Assessment of miniaturized ultrasound-powered implants: an in vivo study

UPIs used in this study were prototypes designed and fabricated by the Stanford University team in collaboration with the U.S. Food and Drug Administration team. Devices consisted of an US receiver, a full-wave bridge rectifier with Schottky diodes for AC to DC conversion (HSMS-282P, Avago Technologies), a 6.8 nF capacitor as a low pass filter, a 2.5 kΩ load resistor, and two stainless steel leads (AS633, Cooner Wire) for measuring device output. The circuit schematics of the UPI is shown in figure 1(A). PZT4, a piezoelectric ceramic material, with dimensions of 1.08  ×  1.08  ×  1.44 mm³ was used as the US power receiver; at its short-circuit resonance, f _sc, (~1 MHz), the receiver can be modeled by Thevenin equivalent circuit with an open circuit voltage, V_oc, and a resistor, R_piezo. The load resistor on the device models a potential medical payload. Lead wires were attached to the device to record DC output in response to US exposure for the study. This prototype can be modified to include functionalities such as neural recording and stimulation, or temperature and pressure sensing, as required by the application.

Zoom In Zoom Out Reset image size

Devices were fabricated by placing all the components onto a flexible printed circuit board (PCB). Then the populated flexible PCBs were encapsulated with polydimethylsiloxane (PDMS, Sylgard 184, Dow Corning) using a custom mold. The encapsulation design incorporated suture tabs to improve stability upon implantation. In addition, the devices were designed to eliminate sharp edges to minimize the risk of irritation. A minute amount of medical grade epoxy (EPO-TEK-301, Epoxy Technology, USA) was applied at the wire and PDMS interface for enforcement. Finally, devices were uniformly coated with 7 µm thick layer of Parylene-C to further enhance encapsulation. Figure 1(B) shows the photo of top view as well as the side view schematic of a fully fabricated UPI with all the components. The UPI prototype excluding the suture tabs and lead wires is measured to be 8.0  ×  2.5  ×  2.8 mm³.

To deliver US power to the UPIs, a 1 MHz unfocused US transducer (A314S, Olympus) was used as the external transmitter. A custom 3D printed housing was attached to the transducer to hold an US gel pad (Aquaflex, Parker Laboratories) to extend the transmission distance during measurements. The acoustic output of the US transducer along with the gel pad in the housing was characterized in a water tank at different locations using a hydrophone (HNC-1500, ONDA), which was spatially adjusted by a three-axis linear stage. Based on the measured voltage on the hydrophone and conversion coefficient provided by the manufacturer, acoustic intensities for different input voltages across the transducer (17, 25, 35, 43, 50, 56, and 78 V peak-to-peak) at 1 MHz were determined.

To generate the desired powering waveform, a custom MATLAB script was used to trigger a signal generator every 20 ms (50 Hz). The signal generator was set to output 500 sinusoidal cycles at 1 MHz at each trigger (500 µs pulse width). Output from the waveform generator was delivered into a 55 dB RF amplifier (1040L, Electronics & Innovation) to drive the US transducer. Each UPI was subjected to one minute of US exposure at the various transducer input voltages for the wireless powering test. To record the device DC output voltage under US exposure, the lead wires from the UPI were connected to a data acquisition unit (NI-USB 6361, National Instruments). The recorded raw data was processed by applying a bandpass filter (10–200 Hz) to reduce noise. Maximum voltages of the filtered data at each trigger were then averaged and recorded. The diagram of wireless measurement setup is illustrated in figure 1(C).

For each of the UPI devices (n  =  7), pre-implant scanning electron microscopy (SEM) images were captured to assess surface integrity, and device outputs were measured in vitro to confirm functionality prior to implantation. High magnification SEM images (Jeol 6390LV) were captured under low vacuum mode with 5 kV of electron acceleration. After SEM imaging, UPIs were placed under the US transducer with a gel pad for in vitro wireless powering characterization using the measurement setup described earlier. To further ensure complete contact between the UPI and transducer, US gel (Aquasonic) was also applied between the gel pad and the devices. Figure 1(D) shows the relationship between acoustic intensities at the center of the gel pad surface and the transducer input voltages, as well as the measured UPI output voltages averaged across all implants at different transducer input voltages. The output voltage from the UPI is around 1 V when 78 V is applied at the transducer. With 2.5 kΩ resistor on the UPI, 0.4 mW usable power is obtained. For applications that need more power and energy, the UPI circuit can be modified to include a large storage capacitor for charge storage. The raw output waveform of 500 µs pulse width at 50 Hz from one of the UPIs under US exposure is shown in figure 1(E).

To record UPI output after implantation, the two lead wires attached to the UPIs were cut to 55 mm in length from the base of the UPIs and threaded through 25 mm long silicone tubing. Ends of the lead wires were soldered to the pins of a connector (PZN-06-VV, Omnetics Corporation). A custom 3D printed connector mount was designed to house the connector; the bottom part of the connector mount was made of Nylon and the top was made with bronze infused stainless-steel. These two parts were bonded with epoxy (Loctite Hysol) and a 20  ×  25 mm piece of Mersilene mesh was attached to the bottom surface. The connector was then inserted into the 3D printed housing and secured using epoxy. Finally, the tissue contacting surface of the connector mount was covered with Kwik-Cast (World Precision Instruments Inc.) to minimize animal discomfort after implantation [17, 18]. Figure 2(A) shows the assembled UPI along with the connector mount. Before implantation surgeries, all UPIs were sterilized using Ethylene Oxide process.

Zoom In Zoom Out Reset image size

This animal study was approved by the Institutional Animal Care and Use Committee (IACUC) at the U.S. Food and Drug Administration, White Oak campus. Experiments were performed on Female Lewis rats (n  =  10), weighing between 230–280 g (5–8 months at the time of surgery), purchased from Charles River Laboratories International Inc. UPIs were implanted in n  =  7 animals, while n  =  3 animals were not subjected to surgical procedures and served as naïve control for histological comparison. Following standard protocol, animals were acclimated for 5 d upon receipt and housed with 12 h light and dark cycles. Before surgeries, animals were anesthetized with 4% isoflurane followed by administration of ketamine (75 mg kg⁻¹) and dexmedetomidine (0.25 mg kg⁻¹) cocktail intraperitoneally. Dehydration was prevented with subcutaneous administration of warm saline (3 ml), and the body temperature was maintained using heating pads. Meloxicam (2 mg kg⁻¹, s.c.) and baytril injectable (7.5 mg kg⁻¹, s.c.) were administered for analgesia and infection control respectively. Upon confirming the loss of toe pinch reflex, the lumbar region along with the right hind limb was shaved and disinfected with alcohol and betadine. Two ~20 mm long skin incisions were made for implantation: one over the lumbar fascia and the other over the right biceps femoris muscle (leg incision). Using blunt scissors, the skin between the two incisions was freed from connective tissue to allow for tunneling of UPIs. The device was first tunneled under the skin from the lumbar incision to the leg incision, and the connector mount was secured to the underlying lumbar fascia using 4-0 Polypropylene sutures (Oasis). The UPI along with a ~15  ×  15 mm Mersilene mesh was then sutured to the underlying muscle as shown in figure 2(B). Skin around the connector mount was closed using 4-0 Polypropylene sutures and staples were applied to the skin over the implant site as shown in figure 2(C). To prevent infection, topical antibiotic ointment was applied to the incision sites. Finally, animals were administered 3 ml of warm saline subcutaneously, and the anesthesia was reversed with atipamezole (0.5 mg kg⁻¹, i.p.). To prevent post-surgical pain and infections, meloxicam (2 mg kg⁻¹, s.c.) was further administered for 2 d after surgery, and trimethoprim-sulfamethoxazole (48 mg kg⁻¹, oral) was administered for 5 d in a red tinted water bottle. Remaining skin sutures and staples were removed 10 d post-surgery.

After two weeks of recovery from surgical procedures, each animal was tested three times per week for 10 weeks. During wireless powering tests, animals were anesthetized with isoflurane (4% induction, 2% maintenance) and placed on a circulating water heating pad. The measurement setup was as described in the previous section. US gel was used between the gel pad and the animals' skin to enhance coupling. Position of the transducer was manually adjusted until a maximum output voltage was obtained for the first measurement. Animals were subjected to one minute of US (500 cycles at 1 MHz at 50 Hz) at each voltage step (17, 25, 35, 43, 50, 56, and 78 V) with a 30 s pause period between each voltage step. The in vivo wireless powering measurement setup is shown in figure 2(D). To maintain consistency for histological analysis, this testing protocol was applied to the animal even when output voltages from the UPIs could not be recorded. Additionally, the rats' contralateral legs were also exposed to the same US levels to study the effects of long-term sonication on muscle and skin in the absence of a UPI.

After 12 weeks of implantation, animals were euthanized with sodium pentobarbital (200 mg kg⁻¹, i.p.) followed by device extraction. Immediately following extraction, lead wires were trimmed from the connector mount and in vitro wireless powering test was conducted again on each of the seven UPIs to check for functionality. After the measurements, devices were submerged in a solution containing 1% Tergazyme (Alconox) in deionized water for 48 h on a laboratory shaker for cleaning biological debris. Devices were then removed from the shaker, rinsed in water, and air dried for SEM characterization to assess the post-explant surface integrity.

Qualitative histology using Hematoxylin and Eosin (H&E) staining was performed to assess the effects of implants and US exposure on surrounding skin and muscle. Skin and muscle around the UPI on the right leg, along with the contralateral (left leg) tissues were harvested and immersion fixed overnight in 4% paraformaldehyde solution at 4 °C. Control skin and muscle tissues from naïve animals (n  =  6 samples, right and left hind limb) were obtained for comparison. After fixation, samples were rinsed in phosphate-buffered-saline and embedded in paraffin (ASP300S, Leica Biosystems). Skin samples were oriented to obtain cross-sections, while muscle samples were oriented for longitudinal sections. Approximately 10 µm thick sections were cut, deparaffinized, and stained with H&E using a Leica ST5020 Multistainer. Bright-field images were acquired at 20  ×  using a Zeiss fluorescent microscope for analysis.

In this study, we focused on assessing the long-term in vivo performance of UPI using a rodent model. The device was encapsulated with PDMS and then coated with an additional thin layer of Parylene-C. The device fabrication, especially for encapsulating the internal components was repeatable using a molding method. The UPI did not have any sharp edges, which was a critical design requirement to avoid animal discomfort upon implantation (figure 4(E)). The lead length of 55 mm was selected for the fully assembled devices based on prior experiments to accommodate for increase in animal size with age. The in vitro wireless testing of UPI confirmed device functionality before implantation with measured output ranging between 0.4 V to 0.7 V across all devices for an acoustic intensity of 1 mW mm⁻² (figure 1(D)). The shape of the output waveform was consistent with circuit simulation (figure 1(E)).

During the entire study, there were no UPI related adverse events. As shown in figure 4, skin over the UPI implant healed completely without any signs of infections. The region around the connector mount healed well without signs of infection, which is critical for interfacing with the UPI implant over long periods of time. More importantly, the connector mount provided a seamless access point to interface with the implanted UPI throughout the experimental period. We also did not observe macroscopic damage to the skin on the contralateral side, which was exposed to US.

The in vivo measurements from implanted UPI during wireless powering experiments are shown in figure 3. The violin plot illustrates the distribution of measured voltage outputs from all functional UPIs. The white dot in the box plot inside the violin plot represents median and the gray bar inside the violin plot represents interquartile range. The measured output voltages varied over testing period, mainly due to misalignment between the transducer and the UPIs, as the transducer had to be repositioned for each test for all seven animals. The orientation of the implanted UPIs might also change slightly due to animal movement. Nonetheless, the measured output waveform shape was maintained throughout the implantation period, indicating that the encapsulation used in the UPIs was able to reliably protect the US receiver and the internal circuitry in an in vivo environment. The in vivo measured output voltages were less than the measured values from in vitro testing in figure 1(D) because the transmission distance and the attenuation were larger due to animal tissue in the powering path.

Zoom In Zoom Out Reset image size

The subplot in figure 3 shows the number of UPIs with measurable device output over time. Three implants consistently generated outputs under US exposure for the whole implantation period, demonstrating UPIs were able to receive US power and function long-term in the animals. Outputs from four devices could not be recorded starting at different time points (Device 1: day 51, Device 2: day 42, Device 4: day 51, Device 5: day 35). Since the event happened abruptly; we initially suspected the issue lay in the interfaces close to device or the connector, perhaps due to shorting or mechanical failure on the connector instead of the devices themselves. This was later confirmed by the in vitro wireless powering test after explanting UPIs. An improved strategy should be used in the future for connector assembly to strengthen the interface between the connector and the lead wires to sustain long-term implantation.

Devices were covered in a fibrotic capsule at the time of harvest as seen in figure 4(B). The incorporation of suture tabs prevented dislodging of the UPIs after implantation. An important aspect for UPIs to be safe and implantable was preventing skin erosion and irritation; the use of Mersilene mesh and rounded shape of the implant encapsulation both helped in this regard. The skin and muscle tissue surrounding the device also showed no signs of macroscopic damage as seen in figures 4(B) and (C). The explanted devices were further characterized for functionality and surface damage with in vitro wireless test and SEM imaging.

Zoom In Zoom Out Reset image size

For post-explant in vitro wireless powering tests, all 7 UPIs were separated from the connector mount assembly before testing. All UPIs were able to generate outputs under US exposure, confirming that the failure in the in vivo stage was due to lead wire related issues at the connector mount interface. Figure 4(D) shows the average and the range of output voltages across all UPI devices under different input voltages across the US transducer. The average output levels measured post-explant increased compared to the ones measured before implantation (figure 1(D)), which could be attributed to the potential difference in transducer placement over the device. Pre- and post-implantation SEM images of the devices are shown in figures 4(E) and (F). It can be observed that the UPI stayed intact, and the materials used did not undergo observable structural changes during the long-term implantation period. Nonetheless, the suture tabs of one device displayed small signs of wear; this deformation is likely due to application of suture at the time of implantation

Muscle and skin surrounding the UPIs were assessed using H&E staining and compared to contralateral (exposed to US) and control (naïve) muscle and skin as shown in figure 5. The epidermal and dermal layer of the skin were visible without signs of excessive fibrosis or formation of necrotic tissues. Similarly, the biceps femoris muscle around the UPI implant did not exhibit characteristics of macroscopic injury when observed at the time of harvest, and the H&E images did not indicate damage across all devices. Both skin and muscle between the three groups were comparable, indicating that the UPI and US exposure did not induce tissue injury. While we did not observe tissue injury in this study, further studies must be conducted to specifically assess tissue injury and establish implant safety.

Zoom In Zoom Out Reset image size