Open Access
Add to BibSonomyAbstract
Systems for infrared reflectance imaging are built with an FT-IR spectrometer, hollow optical fibers, and a high-speed infrared camera. To obtain reflectance images of biological samples, an optical fiber probe equipped with a light source at the distal end and a hybrid fiber probe composed of fibers for beam radiation and ones for image detection have been developed. By using these systems, reflectance spectral images of lipid painted on biomedical hard tissue, which provides reflectance of around 4%, are successfully acquired.
© 2015 Optical Society of America
1. Introduction
Fourier transform infrared (FT-IR) spectroscopy is one of the major tools for biomedical applications because many biomedical molecular markers exist in the mid-infrared region. This technique enables us to investigate chemical bonds of biomedical molecules such as lipids, proteins, nucleic acids, etc. rapidly and noninvasively. In addition, only small amounts of a sample are necessary for analysis [1,2]. Many reports reveal that, by utilizing FT-IR techniques, changes in biomedical samples can be detected via analyzing absorption spectra to identify malignant and benign tissues [2–5]. Furthermore, a system of Fourier transform-infrared imaging spectroscopy (FT-IRIS), which involves the combination of an FT-IR microscope system with a multichannel detector, rapidly collects spectra from multiple points on a sample [6–8]. This is a useful tool that has resulted in highly reliable detection of early tumors [9] and quantitative pathological diagnoses [10].
In these FT-IRIS systems, however, usually analysis is executed in a sample compartment, and only limited parts of the surface of body can be observed in vivo [11,12]. For measuring inner organs such as circulatory and digestive systems in vivo, an imaging probe that can be inserted into an endoscope is essential. Although an optical fiber probe is considered a proper option, a common silica-core optical fiber cannot be used due to its extremely strong material absorption in the mid-infrared region. Instead, chalcogenide glass-core fibers [13] and silver-halide crystalline fibers [14,15] have been developed so far. However, fiber probes made of these materials cannot be applied in medical purposes because of the toxicity or chemical instability of the fiber materials. For delivery of mid-infrared images, we have developed a bundle of hollow optical fibers and show its capability of thermal image guiding of body temperature samples [16]. The hollow optical fiber consists of a thin glass capillary tube, a thin metal film on its inner wall, and dielectric film top coating on the metal to enhance reflectivity [17]. A hollow optical fiber has high flexibility and nontoxicity, and the air core structure prevents infrared light from material absorption and results in low loss transmission in the fiber. Therefore, the hollow optical fiber is considered an ideal candidate for endoscopic applications.
By using an optical fiber probe consisting of a hollow optical fiber and an attenuated total reflection prism, we have succeeded in analyzing biomedical samples [18,19]. However, it was a single-point measurement, and a scanning mechanism was necessary to obtain an image, although it is difficult to equip such a mechanism at the distal end of endoscope. Therefore, we proposed a spectral imaging system based on a hollow optical fiber bundle [16], which transmits an infrared radiation image to expand the use of the fiber probe to multipoint measurements. We have constructed a system consisting of a fiber bundle, an FT-IR spectrometer, and a high-speed infrared camera and succeeded in obtaining a transmittance spectral imaging of biological samples [20]. For endoscopic spectral imaging, however, infrared light reflected from the surface of the sample should be collected. In this paper, we propose two types of hollow-optical fiber probes for reflectance spectral imaging and construct a system utilizing these fiber probes.
2. Spectral Imaging with FT-IR and Hollow Optical Fiber Bundle
We first tried to obtain a reflectance spectral image by using a measurement system, as shown in Fig. 1. We reconstructed the system for a transmission imaging system, as shown in [20], by inserting a beam splitter between the FT-IR and the hollow-optical fiber bundle. In this system, infrared light emitted from the FT-IR spectrometer (Bruker Vertex 70) is modulated by a Michelson’s interferometer inside the spectrometer. The emitted light is focused onto a hollow-optical fiber bundle (19 fibers, 0.32 mm bore diameter, and 30 cm length), and reflected light from the sample is collected by the fiber bundle and detected by the high-speed infrared camera (FLIR SC-4000), which is composed of a focal plane array (FPA) of InSb. The FPA has pixels, and the detection wavelength range is from 3 to 5 μm. The frame rate is as high as 1000 fps, which is high enough to capture precise interferograms of each pixel. The whole scan took 5.3 s.
Figure 2 shows an example of the interferogram obtained at one of the pixels. By calculating the Fourier transform of this interferogram, a reflected power spectrum is obtained. We used a gold mirror as a reference, and, by calculating the Fourier transform of all the pixels, a spectral image of the sample is acquired.
In Fourier-transform calculations, the apodization function of was applied, as indicated in Eq. (1), because the working range of the movable mirror in Michelson’s interferometer is not infinity:
In Eq. (1), is the frequency spectrum of light power, is light intensity at the optical path difference of , and is the total difference in the optical path. We chose the Happ–Genzel function [21], shown in Eq. (2), as the apodization function:
Although we attempted to obtain the reflectance spectra of a porcine sample, due to extremely low reflectance of biological samples (0.5%–1%) [22], it was difficult to obtain a characteristic spectrum of the sample. From the measured result, we evaluated SNR of the above system as 6.2. Main causes of this low SNR are the 6 dB loss at the beam splitter and the transmission loss of the fiber bundle, which is around 12 dB for a round trip.
3. Fiber Probe Equipped with an Infrared Light Source at the Distal End
To improve the system’s SNR, here we propose an optical fiber probe equipped with a light source at the distal end. As shown in Fig. 3, a spherical light absorber is set in the center of probe end, and the absorber is irradiated with a laser beam with a power of 500 mW delivered by the centered hollow optical fiber. Then, temperature of the heated absorber is elevated to around 773 K, and it radiates a broadband infrared beam as a blackbody. The sample is set in front of the probe, and, because of radiation in a semispherical direction, a wide range of the sample is irradiated with the infrared beam. The reflected beam from the sample is then collected by the fiber elements surrounding the light source and propagated to an FT-IR spectrometer (Newport MIR8035). After passing through the interferometer in the FT-IR spectrometer, the beam is detected by the high-speed infrared camera (FLIR SC6000) for spectral imaging construction. In this system, a beam splitter becomes unnecessary, and the infrared beam transmits the fiber only one way.
As shown in the inset of Fig. 3, the probe is composed of a metallic ball absorber, a hollow-optical fiber for laser propagation, hollow-optical fibers for image propagation, a ceramic waveguide tube. As the absorber, we used a metal ball of 1 mm diameter, which is covered by blackbody coating, thus providing a high absorption coefficient for the laser beam. The absorber was fixed on the inside of a stainless tube and is mounted on the fiber for laser propagation. The hollow optical fiber for laser propagation is with a silver and an AgI coating the inner wall, and the thickness of the AgI film is optimized for laser propagation. The length is 50 cm and the inside diameter (I.D.) is 0.7 mm. We first measured the radiation spectrum of the absorber using a monochrometer with a resolution of 0.2 μm and a bolometer detector with a spectral range of 6–14 μm. A measured spectrum is shown in Fig. 4. For comparison, a power spectrum of FT-IR spectrometer measured at the output end of the fiber bundle is also shown in Fig. 4. It is shown that the radiation intensity of the proposed light source is around 10 times higher than that of the FT-IR beam. Hollow optical fibers for image propagation is with only a silver film coated on the inside. The length is 30 cm and the inside diameter is 0.32 mm. In contrast to the fiber for transmission of the laser, the silver-only coated fiber provides broad low losses in the mid-infrared region.
The ceramic waveguide tube shown in the inset of Fig. 3 is an alumina ceramic tube coated with a silver film on the inside. It is used for raising the illumination efficiency on a sample by reflecting the radiated beam by the metal coating. Beside, it decreases thermal influence of the light source as a heat isolator. In our measurements, the end of the fiber probe touches the sample surface, and the distance between the light source and the sample is called the stand-off distance. The illumination pattern varies with stand-off distances; thus, power density of the beam radiated on the sample also varies. With small stand-off distances, the whole area of sample cannot be illuminated. In an opposite case when the distance is too large, the power of reflectance light attenuates due to the loss of the ceramic waveguide. The relationship between detected power intensity and stand-off distance is shown in Fig. 5. It indicates that the power intensity is maximized at the stand-off distance of 4 mm, and we adopted this distance for the following experiments.
As a sample for system evaluation, we used a metallic mirror whose surface is partially coated with a thin layer of porcine fat. Figure 6 shows reflection spectra measured by fiber elements that correspond to parts of the mirror surface with and without a fat layer. In the spectral imaging procedure, reflection spectra of all the fiber elements are first obtained, and then some characteristic components are picked up. In these spectra, absorption of fat, which originated from the C–H bonds, is clearly seen at around 3.5 μm. By integrating the spectral reflectance from 3.4 to 3.7 μm, the reflectance spectral image of fat is constructed by mapping the intensities, as shown in Fig. 7. The border of the fat layer is clearly seen in the reflectance image, although there is an inevitable defect due to the light source at the center of image.
In this system, we proposed a fiber probe with a broadband light source, which is different from the system used in the preliminary test. Thus, we could raise the power of the light source by increasing power of the laser, and it is possible to obtain much higher power than FT-IR. However, we evaluated the SNR of the system from the above results, and it was only 9.3, which was far below the expected value. One of the main reasons of this low SNR is distortion in infrared images in the FT-IR interferometer. In this system, an interferometer is allocated at the end of fiber bundle for image transmission. When the size of transmitted beam is not very small, the images are distorted because light path length is different between the horizontal and vertical axis due to the flat beam splitter, which is obliquely placed on the light axis in the interferometer.
4. Hybrid Hollow Optical Fiber Probe for Spectral Imaging
To solve the above problems, which are a centered large defect and distortion of observed images, a fiber probe composed of two types of hollow optical fiber is proposed. A schematic of the probe structure and the overview of measurement system are shown in Fig. 8. Hollow optical fibers with a large bore size of 0.7 mm are used to deliver an infrared beam from an FT-IR spectrometer. Due to the large core size, the transmission loss of the fiber can be kept low. A bundle of small bore, 0.32 mm diameter fibers is for transmission of reflected infrared images. The small diameter provides detection of high-resolution images. These two types of fiber are combined, and the bundle for image transmission is surrounded by the large bore fibers at the end of the probe. All of these fibers are coated with silver and AgI film on the inner wall for low transmission losses. On the end of the fibers, a tapered glass hollow optical waveguide with a silver inner coating is mounted to efficiently irradiate the center region in front of the probe with an infrared beam.
In this system, the images are not distorted because the FT-IR interferometer is located ahead of the fibers for illumination. In addition, no defect appears detected in the images. For sample preparation, a thin layer of porcine fat is painted on a metal reflector in an “” shape. Measured spectra of the reflector with and without the fat layer are shown in Fig. 9. An absorption of C–H bonds is clearly seen in the spectrum of the fat layer. Figure 10 shows a measured reflectance image of the sample whose visible image is shown left. The reflectance image is obtained by integrating the absorption spectra from 3.4 to 3.7 μm. In this image, the contrast of fat and reflector is clearly seen, and the L-shape fat layer is observed. The calculated SNR of the spectra is 36.4, which is six times higher than the one acquired by the setup shown in Fig. 1.
Then, we prepared a sample of a human incisor tooth (specular reflectance 4%) whose surface was partially coated with a thin layer of oil. The measured spectra of a sample are shown in Fig. 11. In spite of the reflectance of the sample, which was much lower than the metal reflector samples, the absorption band appears on the spectrum of the fat layer, and the calculated SNR of spectra was 9.0. Figure 12 shows an observed image, and the border of the fat layer is clearly seen by mapping the integrated reflectance in the C–H absorption band.
Finally, to check the feasibility of the system for biomedical soft tissue, we prepared a sample of porcine tissues with lean and fat. The specular reflectance of these soft tissues are less than 1%. In the experiment, four hollow-optical fibers with a larger bore size of 1.0 mm were used for infrared light radiation to raise the radiation power, and, as a result, twofold SNR was obtained. Before measurement, the sample was dried, and the surface was flattened to suppress the surface scattering.
The fiber probe was located on the border of lean and fat. In the measured spectra of the sample, the absorption peaks of fat appear on the spectra, and these peaks did not exist in the spectra of lean. The calculated SNR of spectra was 3.8. Figure 13 shows an observed image, and the border of the fat and lean is clearly observed by mapping the integrated reflectance in the C–H absorption band.
5. Conclusion
Reflectance spectral imaging systems based on two types of hollow optical fiber probes propagating an infrared radiation image are proposed. First, we fabricated a hollow optical fiber based probe with a small infrared light source installed on its end. By placing the light source at the end, a beam splitter used in the former system, which caused an inevitable loss, is not necessary. In addition, transmission due to the fiber becomes half of the former. Although the radiated power of this system was 10 times higher than the one transmitted in the fiber from the source in the FT-IR spectrometer, the transmitted images show a centered large defect. Furthermore, the obtained image was distorted because the FT-IR interferometer was located after image transmission, and, as a result, the spectral SNR was worse than expected. This problem may be solved by decreasing the image size in the interferometer by using a lens system. In principle, the system is not affected by the transmission loss of the fiber before radiation on to the sample; therefore, much higher power density of infrared light than the one from the hybrid probe will be obtained.
Second, a hybrid fiber probe is designed to solve these problems. This probe comprises large-core hollow fibers for delivery of illumination beam and a bundle of small-core fibers for image transmission. With this hybrid structure, the spectral SNR were drastically improved, and a hard tissue sample with low reflectance of around 4% was successfully imaged. Furthermore, with the probe, although the SNR is low, a soft tissue sample was also successfully imaged, and the system is possible to be used in the practical biomedical tissue measurement.
To improve the resolution and flexibility of the probe, fibers with much smaller diameters are necessary, although thinner fibers provide higher losses. To reduce the losses of hollow-optical fibers, we have worked on the fibers with a multilayer coating on the inside, and we are currently working on a design for broadband transmission. To improve spatial resolution, thus keeping the system SNR high enough for imaging, we are trying to use a quantum cascade laser [22] and infrared tunable lasers [23], which emit much higher power than the FTIR system. These systems can be used after acquiring a spectral reflectance image to choose specific components because of the narrower wavelength band of these lasers.
References
1. Z. Movasaghi, S. Rehman, and I. Rehman, “Fourier transform infrared (FTIR) spectroscopy of biological tissues,” Appl. Spectrosc. Rev. 43, 134–179 (2008). [CrossRef]
2. S. Argov, J. Ramesh, A. Salman, I. Sinelnikov, J. Goldstein, H. Guterman, and S. Mordechai, “Diagnostic potential of Fourier-transform infrared microspectroscopy and advanced computational methods in colon cancer patients,” J. Biomed. Opt. 7, 248–254 (2002). [CrossRef]
3. P. Wong, R. Wong, T. Caputo, T. Godwin, and B. Rigas, “Infrared spectroscopy of exfoliated human cervical cells: evidence of extensive structural changes during carcinogenesis,” Proc. Natl. Acad. Sci. USA 88, 10988–10992 (1991).
4. D. Maziak, M. Do, F. Shamji, S. Sundaresan, D. Perkins, and P. Wong, “Fourier-transform infrared spectroscopic study of characteristic molecular structure in cancer cells of esophagus: an exploratory study,” Cancer detection and prevention 31, 244–253 (2007).
5. R. Sahu and S. Mordechai, “Fourier transform infrared spectroscopy in cancer detection,” Future Oncol. 1, 635–647 (2005).
6. S. Kazarian and K. Chan, “Applications of ATR-FTIR spectroscopic imaging to biomedical samples,” Biochim. Biophys. Acta 1758, 858–867 (2006). [CrossRef]
7. X. Bi, G. Li, S. Doty, and N. Camacho, “A novel method for determination of collagen orientation in cartilage by Fourier transform infrared imaging spectroscopy (FT-IRIS),” Osetoarthritis Cartilage 13, 1050–1058 (2005).
8. C. Johnson, N. Pleshko, M. Achary, and R. Suri, “Rapid and sensitive screening of 17β-estradiol estrogenicity using Fourier transform infrared imaging spectroscopy (FT-IRIS),” Environ. Sci. Technol. 48, 4581–4587 (2014). [CrossRef]
9. J. Anastassopoulou, E. Boukaki, C. Conti, P. Ferraris, E. Giorgini, C. Rubini, S. Sabbatini, T. Theophanides, and G. Tosi, “Microimaging FT-IR spectroscopy on pathological breast tissues,” Vibr. Spectrosc. 51, 270–275 (2009).
10. R. Mendelsohn, E. Paschalis, P. Sherman, and A. Boskey, “IR microscopic imaging of pathological states and fracture healing of bone,” Appl. Spectrosc. 54, 1183–1191 (2000).
11. L. Mostaço-Guidolin, L. Murakami, A. Nomizo, and L. Bachmann, “Fourier transform infrared spectroscopy of skin cancer cells and tissues,” Appl. Spectrosc. Rev. 44, 438–455 (2009).
12. B. Balázsa, G. Farkasa, O. Berkesic, R. Gyulaid, S. Berkóa, M. Budai-Szűcsa, P. Szabó-Révésza, L. Keménye, and E. Csányi, “Protein structure is changed in psoriatic skin on the unaffected region—Imaging possibility with ATR-FTIR spectroscopy,” Microchem. J. 117, 183–186 (2014).
13. B. Bureau, C. Boussard, S. Cui, R. Chahal, M. Anne, V. Nazabal, O. Sire, O. Loréal, P. Lucas, V. Monbet, J. Doualan, P. Camy, H. Tariel, F. Charpentier, L. Quetel, J. Adam, and J. Lucasa, “Chalcogenide optical fibers for mid-infrared sensing,” Opt. Eng. 53, 027101 (2014). [CrossRef]
14. Y. Lavi, A. Millo, and A. Katzir, “Thin ordered bundles of infrared-transmitting silver halide fibers,” Appl. Phys. Lett. 87, 241122 (2005). [CrossRef]
15. A. Millo, L. Lobachinsky, and A. Katzir, “Single-mode index-guiding photonic crystal fibers for the middle infrared,” IEEE Photon. Technol. Lett. 20, 869–871 (2008).
16. Y. Matsuura and K. Naito, “Flexible hollow optical fiber bundle for infrared thermal image,” Biomed. Opt. Express 2, 65–70 (2011). [CrossRef]
17. Y. Matsuura, M. Saito, M. Miyagi, and A. Hongo, “Loss characteristics of circular hollow wave guide for incoherent infrared light,” J. Opt. Soc. Am. A 6, 423–427 (1989).
18. S. Kino and Y. Matsuura, “Nontoxic and chemically stable hollow optical fiber probe for Fourier transform infrared spectroscopy,” Appl. Spectrosc. 61, 1334–1337 (2007). [CrossRef]
19. S. Kino, Y. Tanaka, and Y. Matsuura, “Blood glucose measurement by using hollow optical fiber-based attenuated total reflection probe,” J. Biomed. Opt. 19, 057010 (2014).
20. C. Huang, S. Kino, T. Katagiri, and Y. Matsuura, “Remote Fourier transform-infrared spectral imaging system with hollow-optical fiber bundle,” Appl. Opt. 51, 6913–6916 (2012). [CrossRef]
21. J. Rabolt and R. Bellar, “The nature of apodization in Fourier transform spectroscopy,” Appl. Spectrosc. 35, 132–135 (1981).
22. P. Bassan, M. Weida, J. Rowletteb, and P. Gardner, “Large scale infrared imaging of tissue micro arrays (TMAs) using a tunable quantum cascade laser (QCL) based microscope,” Analyst 139, 3856–3859 (2014). [CrossRef]
23. S. Kumar, A. Esteban-Martin, T. Ideguchi, M. Yan, S. Holzner, T. Hänsch, N. Picqué, and M. Ebrahim-Zadeh, “Few-cycle, broadband, mid-infrared optical parametric oscillator pumped by a 20-fs Ti:sapphire laser,” Laser Photon. Rev. 8, 86–91 (2014).
