Ultra-fast line-field swept source scanning optical coherence elastography

Abstract

The drawbacks of repeated excitations and long imaging time in wave-based optical coherence elastography limited its application in ophthalmology. In this study, we put forward a swept source scanning optical coherence elastography (SSS-OCE) system based on the wavelength-dependence scanning technique. A dispersive element was employed to achieve ultra-fast line scanning from swept light source. Through a series of experiments on agar phantoms with varying concentrations, in situ porcine corneas under different intraocular pressures, and human eyes in vivo, we have demonstrated that SSS-OCE has the capability to rapidly evaluate the elasticity of ocular tissues in vivo with single excitation and offers the advantages of a simple structure and short measurement time.

1. Introduction

Optical coherence elastography (OCE), a technology derived from optical coherence tomography (OCT), offers the capability to acquire biomechanical properties with high resolution, sensitivity, and sub-millimeter precision [1–4]. This technique holds immense significance in the biomedical field for conducting elasticity evaluations in various human tissues, including the cornea [5–8], iris [9], artery [10,11], and skin [12,13], which is crucial for monitoring and understanding disease progression.

The cornea is characterized by distinct structural attributes, namely a relatively large width and a thin thickness. Compared to other techniques, OCE provides high spatial resolution and a wide field of view. These characteristics of OCE make it very beneficial to characterize the elastic changes of corneal tissue. For the quantification of tissue elastic properties, OCE is founded on several underlying principles, specifically vibrational amplitude comparison [14], elastic wave measurement [15,16], and resonant frequency measurement [17,18]. Among them, elastic wave measurement is most widely adopted in relevant research and applications because the vibrational amplitude comparison merely yields relative elasticity data, whereas the procedure of resonant frequency measurement is frequently intricate [19]. In wave-based OCE, the elastic wave is stimulated on the target tissue by piezoelectric transducer (PZT) [20,21], air-puff device [22–25], or acoustic radiation force [26–29]. OCT is employed to detect the propagation velocity of the elastic waves, thereby obtaining the mechanical properties of the tissue [30,31].

However, on account of the imaging methodology of point-scanning for OCT, conventional wave based OCE employs M-B scanning pattern to track the propagation of elastic wave [32–34], which necessitates multiple repeated stimulations on the sample. Repeated stimulation and prolonged imaging time will bring discomfort for subjects in ophthalmic measurements, as well as uncertainty of measurement caused by eye closure and movement. These limitations hinder the widespread adoption of this technology in medical clinical applications, particularly in ophthalmology.

To address this challenge, recent researches taken advantage of high speed mode-locked swept-source laser [35,36], as well as line-field spectral-domain OCT [37], to achieve high acquisition frame rate of OCT and hence single excitation. However, these methods are characterized by high costs for mode-locked swept-source or high-speed area-array camera and complex system designs, which prompt the need for simpler and more effective approaches of fast OCE for ocular application.

In this study, we introduce a novel swept source scanning optical coherence elastography (SSS-OCE) system based on swept source OCT and wavelength dispersion, achieving ultra-fast line field measurement with single excitation and simple system structure. To demonstrate the SSS-OCE system’s capacity of detecting tissue elastic changes, measurements were conducted on agar phantoms with different concentrations and in situ porcine corneas at various intraocular pressures. Ultimately, we conducted SSS-OCE measurements in vivo on five subjects under two distinct intraocular pressure (IOP) conditions because of diurnal variation to confirm the effectiveness of the SSS-OCE system in assessing human eyes.

2. Methods

2.1. Experimental setup

The schematic diagram of the SSS-OCE is depicted in Fig. 1. The SSS-OCE system was modified from a conventional swept source OCT, which utilized a swept-source laser with a bandwidth of 100 nm at 1050 nm and a scan repetition rate of 100 kHz (Axsun Technologies in Massachusetts, USA). Notably, a reflective grating (43-852, Edmund Optics, Arizona, USA) with a groove density of 1200 lines/mm was incorporated into the sample arm after the collimator. Subsequently, the first-order diffracted beam from the grating formed a line-focused beam pattern and scanned while the light source was swept by wavelength. An object lens (AC254-050-AB-ML, Thorlabs, New Jersey, USA) with 50 mm focal length generated the line focus with a length of 6.285 mm at a total power of 1.0 mW on the sample. The returning beams from the two arms first passed through the fiber circulators, then were recombined by a 50/50 optical coupler, and detected by a high speed InGaAs detector for 1060 nm with a bandwidth of 400 MHz (PDB471C-AC, Thorlabs, Newton, New Jersey, USA). A solenoid valve-controlled air pulse system was used for targeted tissue excitation. To generate pulsed excitation forces, a microliter dispensing solenoid valve (The Lee Company, Connecticut, USA) was driven by a single burst sinusoidal wave. The nozzle with an inner diameter of 0.85 mm and an adjustable pointing angle of 45 to 60 degrees was positioned ∼1 cm away from the limbus. In our measurement, the air supply pressure is 0.15 MPa, and the excitation force generated on the ocular surface is about 1.73 mN. Further details about the OCT and air pulse system are provided in our previous study [38,39].

Fig. 1.

The schematic diagram of SSS-OCE system.

When measuring, 512 frames were consecutively acquired after the air pulse excitation, forming a 2D data set of 1024 × 512 points. The 1024 points in each frame represent the axis depth positions of light beam scanning in spatial dimension caused by wavelength-dependent dispersion. 512 points in temporal dimension represent the interference signals at different times in each position. With the line scanning rate of 100 kHz, a single measurement can be completed in 5 ms.

2.2. Tissue mimicking material samples

In this study, agar phantoms (Regular Agarose G-10, BIO WEST, CEDE COMPANY LTD, Hong Kong) with varying concentrations (0.8%, 1.0%, and 1.2% w/w) were prepared using standard methodologies, with the addition of fat emulsions to enhance scattering. For each concentration, three samples were prepared, and each sample was measured eight times to ensure data reliability. To validate the accuracy of the SSS-OCE technique, mechanical pressure-strain testing for the agar phantoms was conducted immediately following the SSS-OCE experiments using a uniaxial mechanical tester (TH-8203A, Suzhou Tophung Machine Equipment Co., Ltd., China). The tensile machine compressed the phantom with a speed of 2 mm/min, the measurement stopped when sample strain reaches 15%, and the built-in software analyzed the stress-strain curve to calculate Young’s modulus. All experiments were conducted at room temperature.

2.3. In situ porcine cornea samples

After validation with phantom, in situ porcine corneas were measured by SSS-OCE. Three porcine eyeballs were obtained from a local slaughterhouse and transported to the laboratory for measurement within 12 hours of removal. The fresh porcine eyes were then stored in 1× phosphate-buffered saline (PBS) solution. To characterize the changes of corneal biomechanical properties under different IOP conditions, an IOP control and monitoring system was used to maintain the IOP of eyeball at a preset value of 10 mmHg, 15 mmHg or 20 mmHg during the measurement, by adjusting the height of the infusion bottle. Measurements were carried out three times within the central corneal region for each eye.

2.4. Participants

To prove the measurement sensitivity of SSS-OCE system for in vivo clinical application, the study focused on the central corneas of five healthy young individuals (two males and three females, aged 24 ± 1 years) recruited from the optometry school of the Wenzhou Medical University. All subjects presented with unremarkable general and ocular health statuses. The exclusion criteria encompassed a history of anterior segment diseases especially KC, subjects with poor ability to maintain fixation, and candidates with blepharophimosis. All procedures were meticulously conducted in strict accordance with the tenets of the Declaration of Helsinki. The study was initiated after the acquisition of informed consent from each individual subject. Throughout the process of image acquisition, the subjects were directed to fixate their gaze on designated visual objects. The study was structured into two separate visits (10:00 AM and 4:00 PM on the same day), and five repetitive SSS-OCE measurements were implemented to each subject for each visit. Before each visit, a non-contact tonometer (TX-20P, Canon Medical Systems, Tokyo, Japan) was used to measure the intraocular pressure and central corneal thickness (CCT) of the subjects to obtain their relevant information.

2.5. Data processing and velocity calculations

As illustrated in Fig. 2, the data preprocessing includes the following steps: (1) Set a sliding window with a width of 64 pixels and step of 5 pixels in spatial dimension and use the sliding window to perform short-time Fourier transform (STFT) on the interference signal. Hence, 193 segments, each containing 32 complex signals, are obtained for each frame data. (2) Consider all segments in the temporal dimension at the same position, constituting a subset of 32 × 512 points. Using the phase-resolved color Doppler (PRCD) algorithm to calculate the phase variations $\mathrm{\Delta }\varphi$ between frames with interval of 5 pixels for each subset as following [40]:

$$\mathrm{\Delta }\varphi (z)=\text{arctan}\left[\frac{\text{Im}(A_{j+1,z})\text{Re}(A_{j,z})-\text{Im}(A_{j,z})\text{Re}(A_{j+1,z})}{\text{Re}(A_{j,z})\text{Re}(A_{j+1,z})+\text{Im}(A_{j+1,z})\text{Im}(A_{j,z})}\right]$$ (1)

where Re(A_j,z) and Im(A_j,z) are, respectively, the real and imaginary part of the complex data at frame number j and axial depth z. (3) After all subsets are processed, the 3D data set of phase variations is projected along the axial direction to form the spatial-temporal phase distribution map. For instance, Fig. 3(a) presents the wave propagation in the 1.6% w/w agar phantom at different times after excitation. Eventually, after phase unwrapping, the further displacement map is obtained.

Fig. 2.

Flow chart of data processing. STFT: short-time Fourier transform. PRCD: phase-resolved color Doppler. TOF: time-of-flight. 2D-DFFT: two-dimensional discrete fast Fourier transform.

Fig. 3.

(a) Wave propagation in a 1.6% w/w agar phantom at different times after excitation. (b) The elastic modulus of homogeneous agar phantoms with varying concentrations (0.8%, 1.0%, and 1.2% w/w) measured and compared by SSS-OCE and uniaxial mechanical testing.

For bulk sample, such as agar phantom, the method of time-of-flight (TOF) is used to calculate the time delay of each segment based on the correlation algorithm [41]. By employing linear fitting on this distance-time curve, the group velocity of Rayleigh wave could be determined. Utilizing the Rayleigh wave model for the semi-infinite bulk phantom, the Young’s modulus E could then be calculated using the equation:

$$E=\frac{2\rho {(1+\nu )}^3}{{(0.87+1.12\nu )}^2}{{\text{c}}_{\text{r}}}^2$$ (2)

where ρ represents the density of the material (specifically, ρ = 1000 kg/m³ for the agar phantom and ρ = 1064 kg/m³ for the cornea), $\nu$  = 0.49 denotes the Poisson’s ratio, which accounts for the near incompressibility of the phantom [42], and c_r refers to the Rayleigh wave velocity.

Cornea is an extremely thin layer, the thickness of which has the same order as the wavelength of shear wave on its surface. A modified Rayleigh-Lamb frequency equation model has been documented to calculate the phase velocity dispersion in cornea [43,44]. A 2D discrete fast Fourier transform was implemented to transform the spatial-temporal displacement domain map into the wavenumber-frequency (k-f) domain map. Subsequently, the frequency dependent phase velocity c_p was derived by choosing the wavenumber at the maximum intensity of corresponding frequency, using the following equation [45,46]:

$$c_p=\frac{\omega }{k}$$ (3)

where c_p represents the phase velocity, ω denotes the angular frequency, and k is the wavenumber. As depicted in Fig. 2, the phase velocity tends to converge all the high-frequency spectrum for Lamb wave model. Significantly, with the air-medium boundary condition, the converged value of c_p matches the Rayleigh wave velocity c_r [32]. In our research, the averaged phase velocity over the frequency range above 1 kHz was analyzed as the Rayleigh wave velocity c_r, and the Young’s modulus was then derived by Eq. (2).

3. Results

3.1. Tissue mimicking agar phantoms

The group velocity of agar phantoms of 0.8%, 1.0%, and 1.2% measured by SSS-OCE were 3.18 ± 0.11 m/s, 4.04 ± 0.19 m/s, and 5.18 ± 0.24 m/s, respectively. Using the previously mentioned Eq. (2), the corresponding elastic modulus values were calculated as 33.36 ± 2.33 kPa, 53.82 ± 4.97 kPa, and 88.34 ± 8.15 kPa. Before performing the intraclass correlation coefficient (ICC) analysis, the repeated measurement values of each group of the agar phantoms were tested for normality using the Shapiro-Wilk test. The results showed normality, which met the normality assumption for ICC analysis. The ICC of the Young’s modulus for all agar phantoms was 0.999 (95% CI, 0.948 to 1.000), indicating an extremely high degree of repeatability. As shown in Fig. 3(b), the result revealed highly consistent between the values obtained from SSS-OCE and mechanical measurements, suggesting that SSS-OCE can perform rapid, high-fidelity imaging with single excitation while maintaining excellent measurement accuracy.

3.2. In situ porcine corneas

As illustrated in Fig. 4(c), the phase velocity of elastic wave propagation exhibited a progressive increase with elevated IOP. Based on the average calculation, the Rayleigh wave velocity of the porcine corneas is 3.95 ± 0.17 m/s at 10 mmHg, 4.98 ± 0.22 m/s at 15 mmHg, and 5.94 ± 0.21 m/s at 20 mmHg. The Young’s modulus values shown the same trend, which were 51.36 ± 4.47 kPa at 10 mmHg, 81.57 ± 7.30 kPa at 15 mmHg, and 116.25 ± 8.39 kPa at 20 mmHg. As shown in Fig. 4(d), a comprehensive series of paired t-tests was conducted for all possible inter-group comparisons (10 vs 15 mmHg, 10 vs 20 mmHg, and 15 vs 20 mmHg), and all comparisons yielded statistically significant results (p < 0.001). The mechanical preload exerted by IOP fundamentally governs corneal strain behavior, demonstrating a direct positive correlation wherein elevated IOP levels induce progressively greater tissue strain. This strain-pressure relationship manifests biomechanically as a nonlinear stiffening response, consistent with the characteristic behavior of collagenous tissues under tensile stress. Many in situ or in vivo studies using OCE have shown a similar increase in corneal stiffness when IOP was elevated [44,47,48].

Fig. 4.

Results of Lamb wave analysis of porcine corneas. (a) Picture of a porcine eye with IOP control system. (b) Frames at different times show the Lamb wave propagation in a porcine cornea after air-puff excitation. (c) A comparison of the elasticity of three in situ porcine corneas under different IOP (10 mmHg, 15 mmHg, and 20 mmHg) measured by SSS-OCE. The error bars in figure represent the standard deviation of the samples. (d) Overall comparison of Young’s modulus of porcine corneas under different IOPs (***p < 0.001).

3.3. Measurement of human corneas at two different IOPs

Due to the diurnal disruption of IOP, the collected data from two visits were categorized into higher IOP and lower IOP groups for each subject. The basic information of the five participants were reported in Table 1 and the range for “higher IOP” was 12.9 to 19.1 mmHg while the range for “lower IOP” was 11.2 to 17.6 mmHg. Figure 5(b) presents a comparison of corneal elasticity at two IOP levels. Notably, the results revealed a statistically significant increase (P < 0.001) in averaged corneal elastic modulus within the higher IOP group (893.18 ± 84.25 kPa) compared to the lower IOP group (487.10 ± 52.16 kPa). The close correlation between IOP levels and corneal elastic modulus verified the accuracy of our SSS-OCE.

Table 1. The basic information of the higher IOP group and the lower IOP group.

Subject	Gender	Age/Year	Eye	CCT/μm	IOP/mmHg
					Higher IOP group	Lower IOP group
1	Male	25	OD	580	19.1	17.6
2	Female	26	OD	500	17.6	16.8
3	Female	24	OD	550	12.9	11.2
4	Female	23	OD	540	16.6	13.6
5	Male	23	OD	540	15.4	14.6

Fig. 5.

Results of Lamb wave analysis of human corneas in vivo. (a) Picture of the measurement process of corneal elasticity in normal subjects based on the SSS-OCE system. (b) Measurement results of SSS-OCE in five human corneas in vivo under higher and lower IOP conditions.

4. Discussion

In this study, the feasibility of the ultra-fast line-field swept source scanning optical coherence elastography system and its associated algorithms were demonstrated using agar phantoms and in situ porcine corneas. By leveraging the Lamb wave analysis algorithm, we successfully accomplished the in vivo measurement of the elastic modulus of human corneas under various IOP conditions. It is proved that SSS-OCE under single excitation can be conveniently implemented for in vivo measurement, which has great potential clinical application value.

The swept source is the key component in our system. A centeral wavelength of 1050 nm is widely used in anterior segment SS-OCT for its optimal balance between tissue penetration and resolution. In the data processing, the windowing size mainly involves the spatial resolution of the line field detected by OCE system. In our system, the bandwidth of swept source was ∼ 100 nm, and 1024 data points were acquired within a single sweep. Theoretically, a narrower window enables finer discrimination of spectral components, and increases the spatial resolution. In our experiment, the spatial resolution was measured on a dual-concentrations agar phantom (1.0% and 1.6%). The result was shown in Fig. 6(a), which indicating a lateral resolution of 0.21 mm with window size of 64 pixels. Based on the same data, a relatively lower resolution (0.45 mm) was also calculated with a window size of 128 pixels. However, if the selected sliding window is excessively small, the number of points available for FFT becomes inadequate, potentially causing signal loss and other artifacts. By measuring a reflective mirror, we characterized the position-resolved sensitivity of our system, as shown in Fig. 6(b). A mean sensitivity of 33 dB was achieved across the line field. This value is expectedly lower than regular OCT because of the totally different scanning models between the two systems. Therefore, a sliding window size of 64 pixels best balances the trade-offs under the specification of swept light source used in our system.

Fig. 6.

SSS-OCE system performance metrics obtained using a 64-pixel sliding window. (a) Lateral resolution analysis. (b) The sensitivity versus position profiles.

The proposed SSS-OCE system demonstrates significant advantages over regular OCE technologies. Compared with traditional OCE using M-B mode scanning that requires repeated excitation [32,49–52], our system benefits from the inherently high sweep speed of the broadband swept laser source, achieving a remarkable improvement in frame rate in line field. With an A-line rate of 100 kHz in our experiment, one frame can be acquired in 10 microseconds. This high-speed imaging capability is sufficient with regard to propagation of elastic wave in tissue, making SSS-OCE be treated as an approimately parallel imaging technology in line-field. Therefore, single excitation can be adopted, and one measurement can be completed in just several milliseconds. The combination of rapid acquisition and single excitation is particularly valuable for in vivo ophthalmic applications.

Furthermore, SSS-OCE is modified from a regular SS-OCT system by adding a dispersion element. It offers advantages of both simpler structure and cost-effectiveness compared with mode-locked swept-source laser based OCE [35] and line-field spectral-domain OCE [37]. However, SSS-OCE has low axial resolution in each window, and the depth-resolved elasticity is not offered in the final results. Fortunately, depth information is not required if we only measure the overall corneal elasticity.

Through the utilization of the dispersion element, the SSS-OCE transforms the regular OCE’s point illumination pattern into a line-focused illumination configuration. This shift in illumination mode of SSS-OCE spreads the light energy across an extended region, reducing the energy density at any single position on the tissue. Therefore, SSS-OCE theoretically permits the employment of a light source with a relatively higher output power in the context of laser safety standards. In practical experiments, a total incident power of 1.0 mW on the sample is sufficient to maintain the sensitivity of phase information, which is far below the ANSI Z136.1 MPE limit for the cornea [37]. By contrast, the mode-locked swept source and the supercontinuum laser mentioned in literature have output powers of 160 mW [35] and 126 mW [37], respectively. When performing in vivo measurements on the human eye, high-power levels pose potential risks to the delicate ocular tissues. Theoretically, we can further enhance the sensitivity of OCE by increasing the power of the swept source within the safety limit.

Given these distinct characteristics, the SSS-OCE technology developed in this research holds substantial theoretical significance and practical value within ophthalmic clinical settings. It offers a novel and safer approach for in vivo OCE assessments, potentially enabling more accurate and reliable diagnosis of various ocular diseases. The main limitation of this work is that only line field was measured based on the current system construction. Two-dimensional area scanning will be achieved by integrating a one-dimensional scanning galvanometer in the sample arm. The proposed improvement is expected to yield a more comprehensive field of view while maintaining micron-level positioning accuracy.

5. Conclusion

In summary, taking advantage of wavelength-dependence sweeping based on dispersion principle, SSS-OCE achieves ultra-fast measurement of elasticity with single excitation, low exposure power and low cost. The capability of this technology for in vivo assessing human corneal elasticity has been demonstrated in our work. SSS-OCE has significant application potential in early diagnosis of keratoconus, as well as in vivo assessment of corneal collagen cross-linking surgery.

Funding

National Natural Science Foundation of China 10.13039/501100001809 ( 62275201); Natural Science Foundation of Zhejiang Province 10.13039/501100004731 ( LY23F050012).

Disclosures

The authors declare that there are no conflicts of interest related to this article.

Data availability

Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.