Wireless ingestible luminal diagnostic capsule that autonomously treats a condition

Abstract

An apparatus for detecting and treating gastrointestinal pathology. The apparatus includes a wireless capsule, where the wireless capsule includes at least one of: an optical imaging system, a thermal imaging system, or a camera. The camera is a narrow band imaging system, a white light imaging system, or both.

Description

MGH 2024-124-03 Quarles 125141.04710     WIRELESS INGESTIBLE LUMINAL DIAGNOSTIC CAPSULE THAT AUTONOMOUSLY TREATS A CONDITION CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] The present application is based on and claims priority from U.S. Patent Application Ser. No. 63 / 625,178, filed on January 25, 2024, and U.S. Patent Application Ser. No. 63 / 570,723, filed on March 27, 2024, the entire disclosures of which are incorporated herein by reference. STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH

[0002] N / A BACKGROUND

[0003] Due to inequities in health care and expensive, inefficient, and resource- intensive medical procedures, diseases are usually detected and treated late when morbidity, mortality, and costs are high. This problem is profound in the GI tract where the current standard of care for diagnosis and treatment is endoscopy, a complex and expensive procedure. Indeed, even though the US health care system covers over $50B / year in endoscopy expenses, ~100,000 people in the US will die each year from preventable GI tract diseases. It would helpful if there were a new approach to detecting, diagnosing, and treating GI tract diseases. Especially helpful would be means to decrease costs by eliminating expensive equipment and resources, physicians, and pathologic tissue analysis. If the new approach could be deployed non-invasively in the primary care office or at home, it will be easier for people to get preventative care to eradicate early-stage disease, avoiding the increased morbidity and mortality of advanced disease. SUMMARY

[0004] Disclosed herein is an autonomous, swallowable capsule that images the entire gastrointestinal (GI) tract and identifies and treats disease in real time, without additional human intervention. This all-in-one functionality will be realized in a wireless capsule, a device with the acronym WILDCAT – Wireless Ingestible Luminal Diagnostic Capsule that Autonomously Treats. The wireless capsule will incorporate advanced microscopic / functional imaging, Artificial Intelligence (AI)-based diagnosis, and laser treatment in a single, battery- powered pill. This pill can be comprised of all or any subset of its technologies to serve as a 1  MGH 2024-124-03 Quarles 125141.04710     colonoscopy surrogate. By enabling GI tract diseases to be identified and treated non- invasively at home, this autonomous capsule will democratize preventative health care and will eliminate the need for complex GI endoscopy procedures, decreasing morbidity, mortality, and costs.

[0005] In a first aspect, provided herein is an apparatus for detecting and treating gastrointestinal pathology. The apparatus includes a wireless capsule, where the wireless capsule includes at least two of: an optical imaging system, a thermal imaging system, or a camera. The camera is a narrow band imaging system, a white light imaging system, or both.

[0006] In another aspect, provided herein is a method for detecting and treating gastrointestinal pathology. The method includes contacting a sample with a wireless capsule. The wireless capsule comprising at least two of: an optical imaging system, a thermal imaging system, or a camera. The camera is a narrow band imaging system, a white light imaging system, or both. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Some embodiments of the disclosure are described herein with reference to the accompanying figures. The description, and figures make apparent to a person having ordinary skill in the art how some embodiments of the disclosure may be practiced. The figures are for the purpose of illustrative discussion and no attempt is made to show structural details of an embodiment in more detail than is necessary for a fundamental understanding of the teachings of the disclosure. Any figures herein are not shown to scale. Where dimensions are given in the text or figures, these dimensions are merely exemplary and do not limit the scope or spirit of the disclosed invention.

[0008] FIG. 1 shows a schematic of an autonomous, wireless capsule device (e.g., WILDCAT) according to an aspect of the disclosure herein, shown configured with OCT, printed circuit board and optoelectronics components. OCT - optical coherence tomography; PCB - printed circuit board; PIC - photonic integrated circuit; RX Tx – receive / transmit radio frequency (RF) transmitter.

[0009] FIG. 2A shows the conceptual components of the wireless capsule, according to an aspect of the disclosure herein.

[0010] FIG.2B shows some of the components for miniaturizing the wireless capsule such that it can be a standalone capsule. 2  MGH 2024-124-03 Quarles 125141.04710

[0011] FIG.3 shows an overview of the Application Specific Integrated Circuit (ASIC) and the four major core blocks that comprise it, according to an aspect of the disclosure herein. Imaging / Therapy I / O interfaces with all imaging and delivers the data to the CPU which then tasks the Neural Processing Unit (NPU) with diagnosis. Upon decision from NPU, the CPU can decide therapy and send images wirelessly using RF Communications.

[0012] FIG.4 shows an example of a tethered motor capsule that requires an operator.

[0013] FIG.5, panel A shows a closeup photograph of the wireless capsule with a 16- element thermistor array (arrow) and a penny for scale, according to an aspect of the disclosure herein.

[0014] FIG.5, panel B shows antennae of the wireless capsule (dark arrow shows one of the antennae) and receiver (light arrow), according to an aspect of the disclosure herein.

[0015] FIG. 5 panel C shows a photograph of a person wearing a receiver belt, according to an aspect of the disclosure herein.

[0016] FIG. 5 panel D shows an example of a thermal readout as a function of time showing two hot spots (arrows), according to an aspect of the disclosure herein.

[0017] FIG.6, panel A shows a photo of the miniature electronics for the wireless OCT system next to a ruler for scale.

[0018] FIG.6, panel B shows an exemplary OCT image of a finger obtained using the compact VSCEL swept source, showing the stratum corneum (sc), epidermis (e), and dermis (d). OCT: A-line / s - 1 kHz, SNR -105 dB.

[0019] FIG. 7, panel A shows an OCT depth projection of normal mucosa (lower arrow) adjacent to a colorectal adenoma (upper arrow), acquired ex vivo.

[0020] FIG. 7, panel B shows an ROC curve for an example OCT ML algorithm for prospectively discriminating malignant potential (MP, adenoma & serrated polyp) from no malignant potential (NMP, hyperplastic polyp & normal mucosa) lesions, ex vivo.44 histology positive and 40 negative OCT images were used to train the ML algorithm and 48 histology positive and 30 negative OCT images from different lesions were used to test and generate the ROC curve. Notably, this preliminary algorithm is sufficiently accurate to satisfy ASGE's PIVI thresholds.

[0021] FIG. 8, panel A shows a photo of example miniature electronics for laser thermal absorption (LTA) ablation at 1940 nm. 3  MGH 2024-124-03 Quarles 125141.04710

[0022] FIG.8, panel B shows a histologic viability slide (blue stain is viable) of fresh swine colon irradiated ex vivo with these components (5 mm beam size, 1 W, 5 sec).

[0023] FIG.8, panel C shows an NTBC slide of fresh swine colon irradiated ex vivo at 660 nm after exogenous absorber therapy (EAT) using topical application of methylene blue (4 mm beam size, 600 mW, 10 sec, 10 µM).

[0024] FIG. 9, panel A shows a photograph of a capsule containing acetic acid and sodium bicarbonate inside a 2.5 cm diameter water-filled tube.

[0025] FIG.9, panel B shows the mechanical actuation allowing the two chemicals to mix, creating a CO2bubble that displaced the surrounding water.

[0026] FIG. 10, panel A shows OCT R-TCE data obtained from the human colon in vivo. A. shows a cross- sectional OCT frame showing clear visualization of the circumference of the entire colonic wall and mucosal architectural morphology, even beneath several mm of interposed fecal matter.

[0027] FIG. 10, panel B shows a 3D flythrough view of the rectum and lower colon demonstrating a clear depiction of the mucosal surface and colonic folds.

[0028] FIG.10, panel C shows a closeup photograph of the R-TCE capsule.

[0029] FIG.11 shows a schematic of the hollow shaft motor and 3D-nanoprinted, dual deflecting optic, enabling compact OCT beam scanning and a large, collimated beam for laser therapy targeting, according to an aspect of the disclosure herein.

[0030] FIG. 12, panel A shows a schematic of the OCT PIC interferometer design, according to reference 4.

[0031] FIG. 12, panel B shows a printed micro-optic interfacing the VSCEL with the PIC.

[0032] FIG.12, panel C shows the sample arm beam is coupled out of the PIC using a printed collimator stacked with a multiplexed phase plate to achieve OCT depth of focus extension.

[0033] FIG. 12, panel D shows a schematic diagram depicting a general OCT configuration, laser wavelength sweep, processing interferograms to A-scans using Fourier transform methods, and how OCT images are obtained.

[0034] FIG. 13 shows a schematic of the trimodal imaging system and endoscopic probe, in accordance with an embodiment of the invention. 4  MGH 2024-124-03 Quarles 125141.04710

[0035] FIG. 14A shows a schematic diagram with the capabilities of the wireless capsule, according to an aspect of the disclosure herein.

[0036] FIG. 14B shows a high-level flow chart of the use of the wireless capsule, according to an aspect of the disclosure herein.

[0037] FIG. 14C shows an imaging and therapy I / O Control block, also referenced as "Analog Front End", functional block diagram, according to an aspect of the disclosure herein. The CPU provides key input control signals for a variety of mixed mode components. Analog input signals are digitized by a 1-100MSPS (Mega-sample per second) 16 bit ADC (Analog to Digital Converter). Abbreviations (ADC = Analog to Digital Converter, DAC = Digital to Analog Converter, MSPS = Mega-samples per second, MUX = multiplexer, TIA = Trans- Impedance Amplifier, I / O = Input Output, CPU = central processing unit, OCT = Optical Coherence Tomography, PIC = Photonics Integrated Circuit).

[0038] FIG.15 shows an example of a dip coating process where a capsule is lowered into resin at 60mm / min, dwelled for 10 seconds, then raised at 60mm / min, before curing and post curing in an oven.

[0039] FIG. 16 shows exemplary dimensions of a wireless thermal capsule and a depiction of capsule wall layers, according to an aspect of the disclosure herein. The sensor is attached to a polycarbonate wall by thermal adhesive tape or paste and then coated in an outer biomedical coating to add resistance to bodily pH changes.

[0040] FIG.17 shows an expandable polymer mesh in wireless capsule, from reference 3.

[0041] FIG. 18 shows a schematic of one example of how images are generated. Position data, thermal readings and the estimated luminal thicknesses are input and used to create final spatial and thermally calibrated images accounting for the impacts of capsule velocity and fecal matter.

[0042] FIG.19 shows a simulation of the thermal diffusion in the GI tract, specifically a (panel (a)) deep tissue inflammation site inside the large intestine propagating heat through a layer of fecal matter, a capsule wall made of biomedical polymer to a final sensor inside an air- filled capsule. The time constant Tau (time to 63.2% of steady state value) was used as a metric to compare the impacts of parametric sweeps of (panel (b)) fecal matter thickness, (panel (c)) water concentration using values from references 27 and 28, (panel (d)) capsule wall thickness, 5  MGH 2024-124-03 Quarles 125141.04710     (panel (e)) inflammation depth, and then in (panel (f)) we show fecal matter thickness vs temperature for different time frames in simulation. Note in (panel (b)) the fecal matter thickness greatly impacts the response time to rise to final temperature. This is calculated with velocity to determine end temperature.

[0043] FIG.20 shows an example of a therapy R&D workflow, in accordance with an embodiment of the disclosure. *Method depends on whether LAT or EAT is being investigated. NTBC - nitroblue tetrazolium chloride histopathologic tissue viability stain. DETAILED DESCRIPTION

[0044] Disclosed herein is an autonomous, swallowable capsule that images the entire gastrointestinal (GI) tract and identifies and treats disease (e.g., colorectal cancer precursors called adenomas) in real time, without additional human intervention. The device can be customized to autonomously identify and treat a variety of pre-cancers in the GI tract (e.g., Barrett’s esophagus precursor to esophageal adenocarcinoma), inflammatory conditions (e.g., inflammatory bowel disease), and ulcers / bleeds.

[0045] This all-in-one functionality can be realized in an autonomous or wireless capsule by creative customization and miniaturization of complex hardware incorporating advanced microscopic / functional imaging, AI-based diagnosis, and laser treatment in a single, battery- powered pill. Components of this wireless capsule include conforming batteries, microstructural and physiological imaging systems, optoelectronic (OE) and machine learning (ML) integrated circuits (IC), and therapeutic lasers. Prototypes of each individual subcomponent have been made and demonstrated on benchtop. By enabling GI tract diseases to be identified and treated non-invasively at home, the wireless capsule can democratize preventative health care and can eliminate the need for complex GI endoscopy procedures, decreasing morbidity, mortality, and costs.

[0046] A high-level schematic of the wireless capsule is shown in FIG.1. The wireless capsule can include a battery, printed circuit boards, an optical imaging system such as an optical coherence tomography (OCT) imaging system including an OCT laser and an OCT detector, an ablation laser, an artificial intelligence microchip, a supercapacitor, a mirror, a thermistor array, a micromotor, stabilization and clearance mechanism, (FIG. 1). In this example, the wireless capsule has an elongated shape with curved ends. It is contemplated that 6  MGH 2024-124-03 Quarles 125141.04710     the wireless capsule can have any shape and size that is conducive to swallowing and passage through the gastrointestinal tract. Creating the wireless capsule is now possible because of recent advancements in battery and motor technology, application specific integrated circuits (ASIC), OE devices, miniature imaging systems and lasers, nanotechnology, and embedded AI (FIG.2A, 2B). Table 1 shows the components of the wireless capsule.

[0047] Table 1. Wireless capsule technology Technology Task(s) Battery Higher capacity battery that conforms to the capsule's shapee a c a e ge ce c oc p ca e c u e a pp ca o pec fic Integrated Circuit (ASIC) which allows the combination of all components necessary for capsule functionality onto a single chip for faster communication and smaller packaging. The ASIC can have with 4 core components as shown in FIG.3: 1) A core microcontroller featuring a CPU, RAM and ROM, 2) A mixed-mode (both digital and analog sub-components) peripheral interface for thermal, OCT and white light imaging and for delivering therapy such as laser ablation, 3) A Neural Processing Unit (NPU) to load a machine learning algorithm to assess images and perform screening and diagnosis to inform therapy and 4) an RF Communication block for transmitting images in a research setting. The ASIC can operate on battery power and enables a full process of imaging, screening, diagnosis and treatment by the wireless capsule.

[0049] OCT is a depth-resolved microscopic imaging technique that can be utilized to increase diagnostic specificity for hot-spots identified by thermal imaging. The optical imaging system can be a self-contained OCT imaging system for obtaining high-resolution images. The OCT imaging system can scan around the circumference of the capsule by a mirror on the shaft of the micromotor, taking cross-sectional images of the entire GI tract as the device traverses the organ. Example images are shown in FIG. 4, obtained using a user-operated, tethered 7  MGH 2024-124-03 Quarles 125141.04710     capsule. As disclosed herein, the wireless capsule is autonomous and does not include a tether to any external device. As shown in FIG.5, the wireless capsule can include antennae.

[0050] The wireless capsule can include a self-contained OCT imaging system. Some of the miniaturized components are shown in FIG.6A. The OCT imaging system can include a wavelength-swept laser source. In some examples, the wavelength-swept laser source can be a wavelength-swept vertical cavity surface emitting laser (VCSEL). The wavelength-swept laser source can provide light having a wavelength ranging between 400 and 1900 nm. In certain examples, the OCT imaging system can include a 1310 nm wavelength-swept VCSEL. The sweep rate of the wavelength-swept laser source can be from 0.1 to 150 kHz. An example of an OCT image obtained using the compact VSCEL swept source is shown in FIG.6B.

[0051] The OCT imaging system can include the mirror as a rotating reflector. The rotating reflector can be attached to and rotated by the micromotor. The micromotor can position the rotating reflector to reflect the laser beam from the OCT laser or the ablation laser. The rotating reflector can direct light from the laser source in a circumferential pattern around an outer perimeter of the wireless capsule to a sample. Additionally, and alternatively, the rotating reflector can direct light reflected from the sample to the OCT detector. The OCT detector can be a photodetector.

[0052] The micromotor can be a stepper motor. In some examples, the micromotor can have between 20 and 60 steps per revolution. In other examples, the micromotor can stop every 5°, 6°, 7°, 8°, 9°, or 10°. For power considerations, the micromotor can operate at 0.5-5 V or less. For size considerations, the micromotor can have a body volume of less than 200 mm3.

[0053] The optical imaging system can include optical components to direct light between the rotating reflector and at least one of the wavelength-swept laser source, the OCT detector, or the ablation laser. The optical components can include nano-printed optics, such as nano-printed light-delivering optics.

[0054] Thermal imaging, which sensitively measures increased metabolic activity that is present in neoplasia and inflammation, may be implemented using a thermistor array distributed around the capsule’s circumference. Detection of local temperature may create two- dimensional maps as the capsule traverses the GI tract. The wireless capsule can include a thermal imaging system. The thermal imaging system can include a thermistor array disposed in a circumferential pattern around the outer perimeter of the wireless capsule. The thermistor 8  MGH 2024-124-03 Quarles 125141.04710     array is configured to obtain temperature information from the sample surrounding the wireless capsule. The temperature information can be used to generate two-dimensional maps of the sample to assist in detecting diseased tissue. An example of thermal imaging is shown in FIG. 5D. An example of the image generation process using thermal data and position data is shown in FIG.18 and example simulation data is shown in FIG.19.

[0055] The thermistor array can include between 1 and 50 elements. For example, the thermistor array can include 16 elements, as shown in the capsule depicted in FIG.5A. Example specifications for the thermistor array are shown in Table. 10. The thermistor array can be arranged as multiple spaced rings of thermistor elements. For example, a first ring of thermistor elements can be arranged at one end of the capsule and a second ring of thermistor elements can be arranged near the opposite end of the capsule. These thermistor rings can be used as markers to provide measurements along the length of the capsule that can also be used to compute velocity, i.e., a speed at which the capsule is moving through a sample.

[0056] The wireless capsule can include a light imaging system or a camera. In some examples, the camera can be a Narrow Band Imaging (NBI) system. In other examples, camera can be an LED illuminated white light imaging system. The camera can be configured for other types of image-enhanced endoscopy such as flexible spectral imaging color enhancement (FICE), blue laser imaging (BLI), and autofluorescence imaging (AFI). The camera can be positioned at one end of the wireless capsule.

[0057] The wireless capsule can include miniature electronics for a laser capable of laser ablation therapy as shown in FIG.8A. The ablation laser can be used to provide heat to a specified sample area to damage or destroy abnormal cells. The rotating reflector can be a dual reflector that can guide the OCT beam to redirect and collimate a therapy laser beam such as from the ablation laser. The ablation laser can be driven by a supercapacitor within the wireless capsule. In some examples, the ablation laser can be a therapy laser to be used for laser absorption therapy (LAT) where water is the target of the laser light. In this example, the ablation laser can have a wavelength of 1940 nm. Example data is shown in FIG.8B. In other examples, the therapy laser can include a 660 nm laser diode to be used for exogenous absorber therapy (EAT), where a safe dye (e.g., methylene blue, 660 nm) is topically applied. The ablation laser is separate from the OCT laser.

[0058] The supercapacitor can store the required energy for the ablation laser. The 9  MGH 2024-124-03 Quarles 125141.04710     capacitance of the supercapacitor can be between 1 F to 20 F. In use, the ablation laser draws from the supercapacitor which discharges at a high current for multiple seconds. The supercapacitor can recharge from the battery between applications of the ablation laser. Table 3 shows an exemplary power and energy analysis including the capacitance required to drive each laser diode to fully treat a colonic mucosa over a -5 mm area, established by preliminary testing ex vivo (FIG.8B).

[0059] The wireless capsule can include a controller. The controller can be in communication with the laser source, the OCT detector, the ablation laser, and the thermistor array. The controller is configured to obtain information from the OCT detector to generate optical image data. The controller can additionally, or alternatively, obtain information from the thermistor array to generate thermal image data.

[0060] Digitized OCT and thermal images can be sent to an integrated AI microchip that can use an ML algorithm to diagnose the images in real time. The controller can use a machine learning algorithm to analyze at least one of the white light image data, narrow band image data, optical image data and the thermal image data to diagnose a condition in a portion of the sample. For example, machine learning algorithm can use the image data to classify the sample as normal or polyp. In some examples, the machine learning algorithm can use the image data to classify the sample as polyp without malignant potential or as polyp with malignant potential. Classification of the sample combined with additional information from the image data (e.g., lesion size or appearance) can be used to make a determination as to whether to treat, leave, or flag a polyp for colonoscopic follow-up. The machine learning algorithm can be used with or without the ablation laser.

[0061] The machine learning algorithm can be capable of real-time analysis of the image data such that the determination to treat can be made while the wireless capsule is in the vicinity of the sample in need of treatment. When disease such as a colonic adenoma is identified, the controller can control the micromotor to direct the ablation laser at the portion of the sample to ablate tissue associated with the diagnosed condition. In some examples, the controller can be in communication with the micromotor to rotate the rotating reflector.

[0062] The wireless capsule can include any combination of one or more of the optical imaging system, the thermal imaging system, and the camera. A power source can be included to provide power to at least one of the optical imaging system, the thermal imaging system, 10  MGH 2024-124-03 Quarles 125141.04710     and the ablation laser. The power source can be a battery or a supercapacitor. In some examples, each of the OCT laser and the ablation laser has an independent dedicated power source. In other examples, each of the optical imaging system, thermal imaging system and the camera has an independent dedicated power source.

[0063] The wireless capsule can include a temperature sensor. The temperature sensor can be in communication with the controller. In some examples, the controller can be configured to sense a change in temperature adjacent to the wireless capsule such as an increase in temperature or a decrease in temperature. In some examples, the controller can place the controller into a shutdown state based on sensing an increase in temperature above a predetermined level.

[0064] A pH sensor can be included in the wireless capsule, for example in the outer shell of the capsule. The pH sensor can include a chip having a surface area of between 3 mm2and 5 mm2and a reference electrode with a 2 mm to 4 mm diameter. The sensitivity of the pH sensor can be at least + / -0.5 pH units and the pH sensor can have a total range of at least pH 0.5-9. The power draw of the pH sensor can be less than 100 µW or less than 50 µW.

[0065] The controller can be in communication with the pH sensor and can be configured to sense a change in pH in the sample region adjacent to the wireless capsule. The controller can determine the approximate location of the capsule based on information received from the pH sensor. Additionally, the controller can activate at least one of the OCT imaging system, the thermal imaging system, or the ablation laser based on sensing the change in pH.

[0066] The wireless capsule can include a stabilization and clearance mechanism. In one example, the mechanism can include one or more reactant wells configured to release reactants to the sample adjacent the wireless capsule. The reactant wells can hold dyes, drugs, or other chemical components. In some examples, the reactant wells can hold chemical components that, when released, combine to generate a gas adjacent to the wireless capsule. For example, one reactant well can hold NaHCO3and a second reactant well can hold an acid, such as C6H8O7 which, when released, combine to form CO2 gas outside the wireless capsule, as shown in FIG.9.

[0067] In an example of a stabilization mechanism, the reactants can combine outside the wireless capsule to form a polymer. The polymer can be a pH sensitive polymer or an absorbent polymer. In particular, the polymer can expand and create friction to stabilize the 11  MGH 2024-124-03 Quarles 125141.04710     wireless capsule.

[0068] The wireless capsule can include at least one coating on an outside portion of the wireless capsule. The coating can be a thermally-conductive coating. Additionally, and alternatively, the coating can be a biocompatible coating as shown in FIG. 16. The biocompatible coating can be resistant to acid, such as stomach acid, or conditions having pH above 0.5 and less than 7. The biocompatible coating can be compatible with high temperatures, such as temperatures between 30 °C and 150 °C. The biocompatible coating can be resistant to repeated sterilization cycles. For example, the biocompatible coating can withstand exposure to steam and / or hydrogen peroxide.

[0069] When forming, the biocompatible coating can be fast curing. The curing process can include elevated temperatures between 30 °C and 150 °C. In some examples, the biocompatible coating can be formed from a two-part epoxy system. In some instances, the biocompatible coating is applied by dip-coating as shown in FIG. 15 and curing to form a biocompatible outer shell on the wireless capsule. In other examples, the biocompatible coating can be applied by painting, wrapping, spraying, or polymerizing.

[0070] The wireless capsule can be used for detecting and treating gastrointestinal pathology. FIG. 14A shows some of the capabilities of the wireless capsule and FIG. 14B shows an example of a wireless capsule work flow. According to an exemplary process, the wireless capsule can be swallowed and the pH sensor is activated, which helps determine the location of the capsule in the GI tract. Once the wireless capsule enters the colon, optical and / or thermal imaging begins using the self-contained OCT imaging system in the capsule or the thermal imaging system. In some examples, stabilization and / or clearance methods can be deployed as needed. As the wireless capsule moves through the GI tract, images can be processed and categorized in real time by the machine learning algorithm using the ASIC. If categorized as diseased, the identified region of the GI tract can be laser treated. The capsule can include a temperature sensor. When temperature decreases because the wireless capsule has exited the GI tract, the imaging processes can cease, and the diagnostic information can be retrieved from the capsule.

[0071] According to an aspect of the disclosure herein, the wireless capsule, having the optical imaging system and the thermal imaging system can be contacted with a sample. In some examples, the wireless capsule is swallowed to enter a gastrointestinal tract in order to 12  MGH 2024-124-03 Quarles 125141.04710     contact the sample. The optical imaging system can employ the OCT imaging system to obtain OCT image data as follows. The controller can control the rotating reflector to direct light from the wavelength-swept laser source in a circumferential pattern around the outer perimeter of the wireless capsule to the sample. The rotating reflector can direct light reflected from the sample to the OCT detector to acquire the optical image data. FIG.14C shows an imaging and therapy I / O Control block functional block diagram.

[0072] The controller can be used to obtain information from the OCT detector to generate optical image data. The controller can be used to obtain information from the thermistor array to generate thermal image data. The controller can use a machine learning algorithm to analyze at least one of the optical image data, the thermal image data, the narrow band imaging data, or the white light data to diagnose a condition in a portion of the sample. Additionally, and alternatively, the controller can be used to obtain information from the camera, such as white light images or narrow band images. Images obtained from the camera can be used to diagnose a condition in the sample.

[0073] In some examples, the wireless capsule includes optical components disposed adjacent to the wavelength-swept laser source, the OCT detector, and the ablation laser. The optical components allow the light can be directed between the rotating reflector and at least one of the wavelength-swept laser source, the OCT detector, or the ablation laser. The optical components can include nano-printed optics, lenses, mirrors, and diffractive elements.

[0074] The method can include using the rotating reflector to direct light from the ablation laser toward the sample. For example, the controller can control the micromotor to direct the ablation laser at the portion of the sample to ablate tissue associated with the diagnosed condition. In other words, the ablation laser can be used to apply heat to the sample in an area of the sample identified by the image data.

[0075] The controller can sense a change in pH adjacent to the wireless capsule based on the pH sensor. If a change in pH is detected, the controller can determine the approximate location of the capsule. The controller can activate at least one of the camera, OCT imaging system or the thermal imaging system to acquire images at the location. Additionally, or alternatively, the controller can activate the ablation laser to treat tissue in the sample at the location. In some examples, such as > 5 mm colonic adenomas that should be removed by endoscopy (10-20% of cases), the capsule will wirelessly transmit data that will inform the 13  MGH 2024-124-03 Quarles 125141.04710     patient that they should follow up with a medical specialist rather than using the ablation laser of the wireless capsule to treat the adenoma.

[0076] The wireless capsule can generate a gas bubble by the mixing of reactants stored in the reactant wells. For example, as the wireless capsule traverses the GI tract, there may be a region of the sample covered with contaminants such as feces, making imaging and / or laser ablation of the sample difficult. In some examples, the imaging data can be used to recognize contaminants. In this situation, the controller can release, using the reactant wells, reactants to generate a gas adjacent to the wireless capsule. This electromechanically actuated reactant mixing can enable CO2gas bubble generation around the capsule to displace intraluminal contaminants for imaging and / or ablation as needed. Additionally, and alternatively, a balloon can be electromechanically inflated around at least a portion of the capsule.

[0077] The controller can sense an increase in temperature adjacent to the wireless capsule based on data obtained from the temperature sensor. The temperature sensor can be used to detect areas of increased metabolic activity or perfusion such as in inflammation, polyps, or cancers in the GI tract.

[0078] In some examples, the controller can place the wireless capsule into a shutdown state based on sensing the increase in temperature above a predetermined level. For example, when the coating is applied to the wireless capsule, curing the coating may require elevated temperatures, and it is beneficial for the wireless capsule to be in a low power state or to be completely powered down.

[0079] The wireless capsule WILDCAT is expected to revolutionize how and when GI tract diseases are detected, diagnosed, and treated. There will be no need to get a referral or suffer endoscopy scheduling delays, as this device can be taken at the subject’s convenience. WILDCAT will be more accurate and objective than endoscopy because it will also obtain functional thermal and three- dimensional (3D) microscopic information from the GI tract and will rely on ML for interpretation of these images. The capsule will decrease costs by eliminating expensive equipment and resources, physicians, and pathologic tissue analysis. Because the capsule can be taken non-invasively in the primary care office or at home, it will be easier for people to get preventative care to eradicate early-stage disease, avoiding the increased morbidity and mortality of advanced disease. These benefits of the capsule will allow superior GI care to be provided to all, including underrepresented individuals who do not have 14  MGH 2024-124-03 Quarles 125141.04710     access to expensive surgical resources and expert physicians.

[0080] Wireless OCT has never been demonstrated in a swallowable pill sized before. AI identification, diagnosis and decision to treat on-board a capsule endoscope based on OCT capsule endoscopy and the combination of OCT capsule endoscopy and thermal imaging capsule endoscopy is novel. Similarly, the combination of a wireless insufflation and laser ablation in a capsule also is novel to the best of our understanding.

[0081] Polymer and mechanical-based stabilization methods of the capsule, phototherapy methods such as laser absorption therapy (LAT) or exogenous absorber therapy (EAT) using a topically applied safe dye (e.g., methylene blue, 660 nm), have not to our knowledge been previously disclosed in a wireless capsule. Nor has en face ML processing of OCT been able to arrive at diagnosis of colon polyp lesions.

[0082] Using off-the-shelf components, individual versions of each of these elements has been demonstrated and it has been shown that they can be miniaturized to fit within a swallowable capsule (12 mm diameter x 30 mm length) and can run off a capsule-sized battery for a typical bowel transit time of 8 hours. This includes the development of 1) an IRB- approved wireless thermal imaging capsule, 2) a miniature, battery-powered OCT system, 3) an ML algorithm that uses OCT and thermal images to diagnose colorectal adenomas with an overall accuracy >95% using a battery-powered AI microchip, 4) a miniature, battery-powered 1950 nm laser with associated supercapacitor-based electronics that can ablate full-thickness colonic mucosa in 1-second, and 5) peri-capsular bubble generation through mechanically- actuated reactant mixing.

[0083] Each of these technologies and functionalities can be combined into a single capsule so that they work together to visualize, diagnose, and treat GI tract disease autonomously and in real time. To achieve this objective, custom components (ASICs, micromotors, optoelectronics, supercapacitors, micro-optics, etc.) can be used that are smaller, more capable, and utilize less battery power (Table 1). Miniaturization can be achieved by integrating related / interacting optics and electronics on OE chips.

[0084] EXAMPLES

[0085] The following are non-limiting examples according to embodiments of the disclosure.

[0086] Example 1: 15  MGH 2024-124-03 Quarles 125141.04710

[0087] Preliminary Data. Using off-the-shelf components, we have developed individual versions of many of the elements required for WILDCAT and have shown that they can be miniaturized to fit within a swallowable capsule (12.8 mm diameter x 32.8 mm length) and can run off a capsule-sized battery for a typical bowel transit time of 8 hours. This preliminary data includes the development of 1) an IRB-approved wireless thermal imaging capsule (FIG. 5), 2) a miniature, battery-powered OCT system (FIG. 6), 3) an ML algorithm that uses OCT images to diagnose colorectal adenomas with an overall accuracy >90% (FIG. 7), 4) a miniature, battery-powered 1940 nm laser with associated supercapacitor- based electronics that can ablate full-thickness colonic mucosa 5- seconds (FIG. 8A, B), mucosal ablation with a 660 nm laser irradiating methylene blue (MB) stained tissue in 10 seconds (FIG. 8C), and 5) peri-capsular bubble generation through mechanically actuated reactant mixing (FIG.9). We also have conducted colonic OCT R-TCE imaging in human subjects (FIG. 10), which will be adapted in this program to test WILDCAT's clearance technologies. Our main technical challenge is to combine all of WILDCAT's subcomponents into a single capsule so that they work together to visualize, diagnose, and treat GI tract disease autonomously and in real time. Regarding capsule size, targeting WILDCAT dimensions of 12.8 mm in diameter x 32.8 mm long, which is comparable in dimensions to that of Medtronic's PillCam™ COLON 2 Capsule, and assuming a 0.2 mm capsule wall thickness, the internal working volume should be 3665 mm3. Using these capsule dimensions, we simulated a mechanical design that places all of the proposed WILDCAT components inside the capsule in a manner that would retain theirfunctionalities. Results show that 130 mm3of free space would still be available inside thecapsule, suggesting that WILDCAT is feasible from a dimensional standpoint.

[0088] Clinical studies using single- functionality, hard-wired capsules attached to colonoscopes via capsule delivery devices will first be conducted to determine the accuracy of the imaging modalities for diagnosing adenomas and the capacity of the ablation technology to eradicate adenomas in humans in vivo. Primary endpoints will be benchtop verification that all components and devices meet required design input specifications and clinical imaging diagnostic accuracy >90% and clinical laser ablation adenoma eradication with >90% efficacy, for example in swine. Once the integrated autonomous capsule is ready for human use, we will conduct pilot human studies to demonstrate safety and feasibility. We will then perform a clinical validation study showing that the autonomous capsule performs at least as well as 16  MGH 2024-124-03 Quarles 125141.04710     colonoscopy for diagnosis and treatment. The final deliverable will be an autonomous capsule that surpasses 90% diagnostic accuracy and adenoma eradication rate thresholds in living human patients.

[0089] Considerations

[0090] One consideration is battery capacity; we believe that moderate improvements in battery efficiency and electronic power consumption are realizable and sufficient to concurrently power all electronics in the capsule. If necessary, periodic battery recharging can be accomplished using magneto-elastic power harvesting.

[0091] Capsule treatment contraindication. In circumstances where the capsule should not be used to treat a lesion, such as > 5 mm colonic adenomas that should be removed by endoscopy (10-20% of cases), the capsule will wirelessly transmit data that will inform the patient that they should follow up with a medical specialist.

[0092] Another consideration is the mean transit time of a diagnostic capsule in the GI tract is -0.5 mm / s and -0.2 mm / s in the colon. Since the autonomous capsule will make diagnoses in real time and laser ablation will take -1 second, we do not anticipate that the capsule will need to be stopped in order to completely treat adenomas ≤ 5 mm or other diseases.

[0093] Example 2:

[0094] Technical approach. Creating this autonomous capsule is now possible because of recent advancements in battery and motor technology, application specific integrated circuits (ASIC), photonic integrated circuits (PIC), miniature imaging systems, sensors, lasers, nano- printed optics, and embedded AI. FIG. 1 is a high-level schematic of WILDCAT that incorporates these elements and FIG. 14B is a flow chart that depicts how the device will operate. First, the capsule will be taken out of the magnetic packaging that keeps the device in an off state. After activation, WILDCAT will be swallowed and will naturally travel down the GI tract. A pH sensor, integrated in the capsule's shell, will provide data that will indicate when the capsule enters the colon, triggering imaging, stabilization, and AI functionalities. Two LED- illuminated white light (WL) cameras will be included in the capsule's endcaps, allowing the device to capture forward-viewing images regardless of its orientation.5Thermal imaging, which measures increased metabolic activity that is present in neoplasia and inflammation, will be implemented using a thermistor array distributed around the capsule's circumference. Detection of local temperature will create two-dimensional "heat" maps as the capsule traverses 17  MGH 2024-124-03 Quarles 125141.04710     the GI tract. Optical coherence tomography (OCT), a depth-resolved microscopic imaging technique, will increase diagnostic specificity for lesions seen via WL and thermal imaging. OCT will be implemented using a wavelength-swept vertical cavity surface emitting laser (VCSEL) that illuminates a PIC; OCT light will be scanned around the circumference of the capsule by a mirrored prism on a hollow-core micro-motor, taking cross-sectional images as the device traverses the organ. Digitized video, OCT and thermal images will be sent to an integrated AI microchip that will use an ML algorithm to diagnose the images in real time. When disease such as a diminutive (< 5 mm) colonic adenoma is identified, the motor will direct high power laser light to ablate the lesion. The wavelength of this light will either be tuned to be absorbed by water (1900-2000 nm), termed laser absorption therapy (LAT) or by a topically applied safe dye (e.g., methylene blue, 660 nm), called exogenous absorber therapy (EAT). Electromechanically actuated balloons or other stabilization mechanisms may control the capsule's velocity and / or clear intraluminal contents if needed. Wireless transmission, activated after the capsule is excreted, will inform the subject and / or primary care provider whether an adenoma was identified, the number and sizes of adenomas, and whether an adenoma was treated.

[0095] Electronics. Electrical components including the battery, ASIC, incorporating a mixed- mode sensor interface, microcontroller (CPU, RAM, ROM), Neural Processing Unit (NPU), and wireless transmission (modulation, PA, matching network), supercapacitor, electronics on a rigid- flex PCB, micro-motor, laser and thermal management electronics, and pH sensor, are critical elements that will need to be customized in order to realize a functional WILDCAT device. During Phase I, these components will be designed and prototyped to meet the initial design input specifications for WILDCAT. In Phase II, these components will be tested to operate in concert with all other WILDCAT elements.

[0096] OCT PIC. A single chip that contains all OCT optics and light detection electronics will be developed. The OCT laser will also be customized and integrated with the PIC using 3D nano-printed coupling optics.

[0097] Imaging. To determine the best complement of imaging technologies to incorporate in the capsule, we will develop a trimodal probe that merges WL imaging (WLI), OCT, and thermal imaging and can be inserted into the accessory port of a colonoscope. After the probe is ready for human use, we will conduct a study in patients undergoing colonoscopy 18  MGH 2024-124-03 Quarles 125141.04710     to obtain trimodal images of all colonic polyps and representative normal regions. Imaged sites will be biopsied and processed for histology. Images will then be input into a multivariate model for predicting the histologic diagnosis, including non-malignant vs. malignant potential colonic tissue classification.

[0098] Clearance. We will conduct research and development of capsule-based methods for clearing intraluminal contents from the capsule’s surroundings. This technology will be incorporated into WILDCAT if it is needed to effectively image or conduct therapy. Methods contemplated include creating a CO2 bubble, inflating a balloon around the capsule, or deploying expandable polymer that absorbs bowel contents. Clearance technology will be tested clinically by incorporating it into an OCT retrograde tethered capsule endomicroscopy (R-TCE) device that we have previously demonstrated in patients for comprehensive colonic OCT imaging.

[0099] Stabilization. To completely eradicate a diminutive adenoma, the capsule's colonic dwell time over the lesion should be no shorter than the therapy duration. As little is known about the instantaneous velocity of capsules in the colon, we will utilize our existing wireless capsule with an integrated pH sensor and antenna-based capsule position tracking platform to study capsule colonic motion in ambulatory humans. In parallel, we will develop methods including integrated inflatable balloons and deployable sponges that can lower the capsule's maximum velocities. If the need is identified from the velocity clinical study, we will select the appropriate capsule stabilization mechanism and will incorporate it in the same wireless capsule. Another clinical study will use this stabilization-enabled capsule to show that its instantaneous velocity does not exceed the requirements for effective colonic imaging and therapy.

[0100] Therapy. LAT and EAT therapy methods will be developed on the benchtop and studied using endoscopic therapy probes on rat models of colorectal adenomas6 and then optimized and validated in swine models of colorectal cancer.7, 8After the most efficacious method for therapy is selected, we will conduct a clinical study using colonoscope-compatible therapeutic probes to demonstrate safety and efficacy.

[0101] Machine Learning. Data for ML will include those collected in our trimodal probe clinical studies, benchtop studies with excised human polyps, swine adenoma model, colonic video capsule endoscopes, and publicly available VCE images and down-sampled WL 19  MGH 2024-124-03 Quarles 125141.04710     videos9, 10obtained from colonoscopes.11-13Machine learning will be initially developed and validated using a software platform that emulates the proposed ML component of the ASIC and then on the actual ASIC itself.

[0102] WILDCAT Development. Using knowledge from the research phase, we will develop a WILDCAT device suitable for human use. Devices will be developed using an industry-standard design control process, fabricated in a class 10000 clean room using GMP, and approved for human studies through an FDA Investigational Device Exemption (IDE).

[0103] WILDCAT Clinical Feasibility. Unsedated bowel-prepped subjects, scheduled to undergo colonoscopic screening, will swallow the WILDCAT devices prior to their colonoscopy exam. After WILDCAT has been eliminated, patients will undergo their scheduled colonoscopy as per the standard of care; any region treated by the capsule will also be biopsied. Endpoints will include accurate determination of the presence / absence of adenomas and their sizes, and complete eradication of diminutive adenomas if present.

[0104] Considerations

[0105] (1) adequate access to validated training data to build "on-board" decision support for the device

[0106] WILDCAT will contain two WL video cameras. WL imaging (WLI) is significant because ML algorithms using VCE and colonoscopy images accurately detect colonic polyps and increase adenoma detection rates.14, 15Moreover, the use of WLI allows us to augment training data through publicly available16-18and proprietary colonoscopy datasets that can be obtained in future licensing agreements. We have also shown >90% accuracy with OCT using < 100 training images (FIG. 7) and thermal imaging has demonstrated high sensitivity for detecting neoplasia.19-21When data from all three imaging techniques (WLI, OCT, and thermal) are used to create a fused model, it should at least perform as well as the best model that uses data from only one of these modalities. The fused model should also require fewer training datasets than a single modality model.22-24

[0107] The trimodal probe imaging study will last 3.5 years, collecting WL, thermal, and OCT image data from -250 patients. Data will be further augmented by utilizing WL, OCT, and thermal images obtained from our studies in rats and swine models of colorectal adenomas. We also will continue to conduct OCT imaging of excised human colorectal polyps. At our current collection rate of -500 / year, we will have collected OCT images of 2000 excised 20  MGH 2024-124-03 Quarles 125141.04710     colorectal polyps (2 million OCT images). Given these considerations, we believe that we will acquire enough data to train a ML model that can be used to demonstrate WILDCAT clinical feasibility. Finally, since WILDCAT will transmit diagnostic WL, OCT, and thermal images after the capsule is excreted, data can accumulate to continue to update and improve WILDCAT's algorithms following scale up and commercialization.

[0108] (2) engineering risks mitigations around battery life

[0109] Since WILDCAT is fully customized device, we will be able to optimize all parameters to fit within our specifications and maximize the battery's size by better conforming it to the capsule's shape. By using rechargeable batteries alternative energy methods such as glucose biofuel cells, wireless power transfer, or piezo-energy harvesters could be used to increase the usable energy and overall lifespan of WILDCAT. Furthermore, our custom ASIC will be designed to control the power-up and sleep cycles of each individual component. As the capsule is swallowed and travels through the upper GI tract and small intestine, the only powered components being utilized will be a small, low-power pH sensor and low-level processing on the ASIC. Once the pH sensor indicates that the capsule has reached the colon, the ASIC will initiate a start-up sequence for the imaging modalities and other critical components. Images will be taken and processed at a rate high enough to obtain complete visualization of the colon, but intermittently enough to conserve energy. For therapy, the laser diode will be driven by a supercapacitor that stores charge, minimizing its effects on the battery's lifespan. Finally, all the capsule's custom electronics (ASIC, sensors, motor, etc.) will be designed to consume lower power than off-the-shelf components.

[0110] (3) adequacy of transit time and on-board capability to fully "visualize" and treat suspicious lesions

[0111] To address this challenge, we will utilize our existing wireless capsule with an integrated pH sensor and capsule position tracking wireless receiver platform to investigate the instantaneous velocity of capsules in the colons of ambulatory human subjects. An antenna array placed on the subject will record timestamped data of the capsule's position and pH; corresponding values for the capsule's instantaneous velocity in the colon will be determined. Results will indicate whether imaging and / or autonomous therapy would be hindered by sudden, rapid motions of the capsule in vivo. This data will provide critical requirements for stabilization and WILDCAT operation. Concurrently, we will develop stabilization 21  MGH 2024-124-03 Quarles 125141.04710     mechanisms that can be deployed in WILDCAT to slow the capsule down, center the device in the lumen, and potentially displace fecal contents during its transit. We anticipate that a combination of stabilization and proper selection of imaging and therapy parameters will enable comprehensive imaging and complete therapy of colonic diminutive adenomas.

[0112] (4) developing evidence base to displace long-established practice norms and funding streams

[0113] The cost of a colonoscopy in the US ranges from $2000-$3000,27, 28not considering expenses associated with complications, patient time / work lost to sedation recovery,29, 30and 5-10% incidence of post- colonoscopy colorectal cancer (PCCRC)31, 32due in part to -20% missed adenoma rate.33,34In volume production, we anticipate that the direct cost of WILDCAT will be similar to that of a VCE device ($500); at this superficial level, the financial advantage of our approach seems clear. Nevertheless, there are many other factors associated with the ultimate cost of WILDCAT (diagnostic accuracy, capsule administration intervals, reimbursement, patient outcomes, etc.).

[0114] Example 3:

[0115] 1 Electronics

[0116] 1.1 Battery

[0117] Rationale. The space, power, and energy demands of WILDCAT's subsystems requires energy storage that is >1.5 kJ, delivery of an instantaneous power >200 mW, and a total length <1 cm. While many power delivery strategies have been investigated for implantsand other deep tissue biomedical electronics,42-45as of today, only a battery meets our necessarypower density and length constraints. The current standard biomedical chemistry for medium current draw devices is Lithium Manganese Dioxide with on-the-market cylindrical packagingof 1 cm length.46

[0118] Overview. A typical wireless capsule draws current at a rate of 5 mA to 10 mA, supporting -30-hour transit through the GI tract. From our previous thermal capsule work, wehave used anoff-the-shelf battery (CR1 / 3N LiMnO2) at 3 V, with continuous current up to 60mA, and 170 mAhcapacity. While this chemistry and packaging supports simple capsules,WILDCAT has higher storage and power requirements and added space limitations. A custom battery with higher energy and power density is important for ensuring that the battery will last for the duration of the capsule's transit through the GI tract while providing adequate power for 22  MGH 2024-124-03 Quarles 125141.04710     each subcomponent.

[0119] Power utilization analysis. WILDCAT is initially intended to be used with bowel-prepped subjects who take boosts during the procedure, resulting in colonic transit times (CTT) between 2.5 hours to 6.5 hours.47, 48A power analysis for a suitable battery is presented in Table 2 for a colonic transit time (CTT) of 9 hours at a capsule- appropriate size of 11.5 mm diameter and 6.69 mm length.

[0120] Table 2. WILDCAT power utilization for a CTT of 9 hours. Power Time Voltage Current Time Energy .4.64 5 .4.258 6.

[0121] 1.2 ASIC

[0122] Rationale.

[0123] An Application Specific Integrated Circuit (ASIC) allows the combination of all components necessary for capsule functionality onto a single chip enabling less distance between components for faster communication, smaller packaging and decreased resistance between high current lines for lower power. The IC design is split into each ASIC sub- component, and carried out in schematic, layout and full system simulation before sending the chip off for fabrication. Because of the high expense and time to perform a tape-out of a custom IC, simulated verification is an important step of this process and so we use industry standard 23  MGH 2024-124-03 Quarles 125141.04710     tools and software.

[0124] Overview.

[0125] The ASIC will be designed with 4 core components: 1) A core microcontroller featuring a CPU, RAM and ROM, 2) A mixed-mode (both digital and analog sub-components) peripheral interface for thermal, OCT and white light imaging and for delivering therapy such as laser ablation, 3) A Neural Processing Unit (NPU) to load a machine learning algorithm to assess images and perform screening and diagnosis to inform therapy and 4) an RF Communication block for transmitting images in a research setting. The ASIC is designed to operate off of battery power and enable a full process of imaging, screening, diagnosis and treatment for the WILDCAT capsule.

[0126] Methods. For the tape-out run, each core block will be first developed independently. Development of sub-units will be split into two phases: schematic level design, and layout. Schematics will be designed in Cadence Virtuoso. CMOS geometries will be optimized in Virtuoso ADE Assembler and Spectre Multi-Mode Simulation. Simulated inputs such as clock signals and sensor data-streams can be input into the SPICE model and coordinated within each block's sub-components. In parallel with SPICE simulations, layout should also be iteratively performed using Virtuoso Layout Suite as placement of components can dictate available geometries and semiconductors available for use as well as determine cross component noise. Following the layout of a core block, additional simulations will be performed characterizing robustness of the system to power line noise and timing variations using Voltus IC Power Integrity, Quantus parasitic extraction simulations, and Tempus Timing Signoff solutions.

[0127] Full system simulation will be accomplished by using Jasper Register Transfer Level (RTL), Xcelium Logic, and Xcelium Mixed Signal simulation which all are tasked by the AI Verisium Manager to locate and determine root bugs. Simulations of the communications between each major core block ensure robustness to cross platform signal timing, coupling of noise between systems and overall adequate power draw.

[0128] Once the system design has been addressed using sub-system and full system verification, the layouts will be run through foundry-guided design rule checks, layout vs. schematic checks to ensure consistency between design documents, and adequate electrical rule checks. The generated files will then be sent to the foundry for a multi-process wafer run. 24  MGH 2024-124-03 Quarles 125141.04710     Once the chips are obtained, standard practice is to test all sub-components of the designed chip using both a breakout PCB and ESD micro-manipulated probing station during functionality. Design inputs informed by clinical studies as well as results from physical IC testing will inform a second tape-out run geared toward a final WILDCAT capsule product and the above procedures will be iterated and repeated to deliver a final chip.

[0129] Testing and validation. testing and validation are all performed by simulating the IC hardware at a gate and transistor level. At the schematic level, analog and mixed mode circuits are verified by simulating performance using the Virtuouso ADE for general simulations. Spectre Multi-Mode simulates harmonics that occur on high speed signals needed for data communication and Radio Frequency signals. Voltus Power allows for in depth power analysis at all levels of chip design. For digital components Xcelium provides a simulation environment to handle several million gate components such as the CPU, RAM, ROM, and NPU. Tempus Timing applies additional testing for signal timing compliance to ensure proper rise, fall and hold times to ensure data stream fidelity in cross component communications. Genus Logic Synthesis assists in generating register-transfer-level (RTL) design necessary for larger sub component logic blocks.

[0130] At the layout level, parasitic losses become critical for both analog and digital components when using advanced node technologies such as TSMC 12nm due to small gate size necessary for high speed causing high parasitic leakage. Parasitics are simulated by Cadence Quantus and play an important role in power analysis with Voltus. Similarly additional noise in the system from layout can affect Timing and signal integrity. For final checks for foundry, the Pegasus DRC = Design Rule Check, LVS = Layout versus Schematic, DFM = Design for Manufacturing are critical to run early in layout design to ensure compatibility.

[0131] Full system verification requires simulations of the entire chip, carrying a high computational load. For this we use the Verisium Manager which schedules stock tests of common pitfalls in interchip communication and power. It then helps guide tests to target areas of highest concern based on results using the Xcelium simulation environment to simulate chip performance and the Jasper Formal to check if the design satisfies requirements using a static analysis technique of formal mathematical methods.

[0132] Post-tape-out: Post tapeout designs functionalities will be assessed using both 25  MGH 2024-124-03 Quarles 125141.04710     a breakout PCB and ESD micro-manipulated probing station. The ESD probes can be connected to our electrical analysis tools such as logic analyzers, oscilloscopes, network analyzers, etc. To characterize RF component fidelity, logic level functionality, and overall mixed mode integrated circuit functionality.

[0133] Milestones.

[0134] ASIC Design Resource Access: The team purchases and sets up access to all IC design software services. This includes delivery of a server module, setup of access to said module, and beginning work on the schematic and layout. SDK kits should also be obtained for TSMC 12nm technology from the designated Multi-Process Wafer foundry.

[0135] Timeline: 1.5 months (In Parallel with Peripheral Specification Finalization)

[0136] Peripheral Specification Finalization: The team determines all specifications to be delivered by vendors for peripherals necessary to design chip. ^ Battery: Battery Voltage, capacity and max discharge rate ^ Super Capacitor: ESR (Equivalent Series Resistance), Capacitance, Maximum voltage ^ Motor: Input voltage, Input current, onboard hall position sensors (if any), and timing diagram for driving. ^ Thermal Ablation Laser: Drive voltage, drive current, capacitance, reverse voltage. ^ O.C.T. Laser: Tuning Voltage Range, drive voltage, drive current, capacitance, reverse voltage, output optical power, bandwidth. ^ O.C.T. Photodetector: Equivalent resistance, capacitance, reverse bias voltage, as well as expected forward current range. ^ White Light Camera: Driving voltage, SPI communication timing and logic specifications ^ Thermal Sensors: Selection of thermistor resistance value. ^ Thermal Management: TEC specifications of driving voltage, driving current vs Qc (Cooling capacity in Watts). ^ pH Sensor: Biasing voltage, forward current, gate capacitance, reference voltage range.

[0137] Timeline: 1.5 months (In Parallel with ASIC Design Resource Access)

[0138] Digital IP Configuration and Sub-component testing: The digital blocks are all generated from purchased IP modules. Success is measured by successful simulations of the subcomponents of these blocks. For RAM / ROM this includes full write / read to all memory 26  MGH 2024-124-03 Quarles 125141.04710     locations, within the CPU, characterization of core clock signals jitter and rise / fall time. I / O output speeds and rise / fall times will similarly be tested along with noise stress testing along major power inputs. Within the NEO unit, successful loading and execution of a basic Linear Neural Net and Convolutional Neural Net will verify functionality. ^ CPU Generation: Based on the I / O requirements, the CPU clock speed, RAM, ROM, cache sizes will be specified. A Tensillica Neo CPU will then be generated using the Tensillica GUI Interface. This creates all schematic and layout documents necessary to begin interfacing with all other components. ^ NEO Neural Processing Unit: Once the imaging I / O and CPU have been generated, a similar GUI interface will be used to generate the NEO Neural Processing Unit schematic and layout components using the Neo 210X architecture.

[0139] Timeline: 1-2 weeks initial generation. Simulated sub-component testing 4-6 weeks.

[0140] Initial Analog Schematic Design and Subcomponent Simulation: All peripherals can be enumerated into several subcomponents (amplifiers, ADC’s, DAC’s), each of which will carry standardized subcomponent testing. Success for each imaging / sensor block is measured by a simulated input image signal and desired output. Success of the therapy interface is successful driving of a simulated photodiode at the specified current. An IC subsystem test report documenting all tests and signals will be generated showing robustness to simulated noise parameters and functionality of each sub-block. ^ Imaging / Therapy Control I / O: Using the peripheral interface specifications the peripheral blocks shown in FIG.14C will be designed for each peripheral. o White Light Serial Interface: SPI Interface of Digital Input Array, Clock driver I / O and power bus ^ Estimated Time: 6 weeks o O.C.T. Interface: DAC for tuning VCSEL wavelength, Transimpedance amplifier for photodetector, 20MSPS 16bit ADC for photodetector, VCSEL Laser driver, Thermal management controller for VCSEL Laser, motor controller ^ Estimated Time: 4 months ^ DAC 4 weeks, ADC 4 weeks, VCSEL Laser driver 2 weeks, Thermal management 2 weeks, motor controller 3 weeks. 27  MGH 2024-124-03 Quarles 125141.04710     o Thermal Interface: 1 MSPS 12 bit ADC, Analog multiplexer for input to ADC ^ Estimated Time: 2 months o pH Interface: 1 MSPS 12 bit ADC, biasing voltage output, impedance matching to probe ^ Estimated Time: 1 month o Therapy Interface: Ablation laser driver, Motor controller ^ Estimated Time: 1 month ^ RF Communication Schematic: Design of RF PA System including clock, modulator, PA Amplifier, Digital input buffer, output analog impedance matching network. o Estimated Time: 2 months.

[0141] Initial Layout Design: Layout of all schematic components will be performed and followed by rigorous sub-unit and full system verification, repeating sample image input simulations and successful communication from Analog components to digital component blocks. Each block’s full system simulation with each peripheral system marks a successful sub-milestone. An IC system test report documenting all tests and signals will be generated showing robustness to simulated noise parameters and functionality of each sub-block. ^ Imaging / Therapy Control I / O: o White Light Serial Interface (4 weeks) o O.C.T. Interface (6 weeks) o Thermal Interface (4 weeks) o Therapy Interface: Ablation laser driver, Motor controller (2 weeks) ^ RF Communication: (6 weeks) ^ CPU Layout: (2 week) ^ NPU Layout: (3 weeks)

[0142] Initial Tape-out Generation: Following successful full system simulation of each core sensor block, tape-out specification documents will be generated and sent off to the foundry service to begin tape out.

[0143] Estimated time: 1 week

[0144] Initial Tape-out: Engineers will coordinate and revise design in accordance with foundry specifications as any changes occur to multi-process wafer service.

[0145] Estimated time: 3 months 28  MGH 2024-124-03 Quarles 125141.04710

[0146] Testing PCB Design: A printed circuit board (PCB) will be designed during tape out to test the incoming IC’s, featuring breakouts of all pins and interconnects with any peripheral devices.

[0147] Estimated time: 1 month (In parallel with Initial Tape-out)

[0148] Testing PCB Fabrication: PCB Gerbers will be sent out to foundry and engineers will work with foundry to ensure fabrication and answering engineering questions.

[0149] Estimated time: 2 months (In parallel with Initial Tape-out)

[0150] Initial IC Testing: Each analog IC block will be tested with imaging input and probed along block to ensure successful functionality up to input to CPU. CPU and NPU will be tested with successful programming. RAM / ROM will be tested on each with successful full read / write. The system will be probed using a microprobe station. An IC test report documenting all tests and signals will be generated showing robustness to simulated noise parameters and functionality of each sub-block.

[0151] Estimated time: 3 months

[0152] IC Redesign: Following inputs from imaging studies, a final decision on the most beneficial of the three imaging modalities to include will be implemented into the IC design. Unused imaging blocks will be removed, and improvements as needed to IC sub-component and full system performance will be implemented. The RF communication block may be replaced with an alternative communication mechanism such as a buzzer to change from research use to consumer. An IC Subsystem test report and IC full system test report will be generated like above milestones. Tape-out specification documents will be generated and sent off to the foundry service to begin tape out.

[0153] Estimated time: 6 months

[0154] Second Tape-out: Engineers will coordinate and revise design in accordance with foundry specifications as any changes occur to multi-process wafer service.

[0155] Estimated time: 3 months

[0156] Second Testing PCB Fabrication: Any needed revisions to the testing PCB shall be implemented and PCB Gerbers will be sent out to foundry for fabrication. Engineers will work with foundry to ensure fabrication and answering engineering questions. Estimated time: 3 months (In parallel with Initial Tape-out) 29  MGH 2024-124-03 Quarles 125141.04710

[0157] Final IC Testing: Each analog IC block will be tested with imaging input and probed along block to ensure successful functionality up to input to CPU. CPU and NPU will be tested with successful programming. RAM / ROM will be tested on each with successful full read / write. The system will be probed using the microprobe station. An IC test report documenting all tests and signals will be generated showing robustness to simulated noise parameters and functionality of each sub-block.

[0158] Considerations.

[0159] Digital Architecture Design: This project requires significant high end digital architecture for both the CPU and NPU. We foresaw development of a CPU or NPU from scratch as having a significant impact on timeline and carrying high risk. As such we budgeted to purchase IP for pre-designed, configurable CPU’s and NPU’s. We specifically chose the Xtensa LX8 CPU and Neo NPU because of their configurability int terms of internal interconnects between memory and I / O giving flexibility for any design changes found necessary during engineering and studies.

[0160] Machine Learning Compatibility: A major time sink with custom NPU’s is compiling neural network modules from high level python or C code down into machine code that is compatible with the designed architecture. We trained a CNN on a MAX78000 chip on a sample of human colon OCT images of varying grades of polyps, adenomas and carcinomas performing with >95% classification accuracy. The MAX78000 processing power, given by Multiply Add Compute (MAC) per cycle was measured at 9207 MAC / cycle or 921 GOPS (Giga-operations per second). We selected the Neo NPU both because it came with a supported compiler to translate Pytorch models into compatible machine code and the device was configurable to any amount of RAM or varying GOPS from 640 GOPS up to 5120 GOPS within our power budget of 3V and 20mA current draw.

[0161] Communications Fidelity: High speed communications internal to high clock speed devices such as CPU, NPU and for ADC’s such as the 20MSPS OCT ADC are subject to various performance mitigating noises largely determined by the layout. We budget for large scale full system simulation using the to ensure proper characterization of critical signals during full system operation and enable iterations of layout pre-tapeout.

[0162] Power Consumption: We are limited in total power consumption by our pill-sized form factor. The Voltus IC Power Integrity Solution allows us to simulate performance from a 30  MGH 2024-124-03 Quarles 125141.04710     battery source in full system simulations and test robustness to various noise sources such as EMG

[0163] Further considerations for alternative directions.

[0164] 1.3 Supercapacitor

[0165] Rationale. While the main source of energy for the wireless capsule is a battery, the proposed methods of laser therapy require a higher electrical power source for multiple seconds of treatment per identified lesion. In comparison to the high energy density of a battery, the high power density of a supercapacitor makes it the ideal solution for intermittently powered components that require high instantaneous currents. Preliminary experiments with off-the- shelf supercapacitors proved that successful tissue ablation can be achieved by driving a high- power laser diode with a supercapacitor charged by a capsule-sized battery (FIG.8). However, the best off-the-shelf supercapacitor had too high of an internal resistance and was large (8 mm diameter and 20 mm length), limiting its use in wireless capsules.

[0166] Design. The custom supercapacitor cell will have an 11.5 mm diameter to match that of our proposed custom battery diameter. The preliminary electrical specifications will be a cell of 5 F to 10 F, 2.7 V, and an equivalent series resistance (ESR) < 500 m^ to achieve a maximum current above 5 A. The capacitor will store the required energy for therapy, discharging at a high current for multiple seconds, and then recharging at a lower current from the battery between lesions. A power and energy analysis in Table 3 shows the capacitance required to drive each laser diode to fully treat the colonic mucosa over a -5 mm area, established by preliminary testing ex vivo (FIG.8B).

[0167] Methods, testing, and validation. Based on our design input specifications, reports on charge / discharge curves over time and the maximum possible current draw of the capacitor cell, demonstrating its energy capacity and power capabilities.

[0168] Considerations. If the super capacitor does not deliver within specifications, custom therapy lasers will be developed with increased efficiency. In so doing, the battery will contribute more power towards directly driving the therapy laser, enabling smaller capacitance and supercapacitor footprint requirements.

[0169] Table 3. Supercapacitor power and energy analysis Equivalent Therapy Wavelength Voltage Current Time Energy r31  MGH 2024-124-03 Quarles 125141.04710     LAT 1940 1.8 4 4 28.8 7.9 EAT 660 2.7 1.1 10 29.7 8.1

[0171] Rationale We will utilize rigid-flex PCBs, the standard for wireless capsule devices, including our own Wireless Thermal Capsule, as they allow for multiple traditional FR4 boards to be connected and integrated into a single board, minimizing noise and losses that discrete connectors and wires introduce. It also allows integration of discrete passive components, such as bypass capacitors, oscillators, and switches that are necessary for the proper function of the ASIC and electronics.

[0172] Overview PCB development will begin when each custom electronic component is specified and designed. Primary considerations will be ASIC footprint, ASIC passive components, PIC footprint, electrical connections / interfaces to the lasers, motor, pH sensor, WL image sensors, battery, and supercapacitor, and mechanical positioning of all components inside the capsule. The PCB will be designed using Altium and sent to a manufacturer for fabrication and assembly.

[0173] Methods The rigid-flex PCB will consist of multiple 10 mm to 11 mm diameter round rigid boards connected with flexible ribbons (e.g., FIGs. 6A, 8A). The ASIC will be packaged into a form that can be easily added to a PCB, and passive components will be added on the same boards, as close as possible to the IC to minimize noise and losses. Bypass capacitors will be added to all power traces from the battery and other regulated sources. An external crystal oscillator will be the main clock for the system. All GPIO pins will be broken out to connect to their associated peripherals, and a magnetic switch will be integrated to easily control the capsule's on / off state when placed in or removed from its packaging. The PIC will be packaged as a ball-grid array (BGA), allowing it to be easily soldered to the PCB. Electrical and mechanical constraints will be considered for interfacing the remaining peripheral subcomponents.

[0174] Testing and validation Boards will be X-rayed to verify that assembly contains no solder shorts. When PCBs are received, we will verify connectivity and functionality using PCB test points. PCBs will be tested with each peripheral in isolation before components are integrated.

[0175] Considerations Multiple PCB design reviews and procurement of PCBs in large 32  MGH 2024-124-03 Quarles 125141.04710     quantities will address potential issues that arise in PCB designs. Predicate device PCBs will be used to take functional shapes and blocks and implement them wherever possible to address these issues well.

[0176] 1.5 Motor

[0177] Rationale. WILDCAT's OCT imaging and side-firing therapy requires circumferential sweeping and aiming of their respective beams. Stepper motors are ideal for optimal WILDCAT imaging / therapy due to their reduced non-uniform rotational distortion (NURD) and capacity for rotational synchronization and discrete positioning. Here we describe our collaboration with Minebea Mitsumi to create a custom stepper motor that reduces the overall length of the device, lowers voltage and current utilization, and provides precision control capabilities.

[0178] Overview. In previous TCE devices, we used the SM34-F20 miniature stepper motor from Minebea that operates at 3 V, has 20 steps per revolution, and is 6 mm in length and 3.4 mm in diameter. It has successfully been used in OCT applications with resolutions and rotational speeds comparable to that of the WILDCAT device. We also developed and clinically tested control schemes to repeatedly stop it every 9° with high precision for laser therapy.49Given the critical space requirements of the WILDCAT device and the potential need for complex optical pathing for dual modality operation (OCT and laser therapy), the motor must be customized. Based on mechanical modeling, a motor with a larger diameter (up to 10 mm) and a reduced overall length including the body and shaft of -4 mm optimizes positioning, NURD, and footprint. Additionally, changing the shaft from a solid metal pin to a hollow core tube will allow a flexible optical layout for WILDCAT where therapy and imaging sources can be placed on opposite sides of the motor, enabling different lensing for each laser beam (FIG.11).

[0179] Methods. A micromotor will be designed by increasing the diameter, to decrease the length of the motor without adversely affecting rotational performance. The current industry standard motor has 163 mm3of body volume (3.4 mm diameter, 4.7 mm length). Our proposed geometry has 196 mm3of body volume (5.0 mm diameter, 2.5 mm length). Both these measurements are approximate and exclude the shaft, but preliminary models suggest that that the modified motor will have enough space for the necessary coils. Work will also be conducted to define and analyze the rotational load of WILDCAT's beam deflector. 33  MGH 2024-124-03 Quarles 125141.04710

[0180] Testing and validation. NURD will be evaluated using a high-speed camera imaging a fiduciary marker attached to the motor's shaft and quantified based on the magnitude of the difference between angular position from one frame to the next compared to the expected change. Targeting testing will use the same setup. Precession will be assessed by attaching a reflector to the shaft of the motor in the path of the sample arm of an OCT system and quantified by OCT image shift over time. Torque will be measured using a force sensor. Power draw will be tested for instantaneous and average current and coil voltage using a multimeter and oscilloscope.

[0181] Considerations A flywheel reflector will be used if the motor cannot provide stable imaging at WILDCAT speeds. This will increase the overall current draw and length of the system, but inertia can be used to smooth out performance. If stopping precision cannot match input specifications (-1 mm at target), then the motor can be geared down at the cost of additional length.

[0182] 1.6 Lasers and Thermal Management

[0183] Rationale. Consideration must be given to current control and thermal management when driving OCT imaging and therapy lasers. Driving a VCSEL for OCT requires a constant current and a tuning voltage synchronized to detection electronics while imaging. The therapy diode needs a high current to reach an optical power that affects treatment. Also, the integrity of the lasers can be compromised if excessive powers are used or if overheating occurs without adequate heat sinking and dissipation. To meet these needs, custom lasers will be developed to decrease overall size, increase power, and optimize thermal management over off-the-shelf options.

[0184] Overview. Two different lasers will be used in the capsule: a 1310 nm wavelength-swept VCSEL used for OCT imaging and a 1940 nm or 660 nm laser diode for LAT or EAT therapy, respectively. We will also dedicate resources to studying the ideal laser driving parameters, the extent of thermal management required for each, the effect of increased temperature on the laser's power, and the power consumption of different cooling processes at body temperature.

[0185] Methods, testing, and validation. We will build setups to conduct power consumption and heating generation tests to characterize the custom laser modules. Using off- the-shelf laser diodes and VCSELs, initial testing will establish the specifications required for 34  MGH 2024-124-03 Quarles 125141.04710     the new lasers and the potential thermal management needed. Most commercially available lasers come with heat sinks built into their packaging to ensure they operate at their advertised specifications; however, given the size and space constraints of WILDCAT, custom laser modules will be sourced to be as small as possible. We will thus need to investigate various options for heat sinking as a custom integrated solution.

[0186] Considerations To mitigate laser overheating, we will work with laser manufacturers to optimize packaging and driving circuitry to improve electric-optical efficiency. We will also investigate different laser driving schemes like pulsed mode operation that allows transient cooling. If heat dissipation remains problematic, we can use Peltier elements to cool the lasers during operation.

[0187] 1.7 pH Sensor

[0188] Rationale One of the biggest challenges for wireless capsules is battery life, and one way to extend battery life is by keeping the capsule in a form of sleep-mode until it reaches the colon. While there are a few known ways to identify entrance into the colon, one of the most well-studied and reliable ways is with a pH sensor. The pH of the GI tract varies throughout, averaging 1.9 to 2.2 in the stomach, 7.0 to 7.3 in the small intestine, and 6.8 to 7.2 in the colon, with distinctive transitions that can be tracked and used to identify these anatomical landmarks.50, 51In addition to being one of the most accurate metrics for colon entry, pH sensors are electrically non-complex and have extremely low power consumption. pH sensors have already been integrated into wireless endoscopy capsules such as the SmartPill,52so their feasibility for this purpose is well established.

[0189] Methods We will target an off-the-shelf or custom pH sensor with a sensitivity of at least + / -0.5 pH units and a total range of at least pH 0.5-9. Power consumption of these sensors is low; the MSFET-3330 pH sensor from Microsens and LioniX has a nominal power draw of 50 µW.53At this power, the pH sensor could run for 10 hours and only use 1.8 J, which is approximately 0.12% of our target battery's total energy. Specifications will also include a small form factor since we intend to incorporate the pH sensor in the capsule's outer shell. Microsens and LioniX produce an Ion-Specific Field-Effect Transistor (ISFET) in a 1.2 mm x 3 mm x 0.3 mm chip and a reference electrode with a 2 mm to 4 mm diameter. Additionally, Zimmer & Peacock produce a solid-state pH sensor in a small package.54pH sensing data will be routed to the custom ASIC, which will dedicate a small amount of its processing to the 35  MGH 2024-124-03 Quarles 125141.04710     analysis of sensor data, detecting WILDCAT's entrance into the colon, and signaling the other (imaging, stabilization, etc.) functionalities to start up. Based on the maturity of this technology, we do not anticipate any significant challenges integrating a pH sensor in the WILDCAT device that can detect its entrance into the colon.

[0190] Testing and validation. The manufacturer will provide power requirements and pH range, sensitivity, and response times for the off-the-shelf sensor samples and the finalized custom sensor. We will conduct range and sensitivity validation tests in-house to confirm that the sensors are able to distinguish pH values with the required range and resolution. Furthermore, the pH sensor will be integrated into the stabilization study capsules and will be cross validated with position tracking data to ensure they are indicating entrance into the colon at the correct time.

[0191] 2. OCT PIC

[0192] Rationale. OCT is a high-resolution, cross-sectional, microscopic imaging technology that shows promise for accurately detecting human colorectal polyps with malignant potential (FIG.7). As such, OCT is an ideal complement to WL and thermal imaging for maximizing diagnostic accuracy. However, current OCT systems, even when miniaturized, are too large to fit within the confines of a swallowable capsule. We therefore have decided to pursue integrating OCT components on a photonic chip, which in addition to being appropriately sized for capsule utilization, does not suffer from alignment challenges that would be associated with a miniature free space OCT system.

[0193] Methods A novel OCT system, meeting the specifications in Table 4, will combine a PIC comprising an optical waveguide-based interferometer, balanced photodetectors, k-clock, and transimpedance amplifiers (FIG.12A). A high-speed, high-power, chip-on-submount, wavelength-swept VCSEL module will be developed to ensure high performance OCT imaging. Ports will be provided on the PIC for VCSEL input light coupling and guiding sample arm light to the imaging optics. The VCSEL source will be optically coupled into the PIC using a waveguide (FIG.12B) printed using the Nanoscribe Quantum X Shape system with the Chip Printing set, combined with their 3D Grayscale Lithography technology (FIG.12B).55To achieve a high lateral resolution over an extended depth of focus (EDOF), we will also use this printer to create a multiplexed phase pattern that generates many closely spaced axial foci on the tissue (Fig.12C).56EDOF OCT imaging will allow us to attain 36  MGH 2024-124-03 Quarles 125141.04710     high lateral resolution regardless of where the capsule resides in relation to the colonic wall.

[0194] Testing and Validation. To ensure 3D printed surfaces do not cause diffraction artifacts on beam paths, we will adhere to the ^ / 10 criteria, requiring the Arithmetical Mean Height (Sa) to be under -131 nm, using a Bruker Surface Profiler to measure Sa on actual or equivalent objects when direct profiling is infeasible. For OCT, we will measure VCSEL-PIC optical input / output coupling and SNR using standard methods. Lateral resolution will be determined using resolution phantoms and beam profiling, and axial resolution and ranging depth will be measured via z-scanning techniques.

[0195] Table 4. Specifications for the OCT PIC module. Specification Value Center

[0196] Considerations If thermal drift alters the alignment of the printed microlenses that couple the VCSEL to the PIC, we will instead couple these two elements using Quantum X 3D printed waveguides.57Should the OCT PIC not be delivered within specifications on time, we will use the Quantum X to print miniature optics for a free space OCT setup that is compatible with WILDCAT. Optical assembly will be performed using automated robotic micromanipulators to carefully handle, place, and bond optical components that cannot be printed directly together.

[0197] 3. Imaging

[0198] Rationale. We are considering using three imaging technologies in the WILDCAT device for detection of colonic neoplasia. These technologies have complementary 37  MGH 2024-124-03 Quarles 125141.04710     strengths towards this clinical application: WLI is well established for ML-based polyp detection14, 58and can view ahead over a large field, OCT provides depth-resolved microscopic images that have been shown to accurately classify neoplastic polyps (FIG. 7), and thermal imaging is a functional readout that is active in cancer19-21and less affected by intraluminal contents. However, there are unknowns that should be addressed before finalizing WILDCAT's design. These include determining: 1) whether thermal imaging is sensitive for diminutive adenomas, 2) whether the small size and power constraints of a wireless capsule compromise WLI and OCT resolution in a manner that affects their diagnostic accuracies, and 3) how the combination of these three imaging modalities improves diagnostic performance.

[0199] Trimodality imaging probe. To answer these questions and determine the most appropriate combination of imaging technologies for WILDCAT, we will develop a flexible trimodality imaging probe that can be used through the accessory port of a colonoscope. This probe will be utilized in a clinical study of -250 subjects where WLI, OCT, and thermal images of polyps seen by colonoscopy will be obtained, biopsied, and sent for histopathology. Device development time will be short, as it is very similar to an existing triple-lumen OCT probe that is being used in clinical studies in our lab (FIG. 13). The central lumen of the probe's sheath will house a mechanically- scanning OCT catheter, connected to our current 1310 nm OCT system and rotary junction. The sheath's two smaller lumens will contain electrical wires for WLI and thermistor electronics located at the tip of the sheath (FIG.13).

[0200] WLI (Table 5), OCT (Table 4), and thermal (Table 6) image data will be configured to match that expected to be acquired by the final WILDCAT device. The only modifications to our existing OCT catheter will be the addition of EDOF optics. The WLI camera, LEDs, and optics will be the same as those that will be used in WILDCAT, comprising the OCHFA10 (Omnivision) module (Table 5). The LEDs will be soldered to wires joined with the pre-terminated OCHFA10 module cable, run through one of the sheath's smaller lumens, and received by an adapter board that will digitally store WL video in real time. The thermal array's wires will route through the sheath's other small lumen. OCT testing will be accomplished as described above. WLI field of view and aberrations will be determined using ISO stock targets for medical endoscopes. Thermal imaging sensitivity will be measured by varying a chip resistor heater from 37°-40°C; spatial thermal resolution will be evaluated using a thermal point source. 38  MGH 2024-124-03 Quarles 125141.04710

[0201] Clinical Study 1 - Trimodality Probe Clinical Study to Image Colorectal Adenomas. Clinical Study 1 will be conducted in two phases with an interim analysis after 1 year of enrollment. Adult subjects (n=245; 50% male, 50% female) undergoing a standard of care, high-definition, white light screening colonoscopy will be enrolled. During the colonoscopy, if a polyp is identified, then the trimodality probe will be inserted through the accessory port of the colonoscope to image the lesion. After imaging, the probe will be removed. Subsequently, the polyps will be resected and sent for histology as per standard of care. Uninvolved colonic mucosa may also be imaged and biopsied during the colonoscopic exam. After enrollment of the first 85 subjects (50% male, 50% female) (Phase 1, Years 1-2), an interim analysis will be performed to determine the suitability of the different modalities for adenoma detection. Features from each imaging modality will be extracted from the trimodality probe images and input into multivariate regression models or Machine Learning (ML) models with feature importance metrics, using histology as the gold standard. Predictive power and feature importance will be used to determine WILDCAT's expected diagnostic performance and the optimal diagnostic imaging modality(ies) to incorporate in WILDCAT's design. Enrollment will continue following the interim analysis for another 2.5 years through Phase 2 (n=160; 50% male, 50% female). All imaging data will be used to train / validate the ML algorithm for WILDCAT's tissue classification / decision-making algorithm.

[0202] Table 5. Specifications for the WLI subsystem. Specification Value

[0203] Table 6. Specifications for the thermistor array. Specification Value39  MGH 2024-124-03 Quarles 125141.04710     Sensitivity 0.05 C Sampling rate 1-2 MHz

[0205] Rationale. Even though WILDCAT is initially intended to be used in people after they have undergone bowel preparation, small amounts fecal contents can adversely affect OCT and WLI. Furthermore, ablation using 1940 nm light relies on water absorption to createthermal damage;59,60significant amounts abluminal water above the lesion could absorbenergy, rendering this therapy less effective. A mechanism for clearing contaminants could thus improve WILDCAT's diagnostic sensitivity and therapeutic efficacy. Moreover, if WILDCAT were to be used without bowel preparation in the future, a fecal clearance mechanism could be essential for its success.

[0206] Methods. To assess the magnitude of this issue, we will first obtain publicly available and archival VCE scans and virtual CT colonoscopy exams. This data will be analyzed to estimate the amount (percentage, thickness, type) of contaminants (feces, water, mucus) that are present in the colon under different bowel prepped and non-bowel prepped conditions. In parallel, we will develop benchtop setups that mimic colonic environments to test candidate clearance technologies, such as absorption-based devices utilizing biocompatible polymers that expand when exposed to water- rich intraluminal contents, moving them out of the way.3, 61, 62Another approach that will be investigated will be mixing reactants (e.g., aceticacid and sodium bicarbonate) inside thecapsule,63, 64 generating CO2gas that can eithersurround the capsule (FIG. 9) or inflate apericapsular balloon.64, 65From our testing, creatinga 4 cm diameter bubble around the capsule would require a reactant volume of 700 mm3thatcan easily fit within the WILDCAT device. After benchtop demonstration of the most feasible approaches, we will modify our R-TCE capsule with these clearance technologies so that we can deploy them in vivo and utilize OCT to determine their effectiveness. Initial testing will be conducted in bowel prepped naive swine. The R-TCE device (FIG. 10) will be advanced into the colon and regions of high fluid content will be identified (FIG.10A). OCT R-TCE images acquired before and after clearance mechanism activation will be used to measure the thickness of fecal contents interposed between the capsule and the bowel wall.

[0207] Clinical Study 2 - Clearance Mechanism Efficacy Clinical Study Using R-TCE. 40  MGH 2024-124-03 Quarles 125141.04710     In this study, the efficacy of the most promising clearance mechanisms will be evaluated using OCT, implemented through a modified R-TCE device. Unsedated adult subjects (n=30; 50% male, 50% female) who have undergone bowel preparation will be enrolled in this study. The modified R-TCE capsule will be placed within the subject and pulled back while imaging. Once sites containing intraluminal contents are identified by OCT imaging, the pullback will stop, the clearance mechanism deployed, and OCT images retaken. Intraluminal content thicknesses pre- and post-clearance will be recorded. This procedure will be repeated 10 times in each subject. To determine efficacy of the clearance mechanisms, OCT-determined intraluminal content thickness / volumes pre-and post- clearance actuation will be measured and compared.

[0208] Considerations Retention can be mitigated by ensuring that the expansion regions are malleable and / or reverse over time. Premature deployment of clearance technologies could be problematic if in the small intestine. Extensive testing of pH-sensitive colonic deployment triggers will be critical.

[0209] 5. Stabilization

[0210] Rationale: Portions of the colon are larger in diameter than the size of the capsule, potentially allowing the capsule to tumble in the organ during transit.66This tumble, and periodic peristaltic movement may precipitate irregular and potentially rapid instantaneous capsule velocities. A mechanism for stabilizing or slowing WILDCAT in the colon would addresses these issues.

[0211] Methods. To establish baseline and extreme transit speeds for non- stabilized capsules, a clinical study will be conducted to measure the instantaneous capsule velocity in real-time. Leveraging our clinically approved wireless thermal capsule platform, we will develop a wireless triangulation capsule with a pH sensor to determine when capsule is in colon (Table 7). An antenna array placed on the subject will continuously record timestamped data of the capsule's position and pH in the body throughout the entire procedure; corresponding values for instantaneous velocity in the colon will be determined. Should results indicate that deceleration is needed to implement effective adenoma therapy (e.g., maximum velocities exceed 1 mm / s), capsule stabilization technologies will be developed and tested. Methods that we will consider for development include the pericapsular balloon64, 65discussed in the Clearance Section 4 above, pH-sensitive absorbent polymers that expand when exposed to colonic fluid contents (FIG.16),3or the use of deployable appendages that expand the capsule's 41  MGH 2024-124-03 Quarles 125141.04710    effective diameter or grip the mucosal surface.67-69Capsule position, velocity, and time to fullexpansion will be measured in phantoms and swine in vivo.

[0212] Table 7. Specifications of the pH triangulation capsule. Spatial resolution 7 mmeless Capsule Instantaneous Velocity Measurement Study. Adult subjects (n=20; 50% male, 50% female) who have undergone bowel prep will be enrolled in the study. The subject will swallow the pH capsule, undergo a bowel boost regimen (e.g., split dose oral sulfate,0.25 L / dose), while continuing their normal daily activities. Signals from the antennae will be digitized, recorded, and used to calculate the capsule's position as a function of time, allowing computation of its instantaneous velocity as the capsule traverses the colon. After the capsule is excreted, the subject will return the receiver belt to the study team (FIG. 5C). Descriptive statistics will be used to summarize instantaneous velocity data, with a focus on the fastest velocities. This data will inform on the need and input specifications for WILDCAT device stabilization mechanisms.

[0214] Clinical Study 4 - Study to Test the Effectiveness of Capsule Stabilization Mechanisms. If the results from Study 3 indicate that high colonic capsule instantaneous velocities require mitigation for effective WILDCAT operation, then Study 4 will test the efficacy of various mechanisms for stabilizing and slowing the capsule's velocities in the colon. The wireless capsule with an integrated pH sensor used in Study 3 will be outfitted with the stabilization mechanism under investigation. Adult subjects (n=30; 50% male, 50% female) who have undergone bowel prep will be enrolled in the study. The subject will swallow the capsule, undergo a bowel boost regimen (split dose oral sulfate, 0.25 L / dose), while continuing their normal daily activities outside of the hospital. The pH sensor will track the progression of the capsule through the GI tract and will actuate and deploy the incorporated stabilization methods once in the colon. Instantaneous velocity of the capsule in the colon will be recorded 42  MGH 2024-124-03 Quarles 125141.04710     as described in Study 3. The primary outcome of this study will be successful demonstration that the capsule stabilization technique can slow the device sufficiently to implement consistent and comprehensive WILDCAT imaging and therapy.

[0215] Considerations Capsule retention and premature deployment risks are discussed in the Clearance Section 4 above. Some mechanisms discussed have the potential to damage the mucosa, meriting extensive preclinical safety testing. Space and power requirements will also need to be considered to ensure that the stabilization mechanisms are compatible with the final WILDCAT design.

[0216] 6. Therapy

[0217] Rationale. We have chosen to utilize laser therapy for adenoma treatment in WILDCAT because the capsule's beam-directing motor enables direct lesion targeting, the laser / electronics fits within in the device, and there is a robust literature demonstrating the capacity of lasers to effectively remove small adenomas.70-74Different laser-based ablationapproaches, including LAT based on water absorption75and EAT based on swallowable agentabsorption76-78will be investigated. LAT is a standard method for diseased tissue treatment inmedicine and does not require application of external agents but may be less efficacious when copious water-containing fluids are interposed between the laser and target. EAT is implemented using laser light that is not absorbed significantly by water but requires topical application of the absorbing agent, which we anticipate will be accomplished via oral administration, as is done with LumeBlue® for chromoendoscopy.79, 80

[0218] Methods. FIG.20 shows the workflow of the therapy research and development conducted here towards a WILDCAT device. LAT (1940 nm) and EAT (MB: 660 nm or gold nanoparticles: 500-800 nm) will be investigated and optimized using phantoms andCOMSOL modeling, tested and improved in rat adenoma model6tissue ex vivo, and optimizedand verified in rat adenoma models in vivo. ln vivo experiments will then be conducted in swineadenoma models7,8using a fiber optic probe (with aiming beam) delivered through anendoscope; polyps will be identified endoscopically and treated. After animal sacrifice, the treated region will be resected for investigation of laser damage extent (depth) and adenomatherapeutic efficacy using NTBC81tissue viability histology. One treatment modality will beselected for the WILDCAT capsule based on the findings from these studies. After the swine in vivo experiment is conducted, we will conduct a clinical study with a colonoscope- 43  MGH 2024-124-03 Quarles 125141.04710     compatible fiber optic probe to determine the safety and efficacy of laser-based adenoma treatment in humans with the selected method.

[0219] Clinical Study 5 - Colonic Adenoma Therapy Safety and Efficacy Study. Laser light-based LAT or EAT will be implemented using a fiber-optic therapy probe inserted through the accessory port of a colonoscope. The therapy probe will have the same therapy- delivery optics as the final WILDCAT device and will implement a visible aiming beam. The laser will operate at wavelengths and energy dosages determined by the animal studies and consistent with WILDCAT implementation. Adult subjects (n=40; 50% male, 50% female) undergoing screening colonoscopy or willing to undergo a research colonoscopy will be enrolled in the study. If EAT therapy is implemented, subjects will be given an oral dose of an exogenous photothermal absorber (e.g. MB or gold nanoparticles) prior to the colonoscopy in a solution or via a delivery capsule such as LumeBlue®.79, 80Subjects will undergo sedated, high-definition white light colonoscopy. Imaging of adenomas will also be conducted with the colonoscopic trimodality imaging probe to augment ML training / validation data. If a diminutive polyp is identified by colonoscopic exam, the therapy probe will be inserted through the accessory port of the colonoscope and used to treat the lesion under direct visualization with guidance from the aiming beam. Normal mucosa will also be ablated to assess treatment of false positives. The treated region will be biopsied and sent for histology. To evaluate safety, the treatment site will be observed for signs of excessive bleeding or perforation and subjects monitored for perforation symptoms. Histopathologic NTBC staining should also show no evidence of damage beyond the submucosa. Successful therapeutic efficacy will be determined as complete eradication of 90% of diminutive adenomas, consistent with gold standard care cold snare and cold biopsy.82,83

[0220] WILDCAT therapy module development. After the therapeutic approach and parameters are selected, we will finalize the WILDCAT capsule's therapy module. An OEM chip-on-carrier laser diode will be integrated into the WILDCAT device. The module will also comprise a supercapacitor, laser triggering mechanism, control system, and nano-printed light- delivering optics. Once an adenoma is identified and determined to be treatable by the ML algorithm, the laser diode will be turned on and therapy light will be sent to the target via optics directed at the mirror-mounted motor (FIGs 1, 11). The therapy laser's triggering and control system will be part of the ASIC mixed-mode I / O control electronics that handles treatment 44  MGH 2024-124-03 Quarles 125141.04710     command generation, triggering, and sending target coordinates to the motor's controller. These operations will be coordinated with other potential functions, including activation of clearance / stabilization mechanisms. The WILDCAT laser will be developed in compliance with FDA standards for medical lasers.

[0221] Considerations. Laser pulsing schemes may be implemented to ensure thermal confinement that may be required for some laser therapy approaches to avoid damage deep to the lesion. If mucus significantly hinders EAT agent deposition or attenuates LAT light, we can administer mucolytics that are selectively delivered to the colon or implement capsule-based mechanical mechanisms that scrape mucus from the luminal surface.84

[0222] 7. Machine Learning Rationale. The overall goal of our ML research is to develop an AI model with best performing WLI, OCT, and thermal imaging methods that can accurately classify colonic tissue: 1) Normal vs. polyp and 2) polyp without malignant potential vs. with malignant potential. Classification of the latter in addition to measurement of lesion size, which is straightforward with OCT, will determine whether to treat, leave, or flag a polyp for colonoscopic follow-up. The models will be capable of real-time inference performance meeting ASGE PIVI thresholds for making resection decisions for diminutive (≤5 mm) colorectal polyps without histology.85ASGE PIVI criteria for novel colonoscopic interventions (approved for diminutive adenomas only)1, 2are a ≥ 90% negative predictive value (NPV) for a lesion with malignant potential can be associated with a "Diagnose and leave" strategy, avoiding complications / costs of excising lesions without malignant potential while a ≥ 90% agreement between surveillance intervals can be associated with a "Resect and discard" strategy, avoiding costs associated with histopathology.

[0223] Overview. We will first develop and test various prototype machine learning models running in a virtual environment provided by Cadence for Neo NPU's adhering to the NPU specifications during the simulated ML development phase. Once the first ASIC NPU has been manufactured, a model will be ready for deployment and testing on physical hardware. As preliminary studies commence during the simulated ML development phase, we will validate our model's development progress with new datasets, providing performance benchmarks on a quarterly basis. Final deliverables will include labeled curated public, animal, and human study datasets for the final imaging modalities and a final ML model with hardware validation benchmarks. 45  MGH 2024-124-03 Quarles 125141.04710

[0224] Methods. Dataset curation and preprocessing. We will begin by curating a training dataset using our lab's internal OCT and thermal imaging human datasets and public WL-VCE and colonoscopy datasets (Table 8).16, 17, 87-90All datasets will be labeled by experts in the respective modalities and validated using gold standards. Data preprocessing and image augmentation techniques including filters, distortion correction, noise reduction, normalization, etc., will be applied during model training and inferencing, standardized, and documented to ensure dataset uniformity.

[0225] Table 8. Currently available datasets for ML training and validation. Description Modality Size Kvasir

[0226] NPU resource allocation. The ML and IC teams will collaborate to specify critical resources such as power budget, RAM, and computing power for the NPU. The ML team will test multiple types of common classification ML architectures such as CNNs,10, 86, 91-93Transformers,94and GANs (Generated Adversarial Networks)9to determinecan function in the given resource environment at the fastest speed.

[0227] Simulated development. All training, quantization, hardware compatibility and initial performance assessment will be done prior to obtaining the NPU hardware using the Neo NPU's design environment. Initial model development will use the curated datasets to iteratively develop ML models of increasing classification accuracy. Pretraining is common on both animal and public datasets for human application across endoscopy classification 46  MGH 2024-124-03 Quarles 125141.04710     methods, yielding significant performance improvements.9, 91, 95As animal and human datasets become available from the studies, we plan to iteratively utilize them to realize a 3-stage transfer learning model. The 3 stages will be trained on the curated datasets, animal model datasets, and human datasets to achieve benchmarked performance of >90% NPV and overall accuracy.

[0228] Hardware development. Once the ASIC arrives with the physical NPU, the team will collaborate with the IC team to upload the model developed in the Neo NPU environment onto the Neo NPU. Benchmarks will be taken to ensure model accuracy performance is maintained and performance is within power and speed specifications. All models will be logged and benchmarked for power consumption, image classification speed, model loss parameters, and classification accuracy agreement, using histology as the gold standard.

[0229] Considerations Issues including NPU power consumption, adequate features in training data, classification speed, and sufficient NPU resources to perform these tasks, will be mitigated by continuously iterating specifications for hardware needs and benchmarking performance and resource usage throughout the development cycle. For dataset adequacy, we have seen in the literature91, 92and our own preliminary data96(See Plans to address challenges highlighted in the preproposal above) that PIVI performance is achievable by WLI and OCT datasets. We plan to incorporate new datasets from animal / clinical studies as they become available to enable feedback to guide model development and ensure adequate feature discrimination.

[0230] 8. WILDCAT Development

[0231] Rationale. In Phase I we will have developed, identified, and validated all components that will be part of the final WILDCAT capsule. Here, in Phase II, we will create a cohesive, functional, and manufacturable WILDCAT device, testing different combinations of components and their interactions on the benchtop and in swine before integrating the final human-use capsule.

[0232] Overview. A robust autonomous capsule workflow (FIG. 14B) is critical for WILDCAT to safely meet the screening / therapeutic needs of end users. By establishing this operational flow, each component can be identified and its interactions with others tested. First, the capsule is swallowed, and the pH sensor is turned on. This sensor runs continuously with low-level power consumption throughout the entire GI tract until it reaches the colon. Once the pH levels have indicated that the capsule has entered the colon, the stabilization and / or 47  MGH 2024-124-03 Quarles 125141.04710     clearance methods (if and when needed) will be deployed to ensure clear and stable imaging and therapy. Colon entry will also trigger the commencement of imaging. To conserve battery power, WL, OCT, and thermal imaging may be used in WILDCAT at differing frequencies and operational phases. WLI will likely act as a continuous red-flag modality and the others will interrogate a flagged lesion to increase specificity. All images will be processed by the NPU in the ASIC, which will identify neoplastic polyps, determine approximate dimensions, and decide whether a polyp will be treated. If the decision is to treat, the therapy laser will be directed to the target location and activated. Imaging will resume following treatment. At any time, if the thermal sensors identify a significant drop in temperature, indicating a transition from body to ambient, the process will cease, and the capsule will wirelessly transmit clinical diagnostic information to an external receiver. This information will be used to determine future screening intervals and whether a follow up with a gastroenterologist is needed for colonoscopic treatment (e.g., larger adenomas).

[0233] Methods. During the development phase, we will implement a step-by-step approach to ensure reliability based on a mode effects analysis (MEA) to define and tune this decision flow and ensure that the controlling CPU in the ASIC follows it reliably. As part of the MEA, these decisions will be tested first in isolation, and then in combinations, and finally in the fully integrated device. Examples of integrated testing combinations are listed in Table 9.

[0234] Table 9. Partial list of integrated WILDCAT element testing. pH sensor and stabilization / clearance method(s)48  MGH 2024-124-03 Quarles 125141.04710     Power consumption of all components and overall battery life Capsule body suitability and fit testing of componentslopment, we will establish design inputs that include manufacturability. Manufacturing will be conducted under the supervision of the Manufacturing Manager using the Tearney Lab's Quality management process in our clean rooms and clinical device manufacturing facilities. During WILDCAT fabrication, the manufacturing team will ensure that each component meets the design inputs and specifications and will design and validate fixtures for each step of the capsule fabrication process. Once all components and fabrication steps are finalized, the manufacturing engineer will establish the final production processes, and an independent engineer will complete process verification. A partial list of capsule qualification tests is presented in Table 10.

[0236] Table 10. Partial list of WILDCAT qualification tests. Software

[0237] Considerations Custom components are set to have first iterations completed by the start of WILDCAT Development; however, if these first iterations do not meet design specifications or need revisions, a second iteration will occur, modestly deferring development timelines. The timeline for WILDCAT development is accelerated and obstacles (e.g., hiring 49  MGH 2024-124-03 Quarles 125141.04710     difficulties, supply chain, or changes in design) may cause delays. These risks will be mitigated by managing the team in an agile manner and by running projects in parallel.

[0238] 9. WILDCAT Clinical Feasibility

[0239] Clinical Study 6 - WILDCAT Device Clinical Efficacy Study. In Phase III of this research program, bowel prepped adult subjects (n=20; 50% male, 50% female), scheduled to undergo a standard of care colonoscopy, will be enrolled in the study. Subjects will swallow the WILDCAT device approximately 1-2 days prior to their scheduled colonoscopy, will undergo a bowel boost regimen (e.g., split dose oral sulfate, 0.25L / dose), and will proceed to continue their normal daily activities outside of the hospital. Subjects will be provided with a hand-held wireless transmission receiver to carry with them that will receive procedural data after the capsule is excreted. Depending on the final therapy modality selected to be used in WILDCAT, the subject may be orally dosed with an exogenous photothermal absorber prior to swallowing the WILDCAT device. When the WILDCAT device has been eliminated from the subject, wireless data transmission will be sent to a receiver. The data collected will indicate whether a colonic polyp or adenoma was identified, a diminutive adenoma was treated, and if there are additional untreated diminutive adenomas or larger adenomas present and their numbers. The receiver will also collect diagnostic images involved in the algorithm's autonomous decision-making process. 1-2 days after the WILDCAT procedure, subjects will undergo their scheduled colonoscopy. During the colonoscopy, all polyps will be identified and excised per the standard of care. Any region treated by the WILDCAT capsule will be visible as a large white spot on the mucosa. These regions will also be biopsied in their entirety. The WILDCAT device will be assessed based on its ability to accurately determine the presence or absence of adenomas, number of adenomas identified, the size of adenomas, and complete eradication of treated diminutive adenomas if present.

[0240] Analyses. Our first analysis will be to estimate WILDCAT's diagnostic accuracy and negative predictive value (NPV) on a per-lesion basis, from which we can compare WILDCAT's performance to ASGE's PIVI thresholds. Even though true positive adenomas treated by WILDCAT will not be viable, they should be able to be diagnosed histologically, even in the presence of histopathological evidence of thermal damage. This diagnosis can be facilitated via P53 and Ki-67 immunohistochemical (IHC) staining. Similarly, false positives can be identified via histopathology and absence of P53 and Ki-67 IHC. True negatives will be 50  MGH 2024-124-03 Quarles 125141.04710     clear via histopathology and false negatives will be diminutive adenomas by colonoscopic biopsy that were not identified by WILDCAT. Additional analyses will include the capacity of WILDCAT to completely eradicate diminutive adenomas that were identified by the device and the accuracy of WILDCAT for measuring adenoma number and size.

[0241] Comparison with other ongoing research indicating advantages and disadvantages of the proposed effort

[0242] Novel powering schemes. A maximum power of 380 mW and a capacity of 990 mWh will be needed for supporting the WILDCAT capsule. Biofuel cells can produce continuous power by breaking down external reactants such as glucose,97-99but current power densities are too low (1-100 mW / cm2) to enable all WILDCAT functionalities. Similarly, mitochondria biofuel100 and biobatteries that use intestinal bacteria and acidic pH to generate power101, 102give power densities less than 100 µW / cm2. Wireless power transfer such as microwave power transfer (MPT) has been explored for mm-sized implants and established to safely deliver 1-5 mW power levels to GI tract depths within FDA and FCC non-ionizing radiation limits.103-105Ultrasound power transfer faces less attenuation by tissue for mm-sized power harvesters and has similarly been shown to achieve 1-3 mW power transfer to deep tissue depths.106, 107Research that combines MPT and ultrasound using magnetoelastic harvesting has shown 8-12 mW at GI depths.108-110Ultrasound and MPT require an external transmitter which is undesirable for a capsule that is administered at home.

[0243] Other forms of imaging. While tethered ultrasound capsules have been explored,111-114no wireless ultrasound capsules have been demonstrated, although several papers detail progress on custom ultrasound ASICs, transducers, and control systems for wireless capsules.111, 115-117A company has formed claiming to have designed a wireless ultrasound capsule, but no data has been published.118Nothing has been reported on wireless capsule-based OCT imaging; however, several research groups have created ophthalmic 830- 840 nm specific PICs and ASICs that contain either subcomponents of OCT or entire OCT systems, miniaturizing OCT to sizes that are potentially compatible with wireless capsule in the future.119-121Capsule thermal sensing to date has been limited to measuring internal body temperature with relatively poor temperature resolutions (±1°C),122-124a magnitude lower than our thermal capsule. A 3-thermistor device called ThermPill measures the temperature of the upper GI tract's luminal surface, but has only been realized as a tethered capsule and tested in 51  MGH 2024-124-03 Quarles 125141.04710     swine esophagus.114, 125X-ray imaging has been produced in a wireless swallowable pill called C-scan that requires swallowing an iodine contrast agent, giving 2-3 mm resolution to detect colon polyps greater than 6 mm without requiring bowel prep.126-129While ultrasound imaging and X-ray imaging do not need bowel prep, these capsule technologies do not have sufficient resolution / contrast for identifying diminutive polyps or pathologic lesion discrimination. WILDCAT deploys a multifaceted approach that combines a low-power functional imaging modality (thermal), an industry-standard and proven technique (WLI), and a microscopic diagnostic modality (OCT). This multimodal optical approach should enable WILDCAT to determine in vivo pathology with sufficient accuracy to merit autonomous, real-time, treatment.

[0244] Clearance and stabilization mechanisms. Two papers in literature consider mixing acid with bicarbonate capsules to create CO2to inflate the GI tract, displace fecal matter, and improve imaging quality by reducing colonic folds,130, 131an approach that we will explore in this program. Prior capsule stabilization research primarily proposes mechanical appendages that extend to grab the lumen and exert pressure to halt capsule movement.132-134One method uses a highly absorbent polymer that reacts with water to expand against the lumen and create friction to stabilize the capsule; the reaction is irreversible so the device passes like stool following expansion.135We will explore similar absorbent polymer methods but with different materials, mechanical properties, and geometries that make elimination more feasible. One of the most promising stabilization methods that we will pursue combines clearance and stabilization functions into one by inflating a pericapsular balloon with CO2. Our laboratory has extensive experience with such balloon-based imaging devices49, 136-139and believe that this development is well within reach.

[0245] AI from VCE data. Machine learning using VCE data is more challenging than with HD colonoscopy video owing to the former's lower image resolution. Approaches to mitigate this issue have focused on generating super-resolution images from VCE data,9, 11-13which we will also do to increase our WLI training data. A wireless white light video capsule with a custom ASIC for performing on-board AI diagnosis using a CNN was recently tested on both swine models and on private human capsule endoscopy datasets.86This capsule demonstrated 95.8% - 99.5% precision for detecting colon polyps of all sizes86but did not report on polyp size or explore polyp type classification. Our proposed ASIC in WILDCAT will use three complementary image streams to not only detect diminutive colonic adenomas 52  MGH 2024-124-03 Quarles 125141.04710     but classify their pathology as well.

[0246] Therapeutic capsules Most therapeutic capsule research involves releasing a liquid or drug payload using internal or external magnetic / RF signal triggering with position tracking.140, 141Notable capsules release microneedles for delivering drug below the mucosal layer.142, 143Compared with these drug-delivery capsules, light-based therapy has the advantage that it can be precisely directed to the target lesion even when the capsule does noes not contact the mucosal surface. Two capsules have been proposed for phototherapy, one that shines LEDs at a wavelength that specifically kills Helicobacter Pylori144and another that describes the possibility of intravascularly injecting MB to selectively stain colorectal cancer, followed by administration of an LED capsule for conducting photodynamic therapy (PDT).145IV injection of an agent is not compatible with in-home use and topical administration of MB for adenomas is not selective.146Thus, LAT and EAT are superior therapeutic approaches for WILDCAT.

[0247] Example 3:

[0248] Outer Dip coating

[0249] Background: We found that adding a biomedical outer coating to our capsules help to give increased chemical resistance to stomach acid. A capsule is dip coated in this coating and cured to add an outer shell. We specifically use Masterbond ep62-1Med which carries a USP Class VI, ISO-10993-5 cytotoxicity certifications for biocompatibility and is certified to be resistant to sterilization methods such as steam and low temperature hydrogen peroxide (Sterris) sterilization methods someone skilled in the art would choose to use. We also could perform high level disinfection now with CIDEX OPA. The magnetic switch normally used to turn the capsule off cannot be used during high temperature cures because the other electronics are prone to become magnetized at temperatures higher than normal operation and could cause the device to stay turned off. As a solution, during elevated temperature curing the capsule electronics (microcontroller and / or thermal sensors) detect when the temperature is higher than normal operation and enter into a shutdown state to conserve battery life.

[0250] References 1. Committee AT, Abu Dayyeh BK, Thosani N, Konda V, Wallace MB, Rex DK, Chauhan SS, Hwang JH, Komanduri S, Manfredi M, Maple JT, Murad FM, Siddiqui UD, Banerjee S. ASGE Technology Committee systematic review and meta-analysis assessing the ASGE PIVI thresholds for adopting real-time endoscopic assessment of the histology of 53  MGH 2024-124-03 Quarles 125141.04710     diminutive colorectal polyps. Gastrointest Endosc.2015;81(3):502 e1- e16. Epub 20150116. doi: 10.1016 / j.gie.2014.12.022. PubMed PMID: 25597420. 2. Rex DK, Kahi C, O'Brien M, Levin TR, Pohl H, Rastogi A, Burgart L, Imperiale T, Ladabaum U, Cohen J, Lieberman DA. The American Society for Gastrointestinal Endoscopy PIVI (Preservation and Incorporation of Valuable Endoscopic Innovations) on real-time endoscopic assessment of the histology of diminutive colorectal polyps. Gastrointest Endosc. 2011;73(3):419-22. doi: 10.1016 / j.gie.2011.01.023. PubMed PMID: 21353837. 3. Dobromir Filip OY-P, Christopher N. Andrews, and Martin P. Mintchev. Self- Stabilizing Colonic Capsule Endoscopy: Pilot Study of Acute Canine Models. IEEE Transactions on Medical Imaging.2011;30(12):2115-25. doi: 10.1109 / TMI.2011.2163165. 4. Yan W, Wei Z, Yang Y, Wu D, Zhang Z, Song X, Qin J, Bi L. Ultra-broadband magneto- optical isolators and circulators on a silicon nitride photonics platform. Optica. 2024;11(3):376- 84. doi: 10.1364 / OPTICA.506366. 5. Tran K. Capsule colonoscopy: PillCam Colon. Issues Emerg Health Technol. 2007(106):1- 4. PubMed PMID: 17957839. 6. Amos-Landgraf JM, Kwong LN, Kendziorski CM, Reichelderfer M, Torrealba J, Weichert J, Haag JD, Chen KS, Waller JL, Gould MN, Dove WF. A target-selected Apc- mutant rat kindred enhances the modeling of familial human colon cancer. Proc Natl Acad Sci U S A. 2007;104(10):4036-41. Epub 20070227. doi: 10.1073 / pnas.0611690104. PubMed PMID: 17360473; PMCID: PMC1805486. 7. Flisikowska T, Merkl C, Landmann M, Eser S, Rezaei N, Cui X, Kurome M, Zakhartchenko V, Kessler B, Wieland H, Rottmann O, Schmid RM, Schneider G, Kind A, Wolf E, Saur D, Schnieke A. A porcine model of familial adenomatous polyposis. Gastroenterology. 2012;143(5):1173-5 e7. Epub 20120801. doi: 10.1053 / j.gastro.2012.07.110. PubMed PMID: 22864254. 8. Flisikowska T, Stachowiak M, Xu H, Wagner A, Hernandez-Caceres A, Wurmser C, Perleberg C, Pausch H, Perkowska A, Fischer K, Frishman D, Fries R, Switonski M, Kind A, Saur D, Schnieke A, Flisikowski K. Porcine familial adenomatous polyposis model enables systematic analysis of early events in adenoma progression. Sci Rep. 54  MGH 2024-124-03 Quarles 125141.04710     2017;7(1):6613. Epub 20170726. doi: 10.1038 / s41598-017-06741-8. PubMed PMID: 28747659; PMCID: PMC5529429. 9. Almalioglu Y, Bengisu Ozyoruk K, Gokce A, Incetan K, Irem Gokceler G, Ali Simsek M, Ararat K, Chen RJ, Durr NJ, Mahmood F, Turan M. EndoL2H: Deep Super- Resolution for Capsule Endoscopy. IEEE Transactions on Medical Imaging. 2020;39(12):4297-309. doi: 10.1109 / TMI.2020.3016744. 10. Wahab H, Mehmood I, Ugail H, Sangaiah AK, Muhammad K. Machine learning based small bowel video capsule endoscopy analysis: Challenges and opportunities. Future Generation Computer Systems.2023;143:191-214. doi: 10.1016 / J.FUTURE.2023.01.011. 11. Kohler T, Haase S, Bauer S, Wasza J, Kilgus T, Maier-Hein L, FeuBner H, Hornegger J, editors. ToF meets RGB: Novel multi-sensor super-resolution for hybrid 3-D endoscopy. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); 2013. 12. Sajjadi MSM, Vemulapalli R, Brown M. Frame-Recurrent Video Super-Resolution. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition.2018:6626-34. 13. Gohshi S, Mori C, Tanioka K, Yamashita H. Real-time Nonlinear Signal Processing Super Resolution of 8K Endoscope Cameras. ICETE. 2019;1. doi: 10.5220 / 0007950603430349. 14. Hassan C, Wallace MB, Sharma P, Maselli R, Craviotto V, Spadaccini M, Repici A. New artificial intelligence system: first validation study versus experienced endoscopists for colorectal polyp detection. Gut. 2020;69(5):799-800. doi: 10.1136 / GUTJNL-2019- 319914. 15. Desai M, Ausk K, Brannan D, Chhabra R, Chan W, Chiorean M, Gross SA, Girotra M, Haber G, Hogan RB, Jacob B, Jonnalagadda S, Iles-Shih L, Kumar N, Law J, Lee L, Lin O, Mizrahi M, Pacheco P, Parasa S, Phan J, Reeves V, Sethi A, Snell D, Underwood J, Venu N, Visrodia K, Wong A, Winn J, Wright CH, Sharma P. Use of a Novel Artificial Intelligence System Leads to the Detection of Significantly Higher Number of Adenomas During Screening and Surveillance Colonoscopy: Results From a Large, Prospective, US Multicenter, Randomized Clinical Trial. Am J Gastroenterol. 2024. Epub 20240223. doi: 10.14309 / ajg.0000000000002664. PubMed PMID: 38235741. 55  MGH 2024-124-03 Quarles 125141.04710     16. Smedsrud PH, Thambawita V, Hicks SA, Gjestang H, Nedrejord OO, Nress E, Borgli H, Jha D, Berstad TJD, Eskeland SL, Lux M, Espeland H, Petlund A, Nguyen DTD, Garcia-Ceja E, Johansen D, Schmidt PT, Toth E, Hammer HL, de Lange T, Riegler MA, Halvorsen P, Smedsrud PH, Thambawita V, Hicks SA, Gjestang H, Nedrejord OO, Nress E, Borgli H, Jha D, Berstad TJD, Eskeland SL, Lux M, Espeland H, Petlund A, Nguyen DTD, Garcia-Ceja E, Johansen D, Schmidt PT, Toth E, Hammer HL, de Lange T, Riegler MA, Halvorsen P. Kvasir-Capsule, a video capsule endoscopy dataset. Scientific Data 20218:1. 2021-05-27;8(1). doi: 10.1038 / s41597-021-00920-z. 17. Leenhardt R, Li C, Mouel J-PL, Rahmi G, Saurin JC, Cholet F, Boureille A, Amiot X, Delvaux M, Duburque C, Leandri C, Gerard R, Lecleire S, Mesli F, Nion-Larmurier I, Romain O, Sacher-Huvelin S, Simon-Shane C, Vanbiervliet G, Marteau P, Histace A, Dray X. CAD-CAP: a 25,000-image database serving the development of artificial intelligence for capsule endoscopy. Endoscopy International Open. 2020 / 03;08(03). doi: 10.1055 / a- 1035-9088. 18. The Gastrolab Image Gallery. 19. Lahiri BB, Bagavathiappan S, Jayakumar T, Philip J. Medical applications of infrared thermography: A review. Infrared Physics & Technology.2012;55(4):221-35. doi: 10.1016 / J.INFRARED.2012.03.007. 20. Mashekova A, Zhao Y, Ng EYK, Zarikas V, Fok SC, Mukhmetov O. Early detection of the breast cancer using infrared technology - A comprehensive review. Thermal Science and Engineering Progress.2022;27:101142-. doi: 10.1016 / J.TSEP.2021.101142. 21. Khomsi Z, Elouerghi A, Afyf A, Bellarbi L. Contribution for the Early Detection of Breast Cancer by a Superficial Thermography Solution. 2020 International Conference on Electrical and Information Technologies, ICEIT 2020. 2020. doi: 10.1109 / ICEIT48248.2020.9113225. 22. Kong W, Li C, Lei Y. Multimodal medical image fusion using convolutional neural network and extreme learning machine. Frontiers in Neurorobotics. 2022;16. doi: 10.3389 / fnbot.2022.1050981. 23. Odusami M, Maskeliunas R, Damasevicius R, Misra S, Odusami M, Maskeliunas R, Damasevicius R, Misra S. Explainable Deep-Learning-Based Diagnosis of Alzheimer's Disease Using Multimodal Input Fusion of PET and MRI Images. Journal of Medical and 56  MGH 2024-124-03 Quarles 125141.04710     Biological Engineering 202343:3.2023-06-20;43(3). doi: 10.1007 / s40846-023-00801-3. 24. Deep multimodal fusion for semantic image segmentation: A survey. Image and Vision Computing.2021 / 01 / 01;105. doi: 10.1016 / j.imavis.2020.104042. 25. Bock DC, Marschilok AC, Takeuchi KJ, Takeuchi ES. Batteries used to power implantable biomedical devices. Electrochimica Acta: Pergamon; 2012. p.155-64. 26. Deng J, Sun X, Peng H. Power supplies for cardiovascular implantable electronic devices. EcoMat.2023 / 06 / 01;5(6). doi: 10.1002 / eom2.12343. 27. Subramanian S, Tangka FKL, Hoover S, Cole-Beebe M, Joseph D, DeGroff A. Comparison of Program Resources Required for Colonoscopy and Fecal Screening: Findings From 5 Years of the Colorectal Cancer Control Program. Prev Chronic Dis. 2019;16:E50. Epub 20190425. doi: 10.5888 / pcd16.180338. PubMed PMID: 31022371; PMCID: PMC6513474. 28. Fisher DA, Princic N, Miller-Wilson LA, Wilson K, Limburg P. Healthcare costs of colorectal cancer screening and events following colonoscopy among commercially insured average-risk adults in the United States. Curr Med Res Opin. 2022;38(3):427-34. Epub 20211219. doi: 10.1080 / 03007995.2021.2015157. PubMed PMID: 34918589. 29. ASGE Colonoscopy: A life-saving procedure. 30. Cohen LB, Wecsler JS, Gaetano JN, Benson AA, Miller KM, Durkalski V, Aisenberg J. Endoscopic sedation in the United States: results from a nationwide survey. Am J Gastroenterol.2006;101(5):967-74. doi: 10.1111 / j.1572-0241.2006.00500.x. PubMed PMID: 16573781. 31. Corley DA, Jensen CD, Marks AR, Zhao WK, Lee JK, Doubeni CA, Zauber AG, de Boer J, Fireman BH, Schottinger JE, Quinn VP, Ghai NR, Levin TR, Quesenberry CP. Adenoma detection rate and risk of colorectal cancer and death. N Engl J Med. 2014;370(14):1298-306. doi: 10.1056 / NEJMoa1309086. PubMed PMID: 24693890; PMCID: PMC4036494. 32. Lee YM, Huh KC. Clinical and Biological Features of Interval Colorectal Cancer. Clin Endosc. 2017;50(3):254-60. Epub 20170321. doi: 10.5946 / ce.2016.115. PubMed PMID: 28320200; PMCID: PMC5475521. 33. Heresbach D, Barrioz T, Lapalus MG, Coumaros D, Bauret P, Potier P, Sautereau D, Boustiere C, Grimaud JC, Barthelemy C, See J, Serraj I, D'Halluin PN, Branger B, 57  MGH 2024-124-03 Quarles 125141.04710     Ponchon T. Miss rate for colorectal neoplastic polyps: a prospective multicenter study of back-to-back video colonoscopies. Endoscopy. 2008;40(4):284-90. doi: 10.1055 / s-2007- 995618. PubMed PMID: 18389446. 34. van Rijn JC, Reitsma JB, Stoker J, Bossuyt PM, van Deventer SJ, Dekker E. Polyp miss rate determined by tandem colonoscopy: a systematic review. Am J Gastroenterol. 2006;101(2):343-50. doi: 10.1111 / j.1572-0241.2006.00390.x. PubMed PMID: 16454841. 35. Dionisio PM, Gurudu SR, Leighton JA, Leontiadis GI, Fleischer DE, Hara AK, Heigh RI, Shiff AD, Sharma VK. Capsule endoscopy has a significantly higher diagnostic yield in patients with suspected and established small-bowel crohn's disease: A meta- analysis. American Journal of Gastroenterology2010. p.1240-8. 36. Eluri S, Paterson A, Lauren BN, O'Donovan M, Bhandari P, di Pietro M, Lee M, Haidry R, Lovat L, Ragunath K, Hur C, Fitzgerald RC, Shaheen NJ. Utility and Cost- Effectiveness of a Nonendoscopic Approach to Barrett's Esophagus Surveillance After Endoscopic Therapy. Clin Gastroenterol Hepatol.2022;20(2):e51-e63. Epub 20210210. doi: 10.1016 / j.cgh.2021.02.013. PubMed PMID: 33581357; PMCID: PMC8352994. 37. Kim CW, Baik SH, Roh YH, Kang J, Hur H, Min BS, Lee KY, Kim NK. Cost- effectiveness of robotic surgery for rectal cancer focusing on short-term outcomes: a propensity score-matching analysis. Medicine (Baltimore). 2015;94(22):e823. doi: 10.1097 / MD.0000000000000823. PubMed PMID: 26039115; PMCID: PMC4616367. 38. Kumar S, Saumoy M, Oh A, Schneider Y, Brand RE, Chak A, Ginsberg GG, Kochman ML, Canto MI, Goggins MG, Hur C, Kastrinos F, Katona BW, Rustgi AK. Threshold Analysis of the Cost-effectiveness of Endoscopic Ultrasound in Patients at High Risk for Pancreatic Ductal Adenocarcinoma. Pancreas. 2021;50(6):807-14. doi: 10.1097 / MPA.0000000000001835. PubMed PMID: 34149034; PMCID: PMC8577312. 39. Garcia-Albeniz X, Hsu J, Bretthauer M, Hernan MA. Effectiveness of Screening Colonoscopy to Prevent Colorectal Cancer Among Medicare Beneficiaries Aged 70 to 79 Years: A Prospective Observational Study. Ann Intern Med. 2017;166(1):18-26. Epub 20160927. doi: 10.7326 / M16-0758. PubMed PMID: 27669524; PMCID: PMC5417337. 40. Nishihara R, Wu K, Lochhead P, Morikawa T, Liao X, Qian ZR, Inamura K, Kim SA, Kuchiba A, Yamauchi M, Imamura Y, Willett WC, Rosner BA, Fuchs CS, Giovannucci E, Ogino S, Chan AT. Long-term colorectal-cancer incidence and mortality after lower 58  MGH 2024-124-03 Quarles 125141.04710     endoscopy. N Engl J Med. 2013;369(12):1095-105. doi: 10.1056 / NEJMoa1301969. PubMed PMID: 24047059; PMCID: PMC3840160. 41. Cancer Stat Facts: Colorectal Cancer 2023. 42. Alrashdan FT, Chen JC, Singer A, Avants BW, Yang K, Robinson JT. Wearable wireless power systems for 'ME-BIT' magnetoelectric-powered bio implants. J Neural Eng. 2021;18(4). Epub 20210726. doi: 10.1088 / 1741-2552 / ac1178. PubMed PMID: 34229314; PMCID: PMC8820397. 43. Chu T-F, Rajendran R, Kuznetsova I, Wang G-J. High-power, non-enzymatic glucose biofuel cell based on a nano / micro hybrid-structured Au anode. Journal of Power Sources.2020;453. doi: doi: 10.1016 / j.jpowsour.2020.227844. 44. Nadeau P, El-Damak D, Glettig D, Kong YL, Mo S, Cleveland C, Booth L, Roxhed N, Langer R, Chandrakasan AP, Traverso G. Prolonged energy harvesting for ingestible devices. Nature Biomedical Engineering.2017;1(3). doi: doi: 10.1038 / s41551-016-0022. 45. Zaeimbashi M, Nasrollahpour M, Khalifa A, Romano A, Liang X, Chen H, Sun N, Matyushov A, Lin H, Dong C, Xu Z, Mittal A, Martos-Repath I, Jha G, Mirchandani N, Das D, Onabajo M, Shrivastava A, Cash S, Sun NX. Ultra-compact dual-band smart NEMS magnetoelectric antennas for simultaneous wireless energy harvesting and magnetic field sensing. Nature Communications.2021;12(1). doi: doi: 10.1038 / s41467-021-23256-z. 46. Bock DC, Marschilok AC, Takeuchi KJ, Takeuchi ES. Batteries used to Power Implantable Biomedical Devices. Electrochim Acta. 2012;84. doi: 10.1016 / j.electacta.2012.03.057. PubMed PMID: 24179249; PMCID: PMC3811938. 47. Sato J, Nakamura M, Watanabe O, Yamamura T, Funasaka K, Ohno E, Miyahara R, Kawashima H, Goto H, Hirooka Y. Prospective study of factors important to achieve observation of the entire colon on colon capsule endoscopy. Therap Adv Gastroenterol. 2017;10(1):20-31. Epub 20161107. doi: doi:10.1177 / 1756283X16673556. PubMed PMID: 28286556; PMCID: PMC5330613. 48. Moen S, Vuik FER, Voortman T, Kuipers EJ, Spaander MCW. Predictors of Gastrointestinal Transit Times in Colon Capsule Endoscopy. Clin Transl Gastroenterol. 2022;13(6):e00498. Epub 20220518. doi: doi:10.14309 / ctg.0000000000000498. PubMed PMID: 35584543; PMCID: PMC9236601. 49. Liang CP, Dong J, Ford T, Reddy R, Hosseiny H, Farrokhi H, Beatty M, Singh K, 59  MGH 2024-124-03 Quarles 125141.04710     Osman H, Vuong B, Baldwin G, Grant C, Giddings S, Gora MJ, Rosenberg M, Nishioka N, Tearney G. Optical coherence tomography-guided laser marking with tethered capsule endomicroscopy in unsedated patients. Biomed Opt Express. 2019;10(3):1207-22. Epub 20190212. doi: 10.1364 / BOE.10.001207. PubMed PMID: 30891340; PMCID: PMC6420285. 50. Wang YT, Mohammed SD, Farmer AD, Wang D, Zarate N, Hobson AR, Hellstrom PM, Semler JR, Kuo B, Rao SS, Hasler WL, Camilleri M, Scott SM. Regional gastrointestinal transit and pH studied in 215 healthy volunteers using the wireless motility capsule: influence of age, gender, study country and testing protocol. Aliment Pharmacol Ther. 2015;42(6):761-72. Epub 20150729. doi: doi: 10.1111 / apt.13329. PubMed PMID: 26223837. 51. Diaz Tartera HO, Webb DL, Al-Saffar AK, Halim MA, Lindberg G, Sangfelt P, Hellstrom PM. Validation of SmartPill(®) wireless motility capsule for gastrointestinal transit time: Intra- subject variability, software accuracy and comparison with video capsule endoscopy. Neurogastroenterol Motil. 2017;29(10):1-9. Epub 20170519. doi: 10.1111 / nmo.13107. PubMed PMID: 28524600. 52. Diaz Tartera HO, Webb DL, Al-Saffar AK, Halim MA, Lindberg G, Sangfelt P, Hellstrom PM. Validation of SmartPill((R)) wireless motility capsule for gastrointestinal transit time: Intra- subject variability, software accuracy and comparison with video capsule endoscopy. Neurogastroenterol Motil. 2017;29(10):1-9. Epub 20170519. doi: 2017 / 10 / / 2017, doi: 10.1111 / NMO.13107. PubMed PMID: 28524600. 53. Microsens. MSFET-3330-2 pH sensor [cited 2024 Feb.23]. 54. Peacock M. World's smallest entirely solid state pH sensor 2020 [cited 2024 Feb. 23]. 55. Dietrich PI, Blaicher M, Reuter I, Billah M, Hoose T, Hofmann A, Caer C, Dangel R, Offrein B, Troppenz U, Moehrle M, Freude W, Koos C. In situ 3D nanoprinting of free- form coupling elements for hybrid photonic integration. Nature Photonics.2018;12(4):241- 7. doi: 10.1038 / s41566-018-0133-4. 56. Zhao J, Winetraub Y, Du L, A VANV, Ichimura K, Huang C, SZ AA, Sarin KY, A DLZ. Flexible method for generating needle-shaped beams and its application in optical coherence tomography. Optica. 2022;9(8):859-67. Epub 20220722. doi: doi.org / 10.1364 / OPTICA.456894. PubMed PMID: 37283722; PMCID: PMC10243785. 60  MGH 2024-124-03 Quarles 125141.04710     57. Flynn C, Cao H, Applegate BE, Tkaczyk TS. Fabrication of waveguide directional couplers using 2-photon lithography. Opt Express. 2023;31(16):26323-34. doi: 10.1364 / OE.495363. PubMed PMID: 37710495; PMCID: PMC10544952. 58. Repici A, Badalamenti M, Maselli R, Correale L, Radaelli F, Rondonotti E, Ferrara E, Spadaccini M, Alkandari A, Fugazza A, Anderloni A, Galtieri PA, Pellegatta G, Carrara S, Di Leo M, Craviotto V, Lamonaca L, Lorenzetti R, Andrealli A, Antonelli G, Wallace M, Sharma P, Rosch T, Hassan C. Efficacy of Real-Time Computer-Aided Detection of Colorectal Neoplasia in a Randomized Trial. Gastroenterology.2020;159(2):512-20.e7. doi: 10.1053 / J.GASTRO.2020.04.062. 59. Theisen-Kunde D, Danicke V, Wendt M, Brinkmann R, editors. Temperature dependence of water absorption for wavelengths at 1920 nm and 1940 nm2009; Berlin, Heidelberg: Springer Berlin Heidelberg. 60. Kaperiz KA, Bogachev VY, Minaev VP, Fedorov DA. The 1940 nm, Laser, Are There Any Advantages? European Journal of Vascular and Endovascular Surgery. 2020;60(4):e73-e4. 61. Waimin JF, Nejati S, Jiang H, Qiu J, Wang J, Verma MS, Rahimi R. Smart capsule for non-invasive sampling and studying of the gastrointestinal microbiome. RSC Advances. 2020;10(28):16313-22. doi: 10.1039 / C9RA10986B. 62. Nejati S, Wang J, Heredia-Rivera U, Sedaghat S, Woodhouse I, Johnson JS, Verma M, Rahimi R. Small intestinal sampling capsule for inflammatory bowel disease type detection and management. Lab on a Chip.2022;22(1):57-70. doi: 10.1039 / D1LC00451D. 63. Gorlewicz JL, Battaglia S, Smith BF, Ciuti G, Gerding J, Menciassi A, Obstein KL, Valdastri P, Iii RJW. Wireless Insufflation of the Gastrointestinal Tract. IEEE transactions on bio- medical engineering.2013;60(5):1225-33. doi: 10.1109 / TBME.2012.2230631. 64. Pasricha T, Smith BF, Mitchell VR, Fang B, Brooks ER, Gerding JS, Washington MK, Valdastri P, Obstein KL. Controlled colonic insufflation by a remotely triggered capsule for improved mucosal visualization. Endoscopy.2014;46(7):614-8. Epub 20140520. doi: 10.1055 / s- 0034-1365497. PubMed PMID: 24845802; PMCID: PMC7047728. 65. Gorlewicz JL, Battaglia S, Smith BF, Ciuti G, Gerding J, Menciassi A, Obstein KL, Valdastri P, Webster RJ, 3rd. Wireless insufflation of the gastrointestinal tract. IEEE Trans Biomed Eng. 2013;60(5):1225-33. Epub 20121129. doi: 10.1109 / tbme.2012.2230631. 61  MGH 2024-124-03 Quarles 125141.04710     PubMed PMID: 23212312; PMCID: PMC3850772. 66. Lee JH, Traverso G, Ibarra-Zarate D, Boning DS, Anthony BW. Ex Vivo and In Vivo Imaging Study of Ultrasound Capsule Endoscopy. J Med Device.2020;14(2):021005. Epub 20200318. doi: 10.1115 / 1.4046352. PubMed PMID: 32431763; PMCID: PMC7172867. 67. Woods S, Constandinou T. Engineering Micromechanical Systems for the Next Generation Wireless Capsule Endoscopy. BioMed Research International. 2015;2015:741867. doi: 10.1155 / 2015 / 741867. 68. Quirini M, Webster RJ, Menciassi A, Dario P. Design of a Pill-Sized 12-legged Endoscopic Capsule Robot. Proceedings 2007 IEEE International Conference on Robotics and Automation.2007:1856-62. 69. Glass P, Cheung E, Sitti M. A legged anchoring mechanism for capsule endoscopes using micropatterned adhesives. IEEE Trans Biomed Eng. 2008;55(12):2759-67. doi: 10.1109 / tbme.2008.2002111. PubMed PMID: 19126455. 70. Dixon JA, Burt RW, Rotering RH, McCloskey DW. Endoscopic argon laser photocoagulation of small sessile colonic polyps. Gastrointest Endosc. 1982;28(3):162-5. doi: 10.1016 / s0016-5107(82)73042-1. PubMed PMID: 6890003. 71. Krasner N. Laser therapy in the management of benign and malignant tumours in the colon and rectum. Int J Colorectal Dis. 1989;4(1):2-5. doi: 10.1007 / BF01648540. PubMed PMID: 2708879. 72. Mathus-Vliegen EM, Tytgat GN. Nd:YAG laser photocoagulation in colorectal adenoma. Evaluation of its safety, usefulness, and efficacy. Gastroenterology. 1986;90(6):1865-73. doi: 10.1016 / 0016-5085(86)90254-4. PubMed PMID: 3699405. 73. Mathus-Vliegen EM, Tytgat GN. The potential and limitations of laser photoablation of colorectal adenomas. Gastrointest Endosc. 1991;37(1):9-17. doi: 10.1016 / s0016-5107(91)70613- 5. PubMed PMID: 1848521. 74. Oxenberg J, Hochwald SN, Nurkin S. Ablative therapies for colorectal polyps and malignancy. Biomed Res Int. 2014;2014:986352. Epub 20140626. doi: 10.1155 / 2014 / 986352. PubMed PMID: 25089281; PMCID: PMC4095981. 75. Norberto L, Polese L, Angriman I, Erroi F, Cecchetto A, D'Amico DF. Laser 62  MGH 2024-124-03 Quarles 125141.04710     photoablation of colorectal adenomas: a 12-year experience. Surg Endosc.2005;19(8):1045- 8. Epub 20050519. doi: 10.1007 / s00464-004-2179-2. PubMed PMID: 15942811. 76. Zhi D, Yang T, O'Hagan J, Zhang S, Donnelly RF. Photothermal therapy. J Control Release. 2020;325:52-71. Epub 20200630. doi: 10.1016 / j.jconrel.2020.06.032. PubMed PMID: 32619742. 77. Hiramatsu Y, Utsumi T, Higuchi H, Hayashi J, Horimatsu T, Nikaido M, Nakanishi Y, Shimizu T, Muto M, Seno H. Thermal tissue damage caused by new endoscope model due to light absorption. J Gastroenterol Hepatol.2022;37(9):1801-5. Epub 20220725. doi: 10.1111 / jgh.15963. PubMed PMID: 35861136. 78. Wei Q, Arami H, Santos HA, Zhang H, Li Y, He J, Zhong D, Ling D, Zhou M. Intraoperative Assessment and Photothermal Ablation of the Tumor Margins Using Gold Nanoparticles. Adv Sci (Weinh). 2021;8(5):2002788. Epub 20210118. doi: 10.1002 / advs.202002788. PubMed PMID: 33717843; PMCID: PMC7927626. 79. Di Stefano AFD, Radicioni MM, Vaccani A, Fransioli A, Longo L, Moro L, Repici A. Methylene blue MMX® tablets for chromoendoscopy. Bioavailability, colon staining and safety in healthy volunteers undergoing a full colonoscopy. Contemporary Clinical Trials.2018;71:96- 102. 80. Repici A, Wallace MB, East JE, Sharma P, Ramirez FC, Bruining DH, Young M, Gatof D, Irene Canto M, Marcon N, Cannizzaro R, Kiesslich R, Rutter M, Dekker E, Siersema PD, Spaander M, Kupcinskas L, Jonaitis L, Bisschops R, Radaelli F, Bhandari P, Wilson A, Early D, Gupta N, Vieth M, Lauwers GY, Rossini M, Hassan C. Efficacy of Per- oral Methylene Blue Formulation for Screening Colonoscopy. Gastroenterology. 2019;156(8):2198-207.e1. doi: 10.1053 / j.gastro.2019.02.001. 81. Lo WCY, Uribe-Patarroyo N, Hoebel K, Beaudette K, Villiger M, Nishioka NS, Vakoc BJ, Bouma BE. Balloon catheter-based radiofrequency ablation monitoring in porcine esophagus using optical coherence tomography. Biomed Opt Express. 2019;10(4):2067-89. Epub 20190328. doi: 10.1364 / BOE.10.002067. PubMed PMID: 31086717; PMCID: PMC6484999. 82. Desai S, Gupta S, Copur-Dahi N, Krinsky ML. A prospective randomized study comparing jumbo biopsy forceps to cold snare for the resection of diminutive colorectal polyps. Surg Endosc. 2020;34(3):1206-13. Epub 20190610. doi: 10.1007 / s00464-019- 63  MGH 2024-124-03 Quarles 125141.04710 06874-z. PubMed PMID: 31183796. 83. Lee CK, Shim JJ, Jang JY. Cold snare polypectomy vs. Cold forceps polypectomy using double-biopsy technique for removal of diminutive colorectal polyps: a prospective randomized study. Am J Gastroenterol. 2013;108(10):1593-600. Epub 20130917. doi: 10.1038 / ajg.2013.302. PubMed PMID: 24042189. 84. Srinivasan SS, Alshareef A, Hwang AV, Kang Z, Kuosmanen J, Ishida K, Jenkins J, Liu S, Madani WAM, Lennerz J, Hayward A, Morimoto J, Fitzgerald N, Langer R, Traverso G. RoboCap: Robotic mucus-clearing capsule for enhanced drug delivery in the gastrointestinal tract. Sci Robot. 2022;7(70):eabp9066. Epub 20220928. doi: 10.1126 / scirobotics.abp9066. PubMed PMID: 36170378; PMCID: PMC10034646. 85. The American Society for Gastrointestinal Endoscopy PIVI (Preservation and Incorporation of Valuable Endoscopic Innovations) on real-time endoscopic assessment of the histology of diminutive colorectal polyps. Gastrointestinal Endoscopy. 2011 / 03 / 01;73(3). doi: 10.1016 / j.gie.2011.01.023. 86. Sahafi A, Wang Y, Rasmussen CLM, Bollen P, Baatrup G, Blanes-Vidal V, Herp J, Nadimi ES, Sahafi A, Wang Y, Rasmussen CLM, Bollen P, Baatrup G, Blanes-Vidal V, Herp J, Nadimi ES. Edge artificial intelligence wireless video capsule endoscopy. Scientific Reports 202212:1.2022-08-12;12(1). doi: 10.1038 / s41598-022-17502-7. 87. KVASIR I Proceedings of the 8th ACM on Multimedia Systems Conference. Proceedings of the 8th ACM on Multimedia Systems Conference. 2017. doi: 10.1145 / 3083187.3083212. 88. Biffi C, Antonelli G, Bernhofer S, Hassan C, Hirata D, Iwatate M, Maieron A, Salvagnini P, Cherubini A. REAL-Colon: A dataset for developing real-world AI applications in colonoscopy2024. 89. Jha D, Smedsrud PH, Riegler MA, Halvorsen P, de Lange T, Johansen D, Johansen HD. Kvasir-SEG: A Segmented Polyp Dataset. MultiMedia Modeling. 2020. doi: 10.1007 / 978-3-030- 37734-237. 90. Koulaouzidis A, Iakovidis DK, Yung DE, Rondonotti E, Kopylov U, Plevris JN, Toth E, Eliakim A, Johansson GW, Marlicz W, Mavrogenis G, Nemeth A, Thorlacius H, Tontini GE. KID Project: an internet-based digital video atlas of capsule endoscopy for research purposes. Endoscopy International Open. 2017 / 06;05(06). doi: 10.1055 / s-0043- 64  MGH 2024-124-03 Quarles 125141.04710 105488. 91. Chen T-H, Wang Y-T, Wu C-H, Kuo C-F, Cheng H-T, Huang S-W, Lee C, Chen T-H, Wang Y-T, Wu C-H, Kuo C-F, Cheng H-T, Huang S-W, Lee C. A colonial serrated polyp classification model using white-light ordinary endoscopy images with an artificial intelligence model and TensorFlow chart. BMC Gastroenterology 2024 24:1. 2024-03- 05;24(1). doi: 10.1186 / s12876-024-03181-3. 92. Shen M-H, Huang C-C, Chen Y-T, Tsai Y-J, Liou F-M, Chang S-C, Phan NN, Shen M-H, Huang C-C, Chen Y-T, Tsai Y-J, Liou F-M, Chang S-C, Phan NN. Deep Learning Empowers Endoscopic Detection and Polyps Classification: A Multiple-Hospital Study. Diagnostics 2023, Vol 13, Page 1473. 2023-04-19;13(8). doi: 10.3390 / diagnostics13081473. 93. Ding Z, Shi H, Zhang H, Meng L, Fan M, Han C, Zhang K, Ming F, Xie X, Liu H, Liu J, Lin R, Hou X. Gastroenterologist-Level Identification of Small-Bowel Diseases and Normal Variants by Capsule Endoscopy Using a Deep-Learning Model. Gastroenterology. 2019;157(4):1044-54.e5. doi: 10.1053 / J.GASTRO.2019.06.025. 94. Masutani EM, Bahrami N, Hsiao A. Deep learning single-frame and multiframe super- resolution for cardiac MRI. Radiology. 2020;295(3):552-61. doi: 10.1148 / radiol.2020192173. 95. Alzubaidi L, Fadhel MA, Al-Shamma O, Zhang J, Santamaria J, Duan Y, Oleiwi SR, Alzubaidi L, Fadhel MA, Al-Shamma O, Zhang J, Santamaria J, Duan Y, R. Oleiwi S. Towards a Better Understanding of Transfer Learning for Medical Imaging: A Case Study. Applied Sciences 2020, Vol 10, Page 4523. 2020-06-29;10(13). doi: 10.3390 / app10134523. 96. Pitris C, Thrapp A, Tearney GJ, Pitris C, Thrapp A, Tearney GJ. Multi-spectral analysis of in vivo optical coherence tomography (OCT) images of colon polyps. 2024 / 03 / 12;12830. doi: 10.1117 / 12.3003424. 97. Choi S, Choi S. Biofuel Cells and Biobatteries: Misconceptions, Opportunities, and Challenges. Batteries 2023, Vol 9, Page 119. 2023-02-08;9(2). doi: 10.3390 / batteries9020119. 98. Chu TF, Rajendran R, Kuznetsova I, Wang GJ. High-power, non-enzymatic glucose biofuel cell based on a nano / micro hybrid-structured Au anode. Journal of Power Sources. 65  MGH 2024-124-03 Quarles 125141.04710 2020;453:227844-. doi: 10.1016 / J.JPOWSOUR.2020.227844. 99. Zhou J, Liu C, Yu H, Tang N, Lei C, Zhou J, Liu C, Yu H, Tang N, Lei C. Research Progresses and Application of Biofuel Cells Based on Immobilized Enzymes. Applied Sciences 2023, Vol 13, Page 5917.2023-05-11;13(10). doi: 10.3390 / app13105917. 100. Recent progress in mitochondrial biofuel cells. Journal of Electroanalytical Chemistry.2023 / 12 / 01;950. doi: 10.1016 / j.jelechem.2023.117881. 101. Rezaie M, Rafiee Z, Choi S. A Biobattery Capsule for Ingestible Electronics in the Small Intestine: Biopower Production from Intestinal Fluids Activated Germination of Exoelectrogenic Bacterial Endospores. Advanced Energy Materials.2023 / 01 / 01;13(1). doi: 10.1002 / aenm.202202581. 102. Nadeau P, El-Damak D, Glettig D, Kong YL, Mo S, Cleveland C, Booth L, Roxhed N, Langer R, Chandrakasan AP, Traverso G. Prolonged energy harvesting for ingestible devices. Nature Biomedical Engineering 2017 1:3. 2017;1(3):1-8. doi: 10.1038 / s41551- 016-0022. 103. Ho JS, Yeh AJ, Neofytou E, Kim S, Tanabe Y, Patlolla B, Beygui RE, Poon ASY, Ho JS, Yeh AJ, Neofytou E, Kim S, Tanabe Y, Patlolla B, Beygui RE, Poon ASY. Wireless power transfer to deep-tissue microimplants. Proceedings of the National Academy of Sciences.2014-6- 3;111(22). doi: 10.1073 / pnas.1403002111. 104. Iqbal A, Sura PR, Al-Hasan M, Mabrouk IB, Denidni TA, Iqbal A, Sura PR, Al- Hasan M, Mabrouk IB, Denidni TA. Wireless power transfer system for deep-implanted biomedical devices. Scientific Reports 202212:1.2022-08-11;12(1). doi: 10.1038 / s41598- 022-18000-6. 105. Jeffrey S. Letter from the FDA to the FCC on Radiofrequency Exposure.2019. p.1- 3. 106. Yi X, Zheng W, Cao H, Wang S, Feng X, Yang Z. Wireless Power Transmission for Implantable Medical Devices Using Focused Ultrasound and a Miniaturized 1-3 Piezoelectric Composite Receiving Transducer I IEEE Journals & Magazine I IEEE Xplore. IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control. 2021;68(12). doi: 10.1109 / TUFFC.2021.3103099. 107. Barrett C, Fu H, Theodossiades S. Electrodynamic Wireless Power and Data Transfer for Implanted Medical Devices I IEEE Conference Publication I IEEE Xplore. 66  MGH 2024-124-03 Quarles 125141.04710 2023 IEEE SENSORS.2023. doi: 10.1109 / SENSORS56945.2023.10325056. 108. Hosur S, Kashani Z, Karan SK, Priya S, Kiani M. MagSonic: Hybrid Magnetic- Ultrasonic Wireless Interrogation of Millimeter-Scale Biomedical Implants with Magnetoelectric Transducer I IEEE Journals & Magazine I IEEE Xplore. IEEE Transactions on Biomedical Circuits and Systems. 2024. doi: 10.1109 / TBCAS.2023.3334166. 109. Alrashdan FT, Chen JC, Singer A, Avants BW, Yang K, Robinson JT. Wearable Wireless Power Systems for "ME-BIT" Magnetoelectric-Powered Bio Implants HHS Public Access. J Neural Eng.2022;18(4). doi: 10.1088 / 1741-2552 / ac1178. 110. Zaeimbashi M, Nasrollahpour M, Khalifa A, Romano A, Liang X, Chen H, Sun N, Matyushov A, Lin H, Dong C, Xu Z, Mittal A, Martos-Repath I, Jha G, Mirchandani N, Das D, Onabajo M, Shrivastava A, Cash S, Sun NX. Ultra-compact dual-band smart NEMS magnetoelectric antennas for simultaneous wireless energy harvesting and magnetic field sensing. Nature Communications 202112:1. 2021;12(1):1-11. doi: 10.1038 / s41467-021- 23256-z. 111. Wang J, Memon F, Touma G, Baltsavias S, Jang JH, Chang C, Rasmussen MF, Olcott E, Jeffrey RB, Arbabian A, Khuri-Yakub BT. Capsule ultrasound device: Characterization and testing results I IEEE Conference Publication I IEEE Xplore. 2017 IEEE International Ultrasonics Symposium (IUS). 2017. doi: 10.1109 / ULTSYM.2017.8092071. 112. Lee JH, Traverso G, Ibarra-Zarate D, Boning DS, Anthony BW. Ex Vivo and In Vivo Imaging Study of Ultrasound Capsule Endoscopy. Journal of Medical Devices. 2020 / 06 / 01;14(2). doi: 10.1115 / 1.4046352. 113. Ultrasound Capsule Endoscopy With a Mechanically Scanning Micro-ultrasound: A Porcine Study. Ultrasound in Medicine & Biology. 2020 / 03 / 01;46(3). doi: 10.1016 / j.ultrasmedbio.2019.12.003. 114. Lay HS, Cummins G, Cox BF, Qiu Y, Turcanu MV, McPhillips R, Connor C, Gregson R, Clutton E, Desmulliez MPY, Cochran S. In-Vivo Evaluation of Microultrasound and Thermometric Capsule Endoscopes I IEEE Journals & Magazine I IEEE Xplore. IEEE Transactions on Biomedical Engineering.2019;66(3). doi: 10.1109 / TBME.2018.2852715. 115. Jeong K, Yun G, Choi J, Choi I, Son J, Hwang JY, Ha S, Je M. A Wide-Bandwidth 67  MGH 2024-124-03 Quarles 125141.04710 Ultrasound Receiver and On-Chip Ultrasound Transmitter for Ultrasound Capsule Endoscopy I IEEE Journals & Magazine I IEEE Xplore. IEEE Journal of Solid-State Circuits.2023;58(10). doi: 10.1109 / JSSC.2023.3276130. 116. Liu Y-T, Chou Y-H, Yang Y, Shung KK, Zhou Q, Wu J-X. Low-power, high- precision ultrasound capsule position tracking technology I IEEE Conference Publication I IEEE Xplore. 2023 IEEE International Ultrasonics Symposium (IUS). 2023. doi: 10.1109 / IUS51837.2023.10307068. 117. Memon F, Touma G, Wang J, Baltsavias S, Moini A, Chang C, Rasmussen MF, Nikoozadeh A, Choe JW, Arbabian A, Jeffrey RB, Olcott E, Khuri-Yakub BT. Capsule ultrasound device I IEEE Conference Publication I IEEE Xplore.2015 IEEE International Ultrasonics Symposium (IUS).2015. doi: 10.1109 / ULTSYM.2015.0168. 118. TECHNOLOGY :: INNURVATION - The Leaders In Endoscopy Technology. 119. Rank E, Eggeling M, Nevlacsil S, Kraft J, Vogelbacher F, Maese-Novo A, Muellner P, Sagmeister M, Hainberger R, Drexler W, Nevlacsil S, Muellner P, Maese-Novo A, Eggeling M, Vogelbacher F, Sagmeister M, Kraft J, Rank E, Drexler W, Hainberger R. Multi-channel swept source optical coherence tomography concept based on photonic integrated circuits. Optics Express, Vol 28, Issue 22, pp 32468-32482.2020-10-26;28(22). doi: 10.1364 / OE.404588. 120. Rank EA, Nevlacsil S, Muellner P, Hainberger R, Salas M, Gloor S, Duelk M, Sagmeister M, Kraft J, Leitgeb RA, Drexler W, Rank EA, Nevlacsil S, Muellner P, Hainberger R, Salas M, Gloor S, Duelk M, Sagmeister M, Kraft J, Leitgeb RA, Drexler W. In vivo human retinal swept source optical coherence tomography and angiography at 830 nm with a CMOS compatible photonic integrated circuit. Scientific Reports 202111:1. 2021-10-26;11(1). doi: 10.1038 / s41598- 021-00637-4. 121. Rank EA, Agneter A, Schmoll T, Leitgeb RA, Drexler W. Miniaturizing optical coherence tomography. Translational Biophotonics. 2022 / 03 / 01;4(1-2). doi: 10.1002 / tbio.202100007. 122. Bongers C, Daanen H, Bogerd C, Hopman M, Eijsvogels T, Bongers C, Daanen H, Bogerd C, Hopman M, Eijsvogels T. Validity, Reliability, and Inertia of Four Different Temperature Capsule Systems. Medicine & Science in Sports & Exercise. 2017-08- 14;50(1). doi: 10.1249 / MSS.0000000000001403. 68  MGH 2024-124-03 Quarles 125141.04710 123. Cassilly D, Kantor S, Knight LC, Maurer AH, Fisher RS, Semler J, Parkman HP. Gastric emptying of a non-digestible solid: assessment with simultaneous SmartPill pH and pressure capsule, antroduodenal manometry, gastric emptying scintigraphy. Neurogastroenterology & Motility. 2008 / 04 / 01;20(4). doi: 10.1111 / j.1365- 2982.2007.01061.x. 124. Diaz Tartera HO, Webb DL, Al-Saffar AK, Halim MA, Lindberg G, Sangfelt P, Hellstrom PM. Validation of SmartPill® wireless motility capsule for gastrointestinal transit time: Intra- subject variability, software accuracy and comparison with video capsule endoscopy. Neurogastroenterology & Motility. 2017;29(10):1-9. doi: 10.1111 / NMO.13107. 125. Lay HS, Qiu Y, Al-Rawhani M, Beeley J, Poltarjonoks R, Seetohul V, Cumming D, Cochran S, Cummins G, Desmulliez MPY, Wallace M, Trolier-McKinstry S, McPhillips R, Cox BF, Demore CEM. Progress towards a multi-modal capsule endoscopy device featuring microultrasound imaging I IEEE Conference Publication I IEEE Xplore. 2016 IEEE International Ultrasonics Symposium (IUS). 2016. doi: 10.1109 / ULTSYM.2016.7728692. 126. Gluck N, Shpak B, Brun R, Rosch T, Arber N, Moshkowitz M. A novel prepless X-ray imaging capsule for colon cancer screening. Gut. 2016-03-01;65(3). doi: 10.1136 / gutjnl-2015- 310893. 127. Kimchy Y, Lifshitz R, Lewkowitz S, Bertuccio G, Arber N, Gluck N, Pickhardt PJ, Kimchy Y, Lifshitz R, Lewkowitz S, Bertuccio G, Arber N, Gluck N, Pickhardt PJ. Radiographic capsule- based system for non-cathartic colorectal cancer screening. Abdominal Radiology 201742:5.2017-01-04;42(5). doi: 10.1007 / s00261-016-1026-y. 128. Gelder REv, Venema HW, Serlie IWO, Nio CY, Determann RM, Tipker CA, Vos FM, Glas AS, Bartelsman JFW, Bossuyt PMM, Lameris JS, Stoker J. CT Colonography at Different Radiation Dose Levels: Feasibility of Dose Reduction1. Radiology. 2002-07- 01;224(1). doi: 10.1148 / radiol.2241011126. 129. Check-Cap I Redefining colorectal cancer screening and prevention. 130. Gorlewicz JL, Battaglia S, Smith BF, Ciuti G, Gerding J, Menciassi A, Obstein KL, Valdastri P, Robert J. Webster I. Wireless Insufflation of the Gastrointestinal Tract. IEEE transactions on bio-medical engineering. 2013 / 05;60(5). doi: 69  MGH 2024-124-03 Quarles 125141.04710 10.1109 / TBME.2012.2230631. 131. Pasricha T, Smith BF, Mitchell VR, Fang B, Brooks ER, Gerding JS, Washington MK, Valdastri P, Obstein KL. Controlled colonic insufflation by a remotely triggered capsule for improved mucosal visualization. Endoscopy. 2014 / 07;46(07). doi: 10.1055 / s- 0034-1365497. 132. Woods S, Constandinou T, Woods S, Constandinou T. Engineering Micromechanical Systems for the Next Generation Wireless Capsule Endoscopy. BioMed Research International.2015 / 7 / 15;2015. doi: 10.1155 / 2015 / 741867. 133. Quirini M, Webster RJ, Menciassi A, Dario. P. Design of a Pill-Sized 12-legged Endoscopic Capsule Robot I IEEE Conference Publication I IEEE Xplore. Proceedings 2007 IEEE International Conference on Robotics and Automation. 2007. doi: 10.1109 / ROBOT.2007.363592. 134. Glass P, Cheung E, Sitti M. A Legged Anchoring Mechanism for Capsule Endoscopes Using Micropatterned Adhesives I IEEE Journals & Magazine I IEEE Xplore. IEEE Transactions on Biomedical Engineering. 2008;55(12). doi: 10.1109 / TBME.2008.2002111. 135. Filip D, Yadid-Pecht O, Andrews CN, Mintchev MP. Self-Stabilizing Colonic Capsule Endoscopy: Pilot Study of Acute Canine Models I IEEE Journals & Magazine I IEEE Xplore. IEEE Transactions on Medical Imaging. 2011;30(12). doi: 10.1109 / TMI.2011.2163165. 136. Gora M, Yoo H, Suter MJ, Gallagher KA, Bouma BE, Nishioka NS, Tearney GJ. Optical frequency domain imaging system and catheters for volumetric imaging of the human esophagus. Photonics Lett Pol. 2011;3(4):144-6. doi: 10.4302 / plp.2011.4.06. PubMed PMID: 22924122; PMCID: PMC3425351. 137. Suter MJ, Jillella PA, Vakoc BJ, Halpern EF, Mino-Kenudson M, Lauwers GY, Bouma BE, Nishioka NS, Tearney GJ. Image-guided biopsy in the esophagus through comprehensive optical frequency domain imaging and laser marking: a study in living swine. Gastrointest Endosc. 2010;71(2):346-53. Epub 20091031. doi: 10.1016 / j.gie.2009.07.007. PubMed PMID: 19879573; PMCID: PMC2857407. 138. Suter MJ, Vakoc BJ, Yachimski PS, Shishkov M, Lauwers GY, Mino-Kenudson M, Bouma BE, Nishioka NS, Tearney GJ. Comprehensive microscopy of the esophagus in 70  MGH 2024-124-03 Quarles 125141.04710 human patients with optical frequency domain imaging. Gastrointest Endosc. 2008;68(4):745-53. doi: 10.1016 / j.gie.2008.05.014. PubMed PMID: 18926183; PMCID: PMC2715833. 139. Wolfsen HC, Sharma P, Wallace MB, Leggett C, Tearney G, Wang KK. Safety and feasibility of volumetric laser endomicroscopy in patients with Barrett's esophagus (with videos). Gastrointest Endosc. 2015;82(4):631-40. Epub 20150506. doi: 10.1016 / j.gie.2015.03.1968. PubMed PMID: 25956472. 140. Active delivery of multi-layer drug-loaded microneedle patches using magnetically driven capsule. Medical Engineering & Physics. 2020 / 11 / 01;85. doi: 10.1016 / j.medengphy.2020.09.012. 141. Zhang X, Chen G, Fu X, Wang Y, Zhao Y. Magneto-Responsive Microneedle Robots for Intestinal Macromolecule Delivery. Advanced Materials. 2021 / 11 / 01;33(44). doi: 10.1002 / adma.202104932. 142. Chen W, Sui J, Wang C. Magnetically Actuated Capsule Robots: A Review I IEEE Journals & Magazine I IEEE Xplore. IEEE Access. 2022;10. doi: 10.1109 / ACCESS.2022.3197632. 143. Chen W, Wainer J, Ryoo SW, Qi X, Chang R, Li J, Lee SH, Min S, Wentworth A, Collins JE, Tamang S, Ishida K, Hayward A, Langer R, Traverso G. Dynamic omnidirectional adhesive microneedle system for oral macromolecular drug delivery. Sci Adv.2022;8(1):eabk1792. Epub 20220105. doi: 10.1126 / sciadv.abk1792. PubMed PMID: 34985942; PMCID: PMC8730401. 144. Tortora G, Orsini B, Pecile P, Menciassi A, Fusi F, Romano G. An Ingestible Capsule for the Photodynamic Therapy of Helicobacter Pylori Infection I IEEE Journals & Magazine I IEEE Xplore. IEEE / ASME Transactions on Mechatronics. 2016;21(4). doi: 10.1109 / TMECH.2016.2536944. 145. Photodynamic Therapy Using Endoscopy Capsule Robot. Academic Journal of Life Sciences.2020;6(7). doi: 10.32861 / ajls.67.93.100. 146. Kiesslich R, Fritsch J, Holtmann M, Koehler HH, Stolte M, Kanzler S, Nafe B, Jung M, Galle PR, Neurath MF. Methylene blue-aided chromoendoscopy for the detection of intraepithelial neoplasia and colon cancer in ulcerative colitis. Gastroenterology. 2003;124(4):880-8. doi: 10.1053 / gast.2003.50146. PubMed PMID: 12671882. 71  MGH 2024-124-03 Quarles 125141.04710

[0251] It will be appreciated by those skilled in the art that while the disclosed subject matter is described above in connection with particular embodiments and examples, the invention is not necessarily so limited, and that numerous other embodiments, examples, uses, modifications and departures from the embodiments, examples and uses are intended to be encompassed by the claims attached hereto. Each reference cited herein is incorporated by reference in its entirety. 72

Claims

MGH 2024-124-03 Quarles 125141.04710     CLAIMS What is claimed is:

1. An apparatus for detecting and treating gastrointestinal pathology, comprising: a wireless capsule comprising at least one of: an optical imaging system, a thermal imaging system, or a camera, wherein the camera is a narrow band imaging system, a white light imaging system, or both.

2. The apparatus of claim 1, wherein the optical imaging system comprises an optical coherence tomography (OCT) imaging system.

3. The apparatus of claim 2, wherein the OCT imaging system comprises: a wavelength-swept laser source, an OCT detector, and a rotating reflector, wherein the rotating reflector directs light from the laser source in a circumferential pattern around an outer perimeter of the wireless capsule to a sample, and wherein the rotating reflector directs light reflected from the sample to the OCT detector.

4. The apparatus of claim 3, wherein the thermal imaging system comprises a thermistor array disposed in a circumferential pattern around the outer perimeter of the wireless capsule, and wherein the thermistor array is configured to obtain temperature information from the sample surrounding the wireless capsule. 73  MGH 2024-124-03 Quarles 125141.04710     5. The apparatus of claim 4, further comprising a controller in communication with at least one of a laser source, the OCT detector, and the thermistor array, wherein the controller is configured to at least one of: obtain information from the OCT detector to generate optical image data, obtain information from the thermistor array to generate thermal image data, or obtain information from the camera to generate white light image data or narrow band imaging data, and use a machine learning algorithm to analyze at least one of the optical image data, the thermal image data, the white light image data, or narrow band imaging data to diagnose a condition in a portion of the sample.

6. The apparatus of claim 5, further comprising an ablation laser, wherein the controller is in communication with the ablation laser, and wherein the rotating reflector directs light from the ablation laser toward the sample.

7. The apparatus of claim 6, further comprising a micromotor coupled to the rotating reflector and in communication with the controller, wherein the micromotor is configured to rotate the rotating the reflector.

8. The apparatus of claim 7, wherein the controller is further configured to: control the micromotor to direct the ablation laser at the portion of the sample to ablate tissue associated with the diagnosed condition.

9. The apparatus of any one of claims 5-8, further comprising optical components to direct light between the rotating reflector and at least one of the wavelength-swept laser source, the OCT detector, or the ablation laser. 74  MGH 2024-124-03 Quarles 125141.04710     10. The apparatus of claim 9, wherein the optical components comprise nano-printed optics.

11. The apparatus of any one of claims 3-10, wherein the wavelength-swept laser source comprises a wavelength-swept vertical cavity surface emitting laser.

12. The apparatus of any one of claims 6-11, further comprising a pH sensor associated with the wireless capsule and in communication with the controller, wherein the controller is further configured to: sense a change in pH adjacent to the wireless capsule based on the pH sensor, and determine the approximate location of the capsule, activate at least one of the OCT imaging system, the thermal imaging system, or the ablation laser based on sensing the change in pH.

13. The apparatus of any one of claims 6-12, further comprising a temperature sensor associated with the wireless capsule and in communication with the controller, wherein the controller is further configured to: sense an increase in temperature adjacent to the wireless capsule based on the temperature sensor, and place the controller into a shutdown state based on sensing the increase in temperature above a predetermined level.

14. The apparatus of any one of the preceding claims, further comprising one or more reactant wells configured to release reactants adjacent to the wireless capsule, wherein the reactants combine outside the wireless capsule to generate a gas or a polymer.

15. The apparatus of any one of the preceding claims, further comprising a power source to power at least one of the optical imaging system or the thermal imaging system. 75  MGH 2024-124-03 Quarles 125141.04710     16. The apparatus of claim 15, wherein the power source comprises at least one of a battery or a supercapacitor.

17. The apparatus of any one of the preceding claims, further comprising at least one of an LED illuminated white light imaging system or the narrow band imaging system on at least one end of the wireless capsule.

18. The apparatus of any one of the preceding claims, further comprising a thermally- conductive coating on an outside portion of the wireless capsule.

19. The apparatus of any one of the preceding claims, further comprising a biocompatible coating on an outside portion of the wireless capsule.

20. The apparatus of claim 19, wherein the biocompatible coating is applied by dip- coating and curing to provide a biocompatible outer shell to the wireless capsule.

21. The apparatus of claim 5, wherein, when generating thermal image data, the controller is further configured to generate at least one of a two-dimensional map of the sample or a three-dimensional image of the sample.

22. A method for detecting and treating gastrointestinal pathology, comprising: contacting a sample with a wireless capsule, the wireless capsule comprising at least one of: an optical imaging system, a thermal imaging system, or a camera, wherein the camera is a narrow band imaging system, a white light imaging system, or both.

23. The method of claim 22, wherein the optical imaging system comprises an optical coherence tomography (OCT) imaging system, and 76  MGH 2024-124-03 Quarles 125141.04710     wherein the method further comprises: obtaining OCT image data from the sample using the OCT imaging system.

24. The method of claim 23, wherein the OCT imaging system comprises: a wavelength-swept laser source, an OCT detector, and a rotating reflector, and wherein the obtaining OCT image data further comprises: directing, using the rotating reflector, light from the laser source in a circumferential pattern around an outer perimeter of the wireless capsule to a sample, and directing, using the rotating reflector, light reflected from the sample to the OCT detector.

25. The method of claim 24, wherein the thermal imaging system comprises a thermistor array disposed in a circumferential pattern around the outer perimeter of the wireless capsule, and wherein the method further comprises: obtaining, using the thermistor array, temperature information from the sample surrounding the wireless capsule.

26. The method of claim 25, wherein the wireless capsule further comprises a controller in communication with the laser source, the OCT detector, and the thermistor array, wherein the method further comprises at least one of: obtaining, using the controller, information from the OCT detector to generate optical image data, obtaining, using the controller, information from the thermistor array to generate thermal image data, obtaining, using the controller, information from the white light imaging system to generate white light data, or 77  MGH 2024-124-03 Quarles 125141.04710     obtaining, using the controller, information from the narrow band imaging system to generate narrow band image data, and directing, using the controller, a machine learning algorithm to analyze at least one of the optical image data, the thermal image data, the white light image data, or the narrow band image data to diagnose a condition in a portion of the sample.

27. The method of claim 26, wherein the wireless capsule further comprises an ablation laser, and wherein the method further comprises: directing, using the rotating reflector, light from the ablation laser toward the sample.

28. The method of claim 27, wherein the wireless capsule further comprises a micromotor coupled to the rotating reflector and in communication with the controller, wherein the method further comprises: rotating, using the micromotor, the rotating the reflector.

29. The method of claim 28, wherein directing light from the ablation laser toward the sample further comprises: controlling, using the controller, the micromotor to direct the ablation laser at the portion of the sample to ablate tissue associated with the diagnosed condition.

30. The method of any one of claims 26-29, wherein the wireless capsule further comprises optical components disposed adjacent to the wavelength-swept laser source, the OCT detector, and the ablation laser, and wherein the method further comprises: directing, using the optical components, light between the rotating reflector and at least one of the wavelength-swept laser source, the OCT detector, or the ablation laser. 78  MGH 2024-124-03 Quarles 125141.04710     31. The method of claim 30, wherein the optical components comprise nano-printed optics.

32. The method of any one of claims 24-31, wherein the wavelength-swept laser source comprises a wavelength-swept vertical cavity surface emitting laser.

33. The method of any one of claims 27-32, wherein the wireless capsule further comprises a pH sensor in communication with the controller, wherein the method further comprises: sensing, using the controller, a change in pH adjacent to the wireless capsule based on the pH sensor, and activating, using the controller, at least one of the OCT imaging system, the thermal imaging system, or the ablation laser based on sensing the change in pH.

34. The method of any one of claims 27-33, wherein the wireless capsule further comprises a temperature sensor in communication with the controller, wherein the method further comprises: sensing, using the controller, an increase in temperature adjacent to the wireless capsule based on the temperature sensor, and placing, using the controller, the controller into a shutdown state based on sensing the increase in temperature above a predetermined level.

35. The method of any one of claims 22-34, wherein the wireless capsule further comprises one or more reactant wells, and wherein the method further comprises: releasing, using the reactant wells, reactants adjacent to the wireless capsule, wherein the reactants combine outside the wireless capsule to generate a gas or a polymer. 79  MGH 2024-124-03 Quarles 125141.04710     36. The method of any one of claims 22-35, wherein the wireless capsule further comprises a power source, and wherein the method further comprises: powering, using the power source, at least one of the optical imaging system or the thermal imaging system.

37. The method of claim 36, wherein the power source comprises at least one of a battery or a supercapacitor.

38. The method of any one of claims 22-37, wherein the wireless capsule further comprises an LED illuminated white light imaging system or the narrow band imaging system on at least one end of the wireless capsule.

39. The method of any one of claims 22-38, wherein the wireless capsule further comprises a thermally-conductive coating on an outside portion thereof.

40. The method of any one of claims 22-39, wherein the wireless capsule further comprises a biocompatible coating on an outside portion thereof.

41. The method of claim 40, wherein the biocompatible coating is applied by dip-coating and curing to provide a biocompatible outer shell to the wireless capsule.

42. The method of claim 26, wherein generating thermal image data further comprises generating at least one of a two-dimensional map of the sample or a three-dimensional image of the sample. 80

Images