Digestive tract detection drug delivery system

Abstract

The invention relates to the technical field of alimentary canal diagnosis and treatment, in particular to an alimentary canal detection drug delivery system which comprises an alimentary canal sensing unit for collecting an image, a pH value, a temperature and pathological tissue pathological characteristic data in an alimentary canal cavity; the lesion positioning analysis unit identifies a lesion area through an image segmentation algorithm, constructs a lesion three-dimensional coordinate model according to physiological parameters, prejudges a lesion displacement trend in combination with an alimentary canal peristalsis rule, analyzes a lesion state in stages, and outputs lesion severity; the response type drug delivery module controls the micro drug delivery device to perform targeted release and adjust the drug release rate based on the lesion coordinates and the intracavity environment parameters; the data collaboration platform shares a detection result and a drug administration record through a real-time interaction interface; the curative effect feedback unit is used for collecting physiological index change and patient symptom improvement data after drug administration, judging the drug onset period and possible adverse reactions, and optimizing drug administration parameters and detection frequency. Therefore, the problems of low targeting precision, poor diagnosis and treatment efficiency and the like in the prior art are solved.

Description

Technical Field

[0001] This invention relates to the field of gastrointestinal diagnosis and treatment technology, specifically a gastrointestinal detection and drug delivery system. Background Technology

[0002] As a crucial facility for ensuring accurate diagnosis and treatment of digestive tract diseases, gastrointestinal monitoring and drug delivery systems are susceptible to problems during clinical application. These issues include wide pH fluctuations within the digestive tract (gastric pH 1-3, intestinal pH 7-8), continuous peristalsis, and dynamic changes in physiological indicators (such as real-time fluctuations in inflammatory factor concentrations). These factors can lead to data lag and distortion, poor drug targeting, and insufficient dosage adaptation, resulting in weak diagnostic and therapeutic synergy, increased risk of drug side effects, and even impacting the treatment outcomes of diseases such as gastric ulcers and inflammatory bowel disease. With the deepening development of precision medicine and personalized treatment concepts, higher demands are placed on gastrointestinal monitoring and drug delivery systems in terms of real-time physiological indicator sensing, dynamic drug regulation, adaptability to complex environments, and integrated diagnostic and therapeutic synergy. There is an urgent need for an integrated system capable of capturing the physiological state of the digestive tract from multiple dimensions, dynamically optimizing drug delivery protocols, and stably adapting to complex digestive tract environments to ensure accuracy and effectiveness in the diagnosis and treatment of diverse digestive tract diseases such as gastritis and enteritis.

[0003] However, traditional gastrointestinal detection and drug delivery technologies have inherent limitations: Their regulation strategies are rigid, relying solely on preset times or single indicators (such as fixed-dose oral administration or single pH threshold triggering) for control. They fail to integrate multi-dimensional physiological information such as real-time inflammatory factor concentrations and mucosal temperature, making it difficult to adapt to the dynamic changes in the gastrointestinal tract's physiological state. This results in poor synergy between detection and drug delivery, and low drug targeting accuracy. Furthermore, the hardware is insufficiently adaptable to the harsh environment of the gastrointestinal tract. Traditional devices are susceptible to gastric acid corrosion and intestinal peristalsis, leading to sensor signal drift and drug release mechanism malfunctions, significantly reducing detection accuracy and drug delivery precision. Response speed; the drug management model is crude, passive release (such as the disintegration of oral tablets) cannot adjust the dosage as needed, and active regulation lacks a dynamic adaptation mechanism, which not only leads to low drug utilization (the drug concentration in the lesion area is only 10%-20% of the total dose), but also easily causes drug overdose to irritate normal mucosa; and there is no redundant design, a single failure of the detection module or the drug delivery module will cause the entire system to fail, and the average annual failure will affect the diagnosis and treatment time for a long time, far exceeding the stringent requirements of precision medicine for the continuity of gastrointestinal diagnosis and treatment. The overall technology faces multiple challenges such as low targeting accuracy, insufficient dynamic coordination and poor diagnosis and treatment efficiency. Summary of the Invention

[0004] This application provides a digestive tract detection and drug delivery system to solve the problems of low targeting accuracy and poor diagnostic and treatment efficiency in the prior art.

[0005] The first aspect of this application provides a digestive tract detection and drug delivery system, comprising: a digestive tract sensing unit, a lesion localization and analysis unit, a responsive drug delivery module, a data collaboration platform, and a efficacy feedback unit; wherein, the digestive tract sensing unit is used to collect images, pH values, temperature, and pathological feature data of lesions within the digestive tract lumen; the lesion localization and analysis unit is used to identify lesion areas through image segmentation algorithms, construct a three-dimensional coordinate model of the lesion based on physiological parameters, predict the displacement trend of the lesion in conjunction with the peristaltic pattern of the digestive tract, analyze the lesion status in stages, and output the severity of the lesion; the responsive drug delivery module is used to control a micro-drug delivery device to perform targeted release based on the lesion coordinates and intraluminal environmental parameters, combined with the severity of the lesion, and adjust the drug release rate; the data collaboration platform is used to share detection results and drug delivery records through a real-time interactive interface; the efficacy feedback unit is used to collect data on changes in physiological indicators and improvement of patient symptoms after drug delivery, determine the drug onset period and possible adverse reactions, and optimize drug delivery parameters and detection frequency.

[0006] Preferably, the digestive tract sensing unit includes an active imaging module, a non-invasive surface monitoring module, and a pathological feature acquisition module. The active imaging module adjusts the lens angle wirelessly to monitor traditional blind spots and generate high-definition image sequences of the lesion area. The non-invasive surface monitoring module collects gastric electrical activity, gastric pH, and abdominal temperature in real time. The pathological feature acquisition module obtains lesion tissue samples through retractable biopsy forceps and analyzes tissue water content and protein concentration in real time using a near-infrared spectral sensor to preliminarily determine the nature of the lesion.

[0007] Preferably, the lesion localization analysis unit includes an image segmentation module, a three-dimensional coordinate localization module, a lesion trend prediction unit, and a lesion grading engine. The image segmentation module uses an improved U-Net deep learning algorithm to segment the digestive tract image and automatically identify lesion areas such as ulcers, polyps, and bleeding points. The three-dimensional coordinate localization module constructs a three-dimensional coordinate model of the lesion based on the image pixel coordinates of the lesion area and the spatial distribution differences of digestive tract physiological parameters to locate the lesion coordinates. The lesion trend prediction unit uses a fluid dynamics model to simulate the displacement trajectory of the lesion with digestive tract peristalsis based on digestive tract peristalsis data, predicting the displacement trend of the lesion within a preset time. The lesion grading engine outputs a lesion severity grade based on pathological feature data and clinical standards, and associates it with a corresponding drug administration regimen library.

[0008] Preferably, the responsive drug delivery module includes a target positioning drive module and a drug release control module. The target positioning drive module is used to receive three-dimensional coordinate model data of the lesion and drive the micro-drug delivery device to move to the lesion area through magnetic control. The drug release control module dynamically adjusts the drug release rate according to the environmental parameters of the digestive tract lumen, and adjusts the single drug delivery dose according to the severity of the lesion.

[0009] Preferably, the data collaboration platform includes a visual interactive interface, a permission management unit, a data sharing module, and a real-time communication unit. The visual interactive interface synchronously displays the drug administration timeline and the curves showing changes in intracavitary environmental parameters, allowing users to view detailed information about lesions. The permission management unit assigns different operating permissions based on user roles: doctors have permissions to annotate lesions and edit drug administration plans; patients can only view their own test results and drug administration records; and nurses have permissions to input patient medication feedback data. The data sharing module connects to the hospital's electronic medical record system and pathology analysis system, linking and sharing test data, drug administration data, and medical record data. The real-time communication unit enables multi-terminal online interaction; doctors can annotate key areas of lesions and add drug administration guidance notes on the interface, and nurses can upload real-time patient feedback information after medication administration.

[0010] Preferably, the efficacy feedback unit includes an indicator acquisition module, an onset period determination unit, an adverse reaction monitoring unit, and a parameter optimization unit. The indicator acquisition module collects data on the pH value in the digestive tract, inflammation-related indicators, and patient symptom scores at multiple time points after drug administration. The onset period determination unit determines whether the drug is effective and records the onset time by comparing changes in indicators before and after drug administration. The adverse reaction monitoring unit combines patient feedback information with changes in physiological indicators to determine whether adverse reactions caused by drug stimulation exist; mild adverse reactions are recorded and continuously observed, while severe adverse reactions trigger a drug discontinuation prompt. The parameter optimization unit adjusts the dosage for patients with slow drug onset and adjusts the drug release rate for patients experiencing adverse reactions based on the efficacy feedback data, while also appropriately adjusting the detection frequency.

[0011] A second aspect of this application provides a method for administering medication for gastrointestinal detection, comprising: acquiring images, pH values, temperature, and pathological feature data of lesions within the gastrointestinal lumen; The system analyzes and processes images, pH values, temperatures, and pathological features of lesions within the digestive tract lumen. Lesion areas are identified using image segmentation algorithms. A three-dimensional coordinate model of the lesion is constructed based on physiological parameters. Combined with the peristaltic patterns of the digestive tract, the system predicts lesion displacement trends, analyzes lesion status in stages, and outputs the severity of the lesion. Based on the lesion coordinates and intraluminal environmental parameters, and considering the severity of the lesion, a micro-drug delivery device is controlled for targeted release, and the dosage and release rate are adjusted. A visual interactive interface synchronously displays the drug delivery timeline, intraluminal environmental parameter change curves, and the adjusted dosage and release rate. It connects to the hospital's electronic medical record system and pathology analysis system to share detection data, drug delivery data, and medical record data, allowing doctors to mark key lesion areas and add medication guidance notes. Simultaneously, it uploads patient feedback information after medication. Based on this patient feedback information, the system compares changes in indicators before and after medication to determine the drug's onset period. It monitors for adverse reactions by combining physiological indicator changes, and adjusts the dosage, release rate, and detection frequency based on efficacy feedback data.

[0012] A third aspect of this application provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement a digestive tract detection and drug delivery method as described in the above embodiments.

[0013] A fourth aspect of this application provides a computer-readable storage medium having a computer program stored thereon, which is executed by a processor to implement a digestive tract detection and drug delivery method as described in the above embodiments.

[0014] A fifth aspect of this application provides a computer program product, including a computer program or instructions, for implementing a digestive tract detection and drug delivery method as described in the above embodiments.

[0015] Therefore, this application has the following beneficial effects: This application's embodiments, through a digestive tract sensing unit combined with active imaging, non-invasive monitoring, and pathological acquisition, overcome the limitations of traditional detection blind spots, achieving comprehensive and accurate acquisition of intraluminal images, physiological parameters, and pathological features. The lesion localization analysis unit, relying on an improved U-Net algorithm and fluid dynamics model, accurately identifies lesion areas and predicts displacement trends, providing precise coordinates for targeted drug delivery. The responsive drug delivery module, through magnetic control and dynamic dose adjustment, solves the problems of poor targeting and fixed dosage in traditional drug delivery, achieving precise and appropriate drug delivery to the lesion area. The data collaboration platform breaks down data barriers between multiple systems, supports real-time interaction among multiple roles, and improves the efficiency of diagnostic and treatment collaboration. The efficacy feedback unit, through multi-node indicator acquisition and dynamic parameter optimization, can promptly determine the drug's onset period and adverse reactions, and can specifically adjust the dosing regimen and detection frequency, effectively reducing medication risks and improving treatment efficacy. Thus, it solves the problems of low targeting accuracy and poor diagnostic and treatment efficiency in existing technologies.

[0016] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0017] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a schematic diagram of a digestive tract detection and drug delivery system provided according to an embodiment of this application; Figure 2 This is a schematic diagram of a digestive tract sensing unit according to an embodiment of this application; Figure 3 This is a schematic diagram of a lesion localization analysis unit provided according to an embodiment of this application; Figure 4 This is a schematic diagram of a responsive drug delivery module according to an embodiment of this application; Figure 5 This is a schematic diagram of a data collaboration platform provided according to an embodiment of this application; Figure 6 This is a schematic diagram of a therapeutic feedback unit provided according to an embodiment of this application; Figure 7 This is a schematic diagram of a digestive tract detection and drug delivery system provided according to an embodiment of this application; Figure 8 This is a flowchart of a digestive tract detection and drug delivery method according to an embodiment of this application; Figure 9 This is a schematic diagram of a gastrointestinal detection and drug administration method according to an embodiment of this application; Figure 10This is a schematic diagram of the structure of an electronic device provided according to an embodiment of this application. Detailed Implementation

[0018] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] The following description, with reference to the accompanying drawings, illustrates a digestive tract detection and drug delivery system according to an embodiment of this application. Addressing the issue of low targeting accuracy mentioned in the background section, this application provides a digestive tract detection and drug delivery system. In this system, a digestive tract sensing unit, combined with active imaging, non-invasive monitoring, and pathological data acquisition, overcomes the limitations of traditional detection blind spots, achieving comprehensive and accurate acquisition of intraluminal images, physiological parameters, and pathological features. A lesion localization analysis unit, relying on an improved U-Net algorithm and fluid dynamics model, accurately identifies lesion areas and predicts displacement trends, providing precise coordinates for targeted drug delivery. A responsive drug delivery module, through magnetic control and dynamic dose adjustment, solves the problems of poor targeting and fixed dosage in traditional drug delivery, achieving precise and appropriate drug delivery to the lesion area. A data collaboration platform breaks down data barriers between multiple systems, supports real-time interaction among multiple roles, and improves the efficiency of diagnostic and treatment collaboration. A efficacy feedback unit, through multi-node indicator acquisition and dynamic parameter optimization, can promptly determine the drug's onset period and adverse reactions, and can specifically adjust the drug delivery plan and detection frequency, effectively reducing medication risks and improving treatment efficacy. Thus, it solves the problems of low targeting accuracy and poor diagnostic and treatment efficiency in the prior art.

[0020] Figure 1 This is a schematic diagram of a digestive tract detection and drug delivery system provided in an embodiment of this application.

[0021] This application provides a digestive tract detection and drug delivery system, the system 10 comprising: The digestive tract sensing unit 100, the lesion localization analysis unit 200, the responsive drug delivery module 300, the data collaboration platform 400, and the efficacy feedback unit 500 are all included.

[0022] The digestive tract sensing unit 100 is used to collect images, pH value, temperature, and pathological features of lesion tissue within the digestive tract lumen; the lesion localization and analysis unit 200 is used to identify lesion areas through image segmentation algorithms, construct a three-dimensional coordinate model of the lesion based on physiological parameters, predict the displacement trend of the lesion by combining the peristaltic law of the digestive tract, analyze the lesion status in stages, and output the severity of the lesion; the responsive drug delivery module 300 is used to control the micro-drug delivery device to perform targeted release and adjust the drug release rate based on the lesion coordinates and intraluminal environmental parameters, combined with the severity of the lesion; the data collaboration platform 400 is used to share detection results and drug delivery records through a real-time interactive interface; and the efficacy feedback unit 500 is used to collect data on changes in physiological indicators and improvement of patient symptoms after drug delivery, determine the drug onset period and possible adverse reactions, and optimize drug delivery parameters and detection frequency.

[0023] It is understood that, in this embodiment, the digestive tract sensing unit, combined with active imaging, non-invasive monitoring, and pathological acquisition, overcomes the limitations of traditional detection blind spots, achieving comprehensive and accurate acquisition of intraluminal images, physiological parameters, and pathological features. The lesion localization analysis unit, relying on the improved U-Net algorithm and fluid dynamics model, accurately identifies lesion areas and predicts displacement trends, providing precise coordinates for targeted drug delivery. The responsive drug delivery module, through magnetic control and dynamic dose adjustment, solves the problems of poor targeting and fixed dosage in traditional drug delivery, achieving precise and appropriate drug delivery to the lesion area. The data collaboration platform breaks down data barriers between multiple systems, supports real-time interaction among multiple roles, and improves the efficiency of diagnostic and treatment collaboration. The efficacy feedback unit, through multi-node indicator acquisition and dynamic parameter optimization, can promptly determine the drug's onset period and adverse reactions, and can also adjust the drug delivery plan and detection frequency accordingly, effectively reducing medication risks and improving treatment efficacy. Thus, it solves the problems of low targeting accuracy and poor diagnostic and treatment efficiency in existing technologies.

[0024] In this embodiment of the application, the digestive tract sensing unit 100 includes: as follows Figure 2 As shown, the active imaging module, non-invasive body surface monitoring module, and pathological feature acquisition module are included.

[0025] Among them, the active imaging module adjusts the lens angle wirelessly to monitor traditional blind spots and generate high-definition image sequences of lesion areas; the non-invasive surface monitoring module collects gastric electrical activity, gastric pH value and abdominal temperature in real time; the pathological feature acquisition module obtains lesion tissue samples through retractable biopsy forceps and analyzes tissue water content and protein concentration in real time with near-infrared spectral sensor to preliminarily determine the nature of the lesion.

[0026] It is understood that the active imaging module in this application generates a high-definition image sequence of the lesion area by means of wireless remote angle adjustment, so as to accurately locate the lesion and capture details; the non-invasive surface monitoring module can collect key physiological indicators such as gastric electrical activity, gastric pH value and abdominal temperature in real time, so as to dynamically grasp the functional status of the digestive tract while avoiding invasive operation damage; the pathological feature acquisition module can conveniently obtain lesion tissue samples through retractable biopsy forceps, and combine near-infrared spectral sensor to analyze tissue water content and protein concentration in real time, so as to quickly and preliminarily determine the nature of the lesion, reduce invasive operation in the diagnosis and treatment process, and balance diagnostic accuracy and patient comfort, so as to effectively help the early detection and early assessment of digestive tract diseases.

[0027] For example, taking home monitoring of patients with chronic gastritis as an example, the non-invasive surface monitoring module allows patients to attach a flexible sensor patch to their abdomen. Without invasive procedures such as intubation or puncture, it can collect three key physiological indicators in real time: gastric electrical activity (reflecting the strength of gastric motility and determining whether there is slow gastric peristalsis or rhythm disorder), gastric pH value (monitoring whether there is gastric acid reflux or abnormal gastric acid secretion), and abdominal temperature (helping to determine whether there is a local inflammatory response). The data is wirelessly transmitted to the patient's mobile APP and the doctor's terminal. The doctor can remotely and dynamically track the patient's daily gastric function status and promptly detect abnormalities that are easily missed during routine visits, such as nighttime gastric acid reflux and insufficient gastric motility after meals. This not only avoids the discomfort caused by invasive monitoring to patients, but also provides continuous and accurate daily data support for the precise adjustment of gastric disease treatment plans.

[0028] In this embodiment of the application, the lesion localization analysis unit 200 includes: as follows Figure 3 As shown, the system includes an image segmentation module, a 3D coordinate positioning module, a lesion trend prediction unit, and a lesion grading engine.

[0029] The image segmentation module uses an improved U-Net deep learning algorithm to segment digestive tract images and automatically identify lesion areas such as ulcers, polyps, and bleeding points. The three-dimensional coordinate localization module is used to construct a three-dimensional coordinate model of the lesion based on the image pixel coordinates of the lesion area and the spatial distribution differences of digestive tract physiological parameters to locate the lesion coordinates. The lesion trend prediction unit uses a fluid dynamics model to simulate the displacement trajectory of the lesion with digestive tract peristalsis based on digestive tract peristalsis data and predicts the displacement trend of the lesion within a preset time. The lesion grading engine outputs the severity grade of the lesion based on pathological feature data and clinical standards and associates it with the corresponding drug administration regimen library.

[0030] Understandably, the embodiments of this application utilize an improved U-Net deep learning algorithm in the image segmentation module to automatically and accurately identify lesion areas such as ulcers, polyps, and bleeding points, reducing errors and missed diagnoses caused by manual identification. The three-dimensional coordinate positioning module combines the pixel coordinates of the lesion with the spatial differences in physiological parameters of the digestive tract to construct a three-dimensional model, which can clearly pinpoint the specific location of the lesion, solving the problem of ambiguous positioning in traditional methods and providing accurate spatial reference for subsequent examinations or treatments. The lesion trend prediction unit, based on the peristalsis law and fluid dynamics model of the digestive tract, can simulate and predict the displacement trajectory of the lesion within a preset time, helping doctors to grasp the dynamics of the lesion in advance and avoid monitoring or treatment delays caused by lesion movement. The lesion grading engine outputs the severity grade of the lesion based on pathological characteristics and clinical standards, and directly links it to the drug administration regimen library, quickly providing doctors with a basis for graded diagnosis and treatment, shortening the decision-making time for diagnosis and treatment, providing doctors with multi-dimensional and high-precision lesion information support, improving the efficiency of the diagnosis and treatment process, and ensuring the pertinence and timeliness of the treatment plan.

[0031] It should be noted that the formula for the improved U-Net deep learning algorithm is as follows:

[0032]

[0033]

[0034] in, This is the total loss function; These are the weighting coefficients; For Dice's loss; Cross-entropy loss; For edge loss; This is the spatial attention weight matrix; Use the Sigmoid activation function; This is a two-dimensional convolution operation; To perform global average pooling on the input feature map F; To perform global max pooling on the input feature map F; Input feature map; The output feature map of the residual block; This is the input feature map for the residual block; This is a convolution operation; For batch normalization.

[0035] The 3D coordinate localization module first acquires the pixel coordinates of the lesion area on the image. Then, combining this with the unique physiological parameters of different parts of the digestive tract, such as the esophagus, stomach, and intestines (e.g., spatial distribution differences in esophageal diameter, gastric curvature, and intestinal orientation), and utilizing camera imaging principles (pinhole imaging model and internal / external parameter conversion), it maps the 2D pixel coordinates to 3D space. Simultaneously, it applies anatomical constraints and corrections to the 3D spatial coordinates using digestive tract physiological parameters. Ultimately, it constructs a precise 3D coordinate model reflecting the actual location of the lesion within the digestive tract. Once the model is built, the module outputs the specific coordinate values ​​of the lesion in 3D space (usually with a specific reference point as the origin, including values ​​in the X, Y, and Z dimensions). These coordinates accurately indicate the actual location of the lesion within the digestive tract, such as its distance from the esophageal inlet, its location within the stomach wall, and whether it is near important blood vessels or organs, providing precise location guidance for surgical navigation, precise puncture, and other diagnostic and treatment procedures.

[0036] The lesion grading engine first compares pathological feature data with grading indicators in clinical standards one by one. Then, based on the matching results, it determines the severity level of the lesion (common levels such as mild, moderate, and severe, or specific disease stages such as stage I and stage II), and outputs the results in an intuitive format (such as reports and grading labels), providing doctors with direct reference for quickly understanding the severity of the condition. Simultaneously, the engine establishes a connection mechanism with a "medication regimen library": this library is built based on evidence-based medicine (treatment regimens proven effective by numerous clinical studies), clinical guidelines, and the correspondence between lesion grading and treatment effects. For example, it specifies the recommended types and dosages of oral medications for "mild lesions" and the combination therapy or injection therapy regimens required for "severe lesions." When the engine outputs the lesion severity grading, it automatically matches the recommended medication regimens for the corresponding grading in the regimen library and presents them to the doctor. This reduces the time cost for doctors to manually search for regimens and avoids the problem of inconsistent treatment plans due to individual experience differences, ultimately assisting doctors in developing treatment plans more efficiently and scientifically.

[0037] Fluid dynamics model formula:

[0038]

[0039]

[0040] in, For fluid density; For fluid velocity vector; The local time rate of change of fluid velocity; The convection term represents the fluid velocity. For fluid pressure; For pressure gradient; The dynamic viscosity of the fluid; For the Laplace operator of fluid velocity; It is the vector of gravitational acceleration; For the quality of the lesion; The velocity vector of the lesion; This refers to the drag force of the fluid on the lesion; The force of gravity acting on the lesion; The buoyancy force acting on the lesion; The acceleration vector of the lesion; For fluid dynamic viscosity; The equivalent diameter of the lesion; The density of the lesion; The volume of the lesion; These are constant coefficients.

[0041] In this embodiment of the application, the responsive drug delivery module 300 includes: Figure 4 As shown, the target positioning drive module and the drug release control module.

[0042] The targeted positioning drive module receives the three-dimensional coordinate model data of the lesion and moves the micro drug delivery device to the lesion area through magnetic control; the drug release control module dynamically adjusts the drug release rate according to the environmental parameters in the digestive tract lumen, and adjusts the single drug dosage according to the severity of the lesion.

[0043] It is understood that the targeted positioning drive module in this application uses the three-dimensional coordinate model data of the lesion to drive the micro-drug delivery device directly to the lesion area in a magnetically controlled manner. This breaks through the limitations of the wide-area diffusion of traditional drug delivery, accurately locks the treatment target, avoids stimulation and damage to normal tissues of the digestive tract, reduces toxic side effects, and ensures that the drug delivery device accurately reaches the lesion, laying the foundation for the effective action of the drug. The drug release control module achieves dynamic adaptation in terms of both rate and dosage. It adjusts the drug release rate according to the environmental parameters of the digestive tract lumen (such as pH value, peristalsis speed, etc.) to ensure that the drug maintains an effective concentration in the lesion area, avoids premature degradation of the drug or release that is too fast / too slow due to environmental differences, improves drug bioavailability, and adjusts the single-dose dosage according to the severity of the lesion. It avoids overdose for mild lesions and underdose for severe lesions, so that the drug delivery plan fits the actual situation of the lesion and the physiological environment of the human body, and performs refined drug delivery. This not only improves the treatment effect, but also reduces drug waste and the risk of improper dosage, and greatly improves the accuracy, safety and effectiveness of drug delivery for digestive tract lesions.

[0044] The drug release control module achieves precise drug delivery through dynamic regulation in two dimensions: environmental adaptation and disease matching. On the one hand, it captures key environmental parameters in the digestive tract lumen in real time, such as pH value (e.g., strong acidity in the stomach, weak neutrality in the small intestine), intestinal peristalsis speed, and intraluminal enzyme concentration. It flexibly adjusts the drug release rate according to the physiological differences in different parts of the body—for example, slowing down release in the acidic stomach to avoid drug degradation, and accelerating release in the fast-peristing small intestine to ensure full absorption of the drug, ensuring that the drug always maintains an effective concentration in the lesion area. On the other hand, it adjusts the single-dose dosage according to the severity of the lesion (mild, moderate, severe) determined in the early stage. Low doses are used for mild lesions to avoid drug overdose and side effects, high doses are used for severe lesions to prevent insufficient dosage from affecting efficacy, and balanced doses are used for moderate lesions. Ultimately, it achieves individualized drug delivery that conforms to the physiological environment of the digestive tract and the actual condition of the lesion, which improves drug efficacy and reduces drug risks and waste.

[0045] For example, in the treatment of gastric ulcers, when a patient uses an oral micro-drug delivery device equipped with a drug release control module, the module first monitors the patient's gastric environment in real time. If parameters such as a strongly acidic environment with a gastric pH of about 1-2 and slow peristalsis are detected, the module will automatically slow down the drug release rate to prevent the drug from being rapidly degraded by gastric acid, ensuring that the drug can slowly dissolve and accurately adhere to the ulcer lesion. At the same time, the module will retrieve the severity data of the "moderate gastric ulcer" diagnosed earlier and adjust the single-dose dosage accordingly. This will prevent the ulcer from healing slowly due to an insufficient dose, or from causing side effects such as nausea and stomach pain due to an excessively high dose. Ultimately, the drug will maintain an effective therapeutic concentration in the ulcer area, achieving safe and efficient gastric ulcer repair.

[0046] In this embodiment of the application, the data collaboration platform 400 includes, as follows: Figure 5 As shown, it includes a visual interactive interface, a permission management unit, a data sharing module, and a real-time communication unit.

[0047] The system includes a visual interactive interface that synchronously displays the drug administration timeline and the curves showing changes in intracavitary environmental parameters, allowing users to view detailed information about lesions. The access control unit assigns different permissions based on user roles: doctors have permission to annotate lesions and edit medication regimens, patients can only view their own test results and medication records, and nurses have permission to input patient medication feedback data. The data sharing module connects to the hospital's electronic medical record system and pathology analysis system, enabling the sharing of test data, medication data, and medical record data. The real-time communication unit facilitates multi-terminal online interaction, allowing doctors to annotate key areas of lesions and add medication guidance notes on the interface, while nurses can upload real-time patient feedback information after medication administration.

[0048] It is understood that this application embodiment presents the drug administration timeline, intracavitary environment parameter change curves, and lesion details simultaneously through a visual interactive interface, allowing users to intuitively obtain key treatment data without having to search through multiple reports, reducing information acquisition costs. Doctors can quickly combine data to judge the drug administration effect, and patients can clearly understand their treatment progress. The permission management unit assigns permissions according to the different roles of doctors, patients, and nurses, clarifying the operational boundaries of each role. This ensures data security (e.g., patients cannot modify the plan, and others cannot view the patient's private data) and avoids misoperations caused by permission confusion (e.g., nurses are only responsible for entering feedback rather than editing the plan), standardizing the medical data usage process. The data sharing module... By connecting to the hospital's electronic medical record and pathology analysis systems, data silos between different systems are broken down, enabling the sharing of test, medication, and medical record data. Doctors can obtain complete patient data without switching between multiple systems, reducing repetitive data entry while ensuring data consistency and providing comprehensive data support for precision diagnosis and treatment. The real-time communication unit supports multi-terminal online interaction, opening up instant communication channels between doctors and nurses. Doctors can quickly mark key areas of lesions and add medication instructions, while nurses can upload patient medication feedback in real time, avoiding treatment deviations caused by information transmission delays and improving cross-role collaboration efficiency. This not only improves the utilization efficiency of diagnostic and treatment data but also ensures the standardization and safety of medical operations and promotes efficient collaboration among multiple roles.

[0049] For example, in a scenario where medication is administered after surgery for a patient with gastrointestinal polyps, the doctor, using the real-time communication unit of the data collaboration platform, observes the pH change curve within the intestinal lumen transmitted by the micro-drug delivery device. Upon noticing that the drug release rate near the lesion is slightly lower than expected, the doctor immediately marks the key coverage area of ​​the polyp wound on the visualization interface and adds a note: "Increase the release rate by 15% for the next administration to ensure the drug concentration in the wound meets the target." The nurse receives this notification instantly and, before administering the next round of medication, adjusts the parameters of the drug release control module according to the note. Simultaneously, the nurse uploads feedback from the patient after medication via the real-time communication unit, confirming that there is "no abdominal distension and normal bowel movements." The doctor reviews the feedback in real-time, confirming that the current adjustment plan is suitable for the patient's condition and requires no further modification. The entire process eliminates the need for offline handover or waiting for report transmission, achieving instant linkage between doctor guidance, nurse execution, and feedback. This avoids medication deviations caused by information delays and allows doctors to monitor the patient's medication response in real time, ensuring the accuracy and safety of postoperative treatment.

[0050] In this embodiment of the application, the therapeutic effect feedback unit 500 includes, as follows: Figure 6 As shown, the module includes an indicator acquisition module, an onset period determination unit, an adverse reaction monitoring unit, and a parameter optimization unit.

[0051] The system includes several modules: an indicator acquisition module that collects data on pH levels in the digestive tract, inflammation-related indicators, and patient symptom scores at multiple time points after drug administration; an onset time determination unit that compares changes in indicators before and after drug administration to determine whether the drug has taken effect and records the onset time; an adverse reaction monitoring unit that combines patient feedback with changes in physiological indicators to determine whether adverse reactions are caused by drug stimulation, recording and continuously observing mild adverse reactions, and triggering a discontinuation prompt for severe adverse reactions; and a parameter optimization unit that adjusts the dosage for patients with slow drug onset and the drug release rate for patients with adverse reactions based on efficacy feedback data, while also appropriately adjusting the detection frequency.

[0052] It is understood that the indicator acquisition module in this application collects pH values, inflammatory markers, and patient symptom scores in the digestive tract at multiple time points after drug administration, constructing a multi-dimensional, full-cycle efficacy evaluation data foundation to avoid the one-sidedness of single-point data and make subsequent efficacy judgments more objective and comprehensive; the onset time determination unit judges drug efficacy by comparing changes in indicators before and after drug administration and records the onset time, promptly clarifying whether the treatment is effective. If the drug is ineffective, it can be quickly detected and adjusted to avoid delaying treatment; if it is effective, it provides a basis for controlling the subsequent treatment rhythm; the adverse reaction monitoring unit combines patient feedback and... The system categorizes adverse reactions by severity (mild cases are monitored, severe cases trigger discontinuation of medication), accurately identifying and managing medication risks. This approach avoids over-intervention in mild cases that could negatively impact treatment, while also preventing safety incidents through discontinuation alerts for severe cases, thus balancing efficacy and safety. The parameter optimization unit adjusts the dosage, release rate, and detection frequency based on feedback data, providing individualized iterations of the treatment plan. It increases the dosage for patients with slow onset of action and adjusts the release rate for patients experiencing adverse reactions, making the plan more suitable for individual circumstances. At the same time, it adjusts the detection frequency to reduce unnecessary examinations, improving treatment efficiency, patient experience, and the accuracy of medication administration.

[0053] For example, in treating a patient with moderate chronic gastritis, the initial dosing regimen was a single 20mg dose with a 2-hour drug release rate, monitored once daily. After 3 days, the parameter optimization unit received efficacy feedback data: inflammatory markers decreased by only 15% (far below the expected 30%), the patient still experienced significant bloating (indicating slow onset of action), and the patient reported mild nausea after taking the medication (a mild adverse reaction). Based on this, the unit immediately made targeted adjustments: increasing the single dose to 25mg to increase the drug concentration at the lesion site and accelerate inflammation resolution; simultaneously slowing the drug release rate to 3 hours to avoid high drug concentrations irritating the gastric mucosa in a short period and alleviate nausea; and increasing the monitoring frequency to twice daily to track the effects of the adjustments in real time. Two days later, new feedback data showed a 40% decrease in inflammatory markers, relief of bloating, and disappearance of nausea. The parameter optimization unit then fine-tuned again, reducing the monitoring frequency to once every two days, ensuring efficacy monitoring while reducing the patient's testing burden, ultimately achieving efficient and safe treatment tailored to the individual patient's condition.

[0054] This application proposes a digestive tract detection and drug delivery system. By combining a digestive tract sensing unit with active imaging, non-invasive monitoring, and pathological data acquisition, it overcomes the limitations of traditional detection blind spots, achieving comprehensive and accurate acquisition of intraluminal images, physiological parameters, and pathological features. A lesion localization and analysis unit, relying on an improved U-Net algorithm and fluid dynamics model, accurately identifies lesion areas and predicts displacement trends, providing precise coordinates for targeted drug delivery. A responsive drug delivery module, through magnetic control and dynamic dose adjustment, solves the problems of poor targeting and fixed dosage in traditional drug delivery, achieving precise and appropriate drug delivery to the lesion area. A data collaboration platform breaks down data barriers between multiple systems, supporting real-time interaction among multiple roles and improving diagnostic and treatment collaboration efficiency. A efficacy feedback unit, through multi-node indicator acquisition and dynamic parameter optimization, can promptly determine the drug's onset period and adverse reactions, and can specifically adjust the drug delivery plan and detection frequency, effectively reducing medication risks and improving treatment efficacy. Thus, it solves the problems of low targeting accuracy and poor diagnostic and treatment efficiency in existing technologies.

[0055] The following will describe a digestive tract detection and drug delivery system through a specific embodiment, such as... Figure 7 As shown, it includes: Mr. Li, a 45-year-old male patient, presented to the gastroenterology department of a tertiary hospital with a 3-month history of recurrent burning pain in the upper abdomen, which worsened after meals and was accompanied by acid reflux. He had no prior history of gastrointestinal surgery or drug allergies, and his Helicobacter pylori test was positive. Based on the patient's symptoms, the doctor initially suspected gastric mucosal lesions and decided to use a gastrointestinal testing and drug delivery system for precise detection and targeted therapy.

[0056] Data Acquisition Process of the Gastrointestinal Sensing Unit: As the system's data entry point, the gastrointestinal sensing unit, through the collaboration of the active imaging module, non-invasive surface monitoring module, and pathological feature acquisition module, completes multi-dimensional data acquisition of the patient's gastrointestinal environment and lesion characteristics. The acquisition process takes approximately 40 minutes, is entirely non-invasive, and is well-tolerated by the patient. The active imaging module uses an 8mm diameter high-definition miniature wireless camera equipped with a medical-grade 2.4GHz wireless transmission module. The lens angle can be adjusted via an external remote control (360° horizontal rotation, ±45° vertical adjustment), focusing on monitoring blind spots easily missed by traditional endoscopy, such as the gastric angle and antrum. After the doctor activates the camera via the external control console, the patient swallows the capsule-shaped camera with warm water. Once inside the stomach cavity, the camera automatically starts shooting, acquiring 2 frames per second of 1920×1080 resolution images and transmitting them to the system terminal in real time. During the filming process, the doctor successfully captured a mucosal defect area with a diameter of about 12mm in the antrum of the stomach by remotely adjusting the lens angle. The image showed that the edge of the defect area was red and swollen, and the bottom was covered with a small amount of white moss-like material, which was initially suspected to be a gastric ulcer lesion. At the same time, the camera generated a high-definition image sequence of 120 frames containing the lesion, providing basic data for subsequent lesion localization. The non-invasive surface monitoring module consists of three medical conductive gel electrode pads and a portable monitor. The electrode pads are attached to the patient's upper abdomen below the xiphoid process, left subcostal region, and right subcostal region, respectively. The surface electrodes collect gastric electrical activity signals (sampling frequency 256Hz) and simultaneously record the frequency and amplitude of gastric peristalsis. The monitor's built-in wireless pH sensor (placed in the gastric antrum with endoscopy assistance, diameter 3mm) collects the gastric pH value in real time. Data shows that the patient's gastric pH value is 1.2 when fasting (normal range 1.0-3.0), and rises to 2.5 30 minutes after a meal. In addition, the module also collects the patient's abdominal surface temperature via an infrared temperature sensor, maintaining it at 36.8-37.1℃ without abnormal temperature fluctuations. All monitoring data are automatically uploaded to the system database at a frequency of 1 minute / time, forming a continuous physiological parameter curve. The pathological feature acquisition module uses a 5mm diameter retractable biopsy forceps (inserted through the working channel of the endoscope). The doctor manipulates the biopsy forceps to obtain two gastric mucosal tissue samples with a diameter of about 2mm in the lesion area located by the active imaging module. At the same time, the near-infrared spectral sensor (detection wavelength 700-1100nm) mounted on the tip of the biopsy forceps directly contacts the lesion surface to analyze the water content and protein concentration of the tissue in real time. The data shows that the water content of the lesion area is 78% (the water content of normal gastric mucosa is about 72%), and the protein concentration is 18g / 100g (the protein concentration of normal gastric mucosa is about 22g / 100g). Combined with the subsequent pathological section examination of the tissue samples (suggesting chronic inflammation of the mucosa with glandular atrophy), the lesion is preliminarily judged to be a chronic gastric ulcer (active phase).

[0057] The lesion localization analysis unit identifies and grades lesions: After receiving images and data transmitted from the digestive tract sensing unit, the lesion localization analysis unit completes the precise localization, displacement prediction, and severity grading of gastric ulcer lesions through image segmentation, 3D localization, trend prediction, and grading engine processing. The entire analysis process takes approximately 15 minutes, providing core decision-making basis for subsequent drug administration. The image segmentation module adopts an improved U-Net deep learning algorithm, which adds an attention mechanism module to the traditional U-Net architecture, enabling more accurate identification of gastric mucosal defect areas. The system inputs a sequence of 120 high-definition images acquired by the active imaging module into the algorithm model. The algorithm first performs grayscale correction and noise removal on the images, then extracts image features (such as the red and swollen areas at the edge of the lesion and the moss-like material at the bottom) through the encoder, and the decoder restores the contour and extent of the lesion based on the feature mapping. Ultimately, the algorithm automatically identified a 12mm diameter ulcer lesion in the gastric antrum with a localization accuracy of 98%, and generated a binary segmented image of the lesion area (white area represents the lesion, black area represents normal mucosa), while also outputting the pixel coordinates of the lesion (with the upper left corner of the image as the origin, and the lesion center coordinates as (480, 360) pixels). The three-dimensional coordinate localization module, based on the pixel coordinates output by the image segmentation module, combined with physiological parameters (such as gastric wall thickness and gastric volume) collected by the non-invasive surface monitoring module, constructed a three-dimensional coordinate model of the lesion. The system first establishes a mapping relationship between pixel coordinates and actual spatial coordinates (1 pixel corresponds to 0.05 mm) by measuring the actual size of the gastric antrum (anteroposterior diameter of approximately 35 mm) using gastroscopy. Then, combining gastric wall thickness data (4 mm in the lesion area, approximately 2 mm in the normal area) and changes in gastric volume (approximately 50 mL on an empty stomach, approximately 150 mL after a meal), a 3D reconstruction algorithm generates the spatial coordinates of the lesion. With the xiphoid process as the origin, the 3D coordinates of the lesion center are (15 mm, 8 mm, -5 mm) (X-axis for left-right direction, Y-axis for up-down direction, Z-axis for anteroposterior direction). These coordinates accurately reflect the actual location of the lesion within the gastric cavity, providing a spatial positioning benchmark for targeted drug delivery. The lesion trend prediction unit uses data on gastrointestinal peristalsis (gastric electrical activity signals collected by the non-invasive surface monitoring module; the patient's gastric peristalsis frequency is 3 times / minute, with an amplitude of 5 mm) and employs a fluid dynamics model to simulate the lesion's displacement trajectory with gastric peristalsis. The model treats food and liquid within the stomach cavity as a viscous fluid and stomach wall peristalsis as periodic contractions. By calculating the interaction forces between the fluid and the stomach wall, it predicts the displacement of the lesion at different time points. Results show that within one hour after a meal, due to increased peristalsis in the antrum, the lesion will shift approximately 6 mm towards the pylorus, with its three-dimensional coordinates becoming (18 mm, 10 mm, -5 mm). Two hours after a meal, as the stomach contents empty into the small intestine, the lesion displacement gradually stabilizes, returning to near its initial coordinates. The system stores this displacement trend curve in a database, providing a predictive basis for the dynamic positioning of subsequent drug delivery devices.The lesion grading engine combines the detection data (tissue water content, protein concentration, pathological section results) from the pathological feature acquisition module with clinical standards to grade the severity of the lesion. The engine first extracts pathological feature data: lesion diameter 12mm (>10mm), mucosal gland atrophy (moderate), Helicobacter pylori positive, and then compares it with the grading indicators in the clinical standards—mild ulcer (diameter <5mm, no glandular atrophy), moderate ulcer (diameter 5-10mm, mild glandular atrophy), and severe ulcer (diameter >10mm, moderate and above glandular atrophy). Finally, it determines the severity of the patient's gastric ulcer as "severe" and automatically associates it with the system's built-in drug regimen library to retrieve the initial drug regimen corresponding to severe gastric ulcer (proton pump inhibitor, single dose 20mg, twice daily, drug release rate slow release over 2 hours). Targeted drug delivery by the responsive drug delivery module: Based on the three-dimensional coordinates, displacement trends, and grading results output by the lesion localization analysis unit, the responsive drug delivery module achieves precise targeted drug delivery to the gastric ulcer lesion through the synergy of the targeted localization drive module and the drug release control module. The first drug delivery is performed 1 hour after the patient's meal (after the lesion displacement has stabilized), and the drug delivery time is approximately 30 minutes. The targeted localization drive module adopts a combination of "external magnetic control + in vivo capsule drug delivery device". The micro drug delivery device is a capsule-shaped structure with a diameter of 9mm and a length of 25mm, which contains a permanent magnet (neodymium iron boron material, magnetic field strength 0.5T), a drug storage chamber (capacity 50mg), and a wireless receiving module. The external magnetic control system consists of 3 sets of electromagnetic coils (placed on the left, middle, and right sides of the patient's upper abdomen, respectively) and a magnetic control console. By adjusting the coil current intensity (0-5A) and direction, a controllable magnetic field is generated to drive the capsule device to move in the gastric cavity. Before drug administration, the system inputs the three-dimensional coordinates (18mm, 10mm, -5mm) of the lesion predicted 1 hour postprandial by the lesion localization analysis unit into the magnetic control console. The console automatically calculates the current parameters of the electromagnetic coil and generates a magnetic field drive path. The doctor observes the position of the capsule device in real time through an external monitoring screen (the capsule's built-in micro-positioning chip transmits position data) and fine-tunes the magnetic field parameters: First, a magnetic field is generated by energizing the left coil, propelling the capsule from the stomach body to the antrum; when the capsule approaches the lesion area (approximately 10mm from the lesion), the current of the middle coil is adjusted to slow the capsule's movement; finally, with the synergistic effect of the three sets of coils, the capsule device precisely stops directly above the lesion (approximately 2mm from the lesion surface), with a positioning error of less than 1mm, completing targeted localization. The drug release control module receives real-time gastric pH data (at this time, the gastric pH is 2.3) and lesion grading results (severe ulcer) transmitted by the non-invasive surface monitoring module, and dynamically adjusts the drug release parameters. The module first determines the drug release rate based on pH value: Since the stomach is highly acidic, rapid drug release can easily lead to its destruction by gastric acid. Therefore, the module controls the sustained-release membrane (enteric-coated material) of the capsule to release the proton pump inhibitor at a rate of 10 mg per hour, ensuring the drug is slowly released over 2 hours to maintain a stable drug concentration in the lesion area. Simultaneously, based on the grading results of severe ulcers, the module sets the single-dose dose at 25 mg (higher than the initial 20 mg) to ensure an effective therapeutic concentration in the lesion area (the effective concentration of the proton pump inhibitor must be ≥0.5 μg / mL). During drug release, the module monitors the drug concentration in the gastric cavity in real time using a drug concentration sensor built into the capsule. Data shows that after 1 hour of release, the drug concentration in the lesion area reaches 0.8 μg / mL, and after 2 hours it remains at 0.6 μg / mL, meeting the treatment requirements. Simultaneously, the module wirelessly transmits the release progress (e.g., "15 mg released, 10 mg remaining") to a data collaboration platform in real time for remote monitoring by doctors.

[0058] Multi-role interaction and data sharing of the data collaboration platform: As the "information hub" of the system, the data collaboration platform enables collaborative interaction among doctors, nurses, and patients through a visual interactive interface, access control, data sharing, and real-time communication functions. It also facilitates data linkage between the system and the hospital's existing information systems, continuously supporting diagnostic and treatment decisions throughout the entire treatment cycle. The visual interactive interface is deployed on the terminal computer in the doctor's office and the tablet computer at the nurse's station. The left side of the interface displays a real-time drug administration timeline (the first administration time is marked as "D113:00", and the next administration time as "D121:00"). The right side shows a line graph of gastric pH changes (D112:00-14:00, pH rises from 1.8 to 2.5 and then falls to 2.3). Below, a high-resolution image of the lesion captured by an active imaging module is embedded. Clicking on the image allows for magnification to view lesion details (such as the degree of redness and swelling at the edges, and the extent of moss coverage). Doctors can intuitively grasp the patient's drug administration progress and physiological parameter changes through the interface, obtaining complete diagnostic and treatment data without switching between multiple systems. The access control unit assigns operation permissions based on user roles: Doctor accounts have the highest permissions and can mark key areas of lesions on the interface (such as marking bleeding points at the edge of ulcers with red circles) and edit dosing regimens (such as adjusting the dosage for the next administration); Nurse accounts only have data entry permissions and can record the patient's physical feedback (such as "no nausea, abdominal distension, and slight relief of upper abdominal pain symptoms") via tablet computer within 30 minutes after the patient takes medication; Patient accounts log in to the platform through the hospital's APP and can only view their own test results (such as ulcer diameter and pH curve) and medication records (such as "D1 13:00, took 25mg proton pump inhibitor"), and cannot modify any data, ensuring data security and privacy protection. The data sharing module connects to the hospital's Electronic Medical Record (EMR) system and pathology analysis system via the hospital's HL7 data interface. On one hand, it retrieves patients' past medical history (e.g., "Taking nifedipine for hypertension 5 years ago") and allergy history from the EMR system, supplementing the system database to provide a reference for adjusting medication regimens. On the other hand, it synchronizes pathological feature data (e.g., tissue water content, protein concentration) from the digestive tract sensing unit and the grading results from the lesion localization analysis unit to the pathology analysis system for pathologists to review and diagnose (the final review result is consistent with the system grading, indicating severe chronic gastric ulcer). Simultaneously, system-generated medication records (e.g., administration time, dosage, release rate) are automatically uploaded to the EMR system, forming a complete medical record and avoiding duplicate data entry. The real-time communication unit supports multi-terminal online interaction and uses an encrypted instant messaging protocol to ensure secure information transmission.At 11:00 PM, the nurse uploaded patient feedback via tablet: "Upper abdominal pain score decreased from 8 (VAS score) to 5." Upon receiving the message in real time, the doctor marked on the visualization interface: "Drug concentration in the lesion area is within target range; pain relief is effective," and added the instruction note: "Maintain 25mg dose for the next administration, keep release rate unchanged." The nurse immediately received the note and confirmed the subsequent medication plan. The patient sent a question via the app: "Can I eat fruit after taking the medication?" The doctor replied within 10 minutes: "You can eat small amounts of low-acid fruits such as bananas and apples, but avoid citrus fruits," achieving efficient communication between doctor and patient.

[0059] The efficacy feedback unit monitors and optimizes the treatment regimen: Throughout the entire treatment cycle (2 weeks), the efficacy feedback unit dynamically adjusts the dosing regimen through multi-time-point data collection, efficacy assessment, adverse reaction monitoring, and regimen optimization. This ensures maximum treatment effectiveness and minimum risk, with three minor regimen adjustments completed, resulting in significant final efficacy for the patients. The indicator collection module sets five key time points for data collection: 12 hours after administration (D21:00), 24 hours (D21:00), 72 hours (D41:00), 7 days (D81:00), and 14 days (D151:00), with each collection taking approximately 20 minutes. The collected data includes: gastrointestinal pH (obtained by a non-invasive surface monitoring module), inflammation-related indicators (C-reactive protein, CRP detected via finger-prick blood collection), and patient symptom scores (VAS score, covering abdominal pain, acid reflux, and bloating). Data collected at D213:00 showed that the gastric pH rose to 3.5 (the drug effectively inhibited gastric acid secretion), CRP decreased from 15 mg / L on D1 to 10 mg / L (normal range <10 mg / L), the abdominal pain score decreased to 4, acid reflux symptoms disappeared, and the bloating score decreased from 6 to 3. These data preliminarily indicate effective treatment. The onset time determination unit compared changes in indicators before and after administration to determine the drug's onset time. The unit compared indicators from D1 (before administration) with those from D2 (24 hours after administration): CRP decreased by 33%, the abdominal pain score decreased by 50%, the gastric pH increased by 113%, and the patient reported "significant relief of postprandial abdominal pain," meeting the criteria for "drug onset" (inflammation markers decreased by ≥30%, symptom score decreased by ≥40%). Therefore, the onset time was recorded as "18 hours after administration," earlier than the expected 24 hours, indicating that the initially adjusted 25 mg dose was appropriate for the patient's condition. The adverse reaction monitoring unit continuously monitored medication risks based on patient feedback and physiological indicators. At 10:00 on day 4, the patient reported "mild abdominal distension, without nausea or vomiting" via the APP. The nurse immediately collected the patient's abdominal signs (no tenderness or rebound tenderness) and retrieved the gastric motility data from the non-invasive surface monitoring module (peristalsis frequency decreased to 2 times / minute, slightly below the normal range), which was determined to be "mild adverse reaction (abdominal distension)". The system only recorded this reaction and did not trigger a medication discontinuation prompt. At the same time, the monitoring frequency of abdominal distension symptoms was increased (from once a day to twice a day). On day 8, the patient reported that the abdominal distension symptoms disappeared, the gastric motility frequency returned to 3 times / minute, and the adverse reaction was resolved. No severe adverse reactions occurred during the entire treatment cycle, and the medication safety was good.Based on efficacy feedback data, the parameter optimization unit adjusted the dosing regimen and testing frequency in three stages: The first adjustment was on day 4 (72 hours after administration). Data showed that CRP was still 10 mg / L (not yet within the normal range), and the abdominal pain score remained at 4 (the rate of decrease was slowing down), indicating a "slowed onset of action." Therefore, the single dose was increased from 25 mg to 30 mg, and the testing frequency was increased from once daily to twice daily (at 8 am and 8 pm respectively) to track the efficacy after the dose adjustment in real time. The second adjustment was on day 8 (7 days after administration). Data showed that CRP had decreased to 8 mg / L (within the normal range), the abdominal pain score had decreased to 2, and abdominal distension had disappeared, indicating "stable efficacy." The testing frequency was reduced from twice daily to once every two days to reduce the patient's testing burden. The third adjustment was on day 12 (11 days after administration). Data showed that all indicators were normal, and the abdominal pain score had decreased to 1, indicating "approaching clinical cure." The single dose was reduced from 30 mg to 25 mg to avoid side effects (such as elevated liver enzymes) caused by long-term high-dose medication. At the end of the treatment cycle (D15), the patient's follow-up examination showed that the diameter of the gastric ulcer lesion had shrunk to 5mm (a 58% reduction from the initial size), the gastric pH value had stabilized at 4.0, CRP had decreased to 5mg / L, and symptoms of abdominal pain, acid reflux, and bloating had completely disappeared (VAS score of 0 for all lesions). Helicobacter pylori testing turned negative, indicating a significant treatment effect. The system automatically generated a efficacy evaluation report, including test data from the entire treatment cycle, records of medication regimen adjustments, and efficacy change curves, providing a reference for subsequent follow-up.

[0060] In summary, this application's embodiments utilize an active imaging module to accurately capture 12mm ulcer lesions in the gastric antrum, which are easily missed by traditional endoscopy. Combined with non-invasive surface monitoring and near-infrared spectroscopy-assisted pathological feature acquisition, it achieves multi-dimensional and precise identification of lesion location, physiological parameters, and lesion nature, providing reliable data support for subsequent treatment. In the drug delivery phase, magnetically controlled targeted drive ensures the capsule device's positioning error is <1mm. Simultaneously, the drug release rate (slow release over 2 hours) and dosage (initial adjustment dose of 25mg) are dynamically adjusted based on the highly acidic gastric environment and the severity of the ulcer, preventing drug degradation by gastric acid and ensuring effective drug concentration in the lesion area while minimizing stimulation of normal tissues. In the collaborative phase, the data collaboration platform integrates diagnostic and treatment data through a visual interface, assigns permissions according to roles, and connects to various systems. The hospital information system enables data sharing and supports real-time interaction across multiple terminals. This avoids the inefficiency of doctors repeatedly searching for data and nurses repeatedly entering data, while also allowing patients to clearly understand their treatment progress and improving the efficiency of doctor-patient collaboration. In the efficacy management phase, by collecting indicators at multiple time points, dynamically determining the onset period (18 hours earlier than expected), accurately monitoring mild adverse reactions (abdominal distension), and optimizing the dosing regimen in three stages (dosage 25mg → 30mg → 25mg, with testing frequency adjusted as needed), the system achieves "maximum efficacy and minimum risk." Ultimately, this resulted in a 58% reduction in the size of the patient's ulcer lesions, Helicobacter pylori clearance, and complete symptom disappearance, with no severe adverse reactions throughout the entire process. This significantly improves treatment effectiveness and safety, while also reducing the physical and time burden on patients through non-invasive testing and reducing unnecessary examinations.

[0061] Next, referring to the accompanying drawings, a digestive tract detection and drug administration method according to an embodiment of this application is described.

[0062] like Figure 8 As shown, this method for administering medication for gastrointestinal testing includes the following steps: In step S101, images, pH values, temperatures, and pathological characteristics of lesions within the digestive tract are acquired.

[0063] It is understood that the embodiments of this application, by acquiring intraluminal images, can intuitively present the location (e.g., gastric antrum, gastric angle) and morphology (e.g., ulcer size, edge condition) of lesions, avoiding the omission of blind spots in traditional detection, and providing visual basis for subsequent lesion localization; acquiring physiological parameters such as pH value and temperature can reflect the real-time environment in the digestive tract lumen (e.g., strongly acidic in the stomach, weakly neutral in the intestine), directly determining the direction of adjustment of subsequent drug release rate, and avoiding drug degradation or inactivation due to unsuitable environment; acquiring pathological feature data of lesion tissue (e.g., tissue water content, protein concentration, pathological section results) can accurately determine the nature (e.g., inflammation, ulcer) and severity (e.g., mild / severe ulcer), providing core basis for lesion grading and drug dosage determination. By comprehensively collecting key data, the risk of misdiagnosis based solely on symptoms is reduced, and complete data support is provided for subsequent lesion localization analysis, targeted drug delivery regimen formulation, and efficacy monitoring, laying the foundation for individualized and precise digestive tract diagnosis and treatment, while also reducing subsequent treatment deviations caused by data gaps, and improving treatment safety and effectiveness.

[0064] In step S102, the images, pH value, temperature and pathological features of the lesion tissue in the digestive tract are analyzed and processed. The lesion area is identified by the image segmentation algorithm. A three-dimensional coordinate model of the lesion is constructed based on physiological parameters. Combined with the peristalsis law of the digestive tract, the displacement trend of the lesion is predicted. The lesion status is analyzed in stages and the severity of the lesion is output.

[0065] Among them, the lesion three-dimensional coordinate model is a model that accurately reflects the actual spatial location of the lesion in the digestive tract lumen, based on the pixel coordinates of the lesion area image and digestive tract physiological parameters (such as gastric wall thickness and gastric cavity volume) through a three-dimensional reconstruction algorithm. It provides a core spatial benchmark for the positioning of subsequent targeted drug delivery devices.

[0066] It is understood that the embodiments of this application utilize a three-dimensional coordinate model of the lesion to transform the two-dimensional image information of the lesion in the digestive tract lumen into precise three-dimensional spatial location data. Combined with the peristaltic laws of the digestive tract, the lesion location is dynamically adapted, providing a key spatial reference for targeted drug delivery and disease monitoring. This allows the micro-drug delivery device to accurately move to the lesion area based on the precise spatial coordinates in the model, significantly reducing positioning errors, preventing drugs from acting on normal tissues, reducing side effects, and improving drug utilization. The peristaltic laws can be combined to predict the lesion displacement trend, enabling the timing and location of drug delivery to dynamically match changes in the lesion, ensuring that the device remains accurately aligned with the lesion during drug delivery and avoiding drug delivery deviations caused by lesion movement. At the same time, the model can also serve as a spatial reference for phased analysis of the lesion status. During subsequent monitoring, the model coordinates and morphology at different stages can be compared to intuitively determine whether the lesion has shrunk or shifted, assisting in the evaluation of treatment effects.

[0067] For example, taking the treatment of a 45-year-old male patient, Mr. Li, with severe antral gastric ulcer, the doctor first constructed a three-dimensional coordinate model of the lesion based on physiological parameters such as gastric wall thickness (lesion area 4mm) and gastric cavity volume (fasting 50mL), combined with the lesion pixel coordinates (480, 360) output by the active imaging module, and determined the three-dimensional coordinates of the lesion center as (15mm, 8mm, -5mm) (with the xiphoid process as the origin). Then, combined with his gastric peristalsis pattern (frequency 3 times / minute, amplitude 5mm), the model predicted that the lesion would shift towards the pylorus to (18mm, 10mm, -5mm) 1 hour after a meal. When administering the drug, the magnetic control system of the responsive drug delivery module adjusted the electromagnetic coil parameters according to the predicted coordinates, driving the capsule drug delivery device to move precisely, and finally making the device stop directly above the lesion with an error of less than 1mm, avoiding positioning deviation caused by the dynamic displacement of the lesion, ensuring that the drug can be accurately released into the lesion area and maintaining an effective therapeutic concentration.

[0068] In step S103, based on the lesion coordinates and intracavitary environmental parameters, combined with the severity of the lesion, the micro-drug delivery device is controlled to perform targeted release, and the drug dosage and drug release rate are adjusted.

[0069] Among them, the micro-drug delivery device is a small device (usually capsule-shaped) used for precise drug delivery in the digestive tract detection and drug delivery system. It has built-in components such as permanent magnets, drug storage chambers and positioning chips. It can be moved to the lesion area by magnetic control and other means, and precisely release drugs according to control signals to achieve targeted drug delivery to digestive tract lesions.

[0070] It is understood that the embodiments of this application use a built-in permanent magnet to precisely move the drug to the lesion area with magnetic control, and then complete the targeted release of the drug according to the control signal. At the same time, the dosage and release rate are adjusted to reduce positioning errors, allowing the drug to act directly on the lesion, avoiding stimulation of normal digestive tract tissues and drug waste. It can dynamically respond to treatment needs (such as slowing down the release rate according to the strong acid environment in the stomach, and increasing the dosage according to the severity of the lesion), ensuring that the lesion area always maintains an effective drug concentration to improve the efficacy and reduce the discomfort of traditional drug administration methods.

[0071] For example, in the treatment of Mr. Li, a 45-year-old patient with severe antral gastric ulcer, medical staff used a capsule-type micro-drug delivery device (with a built-in permanent magnet, a 25mg proton pump inhibitor storage chamber, and a positioning chip). First, based on the lesion's three-dimensional coordinate model, the coordinates of the lesion were predicted one hour after a meal (18mm, 10mm, -5mm). The parameters of the electromagnetic coil were adjusted by external magnetic control, driving the device to precisely stop directly above the lesion with an error of <1mm. Subsequently, the device, combined with the strongly acidic environment of 2.3 in the stomach, adjusted the drug release rate to a slow release over 2 hours to avoid drug degradation. At the same time, the single dose of 25mg was determined according to the severity of the ulcer. During the release process, the drug concentration in the lesion area was monitored in real time by the built-in sensor (maintaining an effective range of 0.6-0.8μg / mL). Ultimately, the drug was precisely applied to the lesion without irritating the normal gastric mucosa, achieving a targeted drug delivery effect.

[0072] In step S104, the drug administration time axis, intracavitary environment parameter change curve, and adjusted drug dosage and drug release rate are displayed synchronously through a visual interactive interface. The system connects to the hospital's electronic medical record system and pathology analysis system to share test data, drug administration data, and medical record data, allowing doctors to mark key areas of lesions and add drug administration guidance notes. At the same time, the system uploads the patient's physical feedback information after medication.

[0073] It is understood that the visual interactive interface of this application embodiment centrally and synchronously displays the drug administration timeline, the curve of changes in intracavitary environmental parameters, and the adjusted drug dosage and release rate, allowing doctors to intuitively grasp key treatment data and quickly determine the suitability of the current plan without having to search through multiple reports or systems. By connecting with the hospital's electronic medical record and pathology analysis system, it can break down data silos, automatically share detection, drug administration, and medical record data, avoid errors caused by manual re-entry, and ensure data consistency and integrity. Doctors can directly mark key areas of lesions, add drug administration guidance notes, and also support uploading patient medication feedback. This not only enables real-time collaboration among multiple terminals (doctors, nurses, etc.) and shortens information transmission delays (such as nurses sending feedback and doctors providing guidance without offline handover), but also allows doctors to optimize plans in a timely manner based on complete data, improve the accuracy of diagnosis and treatment, and provide a clear and complete data chain for subsequent diagnosis and treatment traceability and efficacy review, further standardizing the diagnosis and treatment process.

[0074] In step S105, based on the patient's physical feedback information after medication, the drug's onset period is determined by comparing the changes in indicators before and after medication. In conjunction with changes in physiological indicators, adverse reactions are monitored. The dosage and drug release rate are adjusted according to the efficacy feedback data, and the detection frequency is adjusted appropriately.

[0075] Among them, efficacy feedback data is a collective term for data collected by the gastrointestinal detection and drug delivery system after drug administration, such as pH value in the gastrointestinal lumen, inflammation-related indicators, patient symptom scores, drug onset time and adverse reactions. It is used to evaluate drug efficacy, monitor drug risks, and provide core basis for subsequent optimization of drug administration parameters and detection frequency.

[0076] It is understood that the embodiments of this application determine the drug's onset period by comparing changes in indicators before and after drug administration, monitor adverse reactions by combining physiological indicators, and adjust the dosage, release rate, and detection frequency based on this, providing data support for the dynamic optimization of the treatment plan, timely detecting whether the drug is effective (if it is ineffective, it can be quickly intervened to adjust and avoid delaying treatment), accurately identifying and managing drug risks (such as distinguishing between mild and severe adverse reactions and balancing efficacy and safety), making the dosing plan more suitable for individual patient responses (such as increasing the dosage for those with slow onset of action and adjusting the release rate for those with adverse reactions), and reducing the burden on patients by adjusting the detection frequency (reducing detection when the efficacy is stable).

[0077] According to the embodiments of this application, a digestive tract detection and drug delivery method is proposed. By combining a digestive tract sensing unit with active imaging, non-invasive monitoring, and pathological acquisition, it overcomes the limitations of traditional detection blind spots, achieving comprehensive and accurate acquisition of intraluminal images, physiological parameters, and pathological features. A lesion localization analysis unit, relying on an improved U-Net algorithm and fluid dynamics model, accurately identifies lesion areas and predicts displacement trends, providing precise coordinates for targeted drug delivery. A responsive drug delivery module, through magnetic control and dynamic dose adjustment, solves the problems of poor targeting and fixed dosage in traditional drug delivery, achieving precise and appropriate drug delivery to the lesion area. A data collaboration platform breaks down data barriers between multiple systems, supports real-time interaction among multiple roles, and improves the efficiency of diagnostic and treatment collaboration. A efficacy feedback unit, through multi-node indicator acquisition and dynamic parameter optimization, can promptly determine the drug's onset period and adverse reactions, and can specifically adjust the drug delivery plan and detection frequency, effectively reducing medication risks and improving treatment efficacy. Thus, it solves the problems of low targeting accuracy and poor diagnostic and treatment efficiency in existing technologies.

[0078] The following will illustrate a method for administering medication for gastrointestinal testing through a specific embodiment, such as... Figure 9 As shown, it includes: A 52-year-old female patient, Ms. Wang, presented to the gastroenterology department of a tertiary hospital with a 4-month history of recurrent left upper abdominal pain, particularly at night accompanied by acid reflux and belching. She had a 3-year history of hypertension (controlled with regular valsartan), no history of gastrointestinal surgery, and no history of drug allergies. Her Helicobacter pylori breath test result was 8.7‰ (normal reference value <4‰, indicating a positive result). Based on her symptoms, the doctor initially suspected gastric mucosal lesions and decided to use a gastrointestinal monitoring and drug delivery system for diagnosis and treatment. The first stage was data acquisition, which took approximately 45 minutes and was primarily non-invasive, with good patient tolerance. Intraluminal image acquisition was performed using a 9mm diameter medical capsule endoscope (equipped with a 2-megapixel high-definition camera and a 2.4GHz wireless transmission module). The patient took the capsule with 30mL of warm water after fasting for 6 hours. The doctor adjusted the capsule lens angle (360° horizontal rotation and ±40° vertical adjustment) via an external control console, focusing on monitoring blind spots easily missed by traditional gastroscopy, such as the lesser curvature and gastric angle. After the capsule enters the stomach cavity, it automatically activates the imaging mode, acquiring two frames per second of 1920×1080 resolution images and transmitting them to the system terminal in real time. After 30 minutes, the terminal successfully received 180 frames of images from inside the stomach cavity, among which 12 frames clearly showed a mucosal defect area approximately 10mm in diameter on the lesser curvature of the stomach body—the defect edges were red and congested, and the base was covered with a thin white coating, consistent with typical imaging characteristics of a gastric ulcer. The system automatically marked this area as a "key lesion sequence," providing a basis for subsequent analysis. pH and temperature monitoring were performed using a non-invasive surface monitor, consisting of three medical conductive gel electrode pads (2cm in diameter) and a portable data logger. Electrodes were applied to the patient's upper abdomen below the xiphoid process, at the intersection of the left midclavicular line and the costal arch, and at the intersection of the right midclavicular line and the costal arch. Two electrodes collected gastric electrical activity signals (sampling frequency 256Hz), simultaneously recording the gastric peristalsis frequency (2.8 times / minute) and amplitude (4.5mm). The third electrode contained a miniature pH sensor (3mm in diameter, fixed near the lesion on the lesser curvature of the stomach with the aid of a gastroscopy) to collect real-time gastric pH values—1.3 on an empty stomach (normal range 1.0-3.0), rising to 2.4 30 minutes after a meal. Simultaneously, the monitor collected abdominal surface temperature via an infrared thermometer, maintaining it consistently between 36.9-37.2℃ without abnormal fluctuations. All data were automatically uploaded to the system database at a frequency of 1 minute, generating continuous pH change curves and temperature trend graphs. The pathological features of the lesion tissue were collected with the assistance of gastroscopy: the doctor inserted a 5mm diameter retractable biopsy forceps (with a near-infrared spectral sensor at the tip, with a detection wavelength of 700-1100nm) into the stomach cavity through the working channel of the gastroscopy, and precisely took two gastric mucosal tissue samples with a diameter of about 2mm from the lesion area marked by the capsule endoscope.During the biopsy, a near-infrared spectroscopy sensor directly contacts the lesion surface to analyze tissue physiological indicators in real time. The test showed that the tissue water content in the lesion area was 76% (the water content of normal gastric mucosa is about 72%), and the protein concentration was 19g / 100g (the protein concentration of normal gastric mucosa is about 22g / 100g). After the tissue sample was sent to the pathology department, HE staining and section examination indicated "chronic inflammation of the gastric mucosa with moderate glandular atrophy", and the preliminary judgment was that the nature of the lesion was chronic gastric ulcer (active phase).

[0079] After receiving the collected data, the lesion localization analysis unit is activated for processing, taking approximately 18 minutes in total. The final output includes lesion severity grading and diagnostic reference information. The image segmentation stage employs an improved U-Net deep learning algorithm. This algorithm adds a "channel attention module" to the traditional U-Net architecture, enhancing feature extraction of lesion edges and abnormal tissue. The system inputs 180 frames of intragastric images acquired by capsule endoscopy into the algorithm model. First, Gaussian filtering removes image noise and grayscale correction unifies brightness. Then, the encoder (containing 4 convolutional layers) extracts image features (such as the congested area at the lesion edge and grayscale differences in the white coating at the bottom). The decoder (containing 4 deconvolutional layers) reconstructs the lesion contour based on feature mapping. Finally, the algorithm automatically identifies a 10mm diameter ulcer lesion on the lesser curvature of the gastric body with a localization accuracy of 97.5%, generating a binary segmentation map of the lesion area (white area represents the lesion, black area represents normal mucosa), and outputting the pixel coordinates of the lesion center (with the upper left corner of the image as the origin, coordinates (520, 410) pixels). The 3D coordinate model of the lesion is constructed based on pixel coordinates, combined with physiological parameters collected by a non-invasive surface monitoring device: The system first measures the actual size of the lesser curvature of the stomach (anteroposterior diameter approximately 32 mm) using a gastroscopy, establishing a mapping relationship between pixels and actual space (1 pixel corresponds to 0.06 mm); then, it substitutes the stomach wall thickness data (3.8 mm in the lesion area, approximately 2.1 mm in the normal area) and changes in stomach volume (approximately 45 mL in fasting condition, approximately 140 mL in postprandial condition), and uses a "voxel reconstruction algorithm" to generate a 3D model of the lesion. With the xiphoid process of the patient's body surface as the origin (X-axis for left-right direction, Y-axis for up-down direction, Z-axis for anteroposterior direction), the model shows that the 3D coordinates of the lesion center are (12 mm, 9 mm, -4 mm). These coordinates accurately reflect the spatial location of the lesion within the stomach cavity, providing a benchmark for subsequent targeted drug delivery. The prediction of lesion displacement trend is based on gastric peristalsis data: the system inputs the peristalsis frequency of 2.8 times / minute and amplitude of 4.5mm collected by a non-invasive surface monitoring device into a fluid dynamics model. This model treats food and liquid in the gastric cavity as viscous fluids (viscosity coefficient of 0.01 Pa·s) and gastric wall peristalsis as periodic contraction movements (contraction cycle of 21 seconds). By calculating the interaction force between the fluid and the gastric wall, the model predicts the displacement trajectory of the lesion at different time points: the results show that within 1 hour after a meal, gastric peristalsis increases, and the lesion will shift approximately 5mm towards the pylorus, with the three-dimensional coordinates becoming (15mm, 11mm, -4mm); 2.5 hours after a meal, gastric contents empty into the small intestine, peristalsis weakens, and the lesion displacement returns to near the initial coordinates (error < 1mm). The system stores this displacement curve in the database, marking "1 hour after a meal is the stable window period for lesion displacement, suitable for drug administration."The system performs phased analysis of lesion status and severity grading simultaneously: It first sets four monitoring phases: before treatment, 24 hours after administration, 72 hours after administration, and 7 days after administration. Each phase is associated with three types of data: image features (lesion size, edge state), physiological parameters (pH value, inflammatory markers), and displacement trend. Then, combined with grading standards, key judgment indicators are extracted: lesion diameter 10mm (at the borderline between "moderate ulcer 5-10mm" and "severe ulcer >10mm", combined with pathological features of "moderate glandular atrophy"), Helicobacter pylori positivity, and tissue protein concentration below normal levels (indicating decreased mucosal repair capacity). The overall severity of the patient's gastric ulcer is determined to be "moderate to severe," and the system automatically references its built-in dosing regimen library, retrieving the corresponding initial regimen (proton pump inhibitor, single dose 20mg, twice daily, drug release rate of 2.5 hours slow release).

[0080] Based on the lesion analysis results, the doctor chose to administer the first dose 1 hour after the patient's meal (the lesion displacement stabilization window). The responsive drug delivery module operated fully automatically, taking approximately 35 minutes. The miniature drug delivery device uses a capsule-shaped structure (9mm in diameter and 26mm in length), containing three core components: a neodymium iron boron permanent magnet (magnetic field strength 0.45T, used for magnetic control), a drug storage chamber (50mg capacity, pre-filled with proton pump inhibitor powder), and a wireless positioning chip (accuracy ±0.5mm, real-time transmission of position data). The external supporting equipment consists of three sets of electromagnetic coils (placed on the left, middle, and right sides of the patient's upper abdomen, respectively, with a diameter of 8cm) and a magnetic control console (including a position monitoring screen and parameter adjustment buttons). Before drug delivery, the "lesion coordinates (15mm, 11mm, -4mm) 1 hour after the meal" predicted by the lesion's three-dimensional coordinate model are input into the magnetic control console. The console automatically calculates the current parameters of the electromagnetic coils (3.2A for the left coil, 2.8A for the middle coil, and 3.0A for the right coil) and generates the magnetic field driving path. Doctors observe the movement trajectory of the capsule device in real time through a position monitoring screen: In the initial stage, the left coil is energized to generate a magnetic field, which propels the capsule from the bottom of the stomach to the stomach body (speed 2mm / s); when the capsule is about 8mm away from the lesion, the current in the middle coil increases, slowing the movement speed to 0.8mm / s; finally, the three sets of coils work together to adjust the direction and intensity of the magnetic field, so that the capsule stops precisely above the lesion (about 2mm away from the lesion surface), with a positioning error of 0.8mm, completing the targeted positioning. Drug dosage and release rate adjustment are based on intracavitary environmental parameters and lesion grading: The system first reads the gastric pH value (currently 2.5, strongly acidic) transmitted in real time by the non-invasive surface monitor. Since rapid drug release is easily destroyed by gastric acid, the drug release control module is set to a "2.5-hour slow release" mode—releasing the drug at a rate of 8 mg per hour through the enteric-coated sustained-release membrane (0.1 mm thick) on the capsule shell. Combined with the lesion grading of "moderate to severe," the module increases the single-dose dose from the initial 20 mg to 24 mg (ensuring an effective drug concentration ≥0.5 μg / mL at the lesion site). During release, the drug concentration sensor built into the capsule monitors the concentration at the lesion site in real time: the concentration reaches 0.7 μg / mL after 1 hour of release and remains at 0.6 μg / mL after 2.5 hours, meeting treatment requirements. Simultaneously, the module wirelessly transmits progress data such as "16 mg released, 8 mg remaining" and "current concentration 0.7 μg / mL" to a data collaboration platform for remote monitoring by physicians.

[0081] As the information hub, the data collaboration platform immediately initiates data integration and interaction functions after the drug administration procedure, enabling efficient linkage between multiple roles (doctors, nurses, and patients) and multiple systems (hospital electronic medical record EMR, pathology analysis system). The visual interactive interface is deployed on three terminals: a desktop computer (19-inch display) in the doctor's office, a tablet computer (12-inch) at the nurses' station, and a hospital app on the patient's mobile phone. The interface layout follows the principle of "layered information display": the left side is the drug administration timeline, marked "First administration time D114:00, next administration time D122:00," and marked with a red dot indicating "Currently in the post-drug monitoring phase"; the right side is divided into upper and lower sections, the upper section showing the gastric pH change curve (D113:00-15:00, pH value rising from 2.1 to 2.5 and then decreasing to 2.3), and the lower section showing the drug administration parameter panel, clearly displaying "Adjusted dose 24mg, release rate 2.5 hours slow release"; a high-resolution image of the lesion taken by capsule endoscopy is embedded at the bottom of the interface, which can be zoomed in to full screen to view details of edge congestion. Doctors can access three core data categories—drug administration progress, environmental parameters, and lesion images—in a single session via a desktop computer interface, without switching to other systems, in just 30 seconds. Data sharing is achieved through the hospital's HL7 standard interface: on one hand, "intracavitary images, pH curves, pathological feature data, lesion grading results, and drug administration parameters" are automatically synchronized to the pathology analysis system. After logging into the system, pathologists can review the nature of the lesion by combining it with previous tissue section reports (the final review result is consistent with the system grading, indicating "moderate to severe gastric ulcer"). On the other hand, the patient's history of hypertension and medication records (valsartan) are retrieved from the EMR system and added to the treatment database, with a notification indicating "the drug interaction between proton pump inhibitors and valsartan needs to be monitored; there are no contraindications, but blood pressure monitoring is necessary." Simultaneously, drug administration records (time, dosage, and rate) are automatically uploaded to the EMR system, integrated with the patient's previous medical records to form a complete treatment file, avoiding errors caused by manual data entry by nurses (such as incorrect dosage or time discrepancies). Efficiently integrated multi-role interaction and feedback upload process: At 6:00 AM on D11 (2 hours after administration), a nurse brings a tablet to the patient's ward and records the patient's feedback—"Left upper abdominal pain score decreased from 7 (VAS score) before administration to 4, no nausea or bloating, and acid reflux symptoms have lessened." After clicking "Submit," the data is synchronized to the doctor's terminal in real time. After reviewing the feedback, the doctor marks the "area of ​​congestion at the edge of the lesion needs special attention" with a red circle on the lesion image on the desktop computer interface, and adds a medication guidance note: "Maintain the dose of 24mg for the next administration, with the release rate unchanged. If the pain score drops below 3, a slight adjustment of the dose may be considered." The note is immediately pushed to the nurse's tablet. After the nurse confirms receipt, the system automatically generates an "interaction log." Patients can log in to the platform via a mobile app to view their medication records, pH curves, and changes in pain scores without having to go to the nurses' station for consultation, thus improving the medical experience.

[0082] The efficacy feedback unit uses a "2-week" complete treatment cycle and sets 5 key monitoring nodes (12 hours after administration, D2 2:00; 24 hours after administration, D2 14:00; 72 hours after administration, D4 14:00; 7 days after administration, D8 14:00; 14 days after administration, D15 14:00). Data is collected at each node, and evaluation and adjustments are completed, achieving "data-driven dynamic optimization" throughout the process. The drug onset period is determined at D2 14:00: The system compares the core indicator before and after administration (D1 and D2)—the inflammatory marker CRP (C-reactive protein) decreased from 14 mg / L on D1 to 9 mg / L on D2 (normal range <10 mg / L), a decrease of 35.7%; the abdominal pain VAS score decreased from 7 to 3, a decrease of 57.1%; the gastric pH value increased from 1.3 to 3.6, an increase of 176.9%; and the patient reported "nocturnal pain disappeared, and only mild discomfort after meals." Based on the system's built-in "onset of action criteria" (inflammation markers decreased by ≥30%, symptom scores decreased by ≥40%), the onset time of the drug was determined to be "16 hours after administration," earlier than the expected 24 hours, indicating that the 24mg dose was appropriate for the patient's condition. Adverse reaction monitoring was conducted throughout the entire cycle: At 10:00 on D4, the patient reported "mild constipation (one bowel movement per day, drier than usual), no other discomfort" via the mobile app. The nurse immediately retrieved the intestinal peristalsis data from the non-invasive surface monitor (frequency 1.8 times / minute, slightly lower than the normal range of 2-3 times / minute), combined with CRP (8mg / L, normal) and abdominal signs (no tenderness), and determined it to be "mild adverse reaction (drug-related slowed intestinal peristalsis)." The system only recorded this reaction and did not trigger a discontinuation prompt. At the same time, "constipation symptoms" were added to the monitoring list, and the nurse followed up with feedback daily via the app; at D7, the patient reported "constipation relieved, bowel movements returned to normal," and the intestinal peristalsis frequency returned to 2.2 times / minute, indicating that the adverse reaction had resolved. The dosing regimen and monitoring frequency were adjusted in three stages: The first adjustment was at 14:00 on day 4. Data showed CRP at 8 mg / L (normal) and an abdominal pain score of 3, but the patient still experienced mild postprandial discomfort. The system determined that "the efficacy was stable but there was still room for optimization," so the single dose was slightly adjusted from 24 mg to 26 mg, and the monitoring frequency was increased from "once daily (9:00 AM)" to "twice daily (9:00 AM and 6:00 PM)" to track the effect of the dose adjustment in real time. The second adjustment was at 14:00 on day 8. New data showed CRP... With a dose of 6 mg / L, an abdominal pain score of 1, and no adverse reactions, the system determined that the treatment was "significant and stable." The testing frequency was reduced from twice daily to "once every two days (9:00 AM)" to reduce the patient's monitoring burden (e.g., reducing the number of electrode patch applications). The third adjustment was made at 14:00 on D12. All indicators were normal (CRP 5 mg / L, abdominal pain score 0, pH stable at 4.0). The system determined that the patient was "approaching clinical cure," and the dose was reduced from 26 mg to 24 mg to avoid the risk of elevated liver enzymes that may be caused by long-term high-dose medication.At the end of the treatment cycle (D15), the patient's follow-up data showed that: capsule endoscopy images showed that the diameter of the ulcer lesion had shrunk to 4mm (60% smaller than the initial size), and the peripheral congestion had disappeared; the gastric pH value was stable at 3.8-4.2; CRP decreased to 4mg / L; symptoms such as abdominal pain, acid reflux, and constipation completely disappeared (VAS score was 0); the Helicobacter pylori breath test value was 2.1‰ (turned negative). An efficacy evaluation report containing three parts, namely "data comparison table of each stage, dosing regimen adjustment record, and efficacy change curve", was automatically generated and synchronized to the EMR system to provide a complete reference for the subsequent one-month follow-up (monthly re-examination).

[0083] In summary, the embodiments of this application utilize capsule endoscopy to monitor blind spots in the digestive tract that are easily missed by traditional gastroscopy, accurately capturing ulcer lesions. The surface monitoring device and near-infrared assisted pathological acquisition enable multi-dimensional data acquisition, significantly improving patient tolerance and providing a comprehensive and reliable data foundation for subsequent diagnosis and treatment, avoiding the bias of diagnosis based solely on symptoms. Optimized deep learning algorithms achieve precise lesion localization, constructing a three-dimensional coordinate model of the lesion based on physiological parameters and predicting the lesion's displacement trend with digestive tract peristalsis. Then, referring to clinical standards, the severity grade of the lesion is output, providing scientific guidance for targeted drug delivery and reducing errors caused by empirical localization. In the targeted drug delivery process, a magnetically controlled micro-drug delivery device achieves precise docking near the lesion, dynamically adjusting the dosage and drug release rate according to the intraluminal environment and disease severity, ensuring that an effective drug concentration is maintained in the lesion area, thus avoiding drug degradation due to environmental factors. Discomfort is reduced, and stimulation of normal digestive tract tissues is minimized, enabling precise drug delivery to the lesion. A multi-terminal visual interactive interface integrates key diagnostic and treatment data, connecting with hospital electronic medical records and pathology analysis systems for data sharing. It also supports real-time interaction among doctors, nurses, and patients, allowing doctors to quickly access complete diagnostic and treatment data, eliminating the need for nurses to manually re-enter information, and enabling patients to independently monitor their treatment progress, significantly improving medical collaboration efficiency and reducing data transmission errors. By monitoring efficacy indicators at multiple time points, the onset of drug action can be determined promptly, mild adverse reactions can be accurately identified and managed, and the dosing regimen and testing frequency can be dynamically adjusted based on efficacy feedback. This ensures treatment effectiveness (ulcer lesion shrinkage, pathogenic bacteria negativity, and complete symptom disappearance) while avoiding the risks of long-term inappropriate medication, reducing unnecessary testing procedures, and alleviating the physical and time burden on patients.

[0084] Figure 10 A schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device may include: The memory 1001, the processor 1002, and the computer program stored on the memory 1001 and capable of running on the processor 1002.

[0085] When the processor 1002 executes the program, it implements a digestive tract detection and drug delivery method provided in the above embodiments.

[0086] Furthermore, electronic devices also include: Communication interface 1003 is used for communication between memory 1001 and processor 1002.

[0087] The memory 1001 is used to store computer programs that can run on the processor 1002.

[0088] The memory 1001 may include high-speed RAM (Random Access Memory) memory, and may also include non-volatile memory, such as at least one disk storage.

[0089] If the memory 1001, processor 1002, and communication interface 1003 are implemented independently, then the communication interface 1003, memory 1001, and processor 1002 can be interconnected via a bus to complete communication between them. The bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 10 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0090] Optionally, in a specific implementation, if the memory 1001, processor 1002, and communication interface 1003 are integrated on a single chip, then the memory 1001, processor 1002, and communication interface 1003 can communicate with each other through an internal interface.

[0091] The processor 1002 may be a CPU (Central Processing Unit), an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of this application.

[0092] This application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method for detecting and administering drugs in the digestive tract.

[0093] Furthermore, this application also provides a computer program product, including a computer program or instructions, which, when executed, implement the above-described digestive tract detection and drug delivery method.

[0094] In the description of this specification, the references to "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0095] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0096] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0097] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any of the following techniques known in the art, or a combination thereof: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0098] Those skilled in the art will understand that all or part of the steps of the methods described in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0099] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A digestive tract detection and drug delivery system, characterized in that, include: The system includes a digestive tract sensing unit, a lesion localization and analysis unit, a responsive drug delivery module, a data collaboration platform, and a treatment efficacy feedback unit; among these components... The digestive tract sensing unit is used to collect images, pH value, temperature and pathological characteristics of diseased tissues within the digestive tract lumen. The lesion localization and analysis unit is used to identify the lesion area through image segmentation algorithm, construct a three-dimensional coordinate model of the lesion based on physiological parameters, predict the displacement trend of the lesion by combining the peristalsis law of the digestive tract, analyze the lesion status in stages, and output the severity of the lesion. The responsive drug delivery module is used to control the micro-drug delivery device to perform targeted release and adjust the drug release rate based on the lesion coordinates and intracavitary environmental parameters, combined with the severity of the lesion. The data collaboration platform is used to share test results and medication records through a real-time interactive interface; The therapeutic feedback unit is used to collect data on changes in physiological indicators and improvement of patient symptoms after drug administration, determine the drug's onset period and possible adverse reactions, and optimize drug administration parameters and detection frequency.

2. The digestive tract detection and drug delivery system according to claim 1, characterized in that, The digestive tract sensing unit includes an active imaging module, a non-invasive surface monitoring module, and a pathological feature acquisition module. The active imaging module adjusts the lens angle wirelessly to monitor traditional blind spots and generate high-definition image sequences of lesion areas. The non-invasive surface monitoring module collects gastric electrical activity, gastric pH, and abdominal temperature in real time. The pathological feature acquisition module obtains lesion tissue samples through retractable biopsy forceps and analyzes tissue water content and protein concentration in real time using a near-infrared spectral sensor to preliminarily determine the nature of the lesion.

3. The digestive tract detection and drug delivery system according to claim 1, characterized in that, The lesion localization and analysis unit includes an image segmentation module, a three-dimensional coordinate localization module, a lesion trend prediction unit, and a lesion grading engine. The image segmentation module uses an improved U-Net deep learning algorithm to segment digestive tract images and automatically identify lesion areas such as ulcers, polyps, and bleeding points. The three-dimensional coordinate localization module constructs a three-dimensional coordinate model of the lesion based on the image pixel coordinates of the lesion area and the spatial distribution differences of digestive tract physiological parameters to locate the lesion coordinates. The lesion trend prediction unit uses a fluid dynamics model to simulate the displacement trajectory of the lesion with digestive tract peristalsis based on digestive tract peristalsis data, predicting the displacement trend of the lesion within a preset time. The lesion grading engine outputs a lesion severity grade based on pathological feature data and clinical standards, and associates it with a corresponding drug regimen library.

4. The digestive tract detection and drug delivery system according to claim 1, characterized in that, The responsive drug delivery module includes a target positioning drive module and a drug release control module. The target positioning drive module receives three-dimensional coordinate model data of the lesion and moves the micro-drug delivery device to the lesion area through magnetic control. The drug release control module dynamically adjusts the drug release rate according to the environmental parameters of the digestive tract lumen, and adjusts the single drug delivery dose according to the severity of the lesion.

5. The digestive tract detection and drug delivery system according to claim 1, characterized in that, The data collaboration platform includes a visual interactive interface, a permission management unit, a data sharing module, and a real-time communication unit. The visual interactive interface synchronously displays the drug administration timeline and the curves showing changes in intracavitary environmental parameters, allowing users to view detailed information about lesions. The permission management unit assigns different operating permissions based on user roles: doctors have permissions to annotate lesions and edit drug administration plans; patients can only view their own test results and drug administration records; and nurses have permissions to input patient medication feedback data. The data sharing module connects to the hospital's electronic medical record system and pathology analysis system, enabling the sharing of test data, drug administration data, and medical record data. The real-time communication unit facilitates multi-terminal online interaction; doctors can annotate key areas of lesions and add drug administration guidance notes on the interface, while nurses can upload real-time patient feedback information after medication administration.

6. The digestive tract detection and drug delivery system according to claim 1, characterized in that, The efficacy feedback unit includes an indicator acquisition module, an onset time determination unit, an adverse reaction monitoring unit, and a parameter optimization unit. The indicator acquisition module collects data on the pH value of the digestive tract, inflammation-related indicators, and patient symptom scores at multiple time points after drug administration. The onset time determination unit determines whether the drug is effective and records the onset time by comparing changes in indicators before and after drug administration. The adverse reaction monitoring unit combines patient feedback information with changes in physiological indicators to determine whether adverse reactions caused by drug stimulation exist; mild adverse reactions are recorded and continuously observed, while severe adverse reactions trigger a drug discontinuation prompt. The parameter optimization unit adjusts the dosage for patients with slow drug onset and adjusts the drug release rate for patients experiencing adverse reactions, while also appropriately adjusting the detection frequency based on the efficacy feedback data.

7. A method for use in a digestive tract detection and drug delivery system according to any one of claims 1-6, characterized in that, The method includes: Acquire images, pH values, temperature, and pathological characteristics of diseased tissues within the digestive tract lumen; The data on images, pH value, temperature and pathological features of lesion tissue in the gastrointestinal lumen are analyzed and processed. The lesion area is identified by the image segmentation algorithm. A three-dimensional coordinate model of the lesion is constructed based on physiological parameters. Combined with the peristalsis law of the gastrointestinal tract, the displacement trend of the lesion is predicted. The lesion status is analyzed in stages and the severity of the lesion is output. Based on the lesion coordinates and intracavitary environmental parameters, combined with the severity of the lesion, the micro-drug delivery device is controlled to perform targeted release, and the drug dosage and drug release rate are adjusted. The system synchronously displays the drug administration timeline, the intracavitary environment parameter change curve, and the adjusted drug dosage and drug release rate through a visual interactive interface. It connects to the hospital's electronic medical record system and pathology analysis system to share detection data, drug administration data, and medical record data, enabling doctors to mark key areas of lesions and add drug administration guidance notes. At the same time, it uploads the patient's physical feedback information after medication. Based on the patient's feedback information after medication, the drug's onset period is determined by comparing the changes in indicators before and after administration. The presence of adverse reactions is monitored in conjunction with changes in physiological indicators. The dosage and drug release rate are adjusted according to the efficacy feedback data, and the detection frequency is adjusted appropriately.

8. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the digestive tract detection and administration method as described in claim 7.

9. A computer-readable storage medium having a computer program or instructions stored thereon, characterized in that, When a computer program or instruction is executed, it implements the digestive tract detection and administration method as described in claim 7.

10. A computer program product, comprising a computer program or instructions, characterized in that, When a computer program or instruction is executed, it implements the digestive tract detection and administration method as described in claim 7.

Images