科技公园Cheng Zhou, Jinlei Jiang, Songwei Huang, Junhao Wang, Xinyuan Cui b, Weicheng Wang, Mingrui Chen, Jiawei Peng c, Nanqing Shi, Bensong Wang, Amin Zhang, Qian Zhang, Qichao Li, Shengsheng Cui, Shenghao Xue d, Wei Wang, Ning Tang e,Daxiang Cui, c,
a School of Sensing Science and Engineering, School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University, Shanghai, 200240, PR China
b Department of Radiology, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, 200025, PR China
c National Engineering Center for Nanotechnology, Shanghai, 200240, PR China
d Department of Prothodontics, Shanghai Stomatological Hospital & School of Stomatology, Shanghai Key Laboratory of Craniomaxillofacial Development and Diseases,
Fudan University, Shanghai, 200001, PR China
e Precision Research Center for Refractory Diseases in Shanghai General Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, 200025, PR China
ARTICLEINFO
Keywords:
Near-infrared fluorescence Capsule endoscopy
Specific gastrointestinal diagnoses Wireless power transmission Magnetic-control
ABSTRACT
Early diagnosis of gastrointestinal (GI) diseases is important to effectively prevent carcinogenesis. Capsule endoscopy (CE) can address the pain caused by wired endoscopy in GI diagnosis. However, existing CE ap- proaches have difficulty effectively diagnosing lesions that do not exhibit obvious morphological changes. In addition, the current CE cannot achieve wireless energy supply and attitude control at the same time. Here, we successfully developed a novel near-infrared fluorescence capsule endoscopy (NIFCE) that can stimulate and capture near-infrared (NIR) fluorescence images to specifically identify subtle mucosal microlesions and sub- mucosal lesions while capturing conventional white light (WL) images to detect lesions with significant morphological changes. Furthermore, we constructed the first synergetic system that simultaneously enables multi-attitude control in NIFCE and supplies long-term power, thus addressing the issue of excessive power consumption caused by the NIFCE emitting near-infrared light (NIRL). We performed in vivo experiments to verify that the NIFCE can specifically “light up” tumors while sparing normal tissues by synergizing with probes actively aggregated in tumors, thus realizing specific detection and penetration. The prototype NIFCE system represents a significant step forward in the field of CE and shows great potential in efficiently achieving early targeted diagnosis of various GI diseases.
1. Introduction
According to Global Cancer Statistics 2020, the incidence and mor- tality rates of gastrointestinal (GI) cancers have reached 15.4% and 16.9%, respectively (Sung et al., 2021), which can be primarily attrib- uted to the low early diagnosis rate of GI diseases (Arnold et al., 2020; Yang et al., 2018). Currently, the most common instrument used for clinical GI diagnosis is the wired endoscopy, which can effectively visualize the GI tract except for part of the small bowel. However, this invasive procedure may cause serious complications, as well as
psychological and physical discomfort; as a result, some patients may be hesitant to undergo such diagnostic procedures during routine medical examinations, which adversely affects the early diagnosis rate (Nam et al., 2021; Romero-V´azquez, 2014). Capsule endoscopy (CE) is a novel, minimally invasive GI disease detection approach that does not require patient sedation and enables the observation of the entire GI tract, demonstrating its potential as an effective alternative to wired endos- copy in certain clinical settings (Abramson et al., 2019; Babaee et al., 2019; Bruaene, 2015; Enns et al., 2017; Mimee et al., 2018; Pennazio et al., 2015; Ramadi et al., 2023). However, the mainstream light source
* Corresponding author.
** Corresponding author.
*** Corresponding author. School of Sensing Science and Engineering, School of Electronic Information and Electrical Engineering, Shanghai Jiao Tong University, Shanghai, 200240, PR China.
E-mail addresses: w775276998@163.com (W. Wang), ntang@sjtu.edu.cn (N. Tang), dxcui@sjtu.edu.cn (D. Cui).
https://doi.org/10.1016/j.bios.2024.116209
Received 19 November 2023; Received in revised form 29 February 2024; Accepted 11 March 2024
Available online 24 March 2024
0956-5663/© 2024 Published by Elsevier B.V.
used in endoscopy is WL, which has weak penetration ability, so it is difficult to identify submucosal lesions (Ciaccio et al., 2015; Dickson et al., 2006). Moreover, in this case, compared with wired endoscopy, the resolution of most existing commercial CE systems is also limited by the wireless transmission capacity (Kav, 2012), which makes them difficult to distinguish the lesions with insignificant changes.
Currently, wired endoscopies combined with fluorescence imaging technology have been demonstrated to be effective in improving the diagnosis rate of GI diseases (Moriichi et al., 2016; Papayan and Akopov, 2018; Watanabe et al., 2017; Zhao et al., 2015), which supports the idea
that fluorescence mode has more advantages than WL mode in imaging detection (Cummins et al., 2019; Kosaka et al., 2011; Moriichi et al., 2011; Yim et al., 2020). However, in terms of fluorescence imaging of CE systems, previous studies have either employed commercial CE com- bined with an external optical system to validate the feasibility of CE for capturing fluorescence (Zhang et al., 2008), or integrated photoelectric detectors such as single photon avalanche detectors (Al-Rawhani et al., 2015; Beeley et al., 2018; Melino et al., 2020), photodiodes (Demos- thenous et al., 2016; Nemiroski et al., 2015), spectral sensor (Alam et al., 2020), instead of image sensors and integrated fluorescent excitation
sources instead of WL sources into CE to induce and capture fluorescence signals. Although the fluorescence intensity and spectral information obtained by the above method can theoretically be used for qualitative diagnosis of lesions. However, the lack of an integrated image sensor and WL sources means that conventional WL images of the diseased area cannot be obtained, resulting in the loss of abundant information such as the specific type, shape, and extent, which may lead to misdiagnosis in actual clinical cases.
Up to now, prototype CE that can capture WL images, stimulate and capture fluorescence images at the same time is still rarely reported, mainly because the integration of WL sources, an image sensor, and fluorescent excitation sources which possess large power consumption into a CE system powered by batteries is a serious challenge. Especially in terms of fluorescent excitation sources, NIRL is the best choice due to its strong penetration and capability to stimulate indocyanine green (ICG), the only fluorescent dye approved by the FDA for clinical diag- nosis (Boni et al., 2014). While, the power consumption of the NIRL sources significantly exceeds that of other visible sources. Therefore, it is very important to use wireless energy supply systems instead of batteries to provide uninterrupted energy supply for CE. In addition, the capa- bility to manipulate the motion of the CE is also crucial, as much pathological information, including their fluorescence, can be missed if the CE is driven only by the peristalsis of the GI tract. However, to our knowledge, there is still no prototype system that simultaneously allows manipulation of the CE motion while providing long-term power, which prevents the further clinical use of potential prototype CE that can capture both WL and near-infrared fluorescence (NIRF) images.
Considering the abovementioned limitations, we developed a novel NIFCE and a synergetic system capable of simultaneously powering the NIFCE and enabling motion manipulation (Fig. 1A). The NIFCE can emit NIRL for stimulating NIR fluorescent dye to emit NIRF at higher wave- lengths and capture the NIRF images while emitting WL and capturing WL images. The resulting images can be transmitted wirelessly to the image receiver. In terms of the synergetic system, it includes a wireless power transmission module (WPTM) and a magnetic-controlled module (MCM), in which WPTM replaces the batteries to wirelessly power the NIFCE, while still achieving precise manipulation of the NIFCE motion through the MCM. Based on the multimodal imaging capabilities, as well as the uninterrupted power supply and precise motion manipulation capabilities of the synergetic system, NIFCE enables targeted diagnosis of specific GI diseases. Importantly, the diagnosis is not limited to the obvious mucosal manifestations, but still has diagnostic capabilities for subtle mucosal lesions and submucosal lesions due to the deeper pene- tration of NIRL and the specificity of NIRF. As a proof of concept, we took gastric cancer (MGC-803) as an example to demonstrate the suc- cessful targeted detection of subcutaneous tumors using the synergistic effect of NIFCE and targeted fluorescent probes, indicating that the NIFCE system can efficiently and targeted diagnose GI lesions, providing a new way for capsule endoscopy to improve the accuracy of gastroin- testinal diagnoses.
2. Materials and methods
2.1. Design and hardware architecture of the collapsible rigid-flex printed circuit board (PCB)
The PCB consists of 4 modules: an image acquisition module, a mi- crocontroller unit, a wireless communication module, and a wireless photoelectric switch module. The PCB and system hardware block dia- gram is shown in Figs. S1 and S2. The detailed descriptions of the 4 modules are presented in supplementary materials.
2.2. Encapsulation of the NIFCE
The overall structure of the NIFCE must be compact to be potentially ingestible. The smallest available size was chosen for most of the
electronic components; for instance, we chose the 0201 package for most resistors. The NIFCE incorporates a collapsible rigid-flex PCB (Fig. S1), including five circular PCBs with maximum diameters of
10.26 mm and flexible printed circuit (FPC) flat cables that connect to each circular PCB. This layout allows the PCB to be folded and compact. When the PCB is assembled with the 3-dimensional receiving coil (3DRC), with a height of 11 mm and a diameter of 10.6 mm, and the internal permanent magnet (IPM), with a height of 5 mm and a diameter of 10 mm, the system can be inserted into a capsule with an 11.9 mm diameter (inner diameter 11.05) and 37.5 mm length (Fig. S4). The optical dome of the capsule is constructed from polycarbonate, and the remaining capsule is made of acrylonitrile butadiene styrene plastic.
2.3. Image receiver
An Android application (app) was developed to receive and browse the images transmitted by the NIFCE in real time and control the NIFCE to switch shooting modes. Nordic Semiconductor provided the source code of the app, and the app we used was modified in Android Studio from the source code. We connected a commercial Android platform to an 8-channel power splitter and 8 FPC antennas to construct an image receiver (Fig. S5), where the multiple antennas can efficiently receive data. In addition, any Android smartphone with the app installed can be used as an image receiver (Fig. S5).
2.4. System timing and data flow
The microcontroller is initialized when the NIFCE is powered on, and the CMOS image sensor is configured by the microcontroller to capture images and send data with an output format of RGB565 and the VGA image sizes through DVP to the microcontroller. The RGB565 image is first converted to YCbCr 4:2:2 format, then encoded as a JPEG- compressed image at a very fast speed by the microcontroller. Finally, the image is sent to the wireless communication module via the SPI and transmitted wirelessly at 2 frames per second (fps).
2.5. Wireless power transmission module (WPTM)
The transmitting terminal of the WPTM consists of a transmitting coil, an inverter control PCB, a fixed vacuum capacitor, an adjustable vacuum capacitor and two DC voltage sources (Figs. S6 and S7). The receiving terminal of the WPTM consists of a 3DRC and a rectifier- regulator PCB (Fig. S8). The detailed descriptions of the WPTM are presented in supplementary materials.
2.6. Magnetic-controlled module (MCM)
The MCM consists of a cylindrical external permanent magnet (EPM), an EPM drive unit (EDU), an electronic control cabinet, and a cylindrical IPM (Figs. S9–S11). The detailed descriptions of the MCM are
presented in supplementary materials.
2.7. Probe fabrication
The probes consisted of liposomes encapsulated with ICG and then externally attached to folic acid. The probes are not the focus of this study and were only used to verify the performance of the NIFCE system and were prepared by referring to our previous work without modifi- cation (Wang et al., 2022). The targeting of folic acid for gastric cancer was demonstrated in our previous work (Chang et al., 2019; Ma et al., 2012; Zhang et al., 2015).
2.8. Measurement
The emission spectral characteristics of the NIR LED were measured by a spectrograph (omin-labda300, Zolix). The emission spectral
characteristics of the white LED were measured by a spectrograph (HP8000Pro, DUOTONE CLOUD). The fluorescence spectrum of ICG was measured by a spectrofluorometer (FL-4600, Hitachi).
2.9. Mask R–CNN
Mask R–CNN is a very flexible deep learning algorithm. We cropped the captured fluorescence images and passed them through multiple convolution layers to gradually extract the image features. The region
proposal network (RPN) was utilized to generate the region of interest (ROI), which potentially includes the target object, and each ROI was assigned a score and a suggestion regarding the bounding box’s location.
The mask generation network (MGN) conducts convolution and
upsampling operations based on each feature in every ROI to generate a corresponding target mask corresponding to the ROI, indicating the pixel positions of the target objects in the image. Additionally, another network branch outputs the classification information.
2.10. Cytotoxicity evaluation
Live/dead staining: To assess the cell compatibility of the NIFCE, human gastric mucosal GES-1 cells were seeded at a density of 5 × 10^4 cells per well in 24-cell plates. After 24 h of culture in medium, the cells
were divided into three groups of three subgroups each. Cells in the first and second subgroups were irradiated for 10 min with NIRL and WL by the NIFCE, respectively. Cells in the third subgroups were left untreated. After irradiation, live/dead working solution was added, and the cells were incubated for 30 min. The cells were then rinsed with PBS, fol- lowed by observation under an inverted fluorescence microscope.
Cell viability assay: To assess the cell compatibility of the NIFCE, human gastric mucosal GES-1 cells were seeded at a density of 3 × 10^4 cells per well in a 96-well plate, the cells were divided into three groups.
Cells in the first and second groups were irradiated for 10 min with NIRL and WL by the NIFCE, respectively. Cells in the third group were left untreated, then cultured in medium for 48 h. The optical density (OD) was measured at 450 nm using the CCK-8 assay at 0 h (before irradia- tion) and 48 h time points.
2.11. Animals and tumor model
Nude mice (5 weeks old, approximately 20 g) were purchased from Gempharmatech Co., Ltd. (China). All experiments conducted on the animals were approved by the Administrative Committee on Animal Research of Shanghai Jiao Tong University (approval number
202201131). To establish the tumor model, MGC-803 cells (4 × 106 cells in 100 μL complete medium) were transplanted subcutaneously into the right leg of the mice. The mice were used for the experiments until the tumor size reached approximately 150 mm3.
2.12. In vivo fluorescence imaging
100 μL of probes was intravenously injected into tumor-bearing mice, and the equivalent ICG concentration for mice was 5 mg/kg.
Then, fluorescence images of mice were recorded by the NIFCE at 1 h, 4 h, 12 h, 24 h and 48 h. As a verification, fluorescence images were recorded by an imaging system (IVIS spectrum, PerkinElmer) at 24 h and 48 h with the following parameters: excitation at 745 nm, emission at 840 nm, and exposure time of 5 s.
3. Results and discussion
3.1. Design of NIFCE with multimodal imaging capabilities
The NIFCE consists of a collapsible rigid-flex PCB with NIRL and WL emission, NIRF and WL image capture, image compression, and wireless image transmission functions (Fig. 1A and Figs. S1 and S2), a 3DRC
(Fig. 1B), a rectifier-regulator PCB (Fig. S8), and an IPM (Fig. 1C), which are encapsulated within an ingestible capsule (Fig. 1A–C and Fig. S4). The key for NIFCE can capture NIRF and WL simultaneously is a band-
stop filter (Fig. 1D), which effectively blocks the NIRL emitted by the NIFCE from entering the CMOS image sensor and enables the NIRF generated by the fluorescent materials (e.g., ICG) to enter the CMOS (Fig. 1D and Figs. S12 and S13), ensuring that weak fluorescence signals
are not drowned out due to the NIFCE’s NIRL emissions. In addition, the
WL emitted by the NIFCE is virtually unobstructed by the filter (Fig. 1D and Fig. S13); therefore, the WL and NIRF can enter the COMS simul- taneously. In such configuration, the NIFCE can obtain different mo- dalities of images by switching between the luminous modes of simultaneously emitting NIRL and WL, alternately emitting NIRL and WL, emitting NIRL only, and emitting WL only under the control of the image receiver (Fig. 1E and Fig. S5 and Movie S2). This configuration is realized by precisely matching the cutoff band of the bandstop filter with the spectra of the WL and NIRL sources (Fig. 1D), which eliminates the need for space-consuming methods such as adding filters to the light sources or using dichroic filters. After irradiating gastric mucosal cells with the NIFCE by emitting WL and NIRL for 10 min, almost no cell death was observed, similar to the observations in the control group without exposure (Fig. 1F and Fig. S14), demonstrating the biocom- patibility of the NIFCE. The results of cell viability assay also showed that there were no significant differences between WL, NIRL and control group after finishing irradiating for 48 h, further indicating the biocompatibility of the NIFCE (Fig. S15).
3.2. Multimodal imaging capabilities characterization of the NIFCE
ICG solutions at varying concentrations were imaged by the NIFCE under dark conditions (Fig. 2A), the calculated quantitative fluorescence intensities were shown in Fig. S16, demonstrating the capability of the NIFCE to stimulate and capture sufficient NIRF of ICG even under a low concentration (0.0005 mg/ml). The multimodal imaging capabilities of the NIFCE in biological tissues were evaluated using porcine gastric tissue drip-spiked with ICG solution under dark conditions (Fig. 2B and Figs. S17 and S18). In the NIRF&WL modality image, the surface mucosal and fluorescence contents of the porcine gastric tissue can be observed simultaneously, maximizing the information included in a single image. In the WL modality image, clear surface mucosal images can be obtained. The NIRF modality image focuses solely on the fluo- rescence signal, showing a high contrast. Based on the images of different modalities, it is evident that the ICG solution (hypothesized as a lesion with no obvious appearance that adsorbed fluorescent mate- rials), which is completely invisible in the WL image but clearly revealed in the NIRF image, creating a distinct differentiation from surrounding tissues, proving that the NIFCE has immense potentials in detecting le- sions that have not yet induced apparent morphological alterations on the mucosal surface.
To assess the penetration capability of the NIFCE, we overlaid an additional piece of porcine gastric tissue onto the ICG solution-dripped porcine gastric tissue and captured images using the NIFCE before and after the overlay process (Fig. 2C and Fig. S19). The fluorescence was clearly visible even after covering the tissue, indicating the potential of the NIFCE to detect submucosal lesions. To evaluate the concrete penetration capability of the NIFCE, seven pieces of porcine gastric tis-
sues with different thicknesses were prepared and dropped with 5 μL of
ICG solution at a concentration of 0.005 mg/ml and then took NIR fluorescence images of the front and back side of the tissues using the NIFCE. Given that the fluorescence has become blurred on the back side of the fifth piece of tissue and is too weak to distinguish on the back side of the sixth piece, the maximum depth of penetration under the current NIFCE configuration is about 3 mm corresponding to the fifth stomach tissue (Fig. S20). To further validate the penetrability of the NIFCE for in vivo diagnosis, ICG was injected into the tumor site in mice, which was then imaged by the NIFCE. Despite the presence of a cutaneous barrier,
the tumor exhibited distinct fluorescence (Fig. 2D and Fig. S21 and Movie S3), demonstrating the remarkable potential of the NIFCE for the diagnosis of submucosal lesions in vivo. We assessed the capability of the NIFCE to image fluorescent substances actively aggregated in lesions in tumor-bearing mice using folate-based targeted fluorescent probes. The probes were administered intravenously into the mice. The images acquired by the NIFCE and the control images acquired by an in vivo imaging system showed that due to the metabolism of the probes, the fluorescence in the mice gradually changed from being distributed throughout the body to only the tumor showing obvious fluorescence, while the healthy tissues around the tumor no longer emitted fluores- cence (Fig. 2E). This result demonstrates that the NIFCE system suffi- ciently stimulated and captured the fluorescence of the residual probes, which remained at the tumor site after 48 h of metabolism. This suc- cessful in vivo demonstration of targeted diagnosis of subcutaneous le- sions with the NIFCE shows the considerable potential of the proposed system for in vivo specific lesion diagnosis in the GI tract.
The high contrast in the fluorescence images enables accurate and rapid automated identification of the fluorescence in the image and segmentation of the fluorescent regions using artificial intelligence (AI) algorithms. We used Mask R–CNN (Fig. 2F), a convolutional neural
network (CNN), to identify fluorescence in the images acquired by the
NIFCE in NIR mode and segment the fluorescent regions with ultrahigh accuracy (Fig. S22). The dataset, including 150 fluorescence images of porcine stomachs captured by the NIFCE, was partitioned into a training set and a validation set, with 100 images allocated for training and the remaining images used for validation. We trained the model 100 times, observing a gradual convergence of the loss function after approxi- mately 20 training iterations (Fig. 2G). Moreover, the mean average
accuracy (mAP) index progressively approached 1 (Fig. 2H), demon- strating the efficacy of our model. We adopted the strategy of manually adjusting the learning rate (lr) in the 16th training iteration (Fig. 2G), aiming to ensure that the model did not fall into local optima during training, which may impact the final result.
3.3. Design and demonstration of the WPTM capable of powering the NIFCE
With the WPTM (Fig. 3A and Figs. S6 and S7), the NIFCE can be powered by a 3DRC instead of batteries; thus, the NIFCE can effectively afford the high-power consumption of NIRL sources and can be de- energized at any time if necessary, ensuring endurance and safe opera- tion. The pulse width modulation (PWM) generator in the inverter control PCB generates two square waves with opposite phases, 50% duty cycles, frequencies of 100 kHz, and voltage of 15 V (Fig. S23). Each of the waves drives a gate driver chip that controls a half-bridge circuit including two N-channel enhancement MOSFETs, forming a full-bridge inverter circuit, which is used to convert direct current (DC) into alter- nating current (AC) to power the transmitting coil. The transmitting coil generates an alternating magnetic field with a frequency of 100 kHz due to the Biot-Savart law. The 3 solenoids in the 3DRC are independent and
generate AC power in the alternating magnetic field according to Far- aday’s law of electromagnetic induction. The three AC signals are con- verted to DC by three rectifiers and then connected in parallel to a DC-
DC voltage regulator chip. This arrangement prevents the induced electromotive force (EMF) generated by the three solenoids from canceling mutually due to their opposite directions. Additionally, by utilizing the voltage clamp of the diodes in the rectifier bridges, only the
solenoid with the highest induced EMF forms a closed loop for energy output, while the induced EMF generated by the other two solenoids cannot form a loop. Consequently, the interference among the magnetic fields generated by the current from the three closely wound solenoids, which can lead to the detuning of the LC loop, is effectively avoided, ensuring resonance and optimal transmission efficiency for the solenoid that outputs energy. The DC with the largest voltage is regulated by the DC-DC voltage regulator and connected to the power input of the NIFCE. Before setting up the system, we conducted finite element simula- tions to analyze the performance of the transmitting and receiving coil. We assessed the magnetic flux density distribution along the axis plane (Fig. 3B) and the section perpendicular to the axis at distances of 0 mm, 40 mm, 80 mm, 120 mm, 160 mm, and 200 mm from the center of the transmitter coil (Fig. S24). The simulation results demonstrate that the distribution of the magnetic flux density is relatively uniform in the internal space of the transmitting coil. In the area between the two so- lenoids, the strength of the magnetic field decreases with increasing distance from the axis, while in the solenoid, the magnetic field strength increases with decreasing distance from the edge of the solenoid. We also evaluated the power responses of the three solenoids in the 3DRC at the center of the transmitting coil (coaxial with the coil), including the total power, consumed power, and output power under different capacitance matching conditions (Fig. 3C and Fig. S25). It is evident that the matching capacitances exert a vital influence on the magnitude of power. After constructing the system, the output power of the 3DRC (fixed capacitance matching) was measured at the center of the trans- mitting coil with each of the three solenoids coaxial oriented towards the transmitting coil, which was less than the maximum simulation
output power (Fig. 3C and D and Fig. S25). The primary reason for this difference is that the simulation obtains the output power of a single solenoid under optimal impedance matching conditions; however, per- fect impedance matching cannot be achieved in practice, and the sole- noids interfere with each other. Additionally, certain idealized assumptions were made when setting up the coil model, such as neglecting the skin effect. We tested the output power of the 3DRC at 12 orientations and 3 representative positions inside the transmitting coil (Fig. 3E and Fig. S26). Similar to the simulation results of the trans- mitting coil (Fig. 3B), we observed that the output power at each orientation was highest at the position with the largest magnetic flux density (15.5, 0, 0), while the output power was the lowest at the po- sition with the smallest magnetic flux density (0, 10, 0). The minimum recorded output power during testing was 374 mW, which is sufficient to supply energy to the NIFCE.
The magnetic field strength of the IPM is significantly higher than that of the EPM in the vicinity of 3DRC; thus, only the effect of the interference of the IPM on energy reception needs to be considered. The output power of the 3DRC was tested at six orientations, both with and without the additional magnetic field generated by the IPM at the center of the transmitting coil (Fig. 3F). When solenoid@3 (outermost dimension) was coaxial with the transmitting coil, the presence of the IPM led to a slight decrease (15%) in the output power. The output power of the 3DRC is essentially unaffected by the IPM when sole- noid@1 (innermost dimension) and solenoid@2 (middle dimension) are coaxial with the transmitting coil. This is because we assembled the IPM coaxially with solenoid@3 of the 3DRC with a certain assembly gap in the interior of NIFCE. Therefore, for the two inner solenoids, the bias
induced by the IPM static magnetic field is symmetrical, which does not have a significant impact on their output power. However, for the so- lenoid@3, it has a magnetic field bias along the IPM axis, which reduces the increased effect of the magnetic flux. Nevertheless, the presence of an assembly gap mitigates the impact of this bias. Therefore, the overall impact of the magnetic field generated by the IPM on the wireless power supply is minimal. The assembled NIFCE without battery was positioned
at five different locations inside the transmitting coil, and its perfor- mance was tested at five deflection angles (0◦, 45◦, 90◦, 135◦, and 180◦). The results show that the NIFCE is supplied with sufficient energy
wirelessly under the WPTM to enable various shooting modes (WL mode, NIRL mode, and WL & NIRL mode) at different positions and angles (Fig. 3G and Fig. S27). The infrared thermograms were taken after 5 min of operation of NIFCE in air, and liquid near body temper- ature respectively under the power supply of WPTM. The results show
that in air, the NIFCE temperature after 5 min of work is only 22.1 ◦C,
with no rapid warming. In water, the temperature of the water is slowly decreasing because the room temperature is lower than the water tem- perature, and the temperature of NIFCE after 5 min of work is still covered by the temperature of water, demonstrating that there is also no obvious warming of NIFCE (Fig. S28). These results indicate that NIFCE will not generate harmful Joule heating to biological tissues during its operation.
3.4. Design and demonstration of the MCM capable of manipulating the NIFCE motion
The MCM is designed to manipulate the position and orientation of the NIFCE (Fig. 4A and Figs. S9–S11), such as controlling the NIFCE to remain at the location where the fluorescence images were captured to
acquire additional images of potential lesions. The two shafts controlling the rotation of the EPM are nested within each other and are differen- tially coordinated through a bevel gear set, thus enabling multidirec- tional rotation of the EPM (Fig. 4A). Before setting up the system, we conducted finite element simulations of the EPM and IPM to determine the magnetic flux density distributions of the EPM and IPM (Fig. 4B and C). Subsequently, the global electromagnetic force and torque of the IPM were calculated to assess the manipulation effect of the EPM on the IPM. We established six motion modes for the EPM, including translation along the X, Y, and Z axes, pole-to-pole translation along the Z axis, and rotation around the X and Z axes. Based on the simulation requirements for different positions and orientations, the model was parametrically scanned; then, a steady-state solver was used to obtain the final results. The changes in the force and torque of the IPM in the X, Y, and Z di-
rections with the movement of EPM were obtained in these six cases (Fig. 4D and E and Figs. S29–S33). The results show that within a certain range, the EPM can effectively control the IPM motion.
(B) and IPM (C) obtained from simulations (T: Magnetic flux density norm). (D and E) The changes in force and torque of the IPM in X, Y, and Z directions with the movement of EPM translation along the Y axes (D) and rotation around the Z axes (E) obtained from simulations (T: Magnetic flux density norm). (F and G) The NIFCE was positioned above (F) and below (G) the empty gastric model and rotated in multiple directions controlled by MCM. (H) The MCM manipulated the movement of NIFCE to various positions within the water-filled gastric model.
We mounted the assembled MCM above the NIFCE and tested its manipulation capability. In a water-free environment, the NIFCE is subjected to the magnetic force and torque exerted by the EPM as well as its own gravity. By adjusting the height of the EPM, the NIFCE can be positioned above and below the gastric model. By precisely controlling the orientation of the EPM, the NIFCE can capture images at different angles (Fig. 4F and G). When the gastric model is filled with water, the NIFCE is subjected to a buoyant force from the water. By adjusting the EPM, the NIFCE can be suspended in the water to capture images and moved to various positions within the gastric model, such as the cardia and pylorus (Fig. 4H and Fig. S34 and Movie S4). The NIFCE can also be controlled to accurately reach to and maintain stationary in the pre- marked target area and point to a specific orientation (Fig. S35 and Movie S5). In addition, the motion of the NIFCE is controlled only by the MCM, the WPTM will not cause additional rotation and displacement of the NIFCE (Fig. S36 and Movie S6).
3.5. Overall integration and demonstration of the synergetic system
We successfully addressed several critical issues to enable seamless compatibility and collaboration among all system modules. For the WPTM, in accordance with Faraday’s law, the energy transmission ef-
ficiency increases as the alternating frequency increases. However, an
excessively high frequency can severely impede communication be- tween the NIFCE and the image receiver. Therefore, to minimize inter- ference from the alternating magnetic field on communication while ensuring that the NIFCE receives sufficient energy, we adjusted the alternating frequency to 100 kHz after several iterations of frequency tuning and impedance matching. Additionally, we constructed a wire-
less communication module with a maximal transmission power of +8
dB m, equipped the image receiver with 8 antennas, and included electromagnetic shielding in the NIFCE. Thus, under the interference of intense magnetic field, the NIFCE can transmit images to the image receiver in real time, and different shooting modes can be switched. Due to the coaxial assembly of the IPM and the solenoid@3 of 3DRC with a certain assembly gap, the alternating magnetic field generated by the WPTM and the static magnetic field generated by the MCM are compatible with each other. Therefore, the integration of the WPTM and the MCM resulted in a synergetic system that could simultaneously supply power and manipulate the motion of the NIFCE (Fig. S37).
After the system was assembled, we first conducted tests in a gastric model (Fig. S38 and Movie S7). The NIFCE successfully established an RF connection with the image receiver and transmitted images wire- lessly in real time under the power supply of the synergetic system. When motion occurred under the control of the synergetic system, both the power supply and the RF signal remained connected. Subsequently, we conducted tests in complete porcine gastric tissue filled with water (Fig. 5A and Movie S8). The NIFCE also synergistically enabled power
supply, motion manipulation, and an RF connection in the presence of a barrier formed by water and biological tissue under the interference of intense magnetic field. And the NIFCE successfully switched the shoot- ing mode under the control of the image receiver to illuminate the whole gastric tissue with NIRL and WL (Fig. 5B and C). Due to the stronger penetrability of NIRL, the gastric tissue appeared brighter. The above results effectively demonstrate the concurrent capabilities of the syn- ergetic system in supplying energy to the NIFCE and manipulating its motion. Furthermore, the alternating and static magnetic fields gener- ated by the synergetic system are compatible and do not interfere with the RF connection, including wireless images transmission and switch- ing the shooting mode.
4. Conclusion
Here, we developed the first NIFCE system that can provide long- term power supply to the NIFCE, enabling the emission of NIRL and the capture of NIRF images, with precise motion control. Therefore, in addition to capturing WL images for routine diagnosis, the NIFCE en- ables targeted diagnosis of specific GI diseases in collaboration with the targeted fluorescent probes (mainly composed of NIR dye and the target ligand). In vivo experiments have demonstrated that the lesions
enriched with probes are stimulated and emitting NIRF under NIFCE’s
NIRL irradiation, which is significantly different behavior than the healthy tissues surrounding the lesions, indicating the specific detection and penetration capabilities of NIFCE. The NIFCE system represents a significant step forward in the field of CE and is anticipated to compensate for the limited resolution of CE by providing specific di- agnoses, thereby improving the diagnosis rate of CE. In the present work, we just fabricated the initial prototype of the NIFCE system and verified its feasibility. Future work is needed to validate its capability to examine specific lesions in the GI tract in vivo. Moreover, since the power supply issue with the NIFCE has been resolved, the high power consumed NIR laser diodes can be integrated into the NIFCE for future potential therapeutic applications.
Since the current research is still in the laboratory stage, the func- tional prototype is built in order to verify the feasibility, so the size is not strictly limited in accordance with clinical requirements. At present, the most space-consuming are the 3DRC and the IPM. In updated NIFCE versions, the shape and material of the magnetic aggregator can be optimized, fewer solenoid@3 can be wound and a thinner frame can be used to shorten the 3DRC length. The length of the IPM can be shortened by using an IPM with higher magnetic force and an EPM with stronger magnetic force and larger size. In addition, the size of NIFCE can also be reduced by utilizing thinner PCBs, minimizing the gap margin in the folded PCB, etc.
CRediT authorship contribution statement
Cheng Zhou: Writing – original draft, Validation, Project adminis- tration, Methodology, Investigation, Data curation, Conceptualization. Jinlei Jiang: Writing – original draft, Visualization, Validation, Meth- odology, Investigation. Songwei Huang: Visualization. Junhao Wang:
Software. Xinyuan Cui: Formal analysis. Weicheng Wang: Validation. Mingrui Chen: Investigation. Jiawei Peng: Validation. Nanqing Shi: Investigation. Bensong Wang: Investigation. Amin Zhang: Investiga- tion, Funding acquisition. Qian Zhang: Investigation, Funding acqui- sition. Qichao Li: Validation. Shengsheng Cui: Investigation.
Shenghao Xue: Visualization. Wei Wang: Validation, Supervision, Funding acquisition, Data curation. Ning Tang: Writing – review & editing, Visualization, Validation, Investigation. Daxiang Cui: Writing – review & editing, Resources, Project administration, Investigation,
Funding acquisition, Conceptualization.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Data availability
Data will be made available on request.
Acknowledgements
Great thanks for the financial support from Shanghai 2020 Science and Technology Innovation Action Plan, China (No. 20142201300), Projects of INTERNATIONAL COOPERATION and Exchanges NSFC (No. 82020108017), China Postdoctoral Science Foundation (No. 2023M732267, No. 2020TQ0191 and 2021M702139), National Facility for Translational Medicine (Shanghai) Open Project Fund (TMSK-2021- 302), National Natural Science Foundation of China (No. 82272821), Natural Science Foundation of Shanghai (No. 22ZR1467600).
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi. org/10.1016/j.bios.2024.116209.
References
Abramson, A., Caffarel-Salvador, E., Khang, M., Dellal, D., Silverstein, D., Gao, Y., Frederiksen, M.R., Vegge, A., Huba´lek, F., Water, J.J., Friderichsen, A.V., Fels, J., Kirk, R.K., Cleveland, C., Collins, J., Tamang, S., Hayward, A., Landh, T., Buckley, S. T., Roxhed, N., Rahbek, U., Langer, R., Traverso, G., 2019. An ingestible self- orienting system for oral delivery of macromolecules. Science 363 (6427), 611–615.
https://doi.org/10.1126/science.aau2277.
Al-Rawhani, M.A., Beeley, J., Cumming, D.R.S., 2015. Wireless fluorescence capsule for endoscopy using single photon-based detection. Sci. Rep. 5 (1), 18591 https://doi. org/10.1038/srep18591.
Alam, M.W., Vedaei, S.S., Wahid, K.A., 2020. A fluorescence-based wireless capsule endoscopy system for detecting colorectal cancer. Cancers 12 (4), 890. https://doi. org/10.3390/cancers12040890.
Arnold, M., Abnet, C.C., Neale, R.E., Vignat, J., Giovannucci, E.L., McGlynn, K.A., Bray, F., 2020. Global burden of 5 major types of gastrointestinal cancer.
Gastroenterology 159 (1), 335–349. https://doi.org/10.1053/j.gastro.2020.02.068.
Babaee, S., Pajovic, S., Kirtane, A.R., Shi, J.Y., Caffarel-Salvador, E., Hess, K., Collins, J. E., Tamang, S., Wahane, A.V., Hayward, A.M., Mazdiyasni, H., Langer, R., Traverso, G., 2019. Temperature-responsive biometamaterials for gastrointestinal applications. Sci. Transl. Med. 11 (488), 13. https://doi.org/10.1126/scitranslmed. aau8581.
Beeley, J., Melino, G., Al-Rawahani, M., Turcanu, M., Stewart, F., Cochran, S., Cumming, D., 2018. Imaging fluorophore-labelled intestinal tissue via fluorescence endoscope capsule. Eurosensors 2, 766. https://doi.org/10.3390/ proceedings2130766.
Boni, L., David, G., Mangano, A., Dionigi, G., Rausei, S., Spampatti, S., Cassinotti, E., Fingerhut, A., 2014. Clinical applications of indocyanine green (ICG) enhanced
fluorescence in laparoscopic surgery. Surg. Endosc. 29 (7), 2046–2055. https://doi. org/10.1007/s00464-014-3895-x.
Bruaene, C.V.d., 2015. Small bowel capsule endoscopy: where are we after almost 15 years of use? World J. Gastrointest. Endosc. 7 (1), 13. https://doi.org/10.4253/wjge. v7.i1.13.
Chang, J., Zhang, A., Huang, Z., Chen, Y., Zhang, Q., Cui, D., 2019. Monodisperse Au@ Ag core-shell nanoprobes with ultrasensitive SERS-activity for rapid identification and Raman imaging of living cancer cells. Talanta 198, 45–54. https://doi.org/10.1016/j.talanta.2019.01.085.
Ciaccio, E.J., Bhagat, G., Lewis, S.K., Green, P.H., 2015. Suggestions for automatic quantitation of endoscopic image analysis to improve detection of small intestinal pathology in celiac disease patients. Comput. Biol. Med. 65, 364–368. https://doi.org/10.1016/j.compbiomed.2015.04.019.
Cummins, G., Cox, B.F., Ciuti, G., Anbarasan, T., Desmulliez, M.P.Y., Cochran, S., Steele, R., Plevris, J.N., Koulaouzidis, A., 2019. Gastrointestinal diagnosis using non- white light imaging capsule endoscopy. Nat. Rev. Gastroenterol. Hepatol. 16 (7), 429–447. https://doi.org/10.1038/s41575-019-0140-z.
Demosthenous, P., Pitris, C., Georgiou, J., 2016. Infrared fluorescence-based cancer
screening capsule for the small intestine. IEEE Trans. Biomed. Circuits Syst. 10 (2), 467–476. https://doi.org/10.1109/tbcas.2015.2449277.
Dickson, B.C., Streutker, C.J., Chetty, R., 2006. Coeliac disease: an update for pathologists. J. Clin. Pathol. 59 (10), 1008–1016. https://doi.org/10.1136/jcp.2005.035345.
Enns, R.A., Hookey, L., Armstrong, D., Bernstein, C.N., Heitman, S.J., Teshima, C., Leontiadis, G.I., Tse, F., Sadowski, D., 2017. Clinical practice guidelines for the use of video capsule endoscopy. Gastroenterology 152 (3), 497–514. https://doi.org/10.1053/j.gastro.2016.12.032.
Kav, T., 2012. Is enteroscopy necessary for diagnosis of celiac disease? World J. Gastroenterol. 18 (31), 4095. https://doi.org/10.3748/wjg.v18.i31.4095.
Kosaka, N., Mitsunaga, M., Longmire, M.R., Choyke, P.L., Kobayashi, H., 2011. Near infrared fluorescence-guided real-time endoscopic detection of peritoneal ovarian cancer nodules using intravenously injected indocyanine green. Int. J. Cancer 129 (7), 1671–1677. https://doi.org/10.1002/ijc.26113.
Ma, J., Huang, P., He, M., Pan, L., Zhou, Z., Feng, L., Gao, G., Cui, D., 2012. Folic acid-
conjugated LaF3:Yb,Tm@SiO2 nanoprobes for targeting dual-modality imaging of upconversion luminescence and X-ray computed tomography. J. Phys. Chem. B 116 (48), 14062–14070. https://doi.org/10.1021/jp309059u.
Melino, G., Accarino, C., Riehle, M., Potter, M., Fineron, P., Annese, V.F., Grant, J.P., Al-
Rawhani, M.A., Beeley, J., Carranza, I.E., Cumming, D.R.S., 2020. Capsule endoscopy compatible fluorescence imager demonstrated using bowel cancer tumours. IEEE Sens. J. 20 (17), 9763–9771. https://doi.org/10.1109/jsen.2020.2990293.
Mimee, M., Nadeau, P., Hayward, A., Carim, S., Flanagan, S., Jerger, L., Collins, J., McDonnell, S., Swartwout, R., Citorik, R.J., Bulovic, V., Langer, R., Traverso, G., Chandrakasan, A.P., Lu, T.K., 2018. An ingestible bacterial-electronic system to monitor gastrointestinal health. Science 360 (6391), 915–918. https://doi.org/ 10.1126/science.aas9315.
Moriichi, K., Fujiya, M., Okumura, T., 2016. The efficacy of autofluorescence imaging in the diagnosis of colorectal diseases. Clin. J. Gastroenterol. 9 (4), 175–183. https:// doi.org/10.1007/s12328-016-0658-3.
Moriichi, K., Fujiya, M., Sato, R., Nata, T., Nomura, Y., Ueno, N., Ishikawa, C., Inaba, Y.,
Ito, T., Okamoto, K., Tanabe, H., Mizukami, Y., Watari, J., Saitoh, Y., Kohgo, Y., 2011. Autofluorescence imaging and the quantitative intensity of fluorescence for evaluating the dysplastic grade of colonic neoplasms. Int. J. Colorectal Dis. 27 (3), 325–330. https://doi.org/10.1007/s00384-011-1311-8.
Nam, J.H., Lee, K.H., Lim, Y.J., 2021. Examination of entire gastrointestinal tract: a
perspective of mouth to anus (M2A) capsule endoscopy. Diagnostics 11 (8), 1367. https://doi.org/10.3390/diagnostics11081367.
Nemiroski, A., Ryou, M., Thompson, C.C., Westervelt, R.M., 2015. Swallowable fluorometric capsule for wireless triage of gastrointestinal bleeding. Lab Chip 15 (23), 4479–4487. https://doi.org/10.1039/c5lc00770d.
Papayan, G., Akopov, A., 2018. Potential of indocyanine green near-infrared
fluorescence imaging in experimental and clinical practice. Photodiagnosis Photodyn. Ther. 24, 292–299. https://doi.org/10.1016/j.pdpdt.2018.10.011.
Pennazio, M., Spada, C., Eliakim, R., Keuchel, M., May, A., Mulder, C., Rondonotti, E.,
Adler, S., Albert, J., Baltes, P., Barbaro, F., Cellier, C., Charton, J., Delvaux, M., Despott, E., Domagk, D., Klein, A., McAlindon, M., Rosa, B., Rowse, G., Sanders, D., Saurin, J., Sidhu, R., Dumonceau, J.-M., Hassan, C., Gralnek, I., 2015. Small-bowel capsule endoscopy and device-assisted enteroscopy for diagnosis and treatment of small-bowel disorders: European Society of Gastrointestinal Endoscopy (ESGE)
Clinical Guideline. Endoscopy 47 (4), 352–386. https://doi.org/10.1055/s-0034-1391855.
Ramadi, K.B., McRae, J.C., Selsing, G., Su, A.R., Fernandes, R., Hickling, M., Rios, B.,
Babaee, S., Min, S.K., Gwynne, D., Jia, N.Z., Aragon, A., Ishida, K., Kuosmanen, J., Jenkins, J., Hayward, A., Kamrin, K., Traverso, G., 2023. Bioinspired, ingestible electroceutical capsules for hunger-regulating hormone modulation. Sci. Robot. 8 (77), 13. https://doi.org/10.1126/scirobotics.ade9676.
Romero-Va´zquez, J., 2014. Capsule endoscopy in patients refusing conventional
endoscopy. World J. Gastroenterol. 20 (23), 7424. https://doi.org/10.3748/wjg. v20.i23.7424.
Sung, H., Ferlay, J., Siegel, R.L., Laversanne, M., Soerjomataram, I., Jemal, A., Bray, F., 2021. Global cancer Statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J. Clin. 71 (3), 209–249. https://doi.org/10.3322/caac.21660.
Wang, R., Song, C., Gao, A., Liu, Q., Guan, W., Mei, J., Ma, L., Cui, D., 2022. Antibody-conjugated liposomes loaded with indocyanine green for oral targeted photoacoustic
imaging-guided sonodynamic therapy of Helicobacter pylori infection. Acta Biomater. 143, 418–427. https://doi.org/10.1016/j.actbio.2022.02.031.
Watanabe, M., Murakami, M., Ozawa, Y., Yoshizawa, S., Matsui, N., Aoki, T., 2017.
Intraoperative identification of colonic tumor sites using a near-infrared fluorescence endoscopic imaging system and indocyanine green. Dig. Surg. 34 (6), 495–501. https://doi.org/10.1159/000458450.
Yang, L., Zheng, R., Wang, N., Yuan, Y., Liu, S., Li, H., Zhang, S., Zeng, H., Chen, W., 2018. Incidence and mortality of stomach cancer in China, 2014. Chin. J. Cancer Res. 30 (3), 291–298. https://doi.org/10.21147/j.issn.1000-9604.2018.03.01.
Yim, J.J., Harmsen, S., Flisikowski, K., Flisikowska, T., Namkoong, H., Garland, M., van
den Berg, N.S., Vilches-Moure, J.G., Schnieke, A., Saur, D., Glasl, S., Gorpas, D., Habtezion, A., Ntziachristos, V., Contag, C.H., Gambhir, S.S., Bogyo, M., Rogalla, S., 2020. A protease-activated, near-infrared fluorescent probe for early endoscopic detection of premalignant gastrointestinal lesions. Proc. Natl. Acad. Sci. USA 118 (1), e2008072118. https://doi.org/10.1073/pnas.2008072118.
Zhang, C., Li, C., Liu, Y., Zhang, J., Bao, C., Liang, S., Wang, Q., Yang, Y., Fu, H., Wang, K., Cui, D., 2015. Gold nanoclusters-based nanoprobes for simultaneous fluorescence imaging and targeted photodynamic therapy with superior penetration and retention behavior in tumors. Adv. Funct. Mater. 25 (8), 1314–1325. https://doi.org/10.1002/adfm.201403095.
Zhang, H., Morgan, D., Cecil, G., Burkholder, A., Ramocki, N., Scull, B., Lund, P.K., 2008. Biochromoendoscopy: molecular imaging with capsule endoscopy for detection of adenomas of the GI tract. Gastrointest. Endosc. 68 (3), 520–527. https://doi.org/10.1016/j.gie.2008.02.023.
Zhao, Z.-Y., Guan, Y.-G., Li, B.-R., Shan, Y.-Q., Yan, F.-H., Gao, Y.-J.-Y., Wang, H., Lou, Z.,
Fu, C.-G., Yu, E.-D., 2015. Detection and miss rates of autofluorescence imaging of adenomatous and polypoid lesions during colonoscopy: a systematic review and meta-analysis. Endosc. Int. Open 3 (3), E226–E235. https://doi.org/10.1055/s-0034-1391708.
