1438
OPTICS LETTERS / Vol. 32, No. 11 / June 1, 2007
Optical coherence tomography with plasmon resonant nanorods of gold Timothy S. Troutman, Jennifer K. Barton, and Marek Romanowski* University of Arizona, Biomedical Engineering, 1501 North Campbell Avenue, Tucson, Arizona 85724, USA *Corresponding author:
[email protected] Received January 2, 2007; revised March 11, 2007; accepted March 28, 2007; posted April 2, 2007 (Doc. ID 78575); published May 1, 2007 We explored plasmon resonant nanorods of gold as a contrast agent for optical coherence tomography (OCT). Nanorod suspensions were generated through wet chemical synthesis and characterized with spectrophotometry, transmission electron microscopy, and OCT. Polyacrylamide-based phantoms were generated with appropriate scattering and anisotropy coefficients (30 cm−1 and 0.89, respectively) to image distribution of the contrast agent in an environment similar to that of tissue. The observed signal was dependent on whether the plasmon resonance peak overlapped the source bandwidth of the OCT, confirming the resonant character of enhancement. Gold nanorods with plasmon resonance wavelengths overlapping the OCT source yielded a signal-to-background ratio of 4.5 dB, relative to the tissue phantom. Strategies for OCT imaging with nanorods are discussed. © 2007 Optical Society of America OCIS codes: 170.4500, 160.3900, 290.5850.
Optical coherence tomography (OCT) is an interferometric imaging technique based on the detection of backscattered near-infrared light. This technique is particularly promising for biomedical imaging applications because it provides noninvasive imaging of living tissues, reasonable penetration depth 共⬃2 mm兲 and high spatial resolution 共⬃10 m兲, superior to present clinical methods of noninvasive imaging such as computed tomography, magnetic resonance imaging, and ultrasound. To address different imaging requirements, OCT can be configured as an in-air system, or the optic pathway can be directed to the sample through a catheter or endoscope [1,2]. Development of contrast agents provides enhancement of OCT with the potential benefit of specific targeting of molecular markers of diagnostic importance [3,4]. Since the signal generated in OCT is based on refractive index mismatches in the sample, one type of suitable agent is a material with large refractive index variations, as exemplified by gas-filled microbubbles [5] or micrometer-sized droplets of oil with embedded silica particles on the interface [6]. Nanoparticles of gold are particularly attractive materials for contrast enhancement in biomedical imaging applications. Such nanoparticles have very strong optical resonances responsible for the vivid colors of their suspensions. These unique plasmon resonances, coupled with the biocompatibility of gold, lead to numerous applications relevant to biomedical research. A concept of paramount importance for such applications is the tunability of optical resonances, whereby structures resonant at selected wavelengths can be generated to address experimental requirements such as illumination source or optical properties of the sample. Highly scattering gold nanoshells based on silica spheres have been shown to enhance OCT contrast in vivo [7] and subsequent attachment of antibodies for the endothelial growth factor receptor resulted in specific targeting of cancer cells [8]. Plasmon resonant gold nanocages of approximate size of 40 nm were synthesized, conjugated with antibodies and targeted to breast cancer 0146-9592/07/111438-3/$15.00
cells [9]. Dumbbell-shaped nanorods of gold have been shown to produce OCT signal with a minimum detectable concentration of 25 g Au/ml in a homogeneous 1.1% suspension of intralipid [10]. Here we describe an application of plasmon resonant nanorods of gold as a contrast agent for OCT imaging. To demonstrate OCT imaging we used a polyacrylamide gel phantom, where optical properties of the gel and spatial distribution of nanorods can be controlled. Cylindrical gold nanorods of various aspect ratios (length-to-diameter) were prepared in a manner similar to that described by Mulvaney [11]. Briefly, gold seed solution was prepared and added in quantities ranging from 6 to 50 l to aliquots of the growth solution containing surfactant and gold chloride. The obtained gold nanorods were redispersed in water, in 5 ml (50 g Au/ml) for spectrophotometric analysis or 0.2 ml (1250 g Au/ml) for transmission electron microscopy (TEM) and OCT. Extinction spectra of nanorod suspensions were taken with a Cary 5E spectrophotometer in the range of 400 to 1350 nm in double beam mode against water. Each spectrum consists of two peaks, one near 530 nm corresponding to the transverse dimension of rods and spherical particles, while the location of the other peak varies with the quantity of added gold seed and corresponds to the longitudinal dimension of rods (Fig. 1). An integrating sphere accessory was used to collect total transmission, scattered transmission and reflectance spectra. These data in combination with Inverse Adding Doubling calculation software [12] were used to determine scattering and absorption coefficients as well as scattering anisotropy. For the peak of longitudinal resonance of a nanorod suspension at 975 nm s , a, and g were 11.8 cm−1, 9.7 cm−1, and −0.0362, respectively. All solutions used were found to have stable spectra for several weeks after imaging; thus there was no evidence of spectral shifts due to aggregation. For TEM 2 l drops of the concentrated suspensions were allowed to dry completely on the surface of a thin, transparent carbon sheet supported by a 200 © 2007 Optical Society of America
June 1, 2007 / Vol. 32, No. 11 / OPTICS LETTERS
Fig. 1. Series of nanorod suspensions characterized in this study. The positions of the plasmon resonance peaks in the red and near-infrared inversely relate to the amount of seed added to initiate growth: (in l) 6 (circle), 8 (⫻), 15 (⫹), 20 (square), 30 (diamond), 40 (inverted triangle), 50 (triangle).
mesh copper grid and were imaged in a Jeol 100CX transmission electron microscope. The most striking feature of the produced nanorods is their highly symmetric cylinderlike shape with a fairly constant diameter (approximately 20 nm) and an average length varying from one preparation to the next (Fig. 2). In general, nanorod suspensions exhibiting plasmon resonance peaks in longer wavelengths (nearinfrared) were composed of a population of higher aspect ratios in electron microscopy images. Artifacts of the production of nanorods include small populations of other shapes of gold nanoparticles. We used a time-domain OCT system similar to one previously described [13]. The OCT system source is composed of two superluminescent diodes. The center wavelength for the system when both diodes are active is 890 nm, with a bandwidth of 150 nm. The resolution of the images produced by the system is 6 m axially and 14 m laterally. A polyacrylamide tissue phantom was prepared to characterize the nanorod suspensions with the OCT system. Compared with intralipid suspensions, polyacrylamide-based phantoms allow for independent adjustment of scattering coefficient and anisotropy by adding polystyrene beads of appropriate size and concentration. Here we prepared gels with 1 m diameter polystyrene spheres at a concentration of one sphere per 1100 m3 to obtain s of 30 cm−1 and g of 0.89. The gels were cast between glass plates to produce layers approximately 400 m thick. When desired, a small void with an approximate volume of
Fig. 2. Examples of TEM of nanorod suspensions resonant at (a) 750 nm and (b) 912 nm. The marker bars indicate 100 nm (original magnification 40,000⫻).
1439
1 l was formed in the gel by placing a nylon tape on the casting plate. Once cast, the gel was cut into small slabs and multiple layers stacked together to form the tissue phantom. The voids were filled with a suspension of nanorods. To avoid the specular reflection from the air–phantom interface the phantoms were tilted at a small angle relative to the inbound beam. Presented in Fig. 3 are cross-sectional images of the tissue phantom. The signal in these OCT images is indicative of the intensity of backreflection at a given pixel. The 1 l voids, clearly seen in these images, were filled with water [Fig. 3(a)], gold nanorods resonant at 750 nm, a wavelength beyond the spectral envelope of the OCT source [Fig. 3(b)], or gold nanorods resonant at 912 nm, closely matching the center of the spectral distribution of the OCT source [Fig. 3(c)]. Each nanorod suspension used here contained the same molar quantity of gold; voids contain approximately 1.25 g of gold. Capillary action between the polyacrylamide sheets caused a small amount of contrast agent to flow out of the voids and coat some of the interfaces between polyacrylamide slabs, providing a thin, observable layer of agent on these surfaces. Saturation effects, seen as vertical lines, are the result of strong reflections encountered from highly reflective regions. Dramatic improvement of the signal is noticed in both the bulk and interfacial presence of the nanorods resonant at 912 nm, matching the center of the OCT source, relative to the signal obtained with nanorods resonant at
Fig. 3. OCT images of different preparations loaded into the polyacrylamide tissue phantom: (a) water, (b) nanorods with resonance at 750 nm, (c) nanorods with resonance at 912 nm. Intensity profiles of the boxed regions are shown at the right and represent average signal intensity (0–255 range, horizontal axis) versus depth (vertical axis). Image dimensions are 3.2 mm by 2 mm.
1440
OPTICS LETTERS / Vol. 32, No. 11 / June 1, 2007 Table 1. Average Pixel Intensities in OCT Images Average Pixel Intensitya Image and Sample
3(a) 3(b) 3(c)
Water Nanorods, 750 nm Nanorods, 912 nm
Void
Lower Layer
Upper Layer
dB Enhancementb
44.2 61.2 72.2
50.9 51.4 49.7
54.7 54.1 58.1
−1.3 1.9 4.5
a
Pixel intensity is averaged over a representative 0.4 mm2 rectangular area of the images shown in Fig. 3. Calculated as 10⫻ log10 共void average pixel intensity/lower layer average pixel intensity兲.
b
750 nm (Fig. 3). Both nanorod suspensions show enhanced signal compared with water. The intensities obtained with nanorods also exceed the intensity of backscattered light from the tissue phantom, and this enhancement is particularly pronounced with nanorods spectrally matching the center of the OCT source (Table 1). In this report we demonstrated contrast enhancement in OCT images using plasmon resonant nanorods of gold. The key features of gold nanorods of interest to OCT imaging are their sharp optical resonance and relatively large reduced scattering coefficient, s共1 − g兲. The full width at half-maximum can be as little as 80 nm and is typically around 120 nm, whereas the value of scattering coefficient exceeds that of absorption coefficient. We obtained significant enhancement of contrast with gold nanorods resonant at a wavelength within the spectral bandwidth of the OCT source. Gold nanorods resonant at the wavelengths outside the spectral envelope of the OCT source contribute little to signal. This observation points toward the possibility of spectral multiplexing in contrast-enhanced OCT imaging, where contributions of contrast agents resonant at various wavelengths can be resolved. This approach is of particular importance in applications where contrast agents are targeted to various molecular markers of interest. We propose that further improvement of contrast can be obtained by differential OCT imaging. In this approach two images at slightly shifted OCT center wavelengths are collected so both images of the tissue (typically containing no resonant structures) remain similar, whereas the contrast agent is visible in one but not the other. In comparison, other nanostructures such as nanoshells have much broader scattering bands [8]. Despite some similarity of shapes, gold nanorods of dumbbell shape appear to be primarily absorbing, underlining the importance of shape for resultant optical properties of gold nanostructures [10]. The very sharp resonances of nanocages are predominantly absorptive [9], producing limited backscattering signal for OCT imaging. In our work we did not attempt to establish the detection limit. It appears, however, that the amount of gold in the imaged volume [Fig. 3(c)], defined as the cross-sectional area of the void times the lateral resolution of the OCT system, is less than 3 ng. Our interest in these contrast agents stems from our long-term goal to develop methods of molecular
imaging with OCT. Molecular targeting can conveniently be introduced by molecules carrying groups such as thiol or amine, which readily coordinate with gold. Targeting ligands, e.g., antibodies or small molecules introduced in this manner will direct nanorods to a molecular marker of interest, such as membrane proteins overexpressed in cells undergoing malignant transformation. Currently, OCT is capable of interrogating sections of a tissue with the penetration depth of 1 – 2 mm, a capability often described as optical biopsy. OCT augmented by the targeted contrast agents may noninvasively produce functional information of diagnostic importance similar to that obtained by the methods of immunohistochemistry. Properties of gold nanorods described here enable improved discrimination of signals generated by the contrast agent, a necessary step toward the development of molecular imaging using OCT. This research was supported by the Arizona Biomedical Research Commission (M. Romanowski). References 1. D. Huang, E. A. Swanson, C. P. Lin, J. S. Schuman, W. G. Stinson, W. Chang, M. R. Hee, T. Flotte, K. Gregory, C. A. Puliafito, and J. G. Fujimoto, Science 254, 1178 (1991). 2. A. G. Podoleanu, Br. J. Radiol. 78, 976 (2005). 3. C. Yang, Photochem. Photobiol. 81, 215 (2005). 4. S. A. Boppart, A. L. Oldenburg, C. Xu, and D. L. Marks, J. Biomed. Opt. 10, 041208 (2005). 5. J. K. Barton, J. B. Hoying, and C. J. Sullivan, Acad. Radiol. 1, S52 (2002). 6. T. M. Lee, A. L. Oldenburg, S. Sitafalwalla, D. L. Marks, W. Luo, F. J. Toublan, K. S. Suslick, and S. A. Boppart, Opt. Lett. 28, 1546 (2003). 7. J. K. Barton, N. J. Halas, J. L. West, and R. A. Drezek, Proc. SPIE 5316, 99 (2004). 8. C. Loo, A. Lin, L. Hirsch, M. Lee, J. Barton, N. Halas, J. West, and R. Drezek, Technol. Cancer Res. Treat. 3, 33 (2004). 9. H. Cang, T. Sun, Z. Li, J. Chen, B. J. Wiley, Y. Xia, and X. Li, Opt. Lett. 30, 3048 (2005). 10. A. Oldenburg, D. A. Zweifel, C. Xu, A. Wei, and S. A. Boppart, Proc. SPIE 5703, 50 (2005). 11. J. Perez-Juste, L. M. Liz-Marzan, S. Carnie, D. Y. C. Chan, and P. Mulvaney, Adv. Funct. Mater. 14, 571 (2004). 12. S. A. Prahl, M. J. C. van Gemert, and A. J. Welch, Appl. Opt. 32, 559 (1993). 13. J. K. Barton, F. Guzman, and A. Tumlinson, J. Biomed. Opt. 9, 618 (2004).
