Acoustic Force Elastography Microscopy

A. Acoustic Force Elastography Microscopy (AFEM) and Reflected Longitudinally Polarized Shear Waves (RLPSWs)

AFEM consists of a highly focused ultrasound transducer, an optical coherence tomography (OCT) system and a x-y translation stage. In this study, a focused ultrasound transducer was used to produce the acoustic radiation force (ARF). The transducer was coupled to the bottom of the Petri dish using ultrasound coupling gel. The ARF was exerted on the surface of materials to generate surface waves that travel outwards on the surface as well as a wave that travels through the thickness of the material, which is defined as a longitudinally polarized shear waves (LPSWs) [47, 48], presented in Fig. 1. The waves that reflect from the bottom of the Petri dish and propagate back to the surface of materials are reflected LPSWs (RLPSWs). The arrival time, t_a, of the RLPSW is determined by the local thickness of the sample, h, and shear modulus, μ, of the material according to the relationship $t_a=0.5h\sqrt{\rho /\mu }$, where ρ is the mass density of the material assumed as 1,000 kg/m³ for biological samples and the 0.5 factor accounts for the round-trip travel distance. Therefore, μ of the material can be given by ρ(0.5h/t_a)², where the term 0.5h/t_a is defined as the velocity of RLPSWs, C_AFEM. Young’s modulus E can be determined by E ≈ 3μ as the localized thickness h and t_a of the material are obtained [49, 50], assuming that the material is isotropic, homogeneous, linear, and nearly incompressible. The spatial resolution of AFEM depends on a lateral focal size of the transducer that can be calculated by the equation: 2.44 × f/# × λ_US and is approximately 510 μm in this study [51], where f/# is the f-number defined as the focal distance divided by the aperture size of the transducer and λ_US is the ultrasound wavelength. Measurement of RLPSWs at the surface of materials do not require optical scatterers, i.e., materials can be transparent, which is common for hydrogel scaffolds in tissue engineering.

The role of the OCT in the AFEM system is to measure particle motion of RLPSWs. OCT has a number of advantages including being a non-contact measurement with micro-scale spatial resolutions, sensitivity to sensing particle displacements in nanometer scales with flexible scan speeds [52]; therefore, it is capable of supporting a finer elastography resolution in the AFEM system. Fig. 2 illustrates the basic optical layout of the spectral-domain optical coherence tomography (SD-OCT) system (TEL320C1, Thorlabs Inc., Newton, NJ, USA). The OCT is equipped with a 1300 nm source with low coherence broadband (236.8 nm of bandwidth) and LK4 lens kit (Thorlabs Inc., Newton, NJ, USA) to produce 20 μm of lateral resolution and 3.5 μm of z-axis resolution (provided by Thorlabs Inc.) in air. A low coherence broadband source is split into a reference beam directed toward a stationary reference mirror and a sample beam directed toward the samples. According to the basic principle of the Michelson interferometer, the back-reflected and back-scattered light from samples and retroreflected light from the reference mirror are recombined by a coupler. A spectral interferogram is formed by spectrometers and data are collected by a frame grabber card. The fast Fourier transform (FFT) is utilized to form an A-scan from the receiver array in the SD-OCT system. Each pixel includes a real value and an imaginary value (in-phase/quadrature, IQ) from which magnitude and phase can be calculated. One-dimensional (1D) autocorrelation was used to calculate the phase information Δφ of particle displacements including RLPSWs and surface waves [53]. The particle velocity of RLPSWs V(z, t) are given by [50, 54]

where f_SR is the OCT scan rate, λ_OCT is center wavelength of light source and n is the refractive index of the material. Generally, the refractive index of biological materials, tissues and tissue-mimicking phantoms ranges from 1.35 to 1.55 [55, 56] and we selected 1.40 as an average value in the study.

A 7.5 MHz focused ultrasound transducer with a f/# of 1.07 was used to produce the acoustic radiation force (ARF). Due to different scales of thickness between tissue-mimicking gelatin hydrogels and scaffolds of tissue engineering, a sinusoidal burst signal with two different pulse durations was used for gelatin hydrogels (500 μs) and for scaffolds (50 μs) to provide signals that could be tracked with sufficient amplitude, which was provided by function generator 2 (33250A, Agilent, Santa Clara, CA, USA). The burst signals were amplified 50 dB by a radiofrequency (RF) power amplifier (240L, Electronics and Innovation, LTD, Rochester, NY, USA) to drive the transducer. Function generator 3 (33500B, Keysight, Santa Rosa, CA, USA) controlled the OCT scan rates and triggering in the external mode. Function generator 1 (33250A, Agilent, Santa Clara, CA, USA) was a master trigger to synchronize the timing for the AFEM entire system.

For measuring tissue-mimicking gelatin hydrogels as the system test stage, the customized scan pattern in AFEM was designed to use a 50 kHz A-line OCT scan rate and 100 μm lateral spacing over a 16 mm field-of-view (FOV) (160 spatial locations) to record both a RLPSW and a surface wave in the first 50 ms after the excitation initiation. To achieve a two-dimensional (2D) elastography measurement of scaffolds in a fast manner for tissue engineering, the lateral spacing was chosen to cover a 2–4 mm FOV (20–40 spatial locations) for each spatial position so that AFEM can take data over a small patch centered about the excitation area for each spatial position to achieve a high throughput measurement with AFEM. The scan pattern of AFEM (Fig. 3) at each individual location, p, a data set was acquired with dimensions of z depth samples and N = i·j is the product of i M-scans and j B-scans (M-B scan) [57] for an elasticity measurement in each location. For 2D scans, the scan pattern was composed of a dataset with the dimensions of z by N by k by s matrix, where k and s are the number of lateral and elevational elastography measurement positions, respectively. The 2D AFEM measurement can reveal more details of the elastic properties of scaffolds in tissue engineering, especially for composite scaffolds due to heterogeneous properties. The customized acquisitions were developed by using a SpectralRadar software development kit (SDK) 4.4 Version provided by Thorlabs Inc. in Microsoft Visual C++ 2019 development environment (Microsoft, Redmond, WA).

B. Tissue-mimicking Gelatin Hydrogel Preparations

We fabricated nine gelatin hydrogels to conduct five independent experiments to evaluate the ability of AFEM. The first experiment was designed to study the behavior of RLPSWs. The second and third experiments were designed to evaluate AFEM performance with the fixed thickness and the fixed stiffness, respectively. The fourth experiment was designed to evaluate the elastic resolution of the AFEM. For the first experimental setup, a 20% v/v homogeneous gelatin hydrogel with 4.5 mm thickness was fabricated, with the thickness measured by a pulse-echo ultrasound test with 10 MHz single element transducer (V312, Olympus, Waltham, MA), by using gelatin powder (gel strength 300 type A, G2500-1KG, Sigma-Aldrich, St. Louis, MO, USA). One gram of titanium dioxide (TiO₂, ReagentPlus ≥ 99%, Sigma-Aldrich, St. Louis, MO, USA) was used to provide optical scatterers for investigating the behaviors of RLPSWs compared with the results by the numerical simulation as the validation.

This comparison can evaluate two questions. First, can elastic properties be evaluated using RLPSWs? Second, can AFEM evaluate elastic properties of tissue-mimicking materials reliably? After the first experimental setup, surface waves were used as one of the validation methods, which were applied to the second and third experimental setups. The second experimental setup included the concentrations of 8% v/v, 13% v/v and 18% v/v gelatin hydrogels with the same thickness, 6 mm (measured by a pulse-echo ultrasound test), for the systematic study to evaluate the relationship between stiffness and RLPSW velocity. The third experimental setup was designed to evaluate the relationship between thickness and RLPSWs by using samples with thicknesses of 4.0, 3.2, and 2.5 mm, measured by a pulse-echo ultrasound test, with the same stiffness as 8% v/v. For each concentration, a total volume of 100 mL tap water in a 500 mL beaker was heated to approximate 70 °C. The gelatin powders and 1 g TiO₂ were added with stirring to the beaker for approximately five minutes to homogenize the solution. The mixed solution was placed in a de-gassing chamber to remove small bubbles in the fluid. After de-gassing, the mixed gelatin solution was poured into regular Petri dishes (100 mm diameter × 10 mm height) and were transferred to a 4 °C refrigerator for 3 hours for congealing. The mechanical properties of the gelatin hydrogels with the room temperature were evaluated by the proposed AFEM.

The fourth experiment was performed to test the elastic resolution of the AFEM. A heterogenous gelatin hydrogel sample with high elastic contrast (10% v/v versus 20% v/v) was fabricated using gelatin powder (gel strength 300 type A, G2500-1KG, Sigma-Aldrich, St. Louis, MO, USA). Due to the advantage of AFEM, scatterers are not needed and the hydrogel can be completely transparent. For 20% gelatin hydrogel, a total volume of 100 mL tap water in a 500 mL beaker was heated to approximate 70 °C and the gelatin powders was added with stirring to the beaker for approximately five minutes to homogenize the solution. The gelatin solution was placed in a de-gassing chamber to remove small bubbles in the fluid and then was poured into a container (15 mm diameter × 10 mm height). The liquid gelatin hydrogel was transferred to a 4 °C refrigerator for 3 hours for congealing. After 3 hours, the above procedures were repeated for 10% gelatin hydrogel. The liquid gelatin hydrogel was poured into the same container to cover the substrate (20% gelatin hydrogel), which was transferred to a 4 °C refrigerator for the other 3 hours for congealing. Two planes with approximately 2 mm distance were sliced by using a tissue slicer blade to make a thin gelatin hydrogel sample.

The fifth experiment was performed to test whether or not the ARF pulse generates vibrations in the bottom of the Petri dish that influence the particle motion of RLPSWs. A 7% gelatin hydrogel was fabricated by the same procedure mentioned in the previous paragraph. The gelatin solution was poured into two Petri dishes, a regular Petri dish and a Petri dish where the bottom was replaced with Mylar film, with a volume of 15 mL in each to create the same thickness of the phantoms. The Mylar film is about 100 μm thick so the acoustic power is considered to be weakly attenuated [58]. The thickness of the phantom was measured as 2.053 mm by a pulse-echo ultrasound test. The material of the Petri dish is polystyrene and the thickness of the bottom is approximately 700 μm.

C. Numerical Simulations

To validate our proposed technique, numerical simulations were performed using the finite element method (FEM). An explicit solver was used to solve the differential elastic wave equations in OnScale (OnScale, Redwood City, California) using a transient excitation. Two three-dimensional models with elastic material properties were considered with 16 mm width × 4.5 mm thickness for the validation of the tissue-mimicking gelatin hydrogels and with 4 mm width × 950 μm thickness for the validation of the scaffolds for bone tissue engineering, respectively. By assuming that the Petri dish is infinitely stiffer than the tested samples, we can use a fixed displacement boundary condition applied at the bottom and sides of the numerical phantoms. The parameters of the numerical simulations were based on the results from the AFEM experiments.

According to the experimental results obtained by using AFEM, the RLPSW velocities in the 20% tissue-mimicking gelatin hydrogels and the neat OPF scaffold were set to be 2.28 m/s and 4.78 m/s, respectively. The simulation time step was 0.02 μs for the gelatin hydrogels and 0.0053 μs for the scaffolds due to the consideration of the sample thickness and the finite element size used. The Young’s modulus E_AFEM can be approximately evaluated by $3\rho C_{AFEM}^2$ [59], where C_AFEM is the velocity of RLPSW and density of the materials ρ was set as 1000 kg/m³. A transient force excitation with the pulse duration of 500 μs (gelatin hydrogel) and with 50 μs (synthetic OPF scaffold) was applied to the surface of the model in an axial direction (red arrow of Fig. 2). Particle velocity motion data were measured from the top surface and further processed. The computation time for the numerical simulations took approximately 25 minutes using a standard personal computer (PC).

D. Porcine Kidney Tissue Preparation

A porcine kidney was excised immediately after sacrifice and frozen in saline. The pig was used for another study that was approved by the Mayo Clinic Institutional Animal Care and Use Committee. The kidney was thawed and placed in saline solution at room temperature. The kidney was sliced in half in the long dimension and reflected. A slice of renal cortex with approximate dimensions of 20 × 20 × 1.5 mm was cut for the experiment.

E. Fabrications of Synthetic Oligo[poly(ethylene glycol)fumarate] (OPF) Scaffolds and OPF Hydroxyapatite (HA) Nano-composite Scaffold for Tissue Engineering

Synthetic hydrogels have been widely used because of the ability to tailor their material properties such as molecular weight for specific tissue engineering applications. In this study, two groups of OPF scaffolds, the neat and composite OPF scaffolds with hydroxyapatite (HA), were fabricated. For each group, four samples were made, where one was used for AFEM measurement and the others were used for DMA tests as the validation. The OPF was synthesized from poly(ethylene glycol) (PEG) with a number average molecular weight (M_n) of 10,000 g/mol. A 50 g amount of PEG was azeotropically distilled in toluene to remove residual water. Dried PEG was dissolved in methylene chloride in a three-neck flask placed in the ice bath with nitrogen, which was followed by the addition of potassium carbonate (K₂Co₃) and distilled fumaryl chloride with 24 hours stirring. For the purification of the synthesized OPF, methylene chloride was removed by rotary evaporation and the resulting product was dissolved in ethyl acetate, precipitated in petroleum ether, cleaned with diethyl ether, and vacuum-dried overnight. For the neat OPF scaffolds, 0.5 g OPF, 0.15 g polyethylene glycol diacrylate (PEGDA) as a cross-linker and 2.5 mg Irgacure 2959 as a photo-initiator were dissolved in 1.15 ml deionized water (dH₂O). For composite scaffolds, 10% w/w HA was fully dispersed in OPF solution before crosslinking. The solution was pipetted between glass-slides fitted with 1 mm thick Teflon spacer and transferred to cure the solution by a 365 nm UV light (Black-Ray Model 100AP, Upland, CA) at an intensity of 8 mW/cm² for creating photo-crosslinked OPF scaffolds. The two-dimensional (2D) elastic properties of the crosslinked OPF scaffolds were characterized by AFEM, and their results were validated by conventional dynamic mechanical analysis (DMA) tests. The fabrication processes in more details can be found in previous reports [44, 45].

F. Dynamic Mechanical Analysis (DMA)

The neat OPF scaffolds and OPF composite scaffolds were cut into 8 mm diameter disks and subjected to compression under displacement control. The elastic properties of the scaffolds were determined using the RSA-G2 dynamic mechanical analysis (TA Instruments, New Castle, DE). The compression rate was set as 0.1 μm/s. The stress-strain data were collected from the tests and Young’s modulus E was determined by a slope of the linear portion of the stress-strain relationship in the 0–5% region.

G. Evaluation of AFEM elastography Spatial Resolution

Two planes with approximately 2 mm distance were sliced using a tissue slicer blade to make a thin gelatin hydrogel sample. A total of 21 measurements with a step of 0.635 mm were performed, and the 11^th measurement was exactly at the boundary. AFEM was performed with a scan rate of 50 kHz with 50 μm lateral spacing. The duration of the acoustic excitation was 50 μs. The wave velocity profile was fit to a sigmoid function as an edge-spread function for the calculation of the elastography spatial resolution of AFEM based on the full-width at half-maximum (FWHM) of the differentiated edge-spread function.

A. AFEM Validation Using Finite Element Analysis and Dispersion Analysis

A 7.5 MHz focused ultrasound transducer with an f-number of 1.07 produced the acoustic radiation force (ARF). The ARF was exerted on the surface of materials to generate waves that travel through the thickness of materials, which are described as longitudinally polarized shear waves (LPSWs) [47, 48], as shown in Fig. 1. The wave that reflects from the bottom of the thin material due to the rigid boundary condition and vertically propagate back to the surface of the material through the same route are described as reflected LPSWs (RLPSWs). The mechanical properties of hydrogels and scaffolds are determined by two parameters: arrival time, t_a, of the RLPSW and the thickness of the sample at the measuring location, h. Arrival time is defined as the time for waves to travel from the surface to the bottom and return back to the surface (Fig. 1). AFEM does not require optical scatterers, i.e., materials can be transparent, which is common for hydrogel scaffolds in tissue engineering.

The snapshots of the LPSW at 0.5 ms and the RLPSW at 4 ms vertically propagating inside a tissue-mimicking 20% gelatin hydrogel with 4.5 mm thickness were observed in Figs. 4(a) and 3(b) recorded by AFEM, respectively. Figures 4(a) and 4(b) clearly show the LPSW propagating from the surface of the phantom to the bottom of the Petri dish, and RLPSW exhibits a wave propagating back to the surface from the bottom of Petri dish (Movie S1). For a model with the same thickness and stiffness as in the AFEM experiments, the simulated wave motion was in a good agreement with the AFEM experimental results, illustrated in Figs. 4(c), 4(d) and Movie S2. According to the comparison between the numerical simulations and the measurement results by AFEM, three different types of wave velocities, the group velocity of surface waves C_g, the velocity of RLPSWs C_AFEM, and the phase velocity of surface waves C_p, can be compared. The group velocity was evaluated by a time-to-peak algorithm across multiple lateral positions of spatiotemporal data (t-x), and phase velocity was calculated by transferring spatiotemporal data into a k-space (f-k) via 2D Fourier transform to obtain a frequency domain representation. The details of group and phase velocity calculation are provided in previous literature [60–62]. Figure 4(e) illustrates the spatiotemporal map containing the surface wave and the RLPSW measured by AFEM for the 20% gelatin hydrogel. In the first experimental setup, the field-of-view (FOV) of AFEM was set as the maximum view (16 mm) to record surface wave propagation for evaluating accurate results and comparing them with RLPSWs for validation. The C_g measured by AFEM was 2.28 m/s, which can be calculated by the slope of the wave indicated by the dark dashed line rectangle in Fig. 4(e). Using AFEM, t_a was recorded as 4.03 ms and the thickness of the phantom was measured as 4.5 mm by a pulse-echo ultrasound test. The C_AFEM was evaluated as 2.23 m/s and the Young’s modulus E_AFEM was 14.91 kPa.

For the numerical simulations, the same elastic modulus (C_g = 2.28 m/s) and thickness (h = 4.5 mm) were used as the simulation parameters. According to the simulation results, C_g = 2.20 m/s, the t_a = 4.08 ms and C_AFEM = 2.21 m/s. The Young’s modulus based on the velocity of RLPSWs in simulations E_Sim was 14.65 kPa, presented in Fig. 4(f). A profile through the entire time dimension (20 ms) at the position of the ARF excitation was shown in Fig. 4(e). Figure 4(f) shows that the RLPSWs between the results measured by AFEM and the numerical simulations were highly consistent. The phase velocity dispersion curves of the experiment and the numerical simulations were displayed in Fig. 4(h) and mean C_p over a bandwidth of 200–1000 Hz is 2.31 ± 0.006 m/s for AFEM and 2.16 ± 0.005 m/s for the numerical simulation. According to the first experimental setup, the proposed AFEM has very good agreement with the numerical simulations. The Young’s modulus measured by using RLPSWs agreed well with those measured by using surface waves. The agreement from two independent methods gives high confidence in the quality of the proposed AFEM for evaluating the mechanical properties of biological samples.

B. Systematic Performance Evaluations of AFEM Using Tissue-mimicking Gelatin Hydrogels

Systematic evaluations of the performance of AFEM were performed using groups of tissue-mimicking gelatin hydrogels with the two design conditions, fixed thickness and fixed stiffness. In Figs. 5(a)–(c), with the fixed thickness of 6 mm, the RLPSWs can be clearly detected by AFEM with t_a of 9.98 ms in 8%, 6.32 ms in 13% and 5.26 ms in 18% gelatin hydrogels demonstrating that t_a is inversely proportional to the stiffness of the phantoms (Fig. 5d). In this experimental setup, the C_AFEM changed from 1.32–2.32 m/s with increasing gelatin hydrogel concentrations (Table I). Consequently, E_AFEM ranged from 5.25–16.14 kPa, presented in Fig. 5 and Table I, which implies that the mechanical properties are associated with the changes of C_AFEM traveling in scaffolds. According to the results from the numerical simulations (Fig. 4), surface waves can be used to validate the results from AFEM. The FOV of AFEM was set as 16 mm to record surface wave propagation. According to C_g in Table I, the Young’s modulus based on surface waves E_Surface ranging from 5.09–18.32 kPa show a highly consistent trend and a good agreement with E_AFEM measurements.

Figure 6 shows the RLPSWs in phantoms with three different thicknesses with a fixed stiffness (8% gelatin hydrogel) evaluated by AFEM. In Figs. 6(a)–(c), the RLPSWs were clearly detected by AFEM and t_a is proportional to the thickness of the phantoms (Fig. 6(d)). In this experimental setup, the C_AFEM stays constant at approximately 0.90 m/s regardless of thickness, with E_AFEM ranging from 2.58–2.85 kPa for the three gelatin hydrogels, presented in Fig. 6 and Table I. The surface waves can be used for validating the AFEM experimental results. Using surface waves, the C_g and E_Surface also show similar measurement values of approximately 0.90 m/s and 2.5 kPa in the three hydrogel phantoms, respectively. The two experimental setups demonstrate that AFEM is capable of evaluating the mechanical properties of materials with various thickness and stiffness. The systematic performance evaluations of AFEM with wave velocities and mechanical properties are summarized in Table I.

C. Evaluation of Mechanical Properties of Ex vivo Porcine Kidney Using AFEM

We quantified the elastic properties of the ex vivo porcine kidney using AFEM. The porcine kidney cortex thickness of 1.53 mm can be determined from an OCT image as shown in Fig. 7(a). The RLPSWs propagating inside the porcine kidney and surface waves propagating on the porcine kidney sample were clearly detected by AFEM. The C_g was 0.86 m/s and t_a of the RLPSW was 3.80 ms in Fig. 7(b). AFEM was repeated with ten acquisitions at the same location, and the ten profiles at the position of the ARF excitation is shown in Fig. 7(c). A consistent value of t_a = 3.80 ms from ten RLPSWs was observed to demonstrate the stability of the RLPSWs as they are traveling inside tissue. The C_AFEM was 0.81 m/s, which shows high consistency with C_g. The E_AFEM and E_Surface were approximately 2.00 kPa and was 2.23 kPa, respectively. Our previous literature reports that the Young’s modulus of ex vivo porcine kidney is around 2–3 kPa [63], which is in good agreement with the results measured by the proposed AFEM.

D. Evaluation of 2D Mechanical Properties of Neat OPF Scaffolds and OPF Hydroxyapatite (HA) Nano-composite Scaffolds for Tissue Engineering Using AFEM

To demonstrate the feasibility of the proposed AFEM for tissue engineering applications, the mechanical properties of the synthetic polymer scaffolds were evaluated. Figure 8 shows AFEM can characterize the stiffness of the neat OPF scaffold with M_n = 10,000 g/mol over a 2D region, and results were validated by numerical simulation and conventional DMA tests. In Fig. 8(a), the neat OPF scaffold can be observed as transparent, so it is difficult to quantify the elastic properties using conventional ultrasound and optical modalities due to the lack of scatterers. The thickness of the neat scaffold at a single position was measured to be approximately 960 μm, as shown in Fig. 8(b). The t_a of the RLPSWs was 0.4 ms as detected by AFEM at the single position of the scaffold (Fig. 8(c)), and the C_AFEM was evaluated as 4.78 m/s. The localized stiffness was determined as approximately 68.63 kPa for the neat scaffold and 66.21 kPa for the numerical simulation. According to the parameters of the scaffold thickness and C_AFEM, the numerical simulation demonstrates that t_a = 0.407 ms for the RLPSWs (Fig. 8(d)) is in accordance with the result from AFEM. AFEM has the ability to measure thickness, velocity of RLPSWs and mechanical properties of the scaffolds over a 2D region. The 2D maps consisted of 25 different scan positions (5 × 5 in x- and y-directions) with a 2.54 mm interval on the neat scaffold as shown in Fig. 8(a) with the dark red color region. The 2D maps of t_a, h, and C_AFEM of the scaffold are shown in Figs. 8(e)–(g). The t_a over the 2D region was 0.31 ± 0.003 ms, h was 0.83 ± 0.006 mm and C_AFEM was 5.35 ± 0.07 m/s. Therefore, the 2D elastic map of the neat scaffold measured using AFEM can be obtained as presented in Fig. 8(h), and E_AFEM = 86.14 ± 2.16 kPa. Three repeated DMA tests were conducted to evaluate the Young’s modulus E_DMA of the neat OPF scaffold as 81.76 ± 7.35 kPa, which shows good agreement with the AFEM results presented in Fig. 8(i).

AFEM is also capable of evaluating the mechanical properties of composite scaffolds for tissue engineering. Figure 8(j) shows the OPF composite scaffold with 10% HA, which were examined by both AFEM and DMA. The h was 859 μm measured by AFEM and t_a of the RLPSW was approximately 0.28 ms, as shown in Fig. 8(k); therefore, the C_AFEM was 6.16 m/s and E_AFEM was 114.00 kPa at a single location. Three measurements at randomly selected locations of the OPF composite scaffold were performed. The C_AFEM and E_AFEM of the OPF composite scaffolds were 6.16 ± 0.17 m/s and 108.00 ± 12.06 kPa by AFEM, respectively, which was validated by the DMA test with E_DMA of 102.80 ± 6.53 kPa (Fig. 8(l)). The wave velocities and mechanical properties of scaffolds are summarized in Table I. The proposed AFEM technique is promising for quantitative mechanical characterization of thin scaffolds without scatterers and with non-contact and non-destructive measurements.