Intelligent prp precision preparation system based on blood analysis and centrifugation

Abstract

The application provides an intelligent PRP precise preparation system based on blood analysis and centrifugation, and belongs to the technical field of medical examination and intelligent control of medical devices. The system comprises an input module, an optimization module and an output module. The input module inputs blood basic parameters; the optimization module optimizes centrifugation prescription parameters through an optimization algorithm, the optimization algorithm takes minimization of a cost function as an objective, searches for optimal centrifugation intensity parameters, centrifugation time parameters, acceleration control parameters and deceleration control parameters under the constraints of centrifuge device physical constraints and clinical safety rules; and the output module outputs the optimized centrifugation prescription parameters. The application realizes automatic generation of individualized centrifugation process parameters directly driven by patient blood baseline data, and improves the consistency and quality of PRP preparation.

Description

Technical Field

[0001] This invention relates to the field of medical testing and intelligent control technology for medical devices, and in particular to an intelligent platelet-rich plasma (PRP) precision preparation system based on communication between a blood analyzer and a centrifuge, which is applicable to clinical scenarios such as orthopedics, sports medicine, medical aesthetics and regenerative medicine. Background Technology

[0002] Platelet-Rich Plasma (PRP) is a product containing concentrated platelets and growth factors extracted from a patient's own blood. It is widely used in orthopedics (e.g., arthritis treatment), dermatology (e.g., wound repair), and cosmetic medicine (e.g., hair regeneration). PRP preparation typically involves blood collection, anticoagulation, and centrifugation. Traditional methods employ a two-step centrifugation process: an initial low-speed separation of red blood cells, followed by a second high-speed concentration of platelets.

[0003] Currently, PRP preparation mainly relies on centrifugation of whole blood samples in one or more stages. In existing technologies, centrifugation parameters (such as centrifugal force and centrifugation time) typically use fixed empirical values ​​or preset protocols. For example, Arthrex Angel... ® The cPRP system is an automated centrifugation device that provides 2-3 times platelet concentration in approximately 15.8 minutes, but its parameters are fixed and cannot be adjusted according to blood parameters. The Celling Biosciences ART PRP system uses a nanoporous fiber system that allows selection of the centrifugation layer, but still relies on fixed parameters. The ProGen PRP system supports different concentration levels, but does not enable real-time blood analysis and dynamic parameter adjustment.

[0004] However, due to significant differences in platelet count (PLT), hematocrit (HCT), white blood cell count (WBC), and plasma protein content among different patients, using uniform centrifugation parameters can easily lead to the following problems: (1) large fluctuations in platelet concentration factor and poor batch-to-batch consistency; (2) unstable platelet recovery rate and high activity loss in some samples; (3) increased risk of white blood cell or red blood cell contamination; (4) the need for repeated manual adjustment of parameters, operation relies on experience, and efficiency is low.

[0005] Meanwhile, while some existing technologies introduce algorithms to optimize the separation process, they mostly focus on controlling the PRP concentration during centrifugation, lacking a mechanism for generating individualized centrifugation parameters based on the patient's baseline blood data, and there is no automatic linkage between blood analysis equipment and centrifugation equipment at the system level.

[0006] For example, patent CN119140292B proposes a closed-loop control method for the PRP separation process. Its core is to train a concentration prediction model using historical data and dynamically adjust system parameters (such as separation rotation speed) through an intelligent optimization algorithm based on the real-time monitored PRP concentration during the separation process, thereby achieving precise control of the target PRP concentration. This method emphasizes real-time feedback and adjustment during the separation process.

[0007] However, existing technologies lack an intelligent PRP preparation system capable of generating personalized centrifugation prescriptions based on patient baseline blood data and automatically controlling the centrifuge via communication. Therefore, there is an urgent need for a precise PRP preparation solution that integrates blood analysis instruments and centrifuges at the system level, driving personalized process settings based on differences in patient baseline blood data. Summary of the Invention

[0008] The purpose of this invention is to provide an intelligent PRP precision preparation system based on blood analysis and centrifugation. By driving the optimization of centrifugation parameters through blood analysis results, the PRP preparation process can be individualized, standardized and automated, thereby improving the platelet concentration factor, recovery rate and preparation consistency.

[0009] The complete technical solution provided by this invention is as follows: This invention provides an intelligent PRP precision preparation system based on blood analysis and centrifugation, comprising: The input module is configured to take in the basic blood parameters of a blood sample. The optimization module is configured to optimize the centrifugation prescription parameters of blood samples using an optimization algorithm; The centrifugation prescription parameters include centrifugation intensity parameters, centrifugation time parameters, acceleration control parameters, and deceleration control parameters; the centrifugation intensity parameter is centrifugal force or equivalent rotational speed, the acceleration control parameter is acceleration slope or acceleration level, and the deceleration control parameter is deceleration slope or deceleration level. The cost function of the optimization algorithm is: Where x is the feature vector of basic blood parameters; u is the centrifugation prescription parameter vector to be optimized. This is a predicted value for platelet concentration factor. This is the predicted recovery rate. This is a predicted value for a white blood cell contamination index. This is the predicted value for red blood cell residue. This is the predicted total processing time. For the concentration factor interval penalty function, This is a penalty function for the recovery rate threshold. , , , , These are the weighting coefficients; The constraints of the optimization algorithm include the physical constraints of the centrifuge equipment and clinical safety rules. The physical constraints of the centrifuge equipment include: centrifugation intensity parameter constraints. Centrifugation time parameter constraints: Acceleration control parameter constraints: Deceleration control parameter constraints: ;in, For the first i Centrifugation intensity parameters of the stage, For the first i Centrifugation time parameters for each stage For the first i Acceleration control parameters for each stage For the first i Phase deceleration control parameters, i =1 or 2, corresponding to the first and second stage centrifugation, respectively; the clinical safety rules include at least one of the following rules: when the hematocrit is higher than the clinical threshold for hematocrit, limit the upper limit of the second stage centrifugation intensity parameter or increase the lower limit of the first stage centrifugation time parameter; when the platelet count is lower than the clinical threshold for platelet count, adjust the optimization target to prioritize recovery rate; when the white blood cell count is higher than the clinical threshold for white blood cell count, increase the weighting coefficient of the white blood cell contamination index. ; The output module is configured to output the optimized centrifugation prescription parameters u. .

[0010] Furthermore, the basic blood parameters include at least the platelet count (…). PLT ), hematocrit ( HCT ) and white blood cell count ( WBC ); optionally, it may also include total plasma protein ( TP ), blood viscosity ( η ) and other parameters.

[0011] Furthermore, the centrifugation prescription parameter vector: .

[0012] Furthermore, the concentration factor interval penalty function: The recovery rate threshold penalty function is: The , , , , The value range is 0.1-10.

[0013] Furthermore, the optimization module includes: The quality prediction model unit is configured to output a PRP quality indicator prediction vector based on the input feature vector x of the basic blood parameters and the centrifugation prescription parameter vector u to be optimized. The quality prediction model unit may employ nonlinear regressors such as gradient boosting trees, random forests, or lightweight neural networks. The parameter search unit is configured to use the cost function With the goal of minimization, the optimal centrifugation recipe parameters are searched under the physical constraints of the centrifuge equipment. The rule constraint unit is configured to perform a final verification of the optimal centrifugation prescription parameters output by the parameter search unit using the clinical safety rule constraints, and generate the optimized centrifugation prescription parameters u. .

[0014] Furthermore, the optimization module also includes a data storage unit configured to record preparation data, forming a traceable log.

[0015] Furthermore, the parameter search unit employs a hybrid iterative optimization algorithm that combines local search and global jump. The hybrid iterative optimization algorithm includes: normalizing and encoding the centrifugal prescription parameter vector u to be optimized into a search variable z; initializing a candidate prescription group; evaluating the fitness of the candidate prescriptions; updating the candidate solutions by learning from the current optimal solution and combining random perturbation; performing a jump operation when it is determined that the solution is trapped in a local optimum; and performing boundary repair and feasibility projection on the updated candidate solutions.

[0016] Furthermore, the search variable The encoding formula is: ,in, , The first i Minimum and maximum values ​​of the prescription parameters for each stage; The candidate prescription fitness evaluation function is: ,in For decoding function, For the first r In the first iteration j There are 10 candidate solutions; The candidate solution update formula is: ,in For the first r The optimal candidate in rounds of iteration, It is a random vector. This is the local search learning rate, ranging from 0.1 to 1.5. This is the random disturbance coefficient, with a value range of 0.01-0.5; The jump operation determination condition is: when consecutiveq Execute when the optimal fitness is not improved in the first iteration, where... q The threshold for the number of consecutive rounds without improvement is 5-50; the jump operation formula is: ,in s This represents the jump step size, with a value ranging from 0.1 to 1.0. The boundary repair operation is as follows: if out-of-bounds components appear after the update... or Execute a bounce or truncation: ; The feasibility projection operation is as follows: perform projection on the hard constraints. The projection rules include: (1) Interval constraint projection: for centrifugal strength Centrifugation time Acceleration control Deceleration control If the corresponding component Exceeding the constraint range If so, the projection is truncated to the nearest boundary point; (2) Coupling constraint projection: For clinical safety rule constraints (such as the upper limit of intensity under the influence of hematocrit HCT), the candidate solution is projected back into the feasible region Ω using the minimum Euclidean distance principle; (3) Calculation formula: .

[0017] Furthermore, the system also includes: The communication module is configured to enable bidirectional data communication between the output module and the centrifuge; The centrifuge is configured to receive the optimized centrifugation prescription parameters u via the communication module. It also automatically executes a multi-stage centrifugation process.

[0018] The communication module supports communication methods including but not limited to: serial communication, Bluetooth communication, Wi-Fi communication, Modbus protocol, or network-based REST interface.

[0019] During centrifugation, the centrifuge controls the rotation speed according to the acceleration and / or deceleration parameters, thereby reducing platelet shear damage and improving preparation consistency.

[0020] Furthermore, the system also includes: The user interface module is configured to display the basic blood parameters and the optimized centrifugation prescription parameters. It displays the centrifugation process status and provides human-computer interaction.

[0021] This invention also provides a method for precise PRP preparation based on the above system, comprising the following steps: (1) Receive basic blood parameters output by the blood analyzer through the input module; (2) The optimization module runs the optimization algorithm, calls the quality prediction model, and optimizes the centrifugation prescription parameters based on the cost function and constraints; (3) Output the optimized centrifugation prescription parameters; (4) The optimized centrifugation prescription parameters are sent to the centrifuge via the communication module; (5) The centrifuge automatically executes a multi-stage centrifugation process based on the received parameters to complete the PRP preparation.

[0022] The present invention also provides a computer-readable storage medium storing a computer program, wherein the computer program is configured to execute the functions of the above-described intelligent PRP precision preparation system based on communication between a blood analyzer and a centrifuge when it is run.

[0023] The present invention has achieved the following beneficial effects: (1) Generate individualized centrifugation prescriptions based on blood analysis results to improve the consistency of PRP preparation; (2) To achieve automatic communication and collaborative control between blood analyzers and centrifuges, reducing manual intervention; (3) Improve platelet recovery rate and reduce contamination risk through multi-objective optimization and rule constraints; (4) Supports data recording and continuous algorithm optimization, applicable to different patient groups and different device platforms.

[0024] Compared with the existing technology CN119140292B, the biggest difference of this invention is that: the "controlled object / closed-loop signal source" is different: this technology is "blood analysis → centrifugation parameters", while the existing technology is "separation process → concentration control"; the "system form" is different: this technology emphasizes "cross-device communication + software platform", while the other technology is more inclined to "algorithm / control method".

[0025] Obviously, based on the above description of the present invention, and according to common technical knowledge and conventional methods in the field, various other modifications, substitutions or alterations can be made without departing from the basic technical concept of the present invention.

[0026] The following detailed embodiments further illustrate the above-described content of the present invention. However, this should not be construed as limiting the scope of the present invention to the following embodiments. All technologies implemented based on the above-described content of the present invention fall within the scope of the present invention. Detailed Implementation

[0027] The raw materials and equipment used in this invention are all known products, obtained by purchasing commercially available products.

[0028] Example 1: Construction and application of an intelligent PRP precision preparation system based on blood analysis and centrifugation I. Training and Construction of Quality Prediction Model In this embodiment, the training and construction process of the quality prediction model is as follows: S1: Data Acquisition and Preprocessing Historical data collection and preparation records are used to form a training dataset. ,in: For the first k Blood baseline feature vectors for each sample; For the first k The actual centrifugation prescription parameters used for each sample; For the first k The actual PRP quality indicators obtained after the sample preparation was completed.

[0029] S2: Model Training Training quality prediction model ,accomplish The training objective is to minimize the multi-output error. Among them, key indicators (such as recovery rate) R leukocyte contamination W Set a higher weight.

[0030] S3: Model Validation and Deployment Evaluate the model's prediction performance using an independent validation set to ensure that the prediction error is within an acceptable range. Then, evaluate the trained model... It is deployed to the quality prediction model unit of the data processing software module for online prediction.

[0031] II. System Construction This invention provides an intelligent PRP (Potentially Differentiated Reproductive Technology) precision preparation system based on blood analysis and centrifugation, comprising: an input module, an optimization module, an output module, a communication module, a centrifuge, and a user interface module. The system generates centrifugation parameters based on blood analysis results, and the centrifuge automatically executes the corresponding multi-stage centrifugation process.

[0032] 1. Input module The system uses a fully automated hematology analyzer (such as the Sysmex or Beckman Coulter series) as the front-end device for input modules to analyze patient whole blood samples and output basic blood parameters, including but not limited to platelet count. PLT ), hematocrit ( HCT ), white blood cell count ( WBC ), total plasma protein ( TP ), blood viscosity ( The blood analyzer can output test results through a standardized data interface, which includes, but is not limited to, HL7 protocol, CSV file, database interface, etc.

[0033] 2. Optimization module Developed using the Python language on a computer (or embedded system) connected to the aforementioned device. This module includes: (1) Quality prediction model unit Deploy the trained gradient boosting tree model It is used to output a PRP quality index prediction vector based on the input blood baseline feature vector x and the centrifugation prescription parameter vector u to be optimized. .

[0034] (2) Parameter search unit After receiving the new blood baseline feature vector x, the following optimization steps are performed: ① Read the device constraint set Ω, including: ; ② Call the quality prediction model ; ③ Construct the comprehensive cost function: in, The penalty term is defined as follows: , ; ④Prescription parameters Normalized encoding yields the search variable : ; ⑤ Initialize the candidate prescription group This includes coding points mapped from a preset protocol library and perturbation sampling around the protocol library. and random sampling within the feasible region; ⑥ Iterative search for the optimal prescription: for each iteration round ,implement: Fitness assessment: ; Candidate Updates: ,in It ranges from 0.1 to 1.5. It is 0.01-0.5; Jump operation: when consecutive q If there is no improvement after 10-30 rounds, proceed. ,in s The range is 0.1-1.0; Boundary repair: ; Feasibility projection: ; ⑦ Rule validation: when HCT Above the threshold HCT high (50%), restrictions The upper limit may be increased. Lower limit; when PLT Below the threshold PLT low (100×10) 9 When / L), a prompt is triggered stating "It may not be possible to achieve 4-6 times the target," and the objective function weights are changed to "recovery rate priority"; when WBC Above the threshold WBC high (10×10) 9 When / L), increase the γ weight to enhance the low leukocyte contamination target; generate the final prescription. .

[0035] (3) Rule constraint unit It is used to limit, correct, or issue prompts on centrifugation prescription parameters when the basic blood parameters or algorithm output parameters exceed preset thresholds.

[0036] (4) Data storage unit Use an SQLite database to record the data x generated each time. And the actual quality indicator vector y, forming a traceable log.

[0037] 3. Output module Optimized centrifugation prescription parameters Output to the communication module.

[0038] 4. Communication module It supports serial communication, Bluetooth communication, Wi-Fi communication, Modbus protocol or network-based REST interface to realize bidirectional communication between the optimization module and the centrifuge.

[0039] 5. Centrifuge Employing a programmable PRP centrifuge (such as the Ortoalresa Plasma 22 which supports Modbus protocol communication), it can receive and execute commands that include multi-stage centrifugal force, time, and acceleration / deceleration levels.

[0040] 6. User Interface Module Developed using PyQt5, this system graphically displays basic blood parameters, algorithm-recommended individualized centrifugation prescription parameters, and expected PRP quality indicators; it provides interactive operations for manual confirmation, modification, or overriding; and it displays the centrifugation process status and execution results.

[0041] III. Application Process: (1) Collect blood from the patient, one copy for routine blood analysis and one copy for PRP preparation; (2) The blood analyzer automatically detects and outputs the patient's baseline blood feature vector. ; (3) The optimization module receives data and runs a parameter optimization algorithm to generate individualized centrifugation prescriptions. And display it on the UI; (4) After the operator confirms that the prescription is correct on the UI, click "Confirm Execution"; (5) The software transmits prescription parameters via the Modbus protocol. Send to centrifuge; (6) The centrifuge automatically performs two-stage centrifugation: the first stage is as follows: , , , Centrifugation with parameters, second stage according to , , , Parameter centrifugation; (7) After centrifugation, the operator removes the PRP and tests the actual quality indicators. The data is entered into the system and stored in the SQLite database.

[0042] In summary, this invention provides an intelligent PRP precision preparation system based on blood analysis and centrifugation. This system integrates input, optimization, output, communication, centrifuge, and user interface modules to construct an automated closed-loop process that directly drives individualized centrifugation from patient blood data. Its core lies in the parameter optimization algorithm, which uses patient baseline blood data as input and is guided by multi-objective quality optimization. Under the constraints of equipment physical and safety limitations, it generates the optimal centrifugation prescription through a strategy of "quality prediction model + multi-objective comprehensive cost function + iterative search + rule verification," without relying on real-time PRP concentration signals during the centrifugation process.

Claims

1. An intelligent PRP precision preparation system based on blood analysis and centrifugation, characterized in that, include: The input module is configured to take in the basic blood parameters of a blood sample. The optimization module is configured to optimize the centrifugation prescription parameters of blood samples using an optimization algorithm; The output module is configured to output the optimized centrifugation prescription parameters u. ; The centrifugation prescription parameters include centrifugation intensity parameters, centrifugation time parameters, acceleration control parameters, and deceleration control parameters; the centrifugation intensity parameter is centrifugal force or equivalent rotational speed, the acceleration control parameter is acceleration slope or acceleration level, and the deceleration control parameter is deceleration slope or deceleration level. The cost function of the optimization algorithm is: Where x is the feature vector of basic blood parameters; u is the centrifugation prescription parameter vector to be optimized. This is a predicted value for platelet concentration factor. This is the predicted recovery rate. This is a predicted value for a white blood cell contamination index. This is the predicted value for red blood cell residue. This is the predicted total processing time. For the concentration factor interval penalty function, This is a penalty function for the recovery rate threshold. , , , , These are the weighting coefficients; The constraints of the optimization algorithm include the physical constraints of the centrifuge equipment and clinical safety rules. The physical constraints of the centrifuge equipment include: centrifugation intensity parameter constraints. Centrifugation time parameter constraints: Acceleration control parameter constraints: Deceleration control parameter constraints: ;in, For the first i Centrifugation intensity parameters of the stage, For the first i Centrifugation time parameters for each stage For the first i Acceleration control parameters for each stage For the first i Phase deceleration control parameters, i =1 or 2, corresponding to the first and second stage centrifugation, respectively; the clinical safety rules include at least one of the following rules: when the hematocrit is higher than the clinical threshold for hematocrit, limit the upper limit of the second stage centrifugation intensity parameter or increase the lower limit of the first stage centrifugation time parameter; when the platelet count is lower than the clinical threshold for platelet count, adjust the optimization target to prioritize recovery rate; when the white blood cell count is higher than the clinical threshold for white blood cell count, increase the weighting coefficient of the white blood cell contamination index. .

2. The system according to claim 1, characterized in that, The basic blood parameters include at least platelet count, hematocrit, and white blood cell count; the centrifugation prescription parameter vector: .

3. The system according to claim 1, characterized in that, The concentration factor interval penalty function: The recovery rate threshold penalty function is: The , , , , The value range is 0.1-10.

4. The system according to claim 1, characterized in that, The optimization module includes: The quality prediction model unit is configured to output a PRP quality indicator prediction vector based on the input feature vector x of the basic blood parameters and the centrifugation prescription parameter vector u to be optimized. ; The parameter search unit is configured to use the cost function With the goal of minimization, the optimal centrifugation recipe parameters are searched under the physical constraints of the centrifuge equipment. The rule constraint unit is configured to perform a final verification of the optimal centrifugation prescription parameters output by the parameter search unit using the clinical safety rule constraints, and generate the optimized centrifugation prescription parameters u. .

5. The system according to claim 4, characterized in that, The parameter search unit employs a hybrid iterative optimization algorithm that combines local search and global jump. The hybrid iterative optimization algorithm includes: normalizing and encoding the centrifugal prescription parameter vector u to be optimized into a search variable z; initializing a candidate prescription group; evaluating the fitness of the candidate prescriptions; updating the candidate solutions by learning from the current optimal solution and combining random perturbation; performing a jump operation when it is determined that the solution is trapped in a local optimum; and performing boundary repair and feasibility projection on the updated candidate solutions.

6. The intelligent PRP precision preparation system according to claim 5, characterized in that, The search variable The encoding formula is: ,in, , The first i Minimum and maximum values ​​of the prescription parameters for each stage; The candidate prescription fitness evaluation function is: ,in For decoding function, For the first r In the first iteration j There are 10 candidate solutions; The candidate solution update formula is: ,in For the first r The optimal candidate in rounds of iteration, For random vectors, This is the local search learning rate, ranging from 0.1 to 1.

5. This is the random disturbance coefficient, with a value range of 0.01-0.5; The jump operation determination condition is: when consecutive q Execute when the optimal fitness is not improved in the first iteration, where... q The threshold for the number of consecutive rounds without improvement is 5-50; the jump operation formula is: ,in s This represents the jump step size, with a value ranging from 0.1 to 1.

0. The boundary repair operation is as follows: if out-of-bounds components appear after the update... or Execute a bounce or truncation: ; The feasibility projection operation is as follows: perform projection on the hard constraints. The projection rules include: (1) Interval constraint projection: for centrifugal strength Centrifugation time Acceleration control Deceleration control If the corresponding component Exceeding the constraint range If so, the projection is truncated to the nearest boundary point; (2) Coupling constraint projection: For clinical safety rule constraints, the candidate solution is projected back into the feasible region Ω using the minimum Euclidean distance principle; (3) Calculation formula: .

7. The intelligent PRP precision preparation system according to claim 1, characterized in that, The system also includes: The communication module is configured to enable bidirectional data communication between the output module and the centrifuge; The centrifuge is configured to receive the optimized centrifugation prescription parameters u via the communication module. It also automatically executes a multi-stage centrifugation process.

8. The intelligent PRP precision preparation system according to claim 1, characterized in that, The system also includes: The user interface module is configured to display the basic blood parameters and the optimized centrifugation prescription parameters. It displays the centrifugation process status and provides human-computer interaction.

9. A method for precise preparation of PRP based on the system described in any one of claims 1 to 8, characterized in that, Includes the following steps: (1) Receive basic blood parameters output by the blood analyzer through the input module; (2) The optimization module runs the optimization algorithm, calls the quality prediction model, and optimizes the centrifugation prescription parameters based on the cost function and constraints; (3) Output the optimized centrifugation prescription parameters; (4) The optimized centrifugation prescription parameters are sent to the centrifuge via the communication module; (5) The centrifuge automatically executes a multi-stage centrifugation process based on the received parameters to complete the PRP preparation.

10. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, wherein the computer program is configured to execute the functions of the intelligent PRP precision preparation system based on communication between a blood analyzer and a centrifuge as described in any one of claims 1 to 8 when it is run.

Images