Anastomosis device

Abstract

This invention discloses a method for implementing site-specific treatment of ankylosing spondylitis. It acquires multimodal imaging and clinical data from patients, constructs an anatomical region division model including the sacroiliac joint, thoracic spine, lumbar spine, cervical spine, and peripheral joints. Based on this model, the patient's spine is divided into at least five independent functional assessment units, each corresponding to a specific anatomical coordinate range. Local inflammatory marker values ​​and range of motion limitations are collected for each functional assessment unit. This invention achieves a systematic digital reconstruction of the pathological space of ankylosing spondylitis through the linkage between the anatomical region division model and the functional assessment unit construction units. The independent division based on anatomical coordinates ensures that subsequently collected local inflammatory marker values ​​correspond to specific anatomical structures, avoiding interference between inflammatory signals from different segments.

Description

Technical Field

[0001] This invention belongs to the field of ankylosing spondylitis treatment technology, specifically a method for organizing and implementing treatment of ankylosing spondylitis by location. Background Technology

[0002] Ankylosing spondylitis is a chronic systemic inflammatory rheumatic bone disease. Lesions are scattered, affecting all segments of the spine, the sacroiliac joints, and peripheral joints of the limbs. The progression of inflammation, bone and joint mechanical damage, and muscle dysfunction vary greatly depending on the anatomical location. Localized treatment is currently the direction of standardized clinical practice. However, existing treatments for ankylosing spondylitis have the following technical problems: First, current clinical treatment protocols for ankylosing spondylitis and the adaptation of regional treatments largely rely on comprehensive inflammatory indicators and overall joint function data for overall assessment. They are based solely on broad categories such as the spine and peripheral joints for coarse regional divisions, lacking a refined anatomical spatial digital breakdown logic. This results in clinical treatment protocols that are highly universal but poorly adaptable to different locations. Furthermore, the existing treatment system is a one-way matching model, relying solely on static data from the initial diagnosis to formulate treatment content. There is no dynamic data feedback or adaptive correction mechanism during the treatment process, and the regional treatment content cannot be iteratively adjusted according to real-time changes in the lesions and the patient's condition.

[0003] Secondly, current technologies cannot achieve three-dimensional independent segmentation of ankylosing spondylitis lesions. The collection of inflammatory functional data from across the entire region is mixed, and inflammatory signals and biomechanical data from different bone and joint segments interfere with each other. It is impossible to collect high-fidelity physiological data for a single anatomical site, resulting in a single basis for treatment based on different locations, severe homogenization of treatment plans, and poor site specificity and clinical repeatability. Furthermore, the existing treatment system lacks a mechanism for linking inflammatory indicators, joint biomechanics, and physiological sensor data for analysis. On the one hand, it cannot dynamically adjust the dosage of drugs administered to different areas based on the drug sensitivity characteristics of the lesions, easily leading to insufficient local treatment efficacy or excessive systemic drug accumulation. On the other hand, it cannot capture real-time physiological signals within the joint cavity and muscle movement, failing to accurately locate and treat high-risk lesion sites. This results in insufficient precision in the implementation of segmented treatment and poor controllability of treatment safety. Summary of the Invention

[0004] In view of the above situation and to overcome the shortcomings of the prior art, the present invention provides a method for sorting out and implementing the treatment content of ankylosing spondylitis by location, so as to at least partially solve the above technical problems.

[0005] The technical solution adopted in this invention is as follows: This invention proposes a method for implementing site-specific treatment for ankylosing spondylitis, comprising the following steps: Step 1: Obtain multimodal imaging data and clinical sign data of patients, and construct an anatomical region division model including the sacroiliac joint, thoracic vertebrae, lumbar vertebrae, cervical vertebrae and peripheral joints; Step 2: Based on the anatomical region division model, the patient's spine is divided into at least five independent functional assessment units, each corresponding to a specific anatomical structure coordinate range; Step 3: Collect local inflammation index values ​​and activity limit values ​​of each functional assessment unit to generate a multidimensional feature vector set; Step 4: Using a pre-defined partial treatment rule library, match the corresponding combination of treatment parameters according to the multi-dimensional feature vector set. The combination of treatment parameters includes physical therapy intensity, drug dosage gradient, and rehabilitation training action sequence. Step 5: Monitor the response data of each functional evaluation unit in real time during the treatment process through the dynamic feedback module. If the response data deviates from the preset threshold range, automatically adjust the value of at least one dimension in the treatment parameter combination of the next cycle to form a closed-loop control path.

[0006] In one embodiment of the present invention, the anatomical region division model adopts a three-dimensional spatial coordinate system, wherein the sacroiliac joint region is defined as a three-dimensional volume unit at the junction of the S1 to S3 vertebral bodies and the iliac wing; the thoracic vertebral region is defined as a stacked volume unit of the T1 to T12 vertebral bodies and their associated ribs; the lumbar vertebral region is defined as a volume unit of the L1 to L5 vertebral bodies and their associated psoas major muscle attachment points; the cervical vertebral region is defined as a volume unit of the C1 to C7 vertebral bodies and their associated atlantoaxial joint structures; and the peripheral joint region is defined as an intracapsular space volume unit of the shoulder, hip, knee, and ankle joints.

[0007] In one embodiment of the present invention, the local inflammatory index values ​​include the mean signal intensity of T2-weighted images in magnetic resonance imaging, the blood flow grading value in ultrasound examination, and the local concentration ratio of C-reactive protein in blood tests; the range of motion limitation values ​​include the difference in forward flexion angle, the difference in lateral bending angle, the difference in extension angle, and the difference in rotation angle; the generation process of the multidimensional feature vector set is to normalize the above-mentioned local inflammatory index values ​​and range of motion limitation values ​​according to preset weight coefficients and then concatenate them to form an n-dimensional vector, where n is equal to the sum of the number of local inflammatory indexes and the number of range of motion limitations.

[0008] In one embodiment of the present invention, the segmented treatment rule base includes at least three levels of matching logic: the first level determines whether to initiate the biological agent intervention process based on whether the inflammation index of the sacroiliac joint region exceeds a preset threshold; the second level determines the ratio of the axial component to the shear force component of the traction force applied in physical therapy based on the ratio of the range of motion limits of the thoracic and lumbar vertebrae; the third level determines the alternation frequency of neck stability training and lower limb resistance training in the rehabilitation training action sequence based on the coupling degree of the response data of the cervical vertebrae and peripheral joints; the drug dosage gradient in the treatment parameter combination is calculated by a piecewise linear function, the slope of which is determined by the inflammation regression rate constant of each functional assessment unit.

[0009] In one embodiment of the present invention, the dynamic feedback module includes a data acquisition submodule, a deviation calculation submodule, and a parameter correction submodule. The data acquisition submodule acquires the body temperature distribution map, muscle tension waveform, and synovial fluid pressure change curve of each functional assessment unit every minute. The deviation calculation submodule performs a difference operation between the acquired data and the baseline data of the previous cycle to generate a residual matrix. The parameter correction submodule locates the abnormal functional assessment unit according to the position of the element with the largest absolute value in the residual matrix, and calls a preset correction algorithm to fine-tune the treatment parameter combination for the next cycle. The fine-tuning operation is limited to changing the physical therapy intensity coefficient or the number of repetitions of rehabilitation training movements corresponding to that unit.

[0010] In one embodiment of the present invention, the specific implementation steps of the biological agent intervention process are as follows: when the inflammation index of the sacroiliac joint region is measured three times consecutively above a set threshold, the intravenous drug delivery channel is activated, the initial infusion rate is set to 0.5 ml / min, and then the blood drug concentration is sampled every 15 minutes. The infusion rate is adjusted according to the sampling results until the blood drug concentration reaches a steady-state plateau. The criterion for determining the steady-state plateau is that the relative error between two consecutive sampling values ​​is less than 5%.

[0011] In one embodiment of the present invention, the method for calculating the ratio of the axial component to the shear force component of the applied traction force is as follows: First, the rate of change of the radius of curvature of the thoracic and lumbar vertebrae is extracted, and then substituted into a preset mechanical equilibrium equation to obtain the theoretical optimal ratio value. Then, the ratio is corrected by combining the patient's weight coefficient, and finally the actual applied traction force vector direction and magnitude are output. The traction force vector direction is deflected at a certain angle along the longitudinal axis of the spine, and the angle value is determined by the relative displacement of the thoracic and lumbar vertebrae.

[0012] In one embodiment of the present invention, the muscle tension waveform is acquired by a surface electromyography sensor array, the sensor array being arranged longitudinally along the paravertebral muscle group, with a sampling frequency of not less than 1000 Hz, and the joint fluid pressure change curve is acquired by an implantable micro pressure sensor, the implantable micro pressure sensor being fixed at a specific anatomical position within the joint cavity, and its signal transmission path being wirelessly encoded and modulated to eliminate motion artifact interference.

[0013] In one embodiment of the present invention, the termination conditions of the closed-loop control path include: the inflammation index of all functional assessment units is lower than the set threshold for seven consecutive days and the activity limit value recovers to more than the percentage of the normal range, or the patient voluntarily applies to stop treatment, or a serious adverse reaction event occurs; the criteria for judging the serious adverse reaction event are that the heart rate fluctuation amplitude exceeds 20% of the baseline value or the blood pressure drops by more than 20 mmHg in the vital signs monitoring data.

[0014] The beneficial effects of the technical solution of this invention are as follows: This invention achieves a digital reconstruction of the pathological space of ankylosing spondylitis through the linkage of an anatomical region division model and functional assessment unit construction units. The sacroiliac joint, thoracic vertebrae, lumbar vertebrae, cervical vertebrae, and peripheral joints are defined as independent three-dimensional volume units. This independent division based on anatomical coordinates ensures that subsequently acquired local inflammatory index values ​​correspond to specific anatomical structures, avoiding mutual interference between inflammatory signals from different segments. The generation process of the multidimensional feature vector set normalizes and concatenates the mean signal intensity of T2-weighted images from magnetic resonance imaging, blood flow grading values ​​from ultrasound examinations, and the local concentration ratio of C-reactive protein from blood tests with differences in flexion angle, lateral bending angle, extension angle, and rotation angle. The resulting n-dimensional vector completely records the physiological state of each functional assessment unit, providing high-fidelity input data for the regional treatment rule base and improving the scientific rigor and repeatability of regional treatment.

[0015] This invention constructs an adaptive treatment decision-making closed loop through the collaborative work of three-level matching logic and data acquisition submodules, deviation calculation submodule, and parameter correction submodule in the partial treatment rule base. The three-level logical judgments are connected through a multi-dimensional feature vector set to form a comprehensive regulatory network from inflammation control to mechanical balance and neuromuscular coordination. After the deviation calculation submodule generates the residual matrix and locates the abnormal function assessment unit, the parameter correction submodule calls a preset correction algorithm to change only the physical therapy intensity coefficient or the number of repetitions of rehabilitation training movements corresponding to the unit, ensuring the targetedness and safety of the treatment plan.

[0016] This invention utilizes the data acquisition capabilities of a surface electromyography (EMG) sensor array and an implantable miniature pressure sensor. Through wireless coding modulation processing and differential computation technology, it achieves lossless transmission and high-precision analysis of physiological signals during treatment. The implantable miniature pressure sensor is fixed at a specific anatomical location within the joint cavity, and its signal transmission path undergoes wireless coding modulation processing to effectively eliminate motion artifact interference and ensure the authenticity of the synovial fluid pressure change curve. The data acquisition submodule collects body temperature distribution maps, muscle tension waveforms, and synovial fluid pressure change curves every minute. The data is differentially processed with the baseline data from the previous cycle to generate a residual matrix reflecting changes in the patient's condition. The location of the element with the largest absolute value in the residual matrix allows the system to quickly identify the most unstable functional assessment unit in the current treatment and trigger the parameter correction submodule for targeted intervention.

[0017] This invention achieves personalized drug intervention by linking the drug dosage gradient calculated using a piecewise linear function with the inflammation regression rate constant of each functional assessment unit. The inflammation regression rate constant is determined by the slope of the local inflammatory index value of each functional assessment unit over time, reflecting the sensitivity of the diseased tissue to the drug. The slope of the piecewise linear function is dynamically adjusted according to the constant. When inflammation subsides rapidly, the slope increases to accelerate the increase of drug dosage and maintain therapeutic efficacy; when inflammation subsides slowly, the slope decreases to prevent excessive drug accumulation and ensure that the blood drug concentration is maintained within the effective therapeutic window.

[0018] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0019] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a flowchart outlining the steps for implementing the site-specific treatment method for ankylosing spondylitis proposed in this invention. Detailed Implementation

[0020] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0021] The following describes, with reference to the accompanying drawings, a method for implementing site-specific treatment of ankylosing spondylitis according to an embodiment of the present invention.

[0022] like Figure 1 As shown, this embodiment of the invention provides a method for implementing site-specific treatment for ankylosing spondylitis, including the following steps: Step 1: Obtain multimodal imaging data and clinical sign data of patients, and construct an anatomical region division model including the sacroiliac joint, thoracic vertebrae, lumbar vertebrae, cervical vertebrae and peripheral joints; Step 2: Based on the anatomical region division model, the patient's spine is divided into at least five independent functional assessment units, each corresponding to a specific anatomical coordinate range. Step 3: Collect local inflammation index values ​​and activity limit values ​​of each functional assessment unit to generate a multidimensional feature vector set; Step 4: Using a pre-defined partial treatment rule library, match the corresponding combination of treatment parameters based on the multi-dimensional feature vector set. The combination of treatment parameters includes physical therapy intensity, drug dosage gradient, and rehabilitation training movement sequence. Step 5: Monitor the response data of each functional evaluation unit in real time during the treatment process through the dynamic feedback module. If the response data deviates from the preset threshold range, automatically adjust the value of at least one dimension in the treatment parameter combination of the next cycle to form a closed-loop control path.

[0023] In specific applications, the model of this invention uses patient multimodal imaging data and clinical sign data as basic inputs to establish the specific coordinate range of each functional assessment unit in three-dimensional space, so that each unit corresponds to a specific anatomical structure position. Based on the anatomical region division model, at least five independent functional assessment units are generated, providing clear physical boundaries for subsequent data acquisition, ensuring that the local inflammatory index value and activity limit value of each unit can be collected independently and without interference.

[0024] Furthermore, in the data acquisition phase, the system extracts local inflammatory marker values ​​and range of motion limitation values ​​for each functional assessment unit, and integrates the discrete data to generate a multidimensional feature vector set. This standardizes and concatenates the originally scattered imaging signals, ultrasound blood flow grading values, and blood concentration ratios with the differences in flexion, lateral bending, extension, and rotation angles of joint range of motion. The formation of the multidimensional feature vector set provides a quantitative basis for subsequent rule matching, enabling pathological states of different sites to enter the processing flow in a unified data format. A pre-set treatment rule library stores treatment strategies for different site combinations. Its internal logic automatically selects the corresponding physical therapy intensity, drug dosage gradient, and rehabilitation training sequence based on the numerical distribution in the multidimensional feature vector set.

[0025] Specifically, the output of the treatment parameter combination drives the actuator. The physical therapy intensity is dynamically set based on the degree of local inflammation, the drug dosage gradient is calculated in segments according to the rate of inflammation resolution, and the rehabilitation training sequence is customized based on activity level limitations. During execution, the dynamic feedback module continuously monitors the response data of each functional assessment unit, including real-time information on body temperature distribution, muscle tension waveforms, and joint fluid pressure changes. If the monitored response data deviates from the preset threshold range, the system immediately initiates a correction procedure, automatically adjusting the value of at least one dimension in the treatment parameter combination for the next cycle. Based on the difference calculation results between the current cycle and the previous cycle, the abnormal unit is located, and a specific algorithm is invoked to fine-tune the physical therapy intensity coefficient or the number of repetitions of the rehabilitation training movements.

[0026] Specifically, the dynamic feedback module acquires the physiological state changes of each functional assessment unit through high-frequency sampling, transforming continuous time series data into discrete deviation signals. The deviation calculation submodule extracts the trend of change using differential operations, while the parameter correction submodule determines the adjustment range based on the trend direction, ensuring that the treatment parameters are always within the effective range. The response data of each functional assessment unit is not only used for parameter adjustment in the current cycle, but also serves as the input condition for rule matching in the next cycle, forming a continuous data flow transmission chain.

[0027] In one specific implementation, the anatomical region division model adopts a three-dimensional spatial coordinate system, wherein the sacroiliac joint region is defined as a three-dimensional volume unit at the junction of the S1 to S3 vertebral bodies and the iliac wing; the thoracic vertebral region is defined as a layered volume unit of the T1 to T12 vertebral bodies and their associated ribs; the lumbar vertebral region is defined as a volume unit of the L1 to L5 vertebral bodies and their associated psoas major muscle attachment points; the cervical vertebral region is defined as a volume unit of the C1 to C7 vertebral bodies and their associated atlantoaxial joint structures; and the peripheral joint region is defined as an intracapsular space volume unit of the shoulder, hip, knee, and ankle joints.

[0028] In practical applications, the anatomical region division model of this invention is constructed using a three-dimensional spatial coordinate system. By defining volumetric units of different anatomical structures, a spatial benchmark for segmented treatment of ankylosing spondylitis is established. The sacroiliac joint region is defined as a three-dimensional volumetric unit at the junction of the S1 to S3 vertebral bodies and the iliac wing, locking the inflammatory response of the sacroiliac joint within a specific skeletal interface. The thoracic spine region is defined as a layered volumetric unit of the T1 to T12 vertebral bodies and their associated ribs. This layering describes the spatial nesting structure of the ribs and vertebral bodies, allowing treatment parameters to cover areas with limited overall thoracic mobility. The lumbar spine region is defined as a volumetric unit of the L1 to L5 vertebral bodies and their associated psoas major muscle attachment points. The psoas major muscle attachment points, as the interface between muscle and bone, are included in the volumetric unit to ensure that treatment for limited lumbar mobility can simultaneously act on both the skeletal structure and the soft tissue attachment points. The cervical spine region is defined as a volumetric unit of the C1 to C7 vertebral bodies and their associated atlantoaxial joint structure. The atlantoaxial joint, as the core pivot of cervical spine rotation, is included in the definition to address the blind spots in the assessment of high cervical spine mobility impairments.

[0029] Furthermore, the establishment of three-dimensional volumetric units ensures that each anatomical region has clearly defined length, width, and height dimensions in space, providing a fixed sampling container for subsequent acquisition of local inflammatory marker values. When the system performs data acquisition, the sensor array scans according to the coordinates of the aforementioned volumetric units, ensuring that the acquired data originates solely from within the target anatomical structure and avoiding the intrusion of signals from adjacent healthy tissues. The three-dimensional volumetric units at the junction of the S1 to S3 vertebral bodies and the iliac wing exist independently in space, isolated from volumetric units in other regions by preset geometric cross-sections, ensuring the purity of sacroiliac joint inflammatory data.

[0030] Furthermore, the volumetric units of the L1 to L5 vertebrae and their associated psoas major muscle attachment points spatially cover the main stress areas of the lumbar spine. The introduction of the psoas major muscle attachment points allows the adjustment of treatment parameters to take into account the impact of muscle traction on spinal stability. The volumetric units of the C1 to C7 vertebrae and their associated atlantoaxial joint structures present a spiral upward shape in space, which is used to match the movement trajectory of neck rotation training.

[0031] Based on the volume units defined above, the system can map the patient's multimodal imaging data into a specific three-dimensional spatial grid. Each grid unit corresponds to a functional assessment unit. The acquisition process of local inflammatory index values ​​depends on the boundary constraints of these volume units to ensure that the detection signal does not cross the anatomical boundary. The measurement of the range of motion limit value is also based on the spatial range of the volume unit. The angle change is calculated by tracking the displacement trajectory of specific anatomical landmarks in three-dimensional space. The generation process of the multidimensional feature vector set transforms these discrete spatial data into a computable numerical sequence, where each value corresponds to a specific anatomical volume unit. The matching logic in the sectional treatment rule base filters according to the coordinate information of these volume units. For example, when a high inflammatory signal is detected in a volume unit in the sacroiliac joint region, the rule base automatically triggers a combination of treatment parameters for the volume unit.

[0032] The dynamic feedback module monitors the response data of each functional assessment unit in real time. The data is mapped back to the original three-dimensional volume unit for comparison. If the response data deviates from the preset threshold range, the system locates the abnormal area based on the coordinate information of the volume unit and adjusts the treatment parameter combination for the next cycle. The three-dimensional volume unit at the connection between the S1 to S3 vertebral body and the iliac wing maintains an independent state record during treatment. Its inflammation resolution rate does not interfere with the volume units of other areas. The stacked volume units of the T1 to T12 vertebral bodies and their attached ribs are used as the overall force-bearing objects in traction therapy. Their stacked structure determines the force transmission path. The volume units of the L1 to L5 vertebral bodies and their attached psoas major muscle attachment points are used as the core stability area in rehabilitation training. The state of their muscle attachment points directly affects the recovery of lumbar spine mobility.

[0033] In one specific implementation, the local inflammatory index values ​​include the mean signal intensity of T2-weighted images in magnetic resonance imaging, the blood flow grading value in ultrasound examination, and the local concentration ratio of C-reactive protein in blood tests; the range of motion limitation values ​​include the difference in forward flexion angle, the difference in lateral bending angle, the difference in backward extension angle, and the difference in rotation angle; the multidimensional feature vector set is generated by normalizing the above local inflammatory index values ​​and range of motion limitation values ​​according to preset weight coefficients and then concatenating them into an n-dimensional vector, where n is the sum of the number of local inflammatory indexes and the number of range of motion limitations.

[0034] In practical applications, the collection of local inflammatory marker values ​​in this invention includes the mean signal intensity of T2-weighted images from magnetic resonance imaging (MRI), blood flow grading values ​​from ultrasound examinations, and the local concentration ratio of C-reactive protein (CRP) from blood tests. These three data sources correspond to the degree of tissue edema, microcirculation perfusion status, and the biochemical level of systemic inflammatory response, respectively. The mean signal intensity of T2-weighted images obtained from MRI directly reflects changes in the water content of the vertebral bone marrow and periarticular soft tissues. High-signal areas identify zones of active acute inflammation. The values ​​are calculated through pixel grayscale statistics and serve as the core basis for quantifying the inflammatory burden. The blood flow grading values ​​from ultrasound examinations utilize color Doppler technology to detect the hemodynamic characteristics of the sacroiliac joint and paravertebral muscles. The blood flow signal grading increases progressively from 0 to 4, with higher values ​​indicating more significant vascular proliferation. The local concentration ratio of CRP from blood tests is calculated by comparing the concentration differences between the patient's peripheral blood and the synovial fluid or tissue fluid at the lesion site. This ratio eliminates the interference of individual basal metabolic levels and more accurately reflects the activity level of local inflammation.

[0035] Furthermore, the extraction of range of motion limitation values ​​includes flexion angle difference, lateral bending angle difference, extension angle difference, and rotation angle difference. These four dimensions comprehensively cover the kinematic functional status of the spine and peripheral joints. The flexion angle difference is obtained by measuring the angle difference between the patient's maximum flexion position and the normal anatomical neutral position, reflecting the degree of flexion restriction of the lumbar and thoracic spine. The lateral bending angle difference records the maximum angle deviation during left and right lateral bending, revealing the stiffness of lateral spinal movement. The extension angle difference measures the angle loss during backward extension, assessing the functional status of the intervertebral joint spaces and posterior facet joints. The rotation angle difference quantifies the axial rotational capacity of the cervical and thoracic spine, reflecting the preservation of spinal torsional function.

[0036] Furthermore, the generation process of the multidimensional feature vector set involves normalizing the aforementioned local inflammatory index values ​​and activity limit values ​​according to preset weight coefficients, and then concatenating them to form an n-dimensional vector, where n is the sum of the number of local inflammatory indicators and the number of activity limits. The normalization process uses a min-max standardization method to map data of different dimensions to the interval between 0 and 1, eliminating the influence of unit differences on subsequent matching algorithms. The preset weight coefficients are determined based on the contribution of each indicator to disease progression. For example, the mean signal intensity of T2-weighted images is given a higher weight in the acute phase, while the local concentration ratio of C-reactive protein is adjusted accordingly in the chronic phase, so that the feature vector can adapt to the needs of different disease stages. The concatenation operation arranges the normalized values ​​in a specific order to form a fixed vector structure. Each dimension of data corresponds to a specific physiological parameter, and the length n of the vector is determined by the total number of indicators involved in the calculation, ensuring the completeness of the feature space.

[0037] Specifically, the generated multidimensional feature vector set serves as the sole input for subsequent rule matching. The correlation between the data within the vector is preserved, and there is an implicit causal relationship between adjacent inflammatory indicators and range of motion limits. For example, high T2 signal intensity is usually accompanied by a large difference in flexion angle, which manifests as a specific geometric distribution pattern in the vector space. The pre-defined partial treatment rule base searches based on this distribution pattern. When several dimensions of the vector simultaneously exceed a threshold, the rule base automatically identifies the corresponding pathological combination type. For instance, when the mean T2 signal intensity and blood flow grade values ​​are both high, and the difference in flexion angle exceeds a critical value, the system determines it to be an acute attack of the sacroiliac joint and triggers the biological agent intervention process.

[0038] In the closed-loop control path, the multidimensional feature vector set is not only used to formulate the initial treatment plan, but also serves as a benchmark for dynamic feedback adjustment. After each treatment cycle, the system re-collects new local inflammatory index values ​​and activity limit values, generates a new multidimensional feature vector set, and performs a difference operation with the vector from the previous cycle. The difference result reflects the trend of change after treatment intervention. If the values ​​of some dimensions in the new vector move in the expected direction, it indicates that the current treatment parameters are effective; if the values ​​of some dimensions fluctuate in the opposite direction, it suggests that the intensity of physical therapy or the drug dosage gradient needs to be adjusted.

[0039] In one specific implementation, the segmented treatment rule base contains at least three levels of matching logic: the first level determines whether to initiate the biological agent intervention process based on whether the inflammation index of the sacroiliac joint region exceeds a preset threshold; the second level determines the ratio of the axial component to the shear force component of the traction force applied in physical therapy based on the ratio of the range of motion limits of the thoracic and lumbar vertebrae; the third level determines the alternation frequency of neck stability training and lower limb resistance training in the rehabilitation training movement sequence based on the coupling degree of the response data of the cervical vertebrae and peripheral joints; the drug dosage gradient in the treatment parameter combination is calculated through a piecewise linear function, the slope of which is determined by the inflammation regression rate constant of each functional assessment unit.

[0040] In practical applications, the present invention utilizes at least three levels of matching logic within the distributed treatment rule base to form a computational framework for treatment decisions. The first level of matching logic correlates the real-time comparison results of the inflammation index of the sacroiliac joint region with a preset threshold value. When the system reads that the inflammation index value of the sacroiliac joint region is higher than the set threshold boundary, the matching engine immediately activates the trigger signal for the biological agent intervention process. The signal is sent directly to the drug administration execution unit without manual confirmation. If the inflammation index value is lower than or equal to the threshold value, the biological agent intervention path is skipped, and the system transfers control to subsequent level logic for calculating conventional physical therapy parameters.

[0041] Furthermore, the second-level matching logic focuses on calculating the proportional relationship between the range of motion limits of the thoracic and lumbar spine segments. The system first extracts the differences in flexion, lateral bending, and extension angles of the thoracic spine segment, and simultaneously extracts the corresponding angle differences of the lumbar spine segment, constructing a ratio calculation model between the two. When the ratio of the range of motion limits of the thoracic and lumbar spine segments is within a specific range, the ratio of the axial component to the shear force component of the applied traction force is dynamically adjusted accordingly. If the stiffness of the thoracic spine is greater than that of the lumbar spine, the proportion of the axial tensile component in the traction force vector is increased to prioritize releasing the adhesions between the rib cage and the vertebral bodies; conversely, if the lumbar spine mobility is more severely restricted, the proportion of the shear force component is increased, focusing on releasing the attachment points of the psoas major muscle and the contractures of the facet joints.

[0042] Furthermore, the third-level matching logic is based on the coupling degree analysis of cervical spine and peripheral joint response data. The system continuously monitors the muscle tension change curve of the cervical spine during rehabilitation training, as well as the synovial fluid pressure change data of the shoulder, hip, knee, and ankle peripheral joints, calculating the correlation coefficient between the two over time as a coupling degree index. When the coupling degree value indicates a high synergistic effect between cervical spine stability and lower limb resistance ability, the rehabilitation training movement sequence automatically sets a high alternation frequency, allowing for rapid switching between neck stability training and lower limb resistance training within a short period, utilizing the principle of neuromuscular reciprocal inhibition to improve overall coordination. If the coupling degree value is low, indicating asynchronous upper and lower limb motor control, the system reduces the alternation frequency and prolongs the duration of a single training movement until the coupling degree value returns to the normal range.

[0043] Specifically, the inflammation regression rate constant is obtained by differential calculation of local inflammatory index values ​​over multiple consecutive cycles, reflecting the sensitivity and metabolic rate of the diseased tissue to the drug. When the inflammation regression rate constant of a certain functional assessment unit is large, the slope of the piecewise linear function increases accordingly, meaning that the increase in drug dose per unit time is accelerated to maintain an effective plateau in blood drug concentration. Conversely, when inflammation subsides slowly, the slope decreases to prevent excessive drug accumulation and adverse reactions. The node positions of the piecewise linear function are preset according to clinical pharmacokinetic parameters to ensure the continuity of drug dose adjustment and avoid fluctuations in blood drug concentration caused by dose mutations.

[0044] In one specific implementation, the dynamic feedback module includes a data acquisition submodule, a deviation calculation submodule, and a parameter correction submodule. The data acquisition submodule acquires body temperature distribution maps, muscle tension waveforms, and synovial fluid pressure change curves for each functional assessment unit every minute. The deviation calculation submodule performs a difference operation between the acquired data and the baseline data of the previous cycle to generate a residual matrix. The parameter correction submodule locates the abnormal functional assessment unit based on the position of the element with the largest absolute value in the residual matrix and calls a preset correction algorithm to fine-tune the treatment parameter combination for the next cycle. The fine-tuning operation is limited to changing the physical therapy intensity coefficient or the number of repetitions of rehabilitation training movements corresponding to the unit. The muscle tension waveform is acquired through a surface electromyography sensor array, which is arranged longitudinally along the paraspinal muscle group with a sampling frequency of not less than 1000Hz. The synovial fluid pressure change curve is acquired through an implantable miniature pressure sensor, which is fixed in a specific anatomical position within the joint cavity. Its signal transmission path is processed by wireless coding modulation to eliminate motion artifact interference.

[0045] In practical applications, the dynamic feedback module of this invention consists of a data acquisition submodule, a deviation calculation submodule, and a parameter correction submodule connected in series to form a closed-loop control link. These three modules work collaboratively to maintain the stability of the treatment process. The data acquisition submodule performs real-time monitoring tasks every minute, simultaneously acquiring three types of physiological signals for each functional assessment unit: body temperature distribution map, muscle tension waveform, and synovial fluid pressure change curve. The body temperature distribution map is generated using infrared thermal imaging technology, reflecting the microcirculation congestion caused by local inflammation. The image data is divided into several grid units and mapped to the spatial coordinates of the anatomical region division model. The muscle tension waveform is acquired using a surface electromyography (SEM) sensor array, arranged longitudinally along the paravertebral muscle groups, covering the three main segments of the neck, chest, and lumbar region. The sampling frequency is set to no less than 1000Hz to ensure the capture of high-frequency muscle fiber tremors and low-frequency fatigue characteristics. The synovial fluid pressure change curve is acquired using an implanted miniature pressure sensor. The sensor is fixed at a specific anatomical location within the sacroiliac joint, knee joint, or shoulder joint cavity to directly sense intra-articular hydraulic fluctuations. Wireless coding and modulation processing is applied throughout the entire signal transmission process. It converts the analog signals collected by the sensor into digital coded sequences and uses encryption algorithms to eliminate motion artifact interference, ensuring the integrity and reliability of the data when the patient is performing rehabilitation training movements.

[0046] Furthermore, the deviation calculation submodule receives the baseline data stored in the previous cycle and the real-time data collected in the current cycle, and performs differential operations on the three types of signals respectively. The difference calculation of the body temperature distribution map is based on the point-by-point subtraction of pixel gray values ​​to generate a two-dimensional matrix reflecting the temperature change gradient; the differential operation of the muscle tension waveform adopts the sliding window algorithm to extract the amplitude difference between adjacent time windows and form a one-dimensional residual sequence; the differential processing of the synovial fluid pressure change curve focuses on the offset between the pressure peak and valley values ​​to generate a scalar sequence reflecting the changes in the mechanical state inside the joint, so that the system can focus on the real physiological response changes generated after treatment intervention. The location logic of the maximum absolute value element is the core of deviation calculation. The system traverses all values ​​of the residual matrix, identifies the element with the largest absolute value and its corresponding row and column index, and the index is directly mapped back to the coordinates of the specific functional assessment unit.

[0047] Furthermore, the parameter correction submodule locates the abnormal function assessment unit based on the position of the element with the largest absolute value in the residual matrix, and initiates a preset correction algorithm to fine-tune the treatment parameter combination for the next cycle. This fine-tuning is limited to two dimensions: the physical therapy intensity coefficient or the number of repetitions of rehabilitation training movements, avoiding simultaneous adjustment of multiple variables that could lead to uncontrollable system states. If an abnormal unit displays severe oscillations in its muscle tension waveform and the residual value exceeds a safe threshold, the parameter correction submodule automatically reduces the physical therapy intensity coefficient corresponding to the unit, decreasing traction or massage intensity to prevent secondary soft tissue injury. If an abnormal unit exhibits persistently elevated synovial fluid pressure and limited mobility, the system increases the number of repetitions of rehabilitation training movements. Low-intensity repetitive movements promote synovial fluid circulation and metabolic waste removal. The correction algorithm includes proportional-integral control logic, linearly adjusting the parameter variation based on the residual value to ensure a smooth transition and prevent sudden parameter changes from triggering new physiological stress responses.

[0048] Specifically, the longitudinal arrangement of the surface electromyography (EMG) sensor array closely matches the anatomical structure of the spine, with each sensor node corresponding to a specific vertebral segment. This allows for clear geographical interpretation of the muscle tension waveform acquisition. The sampling frequency of over 1000Hz not only meets the bandwidth requirements of EMG signals but also allows the system to distinguish between active contraction and passive traction states through frequency domain analysis, improving diagnostic reliability. The wireless coding modulation technology of the implantable miniature pressure sensor solves the problem of interference in signal transmission within the body, while the motion artifact elimination algorithm uses adaptive filtering technology to remove high-frequency noise generated by limb movement, ensuring that the synovial fluid pressure change curve accurately reflects the pathological state.

[0049] In one specific implementation, the specific steps of the biological agent intervention process are as follows: when the inflammation index of the sacroiliac joint region is measured three times consecutively above the set threshold, the intravenous drug delivery channel is activated, the initial infusion rate is set to 0.5 ml / min, and then the blood drug concentration is sampled every 15 minutes. The infusion rate is adjusted according to the sampling results until the blood drug concentration reaches the steady-state plateau. The criterion for determining the steady-state plateau is that the relative error between two consecutive sampling values ​​is less than 5%.

[0050] The method for calculating the ratio of the axial component to the shear force component of the applied traction force is as follows: First, extract the rate of change of the radius of curvature of the thoracic and lumbar vertebrae, substitute it into the preset mechanical equilibrium equation, solve to obtain the theoretical optimal ratio value, and then correct it in combination with the patient's weight coefficient, and finally output the actual applied traction force vector direction and magnitude; the traction force vector direction is deflected at a certain angle along the longitudinal axis of the spine, and the angle value is determined by the relative displacement of the thoracic and lumbar vertebrae.

[0051] In practical applications of this invention, the initiation logic of the biological agent intervention process is based on continuous monitoring data of the inflammation index in the sacroiliac joint region. The system performs three consecutive measurements of the inflammation index and compares each measurement with a set threshold. When all three measurements show values ​​higher than a preset critical value, an acute inflammatory outbreak is identified, and the system immediately activates the intravenous drug delivery channel, cutting off the conventional physical therapy pathway and concentrating resources on drug infusion. The initial infusion rate is locked at a baseline value of 0.5 ml / min, calculated based on the initial loading amount in the pharmacokinetic model. Subsequently, the system enters a dynamic adjustment phase, performing blood drug concentration sampling at fixed intervals of 15 minutes. The sampling points are located in the peripheral venous circuit, and sample data is read in real time using a portable biochemical analyzer. The sampled blood drug concentration values ​​are compared with the target steady-state plateau reference value. If the current concentration is lower than the target value, the infusion rate is increased proportionally; if it is higher than the target value, the infusion rate is decreased until the steady-state plateau criteria are met, i.e., the relative error between two consecutive sampling values ​​is less than 5%. When the relative error falls within the range, the system confirms that the blood drug concentration has entered a stable fluctuation range. At this time, the current infusion rate is maintained to ensure that the drug maintains a constant therapeutic level in the body and avoids toxic side effects or insufficient efficacy caused by excessive concentration fluctuations.

[0052] Furthermore, determining the ratio of the axial component to the shear force component of the applied traction force relies on the analysis of the rate of change of the radius of curvature of the thoracic and lumbar vertebrae. The system first extracts spatial coordinate data from the T1-T12 vertebral sequence of the thoracic spine and the L1-L5 vertebral sequence of the lumbar spine from a three-dimensional coordinate system. It then calculates the rate of change of the radius of curvature of each segment during motion. This rate of change reflects the degree of curvature and the rate of change of the spine at different segments, and is a key geometric parameter for assessing spinal stiffness. The ratio represents the optimal distribution of axial tensile and shear forces that maximizes adhesion release and minimizes the risk of injury under the current spinal morphology. Subsequently, a patient weight coefficient is introduced to correct the theoretically optimal ratio. This weight coefficient is calculated based on the ratio of the patient's actual weight to their standard weight, used to compensate for differences in traction force requirements among patients of different body types. The corrected value is ultimately converted into the actual applied traction force vector direction and magnitude, ensuring that the traction force meets both mechanical balance requirements and adapts to the individual physiological characteristics of the patient.

[0053] Specifically, the traction force vector has a clear spatial orientation. Its deflection angle along the longitudinal axis of the spine is determined by the relative displacement of the thoracic and lumbar vertebrae. The system monitors the relative displacement of the thoracic and lumbar vertebrae in the vertical direction in real time. When there is a significant difference in displacement, it indicates obvious segmental misalignment or scoliosis of the spine. In this case, the deflection angle of the traction force vector is increased accordingly to generate a targeted corrective torque. If the displacement of the thoracic and lumbar vertebrae tends to be consistent, it indicates that the overall rigidity of the spine is relatively strong or in a relatively straight state. The direction of the traction force vector is close to the longitudinal axis of the spine, mainly exerting a longitudinal stretching effect. The magnitude of the traction force also adjusts with the change in relative displacement. The greater the displacement, the greater the resistance to be overcome, and the system automatically increases the amplitude of the traction force to ensure the therapeutic effect.

[0054] In one specific implementation, the termination conditions of the closed-loop control path include: the inflammation index of all functional assessment units is below the set threshold for seven consecutive days and the activity limit value recovers to more than the percentage of the normal range, or the patient voluntarily requests to stop treatment, or a serious adverse reaction event occurs; the criteria for judging a serious adverse reaction event are that the heart rate fluctuation amplitude exceeds 20% of the baseline value or the blood pressure drops by more than 20 mmHg in the vital signs monitoring data.

[0055] In practical applications, the termination logic of the closed-loop control path in this invention is composed of multiple conditions. The system continuously monitors the inflammation index and range of motion limitation values ​​of all functional assessment units. When the indicators meet the condition of being below the set threshold for seven consecutive days and the range of motion limitation value recovers to more than a percentage of the normal range, the treatment goal is determined to be achieved, and the termination procedure is automatically triggered. The continuous monitoring of the inflammation index covers multiple areas including the sacroiliac joint, thoracic spine, lumbar spine, and cervical spine. The value of each area must simultaneously meet the condition of being below the threshold. Any rebound in any single area will cause the timer to reset, ensuring that the condition is completely stable. The degree of recovery of the range of motion limitation value is calculated by the ratio of the difference between the flexion, lateral bending, extension, and rotation angles to the standard anatomical values. The ratio must continuously exceed the preset normal range percentage, such as 85% or 90%, to confirm that the spinal motor function has substantially recovered. The seven-day time window is required to eliminate the risk of misjudgment caused by short-term fluctuations and ensure the durability and stability of the therapeutic effect. Only when the data for multiple days shows a benign trend will the system allow the treatment process to end.

[0056] Furthermore, the patient's initiative to discontinue treatment is considered another termination condition, granting the patient autonomy in the treatment process. The system has a built-in human-computer interface, allowing patients to submit a discontinuation command at any time via their terminal device. Once the command is issued, the system immediately pauses the current medication infusion and physical therapy procedures and initiates the exit process. Upon receiving the discontinuation command, the system records all parameter data for the current treatment phase and generates a phase summary report for doctors' subsequent evaluation, ensuring that treatment interruptions do not cause information gaps.

[0057] Furthermore, the determination of serious adverse events constitutes a third category of conditions for mandatory termination. This is based on real-time anomaly analysis of vital sign monitoring data. The system continuously collects heart rate and blood pressure data through wearable sensors or implantable devices, dynamically comparing them with the patient's baseline physiological values. When abnormal heart rate fluctuations exceed 20% of the baseline value, the system immediately identifies it as a cardiovascular instability signal; when a drop in blood pressure exceeds 20 mmHg, it is considered a precursor to circulatory failure. Both situations are defined as serious adverse events, triggering the system's emergency circuit breaker mechanism. Once the circuit breaker is triggered, the intravenous drug delivery channel is immediately closed, the traction application device releases its load, and the rehabilitation training robot stops moving, preventing further deterioration of the condition.

[0058] Specifically, the criteria for determining serious adverse events have undergone rigorous clinical validation. Thresholds of 20% heart rate fluctuation and 20 mmHg blood pressure drop effectively distinguish between physiological fluctuations and pathological crises. The system employs a moving average filtering algorithm when processing vital sign data to eliminate transient noise caused by exercise or respiration, ensuring the accuracy of alarm signals. When the aforementioned thresholds are exceeded, the system not only terminates the process but also simultaneously sends a high-priority alarm to the medical team, including the patient's current vital sign curve, medication records, and treatment parameters, enabling rapid intervention by healthcare professionals.

[0059] The termination conditions within the closed-loop control path are logically mutually exclusive and prioritized. The determination of serious adverse events has the highest priority, immediately overriding other termination conditions upon their occurrence. Patient-initiated cessation of treatment is secondary, taking effect only when no serious adverse events occur. The attainment of target inflammation indices and activity limits are routine termination conditions, requiring long-term data accumulation to trigger. This hierarchical termination logic ensures the safety of the treatment process, respecting patient wishes while strictly controlling medical risks and guaranteeing the scientific achievement of treatment goals. Subsequently, the system generates a detailed termination report, listing the reason for termination, final treatment parameters, and recommended follow-up plans. If the termination reason is disease target achievement, the report emphasizes the importance of maintenance treatment; if the termination reason is an adverse reaction, the report records the time and characteristics of the event in detail, providing a basis for adjusting subsequent treatment plans. After data is uploaded to the cloud database, it supports remote medical teams for access and analysis, achieving transparency and traceability of the treatment process.

[0060] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0061] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. A method for implementing site-specific treatment for ankylosing spondylitis, characterized in that, Includes the following steps: Step 1: Obtain multimodal imaging data and clinical sign data of patients, and construct an anatomical region division model including the sacroiliac joint, thoracic vertebrae, lumbar vertebrae, cervical vertebrae and peripheral joints; Step 2: Based on the anatomical region division model, the patient's spine is divided into at least five independent functional assessment units, each corresponding to a specific anatomical structure coordinate range; Step 3: Collect local inflammation index values ​​and activity limit values ​​of each functional assessment unit to generate a multidimensional feature vector set; Step 4: Using a pre-defined partial treatment rule library, match the corresponding combination of treatment parameters according to the multi-dimensional feature vector set. The combination of treatment parameters includes physical therapy intensity, drug dosage gradient, and rehabilitation training action sequence. Step 5: Monitor the response data of each functional evaluation unit in real time during the treatment process through the dynamic feedback module. If the response data deviates from the preset threshold range, automatically adjust the value of at least one dimension in the treatment parameter combination of the next cycle to form a closed-loop control path.

2. The method for implementing the site-specific treatment of ankylosing spondylitis according to claim 1, characterized in that, The anatomical region division model adopts a three-dimensional spatial coordinate system, wherein the sacroiliac joint region is defined as a three-dimensional volume unit at the junction of the S1 to S3 vertebral bodies and the iliac wing; the thoracic vertebral region is defined as a layered volume unit of the T1 to T12 vertebral bodies and their associated ribs; the lumbar vertebral region is defined as a volume unit of the L1 to L5 vertebral bodies and their associated psoas major muscle attachment points; the cervical vertebral region is defined as a volume unit of the C1 to C7 vertebral bodies and their associated atlantoaxial joint structures; and the peripheral joint region is defined as an intracapsular volume unit of the shoulder, hip, knee, and ankle joints.

3. The method for implementing the site-specific treatment of ankylosing spondylitis according to claim 1, characterized in that, The local inflammatory index values ​​include the mean signal intensity of T2-weighted images in magnetic resonance imaging, the blood flow grading value in ultrasound examination, and the local concentration ratio of C-reactive protein in blood tests; the range of motion limitation values ​​include the difference in forward flexion angle, the difference in lateral bending angle, the difference in backward extension angle, and the difference in rotation angle; the generation process of the multidimensional feature vector set is to normalize the above local inflammatory index values ​​and range of motion limitation values ​​according to preset weight coefficients and then concatenate them to form an n-dimensional vector, where n is equal to the sum of the number of local inflammatory indexes and the number of range of motion limitations.

4. The method for implementing the site-specific treatment of ankylosing spondylitis according to claim 1, characterized in that, The segmented treatment rule base contains at least three levels of matching logic: the first level determines whether to initiate the biological agent intervention process based on whether the inflammation index of the sacroiliac joint region exceeds a preset threshold. The second level determines the ratio of the axial component to the shear force component of the traction force applied during physical therapy based on the ratio of the range of motion limits of the thoracic and lumbar vertebrae. The third level determines the alternation frequency of neck stability training and lower limb resistance training in the rehabilitation training sequence based on the coupling degree of the response data of the cervical vertebrae and peripheral joints. The drug dosage gradient in the treatment parameter combination is calculated by a piecewise linear function, and its slope is determined by the inflammation resolution rate constant of each functional assessment unit.

5. The method for implementing the site-specific treatment of ankylosing spondylitis according to claim 1, characterized in that, The dynamic feedback module includes a data acquisition submodule, a deviation calculation submodule, and a parameter correction submodule. The data acquisition submodule collects body temperature distribution maps, muscle tension waveforms, and synovial fluid pressure change curves for each functional assessment unit every minute. The deviation calculation submodule performs a difference operation between the collected data and the baseline data of the previous cycle to generate a residual matrix. The parameter correction submodule locates the abnormal functional assessment unit based on the position of the element with the largest absolute value in the residual matrix and calls a preset correction algorithm to fine-tune the treatment parameter combination for the next cycle. The fine-tuning operation is limited to changing the physical therapy intensity coefficient or the number of repetitions of rehabilitation training movements corresponding to that unit.

6. The method for implementing the site-specific treatment of ankylosing spondylitis according to claim 4, characterized in that, The specific implementation steps of the biological agent intervention process are as follows: when the inflammation index of the sacroiliac joint region is measured three times consecutively above the set threshold, the intravenous drug delivery channel is activated, the initial infusion rate is set to 0.5 ml / min, and then the blood drug concentration is sampled every 15 minutes. The infusion rate is adjusted according to the sampling results until the blood drug concentration reaches the steady-state plateau period. The criterion for determining the steady-state plateau period is that the relative error between two consecutive sampling values ​​is less than 5%.

7. The method for implementing the site-specific treatment of ankylosing spondylitis according to claim 4, characterized in that, The method for calculating the ratio of the axial component to the shear force component of the applied traction force is as follows: First, the rate of change of the radius of curvature of the thoracic and lumbar vertebrae is extracted and substituted into the preset mechanical equilibrium equation to obtain the theoretical optimal ratio value. Then, the ratio is corrected by combining the patient's weight coefficient, and finally the actual applied traction force vector direction and magnitude are output. The traction force vector direction is deflected at a certain angle along the longitudinal axis of the spine, and the angle value is determined by the relative displacement of the thoracic and lumbar vertebrae.

8. The method for implementing the site-specific treatment of ankylosing spondylitis according to claim 5, characterized in that, The muscle tension waveform is acquired by a surface electromyography sensor array, which is arranged longitudinally along the paravertebral muscle group, with a sampling frequency of not less than 1000 Hz. The joint fluid pressure change curve is acquired by an implantable micro pressure sensor, which is fixed in a specific anatomical position within the joint cavity. Its signal transmission path is wirelessly encoded and modulated to eliminate motion artifact interference.

9. The method for implementing the site-specific treatment of ankylosing spondylitis according to claim 1, characterized in that, The termination conditions of the closed-loop control path include: the inflammation index of all functional assessment units is below the set threshold for seven consecutive days and the activity limit value recovers to more than the percentage of the normal range, or the patient voluntarily applies to stop treatment, or a serious adverse reaction event occurs; the criteria for judging the serious adverse reaction event are that the heart rate fluctuation exceeds 20% of the baseline value or the blood pressure drops by more than 20 mmHg in the vital signs monitoring data.

Images