We present a three-dimensional structured tissue-mimicking phantom for use in optical coherence tomography (OCT). The phantom was fabricated from a silicone matrix and titanium dioxide additive using a lithographic casting method capable of producing a wide range of well-defined geometries with optical contrast and mesoscopic feature sizes relevant to OCT. We describe the fabrication, characterization and OCT imaging of two phantoms and demonstrate their utility in assessing the performance of a spatial-diversity speckle reduction technique. Such phantoms will be important in the development of standards in OCT, as well as in enabling quantitative performance assessment.
©2011 Optical Society of America
Tissue-mimicking objects, known as phantoms, are important tools in biomedical imaging for calibration, development of standards, investigation of new techniques, and validation of theoretical predictions . Although numerous materials have been proposed for phantoms, including silicone , polyvinyl alcohol (PVA) gels  and fibrin , to date, the vast majority of them have lacked a well-defined three-dimensional (3D) structure. Phantoms have featured either homogeneous optical scattering or consisted of simple layers  with no 2D or 3D structure. The absence of known features has made it difficult to quantify the performance of 3D optical imaging techniques such as optical coherence tomography (OCT) . Phantoms containing 2D geometrical variation, including wave-like structures and narrow linear channels, have recently been demonstrated. These phantoms are based on a simple modular fabrication method and represent a significant advance , but the geometries that can be realized are limited. A flexible fabrication method for 3D phantoms with structure of variable but known size (and spatial frequency content) on a “mesoscopic” scale, i.e., intermediate between the resolution and range of the system, would make feasible the realization of a wide range of phantoms. Such phantoms will enable the assessment of OCT feature contrast and resolution throughout an imaging volume under realistic scattering conditions, as well as, for example, their respective enhancement and degradation, when speckle reduction techniques  are applied.
In this paper, we present, to our knowledge, the first versatile technique for the fabrication of mesoscopic structured 3D phantoms for OCT. Fabrication is based on a master produced by lithography, which is used in a two-step casting procedure employing a silicone elastomer, which provides a transparent matrix, and titanium dioxide (TiO2) particles, which provide optical scattering. This method provides well-defined, reproducible and independent control over structure and optical properties. We describe the phantom’s design and fabrication, its characterization using OCT, and demonstrate its use in assessing feature contrast enhancement and the associated effects on image resolution caused by focus-compounding speckle reduction .
The soft lithography technique known as replica molding  was used to fabricate the phantom. Replica molding involves fabricating a master and using it to cast the phantom. Below, we describe the phantom’s design, materials and fabrication.
The structured 3D phantoms were designed to form a rectangular block (with nominal dimensions 20 mm × 30 mm × 4 mm) constructed from two castings, shown schematically (not to scale) in blue and grey in Fig. 1(a) . The 3D features, the letters “O B E L”, protrude from the first (feature) casting (blue in Fig. 1(a)) and are encased on both sides by the second (embedding) casting (gray in Fig. 1(a)). The letter dimensions ranged from 250 μm in height h to between 180 μm and 220 μm in width. The smallest features were 45 μm wide and the spacing between the letters was 50 μm. The feature protrusion length d was designed to be nominally 250 μm. The depth z 0 from the top surface of the casting to the features was designed to be nominally 100 μm.
The phantoms were designed to be imaged from the top and the dimensions were chosen so that, for the intended range of scatterer concentrations, the bottom of the letters could be imaged. The yellow dashed outline in Fig. 1(a) indicates the orientation of an OCT B-scan.
The phantom materials comprised a silicone elastomer containing a homogeneous distribution of optical scatterers. The elastomer was a two-part polydimethylsiloxane (PDMS) silicone (Dow Corning, Sylgard® 184 silicone elastomer) comprising a liquid pre-polymer (Part A) and a curing agent (Part B), employed in a 10:1 ratio. Optical scattering was introduced using TiO2 particles (Sigma-Aldrich). The desired concentration of TiO2 particles was injected into Part A of the silicone, manually mixed with a glass stirring rod, and then placed in an ultrasonic bath for 90 minutes to ensure an evenly distributed suspension before mixing with Part B.
Two phantoms were fabricated based on the design described above, both with a TiO2 particle concentration in the feature casting of 10 mg/mL. The first phantom (Phantom I) contained no scatterers in the embedding casting. The second phantom (Phantom II) was fabricated with a 3 mg/mL concentration of TiO2 particles in the embedding casting except inside the holes of the letter “B”. These holes were filled with transparent epoxy of similar refractive index to the elastomer (n = 1.41), thus, providing an area of transition from very low to high scattering. The scattering coefficients calculated using Mie theory were 13.4 mm−1 for the feature casting and 4.3 mm−1 for the embedding casting, which are representative of those found in tissue.
The 250-μm deep-feature master was produced via UV photolithography using a negative photoresist (MicroChem SU-8 50). The process involved a laser-printed photolithographic mask (chrome layer on a soda-lime glass substrate), on which the letters “OBEL”, reversed, were formed from the chrome material not removed by the laser (2 μm minimum feature size, 200 nm repeatability (tolerance)). The laser printing sets the overall minimum feature size (resolution) of the phantom. The photoresist was spun onto a 3” silicon wafer to obtain the required thickness and then placed in contact with the photolithographic mask, which acted as a spatial filter. Development of the photoresist followed timed exposure to UV light at 365 nm, resulting in a master with the negative imprint of the desired features.
The master was clamped onto a Teflon frame of dimensions 5 mm × 10 mm × 3 mm, to which the first silicone mix was added (blue in Fig. 1(a)). The master and the frame were previously hydrophobized with hexamethyldisil azane (HMDS) for easy cast extraction. The frame allowed for outgassing and curing, providing castings free of air bubbles and scatterer clumps, whilst maximizing adhesion to the master. The feature casting was then placed in a separate larger outer frame, also allowing for outgassing and curing, and a second silicone mix was added (the embedding casting). For the feature casting, outgassing was performed at 80 kPa at room temperature for 16 hours and then curing at 65°C for 3.5 hours. For the embedding casting, outgassing was performed at 80 kPa at room temperature for 2 hours and then at 65°C for 2.5 hours. A longer outgassing and curing time for the feature casting was used to increase its hardness, minimizing the risk of tearing the features when extracting the casting from the master.
Figure 1(b) shows a photo of Phantom I beside an Australian 5 cent coin. The feature casting topography was measured after the first step and before the second step, using an optical profilometer (Veeco NT9100) with z resolution 0.1 nm, x-y resolution 0.49 μm, and repeatability 0.05 nm. The result is shown in Fig. 1(c). The measured feature protrusion length d is 270 μm, which exceeds the design value by 20 μm, due to an exposure irradiance slightly higher than optimal. In Figs. 1(d) and 1(e), photomicrographs of different planes of Phantom I are presented. It is apparent from these images that the substrate and letter surfaces are not perfectly parallel and that the letter spacings are very slightly undercut, which is less apparent from the profilometry.
2.4 OCT imaging
3D OCT images of the phantoms were acquired using a swept-source OCT system (Thorlabs OCS1300SS) with a central wavelength of 1325 nm and a bandwidth of 100 nm, with a theoretical axial resolution of 7.7 μm in air, theoretical lateral resolution of 7.5 μm, and a numerical aperture of 0.056. The phantoms were imaged from the top, in free space, with the beam focus set to the top of the letters. The images are displayed on an 8-bit logarithmic grayscale corresponding to a reflectivity dynamic range of approximately 60 dB.
3. Results and discussion
Figures 2(a) -2(c) show three cross-sectional planes from the 3D-OCT data set for Phantom I. Figure 2(d) shows the orientation and location of these planes with respect to the features. Figure 2(d) (Media 1) shows a sequential multiplanar view and a fly-through of the 3D solid rendering of the phantom. Figures 2(e) and 2(f) show volumetric reconstructions of OCT data sets for Phantoms I and II, respectively. These visualizations demonstrate the full 3D structure of the phantom. Close inspection reveals the speckle patterns within the letters are visible.
The images of Phantom I (Figs. 2 (a)-(e)) reveal that the top of the embedding casting (arrowhead 1 in Fig. 2(b)) is not level with the top of the feature casting (arrowhead 2 in Fig. 2(b)), thus offsetting the depth z 0, nominally 100 μm, by approximately 200 μm. This offset was eliminated in Phantom II. Figure 2(b) reveals a slight meniscus at the top of the feature casting. For Phantom II, z 0 is approximately 100 μm. The bright line visible in Fig. 2(f) is the specular reflection from the top of the embedding casting, i.e., the top of the phantom.
To demonstrate the utility of our phantoms, speckle reduction using focus compounding  was applied to Phantom II. In this technique, decorrelation is introduced between B-scans by varying the axial position of the focus within the sample (20 B-scans each offset by 10 μm were used). Speckle reduction is achieved by incoherently averaging the decorrelated B-scans. Figure 3(a) shows a single B-scan and Fig. 3(b) shows the result of the incoherent averaging. Figure 3(c) shows a photomicrograph for reference. In Figs. 3(d) and 3(e), a close-up of part of the letters “O” and “B” is presented for the single B-scan and speckle-reduced cases. Figure 3(f) shows a plot of the pixel intensities as a function of the lateral position (x-axis) at a depth of 150 μm, corresponding to the red and blue lines in Figs. 3(d) and Fig. 3(e), respectively. A 1.80-fold speckle contrast reduction  was obtained, consistent with the noticeably much smoother blue line. In addition, signal highs (corresponding to the positions of the letters “O” and “B”) and lows (the gaps between the letters) are more clearly defined for the blue curve. The transition between the area with no scattering inside the hole of the letter “B” and the right-hand side of the letter “B”, visible at the lateral position of 250 μm, provides a good step response test for assessing image resolution. The angle of inclination was calculated using the profilometry data to be 89°. The deviation from 90° is due to slight imperfection in the master fabrication. However, the step response is within one lateral resolution element of OCT B-scan, which is adequate.
The relative signal rise time is a quantitative measure of the degradation of spatial (lateral) resolution. The measured 10-90% rise-time ratio of the blue to the red curve is 1.75. Such quantitative assessment of the loss of resolution in speckle reduction techniques  is not possible for biological and other samples with inherently unknown 3D feature sizes.
The fabrication technique presented here is very versatile. The smallest feature size is set by the optical lithography system resolution (2 μm) and the largest feature size can exceed 1 mm. The two-step casting procedure requires carefully manufactured frames. Error in the frame dimensions caused the most prominent deviation of the feature distance, z 0, from the design requirement in Phantom I. The phantoms have an expected two-year lifetime set by the properties of the elastomer. The technique allows for incorporation of other scatterers and absorbers , and provides scope for use of more durable matrices, e.g., polyurethane .
Structured 3D phantoms are likely to be of considerable value in various aspects of OCT. They will be invaluable in the development of calibration procedures [13–16] and standards for inter-comparison . Beyond the quantification of speckle reduction demonstrated here, they could be used in the study of beam propagation and multiple scattering effects  and other imaging artifacts due to beam-sample interactions. By varying the stiffness between the feature and the embedding castings, the phantom could also be used in assessing the performance of optical coherence elastography .
We have presented a novel structured 3D phantom for use in OCT, which was fabricated from silicone and TiO2 particles in a two-stage casting process using a photolithography master. The phantom was designed to exhibit mesoscopic feature sizes and its fabricated dimensions matched well with design to within the tolerance, which was below the OCT system resolution in the x-z plane. We demonstrated the utility of the phantom by quantitatively evaluating the spatial resolution and speckle contrast in a focus compounding speckle reduction scheme. We expect such phantoms to prove invaluable in providing for the first time quantitative reproducible benchmark standard imaging targets for OCT.
The authors acknowledge Mr Simon Doe of the Ian Wark Research Institute at the University of South Australia and Facility Manager of the South Australian node of the Australian National Fabrication Facility (ANFF-SA). The phantom fabrication (master, frames and castings) was performed at the ANFF-SA node and the photolithographic mask was produced at the Optofab node of the ANFF (NSW) under the National Collaborative Research Infrastructure Strategy.
References and links
2. C. E. Bisaillon, G. Lamouche, R. Maciejko, M. Dufour, and J. P. Monchalin, “Deformable and durable phantoms with controlled density of scatterers,” Phys. Med. Biol. 53(13), N237–N247 (2008). [CrossRef] [PubMed]
3. C. U. Devi, R. M. Vasu, and A. K. Sood, “Design, fabrication, and characterization of a tissue-equivalent phantom for optical elastography,” J. Biomed. Opt. 10(4), 44020 (2005). [CrossRef] [PubMed]
4. B. F. Kennedy, S. Loitsch, R. A. McLaughlin, L. Scolaro, P. Rigby, and D. D. Sampson, “Fibrin phantom for use in optical coherence tomography,” J. Biomed. Opt. 15(3), 030507 (2010). [CrossRef] [PubMed]
6. 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, “Optical coherence tomography,” Science 254(5035), 1178–1181 (1991). [CrossRef] [PubMed]
7. D. M. de Bruin, R. H. Bremmer, V. M. Kodach, R. de Kinkelder, J. van Marle, T. G. van Leeuwen, and D. J. Faber, “Optical phantoms of varying geometry based on thin building blocks with controlled optical properties,” J. Biomed. Opt. 15(2), 025001 (2010). [CrossRef] [PubMed]
8. J. M. Schmitt, S. H. Xiang, and K. M. Yung, “Speckle in optical coherence tomography,” J. Biomed. Opt. 4(1), 95–105 (1999). [CrossRef]
9. D. P. Popescu, M. D. Hewko, and M. G. Sowa, “Speckle noise attenuation in optical coherence tomography by compounding images acquired at different positions of the sample,” Opt. Commun. 269(1), 247–251 (2007). [CrossRef]
10. Y. Xia and G. M. Whitesides, “Soft lithography,” Angew. Chem. Int. Ed. 37(5), 550–575 (1998). [CrossRef]
11. A. E. Desjardins, B. J. Vakoc, G. J. Tearney, and B. E. Bouma, “Speckle reduction in OCT using massively-parallel detection and frequency-domain ranging,” Opt. Express 14(11), 4736–4745 (2006). [CrossRef] [PubMed]
14. A. Agrawal, T. J. Pfefer, N. Gilani, and R. Drezek, “Three-dimensional characterization of optical coherence tomography point spread functions with a nanoparticle-embedded phantom,” Opt. Lett. 35(13), 2269–2271 (2010). [CrossRef] [PubMed]
15. P. D. Woolliams and P. H. Tomlins, “Estimating the resolution of a commercial optical coherence tomography system with limited spatial sampling,” Meas. Sci. Technol. 22(6), 065502 (2011). [CrossRef]
16. P. H. Tomlins, G. N. Smith, P. D. Woolliams, J. Rasakanthan, and K. Sugden, “Femtosecond laser micro-inscription of optical coherence tomography resolution test artifacts,” Biomed. Opt. Express 2(5), 1319–1327 (2011). [CrossRef] [PubMed]
17. R. J. Nordstrom, “Phantoms as standards in optical measurements,” Proc. SPIE 7906, 79060H–5,(2011). [CrossRef]
18. T. R. Hillman, A. Curatolo, B. F. Kennedy, and D. D. Sampson, “Detection of multiple scattering in optical coherence tomography by speckle correlation of angle-dependent B-scans,” Opt. Lett. 35(12), 1998–2000 (2010). [CrossRef] [PubMed]
19. B. F. Kennedy, X. Liang, S. G. Adie, D. K. Gerstmann, B. C. Quirk, S. A. Boppart, and D. D. Sampson, “In vivo three-dimensional optical coherence elastography,” Opt. Express 19(7), 6623–6634 (2011). [CrossRef] [PubMed]