Thank you for visiting Nature.com. The browser version you are using has limited CSS support. For the best experience, we recommend that you use an updated browser (or disable Compatibility Mode in Internet Explorer). In the meantime, to ensure continued support, we will render the site without styles and JavaScript.

It has recently been demonstrated that the use of ultrasound increases tissue yield in ultrasound-assisted fine needle aspiration (USeFNAB) compared to conventional fine needle aspiration (FNAB). To date, the relationship between bevel geometry and tip movement has not been thoroughly studied. In this study, we investigated the properties of needle resonance and deflection amplitude for various needle bevel geometries with different bevel lengths. Using a conventional 3.9 mm beveled lancet, the tip deflection power factor (DPR) in air and water was 220 and 105 µm/W, respectively. This is higher than the axisymmetric 4mm beveled tip, providing 180 and 80 µm/W DPR in air and water, respectively. This study highlights the importance of the relationship between the bending stiffness of the bevel geometry in the context of different means of insertion, and therefore may provide insight into methods for controlling post-piercing cutting action by changing the needle bevel geometry, which is important. for a USeFNAB application is critical.

Fine-needle aspiration biopsy (FNA) is a method of obtaining tissue samples for suspected pathology1,2,3 using a needle. The Franseen tip has been shown to provide higher diagnostic performance than conventional lancet4 and Menghini5 tips. Axisymmetric (i.e. circumferential) slopes are also suggested to increase the likelihood of histopathologically adequate specimens.

During a biopsy, a needle is passed through layers of skin and tissue to gain access to suspicious lesions. Recent studies have shown that ultrasound can reduce the penetration force required to access soft tissues7,8,9,10. Needle bevel geometry has been shown to affect needle interaction forces, for example, longer bevels have been shown to have lower tissue penetration forces11. After the needle has penetrated the surface of the tissue, i.e. after puncture, the cutting force of the needle can be 75% of the interaction force of the needle with the tissue12. It has been shown that in the post-puncture phase, ultrasound (ultrasound) increases the efficiency of diagnostic soft tissue biopsy. Other ultrasound-enhanced bone biopsy techniques have been developed for taking hard tissue samples, but no results have been reported that improve biopsy yield. Numerous studies have also confirmed that mechanical displacement increases when subjected to ultrasonic stress16,17,18. While there are many studies on axial (longitudinal) static forces in needle-tissue interactions19,20, there are limited studies on the temporal dynamics and geometry of needle bevel under ultrasonic FNAB (USeFNAB).

The aim of this study was to investigate the effect of different bevel geometries on the movement of the needle tip in a needle driven by ultrasonic bending. In particular, we investigated the effect of the injection medium on needle tip deflection after puncture for traditional needle bevels (i.e., USeFNAB needles for various purposes such as selective aspiration or soft tissue acquisition.

Various bevel geometries were included in this study. (a) The Lancet specification complies with ISO 7864:201636 where \(\alpha\) is the primary bevel, \(\theta\) is the rotation angle of the secondary bevel, and \(\phi\) is the secondary bevel angle. , when rotating, in degrees (\(^\circ\)). (b) Linear asymmetrical single step chamfers (called “standard” in DIN 13097:201937) and (c) Linear axisymmetric (circumferential) single step chamfers.

Our approach starts by modeling the change in bending wavelength along the bevel for conventional lancet, axisymmetric, and asymmetric single-stage bevel geometries. We then calculated a parametric study to examine the effect of pipe slope and length on the mechanical fluidity of the transfer. This is necessary to determine the optimal length for making a prototype needle. Based on the simulation, needle prototypes were made and their resonant behavior was experimentally characterized by measuring the voltage reflection coefficients and calculating the power transfer efficiency in air, water and 10% (w/v) ballistic gelatin, from which the operating frequency was determined. Finally, high-speed imaging is used to directly measure the deflection of the bending wave at the tip of the needle in air and water, as well as to estimate the electrical power delivered at each oblique angle and the geometry of the deflection power ratio (DPR) to the injected medium. .

As shown in Figure 2a, use a 21 gauge tube (0.80 mm OD, 0.49 mm ID, tube wall thickness 0.155 mm, standard wall) to define the needle tube with tube length (TL) and bevel angle (BL) in accordance with ISO 9626:201621) in 316 stainless steel (Young’s modulus 205 \(\text {GN/m}^{2}\), density 8070 kg/m\(^{3}\) and Poisson’s ratio 0.275 ).

Determination of the bending wavelength and tuning of the finite element model (FEM) for needle and boundary conditions. (a) Determination of bevel length (BL) and pipe length (TL). (b) Three-dimensional (3D) finite element model (FEM) using a harmonic point force \(\tilde{F}_y\vec {j}\) to drive the needle proximally, deflect the point, and measure velocity at the tip (\ ( \tilde {u}_y\vec {j}\), \(\tilde{v}_y\vec {j}\)) to calculate the transfer of mechanical fluidity. \(\lambda _y\) is defined as the bending wavelength relative to the vertical force \(\tilde{F}_y\vec {j}\). (c) Definitions of the center of gravity, the cross-sectional area A, and the moments of inertia \(I_{xx}\) and \(I_{yy}\) around the x and y axes, respectively.

As shown in fig. 2b,c, for an infinite (infinite) beam with cross-sectional area A and at a wavelength greater than the beam’s cross-sectional size, the bent (or bent) phase velocity \( c_{EI }\) is determined by 22:

where E is Young’s modulus (\(\text {N/m}^{2}\)), \(\omega _0 = 2\pi f_0\) is the excitation angular frequency (rad/s), where \( f_0 \ ) is the linear frequency (1/s or Hz), I is the moment of inertia of the area around the axis of interest\((\text {m}^{4})\), \(m’=\ rho _0 A\ ) is the mass on unit length (kg/m), where \(\rho _0\) is the density\((\text {kg/m}^{3})\) and A is the cross section of the beam area (xy plane) (\(\ text {m}^{2}\)). Since the force applied in our example is parallel to the vertical y-axis, i.e. \(\tilde{F}_y\vec {j}\), we are only interested in the regional moment of inertia around the horizontal x-axis, i.e. \(I_{xx}\), so:

For the finite element model (FEM), a pure harmonic displacement (m) is assumed, so the acceleration (\(\text {m/s}^{2}\)) is expressed as \(\partial ^2 \vec { u}/ \ partial t^2 = -\omega ^2\vec {u}\) as \(\vec {u}(x, y, z, t): = u_x\vec {i} + u_y\ vec {j } + u_z\vec {k}\) is a three-dimensional displacement vector given in spatial coordinates. Instead of the latter, in accordance with its implementation in the COMSOL Multiphysics software package (versions 5.4-5.5, COMSOL Inc., Massachusetts, USA), the finite deformation Lagrangian form of the momentum balance law is given as follows:

where \(\vec {\nabla}:= \frac{\partial}}{\partial x}\vec {i} + \frac{\partial}}{\partial y}\vec {j} + \frac{ \partial }{\partial z}\vec {k}\) is the tensor divergence operator, \({\underline{\sigma}}\) is the second Piola-Kirchhoff stress tensor (second order, \(\ text { N/ m}^{2}\)) and \(\vec {F_V}:= F_{V_x}\vec {i}+ F_{V_y}\vec {j}+ F_{V_z}\vec {k}\) is the body force vector (\(\text {N/m}^{3}\)) for each deformed volume, and \(e^{j\phi }\) is the phase angle vector\(\ phi\ ) ( glad). In our case, the volume force of the body is zero, our model assumes geometric linearity and a small purely elastic deformation, i.e. , where \({\underline{\varepsilon}}^{el}\) and \({\underline{\varepsilon}}\) are elastic strain and total strain (second order, dimensionless), respectively. Hooke’s constitutive isotropic elasticity tensor \(\underline{\underline{C}}\) is computed using Young’s modulus E (\(\text {N/m}^{2}\)) and Poisson’s ratio v is determined, so i.e. \(\underline{\underline{C}}:=\underline{\underline{C}}(E,v)\) (fourth order). So the stress calculation becomes \({\underline{\sigma}} := \underline{\underline{C}}:{\underline{\varepsilon}}\).

The calculation uses a 10-node tetrahedral element with an element size \(\le\) of 8 µm. The needle is modeled in vacuum, and the value of the transferred mechanical mobility (ms-1 N-1) is defined as \(|\tilde{Y}_{v_yF_y}|= |\tilde{v}_y\vec { j}|/ |\ tilde{F}_y\vec {j}|\)24, where \(\tilde{v}_y\vec {j}\) is the output complex velocity of the handpiece and \( \ tilde{F}_y\ vec {j }\) is a complex driving force located at the proximal end of the tube, as shown in Figure 2b. Translate the mechanical fluidity in decibels (dB) using the maximum value as a reference, i.e. \(20\log _{10} (|\tilde{Y}|/ |\tilde{Y}_{max}|) \ ) . All FEM studies were carried out at a frequency of 29.75 kHz.

The design of the needle (Fig. 3) consists of a conventional 21-gauge hypodermic needle (Cat. No. 4665643, Sterican\(^\circledR\), outer diameter 0.8 mm, length 120 mm, AISI 304 stainless chromium-nickel steel , B. Braun Melsungen AG, Melsungen, Germany) equipped with a plastic Luer Lock sleeve made of polypropylene at the proximal end and suitably modified at the end. The needle tube is soldered to the waveguide as shown in Fig. 3b. The waveguides were printed on a stainless steel 3D printer (EOS 316L stainless steel on an EOS M 290 3D printer, 3D Formtech Oy, Jyväskylä, Finland) and then attached to the Langevin sensor using M4 bolts. The Langevin sensor consists of 8 piezoelectric ring elements loaded at both ends with two masses.

The four types of tips (photo), a commercially available lancet (L) and three manufactured axisymmetric single-stage bevels (AX1-3) were characterized by bevel lengths (BL) of 4, 1.2 and 0.5 mm, respectively. (a) Close-up of the finished needle tip. (b) Top view of four pins soldered to the 3D printed waveguide and then connected to the Langevin sensor with M4 bolts.

Three axisymmetric bevel tips (Fig. 3) were manufactured (TAs Machine Tools Oy) with bevel lengths (BL, as defined in Fig. 2a) of 4.0, 1.2 and 0.5 mm, corresponding to \(\approx) 2 \(^ \circ\), 7\(^\circ\) and 18\(^\circ\) respectively. The mass of the waveguide and needle is 3.4 ± 0.017 g (mean ± sd, n = 4) for bevels L and AX1-3, respectively (Quintix\(^\circledR\) 224 Design 2, Sartorius AG, Göttingen, Germany) . For the L and AX1-3 bevels in Figure 3b, the total length from the tip of the needle to the end of the plastic sleeve was 13.7, 13.3, 13.3, and 13.3 cm, respectively.

For all needle configurations, the length from the tip of the needle to the tip of the waveguide (i.e., to the weld area) was 4.3 cm, and the needle tube was oriented with the cut upwards (i.e., parallel to the Y axis), as shown in the figure. c (Fig. 2).

A custom script in MATLAB (R2019a, The MathWorks Inc., Massachusetts, USA) running on a computer (Latitude 7490, Dell Inc., Texas, USA) was used to generate a linear sinusoidal sweep from 25 to 35 kHz for 7 seconds, passing A digital-to-analog (DA) converter (Analog Discovery 2, Digilent Inc., Washington, USA) converts to an analog signal. The analog signal \(V_0\) (0.5 Vp-p) was then amplified with a dedicated radio frequency (RF) amplifier (Mariachi Oy, Turku, Finland). Falling amplified voltage \({V_I}\) from the RF amplifier with an output impedance of 50 ohms is fed to a transformer built into the needle structure with an input impedance of 50 ohms. Langevin transducers (front and rear heavy-duty multilayer piezoelectric transducers) are used to generate mechanical waves. The custom RF amplifier is equipped with a dual-channel standing wave power factor (SWR) meter that records the incident \({V_I}\) and reflected amplified voltage\(V_R\) in analog-to-digital (AD) mode. with a sampling rate of 300 kHz Converter (analogue Discovery 2). The excitation signal is amplitude modulated at the beginning and at the end to prevent overloading the amplifier input with transients.

Using a custom script implemented in MATLAB, the frequency response function (FRF), i.e. \(\tilde{H}(f)\), was estimated offline using a two-channel sinusoidal sweep measurement method (Fig. 4), which assumes linearity in time. invariant system. In addition, a 20 to 40 kHz band pass filter is applied to remove any unwanted frequencies from the signal. Referring to the theory of transmission lines, in this case \(\tilde{H}(f)\) is equivalent to the voltage reflection coefficient, i.e. \(\rho _{V} \equiv {V_R}/{V_I}\) \) decreases to \({V_R}^ 2 /{V_I}^2\ ) equals \(|\rho _{V}|^2\). In cases where absolute electrical power values are required, incident power \(P_I\) and reflected power \(P_R\) power (W) are calculated by taking the rms value (rms) of the corresponding voltage, for example. for a transmission line with sinusoidal excitation \( P = {V}^2/(2Z_0)\)26, where \(Z_0\) is equal to 50 \(\Omega\). The electrical power supplied to the load \(P_T\) (i.e., the inserted medium) can be calculated as \(|P_I – P_R |\) (W RMS), as well as the power transfer efficiency (PTE) and percentage ( %) can be determined how the shape is given, so 27:

The acicular modal frequencies \(f_{1-3}\) (kHz) and their corresponding power transfer factors \(\text {PTE}_{1{-}3} \) are then estimated using the FRF. FWHM (\(\text {FWHM}_{1{-}3}\), Hz) estimated directly from \(\text {PTE}_{1{-}3}\), from Table 1 A one-sided linear spectrum is obtained at the described modal frequency \(f_{1-3}\).

Measurement of the frequency response (AFC) of needle structures. A sinusoidal two-channel sweep measurement25,38 is used to obtain the frequency response function \(\tilde{H}(f)\) and its impulse response H(t). \({\mathcal {F}}\) and \({\mathcal {F}}^{-1}\) represent the Fourier transform of digital truncation and its inverse, respectively. \(\tilde{G}(f)\) means the product of two signals in the frequency domain, e.g. \(\tilde{G}_{XrX}\) means the inverse scan product\(\tilde{ X} r (f)\ ) and drop voltage \(\tilde{X}(f)\) respectively.

As shown in Figure 5, the high-speed camera (Phantom V1612, Vision Research Inc., NJ, USA) is equipped with a macro lens (MP-E 65mm, \(f\)/2.8, 1-5\). (\times\), Canon Inc., Tokyo, Japan), to record tip deflections during bending excitation (single-frequency, continuous sinusoid) at frequencies of 27.5-30 kHz. To create a shadow map, a cooled element of a high intensity white LED (part number: 4052899910881, white LED, 3000 K, 4150 lm, Osram Opto Semiconductors GmbH, Regensburg, Germany) was placed behind the tip of the needle.

Front view of the experimental setup. Depth is measured from the surface of the medium. The needle structure is clamped and mounted on a motorized transfer table. Use a high speed camera with a high magnification lens (5\(\x\)) to measure oblique angle deviation. All dimensions are in millimeters.

For each type of needle bevel, we recorded 300 frames of a high-speed camera measuring 128 \(\x\) 128 pixels, each with a spatial resolution of 1/180 mm (\(\approx) 5 µm), with a temporal resolution of 310,000 frames per second. As shown in Figure 6, each frame (1) is cropped (2) such that the tip of the needle is in the last line (bottom) of the frame, and the histogram of the image (3) is calculated, so the Canny thresholds of 1 and 2 can be determined. Then apply Canny edge detection 28(4) with Sobel operator 3 \(\times\) 3 and compute positions for non-hypotenuse pixels (labeled \(\mathbf {\times }\)) without cavitation 300 time steps. To determine the range of tip deflection, calculate the derivative (using the central difference algorithm) (6) and determine the frame (7) that contains the local extremes (i.e. peak) of the deflection. After a visual inspection of the cavitation-free edge, a pair of frames (or two frames with an interval of half time) was selected (7) and the deflection of the tip was measured (denoted as \(\mathbf {\times } \) ). The above is implemented in Python (v3.8, Python Software Foundation, python.org) using the OpenCV Canny edge detection algorithm (v4.5.1, open source computer vision library, opencv.org). Finally, the deflection power factor (DPR, µm/W) is calculated as the ratio of the peak-to-peak deflection to the transmitted electrical power \(P_T\) (Wrms).

Using a 7-step algorithm (1-7), including cropping (1-2), Canny edge detection (3-4), calculation, measure the pixel position of the tip deflection edge using a series of frames taken from a high-speed camera at 310 kHz ( 5) and its time derivative (6), and, finally, the range of tip deflection is measured on visually checked pairs of frames (7).

Measured in air (22.4-22.9°C), deionized water (20.8-21.5°C) and 10% (w/v) aqueous ballistic gelatin (19.7-23.0°C , \(\text {Honeywell}^{ \ text {TM}}\) \(\text {Fluka}^{\text {TM}}\) Bovine and Pork Bone Gelatin for Type I Ballistic Analysis, Honeywell International, North Carolina, USA). Temperature was measured with a K-type thermocouple amplifier (AD595, Analog Devices Inc., MA, USA) and a K-type thermocouple (Fluke 80PK-1 Bead Probe No. 3648 type-K, Fluke Corporation, Washington, USA). Use a vertical motorized Z-axis stage (8MT50-100BS1-XYZ, Standa Ltd., Vilnius, Lithuania) to measure depth from the media surface (set as the origin of the Z-axis) with a resolution of 5 µm per step.

Since the sample size was small (n = 5) and normality could not be assumed, the two-sample two-tailed Wilcoxon rank sum test (R, v4.0.3, R Foundation for Statistical Computing, r-project.org) was used to compare the amount of variance needle tip for various bevels. Three comparisons were made for each slope, so a Bonferroni correction was applied with an adjusted significance level of 0.017 and an error rate of 5%.

Reference is made to Fig. 7 below. At 29.75 kHz, the curved half wavelength (\(\lambda _y/2\)) of a 21-gauge needle is \(\approximately) 8 mm. The bending wavelength decreases along the slope as it approaches the tip. At the tip \(\lambda _y/2\) there are stepped bevels of 3, 1 and 7 mm, respectively, for ordinary lancets (a), asymmetric (b) and axisymmetric (c). Thus, this means that the lancet will differ by \(\about\) 5 mm (due to the fact that the two planes of the lancet form a point of 29.30), the asymmetrical slope will vary by 7 mm, and the symmetrical slope by 1 mm. Axisymmetric slopes (the center of gravity remains the same, so only the wall thickness actually changes along the slope).

Application of the FEM study at 29.75 kHz and the equation. (1) Calculate the bending half-wave change (\(\lambda _y/2\)) for lancet (a), asymmetric (b) and axisymmetric (c) oblique geometry (as in Fig. 1a,b,c). ). The average \(\lambda_y/2\) for the lancet, asymmetric, and axisymmetric slopes is 5.65, 5.17, and 7.52 mm, respectively. Note that tip thickness for asymmetric and axisymmetric bevels is limited to \(\approx) 50 µm.

Peak mobility \(|\tilde{Y}_{v_yF_y}|\) is a combination of optimal tube length (TL) and inclination length (BL) (Fig. 8, 9). For a conventional lancet, since its size is fixed, the optimal TL is \(\approx\) 29.1 mm (Fig. 8). For asymmetric and axisymmetric slopes (Fig. 9a, b, respectively), the FEM study included BL from 1 to 7 mm, so the optimal TL ranges were from 26.9 to 28.7 mm (range 1.8 mm) and from 27.9 to 29.2 mm (range 1.3 mm). ) ), respectively. For asymmetric slopes (Fig. 9a), the optimal TL increased linearly, reaching a plateau at BL 4 mm, and then sharply decreased from BL 5 to 7 mm. For axisymmetric slopes (Fig. 9b), the optimal TL increases linearly with BL elongation and finally stabilizes at BL from 6 to 7 mm. An extended study of axisymmetric slopes (Fig. 9c) showed a different set of optimal TLs located at \(\approximately) 35.1–37.1 mm. For all BLs, the distance between two sets of optimal TLs is \(\approx\) 8 mm (equivalent to \(\lambda _y/2\)).

Lancet transmission mobility at 29.75 kHz. The needle tube was flexed at a frequency of 29.75 kHz, the vibration was measured at the end and expressed as the amount of transmitted mechanical mobility (dB relative to the maximum value) for TL 26.5-29.5 mm (0.1 mm step).

Parametric studies of the FEM at a frequency of 29.75 kHz show that the transfer mobility of the axisymmetric tip is less affected by changes in the length of the tube than its asymmetric counterpart. Bevel length (BL) and pipe length (TL) studies for asymmetric (a) and axisymmetric (b, c) bevel geometries in frequency domain studies using FEM (boundary conditions are shown in Figure 2). (a, b) TL ranged from 26.5 to 29.5 mm (0.1 mm step) and BL 1-7 mm (0.5 mm step). (c) Extended axisymmetric oblique angle study including TL 25-40mm (0.05mm step) and 0.1-7mm (0.1mm step) which reveals the desired ratio \(\lambda_y/2\) Loose moving boundary conditions for a tip are satisfied.

The needle structure has three natural frequencies \(f_{1-3}\) divided into low, medium and high modal regions as shown in Table 1. The PTE size is shown in Figure 10 and then analyzed in Figure 11. Below are the results for each modal area:

Typical recorded instantaneous power transfer efficiency (PTE) amplitudes obtained using sinusoidal excitation with swept frequency at a depth of 20 mm for a lancet (L) and axisymmetric slopes AX1-3 in air, water and gelatin. A one-sided spectrum is shown. The measured frequency response (300 kHz sample rate) was low-pass filtered and then downsampled by a factor of 200 for modal analysis. The signal-to-noise ratio is \(\le\) 45 dB. The PTE phase (purple dotted line) is shown in degrees (\(^{\circ}\)).

The modal response analysis is shown in Figure 10 (mean ± standard deviation, n = 5) for the L and AX1-3 slopes in air, water, and 10% gelatin (20 mm depth) with (top) three modal regions (low, medium, high). ), and their corresponding modal frequencies\(f_{1-3}\) (kHz), (average) energy efficiency\(\text {PTE}_{1{-}3 }\) uses design equations. (4) and (bottom) are the full width at half the maximum measured value \(\text {FWHM}_{1{-}3}\) (Hz), respectively. Note that when recording a low PTE, i.e. in the case of an AX2 slope, the bandwidth measurement is omitted, \(\text {FWHM}_{1}\). The \(f_2\) mode is considered to be the most suitable for comparing the deflection of inclined planes, as it demonstrates the highest level of power transfer efficiency (\(\text {PTE}_{2}\)), up to 99% .

First modal region: \(f_1\) does not depend much on the media type inserted, but depends on the bevel geometry. \(f_1\) decreases with decreasing bevel length (27.1, 26.2 and 25.9 kHz for AX1-3, respectively, in air). The regional averages \(\text {PTE}_{1}\) and \(\text {FWHM}_{1}\) are \(\approx\) 81% and 230 Hz respectively. \(\text {FWHM}_{1}\) was the highest in gelatin from Lancet (L, 473 Hz). Note that \(\text {FWHM}_{1}\) for AX2 in gelatin cannot be estimated due to the low magnitude of the reported frequency responses.

The second modal region: \(f_2\) depends on the type of paste and bevel media. In air, water and gelatin, the average \(f_2\) values are 29.1, 27.9 and 28.5 kHz, respectively. The PTE for this modal region also reached 99%, the highest among all measurement groups, with a regional average of 84%. The area average \(\text {FWHM}_{2}\) is \(\approx\) 910 Hz.

Third modal region: \(f_3\) The frequency depends on the type of insertion medium and bevel. Average \(f_3\) values are 32.0, 31.0 and 31.3 kHz in air, water and gelatin, respectively. \(\text {PTE}_{3}\) has a regional average of \(\approximately\) 74%, the lowest of any region. The regional average \(\text {FWHM}_{3}\) is \(\approximately\) 1085 Hz, which is higher than the first and second regions.

* The following refers to Fig. 12 and Table 2. The lancet (L) deflected the most (with high significance to all tips, \(p<\) 0.017) in both air and water (Fig. 12a), achieving the highest DPR (up to 220 µm/W in air). 12 and Table 2. The lancet (L) deflected the most (with high significance to all tips, \(p<\) 0.017) in both air and water (Fig. 12a), achieving the highest DPR (up to 220 µm/ W in air).*

Tip bending amplitude measurements (mean ± standard deviation, n = 5) for L and AX1-3 chamfers in air and water (depth 20 mm) revealed the effect of changing chamfer geometry. The measurements are obtained using continuous single frequency sinusoidal excitation. (a) Peak deviation (\(u_y\vec {j}\)) at the vertex, measured at (b) their respective modal frequencies \(f_2\). (c) Power transmission efficiency (PTE, rms, %) as an equation. (4) and (d) Deviation power factor (DPR, µm/W) calculated as peak deviation and transmit power \(P_T\) (Wrms).

Typical shadow plot of a high-speed camera showing the total deflection of the lancet tip (green and red dotted lines) of the lancet (L) and axisymmetric tip (AX1-3) in water (depth 20mm), half cycle, drive frequency \(f_2\) (frequency 310 kHz sampling). The captured grayscale image has dimensions of 128×128 pixels with a pixel size of \(\approximately) 5 µm. Video can be found in additional information.

Thus, we modeled the change in bending wavelength (Fig. 7) and calculated the mechanical mobility for transfer for conventional lanceolate, asymmetric, and axial combinations of tube length and bevel (Fig. 8, 9). Symmetrical beveled geometry. Based on the latter, we estimated the optimum tip-to-weld distance to be 43 mm (or \(\approx\) 2.75\(\lambda_y\) at 29.75 kHz) as shown in Figure 5, and fabricated three axisymmetric bevels with different bevel lengths. We then characterized their frequency responses compared to conventional lancets in air, water, and 10% (w/v) ballistic gelatin (Figures 10, 11) and determined the best case for comparing tilt deflection mode. Finally, we measured tip deflection by bending wave in air and water at a depth of 20 mm and quantified the power transfer efficiency (PTE, %) and deflection power factor (DPR, µm/W) of the injected medium for each tilt. type (Fig. 12).

The results show that the tilt axis of the geometry affects the amplitude deviation of the tip axis. The lancet had the highest curvature and also the highest DPR compared to the axisymmetric bevel, while the axisymmetric bevel had a smaller mean deviation (Fig. 12).

In experimental studies, the magnitude of the reflected flexural wave is affected by the boundary conditions of the tip. When the needle tip was inserted into water and gelatin, \(\text {PTE}_{2}\) averaged \(\approx\) 95% and \(\text {PTE}_{2}\) averaged the values are 73% and 77% (\text {PTE}_{1}\) and \(\text {PTE}_{3}\), respectively (Fig. 11). This indicates that the maximum transfer of acoustic energy to the casting medium (for example, water or gelatin) occurs at \(f_2\). Similar behavior was observed in a previous study using simpler device structures at frequencies of 41-43 kHz, where the authors demonstrated the voltage reflection coefficient associated with the mechanical modulus of the intercalated medium. The penetration depth32 and the mechanical properties of the tissue provide a mechanical load on the needle and are therefore expected to influence the resonant behavior of the UZeFNAB. Therefore, resonance tracking algorithms such as 17, 18, 33 can be used to optimize the power of the sound delivered through the stylus.

Bend wavelength modeling (Fig. 7) shows that axisymmetric has higher structural stiffness (i.e. higher bending stiffness) at the tip than lancet and asymmetric bevel. Derived from (1) and using the known velocity-frequency relationship, we estimate the bending stiffness of the lancet, asymmetric and axisymmetric tips as slopes \(\approximately) 200, 20 and 1500 MPa, respectively. This corresponds to (\lambda _y\) 5.3, 1.7 and 14.2 mm at 29.75 kHz, respectively (Fig. 7a–c). Considering the clinical safety of the USeFNAB procedure, the influence of geometry on the stiffness of the bevel design needs to be evaluated34.

The study of the parameters of the bevel and the length of the tube (Fig. 9) showed that the optimal TL range for the asymmetric (1.8 mm) was higher than for the axisymmetric bevel (1.3 mm). In addition, the mobility plateau ranges from 4 to 4.5 mm and from 6 to 7 mm for asymmetric and axisymmetric tilt, respectively (Fig. 9a, b). The practical relevance of this finding is expressed in manufacturing tolerances, for example, a lower range of optimal TL may imply a need for higher length accuracy. At the same time, the yield platform provides a greater tolerance for the choice of slope length at a given frequency without significantly affecting the yield.

The study includes the following limitations. Direct measurement of needle deflection using edge detection and high-speed imaging (Figure 12) means that we are limited to optically transparent media such as air and water. We would also like to point out that we did not use experiments to test the simulated transfer mobility and vice versa, but used FEM studies to determine the optimal length of the manufactured needle. From the point of view of practical limitations, the length of the lancet from tip to sleeve is 0.4 cm longer than other needles (AX1-3), see fig. 3b. This may have affected the modal response of the acicular structure. In addition, the shape and volume of waveguide lead solder (see Figure 3) can affect the mechanical impedance of the pin design, resulting in errors in mechanical impedance and bending behavior.

Finally, we have experimentally demonstrated that the bevel geometry affects the amount of deflection in USeFNAB. In situations where a higher deflection amplitude can have a positive effect on the effect of the needle on the tissue, for example, cutting efficiency after puncture, a conventional lancet can be recommended for USeFNAB, since it provides the greatest deflection amplitude while maintaining sufficient rigidity at the tip of the design. In addition, a recent study has shown that greater tip deflection can enhance biological effects such as cavitation, which may help develop applications for minimally invasive surgical interventions. Given that increasing total acoustic power has been shown to increase biopsy yield from USeFNAB13, further quantitative studies of sample yield and quality are needed to assess the detailed clinical benefit of the studied needle geometry.

Frable, WJ Fine needle aspiration biopsy: a review. Humph. Sick. 14:9-28. https://doi.org/10.1016/s0046-8177(83)80042-2 (1983).