Method and system for quantitatively evaluating effects of thyroid function abnormalities on cardiovascular system

Abstract

The application discloses a method and system for quantitatively evaluating the influence of abnormal thyroid function on the cardiovascular system, and relates to the technical field of medical health data analysis. The method comprises the following steps: obtaining a thyroid test sequence and cardiovascular observation data of a target object; extracting the thyroid-stimulating hormone value and the free thyroxine value of each test point in the thyroid test sequence; and converting the values into a first deviation and a second deviation relative to a preset reference value. The application initializes the thyroid effect quantity with the FT4 deviation at the effective starting point, recursively updates the quantity in the effective interval, and restricts the update range by the TSH change amplitude in the lag window and the response type, thereby solving the defects of the static correlation or fixed lag method, such as the difficulty in describing the nonlinear hysteresis of the thyroid axis and the delayed response of the target organ, the jump of the effect estimation, and the susceptibility to noise interference.

Description

Technical Field

[0001] This invention relates to the field of medical and health data analysis technology, and more specifically, this application relates to a method and system for quantitatively assessing the impact of thyroid dysfunction on the cardiovascular system. Background Technology

[0002] Thyroid dysfunction has a significant impact on the cardiovascular system, and its quantitative assessment is often necessary in clinical follow-up and continuous monitoring scenarios. Existing data sources typically include: thyroid function test data, represented by thyroid-stimulating hormone (TSH) and free thyroxine (FT4) (with free triiodothyronine (FT3) as an alternative), and cardiovascular vital signs data, represented by heart rate, systolic blood pressure, diastolic blood pressure, and pulse pressure. Thyroid test data are often low-frequency, discrete, and unequally spaced time-stamped sequences, while cardiovascular vital signs data are often high-frequency continuous or quasi-continuously sampled, accompanied by significant random fluctuations and short-term perturbations. The two types of data exhibit significant inconsistencies in temporal resolution and sampling methods.

[0003] Meanwhile, the thyroid axis has feedback regulation characteristics, and there are often asynchronous changes and lag effects between TSH and FT4 / FT3. The effects of thyroid hormones on the cardiovascular system may also manifest as latency and cumulative effects, making it difficult to directly align the timing of laboratory changes with the timing of physical signs in the time dimension. Existing analytical methods often use static cross-sectional correlation, simple linear regression, or fixed lag matching, which often make it difficult to stably define the effective response interval related to changes in thyroid status against the background of high-frequency physical sign noise. Furthermore, when there are large differences in baseline and sensitivity among different individuals, misjudgment or omission is likely to occur, affecting the consistency and interpretability of the assessment.

[0004] Therefore, existing technologies need to address the following technical issues: Without relying on new specialized equipment, how to utilize low-frequency discrete thyroid test sequences and high-frequency continuous cardiovascular observation data to determine the effective starting point of thyroid state changes and the starting point of cardiovascular response, considering thyroid axis feedback hysteresis and target organ response delay, thereby determining the effective response interval and achieving reproducible quantitative characterization; simultaneously, how to introduce an adaptive correction mechanism oriented towards individual differences to reduce false positives and missed judgments caused by baseline differences and noise disturbances, so that the assessment results can more reliably reflect the dynamic impact of thyroid dysfunction on the cardiovascular system. Therefore, a quantitative assessment method and system for the impact of thyroid dysfunction on the cardiovascular system are proposed to solve this problem. Summary of the Invention

[0005] To address the aforementioned technical problems, this technical solution provides a quantitative assessment method and system for the impact of thyroid dysfunction on the cardiovascular system, resolving the issues raised in the background section.

[0006] To achieve the above objectives, the technical solution of the present invention is as follows:

[0007] Firstly, this application provides a method for quantitatively assessing the impact of thyroid dysfunction on the cardiovascular system, the method comprising:

[0008] Obtain the thyroid test sequence and cardiovascular observation data of the target subject, extract the thyroid-stimulating hormone value and free thyroxine value of each test point in the thyroid test sequence, and convert them into the first deviation and the second deviation relative to the preset reference value, respectively.

[0009] The effective starting point is determined based on the direction of change of the first deviation between adjacent test points and the overall direction of change of the second deviation within the hysteresis window starting from the test point, and the front window and back window are set based on the effective starting point.

[0010] Based on cardiovascular observation data, the baseline of each sub-parameter in the anterior window is determined and the relative difference in the posterior window is calculated. The difference is filtered and fused into a composite difference value according to preset combination conditions. The extreme value of the composite difference value is taken as the response starting point and the response type is determined. Each sub-parameter includes at least heart rate, systolic blood pressure and diastolic blood pressure.

[0011] The effective interval is determined based on the effective starting point and the response starting point. The thyroid effect quantity is initialized with the second deviation at the effective starting point and updated sequentially within the effective interval to generate the thyroid effect quantity sequence. The thyroid effect quantity at each time point is updated by combining the thyroid effect quantity at the previous time point with the change of the second deviation. The update magnitude is constrained by the change of the first deviation within the hysteresis window and the reverse change of the thyroid effect quantity is restricted according to the response type.

[0012] The response intensity of the difference synthesis quantity relative to the thyroid effect quantity sequence within one or more effective intervals is statistically analyzed, smoothed and summarized to obtain the individual sensitivity coefficient, and the amplitude of the thyroid effect quantity is corrected using the individual sensitivity coefficient to generate the thyroid effect surrogate quantity sequence.

[0013] Secondly, this application provides a quantitative assessment system for the impact of thyroid dysfunction on the cardiovascular system, used to implement the quantitative assessment method for the impact of thyroid dysfunction on the cardiovascular system described in any of the above claims, including:

[0014] The data extraction and conversion module is used to acquire the thyroid test sequence and cardiovascular observation data of the target object, extract the thyroid-stimulating hormone value and free thyroxine value of each test point in the thyroid test sequence, and convert them into the first deviation and the second deviation relative to the preset reference value, respectively.

[0015] The effective starting point identification module is used to determine the effective starting point based on the direction of change of the first deviation between adjacent test points and the overall direction of change of the second deviation within the hysteresis window starting from the test point, and to set the front window and the back window based on the effective starting point.

[0016] The cardiovascular response analysis module is used to determine the baseline of each sub-parameter in the front window and calculate the relative difference in the back window based on cardiovascular observation data. It filters and merges the sub-parameters into a composite difference value according to preset combination conditions, takes the extreme value of the composite difference value as the response start point and determines the response type. Each sub-parameter includes at least heart rate, systolic blood pressure and diastolic blood pressure.

[0017] The effect size generation module is used to determine the effective interval based on the effective start point and the response start point, initialize the thyroid effect size with the second deviation at the effective start point, and recursively update and generate the thyroid effect size sequence within the effective interval according to the time sequence.

[0018] The surrogate quantity synthesis module is used to statistically analyze the response intensity of the difference synthesis quantity relative to the thyroid effect quantity sequence within one or more effective intervals, smooth and summarize it to obtain the individual sensitivity coefficient, and use the individual sensitivity coefficient to correct the amplitude of the thyroid effect quantity to generate the thyroid effect surrogate quantity sequence.

[0019] Thirdly, this application provides a computer device including a memory and a processor, the memory storing code, and the processor being configured to acquire the code and execute the above-described quantitative assessment method for the impact of thyroid dysfunction on the cardiovascular system.

[0020] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for quantitatively assessing the impact of thyroid dysfunction on the cardiovascular system.

[0021] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0022] This application converts TSH and FT4 into deviations from a relative reference value. It uses a hysteresis criterion that TSH changes continuously in the same direction and FT4 shows an inverse trend within the hysteresis window to determine the effective starting point of thyroid status changes. It also establishes an individual sign baseline by combining a baseline window. The difference between the sign parameters and the baseline is calculated within the search window and fused into a composite quantity to locate the cardiovascular response starting point. The effective starting point and the response starting point jointly define the effective interval, thus solving the time alignment problem between low-frequency discrete test data and high-frequency continuous sign data.

[0023] This application initializes the thyroid effect quantity with the FT4 deviation at the effective starting point and updates it sequentially within the effective interval. This makes the update amplitude constrained by the TSH change amplitude within the hysteresis window and restricts the reverse change in combination with the response type. It also suppresses short-term rebounds and reverse fluctuations that do not meet the persistence requirement. This solves the defects of static correlation or fixed hysteresis methods in that they are difficult to characterize the nonlinear hysteresis of the thyroid axis and the delayed response of the target organ, which leads to jumps in effect estimation and susceptibility to human symptom noise interference. It achieves a stable output of a continuous thyroid effect state sequence that conforms to the physiological hysteresis characteristics.

[0024] This application obtains an individual sensitivity coefficient by statistically analyzing the response intensity across the effective interval, and then performing anomaly removal and baseline stability weighting. This coefficient is used to adaptively correct the thyroid effect magnitude to generate a thyroid effect proxy sequence. This overcomes the high false positive and false negative rates caused by differences in individual baselines and sensitivity in general thresholds or models, and achieves adaptive quantification and reliable assessment of individualized dynamic impact processes. Attached Figure Description

[0025] The disclosure of this invention is illustrated with reference to the accompanying drawings. It should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of this invention. Wherein:

[0026] Figure 1 This is a flowchart of the quantitative assessment method for the impact of thyroid dysfunction on the cardiovascular system proposed in this invention;

[0027] Figure 2 This is a structural block diagram of the quantitative assessment system for the impact of thyroid dysfunction on the cardiovascular system proposed in this invention. Detailed Implementation

[0028] It is readily understood that, based on the technical solution of this invention, those skilled in the art can propose various interchangeable structural methods and implementations without altering the essential spirit of the invention. Therefore, the following detailed embodiments and accompanying drawings are merely illustrative examples of the technical solution of this invention and should not be considered as the entirety of the invention or as limitations or restrictions on the technical solution of this invention.

[0029] Reference Figure 1 As shown, this application proposes a method for quantitatively assessing the impact of thyroid dysfunction on the cardiovascular system, characterized in that the method includes:

[0030] S1. Obtain the thyroid test sequence and cardiovascular observation data of the target object, extract the thyroid-stimulating hormone value and free thyroxine value of each test point in the thyroid test sequence, and convert them into the first deviation and the second deviation relative to the preset reference value, respectively.

[0031] In this embodiment, the thyroid test sequences of the target patient within a preset analysis period (e.g., the past 6 months) are obtained from the hospital's laboratory system: each test includes a timestamp. and thyroid-stimulating hormone Free thyroxine The measured values ​​are obtained from monitoring systems, follow-up systems, or wearable devices, representing cardiovascular observation data at each sampling time. Record heart rate systolic blood pressure diastolic blood pressure All timestamps are formatted uniformly, obvious errors or duplicate records are removed, and they are arranged in ascending order of time to form a data sequence with timestamps.

[0032] For example, the calculation methods for the first and second deviations relative to the preset reference values ​​are as follows:

[0033] Set the default reference value for TSH , This can be the center value of the TSH reference range (e.g., the reference range is 0.4-4.0). The center value is 2.2. The first deviation can be calculated using the median of a stable historical period for an individual, or a weighted combination of both. :

[0034] ;

[0035] Set the default reference value for FT4 , It can be the center value of the FT4 reference range (e.g., the reference range is 12-22). The center value is 17. The second deviation was calculated using a similar method. :

[0036] ;

[0037] This eliminates the influence of dimensions, making the deviations of different individuals comparable and forming a time-stamped sequence of deviations. ;

[0038] In a specific example, a patient underwent 8 thyroid function tests within 6 months. The extracted deviation sequence is: {(2025-01-25,-0.12,+0.08),(2025-02-15,+0.05,+0.03),(2025-03-08,+0.18,-0.04),(2025-03-22,+0.25,-0.09),(2025-04-14,+0.30,-0.12),(2025-05-18,+0.28,-0.10),(2025-06-20,+0.32,-0.14),(2025-07-15,+0.35,-0.16)};

[0039] S2. Determine the effective starting point based on the direction of change of the first deviation between adjacent test points and the overall direction of change of the second deviation within the hysteresis window starting from the test point, and set the front window and back window based on the effective starting point.

[0040] It should be noted that the front window is used to establish a baseline for cardiovascular parameters, covering intraday fluctuations and maintaining relative stability to avoid introducing state drift due to excessive length; the back window is used to calculate relative differences and detect cardiovascular responses, covering common timescales of target organ response latency and cumulative effects; the front window is... The rear window is , For the time corresponding to the valid starting point, , These are the preset lengths for the front and rear windows, with typical values ​​of 7 days and 14 days respectively.

[0041] S3. Based on cardiovascular observation data, determine the baseline of each sub-parameter in the anterior window and calculate the relative difference in the posterior window. Filter and merge the sub-parameters into a composite difference value according to preset combination conditions. Take the extreme value of the composite difference value as the response start point and determine the response type. Each sub-parameter includes at least heart rate, systolic blood pressure, and diastolic blood pressure.

[0042] It should be noted that the baseline for sub-parameters such as heart rate, systolic blood pressure, and diastolic blood pressure is calculated within the anterior window: the baseline can be the median, mean, or trimmed mean. In this embodiment, the median is selected as the baseline, and the baseline fluctuation (such as the standard deviation) is calculated for subsequent baseline stability assessment.

[0043] S4. Determine the effective interval based on the effective starting point and the response starting point, initialize the thyroid effect quantity with the second deviation at the effective starting point, and generate the thyroid effect quantity sequence by recursively updating it in time sequence within the effective interval; wherein, the thyroid effect quantity at each time point is updated by the thyroid effect quantity at the previous time point combined with the change of the second deviation, the update magnitude is constrained by the change of the first deviation within the hysteresis window, and the reverse change of the thyroid effect quantity is restricted according to the response type.

[0044] It should be noted that this is based on a valid starting point. and the starting point of the response Determine the effective range The effective interval refers to the time period from the onset of thyroid dysfunction to a significant cardiovascular response. Within this interval, a thyroid effector sequence is constructed.

[0045] The first implementation method is to use the valid starting point and the response starting point as endpoints. ;

[0046] The second implementation method is to expand forward or backward according to preset rules to cover the complete response process, such as: ,in, To extend the duration (e.g., 15-30 days, with a typical value of 20 days) and capture the sustained effects after the response initiation point, this embodiment adopts the second method;

[0047] S5. Statistically analyze the response intensity of the difference synthesis quantity relative to the thyroid effect quantity sequence within one or more effective intervals, smooth and summarize to obtain the individual sensitivity coefficient, and use the individual sensitivity coefficient to correct the amplitude of the thyroid effect quantity to generate the thyroid effect surrogate quantity sequence.

[0048] Specifically, this application achieves stable quantification of the dynamic cardiovascular response induced by thyroid dysfunction without the need for additional equipment by aligning the effective start point / response start point intervals and constructing a hysteresis-constrained recursive effect sequence. It also reduces misjudgments caused by baseline differences through individual sensitivity coefficient correction. It should be noted that the above method retrospectively processes the acquired follow-up data, and each step is executed by a computer (processor). The output thyroid effect proxy sequence and auxiliary grading reference information are used for auxiliary assessment and manual review during the follow-up process.

[0049] In an optional embodiment, the effective starting point is determined based on the direction of change of the first deviation between adjacent test points and the overall direction of change of the second deviation within a hysteresis window starting from the test point, specifically including:

[0050] It should be noted that, in the thyroid test sequence, a valid starting point is identified where the TSH deviation changes in the same direction and the FT4 deviation changes in the opposite direction within the hysteresis window. This step is used to capture a starting point with clear thyroid axis driving characteristics, reflecting the physiological hysteresis coupling relationship between TSH and FT4. The specific implementation method is as follows:

[0051] Traverse the three adjacent test points of the thyroid test sequence, calculate the first change and the second change of the first deviation in the two adjacent time intervals, and when the two have the same sign and the absolute value is not less than the preset same-direction threshold, determine the second test point as the candidate starting point.

[0052] Specifically, three adjacent test points are extracted sequentially from the thyroid test sequence. Calculate the change in the first deviation over two adjacent time intervals:

[0053] First change Defined as: ;

[0054] Second change Defined as: ;

[0055] In the specific example above, for the 3rd test point (2025-03-08): , ;

[0056] when and If the signs are the same and their absolute values ​​are not less than the preset same-direction threshold (which can be 0.05), the second test point will be... If the same sign is used as a candidate starting point, it means that the TSH deviation shows the same trend in two consecutive time intervals, reflecting the continuous rise or fall of TSH level.

[0057] In the specific example above, the third test point (2025-03-08) meets the requirements. and ,and , Therefore, the third test point was determined as a candidate starting point. In this example, the second test point (2025-02-15) also met the candidate starting point condition, but according to the deduplication rule of multiple valid starting points in the same direction segment in the following text, the cumulative change of its first deviation (0.17) is less than that of the third test point (0.30). Therefore, the third test point was finally selected as the only valid starting point.

[0058] The second deviation sequence is extracted within the lag window starting from the candidate starting point. If the number of test points exceeds the preset number, the linear regression slope is calculated as the trend value. If it is lower, the difference between the second deviation at the end of the lag window and the second deviation at the beginning is taken as the trend value. The lag window is a time window of preset length.

[0059] In this method, a linear regression is performed with the timestamp as the independent variable and the second deviation as the dependent variable, and the slope is taken as the trend value. For example, the preset length can be 14-65 days, and the typical value can be 45 days.

[0060] In the specific example above, the third test point (2025-03-08) is taken as the candidate starting point, and the lag window is [2025-03-08, 2025-04-22]. The lag window contains the third, fourth, and fifth test points.

[0061] When the sign of the trend quantity is opposite to the sign of the first change quantity, and the absolute value of the trend quantity is not less than the preset reverse threshold, the candidate starting point is confirmed as a valid starting point; wherein, the preset reverse threshold can be 0.05;

[0062] When the second deviation sequence exhibits more than a preset number of difference sign flips within the hysteresis window, the determination of the candidate starting point as a valid starting point is cancelled. The preset number can be 2. A difference sign flip means that the FT4 deviation frequently bounces back within the hysteresis window, which does not conform to a stable trend change and may be due to measurement error or transient physiological fluctuations.

[0063] If multiple candidate starting points are identified as valid starting points within the same direction of change segment, the valid starting point with the largest cumulative change in the first deviation amount is selected as the only valid starting point. The cumulative change amount is defined as the absolute value of the cumulative change in the first deviation amount from the start of the same direction of change segment to the candidate starting point, which ensures that the starting moment when the TSH deviation amount change is most significant is captured.

[0064] Through the above technical solution, this embodiment improves the robustness and repeatability of valid starting point determination by reverse verification and reversal suppression of the trend of candidate starting points, and provides a stable anchor point for subsequent effective interval alignment.

[0065] In an optional embodiment, the difference composite quantity is filtered and fused according to preset combination conditions, the extreme value moment of the difference composite quantity is taken as the response starting point, and the response type is determined, specifically including:

[0066] It should be noted that the relative difference is obtained by normalizing the difference between the sub-parameter value and the baseline at each sampling time within the effective interval.

[0067] Multiple sets of preset combination conditions are set. The preset combination conditions include the sign constraint and magnitude constraint of the relative difference of each dimension sub-parameter, and their associated response types. The response types include at least a first response type and a second response type.

[0068] The first response type is where the relative difference in heart rate and the relative difference in pulse pressure are in the same direction and both exceed the corresponding preset difference threshold, and satisfy the preset relationship between the relative difference in diastolic blood pressure and the relative difference in pulse pressure; the second response type is where the relative difference in diastolic blood pressure exceeds the corresponding preset difference threshold, and satisfies the preset relationship between the relative difference in pulse pressure and the relative difference in heart rate; the relative difference in pulse pressure is determined by the relative difference in systolic blood pressure and the relative difference in diastolic blood pressure.

[0069] For example, the specific constraints of the first response type and the second response type:

[0070] The first response type is defined as: relative difference in heart rate. Relative difference with pulse pressure Both are in the same direction and exceed the threshold. , And satisfy the relative difference in diastolic blood pressure and The pre-defined relationship;

[0071] The second response type is defined as: Exceeding the threshold and satisfy and The preset relationship; where pulse pressure , This is the relative difference in pulse pressure, i.e. , This represents the baseline of the anterior window intra-pulse pressure. This represents the relative difference in systolic blood pressure.

[0072] The preset relationships can be as follows:

[0073] when At that time, the first response type requires or ( Typical value is 0.6, range 0.4-0.8); Second response type requirements or ( (Typical value is 0.5, range 0.3-0.8); used to constrain the difference between pulse pressure dominance and diastolic pressure dominance, and suppress the same-direction drift caused by cuff measurement error;

[0074] The sampling time that meets any preset combination of conditions in the back window is selected, and the relative differences of each dimension of the sampling time are fused according to the first preset weight to obtain the difference synthesis amount. At the same time, the response type corresponding to the preset combination of conditions is associated. The sampling time is the timestamp of the cardiovascular observation data collection.

[0075] For example, the difference synthesis quantity can be:

[0076] ;

[0077] In the formula, This is the difference synthesis quantity. , , The first preset weight can be 0.35, 0.35, or 0.30 respectively;

[0078] The search identifies continuous segments where the difference synthesis quantity satisfies a preset continuity condition, and selects the sampling moment with the largest absolute value of the difference synthesis quantity as the response starting point. If multiple extreme values ​​with equal absolute values ​​exist, the earliest one is selected. The preset continuity condition can be: at least n consecutive sampling moments (n can be 5) satisfy the difference synthesis quantity... The signs are consistent and the amplitude exceeds the preset combined amplitude threshold (e.g., 0.05).

[0079] Through the above technical solution, this embodiment achieves robust identification of the starting point and response mode of cardiovascular response in a noisy background by screening the combination conditions of multidimensional signs and locating the extreme value of the synthetic quantity, providing a reliable basis for constructing the effective interval and recursively estimating subsequent effects.

[0080] In an optional embodiment, the thyroid effector is initialized with a second deviation at the effective starting point, and a thyroid effector sequence is generated sequentially within the effective interval, specifically including:

[0081] For the second deviation of each test point in the thyroid test sequence, the second deviation is mapped to the sampling time with the smallest time difference according to the timestamp, resulting in several mapped sampling times. The remaining sampling times within the effective interval are then completed using linear interpolation. The time point is the sampling time after mapping and completion. Similarly, for the first deviation of each test point, the first deviation is mapped to the sampling time in the same way, and linear interpolation is performed on the remaining sampling times (forward hold or backward hold is used when adjacent mapped points are insufficient for interpolation), resulting in a first deviation sequence coaxial with the sampling times to support the calculation of hysteresis constraints on the sampling time scale. When two adjacent mapped sampling times are insufficient for linear interpolation, forward hold or backward hold is used to complete the corresponding sampling times.

[0082] The process of completing the linear interpolation is as follows: for each test point During the effective period Search for and within all sampling times Sampling time with the smallest time difference ;Will Mapped to this sampling time, denoted as For the effective period For sampling moments that are not mapped, linear interpolation is used to complete them.

[0083] Let the two adjacent mapped sampling times be . and The corresponding second deviation is and Then located at and sampling time between The second deviation is:

[0084] ;

[0085] At each sampling time in the effective interval, the update amount is obtained by the change of the second deviation amount, and the hysteresis constraint amount is generated based on the change of the first deviation amount within the hysteresis window starting from the sampling time.

[0086] The update quantity is truncated using the hysteresis constraint and then superimposed on the thyroid effect quantity at the previous sampling time to obtain a candidate thyroid effect quantity. The absolute value of the truncated update quantity does not exceed the hysteresis constraint.

[0087] This embodiment adopts the following specific implementation method:

[0088] The second deviation of each test point in the thyroid test sequence Mapped to the sampling time with the smallest time difference from the timestamp. This yields the second deviation at several mapped sampling times, and applies it to the effective interval. The remaining sampling times within the time frame are completed using linear interpolation;

[0089] At the effective starting point Corresponding sampling time At the second deviation amount Initialize thyroid effector : ;

[0090] Effective period Each sampling time within (Recursively) Perform the following update steps:

[0091] V1. Generates an update value based on the change in the second deviation. : in, The current sampling time The second deviation, The next sampling time The second deviation;

[0092] V2. At the sampling time Lag window starting from ( Calculate the change range of the first deviation within its preset length to obtain the hysteresis constraint. ;

[0093] The magnitude of the first deviation can be defined in several ways:

[0094] Method 1 (Cumulative Changes): , which represents the total change in the first deviation within the hysteresis window.

[0095] Method 2 (range): This indicates the maximum value of the first deviation within the hysteresis window. and minimum value difference;

[0096] Method 3 (Weighted Change Amplitude): Take the absolute value of each change of the first deviation within the lag window and then sum them up with weights; This embodiment uses Method 1 (Cumulative Change). The lag constraint reflects the constraint strength of TSH change on the FT4 effect update. When the TSH change amplitude within the lag window is large, it indicates that the thyroid axis regulation is active and allows a large FT4 effect update amplitude. Conversely, the update amplitude is constrained.

[0097] V3. Use hysteresis constraints to truncate the update quantity to obtain the truncated update quantity. :

[0098] ;in, The function will Limited to Within the range:

[0099] The cutoff mechanism ensures that the update magnitude of the thyroid effector does not exceed the change magnitude of TSH within the hysteresis window, reflecting the constraining effect of thyroid axis feedback regulation.

[0100] V4. When any of the following conditions are met, the candidate thyroid effect size is corrected to the thyroid effect at the previous sampling time to suppress unreasonable effect size changes:

[0101] The second deviation occurs when the trend determined by the second deviation sequence within the hysteresis window corresponding to the sampling time is opposite and the duration is less than the preset minimum duration. The trend is calculated using the same method as linear regression slope or endpoint difference. The rebound is defined as the second deviation being opposite in sign to the trend, and the preset minimum duration can be 3 days.

[0102] When the response type is the first response type, the direction of change of the candidate thyroid effect size is opposite to the main trend direction of the difference synthesis size, and the preset persistence condition is met.

[0103] It should be noted that, at the current sampling time, the direction of change of the candidate thyroid effect quantity compared to the previous sampling is determined to be either increasing or decreasing. The continuous segment where the response starts is located is used as the defining range. The main trend direction of the difference synthesis quantity within this range is characterized as continuously increasing or continuously decreasing. The main trend direction is determined by the sign of the linear regression slope of the difference synthesis quantity with respect to time within this continuous segment, or the sign of the difference synthesis quantity between the beginning and end of this segment. When the current direction of change of the candidate thyroid effect quantity is opposite to the main trend direction of the difference synthesis quantity (i.e., the main trend increases while the thyroid effect quantity decreases, or the main trend decreases while the thyroid effect quantity increases), the judgment condition that the candidate thyroid effect quantity shows the opposite trend to the main trend of the difference synthesis quantity is triggered, and the preset persistence condition must be met (e.g., the opposite for 5 consecutive days).

[0104] Furthermore, the first response type corresponds to the typical pattern of hyperthyroidism or hypothyroidism, and the differential synthesis amount should be consistently positive or consistently negative. If the thyroid effector shows a long-term reverse change, which is inconsistent with the observed cardiovascular response, it should be corrected.

[0105] When the response type is the second response type, the direction of change of the candidate thyroid effect size changes from the same to the opposite direction of the direction of the observed difference synthesis size in the same period, and the preset persistence condition is met.

[0106] It should be noted that the current direction of change of the candidate thyroid effect quantity is compared with the direction of the observed difference composite quantity during the same period. If the two change from being consistent (in the same direction) to being opposite (in reverse), that is, the consistency of direction is reversed, and the preset persistence condition is met (such as being opposite for 5 consecutive days), then the correction condition is determined to be met.

[0107] Furthermore, the second response type corresponds to a response pattern dominated by changes in diastolic blood pressure. The direction of the differential synthesis should remain stable. If the direction of change of the candidate thyroid effector changes from the same to the opposite direction of the differential synthesis, which is inconsistent with the observed cardiovascular response, it should be corrected.

[0108] The corrected candidate thyroid effect size is determined as the thyroid effect size at the current sampling time, and the thyroid effect size sequence composed of each sampling time within the effective interval is output.

[0109] Through the above technical solution, this embodiment obtains the coaxial second deviation sequence through mapping interpolation, and constructs the continuous effect quantity sequence with hysteresis constraint truncation and bounce correction, which improves the stability and physiological consistency of thyroid effect estimation and reduces erroneous updates caused by short-term noise.

[0110] In an optional embodiment, the response strength of the synthesized difference relative to the change in thyroid effect size within multiple effective intervals is statistically analyzed, and the individual sensitivity coefficient is obtained by smoothing and summarizing. Specifically, this includes:

[0111] Based on the thyroid effect quantity sequence, within each effective interval, the difference change of the thyroid effect quantity and the difference change of the composite difference are calculated for adjacent sampling times and form change pairs. Change pairs with an absolute value of the difference change of the thyroid effect quantity less than the preset minimum change threshold are removed.

[0112] For the retained change pairs, the absolute value of the ratio of the difference in the composite difference to the difference in the thyroid effect is calculated as the instantaneous response intensity. Outliers with instantaneous response intensities exceeding the preset response range are removed, and the response intensity of the effective interval is determined by the median of the remaining instantaneous response intensities.

[0113] It should be noted that the preset minimum change threshold can be 0.01; the instantaneous response intensity is taken as the absolute value of the ratio, so that the two types of hyperthyroidism with elevation and hypothyroidism with reduction are measured in the same direction. The preset response range can be [0.1, 5]. The response intensity exceeding this range may be due to measurement error or abnormal physiological state; the median of the remaining instantaneous response intensity is used as the response intensity of the effective interval. The median has better robustness than the average value and is less affected by extreme values.

[0114] The normalized fluctuation of each dimension is obtained by calculating the ratio of the standard deviation of the fluctuation of the baseline of each sub-parameter to its corresponding baseline within the front window corresponding to each effective interval. The baseline stability is obtained by summing the values ​​according to the second preset weight, adding the preset positive number, and taking the reciprocal. The individual sensitivity coefficient is obtained by weighted averaging the response intensity of each effective interval.

[0115] For example, the normalized fluctuation of the baseline of each sub-parameter is calculated within the corresponding front window of each effective interval, which is the ratio of the standard deviation of the heart rate baseline to the heart rate baseline. The ratio of the standard deviation of systolic blood pressure to the baseline systolic blood pressure The ratio of the baseline standard deviation of diastolic blood pressure to the baseline diastolic blood pressure The corresponding second preset weights can be 0.4, 0.3, and 0.3, respectively;

[0116] Baseline stability ,in, A positive number (0.01 is acceptable) is preset to avoid division by zero; the higher the baseline stability, the more stable the cardiovascular parameters within the anterior window, the more reliable the response strength of the effective interval, and the higher the weight in the aggregation.

[0117] To smooth the individual sensitivity coefficient, you can use the moving median, exponential smoothing, or other smoothing methods. If there are only a few (e.g., less than 5) effective intervals, you can directly use the weighted average result.

[0118] Through the above technical solution, this embodiment forms a sensitivity coefficient oriented towards individual differences by statistically analyzing the robust response intensity of multiple effective intervals and weighting it with baseline stability, and provides a basis for effect size magnitude correction, thereby reducing misjudgment and missed judgment under the general threshold.

[0119] In an optional embodiment, the magnitude of the thyroid effector is corrected using the individual sensitivity coefficient to generate a thyroid effector proxy, specifically including:

[0120] Within each effective interval, the thyroid effect size sequence is scaled proportionally according to the individual sensitivity coefficient to obtain the initial effect proxy sequence;

[0121] A physiological adaptation function with the duration of the effective interval as the independent variable is constructed, and a decay factor that gradually decreases over time and has a preset lower limit is calculated.

[0122] It should be noted that the physiological adaptation function is used to apply drift suppression (repeated measurement suppression) that increases with duration to the long-term accumulated effect proxy within the effective interval, in order to reduce the impact of long-term baseline drift and slowly varying noise on the proxy; this embodiment uses an exponential decay function with a lower limit:

[0123] ;

[0124] In the formula, As the attenuation factor, The effective period has lasted for the duration of the period, i.e., from the effective start date. up to the current sampling time The time interval, This is the decay rate parameter (which can be taken as 0.01 / day). Set a preset lower limit (which can be 0.5) to ensure that the attenuation factor is not lower than this value;

[0125] When the duration of the effective interval exceeds the preset trigger duration, the initial effect proxy sequence is attenuated and smoothed using an attenuation factor, and fluctuations that are inconsistent with the response type sign constraint and do not last longer than the preset minimum duration are removed, and the final thyroid effect proxy sequence is output.

[0126] For example, the preset trigger duration can be 15-30 days, with a typical value of 21 days;

[0127] Through the above technical solution, this embodiment introduces duration-driven adaptive decay and sign consistency cleaning on the basis of individual sensitivity correction, so that the final thyroid effect proxy can reflect the short-term cumulative effect and suppress long-term drift and short-term noise, thereby improving the consistency of interpretation and the usability of long-term follow-up.

[0128] In an optional embodiment, after generating the thyroid effector surrogate sequence, a grading analysis is further included, specifically:

[0129] Detect target segments in the thyroid effect proxy sequence that meet the preset proxy conditions, and extract at least one of the extreme value amplitude, duration and interval integral in each target segment to generate a segment intensity index;

[0130] For example, the preset proxy quantity condition can be defined as: the absolute value of the thyroid effect proxy quantity of at least m consecutive sampling points (m can be 7 points, corresponding to 7 days) in the thyroid effect proxy quantity sequence exceeds the preset proxy quantity amplitude threshold (which can be 0.08).

[0131] The extreme value amplitude is the maximum absolute value of the thyroid effect proxy within the target segment, reflecting the peak intensity of the effect; the duration is the time span of the target segment, reflecting the duration of the effect; the interval integral is the integral of the absolute value of the thyroid effect proxy within the segment over time, reflecting the cumulative intensity of the effect.

[0132] The individual impact score is obtained by weighting and summing the intensity indices of each segment, and compared with the preset grading threshold to output auxiliary grading reference information on the impact of thyroid dysfunction on the cardiovascular system. At the same time, the start and end times of the target segment and the corresponding segment intensity index are output as explanatory information.

[0133] It should be noted that the intensity indices of each segment are weighted and summarized. The segment weights are determined based on the time distance between the target segment and the effective starting point of the corresponding effective interval, as well as the baseline stability of the interval to which the segment belongs: First, the absolute value of the time difference between the center time of the target segment and the effective starting time is calculated as the time distance, and the time distance is converted into a distance weight according to the monotonic decay rule (the smaller the time distance, the larger the distance weight); the baseline stability is converted into a stability weight according to the monotonic gain rule (the larger the baseline stability, the larger the stability weight); finally, the distance weight and the stability weight are combined to obtain the segment weight, and after normalization, the sum of the segment weights is made to be 1.

[0134] The grading threshold is set as follows (Individual Impact Score): Low Impact Level Medium impact level High impact level ;

[0135] The low impact level indicates that the overall response intensity of the target object's cardiovascular response relative to the thyroid effect proxy within the target segment is weak or insufficient, indicating a low degree of dynamic impact of thyroid dysfunction on the cardiovascular system; the medium impact level indicates that the overall response intensity is within the common response range and has a certain degree of persistence, suggesting the existence of a quantifiable dynamic impact; the high impact level indicates that the overall response intensity is significantly enhanced and / or the duration is long, suggesting a high degree of dynamic impact of thyroid dysfunction on the cardiovascular system; and the start and end times of the target segment that generated the score and its segment intensity index are output for interpretation and verification.

[0136] Through the above technical solution, this embodiment achieves interpretable quantification of the degree of influence on individuals by extracting and grading the surrogate quantity in fragments, and provides verifiable interpretive information with start and end times and fragment intensity indicators, thereby improving the usability of the results in clinical applications.

[0137] See Figure 2 As shown, this scheme proposes a quantitative assessment system for the impact of thyroid dysfunction on the cardiovascular system, used to implement the aforementioned quantitative assessment method for the impact of thyroid dysfunction on the cardiovascular system, including:

[0138] The data extraction and conversion module is used to acquire the thyroid test sequence and cardiovascular observation data of the target object, extract the thyroid-stimulating hormone value and free thyroxine value of each test point in the thyroid test sequence, and convert them into the first deviation and the second deviation relative to the preset reference value, respectively.

[0139] The effective starting point identification module is used to determine the effective starting point based on the direction of change of the first deviation between adjacent test points and the overall direction of change of the second deviation within the hysteresis window starting from the test point, and to set the front window and the back window based on the effective starting point.

[0140] The cardiovascular response analysis module is used to determine the baseline of each sub-parameter in the front window and calculate the relative difference in the back window based on cardiovascular observation data. It filters and merges the sub-parameters into a composite difference value according to preset combination conditions, takes the extreme value of the composite difference value as the response start point and determines the response type. Each sub-parameter includes at least heart rate, systolic blood pressure and diastolic blood pressure.

[0141] The effect size generation module is used to determine the effective interval based on the effective start point and the response start point, initialize the thyroid effect size with the second deviation at the effective start point, and recursively update and generate the thyroid effect size sequence within the effective interval according to the time sequence.

[0142] The surrogate quantity synthesis module is used to statistically analyze the response intensity of the difference synthesis quantity relative to the thyroid effect quantity sequence within one or more effective intervals, smooth and summarize it to obtain the individual sensitivity coefficient, and use the individual sensitivity coefficient to correct the amplitude of the thyroid effect quantity to generate the thyroid effect surrogate quantity sequence.

[0143] In another embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above embodiments.

[0144] In one embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps described above.

[0145] In one embodiment, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and executes the computer instructions, causing the computer device to perform the steps described above.

[0146] The technical scope of this invention is not limited to the content described above. Those skilled in the art can make various modifications and variations to the above embodiments without departing from the technical concept of this invention, and all such modifications and variations should fall within the protection scope of this invention.

Claims

1. A method for quantitatively assessing the impact of thyroid dysfunction on the cardiovascular system, characterized in that, The method includes: Obtain the thyroid test sequence and cardiovascular observation data of the target subject, extract the thyroid-stimulating hormone value and free thyroxine value of each test point in the thyroid test sequence, and convert them into the first deviation and the second deviation relative to the preset reference value, respectively. The effective starting point is determined based on the direction of change of the first deviation between adjacent test points and the overall direction of change of the second deviation within the hysteresis window starting from the test point, and the front window and back window are set based on the effective starting point. Based on cardiovascular observation data, baselines for each sub-parameter within the anterior window are determined, and relative differences within the posterior window are calculated. Multiple preset combination conditions are set, and these conditions are filtered and fused into a composite difference value. The extreme value of the composite difference value is taken as the response starting point, and the response type is determined. Each sub-parameter includes at least heart rate, systolic blood pressure, and diastolic blood pressure. The preset combination conditions include sign and amplitude constraints on the relative differences of each sub-parameter, and associate them with response types. The response types include at least a first response type and a second response type. The first response type is where the relative difference in heart rate and the relative difference in pulse pressure are in the same direction and both exceed the corresponding preset difference thresholds, and satisfy the preset relationship between the relative difference in diastolic blood pressure and the relative difference in pulse pressure. The second response type is where the relative difference in diastolic blood pressure exceeds the corresponding preset difference threshold, and satisfies the preset relationship between the relative difference in pulse pressure and the relative difference in heart rate. The relative difference in pulse pressure is determined by the relative difference in systolic blood pressure and the relative difference in diastolic blood pressure. The effective interval is determined based on the effective starting point and the response starting point. The thyroid effect quantity is initialized with the second deviation at the effective starting point and updated sequentially within the effective interval to generate the thyroid effect quantity sequence. The thyroid effect quantity at each time point is updated by combining the thyroid effect quantity at the previous time point with the change of the second deviation. The update magnitude is constrained by the change of the first deviation within the hysteresis window and the reverse change of the thyroid effect quantity is restricted according to the response type. The process of initializing the thyroid effect quantity with the second deviation at the effective starting point and then recursively updating the thyroid effect quantity sequence within the effective interval includes: For the second deviation of each test point in the thyroid test sequence, the second deviation of several mapped sampling timestamps are mapped to the sampling time with the smallest time difference, and the remaining sampling times in the effective interval are completed by linear interpolation. The time point is the sampling time after mapping and completion; the sampling time is the timestamp of the cardiovascular observation data collection. At each sampling time in the effective interval, the update amount is obtained by the change of the second deviation amount, and the hysteresis constraint amount is generated based on the change of the first deviation amount within the hysteresis window starting from the sampling time. The update quantity is truncated using the hysteresis constraint and then superimposed on the thyroid effect quantity at the previous sampling time to obtain a candidate thyroid effect quantity. The absolute value of the truncated update quantity does not exceed the hysteresis constraint. The candidate thyroid effect quantity is corrected to the thyroid effect quantity of the previous sampling time when any of the following conditions are met: the second deviation quantity shows a rebound that is opposite to the trend quantity determined by the second deviation quantity sequence within the hysteresis window corresponding to the sampling time and lasts for less than the preset minimum duration; when the response type is the first response type, the change direction of the candidate thyroid effect quantity is opposite to the main trend direction of the difference synthesis quantity and meets the preset duration condition; when the response type is the second response type, the change direction of the candidate thyroid effect quantity changes from the same to the opposite direction of the observed difference synthesis quantity in the same period and meets the preset duration condition; the corrected candidate thyroid effect quantity is determined as the thyroid effect quantity of the current sampling time, and the thyroid effect quantity sequence composed of each sampling time within the effective interval is output; The response intensity of the difference synthesis quantity relative to the thyroid effect quantity sequence within one or more effective intervals is statistically analyzed, smoothed and summarized to obtain the individual sensitivity coefficient, and the amplitude of the thyroid effect quantity is corrected using the individual sensitivity coefficient to generate the thyroid effect surrogate quantity sequence.

2. The method according to claim 1, characterized in that, The effective starting point is determined based on the direction of change of the first deviation between adjacent test points, and the overall direction of change of the second deviation within the hysteresis window starting from the test point. Specifically, this includes: Traverse the three adjacent test points of the thyroid test sequence, calculate the first change and the second change of the first deviation in the two adjacent time intervals, and determine the second test point as the candidate starting point when the two have the same sign and the absolute value is not less than the preset same-direction threshold. The second deviation sequence is extracted within the lag window starting from the candidate starting point. If the number of test points exceeds the preset number, the linear regression slope is calculated as the trend value. If it is lower, the difference between the second deviation at the end of the lag window and the second deviation at the beginning is taken as the trend value. The lag window is a time window of preset length. When the sign of the trend quantity is opposite to the sign of the first change quantity, and the absolute value of the trend quantity is not less than the preset reverse threshold, the candidate starting point is confirmed as a valid starting point. When the second deviation sequence exhibits more than a preset number of difference sign flips within the hysteresis window, the determination of the candidate starting point as a valid starting point is cancelled. If multiple candidate starting points are identified as valid starting points within the same direction of change, then the valid starting point with the largest cumulative change in the first deviation amount is selected as the sole valid starting point.

3. The method according to claim 1, characterized in that, The differential composite quantity is selected and merged according to preset combination conditions. The extreme value of the differential composite quantity is taken as the response starting point and the response type is determined. Specifically, this includes: The sampling time that meets any preset combination of conditions in the back window is selected, and the relative differences of each dimension of the sampling time are fused according to the first preset weight to obtain the difference synthesis amount, and at the same time, the response type corresponding to the preset combination of conditions is associated. Search for continuous segments where the difference synthesis quantity satisfies the preset continuity condition, and take the extreme value moment of the difference synthesis quantity as the response starting point.

4. The method according to claim 1, characterized in that, The response strength of the combined difference relative to changes in thyroid effect size within multiple effective intervals was statistically analyzed, and the individual sensitivity coefficient was obtained by smoothing and summarizing. Specifically, this included: Based on the thyroid effect quantity sequence, within each effective interval, the difference change of the thyroid effect quantity and the difference change of the composite difference are calculated for adjacent sampling times and form change pairs. Change pairs with an absolute value of the difference change of the thyroid effect quantity less than the preset minimum change threshold are removed. For the retained change pairs, the absolute value of the ratio of the difference in the composite difference to the difference in the thyroid effect is calculated as the instantaneous response intensity. Outliers with instantaneous response intensities exceeding the preset response range are removed, and the response intensity of the effective interval is determined by the median of the remaining instantaneous response intensities. The normalized fluctuation of each dimension is obtained by calculating the ratio of the standard deviation of the fluctuation of the baseline of each sub-parameter to its corresponding baseline within the front window corresponding to each effective interval. The baseline stability is obtained by summing the values ​​according to the second preset weight, adding the preset positive number, and taking the reciprocal. The individual sensitivity coefficient is obtained by weighted averaging the response intensity of each effective interval.

5. The method according to claim 4, characterized in that, The thyroid effector surrogate is generated by correcting the magnitude of the thyroid effector using the individual sensitivity coefficient, specifically including: Within each effective interval, the thyroid effect size sequence is scaled proportionally according to the individual sensitivity coefficient to obtain the initial effect proxy sequence; A physiological adaptation function with the duration of the effective interval as the independent variable is constructed, and a decay factor that gradually decreases over time and has a preset lower limit is calculated. When the duration of the effective interval exceeds the preset trigger duration, the initial effect proxy sequence is attenuated and smoothed using an attenuation factor, and fluctuations that are inconsistent with the response type symbol constraint and whose duration is less than the preset minimum duration are removed, and the final thyroid effect proxy sequence is output.

6. The method according to claim 5, characterized in that, After generating the thyroid effector surrogate sequence, a grading analysis is also performed, specifically including: Detect target segments in the thyroid effect proxy sequence that meet the preset proxy conditions, and extract at least one of the extreme value amplitude, duration and interval integral in each target segment to generate a segment intensity index; The individual impact score is obtained by weighting and summing the intensity indices of each segment, and compared with the preset grading threshold to output auxiliary grading reference information on the impact of thyroid dysfunction on the cardiovascular system. At the same time, the start and end times of the target segment and the corresponding segment intensity index are output as explanatory information.

7. A quantitative assessment system for the impact of thyroid dysfunction on the cardiovascular system, characterized in that, A method for quantitatively assessing the impact of thyroid dysfunction on the cardiovascular system as described in any one of claims 1-6, comprising: The data extraction and conversion module is used to acquire the thyroid test sequence and cardiovascular observation data of the target object, extract the thyroid-stimulating hormone value and free thyroxine value of each test point in the thyroid test sequence, and convert them into the first deviation and the second deviation relative to the preset reference value, respectively. The effective starting point identification module is used to determine the effective starting point based on the direction of change of the first deviation between adjacent test points and the overall direction of change of the second deviation within the hysteresis window starting from the test point, and to set the front window and the back window based on the effective starting point. The cardiovascular response analysis module is used to determine the baseline of each sub-parameter in the front window and calculate the relative difference in the back window based on cardiovascular observation data. It filters and merges the sub-parameters into a composite difference value according to preset combination conditions, takes the extreme value of the composite difference value as the response start point and determines the response type. Each sub-parameter includes at least heart rate, systolic blood pressure and diastolic blood pressure. The effect size generation module is used to determine the effective interval based on the effective start point and the response start point, initialize the thyroid effect size with the second deviation at the effective start point, and recursively update and generate the thyroid effect size sequence within the effective interval according to the time sequence. The surrogate quantity synthesis module is used to statistically analyze the response intensity of the difference synthesis quantity relative to the thyroid effect quantity sequence within one or more effective intervals, smooth and summarize it to obtain the individual sensitivity coefficient, and use the individual sensitivity coefficient to correct the amplitude of the thyroid effect quantity to generate the thyroid effect surrogate quantity sequence.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the method described in any one of claims 1-6.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-6.

Images