Individualized nutrition intervention system for local advanced pancreatic cancer patient

Abstract

The invention discloses an individualized nutrition intervention system for a patient with local advanced pancreatic cancer, which comprises a multi-dimensional dynamic monitoring module connected with various monitoring devices and used for collecting metabolism, physiology and tumor specificity omics data of the patient in real time; the monitoring device comprises a wearable physiological monitoring device, a portable blood detection device, a metabolism monitoring device and a tumor specificity omics data detection device, and the dynamic adaptability intervention module utilizes an AI algorithm and a machine learning model. According to the individualized nutrition intervention system for the patient with the local advanced pancreatic cancer, various low-cost monitoring devices are connected, metabolism, physiology and tumor specificity omics data of the patient are comprehensively collected in real time, nutrition risks possibly occurring in the treatment process can be more sensitively captured, a reliable basis is provided for timely adjusting a nutrition intervention scheme, and the patient experience is improved. The problem of insufficient monitoring data in the prior art is effectively solved, the nutrition evaluation accuracy is improved, and rich data support is provided for personalized nutrition intervention.

Description

Technical Field

[0001] This invention relates to the field of medical and health technology, specifically to a personalized nutritional intervention system for patients with locally advanced pancreatic cancer. Background Technology

[0002] Patients with locally advanced pancreatic cancer often face complex physiological and pathological changes and malnutrition. Reasonable nutritional intervention is of great significance in improving their quality of life, enhancing treatment tolerance, and promoting recovery.

[0003] Existing nutritional intervention programs and systems are mostly based on general nutritional guidelines or simple physiological indicator monitoring, lacking precise consideration for patients with locally advanced pancreatic cancer. They only provide general nutritional advice based on the patient's basic physical information, failing to dynamically track their real-time metabolic status, changes in tumor-specific indicators, and physiological responses during treatment. Furthermore, they often neglect the safety and emergency treatment needs of the patient's home environment and have insufficient response capabilities to crisis events.

[0004] Existing patents cover nutritional assessment and remote monitoring of cancer patients, but they focus on single-dimensional data collection, lack multi-dimensional data integration and in-depth analysis, fail to fully utilize artificial intelligence algorithms and machine learning models to dynamically optimize nutritional intervention programs, and have imperfect emergency response mechanisms, making it impossible to take effective measures to protect the lives of patients when they experience nutritional crises.

[0005] In summary, existing technologies for nutritional intervention in patients with locally advanced pancreatic cancer have the following main shortcomings: first, they fail to achieve comprehensive monitoring and integration of multi-dimensional data; second, they lack a dynamic adaptive nutritional intervention program adjustment mechanism based on advanced algorithms; and third, they are insufficient in handling emergency situations in the patient's home environment. This invention aims to propose a novel personalized nutritional intervention system that provides patients with precise, real-time, and safe nutritional support through the synergistic effect of multi-dimensional dynamic monitoring, intelligent algorithm-driven dynamic adaptive intervention, and automatic emergency response modules. Summary of the Invention

[0006] (a) Technical problems to be solved

[0007] To address the shortcomings of existing technologies, this invention provides an individualized nutritional intervention system for patients with locally advanced pancreatic cancer. It has the advantages of comprehensive and personalized nutritional management plans and solves the problems of deficiencies in multi-dimensional monitoring, dynamic adaptive intervention, and emergency treatment.

[0008] (II) Technical Solution

[0009] To achieve the above objectives, the present invention provides the following technical solution: a personalized nutritional intervention system for patients with locally advanced pancreatic cancer, comprising the following steps:

[0010] 1. Multi-dimensional dynamic monitoring module

[0011] 1.1 Device Connection and Data Collection: Connects to multiple monitoring devices to collect patient data in real time, including wearable physiological monitoring devices (such as smart bracelets and smartwatches), portable blood testing devices, metabolic monitoring devices, and tumor-specific omics data detection devices. Portable devices include body fat scales, blood glucose meters, and ketone meters. Their data transmission frequencies are three times a day, once every 30 minutes, and twice a day, respectively. The data quality verification mechanism includes data integrity checks, accuracy verification, and timeliness assessment.

[0012] Metabolic monitoring equipment is based on the indirect calorimetry method, using the formula:

[0013] 1.2 Calculate Resting Energy Expenditure (REE):

[0014] REE = 3.94 V O2 +1.11·V CO2 —2.17. Urinary nitrogen excretion

[0015] 1.3 Tumor-specific omics data detection equipment: Circulating tumor DNA is detected using real-time quantitative PCR technology, with a detection limit of 10 copies / μL;

[0016] 1.4 Wearable devices collect physiological data such as heart rate and blood pressure; portable blood testing devices detect blood biochemical data such as blood glucose; metabolic monitoring devices measure energy consumption through indirect calorimetry; and tumor-specific omics data detection devices detect data such as tumor markers.

[0017] 1.5 Data Preprocessing: The collected data undergoes preprocessing, including data cleaning, fusion, and standardization. Data cleaning removes outliers and noisy data; data fusion correlates different types of data in terms of time and physiological significance; data standardization converts data collected from different devices into a format and unit that the system can recognize.

[0018] 2. Dynamic Adaptive Intervention Module

[0019] 2.1 Data Feature Extraction and Analysis: Feature extraction is performed on the preprocessed data to construct patient feature vectors. For example, heart rate variability features are extracted from heart rate data, systolic and diastolic blood pressure fluctuation range features are extracted from blood pressure data, and resting and activity energy consumption ratio features are extracted from energy consumption data. At the same time, a comprehensive patient feature vector is constructed by combining patient treatment variables and basic information.

[0020] 2.2. Deep Learning-Based Nutritional Needs Prediction Model: A nutritional needs prediction model is constructed using a Long Short-Term Memory (LSTM) network. The model input includes historical nutritional data of patients and current multi-dimensional monitoring data. Through training, it learns the dynamic changes in patients' nutritional needs and predicts the future nutritional requirements of patients. The prediction model formula is as follows:

[0021] Y t+h =LSTM(X) t ,Y t―1 )

[0022] Among them, Y t+h X represents the predicted nutritional requirements of the patient at a future time t+h. t Y represents the multi-dimensional monitoring data at the current time t. t―1 This represents the patient's historical nutritional needs data.

[0023] 2.3 Optimization of Nutritional Intervention Programs Based on Multi-Objective Optimization Algorithms: Based on the prediction of patients' nutritional needs, multi-objective optimization algorithms (such as genetic algorithms) are used to optimize nutritional intervention programs. Optimization objectives include meeting basic nutritional needs, improving nutrient utilization efficiency, reducing the risk of malnutrition, and adapting to individual taste preferences. The optimization algorithm uses nutritional formula parameters as variables to find the optimal solution under constraints. The multi-objective optimization algorithm uses the ratio of protein (P), carbohydrates (C), and fat (F) as decision variables, and the optimization objective is: [Formula omitted].

[0024]

[0025] 0.15 ≤ P / (P+C+F) ≤ 0.3

[0026] 3 Automatic Emergency Response Module

[0027] 3.1 Construction of the Crisis Event Identification Rule Base: The rule base was jointly constructed by medical experts, nutritionists and data scientists, covering a variety of nutrition-related crisis events and their identification rules. Each rule clarifies the level of the crisis event and judges the occurrence of the crisis event based on the patient's physiological indicators and clinical symptoms.

[0028] 3.2 Automatic Triggering and Execution of Emergency Response Procedures: Upon identifying a crisis event, the system immediately activates the corresponding emergency response procedures. For mild malnutrition symptoms, nutritional supplementation suggestions are provided; for severe crisis events, instructions on nasogastric feeding are given and medical personnel are notified.

[0029] 3.3 Emergency Response Effectiveness Evaluation and Feedback: During emergency response, the system continuously monitors changes in the patient's physiological indicators and evaluates the effectiveness of the emergency response in real time. If the physiological indicators are normal, successful cases are recorded and rules are optimized; if there is no improvement or a new crisis occurs, the event level is escalated and higher-level emergency measures are taken.

[0030] 3.4 The automatic emergency response module has a built-in false alarm prevention mechanism, including multiple confirmations and multi-indicator joint judgment. It also includes a manual confirmation process and operation record traceability. When identifying vomiting symptoms, the false alarm prevention mechanism combines data from rapid weight loss measured by a body fat scale and blood glucose level fluctuations detected by a blood glucose meter for comprehensive judgment. Only when multiple relevant indicators simultaneously meet the warning trigger conditions is it considered a valid warning and the emergency response process initiated.

[0031] (III) Beneficial Effects

[0032] Compared with existing technologies, this invention provides a personalized nutritional intervention system for patients with locally advanced pancreatic cancer, which has the following beneficial effects:

[0033] 1. This personalized nutritional intervention system for patients with locally advanced pancreatic cancer, by connecting to multiple low-cost monitoring devices, collects patients' metabolic, physiological and tumor-specific omics data in real time and comprehensively. It can more sensitively capture nutritional risks that may occur during treatment, provide a reliable basis for timely adjustment of nutritional intervention plans, effectively solve the problem of insufficient monitoring data in existing technologies, improve the accuracy of nutritional assessment, and provide rich data support for personalized nutritional intervention.

[0034] 2. This personalized nutritional intervention system for patients with locally advanced pancreatic cancer utilizes advanced AI algorithms and machine learning models. The system dynamically adjusts the nutritional intervention plan based on real-time monitoring data and treatment variables. The nutritional intervention plan optimization model based on deep learning algorithms can automatically adjust the proportion and intake of nutrients according to individual patient characteristics and real-time data. Compared with traditional fixed-mode nutritional intervention, it can better meet the nutritional needs of patients at different treatment stages and physiological states, improve patients' tolerance to treatment and recovery speed, and reduce errors and delays caused by manual adjustment of the plan, providing more timely and effective nutritional support.

[0035] 3. This personalized nutritional intervention system for patients with locally advanced pancreatic cancer is based on a comprehensive and rigorously validated rule base. The system automatically identifies potential nutritional crisis events and quickly activates emergency nutritional support pathways. In a home environment, when patients experience severe hypoglycemia, electrolyte imbalances, or other crisis situations, the system automatically guides patients or their families on the correct administration of nasogastric feeding solutions and promptly notifies medical staff. This function fills a gap in existing technologies regarding patient safety at home, reduces the risk of complications caused by untimely handling of nutritional crises, and improves the safety and quality of life of patients in their home environment. Compared to traditional methods that rely on patients or their families to make their own judgments and handle crisis events, the system can identify crises more quickly and accurately, provide standardized emergency response measures, and ensure that patients receive effective nutritional support and medical assistance in critical moments. Attached Figure Description

[0036] Figure 1 This is a flowchart illustrating the device connection and data collection process of an individualized nutritional intervention system for patients with locally advanced pancreatic cancer, as proposed in this invention.

[0037] Figure 2 This is a data preprocessing flowchart for an individualized nutritional intervention system for patients with locally advanced pancreatic cancer proposed in this invention;

[0038] Figure 3 This is a flowchart illustrating the nutritional requirement prediction process for an individualized nutritional intervention system for patients with locally advanced pancreatic cancer, as proposed in this invention.

[0039] Figure 4 This is a flowchart of the optimized nutritional intervention program for a personalized nutritional intervention system for patients with locally advanced pancreatic cancer proposed in this invention.

[0040] Figure 5 This is an automated emergency response flowchart for an individualized nutritional intervention system for patients with locally advanced pancreatic cancer, as proposed in this invention. Detailed Implementation

[0041] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0042] Multi-dimensional dynamic monitoring module

[0043] Device connectivity and data collection: Patients are equipped with a variety of monitoring devices. Wearable physiological monitoring devices transmit physiological data via Bluetooth Low Energy technology; portable blood testing devices automatically upload blood test results; metabolic monitoring devices transmit energy consumption data in real time; and tumor-specific omics data detection devices transmit tumor-related data.

[0044] Data preprocessing: The system's data preprocessing unit cleans, merges, and standardizes data, removes outlier data, associates different types of data, and unifies data formats and units.

[0045] Dynamic adaptive intervention module

[0046] Data feature extraction and analysis: The system extracts data features to construct patient feature vectors, including features such as heart rate variability, blood pressure fluctuation range, and energy consumption ratio, and combines treatment variables and basic information to form a comprehensive feature vector.

[0047] A deep learning-based nutritional requirement prediction model: The prediction model is constructed using LSTM, and the state transition equation of the LSTM model is as follows:

[0048] h t =σ(W h ·[h t―1 ,x t ]+b h ),c t =f t ⊙c t―1 +i t ⊙tanh(W c ·[h t―1 ,x t ]+b c )

[0049] Input historical and current data to predict future nutritional needs. During the training process, mean squared error is used as the loss function, and the backpropagation algorithm is used to adjust parameters and optimize model performance.

[0050] Nutritional intervention program optimization based on multi-objective optimization algorithm: The genetic algorithm is used to optimize the parameters of the nutritional program, generate candidate programs and evaluate and screen them to determine the optimal program that meets the nutritional needs of patients and adapts to their taste preferences.

[0051] Automatic emergency response module

[0052] Crisis event identification rule base construction: The rule base clarifies the crisis event level and identification rules, such as rules for severe hypoglycemia and electrolyte disturbance events.

[0053] Emergency response procedures are automatically triggered and executed: After a crisis event is identified, the system initiates emergency response, pushes suggestions and instructions for nasogastric feeding, and notifies medical staff;

[0054] Emergency response effectiveness evaluation and feedback: The system monitors physiological indicators to evaluate the effectiveness. If the results are normal, the system records and optimizes the response; if the results are abnormal, the response measures are upgraded.

[0055] The rule base contains three crisis levels:

[0056] Level 1 Crisis: Serum potassium > 6.0 mmol / L and ECG shows peaked T waves. Triggering mechanism:

[0057] Emergency measures = {nasogastric feeding of low-potassium nutritional solution, notify emergency physician, initiate remote electrocardiogram monitoring};

[0058] Level 2 Crisis: Blood glucose 3.0-3.9 mmol / L. Triggering Procedure: Emergency Measures = {Push oral glucose instructions, retest blood glucose every 15 minutes} The system executes the emergency procedure through a state machine model, assesses the effectiveness of the treatment every 5 minutes, and upgrades the crisis level if the indicators do not improve.

[0059] The emergency response module includes suggestions for temporary dietary adjustments, emergency contact information, indications for immediate medical attention, and a one-click connection to remote medical services. Temporary dietary adjustments are generated based on the patient's symptoms and risk level. For example, for a level one warning of intestinal obstruction risk, the suggestion is to immediately stop eating, take oral medication, use castor oil as a laxative, and go to the hospital as soon as possible.

[0060] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0061] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A personalized nutritional intervention system for patients with locally advanced pancreatic cancer, characterized in that, include: Multi-dimensional dynamic monitoring module: Connects to multiple monitoring devices to collect real-time metabolic, physiological and tumor-specific omics data of patients; The monitoring equipment includes wearable physiological monitoring devices, portable blood testing devices, metabolic monitoring devices, and tumor-specific omics data detection devices; Dynamic Adaptive Intervention Module: Utilizing AI algorithms and machine learning models, this module dynamically adjusts nutritional intervention plans based on real-time monitoring data and treatment variables. The AI ​​algorithms include deep learning algorithms and multi-objective optimization algorithms, used to construct nutritional demand prediction models and optimize nutritional intervention plans. On a cloud server platform, the trained AI model (an LSTM-based deep learning model) is deployed as an independent microservice module, packaged and managed using Docker containerization technology to ensure stable operation and good scalability across different server environments. The model service interacts with other system modules through a RESTful API interface, receiving real-time data from the multi-dimensional dynamic monitoring module, performing predictive analysis, and returning results. Automatic emergency response module: Based on a rule base, it automatically identifies crisis events and activates emergency nutritional support pathways to ensure the safety of the patient's home environment; the rule base was jointly built by medical experts, nutritionists and data scientists, covering a variety of nutrition-related crisis events and their identification rules.

2. The personalized nutritional intervention system for patients with locally advanced pancreatic cancer according to claim 1, characterized in that: The data collected by the multi-dimensional dynamic monitoring module includes heart rate, blood pressure, blood glucose, energy consumption, tumor markers, and circulating tumor DNA. The subjective feeling data collected by the multi-dimensional dynamic monitoring module is recorded through quantitative assessment tools provided by the mobile application interface, such as using the visual analog scale to quantitatively score subjective symptoms such as nausea and pain from 0 to 10.

3. The personalized nutritional intervention system for patients with locally advanced pancreatic cancer according to claim 1, characterized in that: The deep learning algorithm in the dynamic adaptive intervention module is a Long Short-Term Memory (LSTM) network, used to construct a nutritional requirement prediction model. The prediction model formula is as follows: Y t+h =LSTM(X t ,Y t―1 ) Among them, Y t+h X represents the predicted nutritional requirements of the patient at a future time t+h. t Y represents the multi-dimensional monitoring data at the current time t. t-1 This represents the patient's historical nutritional needs data.

4. The personalized nutritional intervention system for patients with locally advanced pancreatic cancer according to claim 1, characterized in that: The multi-objective optimization algorithm in the dynamic adaptive intervention module is a genetic algorithm, used to optimize the parameters of the nutritional intervention program. The multi-objective optimization algorithm formula is as follows: min f(X)=[f1(X),f2(X),…,f n (X)] TPI ne w=min(BI×(1+ΔP / 100),P max ) Where f(X) represents the optimization objective function vector, f i (X) represents the i-th optimization objective, X represents the parameter vector of the nutritional intervention program, (P max (This is the upper limit for protein intake, dynamically calculated based on renal function indicators).

5. The personalized nutritional intervention system for patients with locally advanced pancreatic cancer according to claim 1, characterized in that: The rule base of the automatic emergency response module covers identification rules for events such as severe hypoglycemia and electrolyte imbalance.

6. The personalized nutritional intervention system for patients with locally advanced pancreatic cancer according to claim 1, characterized in that: When the emergency nutritional support pathway is activated, the automatic emergency response module automatically guides the operation of the nasogastric feeding solution and notifies medical staff.

7. The personalized nutritional intervention system for patients with locally advanced pancreatic cancer according to claim 1, characterized in that: It also includes a data preprocessing unit, which is used to clean, merge and standardize the data collected by the multi-dimensional dynamic monitoring module.

8. The personalized nutritional intervention system for patients with locally advanced pancreatic cancer according to claim 1, characterized in that: When optimizing nutritional intervention programs, the dynamic adaptive intervention module constructs a patient feature vector by combining patient treatment variables and basic information.

9. The personalized nutritional intervention system for patients with locally advanced pancreatic cancer according to claim 1, characterized in that: The automatic emergency treatment module continuously monitors changes in the patient's physiological indicators and evaluates the effectiveness of the emergency treatment in real time.

Images