May 15, 2012 / Vol. 37, No. 10 / OPTICS LETTERS
1625
Noncontact all-optical measurement of corneal elasticity Chunhui Li,1,2 G. Guan,1,2 Z. Huang,2 M. Johnstone,3 and R. K. Wang1,3,* 1
Department of Bioengineering, University of Washington, 3720 15th Avenue NE, Seattle, Washington 98195, USA 2
Division of Mechanical Engineering and Mechatronics, University of Dundee, Dundee DD1 4HN, Scotland, UK 3
Department of Ophthalmology, University of Washington, 325 9th Avenue, Seattle, Washington 98104, USA *Corresponding author:
[email protected] Received February 22, 2012; revised March 28, 2012; accepted March 28, 2012; posted March 30, 2012 (Doc. ID 163495); published May 7, 2012
We report on a noninvasive and noncontact all-optical method to measure the elasticity of the cornea. We use a pulsed laser to excite surface acoustic waves (SAW) that propagate on the corneal surface, then use a phase-sensitive optical coherence tomography system to remotely record the SAWs from which the corneal elasticity is estimated. In addition, the system is able to provide real-time tomographic images of the cornea being examined, an important consideration for clinical studies. While precisely maintaining a range of intraocular pressures (IOP), a series of measurements is performed on ex vivo intact primate eyes. The measurement results not only demonstrate the feasibility of the proposed system to remotely measure the corneal elasticity, but also suggest a strong correlation between the corneal stiffness and the true IOP. © 2012 Optical Society of America OCIS codes: 170.4460, 240.6690, 170.0170.
Accurate knowledge of the mechanical property of the cornea is important for a number of clinical applications in ophthalmology, including early detection of corneal ectatic disease [1], patient-specific biomechanical optimization in corneal and kerato-refractive surgery [2], improved accuracy in the measurement of intraocular pressure (IOP) [3], and better characterization of corneal wound healing [4]. Thus, there is demand for techniques that are capable of accurately assessing the corneal elasticity (Young’s modulus) in situ in a nondestructive fashion. To meet this demand, a number of useful techniques have been developed to evaluate the mechanical properties of the cornea. The ultrasound elasticity microscope is widely used to provide strain imaging induced by a compression plate on the corneal surface of eyes ex vivo [5]. In addition, real- time ultrafast and high-resolution ultrasonic systems have been reported to perform supersonic shear imaging of the cornea, leading to quantitative mapping of corneal elasticity [6]. Although promising, the ultrasonic-based approaches require physical contact between the transducer and the cornea, which is often not desirable in clinical situations. Furthermore, the contact-induced stresses complicate the interpretation of final results. In the current clinical environment, the most popular method for evaluating corneal stiffness is by ocular response analyzer (ORA) [7]. The ORA is based on the generation of a noncontact transient air-pulse and a dynamic electro-optical system used to record two applanation pressures at the surface of the cornea, providing a parameter called corneal hysteresis (CH) that is believed to correlate with the IOP and the corneal stiffness. The use of the CH as an indicator for corneal stiffness has proven to be inaccurate [8] because this parameter is known to be confounded by the coupling between the true IOP and corneal viscoelastic properties, and by the dynamic corneal geometry (e.g., thickness), a parameter that ORA is unable to provide. A recent report using state-of-the-art optical coherence tomography (OCT) to image the depthresolved dynamic deformation of cornea induced by the rapid air-pulse [9] sheds a new light into ORA as a tool for estimating the mechanical properties of the cornea. However, the necessity of applanation still requires large cornea displacements (>1 mm). Such displacements un0146-9592/12/101625-03$15.00/0
doubtedly make the cornea respond in a nonlinear fashion to the mechanical stress. Additionally, patients may experience considerable discomfort. Being simple, noncontact, and nondestructive, laserinduced surface acoustic waves (SAW) have been widely used in metrology to estimate the elastic properties of industrial materials [10]. However, its application to biological tissues is less explored. Recently, our group pioneered the use of laser-induced SAW to evaluate Young’s modulus of heterogeneous soft tissue phantoms [11]. By recording the SAW using a low-coherence interferometer at two locations with a known separation distance, the SAW velocity can be evaluated, from which the elasticity of the sample can be calculated using a procedure involving inverse calculations. The advantages of using the SAW approach include: (1) SAW excitation and measurement are all done optically, and (2) the deformation experienced by the sample is less than one micrometer and is typically less than 100 nm. This Letter explores the utility of the laser-induced SAW to measure the corneal elasticity. We use a phase-sensitive OCT (PhS-OCT) to record remotely the SAWs at the corneal surface, from which the corneal elasticity is deduced. A benefit PhS-OCT offers, in addition to being noninvasive and noncontact, is real-time simultaneous tomographic imaging of the cornea. The feasibility of the proposed all-optical approach is tested on ex vivo intact primate (monkey) eyes under conditions of controlled IOP. The schematic of the system setup used in this study is shown in Fig. 1. The system includes three main parts: SAW generation, SAW detection, and a perfusion system to precisely control IOP. A pulsed and solid-state Nd:YAG laser (532 nm) was used to generate SAWs propagating at the surface of the cornea. The laser pulse was set to a duration of 6 ns with an average energy of ∼2–2.5 mJ and repetition rate of 1 Hz (1 sec). A cylindrical lens was employed to generate a line source impinging onto a region near the limbus, where a thin layer of black ink was applied in order to facilitate the SAW generation because of the strong absorption of light by the ink. The SAWs were detected by a PhS-OCT system, similar to that reported previously [12,13]. Very briefly, the PhSOCT worked in a spectral domain configuration in which © 2012 Optical Society of America
1626
OPTICS LETTERS / Vol. 37, No. 10 / May 15, 2012
Fig. 1. Schematic of system setup consisting of three main parts: laser SAW generation, PhS-OCT detection, and perfusion system to precisely control IOP of the sample.
The IOP of the eyes was precisely controlled by a mechanical perfusion system that consisted of two perfusion reservoirs (see the lower part of Fig. 1). Reservoir 1 infused fluid into the anterior chamber by means of tubing connected to a needle that passed though the limbus into the chamber. A second reservoir led through tubing to the vitreous chamber. An additional needle inserted into the anterior chamber at the limbus was attached through tubing to a pressure transducer, assuring continuous on-line recording of the true IOP within the eye. Each eye was perfused at one of five different pressures (5, 7, 10, 15, and 20 mm Hg). Initially, the eye was subjected to the loading force of pressure by increasing IOP stepwise from 5 to 20 mm Hg, and followed by unloading stepwise back to 5 mm Hg. The loading experimental protocol enabled us to study the dynamic response of the cornea to IOP and its influence on corneal mechanical properties. With the OCT probing beam positioned at each measuring location on the cornea, and the perfusion system controlled at one set value of IOP, the detection system acquired ten repeated M-mode OCT signals, which were then averaged to give the final SAW with improved signalto-noise ratio (SNR). The SAW group velocity was then calculated by the known distance between two adjacent locations divided by the time gap of the detected SAW energy peaks. The SAW group velocity (C R ) gives the general mechanical properties of the material. The relationship between C R and corneal elasticity can be approximated as p (1) C R 0.87 1.12υ E∕2ρ1 υ∕1 υ;
a superluminescent diode with a central wavelength of 1310 nm and a spectral bandwidth of 46 nm was used as the probing light source. The system had axial and transverse resolutions of ∼16 and ∼25 μm (in air), respectively. The sampling/imaging rate was set at ∼47; 000 A-lines per second, sufficient to capture the SAW signals (up to 23.5 kHz). With this imaging rate, the system had a measured dynamic range of ∼100 dB at a 0.5 mm depth. The system provided a phase-sensitivity to the sample displacement of ∼50 pm when working in M-scan mode [14]. The roles PhS-OCT played here (Fig. 1) are two-fold: (1) to provide real-time imaging of the cornea that enables the visualization and quantification of 3D geometry of the sample (Fig. 2), and (2) to detect the SAWs at the surface of the cornea that enable the measurement of corneal mechanical property. When in the SAW detection mode, the scanner in the OCT system steered the probe beam step by step to locations 1 to 3 mm away from the excitation beam, with a step distance of 0.5 mm (Fig. 2). At each step, the probe beam was kept stationary for ∼1 sec while the OCT continuously captured the interference signals, i.e., M-mode scanning. A computer was used to synchronously control the acquisition of surface wave signals, the laser pulse excitation, and OCT imaging. To test the feasibility of the proposed system of measuring the corneal elasticity and to determine the relationship between elasticity and the true IOP, we used enucleated primate eyes that were obtained from the University of Washington Primate Center within 5 hours of death.
where E denotes Young’s modulus, ν is Poisson’s ratio, and ρ is the mass-density of the sample. Fig. 3 shows typical laser-induced SAW waveforms recorded at two different IOP settings of 5 mm Hg (left) and 15 mm Hg (right), respectively. Typical phase due to SAW as measured by the PhS-OCT system is ∼1–2 radians, corresponding to a displacement of ∼50–100 nm. Fig. 3 demonstrates that the SAW is moving away from the excitation location because the arrival time of the SAW becomes longer as the probe beam is moved farther away. The waveform dispersion is observed to be minimal, indicating that the recorded SAW is most likely dominated by the stromal layer of the cornea that may be considered as mechanically homogeneous. Most importantly, the SAW travels much faster at an IOP of 15 than at 5 mm Hg, demonstrating that the cornea becomes stiffer (higher Young’s modulus) when the IOP is higher. This
Fig. 2. (Color online) Typical OCT cross-section image of the sample and a brief sketch of the locations of laser excitation beam and OCT probe beams for measuring SAWs.
Fig. 3. Typical laser-induced SAWs recorded by PhS-OCT at the cornea surface with the IOP controlled at 5 mm Hg (left) and 15 mm Hg (right).
May 15, 2012 / Vol. 37, No. 10 / OPTICS LETTERS
observation agrees very well with the mathematical modeling framework in which the mechanical property was predicted to vary with true IOP, such that a stiffer cornea is manifested at higher levels of true IOP [7]. The measured SAW group velocities and the corresponding Young’s moduli of a cornea examined during one cycle of loading and unloading in response to true IOP values are shown in Fig. 4. To estimate Young’s modulus, Eq. (1) was used by assuming that Poisson’s ratio is 0.49 and density is 1000 kg∕m3 for monkey corneas. The estimated values of Young’s moduli are within the range reported previously (between 200 and 2000 kPa) under the similar experimental conditions [5,6,15]. We observed the SAW group velocity [Fig. 4(a)] and the estimate of Young’s modulus [Fig. 4(b)] are strongly correlated with the true IOP. The corneal elasticity increases nonlinearly with the increased loading of the true IOP. While the elasticity decreases during the IOP unloading, it does not follow the loading trajectory. It is well known that viscoelastic material has both viscosity and elasticity, resulting in energy dissipation in such a material when a stress is applied. Energy dissipation leads to hysteresis during a stress—strain cycle [8]. As the cornea is a typical viscoelastic tissue, this explains the hysteresis in Fig. 4 during a loading and unloading cycle of IOP in our experiment. The proposed system provides both remote excitation and remote detection of the SAW. This noncontact, noninvasive system may thus be a useful tool for the investigation of mechanisms by which true IOP affects corneal stiffness. The system also offers the ability to assess corneal mechanical responses to corneal interventions, such as kerato-refractive surgery and IOP elevation associated with glaucoma. Notably, the proposed system is capable of providing, via OCT, real-time tomographic images of the examined cornea, making simultaneous quantification of the elastic modulus and geometrical parameters, such as corneal thickness, possible. Theoretical modeling [7,8] indicates that the accuracy of measured true IOP is not only dependent on the mechanical properties of the cornea, but also on the corneal thickness and curvature. Such complex relationships currently prevent reliable measurement of the true IOP in clinical settings [8]. The proposed technology may permit better approximation of the true IOP by applying appropriate modeling to simultaneous measurements of corneal elasticity and geometry. The maximum measurable SAW group velocity is 23.5 m∕s, limited by the imaging speed (47 kHz) of the OCT detection system. This limitation prevents measurement of elasticity of the epithelial layer. A sampling rate of at least 100 kHz would be required to record the SAW,
Fig. 4. Estimated SAW group velocities and corresponding Young’s moduli during one cycle of IOP loading and unloading. The error bars denote standard deviation.
1627
since the thickness of the epithelial layer is typically 20–50 μm. A state-of-the-art OCT system has been reported with an imaging rate of >100; 000 A-lines per second [16]. Thus, in the future, if such an ultrafast OCT is employed to record the SAWs using our proposed system, measurement of mechanical property of the epithelial layer will be achievable. In the current system, we used a pulsed 532 nm light source to generate the SAW. A thin layer of black ink was applied to the surface of cornea because the cornea has low absorption at the 532 nm wavelength. This wavelength is clearly not ideal for in vivo corneal studies. This problem may be resolvable by employing an ultraviolet or else an infrared laser source, such as those routinely used in a clinical environment. In addition, the total measurement time was about 50 s, largely limited by the repetition rate of the laser system (1 Hz). However, a kilohertz repetition rate pulsed laser is now commercially available that is suitable for generating SAW within tissue. Thus, we expect that real-time measurement of elasticity would be feasible if this newer laser is used. We have demonstrated that the combination of laser-induced SAW and PhS-OCT can be successfully utilized to estimate the corneal elasticity under stepwise IOP changes. We have shown that Young’s modulus of the cornea varies with the true IOP and experiences a hysteresis during the IOP loading and unloading cycle. Contrary to ORA, the only method used today in clinical practice, our method gives access to the real Young’s modulus and simultaneous assessment of the geometrical properties of the examined cornea. Because the proposed system utilizes an OCT device as the SAW detector, integration of the device into current OCT systems can be envisioned. References 1. W. J. Dupps, Jr., J. Refract. Surg. 21, 186 (2005). 2. I. F. Comaish and M. A. Lawless, J. Cataract Refract. Surg. 28, 2206 (2002). 3. J. Liu and C. J. Roberts, J. Cataract Refract. Surg. 31, 146 (2005). 4. M. R. Bryant, K. Szerenyi, H. Schmotzer, and P. J. McDonnell, Invest. Ophthalmol. Vis. Sci. 35, 3022 (1994). 5. G. Wollensak, E. Spoerl, and T. Seiler, J. Cataract Refract. Surg. 29, 1780 (2003). 6. M. Tanter, D. Touboul, J. Gennisson, J. Bercoff, and M. Fink, IEEE Trans. Med. Imaging 28, 1881 (2009). 7. G. Orssengo and D. Pye, Bull. Math. Biol. 61, 551 (1999). 8. A. Kotecha, Surv. Ophthalmol. 52, S109 (2007). 9. D. Alonso-Caneiro, K. Karnowski, B. J. Kaluzny, A. Kowalczyk, and M. Wojtkowski, Opt. Express 19, 14188 (2011). 10. A. Neubrand and P. Hess, J. Appl. Phys. 71, 227 (1992). 11. C. H. Li, Z. H. Huang, and R. K. K. Wang, Opt. Express 19, 10153 (2011). 12. C. Li, G. Guan, R. Reif, Z. Huang, and R. K. Wang, J. R. Soc. Interface 9, 831 (2012). 13. P. Li, L. An, T. T. Shen, M. Johnstone, and R. K. Wang, Biomed. Opt. Express 2, 3109 (2011). 14. R. K. Wang and A. L. Nuttall, J. Biomed. Opt. 15, 056005 (2010). 15. N. E. K. Cartwright, J. R. Tyrer, and J. Marshall, Invest. Ophthalmol. Vis. Sci. 52, 4324 (2011). 16. L. An, P. Li, T. T. Shen, and R. K. Wang, Biomed. Opt. Express 2, 2770 (2011).
