An Approach for Needle Based Optical Coherence Elastography Measurements

Abstract

While navigation and interventional guidance are typically based on image data, the images do not necessarily reflect mechanical tissue properties. Optical coherence elastography (OCE) presents a modality with high sensitivity and very high spatial and temporal resolution. However, OCE has a limited field of view of only 2–5 mm depth. We present a side-facing needle probe to image externally induced shear waves from within soft tissue. A first method of quantitative needle-based OCE is provided. Using a time of flight setup, we establish the shear wave velocity and estimate the tissue elasticity. For comparison, an external scan head is used for imaging. Results for four different phantoms indicate a good agreement between the shear wave velocities estimated from the needle probe at different depths and the scan head. The velocities ranging from 0.9–3.4 m/s agree with the expected values, illustrating that tissue elasticity estimates from within needle probes are feasible.

1 Introduction

While medical imaging is a prerequisite for navigation and guidance of interventional procedures, the respective gray values do not necessarily reflect mechanical tissue properties. Yet, elastic tissue properties are of interest in a number of scenarios, including palpation to identify lesions. Different tissue elasticity also affects needle based interventions, e.g., experienced physicians often feel what type of tissue they are penetrating. Particularly, needle-tissue interaction often results in tissue deformation or needle deflection [1].

Different approaches have been proposed to establish elastic tissue properties from image data. Typically, an externally induced compression or vibration is employed and the tissue response is obtained from interpretation of ultrasound or magnetic resonance imaging signals. Another approach is based on optical coherence tomography (OCT) with its high spatial and temporal resolution. Hence, optical coherence elastography (OCE) is a promising tool to analyze the tissue micro structure.

In recent OCE studies, mechanical waves are excited within tissue samples and the transverse shear wave component is measured using an external OCT probe. For example, an acoustic wave [2, 3], an oscillating piezoelectric actuator, or a vibrator is used to emit the mechanical waves. In [3], the induced shear wave propagates along the OCT beam and the shear wave velocity is estimated by means of an intensity based Doppler variance imaging [4]. In [5], a four channel OCT system is used to determine velocity of the perpendicular travelling shear wave as function of time and distance from the mechanical wave origin. In [6] a forward facing OCE needle probe is used in order to measure tissue deformation under load, where quantitative elasticity estimates are not obtained.

While Zhu et al. demonstrated that acoustic radiation force excitation is feasible [3], so far the OCT images have been obtained from outside the tissue. Hence, OCE is severely limited by the small imaging depth of typically less than 2 mm in scattering tissue. In this study, we propose an approach to measure the shear wave velocity from within the tissue sample. A side-facing needle probe with an outer diameter of 0.81 mm is presented. OCE using the needle at different depths is compared to OCT images of a scan head. The shear wave velocities obtained with these setups are compared. Considering a range of typical soft tissue properties [7] we demonstrate that OCE from within a needle probe is feasible.

2 Materials and Methods

2.1 Shear Wave Propagation

Shear wave propagation in soft tissues is described by wave equation

$$\begin{aligned} \rho \frac{\partial ^2 \varvec{u}}{\partial t^2} = \mu \nabla ^2 \varvec{u} + (\lambda + \mu ) \nabla (\nabla \varvec{u}) \end{aligned}$$

(1)

where the density $\rho$, the shear modulus $\mu$, and the Lame constant $\lambda$ are related to the material properties of the tissue [8]. Assuming a divergence free field the transverse component $\varvec{u}_\text {T}$ follows as

$$\begin{aligned} 0 = (\rho \frac{\partial ^2}{\partial t^2} - \mu \nabla ^2) \varvec{u}_\text {T} \end{aligned}$$

(2)

Hence, shear wave propagation and elasticity are related, i.e., for a known density $\rho$ the shear wave velocity $c_\text {s}$ can be expressed with respect to the shear modulus $\mu$

$$\begin{aligned} c_\text {s} = \sqrt{\frac{\mu }{\rho }} \end{aligned}$$

(3)

Using the definition of the elastic modulus E and assuming a Poisson’s ratio of $\nu = 0.5$ for soft tissues [3] the elasticity can be determined as

$$\begin{aligned} E = 2\mu (1+\nu ) = 2 c_\text {s}^2\rho (1+\nu ) = 3 \rho c_\text {s}^2. \end{aligned}$$

(4)

Fig. 1.

Left: spherical shear wave propagation. Starting at position $x_p$ the wave subsequently passes through points $x_i$. Right: Resulting wave amplitude at position $x_0$ over time. The wave front $u_0$ is followed by the reflected waves $u_1$ and $u_2$ (dashed).

Fig. 2.

Intensity M-scan (A), phase M-scan (B), and estimated relative phase amplitude (C) for scan head measurement. In (A) and (C), the shear wave front is shown as a dashed line, reflections are highlighted with gray background.

We measure $c_\text {s}$ using an optical setup. A sinusoidal burst signal of frequency $f_\text {p}=50$ Hz and resulting amplitude $d = 60\,\upmu$m is used to initiate a shear wave propagating through the tissue. We assume a spherical wave with constant velocity $c_\text {s}$. Figure 1 illustrates that the initial wave front is followed by subsequent waves of the burst and reflections. Hence, the tissue motion at different distances from the excitation point $x_\text {p}$ can be detected, and the run-time of the wave can be used to obtain its average velocity. However, we do not know the exact point of excitation and the measurements are subject to system latencies. Therefore, we measure at multiple points along one line and consider differences.

2.2 Optical Shear Wave Detection

We employ OCT to measure the small tissue motion due to the shear wave. Particularly, Doppler mode OCT is sensitive to small changes in the phase signal and the run-time $t_\text {s}$ between excitation and shear wave appearance in the OCT phase signal is obtained as follows. First, the actuator trigger timestamp $t_\text {0}$ is established. Second, the OCT system trigger is read and used to assign a timestamp $t_\text {j}$ for all $j=1,\ldots ,n$ OCT A-scans. Finally, the OCT phase signal is filtered and the first A-scan s exceeding a predefined threshold is identified, such that $t_s=t_\text {j}-t_\text {0}$ is the measured system run-time, including latency.

During Doppler OCT measurements we acquire A-scans successively at the same measurement position $x_i$. The phase differences are estimated for two consecutive A-scans. Figure 2 illustrates the resulting intensity and phase M-scans showing a shear wave pattern. Based on the phase M-scan, the relative amplitude of the phase is determined and the first maximum exceeding the threshold is considered the wave front, compare Fig. 2.

2.3 Needle Probe Design

In the past, designs of fiber probes based on a spacer, a GRIN-fiber, and a prism were presented [9]. Our imaging probe is made from a single mode fiber (SMF-28, Thorlabs) as the light guiding fiber. For widening the beam, a step index multimode fiber (FG200LEA, Thorlabs) with a cladding diameter of 220 m was spliced to the SMF-28 fiber and precisely cleaved at a length of 600 m with an Automated Glass Processor (GPX3800, Thorlabs). In the same way, the GRIN fiber (G200/220, Fiberware) with a numerical aperture 0.22 and a cladding diameter of 220 m was attached. For deflecting the beam at $90^{\circ }$ a second piece of multimode fiber (FG200LEA, Thorlabs) is spliced to the end of the GRIN-fiber and polished at an angle of $45^{\circ }$ (Fig. 3). The fiber probe is covered by a tube and glued into our 21 gauge needle probe with a drilled imaging window.

Fig. 3.

Schematic and images of the side-facing needle probe.

2.4 Experimental Setup and Calibration

Our experimental setup consists of the needle probe, a piezoelectric actuator to realize the excitation, and a container for gelatine tissue phantoms. In order to compare the needle based measurements to conventional surface scanning, we also included the OCT scan head of a commercially available OCT device (Telesto I, Thorlabs). For calibration purposes and in order to obtain measurements at well defined distances, the needle probe is attached to a micro-motion stage while the OCT scan head is positioned using a hexapod. The OCT scan direction is normal to the tissue surface. Either the needle probe or the scan head is attached to the OCT system. An amplified burst signal of a function generator is used to control the piezo actuator. The trigger signals are acquired by a System-on-Chip (SoC) oscilloscope. The overall setup is shown in Fig. 4.

Fig. 4.

Experimental setup: gelatine tissue sample (D) positioned underneath OCT scan head (B); piezoelectric actuator (C) positioned on the tissue surface. The needle probe (A) is inserted along the x-axis into the phantom. Scan head and needle A-scans are aligned normal to the surface shown as black dashed line. A sinusoidal burst is generated and continuously applied to the piezoelectric actuator. A System-on-Chip (SoC) oscilloscope records the signals of the function generator and the OCT trigger.

Different gelatine tissue phantoms were placed in containers of known geometries and fixation points. In order to align the A-scans recorded with the scan head and the needle probe perpendicular to a desired measurement line a 3D printed calibration rig is used. The rig is visible in OCT image data and the webcam integrated in the scan head, and it is also used to align the actuator before it is retracted by 15 mm using the motion stage. Likewise, the needle probe is manually aligned to the desired z-layer and automatically moved forward using a stepper motor.

2.5 Experimental Parameters

We used gelatine phantoms of different density and with added TiO$_2$ (Table 1). For each phantom, we acquired data with the scan head and with the needle probe at three different depths ($\varDelta z_1 = 10$ mm, $\varDelta z_2 = 6$ mm, $\varDelta z_3 = 4$ mm) in the tissue (Fig. 5). The run-time $t_s$ was established at 21 distances ($15.75 \le x_i \le 24.25$) from the piezo actuator. At each position $x_i$ five Doppler OCT M-scans with $N=5000$ A-scans were acquired. Applying OCT frequency $f_\text {OCT} = 5.5$ kHz and actuator modulation frequency $f_\text {mod} =9$ Hz, during every M-scan allocation two shear waves are excited. Each series was repeated three times.

Fig. 5.

Pictures of the experimental setup. (Left) The needle probe micro-motion stage (A) provides an accurate guidance of the needle probe. The scan head (B) is mounted on a hexapod robot (E) to afford a precise positioning. An additional micro-motion stage is used to arrange the piezoelectric actuator (C). The phantom container (D) is fixed on the optical table. (Right) OCE needle probe inserted in the phantom container. The insertion heights $\varDelta z_i$ are highlighted.

3 Results

We evaluate the time of flight for the described measurement series and determine the shear wave velocity based on the slope of the time difference $t_s$ over measurement points $x_i$. In Fig. 6 for the needle placed at $\varDelta z_1 = 10$ mm the resulting mean and standard deviation values of the estimated time $t_s$ are shown exemplary. The slope of the data points increases with decreased phantom density $\rho _i$.

Fig. 6.

Mean and standard deviation of time differences $t_s$ over distances $x_i$ measured with the needle probe at depth $\varDelta z_1 = 10$ mm. The slopes of the mean values correspond to the shear wave velocity $c_\text {s}$ [m/s] of the imaged phantoms with densities $\rho _i$. With increasing phantom density the shear wave velocity increases.

Table 1. We image four phantoms with different densities. The gelatine concentrations (C in $\%$) and related densities ($\rho$ in [g/ml]) are listed below. Estimated mean shear wave velocity ($c_\text {s}$ in [m/s]) with standard deviations $\sigma$ and elastic modulus (E in [kPa]) are listed for the imaged phantoms $\rho _i$.

In Fig. 7 the estimated mean shear wave velocities are shown for different gelatine density values. An average slope is determined from the measurements in different depths. The largest deviations from the estimated mean slope are visible for phantom 1.

Finally, using the estimated shear wave velocities the related elasticity moduli of phantoms $E_i$ are determined and listed in Table 1. The estimated elasticity moduli range from 0.1 to 3.16 kPa.

4 Discussion

The proposed setup leads to shear wave velocities in the range of $c_\text {s} =$$0.9-3.4$ m/s for gelatine phantoms with concentrations ranging from 4 to $10\%$. These values are in good agreement with the literature [7]. Figure 7 indicates a linear relationship between shear wave velocity and gelatine concentration. However, with increasing density the standard deviation of the measured shear wave velocity increases. As for higher density of the gelatine, the shear wave velocity increases and phase wrapping occurs, i.e., wave front detection may fail. Hence, the amplitude of the piezoelectric actuator should be adapted for the densities.

Fig. 7.

Resulting shear wave velocity $c_\text {s}$ as function of gelatine concentration measured with OCE needle in different depths $\varDelta z_i$ and scan head. According to the assumed linear behaviour of concentration and velocity a line is fitted (slope).

The high density phantom 1 shows the largest difference in velocities measured through needles and scan head (Table 1). In contrast, the values for the phantoms 3 and 4 agree rather well at all needle depths.

Moreover, the differences are considerably less for the lower needle positions.

Combining the needle probe with remote excitation, e.g., using acoustic radiation force [3] would provide a new tool to estimate tissue elasticity to guide needle based interventions.

5 Conclusion

We propose a method to measure shear wave velocities and related elasticity moduli using an OCE needle probe from within the tissue. Even in larger tissue depths realistic shear wave velocity values are determined. In combination with shear wave excitation in the needle’s proximity the approach can be employed to determine quantitative elasticity properties in needle based interventions.