An ultrasound surface wave technique for assessing skin and lung diseases

Abstract

Systemic sclerosis (SSc) is a multi-organ connective tissue disease characterized by immune dysregulation and organ fibrosis. Severe organ involvement, especially of the skin and lung, is the cause of morbidity and mortality in SSc. Interstitial lung disease (ILD) includes multiple lung disorders in which the lung tissue is fibrotic and stiffened. The purpose of this study was to translate ultrasound surface wave elastography (USWE) for assessing patients with SSc and/or ILD via measuring surface wave speeds of both skin and superficial lung tissue. 41 patients with both SSc and ILD and 30 healthy subjects were enrolled in this study. An external harmonic vibration was used to generate the wave propagation on the skin or lung. Three excitation frequencies of 100 Hz, 150 Hz, and 200 Hz were used. An ultrasound probe was used to measure the wave propagation in the tissue noninvasively. Surface wave speeds were measured on the forearm and upper arm of both left and right arms as, well as, the upper and lower lungs through six intercostal spaces of patients and healthy subjects. Viscoelasticity of the skin was calculated by the wave speed dispersion with frequency using the Voigt model. The magnitudes of surface wave speed and viscoelasticity of patients’ skin were significantly higher than those of healthy subjects (p<0.0001) for each location and each frequency. The surface wave speeds of patients’ lung were significantly higher than those of healthy subjects (p<0.0001) for each location and each frequency. USWE provides a noninvasive and nonionizing technique to measure both skin and lung surface wave speed, which may be useful for quantitative assessment of SSc and/or ILD.

INTRODUCTION

Systemic sclerosis (SSc), also termed scleroderma, is a multi-organ connective tissue disease characterized by immune dysregulation and organ fibrosis (Steen and Medsger 2000). Thickening of the skin, often the earliest affected organ, is considered an early marker of disease activity in SSc (Steen and Medsger 2000). Severe organ involvement, especially of the skin and lung, is the cause of morbidity and mortality in SSc. The degree of skin involvement is a predictor of mortality (Clements, et al. 1990). Improvement in skin stiffness is associated with improved survival in many clinical trials (Steen and Medsger 2001). Patients who do not develop severe organ involvement in the first few years are less likely to develop life-threatening involvement later throughout the course of the disease. The Modified Rodnan Skin Score (MRSS) is the standard skin assessment tool in the majority of clinical studies of SSc (Abignano, et al. 2011). It shows that patients with improved MRSS after two years of treatment had improved survival (Steen and Medsger 2001). The MRSS is commonly used as an outcome measure in clinical trials (Clements, et al. 2000). However, the MRSS is a palpation method, which is subjective, and thus, its accuracy is user-dependent (Clements, et al. 1995). Moreover, it is difficult to measure the change of skin stiffness over time using palpation (Postlethwaite, et al. 2008).

Skin stiffness can be measured using durometry (Kissin, et al. 2006, Merkel, et al. 2008), indentation (Boyer, et al. 2009, Pailler-Mattei, et al. 2008, Serup 1995), and cutometers (Hendriks, et al. 2006, Smalls, et al. 2006). Durometry uses a piston-spring-dial handheld apparatus to measure skin hardness. However, durometry readings are affected by the experience of users. Indentation is the technique of measuring elasticity by indenting the material. The cutometer measures skin displacement in response to a suction force. Because its measurements depend on the interactions between a probe and the skin, deeper skin layers must be measured with larger probes (Hendriks, et al. 2006). Notably, these techniques cannot evaluate subcutaneous tissue.

Patients with ILD have fibrotic and stiff lungs leading to symptoms, especially dyspnea, and may eventually lead to respiratory failure (Coultas, et al. 1994). Many ILDs typically are distributed in the peripheral, subpleural regions of the lung (Desai, et al. 2004, Wells, et al. 1993). High-resolution computed tomography (HRCT) is the clinical standard for diagnosing lung fibrosis (Mathieson, et al. 1989, Verschakelen 2010), but it substantially increases radiation exposure for patients. Various scanning techniques were proposed to reduce the dose (Mayo 2009). Intercostal ultrasound can be used to avoid ionizing radiations during follow-ups (Delle Sedie, et al. 2012). However, it is not able to quantify the stiffness of lung.

Currently, no clinical approach is available to noninvasively quantify and evaluate the progression and development of SSc and ILD. Therefore, there is a pressing need to develop an accurate and reproducible clinical technique of quantification, tracking of the disease, and developing effective treatment for sclerosis and fibrotic disorders. (Lott and Girardi 2011). This research aims to translate an ultrasound surface wave elastography (USWE) technique into clinical use for quantitatively assessing patients with SSc and interstitial lung disease (ILD).

METHOD

Raleigh Surface Wave Equation

Surface wave propagation can be analyzed as wave propagation in a semi-infinite linear elastic medium under a local harmonic excitation on the surface. The equation for wave propagation in an isotropic and elastic medium is (Miller 1954)

$$(\lambda +2\mu )\nabla \nabla .\overrightarrow{u}-\mu \nabla \times \nabla \times \overrightarrow{u}=\rho \frac{{\partial }^2\overrightarrow{u}}{\partial t^2},$$ (1)

where u⃗ is the displacement vector, ρ is the mass density, and λ and μ are, respectively, the Lamé coefficients of the medium. For linear viscoelastic material, λ = λ₁ + ∂λ₂/∂t and μ = μ₁ + ∂μ₂/∂t, where λ₁, λ₂, μ₁ and μ₂ are the coefficients of volume compressibility, volume viscosity, shear elasticity and shear viscosity, respectively.

The surface wave propagation can be solved in the cylindrical polar coordinate system as shown in Fig. 1. Consider a harmonic force excitation with a uniform stress on the surface of the medium in the circular region of r ≤ a. The displacement fields are derived in the r and z directions at any location in and on the surface of the medium as (Miller 1954),

$$u_z=\frac{a}{\mu }\underset{0}{\overset{\infty }{\int }}\frac{\sqrt{({\xi }^2-1)}{J}_1(\xi k_{1}a)}{F_0(\xi )}\{2{\xi }^2{e}^{-k_{1}z\sqrt{({\xi }^2-{\eta }^2)}}+({\eta }^2-2{\xi }^2)e^{-k_{1}z\sqrt{({\xi }^2-1)}}\}{J}_0(\xi k_{1}r)d\xi$$

$$u_r=\frac{a}{\mu }\underset{0}{\overset{\infty }{\int }}\frac{\xi J_1(\xi k_{1}a)}{F_0(\xi )}\{2\sqrt{({\xi }^2-1)}\sqrt{({\xi }^2-{\eta }^2)}{e}^{-k_{1}z\sqrt{({\xi }^2-{\eta }^2)}}+({\eta }^2-2{\xi }^2)e^{-k_{1}z\sqrt{({\xi }^2-1)}}\}{J}_1(\xi k_{1}r)d\xi ,$$ (2)

where a is the radius of the distributed stress, and ξ is the integration parameter in the wave number domain, which has been normalized with respect to k₁. The divisor function of the integration functions is $F_0(\xi )={(2{\xi }^2-{\eta }^2)}^2-4{\xi }^2\sqrt{({\xi }^2-1)}\sqrt{({\xi }^2-{\eta }^2)}$, where $\eta =k_2/k_1=\sqrt{\{2(1-\sigma )/(1-2\sigma )\}},k_1=\omega \sqrt{\rho /(\lambda +2\mu )},k_2=\omega \sqrt{\rho /\mu }$, where ω is the angular frequency, ρ is the density, σ is the Poisson’s ratio for the medium, k₁ and k₂ denote the wave numbers for compression and shear wave propagation, respectively, J₀ and J₁ refer to Bessel functions of the first kind.

Figure 1.

Schema of surface wave generation and detection on the skin. The skin surface is on the plane of x and y coordinates. The surface wave on the skin is generated by a handheld electromechanical shaker through a ball-tip applicator on the skin. The vibration excitation is typically 0.1 second harmonic signal at a frequency between 100 Hz and 200 Hz. The resulting surface wave propagation on the skin is measured using an ultrasound probe. A standoff gel pad is used between the ultrasound probe and skin to improve imaging of the skin. The surface wave speed depends on the local elastic properties of the skin and independent of the location of wave generation.

The wave displacement fields can be solved with equation (2). However, using the displacements to estimate the elastic properties of the medium depends on the excitation and boundary conditions. Since the wave propagation is dependent on local medium properties, we used the surface wave speed measurement to estimate the viscoelastic properties of the medium. The surface wave speed can be solved by (Zhang and Greenleaf 2007)

$${(2{\xi }^2-{\eta }^2)}^2-4{\xi }^2\sqrt{({\xi }^2-1)}\sqrt{({\xi }^2-{\eta }^2)}=0.$$ (3)

Equation (3) is a bi-fourth equation of ξ, however, only the solution for which ξ is real, ξ > η and ξ > 1 is the right solution for the surface wave.

Equation (3) has been studied by (Nkemzi 1997) and us (Zhang and Greenleaf 2007). Several approximations have been proposed to solve the surface wave speed (Royer and Clorennec 2007, Twal, et al. 2014, Viktorov 1976, Vinh and Malischewsky 2007). The following equation approximates the relationship between the shear wave speed c₂ and the surface wave speed c_s (Royer and Clorennec 2007),

$$\frac{c_2}{c_s}=\sqrt{\frac{0.58+K}{0.44+K}},$$ (4)

where K = σ / (1 − σ), σ is the Poisson’s ratio of the medium. The speed of shear wave is slightly greater than that of surface wave for all materials from metals to soft tissues. Because soft tissues are generally incompressible and their Poisson’s ratios are in a narrow region between 0.45 and 0.50, c₂ = 1.05 c_s can be obtained from equation (4). The surface wave speed can be related to the elastic modulus of tissue as (Zhang and Greenleaf 2007),

$$c_s=\frac{1}{1.05}\sqrt{\frac{\mu }{\rho }}$$ (5)

where μ is the shear elasticity in Pascal and ρ is the mass density of the tissue in kg/m³.

Voigt’s Viscoelastic Model – Wave Dispersion Curve

For soft tissue under low frequency harmonic excitation, the Voigt’s model, which consists of a spring of elasticity μ₁ and a damper of viscosity μ₂ connected in parallel, has been proven to be effective in modeling the linear viscoelastic materials (Catheline, et al. 2004, Chen, et al. 2009, Prim, et al. 2016, Zhou, et al. 2017). The wave dispersion curve of wave speed c_s with respect to the excitation frequency ω can be formulated by,

$$c_s=\frac{1}{1.05}\sqrt{\frac{2({\mu }_1^2+{\omega }^2{\mu }_2^2)}{\rho ({\mu }_1+\sqrt{{\mu }_1^2+{\omega }^2{\mu }_2^2})}}.$$ (6)

Cross-spectrum - Phase Gradient Method

In USWE, a 0.1s harmonic vibration at a frequency is generated on the skin and the resulting time response of the skin is measured using an ultrasound probe. Let s₁(t) and s₂(t) represent the displacement responses at two locations on the skin, the phase change of surface wave propagation over the two locations can be calculated with a cross-spectrum method. The cross-spectrum S(f) of two signals s₁(t) and s₂(t) is defined as (Hasegawa and Kanai 2006),

$$S(f)=S_1^{\ast }(f)·S_2(f)=|S_1(f)·S_2(f)|·e^{-j\mathrm{\Delta }\phi (f)},$$ (7)

where S₁(f) and S₂(f) are the Fourier transforms of s₁(t) and s₂(t), respectively; * denotes the complex conjugate and Δϕ(f) is the phase change between s₁(t) and s₂(t) over distance at frequency f.

The phase change of surface wave with distance is used to measure the surface wave speed,

$$c_s=2\pi f|\mathrm{\Delta }r/\mathrm{\Delta }\varphi |,$$ (8)

where Δr is the radial distance of two measuring locations, Δϕ is the wave phase change over distance, and f is the frequency.

The estimation of wave speed can be improved by using multiple phase change measurements over distance. The regression of the phase change Δϕ with distance Δr can be obtained by “best fitting” a linear relationship between them, and the equation is

$$\mathrm{\Delta }\varphi =\alpha \mathrm{\Delta }r+\beta ,$$ (9)

where Δϕ denotes the value of Δϕ on the regression for a given distance of Δr, and α is the regression parameter.

The surface wave speed can be estimated by

$$c_{\mathit{\text{sr}}}=2\pi f|\mathrm{\Delta }r/\mathrm{\Delta }\varphi |=2\pi f/\alpha ,$$ (10)

where c_sr is the estimation of wave speed from the regression analysis.

Human Study Protocol

Human studies were approved by the Mayo Clinic Institutional Review Board (IRB). Each participant completed an informed consent form. Patients were enrolled in this research based on their clinical diagnoses. 41 patients with SSc and ILD were enrolled from Mayo Clinic Departments of Rheumatology and Pulmonary and Critical Care Medicine. These patients are confirmed ILD patients with clinical assessments together with pulmonary function test and high resolution CT scans. These patients also have confirmed clinic diagnosis of SSc. The American College of Rheumatology (ACR) (1980) developed classification criteria for SSc in 1980. The diagnosis requires either 1) the major criterion of proximal scleroderma, as judged by palpating or simply observing the skin; or 2) 2 minor criteria such as sclerodactyly, digital pitting scars, or loss of substance from the finger pad, and bibasilar pulmonary fibrosis. The patients enrolled in this study typically have advanced disease because fibrosis already involves the two organs of skin and lung. Further analysis of disease severity and grading using the surface wave speed measures is being carried out and will be published soon. Patients’ mean age was 61.88 years (range 37–82, 13 male and 28 female). 30 healthy subjects were enrolled as controls if they did not have any skin or lung diseases. Controls’ mean age was 45.43 years (range 22–73, 14 male and 16 female).

A subject was tested in a sitting position while his/her left or right forearm and upper arm were placed horizontally on a pillow in a relaxed state. The skin of both left and right forearms and upper arms of subjects was tested (Fig. 2). These locations were in the central part of the arms and on the dorsal sides. A 0.1-second harmonic vibration was generated by the indenter of the handheld shaker (Model: FG-142, Labworks Inc., Costa Mesa, CA 92626, USA) on the skin of the forearm or upper arm of the subject. The excitation force from the indenter was much less than 1 Newton and the subject only felt a small vibration on his/her skin. A Verasonics ultrasound system (Verasonics V1, Verasonics, Inc., Kirkland, WA 98034, USA) with an ultrasound probe of L11-4 with a central frequency of 6.4 MHz was used for detecting the surface wave motion of the skin. The probe had an elevation aperture of 38.4 mm and was customized to provide a 50-mm focal depth in the elevation axis. Elements were spaced at 30 µm intervals along the azimuthal axis. All images were acquired at a 50 V transmission voltage. Images of the skin and lung tissue were acquired by compounding 11 successive angles at a pulse repetition frequency (PRF) of 2 kHz. In order to improve the imaging quality of skin, an ultrasound gel pad standoff by Aquaflex® (Parker Laboratories, Inc., Fairfield, NJ 07004, USA) was placed between probe and skin. Surface wave speed on the skin was measured by determining the change in wave phase with distance along the skin. Figure 3 shows representative B-mode images of the skin for a patient and a healthy subject. The top dark area of the image shows the Aquaflex® standoff gel pad. On the skin surface, eight locations were used to measure the surface wave speed of skin. The skin motion velocity was in response to the external vibration excitation induced by the handheld vibrator. Using the skin motion at the first location as a reference, the wave phase delay of the skin motions at the remaining locations, relative to the first location, was used to measure the surface wave speed. The surface wave speed was measured at three excitation frequencies of 100 Hz, 150 Hz, and 200 Hz. Three measurements were performed at each location and at each frequency. A small tissue motion in tens of µm was enough for sensitive ultrasound detection of the generated tissue motion. The 100 Hz excitation signal is stronger than those of higher frequency excitations. The higher frequency waves have smaller wave length but decay more rapidly over distance than the lower frequency waves. The frequency ranges chosen in this study consider the wave motion amplitude, spatial resolution, and wave attenuation.

Figure 2.

Schematic of skin locations tested on human subjects. Square box indicates the tested positions. Probe orientation is aligned with the arm longitudinal axis. Left and right forearms and upper arms were labelled as L4, L5, R4, and R5.

Figure 3.

Representative B-mode images of skin for a patient (a) and a healthy subject (b). An ultrasound gel pad standoff was used to improve imaging of the skin. The top dark area in the image was associated with the gel pad. Eight locations on the skin surface were used to measure the surface wave speed of skin.

We studied repeatability and reproducibility of the USWE measurements on the forearm of a healthy subject. Ten measurements of wave speed were performed at 100 Hz, 150 Hz, and 200 Hz. The inter-rater reliability was evaluated by two rater’s measurements. The inter-class correlations (ICCs) were 0.94, 0.96, and 0.98 for wave speeds at 100 Hz, 150 Hz, and 200 Hz, respectively. The intra-rater reliability was evaluated by the same rater’s measurements at one day and one week later. At that time, the ICCs were 0.95, 0.91, and 0.95 for wave speeds at 100 Hz, 150 Hz, and 200 Hz, respectively. In these measurements, there were no statistical differences of the mean measurements between the intra-rater and inter-rater data. Reproducibility and repeatability is considered “good” for ICCs ≥0.60 and ≤0.74, and “excellent” for ICCs ≥0.75 and ≤1.00 (Cicchetti 1994).

Both lungs of the subject were tested through six intercostal spaces. The upper anterior lungs were tested at the second intercostal space in the mid-clavicular line. The lower lateral lungs were tested at one intercostal space above the level of the diaphragm in the mid-axillary line. The lower posterior lungs were tested at one intercostal space above the level of the diaphragm in the mid-scapular line. The indenter of the handheld shaker is placed on the chest wall in an intercostal space. The same ultrasound system and probe were used for lung testing. The ultrasound probe is positioned about 5 mm away from the indenter in the same intercostal space to measure the resulting surface wave propagation on the lung. The lung was tested at total lung capacity in which the subject took a deep breath and held for a few seconds. Ultrasound imaging was used to identify the lungs and select appropriate intercostal spaces to measure the upper and lower lungs. Three measurements were performed at each location and at each frequency. Typical testing lasted about 30 minutes.

In lung testing, a direct vibration excitation on the lung surface is not possible. Instead, the surface wave propagation on the lung is induced by a vibration excitation on the chest wall. The resulting wave propagates through the intercostal muscle and on the surface of the lung. We previously demonstrated that surface wave propagation on the lung can be generated by a vibration excitation on the surface of muscle in an ex vivo muscle-lung model (Zhang, et al. 2016).

Statistical analysis

An unpaired, two-tailed t-test between the healthy subjects and patients was conducted to compare the sample means. Differences in mean values were considered significant when p<0.05.

RESULTS

Figure 4 shows a representative wave speed at 100 Hz for a healthy subject and a patient at the same location. The representative surface wave speeds were, respectively, 3.69 m/s and 1.97 m/s for the patient and the healthy subject. Three measurements were made for each frequency and at each location. Table 1 shows representative measurements of surface wave speed for a healthy subject and a patient. The surface wave speed is shown in the format of mean ± SD for the three measurements at each location and each frequency. The forearm and upper arm are designated by numbers 4 and 5, respectively. The right and left arms are designated by letters R and L, respectively. Therefore, R4 represents the skin of the right forearm.

Figure 4.

The wave phase delay of the remaining locations, relative to the first location, is used to measure the surface wave speed. The slope of phase delay with distance measures the wave speed using equation (7). Representative examples of wave speed at 100 Hz for a patient (a) and a healthy subject (b).

Table 1.

Representative surface wave speeds of a healthy subject and a patient at four locations of skin at 100 Hz, 150 Hz and 200 Hz.

Healthy [m/s]	R4	R5	L4	L5
100 Hz	1.97 ± 0.31	1.7 ± 0.28	1.83 ± 0.04	1.79 ± 0.07
150 Hz	2.31 ± 0.08	2.5 ± 0.13	2.33 ± 0.25	2.52 ± 0.41
200 Hz	2.82 ± 0.19	2.99 ± 0.18	3.04 ± 0.37	3 ± 0.19
Patients	R4	R5	L4	L5
100 Hz	3.42 ± 0.27	2.22 ± 0.25	2.82 ± 0.31	2.47 ± 0.22
150 Hz	4.52 ± 0.59	2.52 ± 0.35	3.6 ± 0.38	3.67 ± 0.25
200 Hz	5.17 ± 0.57	3.04 ± 0.23	5.36 ± 0.55	5.78 ± 0.64

A comparison of wave speeds between 30 healthy subjects and 41 patients is shown in Figure 5 for 100 Hz, 150 Hz, and 200 Hz. The p-values for the t-test were less than 0.0001 for all four locations and three frequencies between the patients and controls, respectively.

Figure 5.

Comparison of wave speeds between healthy subjects and patients at four locations. Surface wave speeds at (a) 100 Hz, (b) 150 Hz, (c) 200 Hz. The p-values were less than 0.0001 for all four locations and three frequencies between the patients and controls, respectively.

Figure 6 shows the comparison of elasticity and viscosity of skin tissue between the 30 healthy subjects and 41 patients. Viscoelasticity is estimated using equation (6) with wave speed measurements at 100 Hz, 150 Hz, and 200 Hz. Most soft tissues have a mass density close to 1.0 g/cm³. In this paper, the mass density of skin is assumed to be 1.0 g/cm³. The p-values for the t-test were less than 0.0001 for the four locations between the patients and controls. Therefore, the magnitudes of both elasticity and viscosity of patients were statistically higher than those of healthy subjects.

Figure 6.

Comparison of elasticity μ1 (a) and viscosity μ2 (b) between healthy subjects and patients at four locations. The p-values were less than 0.0001 for the four locations between the patients and controls.

Figure 7 shows representative ultrasound images of superficial lung tissue for a patient and a healthy subject. The lung surface of a healthy subject is typically smooth while a patient’s lung surface is relatively rough. Table 2 shows representative measurements of surface wave speed of lung for a healthy subject and a patient. The three intercostal spaces are designated by a number from 1 to 3. The upper anterior lung is designated by 1. The lower lungs at the lateral and posterior positions are designated by 2 and 3, respectively. The right and left side of the lung are designated by letters R and L, respectively. Therefore, L1 represents the left anterior lung in the second intercostal space.

Figure 7.

Representative B-mode images of a lung for a patient (a) and a healthy subject (b). The lung surface of a healthy subject is typically smooth while a patient’s lung surface is relatively rough.

Table 2.

Representative surface wave speeds of a healthy subject and a patient through six intercostal spaces at 100 Hz, 150 Hz and 200 Hz.

healthy [m/s]	R1	R2	R3	L1	L2	L3
100	1.98 ± 0.07	1.83 ± 0.26	2.01 ± 0.23	2.07 ± 0.39	2.09 ± 0.3	1.95 ± 0.32
150	2.63 ± 0.46	2.65 ± 0.15	2.56 ± 0.16	2.61 ± 0.16	2.68 ± 0.12	2.52 ± 0.42
200	3.18 ± 0.58	3.28 ± 0.25	3.14 ± 0.13	2.98 ± 0.36	3.03 ± 0.2	3.1 ± 0.45
patient [m/s]	R1	R2	R3	L1	L2	L3
100	3.26 ± 0.41	2.78 ± 0.34	3.25 ± 0.15	2.8 ± 0.39	2.79 ± 0.95	2.97 ± 0.26
150	3.86 ± 0.43	3.59 ± 0.41	4.36 ± 0.43	3.35 ± 0.61	3.13 ± 0.85	3.57 ± 0.02
200	4.19 ± 1.61	4.57 ± 0.13	5.79 ± 0.64	5.51 ± 0.51	4.63 ± 1.27	4.35 ± 1.03

A comparison of wave speeds of lung surface between 30 healthy subjects and 41 patients is shown in Figure 8 for 100 Hz, 150 Hz, and 200 Hz. The p-values for the t-test were less than 0.0001 for all intercostal spaces and for three frequencies between the patients and controls, respectively.

Figure 8.

Comparison of wave speeds between healthy subjects and patients through six intercostal spaces. Surface wave speeds were measured at (a) 100 Hz, (b) 150 Hz, (c) 200 Hz. The p-values were less than 0.0001 for all intercostal spaces and for three frequencies between the patients and controls.

DISCUSSION

The aim of this study was to translate an ultrasound surface wave elastography (USWE) technique into clinical use for quantitatively assessing patients with SSc and ILD. A high pulse repetition rate of 2000 frame/s was used to detect tissue motion in response to the excitations of 100, 150, and 200 Hz. A Verasonics ultrasound system was used to collect up to a few thousand imaging frames per second by using a plane-wave pulse transmission method. The skin motion velocities at these locations were measured in the normal direction of skin using the ultrasound tracking beams through those locations (Hasegawa and Kanai 2006, Zhang 2011). The wave speed on the skin was measured by analyzing ultrasound data directly from the skin. Therefore, the wave speed measurement was local and independent of the location and amplitude of excitation. The surface wave speeds of skin of both left and right forearms and upper arms of both healthy controls and SSc patients were tested at three excitation frequencies and their viscoelasticity were evaluated. The surface wave speeds of lung at six intercostal locations were measured at three excitation frequencies between healthy controls and SSc patients.

The results obtained in this study are in agreement with the published results in literature. We found the viscoelasticity of SSc patients is statistically higher than that of healthy controls. Skin of the peri-oral region of SSc patients was stiffer than that of healthy controls (Cannaò, et al. 2014). Our earlier pilot work using an optical-based technique indicated that skin viscoelasticity was a more sensitive measure than palpation for assessing SSc (Zhang 2011). Our technology was favorably reviewed by Yale dermatologists (Lott and Girardi 2011), who commented that this novel technology could benefit patients through new multimodal paradigms for assessing SSc. Ultrasound elastography has been used to assess skin elasticity between SSc patients and healthy controls (Iagnocco, et al. 2010). However, it only provides semi-quantitative scales to characterize the elasticity of skin and it is difficult to differentiate patients from healthy subjects. USWE quantifies the viscoelasticity of skin tissue for SSc patients and provides accurate biomarkers for evaluating disease progression.

Optical coherence tomography (OCT) (Abignano, et al. 2013, Shazly, et al. 2015) and OCT-based elastography techniques (Li, et al. 2014, Nguyen, et al. 2014) provide high-spatial resolution of skin but cannot measure deep subcutaneous tissue. The imaging penetration of OCT in skin can be up to 1.5 mm (Liang and Boppart 2010). Fibrosis not only affects skin but also subcutaneous tissue (Li, et al. 2007). One advantage of USWE is its capability to measure subcutaneous tissue (Kubo, et al. 2017). In the current setup, subcutaneous tissues can be measured up to 45 mm (Figure 2, 6). USWE may be used to assess skin, subcutaneous connective tissue, and muscle for patients with SSc.

Most ultrasound-based elastography techniques use ultrasound radiation force (URF) to generate tissue motion. Acoustic radiation force impulse imaging (ARFI) has been recently used to assess skin fibrosis (Hou, et al. 2015, Lee, et al. 2015). To generate sufficient tissue motion using URF, relatively high-intensity ultrasound energy is needed. Although URF has been used in most organs including the liver, URF should not be applied to the lung. In vivo animal lung studies demonstrated that the relatively high-intensity ultrasound energy may cause alveolar hemorrhage or lung injury (Zachary, et al. 2006). In addition, long periods of high-intensity ultrasound may cause damage to the ultrasound system, e.g., as high voltage drop and probe element damage. In USWE, the surface wave on the lung is safely generated by a local mechanical vibration on the chest. Diagnostic ultrasound is only used for detecting surface wave propagation on the lung. Therefore, the USWE technique is a safe method for lung testing and screening patients. Moreover, because URF cannot be directly generated on surface tissue, a standoff pad is needed for exciting the skin. However, the standoff pad decays the URF and complicates the URF on the skin. In USWE, the mechanical excitation is directly applied to the skin and the standoff pad is only used to improve imaging of the skin. USWE provides a safe and simple way to generate and detect wave propagation on the skin. One advantage of USWE compared with most ultrasound elastography techniques is that the wave propagation is generated by a 0.1 second harmonic excitation rather than a short pulse using URF. The wave speed is measured quickly and accurately at a frequency with high signal-to-noise ratio. The tissue motion can be safely generated at a level of 20 µm using USWE. The radiation force ultrasound generated tissue notion is typically about 1~2 µm.

In USWE, the standoff pad is only used to improve imaging of the skin. We do not use the probe to generate the pressure on the skin such as for getting a tissue stain imaging. In our testing, the gel pad sits on the skin surface naturally. The ultrasound probe is attached to the gel pad using ultrasound gel. We keep this experimental setup for each patient. We do not study the pressure effect on the skin measurements. However, we may investigate it by carefully controlling the generated pressure on the skin and measuring the associated surface wave speed changes. This may be useful to study nonlinear or hyperelastic properties of the skin.

The heart beat or tissue movement should not affect our skin and lung testing. The lung is tested at total lung volume when a subject takes a deep breath and holds for a few seconds. The arm’s skin is tested at normal breathing and the arm is rested on a pillow. We measure the surface wave speed at a given excitation frequency between 100 Hz and 200 Hz, which is typically higher than the frequency of a heart beat or tissue movement. We use triggering for acquisition. The trigger is used to synchronize the ultrasound wave measurement with the external vibration excitation.

In this study, we measured both skin and lung surface wave speeds of patients with SSc and ILD. High-resolution computed tomography (HRCT) is the clinical standard for diagnosing and characterizing lung fibrosis (Mathieson, et al. 1989, Verschakelen 2010). However, HRCT involves ionizing x-ray radiation exposure to patients. USWE provides a noninvasive and safe method for measuring superficial lung tissues for assessing lung fibrosis (Kalra, et al. 2017, Zhang, et al. 2017). USWE may be useful as a screening tool to assess patients with ILD and compliment HRCT as a follow up tool for assessing the progression of the disease.

We cannot find the data of mass density of the lungs of ILD patients. The lung density is also dependent on the pulmonary pressure. In this research, a subject is tested at total lung capacity (TLC) when taking a deep breath and holding. In a paper using an x-ray technique (Garnett, et al. 1977), lung density was averaged to be 0.32 g/cm³ for healthy lungs. The measurements were made with the patient sitting either on a chair or in bed and breathing quietly. Lung density was 0.33–0.93 g/cm³ for patients with pulmonary congestion and edema. In an ex vivo study on sheep lungs (Jahed, et al. 1989), the lung density of 0.19–0.26 g/cm³ was used. We expect that the lung density of ILD would be higher than that of healthy lungs. However, we were not able to find the data of lung density of ILD patients. We do not calculate lung viscoelasticity from the surface wave speeds in this paper. The surface wave speed alone provides good separation between the healthy subjects and the patients with ILD in this research. We are very interested in studying the lung density of patients with ILD and the variation of lung density with the pulmonary pressure.

USWE provides a novel, safe, noninvasive, and nonionizing technique for measuring skin and lung stiffness which may be useful for quantitative and effective assessment of both skin and lung diseases. We will further test the clinical utility of USWE in multiple fibrotic disorders. Many systemic diseases, including systemic sclerosis (SSc) (Steen and Medsger 2000), graft-vs-host disease (GVHD) (Hausermann, et al. 2008), peripheral vascular disease (Spentzouris and Labropoulos 2009), and diabetic sclerosis (Van Hattem, et al. 2008) are associated with skin stiffening due to fibrosis.

CONCLUSION

USWE provides a noninvasive and nonionizing technique to measure the surface wave speeds of both skin and superficial lung tissue. In this study, USWE was used to measure both skin and lung surface wave speeds for 41 patients with both SSc and ILD and 30 healthy subjects. The surface wave speeds were measured at four locations on the arms for three excitation frequencies of 100 Hz, 150 Hz, and 200 Hz. Viscoelasticity of the skin was calculated by the wave speed dispersion with frequency using the Voigt model. The magnitudes of surface wave speed and viscoelasticity of patients’ skin were significantly higher than those of healthy subjects (p<0.0001). The upper and lower lungs were measured through six intercostal spaces for patients and healthy subjects. The magnitudes of surface wave speed of patient’s lung were significantly higher than those of healthy subjects (p<0.0001) for each location and each frequency. USWE may be useful for quantitative assessment of SSc and/ or ILD. We will investigate a larger population of ILD patients to study the relationship between skin stiffness and lung stiffness for patients with and without SSc.