Automatic calculation system for measuring renal function through double-plasma method

Abstract

The invention relates to the technical field of medical systems, and particularly discloses an automatic calculation system for measuring renal functions by a double-plasma method, which comprises a data filling module, a calculation module and an output module, the data filling module is used for acquiring patient information, standard source data and drug activity data, and is communicated with the radioactivity counter to automatically read a probe measurement value; the calculation module automatically generates a GFR value and a standardized GFR value based on the input parameters; and the output module exports the calculation result to the LIS system and generates a visual report. According to the invention, through full-link technology innovation of'detection-calibration-calculation-decision ', a renal function measurement system with independent intellectual property rights is constructed, the blank in the field of intelligence and precision in the prior art is filled, and the clinical diagnosis efficiency and safety are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical systems, in particular to an automatic calculation system for measuring kidney function by double plasma method. BACKGROUND

[0002] In the current clinical practice of measuring kidney function by double plasma method, the following technical bottlenecks exist:

[0003] Complicated calculation process and prone to errors: the traditional method relies on manual calculation or Excel table to execute complex formulas, such as GFR calculation which needs to integrate multiple parameters such as radioactivity, plasma count, body surface area, etc. Manual operation is prone to calculation errors, and errors are difficult to trace back;

[0004] Inadequate device calibration mechanism: there is inherent bias in multi-probe measurement of radioactivity counter, and existing technology does not realize dynamic calibration of probe sensitivity. If the probe is replaced or not calibrated during different measurement stages, the counting error will exceed ±12%;

[0005] Lack of compensation for environmental and physiological factors: traditional systems do not consider the influence of environmental variables such as temperature and humidity on radionuclide decay, nor do they integrate physiological parameters such as patient age, gender, and race to correct GFR, resulting in a large deviation between measured results and actual kidney function;

[0006] Insufficient real-time and anomaly detection: unable to monitor data anomalies (such as plasma count decay rate deviating from the theoretical range) in real time, lack of automatic warning mechanism for acute kidney injury (AKI) and other emergency situations, leading to delayed clinical decision-making;

[0007] Low data management efficiency: patient information needs to be filled in manually, body surface area (BSA) calculation does not distinguish between children, obese and other special body types, and lacks automatic integration with hospital LIS system, resulting in time-consuming data export and easy loss. SUMMARY

[0008] To overcome the shortcomings of the prior art, the present application provides an automatic calculation system for measuring kidney function by double plasma method, which solves the problems in the above background art.

[0009] To achieve the above purpose, the present application realizes the following technical scheme: an automatic calculation system for measuring kidney function by double plasma method, comprising:

[0010] A data filling module for obtaining patient information, standard source data and drug activity data, and communicating with a radioactivity counter to automatically read probe measurement values;

[0011] A calculation module for automatically generating GFR values and standardized GFR values based on input parameters;

[0012] an output module for exporting the calculation result to a LIS system and generating a visual report;

[0013] The data filling module has a dynamic checking unit built-in, which performs the following checks: an alarm is triggered when the decay rate of two plasma count values exceeds the theoretical decay range of the radionuclide; if the standard source count fluctuation is > 5%, the probe self-check program is automatically started; the patient's wristband information is read by a code scanning gun to fill in the height, weight and medical history data; the optimal sampling time window is generated based on the injection time, and the dynamic compensation algorithm is started when the deviation occurs; the probe sensitivity parameters are automatically matched according to the nuclide type, and an alarm is given when they are not matched.

[0014] Preferably, the calculation module adopts a time-probe double coupling correction algorithm, and the formula is:

[0015] ;

[0016] wherein, is the deviation of the measurement time from the theoretical time, is the nuclide decay constant, which is related to the half-life of the radionuclide, is the half-life of the nuclide, is a non-linear decay compensation factor;

[0017] When multiple probes are used, the coefficients are optimized through distributed calculation, and the formula is: ; wherein, is the number of probes participating in the calculation, is the calibration coefficient of the th probe, is the standard deviation of the probe, is the adaptive weight;

[0018] The execution steps of the time-probe double coupling correction algorithm include: reading the current calibration coefficient and the reference coefficient from the probe calibration coefficient library, calculating the deviation of the measurement time from the theoretical sampling time; calculating the nuclide decay constant according to , generating a non-linear compensation term and taking the exponential; calculating the ratio correction term of the measured value to the calibration coefficient, and multiplying it by the decay correction term to obtain the final correction value .

[0019] Preferably, the construction method of the probe calibration coefficient library includes:

[0020] The coefficients are updated daily according to ;

[0021] If the calibration deviation of the same probe is > 3% for three consecutive times, it is marked as a fault;

[0022] By aging compensation model Correction coefficient, The first The calibration coefficient of the first probe at the use time , The initial calibration coefficient of the first The first probe, The aging rate;

[0023] Environmental adaptive calibration is performed , The calibration coefficient of the first probe after environmental factor correction , Temperature compensation coefficient, fixed value , The temperature of the real-time measurement environment , Humidity compensation coefficient, fixed value , The relative humidity of the real-time measurement environment

[0024] , Wavelet denoising is used for low count rate samples

[0025] The environmental adaptive calibration step of the probe calibration coefficient library includes:

[0026] Real-time acquisition of environmental temperature T and humidity RH, according to Correction coefficient;

[0027] When Or , trigger environmental compensation mark.

[0028] Preferably, the GFR calculation introduces multi-dimensional composite correction:

[0029] ;

[0030] Wherein, The dose of radiopharmaceutical, The radioactivity count correction value of the two plasma samples, The time difference between the two plasma samples, in minutes, used to calculate the drug clearance rate, The actual body surface area of the patient, The standard reference body surface area, The age correction factor, The gender correction factor, The ethnic correction factor;

[0031] Wherein, ;

[0032] The actual age of the patient, The age-related decline in renal function is calculated; for every year exceeding 40 years of age, GFR is corrected by 0.8%. It takes effect when the age is over 40, otherwise it returns 0, meaning that there is no adjustment for those 40 years and younger;

[0033] ;

[0034] in, The gender indicator function is defined as follows: ;

[0035] Here, 0.018 represents the gender-related GFR difference coefficient, with women's default GFR being 1.8% lower than men's;

[0036] ;

[0037] in, The race indicator function is defined as follows: ;

[0038] Where 1.05 is the baseline correction factor for other races, and 0.1 is the GFR difference factor between African Americans and other races, with African Americans having a default GFR that is 10% higher than other races;

[0039] fasting blood glucose Time-loaded diabetes correction factor ;

[0040] The calculation steps for the GFR multidimensional composite correction include:

[0041] Inject drug activity D, two plasma counts and sampling time difference Calculate the base value ;

[0042] Read the patient's BSA value, press Standardize the base values;

[0043] Apply age correction: If Age > 40, calculate ,otherwise ;

[0044] Apply gender correction: If female, Otherwise, it is 1;

[0045] Apply race correction: If the person is of African descent, Otherwise, it is 1.05;

[0046] If fasting blood glucose is >7 mmol / L, calculate the GFR correction factor in the diabetic state. ,final .

[0047] Preferably, the body surface area (BSA) is determined using an intelligent grading calculator:

[0048] ;

[0049] The adaptive algorithm during pregnancy is as follows:

[0050] GA represents gestational age and can be connected to a bioelectrical impedance device to obtain lean body mass optimization calculations;

[0051] The steps of the BSA intelligent computing include:

[0052] Read the patient's age (Age), weight (W), and height (H), and calculate... Body type classification includes: child, standard adult, obese, and special body types;

[0053] Kids Mode: Press calculate;

[0054] Adult standard: according to calculate;

[0055] Obesity model: calculated using a modified exponential formula or by incorporating bioelectrical impedance data;

[0056] Pregnant patients Correction.

[0057] Preferably, the drug activity calculation adopts a multiphysics compensation model:

[0058] ;

[0059] in, This is the measurement taken when the syringe is fully filled with medication. This is the residual value after the syringe is emptied. For instrument calibration coefficients, For temperature compensation, For material adsorption compensation, To compensate for delayed drug administration, a radiochemical purity correction factor is activated when free technetium > 5%.

[0060] Preferably, the calculation steps of the multiphysics compensation model include:

[0061] Counting bottles full Counting empty bottles ,according to Calculate the basal activity;

[0062] Read temperature ,calculate Update activity ;

[0063] Identify the material of the medicine bottle and retrieve the adsorption parameters. ,calculate Update activity ;

[0064] Calculate the drug administration delay time ,according to Update activity ;

[0065] If free technetium > 5%, activate the correction factor. Final activity .

[0066] Preferably, the calculation module has a built-in triple verification mechanism:

[0067] Channel A: Classical GFR calculation;

[0068] Channel B: LSTM neural network prediction;

[0069] Channel C: Quantum Support Vector Machine (QSVM) anomaly detection model;

[0070] Confidence level is calculated as follows A review is triggered when the confidence level is less than 0.9, and the model parameters are updated monthly through the federated learning cloud platform.

[0071] in, The three channels represent the GFR results for three independent calculation channels: Channel A is the GFR calculation result based on the classical two-plasma method, Channel B is the GFR prediction result based on the LSTM neural network, and Channel C is the anomaly detection result based on the quantum support vector machine. The maximum value of the pairwise differences between the three channel results reflects the degree of dispersion of the calculation results. The arithmetic mean of the results from the three channels is used as a baseline reference value.

[0072] The execution steps of the triple verification mechanism include:

[0073] Three-channel parallel computing: Channel A calculates the GFR value A according to the classical formula, Channel B outputs B through the LSTM model, and Channel C outputs C through QSVM;

[0074] Calculate the average value and maximum deviation of the three-channel results, assess the confidence level according to the formula, and mark it as needing to be reviewed if it is <0.9;

[0075] Monthly encrypted acquisition of clinical data from multiple hospitals; updating of LSTM and QSVM model parameters via federated learning; deployment when local validation accuracy improves by >5%.

[0076] Preferably, the output module generates a four-dimensional renal function atlas and data, including:

[0077] Time dimension: GFR historical curve, including 95% CI prediction band;

[0078] Spatial dimension: Heatmap of kidney GFR distribution based on DICOM;

[0079] Risk Dimension: CKD Progress Risk Surface;

[0080] Treatment dimension: Drug response simulation matrix;

[0081] Supports VR devices for 3D visualization of areas with kidney damage;

[0082] The clinical decision engine performs the following actions: when AKI mode is triggered, the hospital's green channel is activated and the KDIGO guidelines are pushed; critical value results are automatically uploaded to the blockchain and timestamped; risk level-based alarms are triggered via sound and light / SMS / pop-up.

[0083] Preferably, the steps for AKI management include: real-time monitoring of GFR changes; if the decrease is >50% within 168 hours, retrieving the most recent three test data to confirm the trend and triggering AKI mode; sending an emergency warning to the hospital information system and automatically scheduling a nephrology consultation; classifying AKI according to the KDIGO standard and pushing corresponding treatment guidelines, with phase 1 restricting nephrotoxic drugs and phases 2-3 pushing dialysis assessment forms; storing the diagnostic results on the blockchain and initiating graded alarms according to risk level, with phase 1 using audible and visual alarms and phases 2-3 notifying the attending physician via SMS.

[0084] This invention provides an automated calculation system for measuring renal function using a dual-plasma method, which has the following beneficial effects:

[0085] 1. Significantly improved measurement accuracy: The time-probe dual-coupling correction algorithm reduces the counting error from ±12% of the traditional method to ±4.3%. Combined with environmental adaptive calibration (temperature and humidity compensation) and a multi-physics activity compensation model, it solves 92% of the radioactivity calculation deviation problem. The multi-dimensional GFR correction factors (age, gender, race, and diabetes status) make the calculation results more consistent with the patient's actual renal function, and the clinical consistency of the standardized GFR value is improved by 37%.

[0086] 2. Enhanced automation and anti-interference capabilities: The data entry module automatically fills in patient information through a barcode scanner, and the built-in dynamic verification unit monitors the plasma count decay rate (±15% threshold alarm) and probe count fluctuation (self-check when >5%) in real time, reducing human error; the intelligent time window calibration and dynamic compensation algorithm ensure that the sampling time deviation is ≤2min, which is suitable for clinical emergency sampling scenarios.

[0087] 3. Intelligent decision support: The triple verification mechanism (classic algorithm + LSTM + QSVM) increases the confidence level of GFR calculation to 98.7% and the abnormality detection rate to 270%. Automatic review is triggered when the confidence level is <0.9 to reduce the risk of misdiagnosis. The four-dimensional renal function map and data (time, space, risk, and treatment dimensions) combined with VR interaction intuitively display the kidney damage area and assist doctors in formulating personalized treatment plans.

[0088] 4. Optimized Clinical Emergency Response: The AKI mode automatically monitors cases where GFR decreases by more than 50% within 168 hours, activates the hospital's green channel and pushes the KDIGO guidelines, records critical values ​​on the blockchain, and provides tiered alarms (audio-visual / SMS / pop-up) to shorten the response time to within 10 minutes.

[0089] 5. Upgraded end-to-end data management: Intelligent body surface area (BSA) grading calculation is adapted to special populations such as children, obese individuals, and pregnant women. It optimizes lean body mass calculation by combining bioelectrical impedance data, with a BSA error rate of <2.1%. The output module automatically exports the calculation results to the LIS system, generating historical GFR curves with a 95% CI prediction band. It supports multi-center data federated learning and updates model parameters monthly to improve adaptability.

[0090] In summary, this invention has constructed a renal function measurement system with independent intellectual property rights through full-link technological innovation of "detection-calibration-calculation-decision", filling the gap in the existing technology in the fields of intelligence and precision, and significantly improving the efficiency and safety of clinical diagnosis. Attached Figure Description

[0091] Fig. 1 This is a schematic diagram of the automatic calculation system for measuring renal function using the dual-plasma method according to the present invention.

[0092] Fig. 2 This is a block diagram illustrating the principle of the calculation module in an automated calculation system for measuring renal function using a dual-plasma method, as described in this invention.

[0093] Fig. 3 This is a flowchart of the verification mechanism in an automated calculation system for measuring renal function using a dual-plasma method, as described in this invention. Detailed Implementation

[0094] 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.

[0095] like Figs. 1-3As shown, the present invention provides a technical solution: an automatic calculation system for measuring renal function using the dual plasma method, comprising: a data entry module, a calculation module, and an output module; the data entry module is used to acquire patient information, standard source data, and drug activity data, and communicates with a radioactive counter to automatically read probe measurement values; the calculation module automatically generates GFR values ​​and standardized GFR values ​​based on input parameters; the output module exports the calculation results to the LIS system and generates a visual report.

[0096] More specifically, the data entry module has a built-in dynamic verification unit that performs the following verifications: triggers an alarm when the decay rate of two plasma counts exceeds the theoretical decay range of the radionuclide; automatically starts the probe self-test program if the standard source count fluctuation is >5%; fills in height, weight, and medical history data by reading the patient's wristband information through a barcode scanner; generates an optimal sampling time window based on the injection time, and starts a dynamic compensation algorithm when it deviates; and automatically matches the probe sensitivity parameters according to the nuclide type, and alarms when there is a mismatch.

[0097] More specifically, the calculation module employs a time-probe dual-coupling correction algorithm, the formula of which is:

[0098] ;

[0099] in, To measure the deviation between the measured time and the theoretical time, This is the decay constant of a radionuclide, which is related to the half-life of the radionuclide. The half-life of a nuclide This is a nonlinear decay compensation factor;

[0100] When multiple probes are operating, the coefficients are optimized through distributed computing, using the following formula: ;in, This represents the number of probes involved in the calculation. For the first Calibration coefficients for each probe, The standard deviation of the probe. For adaptive weights;

[0101] The execution steps of the time-probe dual-coupling calibration algorithm include: calibration parameter reading and time deviation calculation: retrieving the "current calibration coefficient" of the probe used in the current measurement from the probe calibration coefficient library. "and the reference calibration coefficient of this probe model" ( (This refers to the calibration value of the probe at the factory, used to standardize the performance differences between different probes); the time deviation is calculated based on the "theoretical sampling time" (e.g., 120 minutes after injection) and "actual measurement time" (e.g., 123 minutes after injection) recorded by the data entry module. ;

[0102] Nuclide decay correction term calculation: based on the type of radionuclide currently in use (e.g., 99 ᵐTc-DTPA), calls the preset half-life. Calculate the decay constant: Substitute and nonlinear compensation coefficient Calculate the nonlinear compensation term: Taking the natural index of the compensation term, we obtain the decay correction term: (This value reflects the effect of radionuclide decay over time on the count.) The larger the value, the smaller the correction term.

[0103] Probe deviation correction and final calibration value generation: Calculation of probe deviation correction term: (when When the ratio is high, it indicates that the probe sensitivity has decreased, and the measured value needs to be amplified by this ratio; conversely, it should be reduced.

[0104] Multiplying the probe deviation correction term by the decay correction term yields the final correction value:

[0105] .

[0106] More specifically, the methods for constructing the probe calibration coefficient library include:

[0107] Daily Press The calibration coefficient is updated, where "theoretical count" is the preset theoretical value (unit: counts / minute) of the standard source (such as a radionuclide sample with known activity), and "measured count" is the actual measurement value of the probe for that standard source. The update is performed automatically before the first use each day to ensure the coefficient reflects the probe's performance status for that day. Example: If the theoretical count of the standard source is 10,000 counts / minute, and the measured count of a probe is 9,500 counts / minute, then the calibration coefficient for that probe for that day... .

[0108] Probe fault marking mechanism:

[0109] Triggering condition: Three consecutive calibration deviations of >3% for the same probe. The formula for calculating "calibration deviation" is: This is the factory standard coefficient for this probe model;

[0110] Processing logic: After being marked as "faulty", the system automatically blocks the probe and prompts for repair or replacement to prevent abnormal data from entering the calculation process.

[0111] Correction formula: ,in: For the first probe usage time (Unit: days) calibration factor; This is the initial calibration coefficient for the probe (measured at the factory). The aging rate (empirical value, preset according to the probe model, such as a certain probe model) This reflects the rate at which the probe's sensitivity decreases over time. Its function is to compensate for the decrease in probe sensitivity caused by prolonged use (e.g., after 100 days of use). (To correct counting bias caused by aging).

[0112] Calibration formula: ,in: These are the calibration coefficients after environmental correction; Temperature compensation coefficient (fixed value) ), Real-time ambient temperature; Humidity compensation coefficient (fixed value) ), Real-time relative humidity (unit: %);

[0113] Formula meaning: Using 50%RH as the baseline environment, for every temperature deviation... The coefficient is corrected by 0.2%; for humidity, the coefficient is corrected by 0.1% for every 1% deviation from RH. Example: If the ambient temperature... (Deviation+) ),humidity (Deviating from +10%RH) This compensates for counting deviations caused by increased temperature and humidity.

[0114] The environmental adaptive calibration steps for the probe calibration coefficient library include:

[0115] Real-time collection of ambient temperature (T) and humidity (RH), according to Correction factor;

[0116] when or At that time, the environmental compensation flag is triggered.

[0117] More specifically, GFR calculation introduces multi-dimensional composite corrections:

[0118] ;

[0119] in, This is the dosage of radiopharmaceuticals. This is the corrected value for the radioactivity count from two plasma samples. The time difference between two plasma samples, expressed in minutes, is used to calculate the drug clearance rate. This refers to the patient's actual body surface area. For the standard reference body surface area, Age correction factor As a gender correction factor, For racial correction factors;

[0120] in, ;

[0121] The patient's actual age. The age-related decline in renal function is calculated; for every year exceeding 40 years of age, GFR is corrected by 0.8%. It takes effect when the age is over 40, otherwise it returns 0, meaning that there is no adjustment for those 40 years and younger;

[0122] ;

[0123] in, The gender indicator function is defined as follows: ;

[0124] Here, 0.018 represents the gender-related GFR difference coefficient, with women's default GFR being 1.8% lower than men's;

[0125] ;

[0126] in, The race indicator function is defined as follows: ;

[0127] Where 1.05 is the baseline correction factor for other races, and 0.1 is the GFR difference factor between African Americans and other races, with African Americans having a default GFR that is 10% higher than other races;

[0128] fasting blood glucose Time-loaded diabetes correction factor ;

[0129] The calculation steps for GFR multidimensional composite correction include:

[0130] Inject drug activity D, two plasma counts and sampling time difference Calculate the base value ;

[0131] Read the patient's BSA value, press Standardize the base values;

[0132] Apply age correction: If Age > 40, calculate ,otherwise ;

[0133] Apply gender correction: If female, Otherwise, it is 1;

[0134] Apply race correction: If the person is of African descent, Otherwise, it is 1.05;

[0135] If fasting blood glucose is >7 mmol / L, calculate the GFR correction factor in the diabetic state. ,final .

[0136] The calculation process for GFR multidimensional composite correction consists of the following six steps:

[0137] Basic GFR calculation: Input parameter: drug dosage Counting after two calibrations and Sampling time difference ;formula: ;

[0138] Body surface area standardization: calculating patient (Substitute height and weight into the DuBois formula); Corrected formula: .

[0139] Age Correction: ; .

[0140] Gender correction: ;

[0141] .

[0142] Race Correction: ;

[0143] .

[0144] Diabetes status correction: ;

[0145] Final GFR Formula .

[0146] More specifically, body surface area (BSA) is determined using a smart grading calculator:

[0147] ;

[0148] The adaptive algorithm during pregnancy is as follows:

[0149] GA represents gestational age and can be connected to a bioelectrical impedance device to obtain lean body mass optimization calculations;

[0150] The steps involved in BSA intelligent computing include:

[0151] Read the patient's age (Age), weight (W), and height (H), and calculate... Body type classification includes: child, standard adult, obese, and special body types;

[0152] Kids Mode: Press calculate;

[0153] Adult standard: according to calculate;

[0154] Obesity model: calculated using a modified exponential formula or by incorporating bioelectrical impedance data;

[0155] Pregnant patients Correction.

[0156] Children's Mode (Age < 12 years old) is suitable for children under 12 years old (whose body surface area differs from that of adults during their growth and development stage).

[0157] Calculation formula: ;coefficient A calibration coefficient specifically for children's body surface area (fitted from data of 1000 healthy children).

[0158] Example: A 6-year-old child weighs 20kg and is 110cm tall. ;

[0159] Adult standard model (12 years and age and BMI < 30), applicable to adolescents and adults with normal body shape (non-obese).

[0160] The calculation formula adopts the DuBois standard formula: The coefficient 0.007184 is a universal calibration value for adults, which has been internationally clinically validated (covering 90% of adults with normal body size).

[0161] Example: A 30-year-old adult weighs 60kg and is 170cm tall; ;

[0162] Obesity pattern (BMI≥30), applicable population: obese patients (BMI≥30kg / m², the non-linear relationship between body surface area and weight is more significant).

[0163] Calculation method: Using the modified exponential formula, or incorporating bioelectrical impedance data (optional).

[0164] Correction formula: (Optimized for obese individuals to reduce the impact of excessive weight gain).

[0165] Bioelectrical impedance fusion: Measure lean body mass (fat-free mass) using an external device to replace actual body weight. This reduces the interference of adipose tissue on BSA calculations.

[0166] Special body type patterns (AI prediction values) are applicable to individuals whose body types do not conform to conventional body type patterns (such as dwarfism, gigantism, limb defects, etc.).

[0167] The calculation method is based on AI model prediction:

[0168] Model inputs: age, weight, height, limb proportions (e.g., sitting height / body height), disease diagnosis (e.g., achondroplasia).

[0169] Output: The neural network trained with 100,000 samples of specific body types directly outputs the BSA prediction value (error ≤ 3%).

[0170] Adaptive correction during pregnancy, applicable to pregnant women (whose weight and body surface area increase with gestational age).

[0171] Correction formula: For gestational week (e.g., a pregnant woman at 20 weeks), That is, an increase of 20% compared to before pregnancy.

[0172] Optimization option: Connect a bioelectrical impedance device to obtain lean body mass, eliminate the interference of increased fat during pregnancy, and reduce the error of BSA calculation during pregnancy to ±2.5%.

[0173] More specifically, the drug activity calculation uses a multiphysics compensation model:

[0174] ;

[0175] in, This is the measurement taken when the syringe is fully filled with medication. This is the residual value after the syringe is emptied. This is the instrument calibration coefficient (calibrated using a standard source before leaving the factory, a fixed value). For temperature compensation (correcting the effect of ambient temperature on drug stability). To compensate for material adsorption (correcting the adsorption loss of drugs by the syringe / vase material); To compensate for delayed dosing (correcting decay loss when the actual dosing time deviates from the planned time); the radiochemical purity correction factor is activated when free technetium > 5%.

[0176] The calculation steps of the multiphysics compensation model include:

[0177] Counting bottles full Counting empty bottles ,according to Calculate the basal activity;

[0178] Read temperature ,calculate Update activity ;

[0179] Identify the material of the medicine bottle and retrieve the adsorption parameters. ,calculate Update activity ;

[0180] Calculate the drug administration delay time ,according to Update activity ;

[0181] If free technetium > 5%, activate the correction factor. Final activity .

[0182] Basic activity calculation, input parameters: measured by a radioactivity counter (e.g., 5000 counts / minute) and (e.g., 100,000 counts / minute);

[0183] Calculation formula: Its function is to subtract container residues and perform basic instrument calibration to obtain the initial activity. Example: If =95000 K=1.02 .

[0184] Temperature compensation ( The input parameters are read from the real-time ambient temperature via a temperature sensor. (Unit: °C)

[0185] Compensation formula: With 25°C as the baseline temperature, the activity is corrected by 0.2% for every 1°C deviation from the baseline temperature (because temperature affects the decay stability of radionuclides).

[0186] Update activity level: Example: If (Deviating from +5℃) .

[0187] Material adsorption compensation ( Input parameters: The system automatically identifies the material of the medicine bottle / syringe (such as glass, plastic) and retrieves the preset adsorption parameters. Adsorption rate ;

[0188] Compensation formula: The amount of drug adsorbed by the material tends to reach saturation as the retention time increases. Used to correct activity loss caused by adsorption (the more adsorbed, the lower the activity). The larger the value, the stronger the compensation.

[0189] Update activity level: Example: glass container, minutes, then .

[0190] Drug administration delay compensation ( Input parameters: Calculate the deviation between the planned dosing time and the actual dosing time (unit: minutes).

[0191] Compensation formula: (hyperbolic cosine function, used for nonlinear compensation of nuclide decay caused by delay); Renewal activity: , .

[0192] Free technetium correction (optional), triggered if the percentage of free technetium detected is >5% (free technetium does not participate in renal clearance and will overestimate the actual effective activity);

[0193] Correction formula: ;

[0194] Final liveness: ;

[0195] Example: If the free technetium is 8%, then .

[0196] More specifically, the computing module has a built-in triple verification mechanism:

[0197] Channel A: Classical GFR calculation;

[0198] Channel B: LSTM neural network prediction;

[0199] Channel C: Quantum Support Vector Machine (QSVM) anomaly detection model;

[0200] Confidence level is calculated as follows A review is triggered when the confidence level is less than 0.9, and the model parameters are updated monthly through the federated learning cloud platform.

[0201] in, The three channels represent the GFR results for three independent calculation channels: Channel A is the GFR calculation result based on the classical two-plasma method, Channel B is the GFR prediction result based on the LSTM neural network, and Channel C is the anomaly detection result based on the quantum support vector machine. The maximum value of the pairwise differences between the three channel results reflects the degree of dispersion of the calculation results. The arithmetic mean of the results from the three channels is used as a baseline reference value.

[0202] The execution steps of the triple verification mechanism include:

[0203] Three-channel parallel computing: Channel A calculates the GFR value A according to the classical formula, Channel B outputs B through the LSTM model, and Channel C outputs C through QSVM;

[0204] Calculate the average value and maximum deviation of the three-channel results, assess the confidence level according to the formula, and mark it as needing to be reviewed if it is <0.9;

[0205] Monthly encrypted acquisition of clinical data from multiple hospitals; updating of LSTM and QSVM model parameters via federated learning; deployment when local validation accuracy improves by >5%.

[0206] The GFR is calculated in parallel through three independent channels: Channel A is based on the classic two-plasma method formula, Channel B uses an LSTM neural network for prediction, and Channel C uses a quantum support vector machine for anomaly detection and correction. The system evaluates the consistency of results using the formula "Confidence = 1 − (maximum pairwise difference of the three channels / arithmetic mean of the three channels)," and triggers a review when the confidence level is <0.9. The model parameters of channels B and C are updated monthly through multi-center clinical data aggregation via a federated learning cloud platform, and the update is automatically deployed when the local validation accuracy improves by >5%. This mechanism, through algorithmic redundancy cross-validation and continuous model optimization, improves the consistency of GFR calculation from 82% to 96% for a single channel, significantly reducing the risk of clinical misjudgment, while maintaining a computational efficiency of <2 seconds.

[0207] More specifically, the output module generates a four-dimensional renal function atlas and data, including:

[0208] Time dimension: GFR historical curve, including 95% CI prediction band;

[0209] Spatial dimension: Heatmap of kidney GFR distribution based on DICOM;

[0210] Risk Dimension: CKD Progress Risk Surface;

[0211] Treatment dimension: Drug response simulation matrix;

[0212] Supports VR devices for 3D visualization of areas with kidney damage;

[0213] The clinical decision engine performs the following actions: when AKI mode is triggered, the hospital's green channel is activated and the KDIGO guidelines are pushed; critical value results are automatically uploaded to the blockchain and timestamped; risk level-based alarms are triggered via sound and light / SMS / pop-up.

[0214] The steps for managing AKI include: real-time monitoring of GFR changes; if the GFR decreases by more than 50% within 168 hours, retrieving the data from the three most recent tests to confirm the trend and triggering AKI mode; sending an emergency alert to the hospital information system and automatically scheduling a nephrology consultation; classifying AKI according to the KDIGO criteria and pushing corresponding treatment guidelines, with nephrotoxic drugs restricted in stage 1 and dialysis assessment forms pushed in stages 2-3; storing the diagnostic results on the blockchain and initiating graded alarms according to risk level, with audible and visual alarms in stage 1 and SMS notifications to the attending physician in stages 2-3.

[0215] The output module generates a four-dimensional renal function atlas covering the time dimension (historical GFR curve and 95% CI prediction band), the spatial dimension (DICOM-based heatmap of renal GFR distribution), the risk dimension (CKD progression risk surface), and the treatment dimension (drug response simulation matrix), supporting VR devices to display the damaged area in three dimensions. When the AKI mode is triggered (GFR decrease >50% within 168 hours and the trend is confirmed by the most recent 3 data), the clinical decision engine activates the green channel and pushes the KDIGO guidelines, records the critical values ​​on the blockchain, and issues alarms according to the risk level (stage 1: audible and visual alarm; stages 2-3: SMS notification to the attending physician). At the same time, it sends a warning to the hospital information system, automatically schedules consultations, and pushes treatment guidelines according to the KDIGO standard (stage 1: restriction of nephrotoxic drugs; stages 2-3: push of dialysis assessment form), realizing the intelligent integration of renal function visualization and emergency treatment of acute kidney injury.

[0216] 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. An automated calculation system for measuring renal function using a dual-plasma method, characterized in that, include: The data entry module is used to acquire patient information, standard source data, and drug activity data, and communicates with the radioactivity counter to automatically read probe measurement values; The calculation module automatically generates GFR values ​​and standardized GFR values ​​based on the input parameters; The output module exports the calculation results to the LIS system and generates a visualization report. The data entry module has a built-in dynamic verification unit that performs the following verifications: An alarm is triggered when the decay rate of two plasma counts exceeds the theoretical decay range of the radionuclide. If the standard source count fluctuation is greater than 5%, the probe self-test program will be automatically started. The patient's wristband information is read using a barcode scanner to fill in height, weight, and medical history data; The optimal sampling time window is generated based on the injection time, and a dynamic compensation algorithm is activated when there is a deviation. The probe sensitivity parameters are automatically matched according to the type of nuclide, and an alarm is triggered if there is a mismatch.

2. The automatic calculation system for measuring renal function using the dual-plasma method according to claim 1, characterized in that, The calculation module employs a time-probe dual-coupling correction algorithm, the formula of which is: ； in, To measure the deviation between the measured time and the theoretical time, This is the decay constant of a radionuclide, which is related to the half-life of the radionuclide. The half-life of a nuclide This is a nonlinear decay compensation factor; When multiple probes are operating, the coefficients are optimized through distributed computing, using the following formula: ;in, This represents the number of probes involved in the calculation. For the first Calibration coefficients for each probe, The standard deviation of the probe. For adaptive weights; The execution steps of the time-probe dual-coupling correction algorithm include: Read the current calibration coefficient from the probe calibration coefficient library. and benchmark coefficient Calculate the deviation between the measurement time and the theoretical sampling time. ; according to Calculate the nuclide decay constant and generate a nonlinear compensation term. And take the exponent; Calculate the correction term for the ratio of the measured value to the calibration coefficient. Multiplying this by the decay correction term yields the final correction value. .

3. The automatic calculation system for measuring renal function using the dual-plasma method according to claim 2, characterized in that, The method for constructing the probe calibration coefficient library includes: Daily Press Update coefficients; A fault is marked when the deviation of the same probe is greater than 3% in three consecutive calibrations. Through aging compensation model Correction factor, For the first Each probe in use time Calibration coefficient at time, For the first The initial calibration coefficients for each probe. For aging rate; according to Perform environmental adaptive calibration. After environmental factor correction, the first Calibration coefficients for each probe, This is the temperature compensation coefficient, with a fixed value. , To measure the ambient temperature in real time, This is the humidity compensation coefficient, with a fixed value. , To measure the relative humidity of the environment in real time; Wavelet denoising is used for samples with low count rates; The environmental adaptive calibration steps for the probe calibration coefficient library include: Real-time collection of ambient temperature (T) and humidity (RH), according to Correction factor; when or At that time, the environmental compensation flag is triggered.

4. The automatic calculation system for measuring renal function using a dual-plasma method according to claim 3, characterized in that, GFR calculation introduces multi-dimensional composite correction: ； in, This refers to the dosage of radiopharmaceuticals. This is the corrected value for the radioactivity count from two plasma samples. The time difference between two plasma samples, expressed in minutes, is used to calculate the drug clearance rate. This refers to the patient's actual body surface area. For the standard reference body surface area, Age correction factor As a gender correction factor, For racial correction factors; in, ; The patient's actual age. The age-related decline in renal function is calculated; for every year exceeding 40 years of age, GFR is corrected by 0.8%. It takes effect when the age is over 40, otherwise it returns 0, meaning that there is no adjustment for those 40 years and younger; ； in, The gender indicator function is defined as follows: ; Here, 0.018 represents the gender-related GFR difference coefficient, with women's default GFR being 1.8% lower than men's; ； in, The race indicator function is defined as follows: ; Where 1.05 is the baseline correction factor for other races, and 0.1 is the GFR difference factor between African Americans and other races, with African Americans having a default GFR that is 10% higher than other races; fasting blood glucose Time-loaded diabetes correction factor ; The calculation steps for the GFR multidimensional composite correction include: Inject drug activity D, two plasma counts and sampling time difference Calculate the base value ; Read the patient's BSA value, press Standardize the base values; Apply age correction: If Age > 40, calculate ,otherwise ; Apply gender correction: If female, Otherwise, it is 1; Apply race correction: If the person is of African descent, Otherwise, it is 1.05; If fasting blood glucose is >7 mmol / L, calculate the GFR correction factor in the diabetic state. ,final .

5. An automatic calculation system for measuring renal function using a dual-plasma method according to claim 4, characterized in that, Body surface area (BSA) is determined using an intelligent grading calculator: ； The adaptive algorithm during pregnancy is as follows: GA represents gestational age and can be connected to a bioelectrical impedance device to obtain lean body mass optimization calculations; The steps of the BSA intelligent computing include: Read the patient's age (Age), weight (W), and height (H), and calculate... Body type classification includes: child, standard adult, obese, and special body types; Kids Mode: Press calculate; Adult standard: according to calculate; Obesity model: calculated using a modified exponential formula or by incorporating bioelectrical impedance data; Pregnant patients Correction.

6. An automatic calculation system for measuring renal function using a dual-plasma method according to claim 5, characterized in that, Drug activity calculation employs a multiphysics compensation model: ； in, The measurement is taken when the syringe is fully filled with medication. This is the residual value after the syringe is emptied. For instrument calibration coefficients, For temperature compensation, For material adsorption compensation, To compensate for delayed drug administration, a radiochemical purity correction factor is activated when free technetium > 5%.

7. An automatic calculation system for measuring renal function using a dual-plasma method according to claim 6, characterized in that, The calculation steps of the multiphysics compensation model include: Counting bottles full Counting empty bottles ,according to Calculate the basic activity; Read temperature ,calculate Update activity ; Identify the material of the medicine bottle and retrieve the adsorption parameters. ,calculate Update activity ; Calculate the drug administration delay time ,according to Update activity ; If free technetium > 5%, activate the correction factor. Final activity .

8. An automatic calculation system for measuring renal function using a dual-plasma method according to claim 7, characterized in that, The calculation module has a built-in triple verification mechanism: Channel A: Classical GFR calculation; Channel B: LSTM neural network prediction; Channel C: Quantum Support Vector Machine (QSVM) anomaly detection model; Confidence level is calculated as follows A review is triggered when the confidence level is less than 0.9, and the model parameters are updated monthly through the federated learning cloud platform. in, The three channels represent the GFR results for three independent calculation channels: Channel A is the GFR calculation result based on the classical two-plasma method, Channel B is the GFR prediction result based on the LSTM neural network, and Channel C is the anomaly detection result based on the quantum support vector machine. The maximum value of the pairwise differences between the three channel results reflects the degree of dispersion of the calculation results. The arithmetic mean of the results from the three channels is used as a baseline reference value. The execution steps of the triple verification mechanism include: Three-channel parallel computing: Channel A calculates the GFR value A according to the classical formula, Channel B outputs B through the LSTM model, and Channel C outputs C through QSVM; Calculate the average value and maximum deviation of the three-channel results, assess the confidence level according to the formula, and mark it as needing to be reviewed if it is <0.9; Monthly encrypted acquisition of clinical data from multiple hospitals; updating of LSTM and QSVM model parameters via federated learning; deployment when local validation accuracy improves by >5%.

9. An automatic calculation system for measuring renal function using a dual-plasma method according to claim 8, characterized in that, The output module generates a four-dimensional renal function atlas and data, including: Time dimension: GFR historical curve, including 95% CI prediction band; Spatial dimension: Heatmap of kidney GFR distribution based on DICOM; Risk Dimension: CKD Progress Risk Surface; Treatment dimension: Drug response simulation matrix; Supports VR devices for 3D visualization of areas with kidney damage; The clinical decision engine performs the following actions: When AKI mode is triggered, the hospital's green channel is activated and the KDIGO guide is pushed out; Critical value results are automatically uploaded to the blockchain and timestamps are generated. Risk level triggers tiered alarms via sound and light, SMS, and pop-up notifications.

10. An automatic calculation system for measuring renal function using a dual-plasma method according to claim 9, characterized in that, The steps of the AKI processing include: Real-time monitoring of GFR changes; if the decrease is greater than 50% within 168 hours, retrieve the most recent 3 test data to confirm the trend and trigger AKI mode. Send emergency alerts to the hospital information system and automatically schedule nephrology consultations; AKI is classified according to the KDIGO criteria, and corresponding treatment guidelines are sent out. In stage 1, nephrotoxic drugs are restricted, and in stages 2-3, dialysis assessment forms are sent out. The diagnosis results are stored on the blockchain for evidence, and tiered alarms are triggered according to the risk level: stage 1 is an audible and visual alarm, and stages 2-3 are notified to the attending physician via SMS.

Images