1. Introduction
Venous thromboembolism (VTE) is a significant global contributor to illness and mortality. It commonly manifests as deep vein thrombosis (DVT) primarily originating from the calf veins. DVT can progress to proximal vein thrombosis, potentially resulting in pulmonary embolism (PE) when dislodged debris reaches the lungs [
1], [
2], [
3]. Annually, approximately 10% of symptomatic PE patients die within an hour of symptom onset, and acute PE accounts for up to 300 000 fatalities in the United States alone [
1], [
4]. Nevertheless, VTE treatment presents notable challenges and financial burdens, with approximately 50% of the patients with symptomatic proximal DVT or PE experiencing recurrent blood clot formation within three months of treatment. Traditional treatment methods involve catheter-directed thrombolysis using tissue plasminogen activators (tPAs) or mechanical thrombectomy [
5], [
6], [
7], [
8]. However, catheter-directed thrombolysis is time-consuming, requiring over 24 h of treatment, and carries a substantial risk of symptomatic bleeding [
5], [
9]. Meanwhile, although mechanical thrombectomy is comparably effective for treating acute, subacute, or chronic DVT without the use of fibrinolytics, it carries the potential for vascular damage [
7], [
8], [
10] and may offer limited clinical improvement than tPA treatment [
11].
To address the therapeutic challenges in thrombolysis, techniques to locally enhance the efficacy of thrombolytic drugs and those that enable minimally invasive mechanical treatment of clots without drugs are needed. Sonothrombolysis, also known as ultrasonography-enhanced thrombolysis, has the potential to fulfill both objectives [
12], [
13]. Pulsed ultrasound (US) is more effective for thrombolysis
in vitro than continuous-wave US. Histotripsy pulses, which produce inertial cavitation, are effective in dissolving retracted clots [
14]. Nevertheless, they have limitations of lesser effectiveness over extracellular structures such as the fibrin matrix, requiring a lose-dose administration of lytic drugs [
15], [
16], [
17]. Usage of US contrast agents including microbubbles (MBs) and nanodroplets (NDs) can lower the cavitation threshold under US excitation, thus improving the sonothrombolysis rate [
18], [
19], [
20], [
21], [
22]. Submegahertz US exposure or diagnostic imaging frequencies can enhance thrombolysis [
23], [
24]. However, the use of external transducers or arrays for treatment is limited by significant attenuation and aberration caused by biological tissues, including frequency-dependent attenuation by soft tissues and blockage by bones [
23], [
24], [
25], [
26], [
27]. Respiratory motion can lead to inadequate acoustic coupling between the skin and transducer interface, which can result in skin burns or vessel damage. Additionally, a significant size difference between the focal zone and the blood vessels can potentially cause moderate collateral damage
in vivo, including coagulative necrosis or hemorrhage [
12], [
28], [
29].
Miniaturized intravascular transducers have been recently introduced to achieve highly efficient
in vitro sonothrombolysis [
18]. Unlike conventional external transducers, intravascular transducers can deliver acoustic energy directly to the thrombi. Thus, MB-mediated intravascular sonothrombolysis is achieved
in vitro with a lower peak-negative pressure. No vessel wall damage was observed in the initial
ex vivo tests [
30]. Moreover, ND-mediated sonothrombolysis shows a higher lysis rate for retracted clot treatment
in vitro [
21], [
31], [
32]. However, the existence of residual vein thrombosis (RVT) after treatment may cause recurrence of the DVT [
33], [
34], [
35], [
36]. Current sonothrombolysis methods have a limitation of a small recanalized vessel size, which is clinically required to be > 3 mm six months after the initial treatment to avoid clot recurrence [
33], [
37], [
38]. Thus, enhancing the efficiency of sonothrombolysis
in vitro and
in vivo is imperative to mitigate the impact of RVT.
This study aims to demonstrate, for the first time in vivo, the intravascular multi-directional transducer enabled sonothrombolysis. In specific, a miniaturized 4-layer PZT-5A (Type III 301, TRS Technologies, Inc., USA) stacks with an aperture size of 1.4 mm × 1.4 mm was developed and integrated into a catheter mediated with NDs, the newly developed ultrasound catheter device was tested for multi-directional sonothrombolysis with both forward- and side-looking (FL and SL) elements, both in vitro and in vivo. In vitro experiments showed the creation of a relatively large flow channel (D > 4 mm) while the in vivo experiments demonstrated successful treatment of a vein with 9 cm long thrombosis, which indicated the device's potential for clinical sonothrombolysis applications.
2. Materials and methods
2.1. Design and fabrication of the transducer and the catheter
An intravascular rotational catheter that included both FL and SL transducers was prepared for the
in vivo treatment (
Fig. 1(a)). Multidirectional stacked transducers were first prepared. With a small aperture size, a single-layer transducer has a relatively large electrical impedance and a less efficient pressure output. Thus, we implemented a 4-layer stacked design for both the FL and SL transducers to increase the capacitance while reducing the electrical impedance, using a method similar reported in Ref. [
18]. The design enhanced the electrical impedance compatibility with the driving electronics, leading to a higher pressure output under identical inputs. For the stack preparation, PZT-5A was chosen as the active layer owing to its high piezoelectric coefficient and good performance. We used Al
2O
3/epoxy with a weight ratio of 1:4 as the isolation layer on the sides of the transducers. The designed center frequency was 850 kHz for FL and SL elements, with a thickness of 250 μm for each layer in the stacks. The poled PZT-5A wafers were used and lapped down to 250 μm with the gold deposition of 200/50 nm Ar/Ti as the electrodes. Conductive silver epoxy (E-Solder 3022, Von-Roll Inc., USA) was employed to create bonding between each adjacent layer by maintaining the pressure with a customized jig and a precise controlled thickness of (25 ± 3) μm for the bonding layer. To provide the required acoustic impedance gradient for efficient acoustic energy penetration, Al
2O
3 particles (50 nm, ALR-1005-01; Pace Technologies, USA) was mixed with epoxy (volume ratio: 25%; EPO-TEK 301; Epoxy Tech. Inc., USA) and centrifuged. The matching layer was then lapped to 0.6 mm for the designed frequency of 850 kHz. Thereafter, the stacks were diced into a dimension of 1.4 mm × 1.4 mm with a dicing saw (DAD 304, DISCO Inc, Japan) to fit into the 10-Fr catheter.
Fig. 1(b) hows the optical photograph of the catheter. On the rear side of the FL stack, a similar stack was attached to a side-looking element. An Al
2O
3/epoxy layer was bonded to the side of the transducer as an insulating barrier, achieving relatively high electrical resistance after 24 h of curing. Then, the side electrodes were formed with silver epoxy with electric resistances below 1 Ω. Subsequently, the FL and SL stacks were connected to a shared coaxial cable using a silver epoxy (E-Solder 3022; Von-Roll Inc., USA). A 13 μm-thick parylene C layer was deposited on the whole structure as the passivation layer with a parylene coater (SCS Labcoter, USA). The operating frequency and pressure output of the transducer were simulated and estimated by using a symmetric model in COMSOL (COMSOL 5.5; COMSOL Inc., USA). The transducer design parameters and material properties are listed in
Table 1.
For catheter rotation, the transducer was assembled with a 200 μm injection tube for MBs/NDs. The entire structure was then fitted to a steel shaft with an inner diameter of 7-Fr and an outer diameter of 10-Fr. The stability of the catheter during rotation was ensured by tight bonding between the shaft and the catheter at a relatively low rotation speed. Consequently, the structure was sealed with epoxy following a 13 μm parylene coating layer and then sealed with epoxy following a 13 μm parylene coating layer. To control the catheter rotation, the steel shaft was coupled to a stepper motor (Nema 17, STEPPERONLINE Inc., USA) using a three-dimensional (3D)-printed linkage (KP3, KINGROON Tech Co., Ltd., China) with PLA filament material (PLA 3D Printer Filament, HATCHBOX, USA). Subsequently, the stepper motor was connected to an Arduino UNO (Arduino, Ivrea, Italy) and a stepper motor shield (Motor/Stepper/Servo Shield; Adafruit, USA). The Arduino and motor shields were fixed inside a 3D-printed housing to protect the circuitry during testing.
2.2. Stacked transducer characterization
The prototype transducer was characterized by its electric impedance response and acoustic pressure output (
Fig. 2(a)). To measure electrical impedance, the transducer was connected to an impedance analyzer (4294A, Agilent Tech. Inc., USA), and the response was measured in the frequency range of 0.2-1.0 MHz. To characterize the pressure output, a function generator (33250A, Agilent Tech. Inc., USA) was used to apply a sinusoidal pulse of 10 cycles per ten ms corresponding to a pulse repetition frequency (PRF) of 100 Hz. Subsequently, the signal was amplified using an RF power amplifier (75A250A, AR, USA) with a power gain of 54 dB before being delivered to the transducer. The pressure output was measured by using a hydrophone (HGL-0085; ONDA Corp., USA) with a 20 dB preamplifier (AH-2020, ONDA Crop., USA), and data were collected using an oscilloscope (DSO7104B; Agilent Technologies, USA). The transmission efficiency was first estimated under different input driving voltages at the focal point in both directions. Subsequently, a 3D motion stage was used to capture the pressure output in a 10 mm × 10 mm region to measure the acoustic pressure field for both the FL and SL elements.
2.3. Blood clot preparation for in vitro test
Subacute clots were prepared
in vitro as previously described [
32]. Briefly, a tygon tube with an inner diameter of 5.6 mm was first prepared to mimic the deep vein (
Fig. 2(b)). As there is no significant difference in the acoustic impedances between the tygon tubes and water, the tygon tube was acoustically transparent to sub-MHz waves, which caused negligible reflection inside the tubing. The acid citrate dextrose bovine blood (Lampire Biological Laboratory Inc., USA) was mixed with the 2.75% w/v calcium chloride solution (Fisher Scientific, USA) in a volume ratio of 10:1. The mixture was then drained into a tygon tube channel and sealed. The containers were immersed in a 37 °C water bath for 3 h and then stored at 4 °C for 7 days to achieve full incubation for subacute clots. For each test, the length of the sample was 6 cm. The weights of the channels were compared before and after clot preparation to estimate clot size.
2.4. MB and ND preparation
The perfluorocarbon MB and ND cavitation agents used in this study were synthesized in-house following established procedures [
31]. Briefly, we conducted a gas-exchange process to replace with decafluorobutane (DFB) gas the air headspace of a lipid solution (1.5 mL) sealed in 3 mL glass vials (Fluoromed, USA). After shaking the vial for 45 s using a Vialmix device (Lantheus Medical Imaging, USA), MBs consisting of phospholipid shells encapsulating the DFB gas cores were produced. This process yielded MBs with particle sizes measuring 0.6-4.0 µm and a mean diameter of 1.1 µm. The MB concentration was approximately 1 × 10
10 particles per mL. To prepare the NDs, in-house lipid-shell DFB MBs were condensed following previously protocols [
31], [
39]. After MBs were produced by vial shaking, the DFB MBs were subjected to condensation under pressure (20 psi, 1 psi = 6.895 kPa) and low temperature (−13 °C). The size of the liquid-core ND particles ranged from 250 to 600 nm, with a mean diameter of 330 nm. The ND particle concentration was similar to that of the MB vials (1 × 10
10 particles per mL). All solutions were further diluted to a concentration of 1 × 10
8 per mL using sterile saline for both the
in vitro and
in vivo tests.
2.5. In vitro sonothrombolysis tests with subacute clots
A flow model was created to emulate a blood vessel using the same tygon tube, as shown in
Fig. 2(b). A fully occluded clot sample was placed in the tube. The hydraulic pressure in the reservoir water tank was adjusted by controlling the height and was maintained at 0.5 kPa during all the tests. The venous flow model started with a tank loaded with saline maintained at a temperature of 37.3 ± 0.3 °C. During the thrombolysis tests, the US transducer was powered by a radiofrequency power amplifier (amplification ratio: 53 dB, model: 75A250A, AR RF/Microwave Instrumentation, USA), and the corresponding input sine wave signal was generated with a function generator (model: 33250A, Agilent Technologies Inc., USA). The burst excitations were generated at a PRF of 1 kHz. Each burst consisted of 25 cycles, corresponding to a low duty cycle of 5%, to prevent any significant temperature increases over the catheter surface. The input peak-to-peak voltage (
Vpp) was 90 V, and the corresponding peak-negative pressure output was set to ∼2.5 MPa. For the accumulation of the contrasts during treatment, a sonication method with 5 s on-and-off pulsing was realized with two functional generators. The catheter promotion speed was controlled at 2 mm∙min
−1, with a rotation speed of 1 r∙min
−1. For the forward-looking element in the forward/side-looking (FSL) transducer, the same design parameters were applied for element size, thickness, and number of layers as those for the single FL transducer. During the
in vitro test, identical input parameters, including pressure output, burst duration, duty cycles, power, and concentration of MB/NDs, were applied to both the FL and FSL transducers, ensuring a fair comparison of the lysis rate and channel opening.
To evaluate the performance of the FSL transducer, three series of tests were conducted: ① the FL transducer without rotation, ② the FSL transducer without rotation, and ③ the FSL transducer with rotation. For each series, the lysis rate was estimated in the following groups: ① control group; ② US thrombolysis alone; ③ MB-mediated sonothrombolysis (MB + US); ④ ND-mediated sonothrombolysis (ND + US); ⑤ MB/ND mixture-mediated sonothrombolysis (MB + ND + US), and ⑥ MB/ND mixture-mediated sonothrombolysis with low-dose tPA (MB + ND + US + tPA). For groups with single contrast agents, the MB/ND concentration was maintained at 108 per mL with an infusion flow rate of 100 μg·min−1. For the MB/ND mixture, the ND/MB ratio was set to 9:1 with a total concentration of 108 per mL for equal comparison. For low-dose tPA treatment, the tPA concentration was set to 0.1 mg·mL−1, with a total dose of 0.3 mg administered for 30 min. In the experimental design, the efficacy of NDs was compared to that of MBs as the basis for the in vivo testing. A small quantity of MBs was added to facilitate tracking of the injection flow during the test, enhancing the contrast under B-mode imaging. Mass reduction of the clot before and after treatment was used to estimate the lysis rate. For each case, four groups of tests were performed for the statistical analysis. Statistical significance was determined using a one-way unbalanced analysis of variance, and Tukey’s honest significant difference test (p < 0.05) was used to estimate the efficiency of sonothrombolysis with different transducers and with different contrast combinations.
2.6. In vivo sonothrombolysis tests with retracted clots
A porcine DVT model was created to test the feasibility of our catheter-directed sonothrombolysis technique
in vivo. The assessment items were as follows: ① real-time monitoring using US guidance, ② catheter feed-in and position control, and ③ determination of sonothrombolysis efficacy using fluoroscopy. Prototype catheters with the same specifications as those used for the
in vitro experiments were used. The treatment setup was similar to that of the
in vitro test setup: two function generators, an RF amplifier, and a micropump. We used a concentration in the safe range [
17], [
40], [
41] with a low dose of tPA (0.1 mg∙mL
−1 with a total administration of 0.3 mg) [
42], [
43], [
44] based on current treatment guidelines for DVT. Treatment efficacy was determined using contrast flow propagation on fluoroscopic images (BV Endura, Philips, Netherlands).
The conditions during sonothrombolysis treatment were the same as those in US pulsing, while concentrations of the MB + ND mixture, tPA dose, and infusion rates were the same as those used for the in vitro efficacy evaluation. A pretreatment fluoroscopic image showing occlusion of the contrast flow was obtained. A prototype catheter was then inserted into a guide sheath (9-Fr, Teleflex AK-09903-CDC, USA) and positioned at the distal end of the target clot under fluoroscopic guidance. The ND + MB + tPA mixture (total concentration: 108 per mL; ND/MB ratio = 9:1) was injected at an infusion rate of 100 µL∙min−1 and sonicated with the 5 s on-and-off pulsing. Treatment status was monitored using US imaging (Lumify L12-4, Philips, Netherlands). The catheter feed speed and 360° rotation were 3 mm∙min−1 and 1 r∙min−1, respectively. Real-time B-mode imaging was used to adjust the catheter tip to ensure a stable position in front of the clot within the focal depth. The position of the C-arm was fixed during the 30 min treatment to capture the fluoroscopy image after the treatment. The removed clot area was quantified by analyzing the difference between the pre- and posttreatment images using ImageJ software (National Institutes of Health, USA).
2.7. In vivo porcine DVT model
The study protocol was approved by the University Committee on the Use and Care of Animals of the University of Michigan (IACUC #PRO00010822). All applicable international, national, and/or institutional guidelines for the care and use of animals were followed.
The porcine DVT model was prepared as previously reported [
44]. Briefly, a 9-Fr sheath (Teleflex AK-09903-CDC, USA) was percutaneously inserted into the femoral vein of mixed breed pigs weighing 30-40 kg. An 8-Fr balloon wedge catheter (Teleflex, AL-07128, USA) was also inserted percutaneously into the sheath at a proximal location (closer to the heart). Another 8-Fr balloon catheter was inserted into the femoral vein through the distally placed (farther from the heart) 9-Fr sheath. The balloon tips of these catheters were positioned under US guidance, allowing for a distance of approximately 2 cm between the balloons. After the inflated balloons caused flow stasis, thrombin (500 units; RECOTHROM, ZymoGenetics, USA) was injected through the distal catheter for hypercoagulation, and the clot was incubated for 2-3 h. Thereafter, the presence of blood clots and blood flow obstruction was confirmed using US imaging. The proximal catheter was removed, and the distal catheter and sheath were left in position to anchor the clot. The remaining distal catheter and sheath were anchored beneath the skin to prevent clot dislodgement. After seven days of clot aging, all pigs were anesthetized, the 9-Fr sheath was exposed to provide vascular access distal to the clot, and catheter-directed sonothrombolysis was performed.
3. Results
3.1. Multidirectional transducer characterization
Fig. 3 shows the measured results for the treatment transducer. Although some resonant modes were induced due to the fabrication errors, the transducer showed a main resonant frequency at 850 Hz with an electrical impedance of 71 Ω. This was expected to match with the excitation system in
Fig. 3(a). The capacitance was 1.9 nF at 1 kHz, with a corresponding dielectric loss of 17 mU. The pressure output for the forward- and side-looking element is shown in
Fig. 3(b). The peak negative pressure (PNP) output was 2.2 and 2.6 MPa at 80 V
Vpp for the forward- and side-looking elements, respectively, corresponding to a sensitivity of 0.0275 MPa∙V
−1 and 0.0325 MPa∙V
−1, respectively.
Fig. 3(c) shows the measured pressure output distributions for the FSL elements. For the forward-looking element, the transducer showed a focal depth of 1.5 mm, with a −6 dB area of 5.3 mm
2. For the side-looking element, the focal depth was 2 mm, with a −6 dB focal size of 13. 2 mm
2. Notably, the matching layer was slightly smaller than a quarter wavelength owing to the dimensional limitations of the SL element. However, the transducer still exhibited a focal zone at 2 mm, with high PNP levels. These were anticipated to be sufficiently high to generate cavitation from the MBs and NDs.
3.2. In vitro sonothrombolysis
The
in vitro test results are shown in
Fig. 4. The effects of the different transducers were first demonstrated by comparing the FL transducer with the FSL transducer without catheter rotation. The results indicated a consistent trend: all treated FSL transducers showed better lysis rates than the FL transducers treated with different combinations of contrast agents. For example, the FSL transducer showed a 32% higher lysis rate than the FL transducer ((12.1 ± 1.3) mg∙min
−1 vs (9.2 ± 0.9) mg∙min
−1) owing to the exitance of the ND + MB mixture. Similar results were obtained when further comparisons between FSL transducers with and without catheter rotation were performed. Compared to non-rotational cases, rotational cases showed a consistent improvement in the lysis rate. Particularly, for the ND + MB mixture group, the lysis rate increased by an additional 10% from (12.1 ± 1.3) to (13.4 ± 1.5) mg∙min
−1.
Fig. 4 shows the different effects of contrast agents on subacute clot sonothrombolysis. For the rotational FSL transducer, the US + ND group showed significantly better lysis rate (> 30%) than the US + MB group (10.1 vs 13.4 mg∙min
−1). Meanwhile, the lysis rate of the US + ND + MB group was comparable to that of US + ND group. Low-dose tPA treatment improved the base lysis rate, with the maximum mass reduction reaching up to 51% with a lysis rate of 16.4 mg∙min
−1 under US + ND + MB + tPA treatment. Thus, by using the rotational FSL transducer, a promising lysis rate was achieved
in vitro using ND + MB mediated sonothrombolysis.
3.3. In vivo sonothrombolysis of retracted clots
The
in vivo results are presented in
Table 2. In all four pigs, US activation was confirmed by applying a 5 s test pulse sequence prior to treatment and then visualizing US interference lines during imaging of the femoral vein where the catheter was inserted (
Figs. 5(a) and
(b)). Infusion of the ND + MB + tPA mixture was also confirmed by increased US scattering at the catheter tip during the 5 s test pulse sequence (
Fig. 5(c)). As shown in
Fig. 5 (green arrows), the side-looking element led to clot erosion, which generated an area with less scatter on US imaging.
Fluoroscopy showed flow restoration and improvement following sonothrombolysis (top row of
Fig. 6) in two of the four pigs (pigs 1 and 3). Blood flow restoration was also confirmed on color Doppler US imaging in pig 3 (bottom row of
Fig. 6). The clot was approximately 9 cm long in pig 1 and 14 cm long in pig 3. The average estimated clot dissolution area was 5.3 mm in diameter for pig 3. Technical problems were encountered in pigs 2 and 4. In pig 2, the target clot propagated proximally up the iliac vein and inferior vena cava, causing insufficient treatment coverage because of the limited length of the prototype catheter. In pig 4, the transducer unit was detached from the catheter housing and electrically disconnected while the prototype catheter was inserted into a guide sheath. These experiments demonstrated the proof-of-concept and feasibility of this technique, although technical improvements in device design and clot formation (e.g., limiting the clot length to allow complete treatment and transducer detachment) are still needed.
4. Discussion
This study demonstrates the feasibility of a rotational, multidirectional catheter for sonothrombolysis of retracted clots, both
in vitro and
in vivo. For the control group, the FSL with rotation showed minimal differences (< 10%) compared with the other two cases, suggesting that catheter rotation itself had a negligible influence on the lysis rate. In addition, all FSL groups were better than the FL groups; this could be attributed to the effect of the FSL transducer on the residual clots on the side of the tube after the FL element treatment. The efficiency of the SL element was enhanced by rotation of the catheter, and this was observed in all rotational FSL cases but not in non-rotational FSL cases. Particularly, the US + ND + MB + tPA group showed the most significant difference in lysis rate using rotational FSL (i.e., 31% higher than non-rotational FSL). The distribution of clot debris with ND-mediated sonothrombolysis showed a safe pattern, consistent with previous
in vitro [
30] and
in vivo [
44], [
45] findings. There was also no evidence of downstream embolization during the recanalization process of the 4 mm channel. All the tested pigs survived, further supporting the safety of the procedure. A debris size assessment will be included in future studies.
Aside from the effect of the transducer itself, the effects of NDs and low-dose tPA were also validated. The lysis rate was higher by 30% in the US + ND group than in the US + MB group. These findings suggest that ND-mediated sonothrombolysis may have an advantage over MB-mediated sonothrombolysis for retracted clots, consistent with previous findings [
31], [
32]. When comparing between the US + ND and US + ND + MB groups, similar lysis rates were observed across all cases involving FL, rotational FSL, and non-rotational FSL. These results are consistent with the fact that NDs had a more significant impact on retracted clot sonothrombolysis when mixed with small amounts of MBs. Considering the ratio between the NDs and MBs, NDs have a more predominant effect than MBs. Among all the groups, the MB + ND + tPA group with FSL rotation achieved the highest lysis rate, resulting in a mass reduction of 52% and a corresponding lysis rate of approximately 16 mg∙min
−1. Compared to non-rotational treatment, rotational treatment of FSL yielded an additional 10% increase in the lysis rate, a considerable improvement. However, it should be noted that rotation of the FSL was associated with an increased standard deviation of 7.1%. This indicates that the rotation of the catheter introduces a more random effect on sonothrombolysis. In the
in vitro study, FSL with rotation exhibited consistent trends of increased lysis rates across all groups. Moreover, ND-mediated sonothrombolysis was more advantageous for retracted clots than was MB-mediated sonothrombolysis. The highest lysis rate was achieved with MB + ND + tPA with FSL rotation, with the lysis rate being significantly better than that in non-rotational treatment. This provided guidance for the
in vivo testing.
The
in vivo feasibility study generated important data regarding the catheter-directed sonothrombolysis technique; the data could guide further technique development. First, the MB + ND mixture was a suitable cavitation agent for US-enhanced DVT treatment. ND-mediated cavitation causes internal clot erosion, facilitating the dissolution of retracted clots [
31], [
32]. In the current study, the MB + ND mixture maintained this benefit with a reduced ND vaporization threshold and facilitated infusion monitoring. Second, we used a significantly lower dose of tPA than in previous US-enhanced sonothrombolysis techniques (< 10%) [
46], and this potentially ameliorated bleeding complications. We previously conducted histological tests to investigate the safety concerns associated with this technique, and the results demonstrated that the blood vessel wall remained undamaged both mechanically and thermally under
ex vivo conditions [
30]. In the current study, no signs of hemorrhage or bleeding were observed during the
in vivo tests in all four pigs. However, given the complexity of the catheter structure, advancement system, and application of a higher excitation pressure, a critical reassessment of the safety of this technique is essential. Future research should focus on investigating temperature variations and blood vessel histology
in vivo to provide a more comprehensive understanding of the safety of this technique.
Although a 4 mm channel was created with the FSL transducer in less than 30 min, a larger channel was expected to prevent revascularization. Other improvements for further in vivo studies and eventual clinical applications were identified. These include the following: ① assembly of a transducer unit to a flexible catheter tip with a soft shaft cover to avoid buckling and dislodgement during catheter insertion and bending; ② incorporation of side-looking MB/ND injection ports for higher sonothrombolysis efficiency over the residual clots; and ③ controlling for manual push-in effects of the catheter. Mechanical push-in effects should be clarified and differentiated from those of cavitation-enhanced sonothrombolysis to establish the efficacy of sonothrombolysis. To direct the catheter during the in vivo test, it was aligned in the sheath direction. This was a suitable approach given the relatively short femoral vein. However, recognizing longer clot sizes in human clinical applications, a guidewire is essential for future catheter designs that are intended to be implanted alongside the MB/ND injection lumen. Collectively, our findings confirm the feasibility of intravascular sonothrombolysis. Specifically, the results show the feasibility of real-time US-guided monitoring of the MB + ND infusion status, pulsing status, and partial flow restoration. Further, the catheter feed-in rate and tip positions could be controlled during treatment, and venous flow could be restored after catheter-directed sonothrombolytic therapy.
5. Conclusions
This study demonstrated the design, fabrication, and characterization of a miniaturized multidirectional transducer with a center frequency of 850 kHz for intravascular sonothrombolysis. A peak-negative pressure over 2.2 MPa is achieved for both the forward- and side-looking elements, with an input of 80 V Vpp. Motorized rotation of the catheter is achieved by integrating the transducer into a shafted catheter, and sonothrombolysis with different contrast agents is performed in vitro. The rotational FSL transducer with MB-ND and low-dose tPA is applied to in vivo subacute clot treatment to validate the efficiency of the sonothrombolysis. In the in vitro test, the rotational multidirectional catheter in the ND + MB + low-dose tPA group significantly improves thrombolysis, with the mass of the retracted clot reduced by 52%, yielding a 37% higher lysis rate than the non-rotational FL transducer treatment. In the in vivo tests, flow restoration is detected by fluoroscopy and color Doppler US images in two of four pigs with > 9 cm long subacute clots. Collectively, this preliminary study using our catheter-directed sonothrombolysis technique shows promising results. Future work will focus on cavitation detection, real-time imaging guidance, and safety of this technique.
Acknowledgements
This work was supported by National Institute of Health (NIH) grants (R01HL141967 and R21EB027304). The authors appreciate the assistance of Brian Velasco with nanodroplet and microbubble formulation.
Compliance with ethics guidelines
Huaiyu Wu, Jinwook Kim, Bohua Zhang, Gabe Owens, Greyson Stocker, Mengyue Chen, Benjamin C. Kreager, Ashley Cornett, Kathlyne Bautista, Tarana Kaovasia, Paul A. Dayton, Zhen Xu, and Xiaoning Jiang declare that they have no conflict interest or financial conflicts to disclose.
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.eng.2024.03.021.
PDF
(2218KB)
