Monitoring and early warning system and computer device after stroke revascularization

Abstract

This disclosure provides a monitoring and early warning system and computer device for stroke revascularization surgery; including: an acquisition module for acquiring a first set of neuroelectrophysiological data and a second set of neuroelectrophysiological data; a first processing module for obtaining a first impairment level characterizing the preoperative neurological function impairment based on the differences between the physiological data of the first affected side and the physiological data of the first healthy side at the same time point in the first set of neuroelectrophysiological data; a second processing module for obtaining a second impairment level characterizing the postoperative neurological function impairment based on the differences between the physiological data of the second affected side and the physiological data of the second healthy side at the same time point in the second set of neuroelectrophysiological data; and a prediction module for predicting the probability of various complications within a predicted postoperative time period. By comparing and analyzing the preoperative and postoperative neuroelectrophysiological data of the affected / healthy side, the risk of complications after revascularization surgery can be predicted.

Description

Technical Field

[0001] This disclosure relates to the field of medical and health monitoring technology, and in particular to a monitoring and early warning system and computer device after vascular recanalization following stroke. Background Technology

[0002] Currently, while revascularization surgery for acute ischemic stroke (such as mechanical thrombectomy) can effectively restore blood flow to large vessels, some patients, even with anatomical recanalization, still face serious complications such as intracranial hemorrhage, malignant cerebral edema, and reperfusion injury, leading to neurological deterioration or even death. Traditional postoperative assessment mainly relies on imaging examinations and clinical scores, which have limitations such as poor timeliness, strong subjectivity, and inability to reflect the brain's functional status in real time. Existing neurophysiological monitoring technologies mostly focus on single-point or single-modal signal analysis, lacking a systematic integration of preoperative baseline and postoperative dynamic changes, making it difficult to provide prospective early warning of complications. Summary of the Invention

[0003] In view of the shortcomings of the prior art described above, the purpose of this disclosure is to provide a monitoring and early warning system and computer device after vascular recanalization in stroke, so as to solve the problems in the related technology.

[0004] The first aspect of this disclosure provides a monitoring and early warning system after vascular recanalization following stroke, comprising:

[0005] The acquisition module is used to acquire the first set of neurophysiological data of the affected and unaffected sides of the patient's brain during a first preset time period before surgery, and the second set of neurophysiological data of the affected and unaffected sides of the brain during a second preset time period after vascular recanalization.

[0006] A first processing module, connected to the acquisition module, is used to obtain a first impairment level characterizing the degree of preoperative neurological function impairment in a patient based on the differences between the physiological data of the first affected side and the physiological data of the first healthy side at the same time point in the first neurophysiological data set; and...

[0007] The second processing module, connected to the acquisition module, is used to obtain a second impairment level that characterizes the degree of postoperative neurological function impairment in patients based on the difference between the second affected side physiological data and the second healthy side physiological data at the same time point in the second neurophysiological data set.

[0008] The prediction module is connected to the acquisition module, the first processing module, and the second processing module respectively; it uses a hierarchical pathological prediction model to predict the probability of each complication during the postoperative prediction period based on the first damage level, the second damage level, the first neuroelectrophysiological data set, and the second neuroelectrophysiological data set.

[0009] The stratified pathological prediction model includes an ischemic injury sub-model, a surgical immediate effectiveness sub-model, and an early warning sub-model.

[0010] The ischemic injury sub-model is used to obtain reversible values ​​characterizing the patient's neurophysiological function based on the first injury level and the second injury level.

[0011] The surgical immediate effectiveness sub-model is used to obtain an effective value characterizing the degree of improvement in neurological function by measuring the changes in the characteristics of the changes from the first neuroelectrophysiological data set to the second neuroelectrophysiological data set within the monitoring time range.

[0012] When the early warning sub-model is implemented as the first early warning sub-model, the first early warning sub-model predicts the prediction result based on the reversible value and the effective value.

[0013] In an embodiment of the first aspect, when the early warning sub-model is implemented as a second early warning sub-model, the acquisition module is further configured to collect real-time neurophysiological data after surgery; the second early warning sub-model predicts the prediction result based on the reversible value, the effective value, and the real-time neurophysiological data.

[0014] In an embodiment of the first aspect, the affected side is the neurophysiological monitoring side corresponding to the cerebral hemisphere on the side where the patient has suffered a stroke; and / or, the unaffected side is the neurophysiological monitoring side corresponding to the contralateral cerebral hemisphere on the side where the patient has not suffered a stroke.

[0015] In the embodiments of the first aspect, the complications include at least one or more of the following: intracranial hemorrhage, malignant cerebral edema, ischemia-reperfusion injury, neurological deterioration, and vascular re-occlusion.

[0016] In an embodiment of the first aspect, the first neurophysiological data in the first neurophysiological data set and / or the second neurophysiological data in the second neurophysiological data set include at least one or more of the following: somatosensory evoked potentials, transcranial motor evoked potentials, and electroencephalograms.

[0017] In the first aspect of the embodiment, a third processing module is further included, connected between the acquisition module and the surgical immediate effectiveness sub-model, for processing the somatosensory evoked potentials, transcranial motor evoked potentials and electroencephalograms to obtain the change characteristics including the patient's rate of change of neural wave amplitude, change of neural signal elicitation state, change of brain rhythm and change of bilateral symmetry, and outputting them to the surgical immediate effectiveness sub-model.

[0018] In an embodiment of the first aspect, the variation features include at least one of the following:

[0019] (1) The rate of change in amplitude between the physiological data of the first affected side and the physiological data of the second affected side;

[0020] (2) Changes in the extraction state of postoperative neurophysiological signals on the affected side based on physiological data from the second affected side;

[0021] (3) Postoperative changes in EEG rhythm of the affected side relative to the healthy side, based on physiological data of the second affected side and physiological data of the second healthy side;

[0022] (4) Changes in bilateral symmetry between the affected and healthy sides after surgery, based on physiological data of the second affected side and physiological data of the second healthy side.

[0023] In an embodiment of the first aspect, an alarm module is further included, connected to the prediction module, for generating a corresponding alarm level based on the prediction result and sending it to external personnel.

[0024] In an embodiment of the first aspect, the acquisition module includes multiple neurophysiological sensors, respectively deployed in the monitoring areas corresponding to the affected and healthy sides of the brain, for collecting the first neurophysiological data set and / or the second neurophysiological data set.

[0025] A second aspect of this disclosure provides a computer apparatus, comprising:

[0026] Processor and memory;

[0027] The memory stores program instructions;

[0028] The processor is configured to execute the program instructions to run the monitoring and early warning system described in any of the preceding embodiments.

[0029] The beneficial effects of this disclosure are: by comparing and analyzing the neurophysiological data of the affected / healthy side before and after the operation, combined with the hierarchical pathological prediction model, it is possible to predict the risk of complications after vascular recanalization in an early, objective, and personalized manner. Attached Figure Description

[0030] Figure 1 A schematic diagram of the structure of a monitoring and early warning system for vascular recanalization in stroke is shown in one embodiment of this disclosure.

[0031] Figure 2 A schematic diagram of the structure of the early warning submodule in a monitoring and early warning system according to an embodiment of the present disclosure is shown.

[0032] Figure 3 A schematic diagram of the structure of a monitoring and early warning system including a third processing module is shown in one embodiment of the present disclosure.

[0033] Figure 4 A schematic diagram of the structure of a monitoring and early warning system including a third processing module is shown in another embodiment of the present disclosure.

[0034] Figure 5 A schematic diagram of the structure of a monitoring and early warning system including a third processing module is shown in another embodiment of the present disclosure.

[0035] Figure 6 This illustration shows a structural diagram of a monitoring and early warning system including an alarm module, according to an embodiment of the present disclosure.

[0036] Figure 7 A schematic diagram of the structure of a computer device according to an embodiment of the present disclosure is shown. Detailed Implementation

[0037] The following specific examples illustrate the implementation of this disclosure. Those skilled in the art can easily understand other advantages and effects of this disclosure from the information disclosed herein. This disclosure can also be implemented or applied through other different specific embodiments, and various details in this disclosure can be modified or changed according to different viewpoints and application modules without departing from the spirit of this disclosure. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of this disclosure can be combined with each other.

[0038] The embodiments of this disclosure will now be described in detail with reference to the accompanying drawings, so that those skilled in the art to which this disclosure pertains can readily implement it. This disclosure may be embodied in many different forms and is not limited to the embodiments described herein.

[0039] In this disclosure, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic represented in connection with that embodiment or example is included in at least one embodiment or example of this disclosure. Furthermore, the specific features, structures, materials, or characteristics represented may be combined in any suitable manner in any one or a group of embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples represented in this disclosure, as well as the features of those different embodiments or examples.

[0040] Furthermore, the terms "first" and "second" are used for illustrative purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the representation of this disclosure, "a set" means two or more, unless otherwise explicitly specified.

[0041] For the purpose of clarity, devices unrelated to the description are omitted, and the same or similar components throughout the specification are given the same reference numerals.

[0042] Throughout this specification, when it is said that a device is "connected" to another device, this includes not only "direct connection" but also "indirect connection" by placing other components in between. Furthermore, when it is said that a device "comprises" a certain constituent element, unless otherwise stated otherwise, this does not exclude other constituent elements, but rather implies that other constituent elements may be included.

[0043] While the terms first, second, etc., are used in some examples herein to refer to various elements, these elements should not be limited by these terms. These terms are used only to distinguish one element from another. For example, first interface and second interface, etc., are used. Furthermore, as used herein, the singular forms “a,” “an,” and “the” are intended to also include the plural forms unless the context indicates otherwise. It should be further understood that the terms “comprising,” “including,” indicate the presence of features, steps, operations, elements, modules, items, kinds, and / or groups, but do not exclude the presence, occurrence, or addition of one or more other features, steps, operations, elements, modules, items, kinds, and / or groups. The terms “or” and “and / or” as used herein are interpreted as inclusive, or mean any one or any combination thereof. Thus, “A, B, or C” or “A, B, and / or C” means “any one of: A; B; C; A and B; A and C; B and C; A, B, and C.” Exceptions to this definition will only occur if the combination of elements, functions, steps, or operations is inherently mutually exclusive in some way.

[0044] The technical terms used herein are for reference only to specific embodiments and are not intended to limit the scope of this disclosure. The singular form used herein includes the plural form unless the statement explicitly indicates otherwise. The word "comprising" as used in this specification means to specify a particular characteristic, region, integer, step, operation, element, and / or component, and does not exclude the presence or addition of other characteristics, regions, integers, steps, operations, elements, and / or components.

[0045] Although not explicitly defined, all terms, including technical and scientific terms used herein, shall have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains. Terms defined in commonly used dictionaries shall be further interpreted as having a meaning consistent with the relevant technical literature and the message of the present disclosure, and shall not be over-interpreted as having an ideal or overly formulaic meaning unless otherwise defined.

[0046] Revascularization after stroke refers to a key treatment measure that, after an acute ischemic stroke (i.e., cerebral infarction), restores blood flow to the occluded cerebral blood vessels as quickly as possible through thrombolysis (such as intravenous alteplase) or endovascular interventional therapy (such as mechanical thrombectomy), in order to salvage the brain tissue in the "ischemic penumbra" that is still in a state of reversible damage, thereby reducing the infarct size and improving neurological prognosis.

[0047] Although revascularization significantly improves patient survival and functional recovery, not all patients who successfully open blood vessels achieve a good prognosis. Some patients may still experience a series of serious complications post-surgery. The root causes are revascularization outside the time window, differences in brain tissue tolerance, poor collateral circulation, and the double-edged sword effect of reperfusion itself. Even if the blood vessel is successfully opened on imaging, if the ischemic time is too long, the penumbra has turned into irreversible necrosis, or microcirculatory dysfunction or uncontrolled activation of the inflammatory cascade may occur, then revascularization may not only fail to save brain tissue but may also trigger secondary damage.

[0048] To address the aforementioned problems, the first aspect of this disclosure provides a monitoring and early warning system after stroke revascularization surgery, which can be referenced. Figure 1 Example, Figure 1 This is a block diagram of the specific structure of the monitoring and early warning system in one embodiment of the present disclosure.

[0049] exist Figure 1 The monitoring and early warning system includes: an acquisition module 100, a first processing module 210, a second processing module 220, and a prediction module 300.

[0050] The acquisition module 100 collects neurophysiological electrical signals from both sides of the patient's brain before and after reperfusion surgery. In some embodiments, the acquisition module 100 collects neurophysiological electrical signals from the patient during a first time period before surgery to obtain a first set of neurophysiological data.

[0051] The acquisition module 100 collects neurophysiological signals from the patient during a second preset time period after recanalization to obtain a second set of neurophysiological data.

[0052] The first processing module 210 and the second processing module 220 are both connected to the acquisition module 100, so as to quantify the first damage level and the second damage level, which characterize the degree of nerve function impairment, based on the difference between the affected side and the healthy side before and after the operation.

[0053] Furthermore, in some embodiments, the affected side is the neurophysiological monitoring side corresponding to the cerebral hemisphere on the side where the patient has suffered a stroke (for example, when the left middle cerebral artery is occluded, the left hemisphere is the affected side, and the evoked potentials recorded by the right limb or ipsilateral cortex are considered as signals from the affected side); the healthy side is the neurophysiological monitoring side corresponding to the contralateral cerebral hemisphere that has not suffered a stroke (i.e., the right hemisphere and its corresponding recording channel). In the absence of other neurological diseases, the healthy side of the brain tissue usually maintains a relatively normal structure and function, and can serve as an individualized internal control.

[0054] The prediction module 300 includes a hierarchical pathological prediction model. The hierarchical pathological prediction model 310 predicts the probability of each complication occurring within the postoperative prediction period based on the first damage level, the second damage level, the first neurophysiological data set, and the second neurophysiological data set. In other words, by inputting the neurophysiological electrical signals acquired by the acquisition module 100, the first damage level, and the second damage level into the prediction module 300, the predicted probability of the aforementioned vascular recanalization complications occurring within the prediction period can be obtained within a preset time after surgery.

[0055] In some embodiments, the complications include at least one or more of the following: intracranial hemorrhage, malignant cerebral edema, ischemia-reperfusion injury, neurological deterioration, and vascular re-occlusion.

[0056] Specifically, intracranial hemorrhage usually occurs after vascular recanalization, especially in patients with large infarct areas and severely damaged brain tissue structures, as the recanalized blood flow impacts the fragile blood vessel walls, leading to bleeding.

[0057] Malignant cerebral edema often occurs secondary to large-area cerebral infarction. It is caused by cell necrosis and the collapse of the blood-brain barrier, which leads to rapidly progressive brain swelling, which can result in brain herniation or even death.

[0058] Ischemia-reperfusion injury is not caused by continuous ischemia, but rather by the fact that after blood flow is restored, the damage to neurons is aggravated by the large-scale generation of free radicals, calcium ion overload, and activation of inflammatory response.

[0059] Postoperative deterioration of neurological function manifests as a decrease in postoperative level of consciousness or further decline in motor / language function, which may be the result of any of the aforementioned pathological processes.

[0060] Re-occlusion refers to the re-blockage of a blood vessel after it has been reopened, often caused by residual thrombus, vasospasm, or endothelial damage, leading to recurrence or worsening of clinical symptoms.

[0061] In summary, if the neurophysiological signals on the affected side are severely suppressed preoperatively, it suggests extensive ischemia leading to vascular bed collapse and microcirculatory structure disruption. After recanalization, the high perfusion rate can easily cause intracranial hemorrhage or rapidly progressing malignant cerebral edema due to the impact on fragile vessels. If the signal does not improve as expected postoperatively or even worsens, it may reflect vascular re-occlusion or no-reflow phenomenon, meaning that although the large vessels are radiographically open, the distal microcirculation has not been effectively perfused. If the signal shows a brief recovery followed by a decline, it strongly suggests blood flow instability or secondary injury, such as ischemia-reperfusion injury or thrombosis recurrence.

[0062] The key to this application's ability to achieve effective prevention after prediction lies in its transformation of early changes in neurophysiological signals into an actionable clinical intervention window. While the aforementioned complications (such as intracranial hemorrhage, malignant cerebral edema, and vascular re-occlusion) have serious consequences, there is often a golden intervention period before typical symptoms appear, during which function has not yet completely collapsed. By dynamically comparing preoperative and postoperative neurophysiological data from the affected / healthy side, risks can be identified hours before abnormal imaging or clinical manifestations, and targeted warnings can be provided based on the pathological characteristics of different complications. For example, when a high risk of hemorrhage is indicated, blood pressure can be strictly controlled and anticoagulation can be suspended; when re-occlusion is warned, vascular assessment can be promptly added or treatment plans adjusted; and when a trend of cerebral edema is identified, dehydration or surgical assessment can be initiated in advance. This mechanism-based early warning allows physicians to shift from passive response to proactive intervention, blocking the pathological cascade reaction at a stage where the damage is still reversible, thereby effectively reducing the incidence or severity of complications and achieving a closed loop of prediction-warning-prevention.

[0063] It is worth noting that the training of the hierarchical pathology prediction model 310 is based on a large amount of neurophysiological data (such as somatosensory evoked potentials, transcranial motor evoked potentials, and electroencephalograms) of stroke patients before and after vascular recanalization, as well as the actual outcomes of postoperative complications (such as whether intracranial hemorrhage or malignant cerebral edema occurred). In some embodiments, the hierarchical pathology prediction model 310 can employ a supervised learning method, using the differences in electrophysiological signals between the affected and healthy sides before and after surgery, the dynamic trends of changes, and clinical variables as inputs, and using the occurrence or risk level of various complications as labels for training. Among the algorithms that can be selected are ensemble learning models (such as XGBoost, LightGBM), deep learning models (such as 1D-CNN, LSTM for processing temporal electrophysiological signals), or multi-task classification networks, to simultaneously predict the incidence probability of multiple complications.

[0064] In some embodiments, the first neurophysiological data in the first neurophysiological data set and / or the second neurophysiological data in the second neurophysiological data set include at least one or more of the following: somatosensory evoked potentials (SSEP), transcranial motor evoked potentials (TcMEP), and electroencephalography (EEG).

[0065] Specifically, SSEP primarily assesses sensory pathways and cortical function, and is highly sensitive to cortical ischemia. A significant decrease or disappearance of its amplitude often indicates structural damage such as malignant cerebral edema or large-area infarction. TcMEP directly monitors motor conduction pathways. If the amplitude drops sharply or cannot be elicited postoperatively, it is often related to vascular re-occlusion or ischemia-reperfusion injury, and can especially provide early warning of complications manifested as motor function deterioration. EEG provides continuous background information on the electrical activity of the entire cerebral cortex, and is more sensitive to diffuse edema, metabolic disorders, or inhibition of neural electrical activity. It can capture the progressive evolution of malignant cerebral edema or sudden electrical failure after reperfusion, and has high temporal resolution. Therefore, the system can achieve more accurate risk warnings based on the characteristics of the target complication.

[0066] exist Figure 1 In this embodiment, the prediction module includes an ischemic injury sub-model 3110.

[0067] The ischemic injury sub-model 3110 is used to obtain reversible values ​​characterizing the patient's neurophysiological function based on the first injury level and the second injury level.

[0068] Specifically, the degree of ischemic injury is the common pathological basis for many serious postoperative complications. When brain tissue experiences prolonged and extensive ischemia due to vascular occlusion, if the damage has progressed to an irreversible stage, even if vascular recanalization is successfully achieved, the destruction of microvascular structure, collapse of the blood-brain barrier, and cellular metabolic failure will still induce a series of secondary damages: fragile vessels in irreversible areas are prone to rupture after blood flow is restored, leading to intracranial hemorrhage; necrotic tissue triggers a strong inflammatory response and fluid extravasation, rapidly developing into malignant cerebral edema; widespread neuronal death directly manifests as deterioration of neurological function; and even if some areas remain viable, endothelial damage and hypercoagulable states caused by severe ischemia may exacerbate ischemia-reperfusion injury or promote vascular re-occlusion. Therefore, the more severe the ischemic injury and the lower its reversibility, the higher the risk of complications for the patient.

[0069] The first damage level and the second damage level mentioned above are obtained by the first processing module 210 and the second processing module 220, respectively. Please refer to... Figure 1 In this embodiment, the first processing module 210 is connected to the acquisition module 100 and is used to obtain a first impairment level that characterizes the degree of preoperative neurological function impairment of the patient based on the difference between the physiological data of the first affected side and the physiological data of the first healthy side at the same time point in the first neurophysiological data set.

[0070] The second processing module 220 is connected to the acquisition module 100 and is used to obtain a second damage level that characterizes the degree of postoperative neurological function impairment in patients based on the difference between the second affected side physiological data and the second healthy side physiological data at the same time point in the second neurophysiological data set.

[0071] Specifically, since stroke is usually a unilateral lesion, the healthy cerebral hemisphere can serve as a relatively normal physiological reference in the absence of other neurological diseases. Therefore, by calculating the ratio, difference, or standardized asymmetry index of signals (such as amplitude, latency, or signal presence) in the physiological data of the affected and healthy sides at the same time point, it is possible to effectively eliminate baseline differences between individuals and quantify the degree of lateral impairment of neurological function.

[0072] The first level of damage represents the degree of functional inhibition of brain tissue before vascular recanalization, reflecting the initial damage state caused by ischemia. If the signal on the affected side has completely disappeared or is extremely attenuated at this time, it indicates that the ischemic time may be long, the core infarction area is large, and there is less salvageable penumbra.

[0073] For example, if the SSEP amplitude on the affected side is significantly lower than that on the healthy side before surgery, the TcMEP threshold is elevated or cannot be elicited, and the EEG background activity is significantly slowed or asymmetrical, it indicates that ischemia has caused significant functional inhibition and outputs a higher first-degree impairment.

[0074] The second level of damage represents the immediate performance of the affected side's function after recanalization (usually in the early postoperative period, such as 30 minutes to several hours). If the signal on the affected side improves significantly and the bilateral asymmetry narrows at this time, it indicates that the recanalization was effective and some previously suppressed but not necrotic neurons have resumed electrical activity, suggesting reversible damage. Conversely, if the signal does not improve or even worsens after the procedure, it indicates that the damage is likely irreversible or that the recanalization failed to provide effective perfusion.

[0075] For example, if the postoperative signal is nearly symmetrical, the amplitude recovers, and the EEG rhythm improves, then the second damage grade is low; otherwise, it remains high.

[0076] It should be noted that the first processing module 210 and the second processing module 220 can be implemented as rule-based threshold models or lightweight machine learning models. For example, by calculating the ratio or difference of neurophysiological signals (such as amplitude, latency, etc.) between the affected and healthy sides at the same time point, the damage level can be directly classified according to a preset clinical threshold (e.g., amplitude ratio <0.5 indicates severe damage), without training. In more complex scenarios, it can also be constructed as a supervised classification or regression model (such as logistic regression, XGBoost, or shallow neural networks), using physiological data from the affected / healthy side as input and expert annotations or imaging / clinical outcomes as labels for training. For example, the damage level can be defined using postoperative DWI infarct volume or NIHSS changes, and the model parameters can be optimized through cross-validation. Regardless of the form adopted, the core objective is to transform the original electrophysiological asymmetry into a stable and standardized first and second damage level, providing a reliable and comparable quantitative basis for subsequent assessment of reversibility and early warning of complications.

[0077] It is worth noting that the ischemic injury sub-model 3110 is trained based on a large amount of historical data from stroke patients. It uses preoperative first-level injury and postoperative second-level injury as input features, and combines real clinical or imaging outcomes reflecting the final extent of brain tissue damage (such as final infarct volume, whether malignant cerebral edema occurs, functional prognosis, etc.) as labels, employing supervised learning methods for modeling. Specifically, if the patient's postoperative neurological function significantly recovers and there are no serious complications, it is marked as "highly reversible"; if the injury continues to worsen or large-area necrosis occurs, it is marked as "lowly reversible." In some embodiments, the ischemic injury sub-model 3110 can use ensemble learning algorithms such as XGBoost and LightGBM, or shallow neural networks, to learn the mapping relationship between changes in injury level and tissue reversibility potential during training, ultimately outputting a continuous "reversible value" between 0 and 1 to quantify the functional recovery capacity of the patient's brain tissue after recanalization.

[0078] exist Figure 1 In this embodiment, the prediction module further includes a surgical immediate effectiveness sub-model 3120, which is used to obtain an effective value characterizing the degree of improvement in neurological function from the first neurophysiological data set to the second neurophysiological data set within the monitoring time range.

[0079] Specifically, effectiveness does not refer to whether the blood vessel is open on imaging, but rather reflects whether the brain tissue achieves a functional effect due to the restoration of blood flow. Effectiveness reflects the probability of complications because it directly reveals the quality of recanalization. For example, high effectiveness indicates good microcirculation perfusion, salvage of the penumbra, and a lower risk of complications for the patient; low effectiveness suggests possible pathological conditions such as no-reflow, re-occlusion, or reperfusion injury, which greatly increases the risk of secondary vascular re-occlusion, malignant cerebral edema, or neurological deterioration.

[0080] For example, if the signal on the affected side recovers significantly after surgery—such as an increase in evoked potential amplitude, a shortened latency, or a re-elicitation of the signal, or if the EEG rhythm becomes more symmetrical and normalized—it indicates that the recanalization is effective and highly effective. Conversely, if the signal does not improve or even worsens after surgery, it indicates that although the large blood vessels may have been opened, the microcirculation has not been effectively perfused or secondary damage has occurred, resulting in low effectiveness.

[0081] It is worth noting that the immediate surgical effectiveness sub-model 3120 can be implemented as a regression or classification model based on temporal feature changes to quantify the degree of improvement in neurophysiological function from preoperative to postoperative. Specifically, the immediate surgical effectiveness sub-model 3120 takes the first and second neurophysiological data sets as inputs, extracts the changes in the affected side signal before and after recanalization, and outputs an effective value (e.g., between 0 and 1) or a graded score. During training, clinical or imaging outcomes are used as supervisory signals: for example, a NIHSS improvement of ≥4 points within 24 hours postoperatively, no complications, or good follow-up mRS (0–2 points) is defined as high effectiveness, and vice versa for low effectiveness; continuous indicators of functional improvement (e.g., ΔNIHSS, infarct growth volume) can also be directly used as regression targets. Thus, in patients, the effectiveness of recanalization can be objectively assessed based solely on preoperative and early postoperative signals, providing crucial evidence for complication warning.

[0082] exist Figure 1 In this embodiment, the prediction module further includes an early warning sub-model 3130. Please refer to [the documentation / reference] for details. Figure 2 Example, Figure 2 Two implementation methods are used for the early warning sub-model 3130.

[0083] As an example, when the early warning sub-model 3130 is implemented as the first early warning sub-model 3131, the first early warning sub-model 3131 predicts the prediction result based on the reversible value and the effective value.

[0084] Specifically, reversibility and effectiveness values ​​can determine the risk probability of various complications because they quantify two core physiological factors that determine the success of stroke recanalization: whether brain tissue still has salvage potential before recanalization (reversibility), and whether functional recovery is truly achieved after recanalization (effectiveness). Extensive clinical evidence shows that the occurrence of different complications is not random, but rather an inevitable result of the combination of reversibility and effectiveness. For example, a low reversibility value indicates large-area irreversible necrosis, fragile vessel walls, and disruption of the blood-brain barrier, making intracranial hemorrhage or malignant cerebral edema likely once recanalized; while a high reversibility value accompanied by a low effective value suggests ineffective perfusion of the penumbra, strongly indicating re-occlusion or ischemia-reperfusion injury. By learning the statistical association between these combinations and actual complication outcomes in historical patient cohorts (e.g., 70% of patients with "reversibility value <0.3 and effective value <0.4" develop malignant cerebral edema), the model transforms pathological mechanisms into data patterns, thus enabling it to independently output the probability of occurrence for each complication.

[0085] It is worth noting that the training objective of the first early warning sub-model 3131 is not to determine the general outcome of whether complications will occur, but rather to predict the probability of occurrence for each specific complication (such as intracranial hemorrhage, malignant cerebral edema, ischemia-reperfusion injury, and vascular re-occlusion). In some embodiments, during training, the system uses historical patient data to individually label each complication as having occurred (e.g., if a patient experiences intracranial hemorrhage but does not experience re-occlusion, the label for "intracranial hemorrhage" is 1, and the label for "vascular re-occlusion" is 0). Reversibility and effectiveness values ​​are used as core input features, and multi-label classification is employed for learning. For example, intracranial hemorrhage is often associated with low reversibility and low effectiveness values, while vascular re-occlusion is more commonly associated with a combination of high reversibility and low effectiveness values. Ultimately, the model outputs the incidence probability of each complication (e.g., "70% risk of intracranial hemorrhage, 20% risk of vascular re-occlusion").

[0086] As an example, when the warning sub-model 3130 is implemented as the second warning sub-model 3132, the acquisition module 100 is also used to collect real-time neurophysiological data after surgery; the second warning sub-model 3132 predicts the prediction result based on the reversible value, the effective value and the real-time neurophysiological data.

[0087] Specifically, the first early warning sub-model 3131 is used to identify inherently high-risk states (such as large-area irreversible infarction with ineffective recanalization) determined by the initial degree of injury and the immediate effect of recanalization, and is suitable for risk stratification immediately after surgery. Real-time neurophysiological data can capture the evolution trend of the condition in the following hours, such as the sudden deterioration of originally stable signals or the relapse after improvement, which are dynamic changes that cannot be covered by static assessment. Therefore, the second early warning sub-model 3132 further improves the system's monitoring sensitivity for delayed, progressive, or sudden complications (such as vascular re-occlusion 2 hours after surgery or malignant cerebral edema progression 12 hours after surgery), realizing a complete closed loop from initial risk assessment to continuous risk tracking. The two work together to enhance the comprehensiveness and timeliness of the early warning.

[0088] It is worth noting that the training of the second early warning sub-model 3132 is based on reversible and effective values, and further incorporates real-time neurophysiological data continuously acquired postoperatively. During training, the model input includes two parts: first, static features, namely the reversible and effective values ​​calculated preoperatively and early postoperatively, used to represent the patient's basic injury state and immediate recanalization effect; second, dynamic features, namely real-time neurophysiological signals at multiple postoperative time points, which reflect the subsequent dynamics of brain function by extracting temporal features such as amplitude variation trends, signal stability, and the evolution of asymmetry between the affected and healthy sides over time. Labels are applied based on whether the patient develops various specific complications (such as intracranial hemorrhage, malignant cerebral edema, vascular re-occlusion, etc.) within the postoperative observation window.

[0089] In some embodiments, the second early warning sub-model 3132 can directly model signal evolution using a temporal neural network (such as LSTM), or it can be converted into structured statistical features and trained using an ensemble tree model (such as XGBoost). Ultimately, the second early warning sub-model 3132 can dynamically output the updated risk probability of each complication at any postoperative time, combining the latest monitoring data, while retaining the rapid assessment advantage of the first early warning sub-model 3131, significantly enhancing the ability to identify delayed or progressive events.

[0090] In some embodiments, the monitoring and early warning system further includes a third processing module 230, connected between the acquisition module 100 and the surgical real-time effectiveness sub-model 3120.

[0091] You can refer to them together. Figures 3 to 5 The various implementation methods are shown. In Figure 3 In this module, the third processing module 230 is located outside the prediction module, serving as an independent preprocessing unit. Figure 4 In this context, it is integrated into the hierarchical pathology prediction model. And... Figure 5 Although it is within the prediction module, it is not included in the structure of the hierarchical pathology prediction model 310.

[0092] Therefore, this application does not limit the specific location of the third processing module 230, and its deployment can be flexibly adjusted according to system architecture, computing resources, or clinical procedures. Regardless of its physical or logical location, the function of the third processing module 230 is to extract and generate change features reflecting the dynamic evolution of neural function based on the corresponding signals (such as somatosensory evoked potentials, transcranial motor evoked potentials, and electroencephalograms) in the first neurophysiological data set (preoperative) and the second neurophysiological data set (postoperative). These features include, but are not limited to, amplitude change rate, signal evoked state transition, EEG rhythm shift, and improvement or deterioration of symmetry between the affected and healthy sides, to quantify the differences in neurophysiological responses before and after recanalization.

[0093] Specifically, the changing characteristics include at least one of the following:

[0094] (1) The rate of change of amplitude between the physiological data of the first affected side and the physiological data of the second affected side.

[0095] The amplitude change rate is used to measure the degree of recovery of neuronal electrical activity on the affected side after recanalization. If the amplitude increases significantly (e.g., an increase of more than 50% in SSEP or TcMEP amplitude), it indicates that the neurons in the ischemic area have regained their electrophysiological function after the blood flow is restored, suggesting that the recanalization is effective; conversely, if the amplitude does not change or further decreases, it indicates that the recanalization has failed to improve or even aggravated functional inhibition, and the effectiveness is low.

[0096] (2) Changes in the extraction state of postoperative neurophysiological signals on the affected side based on physiological data of the second affected side.

[0097] For example, the successful elicitation of a preoperatively absent signal is a clear indicator of neural function reconstruction, suggesting high effectiveness. Conversely, if the signal cannot be elicited postoperatively, or if a previously elicitable signal disappears, it suggests a persistent interruption or secondary damage to the conduction pathway, indicating poor effectiveness. This shift in the "present / absent" state is a highly reliable indicator of whether recanalization has been effective.

[0098] (3) Postoperative EEG rhythm changes of the affected side relative to the healthy side, based on physiological data of the second affected side and physiological data of the second healthy side.

[0099] The symmetry of EEG rhythms (such as α and β wave power) is an important indicator of the integrity of cortical function. If the rhythm on the affected side is close to that on the healthy side after surgery (such as a decrease in slow waves and a recovery of fast waves), it indicates that cortical metabolism and synchronicity have improved, and the effectiveness is high. If the affected side continues to show diffuse slow waves or rhythm inhibition, it reflects that the cortical function has not recovered or that edema has progressed, and the effectiveness is low.

[0100] (4) Changes in bilateral symmetry between the affected and healthy sides after surgery, based on physiological data of the second affected side and physiological data of the second healthy side.

[0101] The balance of functional recovery is quantified by calculating the amplitude ratio, power ratio, or asymmetry index of the affected / healthy side. Higher symmetry (e.g., a ratio close to 1) indicates that bilateral function tends to be balanced after reperfusion, indicating good effectiveness; if asymmetry persists or worsens, it suggests insufficient perfusion or damage extension on the affected side, indicating poor effectiveness.

[0102] In summary, these four types of change characteristics, from the four dimensions of amplitude recovery, signal presence, rhythm normalization, and bilateral balance, together constitute the determination of whether recanalization is truly effective, so as to obtain an effective value.

[0103] Please refer to Figure 6 In the example, Figure 6 In this embodiment, the monitoring and early warning system further includes an alarm module 400, which is connected to the prediction module and is used to generate a corresponding alarm level based on the prediction result and send it to external staff.

[0104] Specifically, the system pre-sets multiple risk thresholds (e.g., risk probability <30% is low risk, 30%–70% is medium risk, and 70% is high risk), and semantically annotates alarm content based on complication types (e.g., intracranial hemorrhage, malignant cerebral edema, etc.). When the prediction result reaches a certain threshold, the alarm module 400 triggers an alarm of the corresponding level (e.g., green alert, yellow warning, or red emergency alarm), and pushes the alarm information in real time to the terminal devices of external staff (e.g., medical workstations, mobile terminals, or monitoring systems). The alarm information not only includes the risk level but can also include high-risk complication types, key evidence (e.g., low effectiveness + low reversibility), and suggested intervention measures, thereby supporting rapid identification of the nature of the risk and targeted treatment.

[0105] In some embodiments, the acquisition module 100 includes multiple neurophysiological sensors deployed on the body surface or cortical monitoring areas (such as scalp electrodes, spinal cord electrodes, or intraoperative cortical electrodes) corresponding to the affected and unaffected sides of the patient's brain, respectively, to synchronously acquire neurophysiological signals reflecting the functional state of both cerebral hemispheres. These sensors can record multimodal signals such as somatosensory evoked potentials (SSEP), transcranial motor evoked potentials (TcMEP), and / or electroencephalograms (EEG) in real time before surgery (for generating the first set of neurophysiological data) and / or after surgery (for generating the second set of neurophysiological data). By pairing and monitoring at symmetrical or functionally corresponding locations on the affected and unaffected sides, the system can effectively eliminate baseline differences between individuals and capture the impact of ischemia on neural conduction pathways.

[0106] Furthermore, all of the aforementioned neuroelectrophysiological data were collected and provided by legitimate institutions in accordance with laws and regulations, such as qualified hospitals and other medical institutions, or by professionals such as doctors and technicians practicing in medical institutions who recorded the data in accordance with clinical standards during the diagnosis and treatment process.

[0107] like Figure 7 The diagram shown illustrates the structure of a computer device according to an embodiment of the present disclosure.

[0108] The computer device 500 may be exemplified as a processing terminal in a cloud platform, such as a server, desktop computer, laptop computer, tablet computer, smartphone, or other terminal.

[0109] The computer device 500 includes a bus 501, a processor 502, and a memory 503. The processor 502 and the memory 503 can communicate with each other via the bus 501. The memory 503 can store program instructions. The processor 502 runs the monitoring and early warning system in the previous embodiment by executing the program instructions stored in the memory 503.

[0110] Bus 501 can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, although only one thick line is used in the diagram, this does not indicate that there is only one bus or one type of bus.

[0111] In some embodiments, processor 502 may be implemented as a central processing unit (CPU), microprocessor unit (MCU), system on chip (System on Chip), or field-programmable array (FPGA). Memory 503 may include volatile memory for temporary data storage during program execution, such as random access memory (RAM).

[0112] The memory 503 may also include non-volatile memory for data storage, such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid-state disk (SSD).

[0113] In some embodiments, the computer device 500 may further include a communicator 504. The communicator 504 is used for communication with external devices. In specific examples, the communicator 504 may include one or more wired and / or wireless communication circuit modules. For example, the communicator 504 may include one or more of, such as a wired network card, a USB module, a serial interface module, etc. The wireless communication protocols followed by the wireless communication module include, for example, Nearfield Communication (NFC) technology, Infrared (IR) technology, Global System for Mobile Communications (GSM), General Packet Radio Service (GPRS), Code Division Multiple Access (CDMA), Wideband Code Division Multiple Access (WCDMA), Time-Division Code Division Multiple Access (TD-SCDMA), Long Term Evolution (LTE), Bluetooth (BT), Global Navigation Satellite System (GNSS), etc.

[0114] This disclosure also provides a computer-readable storage medium storing program instructions that, when executed, implement the operation of the monitoring and early warning system in any of the previous embodiments.

[0115] That is, the system in the above embodiments is implemented as software or computer code that can be stored in a recording medium (such as CD ROM, RAM, floppy disk, hard disk or magneto-optical disk), or as computer code that is originally stored in a remote recording medium or a non-transitory machine-readable medium and will be stored in a local recording medium after being downloaded via a network. Thus, the system represented herein can be stored on such software processing on a recording medium using a general-purpose computer, a dedicated processor or programmable or dedicated hardware (such as ASIC or FPGA).

[0116] This disclosure also provides a computer program product, including: program instructions for executing the monitoring and early warning system described in any of the above embodiments.

[0117] The above embodiments are merely illustrative of the principles and effects of this disclosure and are not intended to limit this disclosure. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of this disclosure. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this disclosure should still be covered by the protection scope of this disclosure.

Claims

1. A monitoring and early warning system after stroke recanalization, characterized in that, include: The acquisition module is used to acquire the first set of neurophysiological data of the affected and unaffected sides of the patient's brain during a first preset time period before surgery, and the second set of neurophysiological data of the affected and unaffected sides of the brain during a second preset time period after vascular recanalization. The first processing module, connected to the acquisition module, is used to obtain a first level of impairment that characterizes the degree of preoperative neurological function impairment in patients based on the difference between the physiological data of the first affected side and the physiological data of the first healthy side at the same time point in the first neurophysiological data set. as well as, The second processing module, connected to the acquisition module, is used to obtain a second impairment level that characterizes the degree of postoperative neurological function impairment in patients based on the difference between the second affected side physiological data and the second healthy side physiological data at the same time point in the second neurophysiological data set. The prediction module is connected to the acquisition module, the first processing module, and the second processing module, respectively. The prediction results of a hierarchical pathological prediction model are based on the first damage level, the second damage level, the first neuroelectrophysiological data set, and the second neuroelectrophysiological data set to predict the probability of each complication during the postoperative prediction period. The stratified pathological prediction model includes an ischemic injury sub-model, a surgical immediate effectiveness sub-model, and an early warning sub-model. The ischemic injury sub-model is used to obtain reversible values ​​characterizing the patient's neurophysiological function based on the first injury level and the second injury level. The surgical immediate effectiveness sub-model is used to obtain an effective value characterizing the degree of improvement in neurological function by measuring the changes in the characteristics of the changes from the first neuroelectrophysiological data set to the second neuroelectrophysiological data set within the monitoring time range. When the early warning sub-model is implemented as the first early warning sub-model, the first early warning sub-model predicts the prediction result based on the reversible value and the effective value.

2. The monitoring and warning system of claim 1, wherein When the warning sub-model is implemented as the second warning sub-model, the acquisition module is also used to collect real-time neurophysiological data after surgery; the second warning sub-model predicts the prediction result based on the reversible value, the effective value and the real-time neurophysiological data.

3. The monitoring and warning system of claim 1, wherein The affected side is the neurophysiological monitoring side corresponding to the cerebral hemisphere on the side where the patient suffered a stroke; and / or, the unaffected side is the neurophysiological monitoring side corresponding to the contralateral cerebral hemisphere on the side where the patient did not suffer a stroke.

4. The monitoring and warning system of claim 1, wherein, The complications include at least one or more of the following: intracranial hemorrhage, malignant cerebral edema, ischemia-reperfusion injury, neurological deterioration, and vascular re-occlusion.

5. The monitoring and warning system of claim 1, wherein, The first neurophysiological data in the first neurophysiological data set and / or the second neurophysiological data in the second neurophysiological data set include at least one or more of the following: somatosensory evoked potentials, transcranial motor evoked potentials, and electroencephalograms.

6. The monitoring and warning system of claim 5, wherein, It also includes a third processing module, connected between the acquisition module and the surgical immediate effectiveness sub-model, for obtaining the change characteristics including the patient's neural wave amplitude change rate, neural signal extraction state change, brain rhythm change and bilateral symmetry change based on the somatosensory evoked potential, transcranial motor evoked potential and electroencephalogram processing, and outputting them to the surgical immediate effectiveness sub-model.

7. The monitoring and warning system of claim 1, wherein The changing characteristics include at least one of the following: (1) The rate of change in amplitude between the physiological data of the first affected side and the physiological data of the second affected side; (2) Changes in the extraction state of postoperative neurophysiological signals on the affected side based on physiological data from the second affected side; (3) Postoperative changes in EEG rhythm of the affected side relative to the healthy side, based on physiological data of the second affected side and physiological data of the second healthy side; (4) Changes in bilateral symmetry between the affected and healthy sides after surgery, based on physiological data of the second affected side and physiological data of the second healthy side.

8. The monitoring and warning system of claim 1, wherein, It also includes an alarm module, which is connected to the prediction module and is used to generate a corresponding alarm level based on the prediction result and send it to external staff.

9. The monitoring and warning system of claim 1, wherein, The acquisition module includes multiple neurophysiological sensors, which are respectively deployed in the monitoring areas corresponding to the affected side and the healthy side of the brain, for collecting the first neurophysiological data set and / or the second neurophysiological data set.

10. A computer apparatus, comprising: include: Processor and memory; The memory stores program instructions; The processor is configured to execute the program instructions to run the monitoring and early warning system as described in any one of claims 1-7.

Images