurement of the amplitude and phase of diffuse photon density wave (DPDW) in the turbid medium containing the object. Continuous-wave MethodsContinuous-wave (cw) Methods of optical
imaging are attractive since the use of compact cw lasers keeps the overall
system hardware simple. The first attempts at using light for tissue imaging,
known alternatively as diaphanagraphy or diaphanoscopy, were cw methods.
Confocal imaging and coherent detection are two commonly used cw techniques.
Confocal imaging has been used to build a laser transmission photoscanning
(LTPS) system for breast examination. The system uses two pairs of
transmitter-receiver fiber collimators-one operating in the red (630nm) and the
other operating in the near infrared (830nm)-and uses the differences between
transmissions at those wavelengths for spectral analysis. The breast cancer
diagnostic ability of the system has been tested in comparison with x-ray
mammography and the results are promising. A similar near-IR diaphanoscopic
system has been used for diagnosing sinusitis.
Coherent detection imaging (CDI) that uses high-sensitivity optical hetrodyne detection has been used to construct images of test phantoms and biological samples, as well as to obtain in vivo tomographic images of the human eye. Imaging horses to bet onSome of the more promising techniques at the current state of the development are the optical Kerr gate and coherence imaging in the time domain, and the diffuse photon density wave method in the frequency domain. Optical Kerr GateThe optical Kerr gate acts as a camera with an ultra fast shutter that is triggered by an intense gating pulse to take a picture the instant the image bearing light emerges and close soon enough to block the diffusive light. A typical OKG experimental arrangement consists of a Kerr active material such as carbon disulfide placed between two crossed polarizers. An intense ultrashort pulse of light is split into two parts. One part illuminates the sample, the other part-the reference or gating pulse-is used to induce a transient birefringence in the Kerr medium. The light that emerges from the sample is made to overlap spatially and temporally with the gating pulse inside the Kerr medium. If the reference pulse is blocked, a very small part of the incident light, typically less then one part in a million, may pass through the crossed polarizers; the gate is �closed�. In the presence of the reference pulse, the fraction of light from the sample that is linearly polarized by the first polarizer becomes elliptically polarized as it passes through the Kerr medium while coincident with the reference pulse. A fraction of this elliptically polarized light can now pass through the analyzer. This may be looked on as the gate being �open�. Thus the light can only pass through the gate when it overlaps both temporally and spatially with the intense reference pulse. The duration of the gating pulse or the recovery time of the Kerr medium, whichever is longer, determines how long the gate remains open. Picosecond and subpicosecond time resolution may be attained. The gate position may be varied by adjusting the arrival of the reference pulse by use of an optical delay line. The image-bearing early light may be extracted in a single shot by adjusting the gate position. A picosecond OKG that uses a picosecond Nd:glass laser, a CS2 Kerr cell, and a two-dimensional (2-D) charge-coupled device (CCD) camera has been successfully used the image test objects hidden inside human breast tissue, Chicken breast tissue, and a suspension of polystyrene spheres in water. While the image was completely blurred without the gate, a nicely resolved image was obtained when the gate selected the initial part of the transmitted pulse. Combined time and space gating was realized by placing the Kerr cell in the arrangement discussed above at the back focal plan of a lens, and an aperture at the center of the focal plane. Another lens placed at the distance equal to its focal length behind the aperture collimated the selected light and directed it to the CCD camera. Compared to simple gate (KFG) provided higher dynamic range, signal-to-noise ratio (SNR), and contrast at a signal level of-10-10 of the illumination intensity. The sensitivity of the technique to detect small changes in optical properties was underscored when it successfully imaged a pure water droplet inside 2% intralipid solution. In a related development, the feasibility of constructing a three-dimensional image was demonstrated by combining two-dimensional shadowgrams formed with a KFG using a backprojection algorithm on a personal computer. Optical Coherence Imaging
Optical coherence imaging (OCI) systems are based on interferometry with low-cohehence light. In an interferometer a beam of light is split into two different paths. The two beams form an interference pattern only when they are coherent-that is, the phase difference between the two remains fixed. If one of the beams interacts with a scattering medium, only the ballistic photons will retain their original phase and interfere with the reference beam. The diffusive component loses its original phase and cannot interfere. Interference thus filters out image-bearing ballistic photons and a fraction of the snake photons. In practice, OCI systems have been implemented both in retroreflection and transillumination geometries using compact fiber-optic Michelson and Mach-Zehnder interferometers, respectively. The use of optical fiber in the sample arms of interferometers ensures special filtering out of some of the scattered light by confocal imaging. In retroreflection OCI, light exiting the single-mode fiber in the sample arm is collimated and directed onto the sample being studied. The backscattered light from the sample is collected by the same fiber and combined in a fiberoptic beamsplitter with the beam reflected from the scanning reference mirror of the Michelson interferometer. Light retroreflected from different depths within the sample are measured by scanning the axial position of the reference mirror and simultaneously recording the interferometric signal amplitude. The two beams interfere only when the two arm path lengths match within the source coherent length. A broadband light source such as a superluminescence diode with short coherence length typicly-10picometers is used in OCI systems. So, longitudinal position of the reflections may be determined. By scanning the sample beam along the transverse dimensions and performing a longitudinal scan at each position, a two-dimensional image may be built up point-by-point. The beam diameter inside the sample determines the lateral resolution of the image. The detection scheme used in OCI is equivalent to optical heterodyne detection and provides high detection sensitivity. In a typical arrangement, using ~500uW power at 830 nm, a detection sensitivity of 94 dB was attained. The transilluminstion OCI system is similar to the retroreflection system in principle, but uses Mach-Zeynder geometry and a high-power femtosecond Tisapphire laser in the experimental arrangement. Using a sophisticated detection scheme, this technique achieves quantum shot noise limited detection sensitivity of 130 dB using an average power of ~150 mW. Even with such an impressive sensitivity, the retroreflection OCI may image through tissue depth of only 3 mm. Strong exponential attenuation of ballistic light is the limiting factor. However, a number of clinically interesting applications for three-dimensional imaging exist in this range. The technique is used to obtain high-resolution images of the human retina, and may prove useful in monitoring diseases such as glaucoma, macular degeneration, and macular edema. It is also used for high-resolution imaging of coronary artery and gastrointestinal tissues, and has potential applications in dermatology. An important drawback of the system is that it builds up the image point by point, which makes it to slow for vivo clinical applications. Diffuse Photon Density Wave MethodDiffuse Photon Density Wave (DPDW) method is a frequency-domain imaging technique that uses the migration characteristics of light for the formation of an image. In this method, light whose intensity is modulated sinusoidally at a high frequency, is incident on the turbid medium. As the photons diffuse through the medium, a periodic distribution of the photon density-a diffuse photon density wave-is set up. DPDWs are characterized by well-defined amplitude and phase at each point in the medium. They propagate with a phase velocity and an attenuation coefficient that depends on the modulation frequency, as well as the absorption and scattering properties of the medium. DPDWs refract, diffract, disperse, scatter, and interfere. The phase shift and amplitude modulation of this wave relative to the source may be measured at any point with proper instrumentation. The differential phase shift and modulation that results from the presence of an inhomogeneity on the medium form the basis for DPDW imaging. The image of the inhomogeneity may be constructed from phase shift and modulation data measured as a function of modulation frequency by use of computer algorithms. The method has been used recently to derive images of a purely absorbing and purely scattering inhomogeneity inside an intralipid solution. A dark horse: Photoacoustic ultrasound imagingThe combination of light and sound to form a novel imaging approach offers new advantages. In a photoacoustic technique, the absorption of pulsed light or chopped cw light by the sample produces sound waves. Once generated, the sound wave can propagate without any significant scattering since its wavelength is larger than the structures in the tissue; it can then be detected sensitively by acoustic transducers. Scattered light is less of a problem in this technique, since any light absorbed by the object enhances the signal level. The difference in absorption properties between the object to be imaged and the intervening medium forms the basis for photoacoustic imaging. Acoustic signals produced at ultrasound frequencies are of particular interest since they may provide submillimeter spatial resolution, and can be detected using highly developed medical ultrasound technology. The Technique shows promise for localization of tumors inside the tissue and diagnostic capability through selective absorption of light by the lesion. The potential of the technique has been demonstrated by detection of test objects inside an intralipid solution and chicken tissue with the use of both pulsed and modulated cw laser light. The inverse process, that is, modulation of light by ultrasound, offers another imaging approach. The beginningAlthough optical imaging and diagnosis of lesions inside the human body is still not a fait accompli, the techniques described in this article indicate that a new era of optical clinical imaging technology is just over the threshold. AcknowledgmentsThe authors thank the City University of New York, New York State Higher Education Applied Technology Program, National Aeronautic and Space Administration Institutional Research Award, New York State Technology Foundation, Mediscience Technology Corp., and the Office of Naval Research for support in part of the biomedical optical imaging research at City College of New York. Top of page OPTICAL IMAGING PATENT PORTFOLIO
|