Standard pathology requires removal of cells or tissue through techniques ranging from fine needle aspiration or core needle biopsy to surgical resection. To allow examination of suspect lesions, a variety of endoscopic imaging techniques may be employed, including fiber-scanning confocal microscopy in reflectance1, 2 and fluorescence3 modes. These techniques use an optical fiber or fiber bundle to view the surface of interior body cavities. Images are created by scanning mechanisms positioned at either the distal or proximal ends of the fiber. These types of systems can capture point-by-point high-resolution images from a large field of view, and the sensitivity of the systems can be very high.4, 5 Recent advances provide similar high-resolution images using simple, low-cost systems based on coherent fiber bundles with images captured using inexpensive digital cameras.6, 7, 8 Image resolution is limited by the individual fibers of the fiber bundle. A fluorescent-mode fiber endoscope developed by Muldoon achieves a resolution of .6, 7 We are developing a high-resolution reflectance-mode endoscope that is flexible, compatible with a needle, and can generate still and video rate images using intrinsic contrast. Reflectance-mode endoscopic imaging can potentially complement fluorescence imaging by capturing information not available in fluorescence mode, and by functioning in situations when fluorescent imaging is not practical.
In this work, we present a needle-based, high-resolution reflectance-mode fiber microscope (RFM) developed for optical tissue biopsy applications. The RFM was assembled on a portable aluminum breadboard to facilitate eventual clinical trials. As shown in Fig. 1 , an air-cooled light-emitting diode (LED) (Luxeon, Brantford, Ontario, Canada; white LED, model LXHL-NWE8; blue LED, model LXHL-NRR8) served as the illumination source to keep the RFM as compact as possible and reduce the system cost. The light from the LED is collimated by an integrated collimating lens and passes through a linear polarizer (Fig. 1), after which the polarized light is reflected by a glass slide positioned at (with respect to the beam trajectory) and projected onto an infinity-corrected objective lens (Newport, Irvine, California, model , ). The objective lens and a fiber chuck (Newport, model FPH-DJ) are positioned with a multimode fiber-coupler positioner (Newport, model F-915T). By careful alignment, the illumination light is focused into the incident end of a outer-diameter fiber image guide positioned by the fiber chuck (Sumitomo Electric U.S.A., Los Angeles, California, model IGN-037/10, fiber elements in a picture area of diameter, ).9 Finally, the imaging end of the fiber image guide is inserted into the core of a needle and positioned in contact with the samples. Light scattered from the samples is collected by the same illumination fiber and passes back through the same objective lens and glass slide described previously (see Fig. 1).
A charge-couple device (CCD) color camera (Edmund Optics, Barrington, New Jersey, model EO-0813C) coupled with a circular polarizer filter (Nikon, ) is used to sample the fiber bundle elements for image acquisition. Because the polarization orientation of the S-polarized illumination light is perpendicular to the plane of incidence with respect to the glass slide (i.e., the XY plane of Fig. 1), a significant portion is reflected toward the incident end of the fiber bundle to be delivered to the imaging end of the fiber bundle. Specular reflectance from the incident end of fiber bundle is also partly reflected away by the glass slide, and the remaining specular reflectance is further suppressed by the circular polarizer in front of the CCD camera that passes P-polarized light with respect to the plane of incidence. Rays that result from multiple scattering events changing their polarization to the parallel direction can be detected by the CCD camera (e.g., light scattered from the sample). Thus, noise due to specular reflectance from the incident end of the fiber bundle is greatly suppressed, and clearer images are collected.
Figure 2 shows a reflectance image of the fiber bundle imaging end when the needle microscope is positioned in air. Light from the white LED that is reflected from the glass-air interface at the imaging end of the fiber bundle is successfully detected by the CCD camera, and individual fibers of the fiber bundle are resolved [Fig. 2]. With the imaging needle of the RFM immersed into a vial of water or phosphate-buffered saline (PBS) solution, the RFM field of view is almost completely dark, because the close match between the refractive indices of water and the glass fiber result in dramatically reduced internal reflection from the imaging end of the fiber bundle.
Figure 2 is obtained with a drop of water covering the tip of the imaging fiber bundle, which mimics the immersion of the fiber tip in water, to provide clear image features. Much less internal reflection occurs from the imaging area covered by water [Fig. 2] compared with the same area when it is exposed to air. Figures 2 and 2 were captured with the same imaging configurations and the index-matching effect is clear. When used in biological or clinical environments, the fiber bundle tip is always in contact with body fluids or tissue. Thus, dipping the fiber bundle tip into water roughly approximates the conditions of realistic imaging applications. In liquid, the reduced internal reflectance leads to a dark field of view and a decrease in background enabling biological imaging. Light scattered by biological structures can be clearly distinguished and cellular images can be obtained. In contrast to the polarized reflectance-mode images in Figs. 2 and 2, images shown in Figs. 2 and 2 were captured with a reduced exposure time under the same imaging conditions, but without the illumination and collection polarizers. The specular reflectance from the optical components makes it impossible to obtain useful information from Figs. 2 and 2.
A white LED was used to acquire the images and plots shown in Figs. 3, 3, 3 and Video 1 . Figure 3 shows an imaging resolution of ( line pairs/mm) obtained upon imaging the United States Air Force resolution target. Figure 3 shows that contrast decreases slightly from low to high resolution, but it also reveals high-frequency components due to the probe fibers ( diam). Figure 3 and Video 1 were obtained by using the same white LED-illuminated RFM to image skin tissue from a green onion. The cell walls of the rectangular green onion skin cells10 are clearly resolved in both the still image and the video. Note that water was used in all the tests to provide refractive index matching.10.1117/1.3465558.110.1117/1.3465558.2
The LED light source is cost effective and easy to change. An air-cooled blue LED was used to image cultured SKBR3 human carcinoma cells. Before imaging, SKBR3 carcinoma cells were cultured in a plate at for . The fiber bundle tip was immersed into the culture medium through a needle and focused onto a group of cells attached to the culture plate. As shown in Fig. 3, the cells can be visualized with the RFM using blue-LED illumination. An endoscopic video clip of the SKBR3 cancer cells is provided in Video 2 . The same RFM was also used to image excised human oral cancer tissue obtained from the Cooperative Human Tissue Network (CHTN, Rockville, Maryland). The tissue was placed in a culture plate with PBS, which was then placed on a 2-D translation stage. The needle and fiber bundle tip were inserted into the tissue sample, and cellular structures were visualized as shown in Fig. 4 and Video 3 . In Video 3, cellular images move in and out of focus as the imaging tip moves closer to or further away from the sample. The same RFM was used to image the reflectance from a dust particle in liquid solution to demonstrate single pixel fiber resolution, as shown in Video 4 .10.1117/1.3465558.310.1117/1.3465558.4
In this work, we describe a high-resolution RFM developed for optical tissue biopsy applications. The resolution of this fiber microscope is limited by the element fiber size of the imaging fiber bundle. For the specific fiber bundle image guide presented, the resolution is . The reported RFM system uses the same fiber bundle guide for both illumination and imaging. A pair of cross-polarized filters is used in addition to the fiber bundle guide to achieve high-resolution reflectance-mode microscopic imaging. The fiber bundle guide is in diameter, making it compatible with current clinical needle tissue biopsy systems. The RFM system can be used without external agents when there is sufficient endogenous imaging contrast. Moreover, with simple filter changes, the RFM system can readily be converted to a fluorescence-mode needle biopsy imaging system similar to that developed by Muldoon 6 Future work will assess whether performance can be improved through the use of simple chemical agents already in routine clinical use.
The authors gratefully acknowledge support from the Department of Defense Congressionally Directed Breast Cancer Research Program Era of Hope Scholar Award (W81XWH-07-1-0428). The NCI Cooperative Human Tissue Network (CHTN) provided specimens used in this project.
K. C. Maitland, A. M. Gillenwater, M. D. Williams, A. K. El-Naggar, M. R. Descour, and R. R. Richards-Kortum, “In vivo imaging of oral neoplasia using a miniaturized fiber optic confocal reflectance microscope,” Oral Oncol.0964-1955 44, 1059–1066 (2008).10.1016/j.oraloncology.2008.02.002Google Scholar
F. Jean, G. Bourg-Heckly, and B. Viellerobe, “Fibered confocal spectroscopy and multicolor imaging system for in vivo fluorescence analysis,” Opt. Express1094-4087 15, 4008–4017 (2007).10.1364/OE.15.004008Google Scholar
V. Dubaj, A. Mazzolini, A. Wood, and M. Harris, “Optic fiber bundle contact imaging probe employing a laser scanning confocal microscope,” J. Microsc.0022-2720 207, 108–117 (2002).10.1046/j.1365-2818.2002.01052.xGoogle Scholar
R. Kiesslich, M. Goetz, M. Vieth, P. R. Galle, and M. F. Neurath, “Technology Insight: confocal laser endoscopy for in vivo diagnosis of colorectal cancer,” Nat. Clin. Pract. Oncol. 4, 480–490 (2007).10.1038/ncponc0881Google Scholar
T. J. Muldoon, M. C. Pierce, D. L. Nida, M. D. Williams, A. Gillenwater, and R. Richards-Kortum, “Subcellular-resolution molecular imaging within living tissue by fiber microendoscopy,” Opt. Express1094-4087 15, 16413–16423 (2007).10.1364/OE.15.016413Google Scholar
T. J. Muldoon, S. Anandasabapathy, D. Maru, and R. Richards-Kortum, “High-resolution imaging in Barrett’s esophagus: a novel, low-cost endoscopic microscope,” Gastrointest. Endosc.0016-5107 68, 737–744 (2008).10.1016/j.gie.2008.05.018Google Scholar
T. J. Muldoon, N. Thekkek, D. Roblyer, D. Maru, N. Harpaz, J. Potack, S. Anandasabapathy, and R. Richards-Kortum, “Evaluation of quantitative image analysis criteria for the high-resolution microendoscopic detection of neoplasia in Barrett’s esophagus,” J. Biomed. Opt.1083-3668 15, 026027 (2010).10.1117/1.3406386Google Scholar
Sumitomo Electric U.S.A., Image Guide Technical Document ⟨http://www.sumitomoelectricusa.com/scripts/products/ofig/image.cfm⟩ (2010).Google Scholar