A method and system for identifying tissue state based on intraoperative ice ball spatiotemporal evolution characteristics

Abstract

The application discloses a kind of based on intraoperative ice ball space-time evolution characteristics and system of tissue state identification method.This method obtains time sequence medical image data and freezing process parameters including ice ball region in the process of cryoablation, extracts the ice ball space-time evolution characteristics including ice ball boundary feature, directional expansion feature and expansion dynamics feature, identifies the local perfusion influence degree (first tissue state parameter) and local heat conduction capacity (second tissue state parameter) of target tissue based on the characteristics and freezing process parameters, and outputs its spatial distribution result.The application promotes ice ball from the traditional coverage boundary display to tissue state identification signal, can identify the tissue perfusion heterogeneity and heat conduction heterogeneity that conventional image is difficult to directly measure in real time during operation, provides objective, quantitative information support for intraoperative decision of cryoablation.

Description

Technical Field

[0001] This invention relates to the fields of tumor ablation therapy, medical image processing and medical device control technology, and in particular to a method and system for tissue state identification based on the spatiotemporal evolution characteristics of intraoperative hockey puck. Background Technology

[0002] Cryoablation and combined cryoablation are important techniques in minimally invasive local tumor treatment. During cryoablation, a cryogenic medium is applied to the target tissue via an ablation probe, creating an ice ball around the target tissue that can be observed in CT, MRI, or ultrasound images. In clinical practice, the ice ball is often considered the visual boundary of the cryoablation coverage. Doctors primarily judge the adequacy of the cryoablation based on whether the ice ball covers the lesion, and accordingly decide whether to continue, stop, or proceed to subsequent treatment steps.

[0003] However, current technologies for utilizing ice hockey are mainly limited to geometric coverage, and have the following shortcomings: Current intraoperative assessment methods typically regard the ice ball as a low-density boundary formed around the lesion, without further utilizing information such as directional expansion differences, boundary irregularities, and changes in expansion speed contained in the ice ball formation process; Differences in local blood perfusion, tissue heat conduction, and tissue heterogeneity within the target tissue can significantly affect the cryoheat transfer process and the final cryotherapy effect. However, these tissue conditions are difficult to measure directly through conventional imaging during the operation and are also difficult to accurately reflect through single-point temperature measurement. Under the same probe structure, the same refrigerant type, and the same freezing excitation conditions, the ice ball morphology formed by different patients, different lesions, and even the same lesion in different directions may still have significant differences. These differences often correspond to tissue state information such as local blood vessel distribution, tissue fibrosis degree, necrotic area distribution, and changes in thermal conductivity. However, the existing technology lacks a technical solution to systematically transform such differences into tissue state recognition results. Due to the lack of intraoperative means to identify perfusion heterogeneity and heat conduction heterogeneity, doctors often rely on experience when judging whether freezing is sufficient, where there may be insufficient cooling, and which directions may be affected by blood vessels, making it difficult to obtain more objective and quantitative intraoperative information support.

[0004] Therefore, it is necessary to provide a new technical solution so that the ice puck is no longer just displayed as a coverage boundary, but can be used as a response carrier formed by the target tissue under controlled freezing stimulation, and the tissue state information of the target tissue can be identified by analyzing the spatiotemporal evolution characteristics of the ice puck. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a method and system for tissue state identification based on the spatiotemporal evolution characteristics of an intraoperative ice puck. This method treats the ice puck as a spatiotemporal response signal formed by the target tissue under controlled freezing stimulation. By extracting the boundary features, directional expansion features, and expansion dynamics features of the ice puck, it identifies the degree of local perfusion influence and local heat conduction capacity of the target tissue, thereby obtaining tissue state information that is difficult to measure directly during surgery.

[0006] The core concept of this invention is to regard the ice puck as a "spatiotemporal response function" generated by the target tissue under controlled freezing stimulation.

[0007] Specifically, under controlled freezing excitation (with known parameters such as refrigerant flow rate, probe temperature, and pipeline pressure), the formation process of the ice puck is jointly influenced by the local blood perfusion and local heat conduction capacity of the target tissue. Therefore, the spatiotemporal evolution characteristics of the ice puck contain encoded information about tissue perfusion and heat conduction states. By establishing an inverse mapping relationship of "excitation-response-state," tissue state parameters can be decoded from the spatiotemporal evolution characteristics of the ice puck.

[0008] The physical basis of this concept lies in: Mechanism of perfusion influence: Local blood flow creates a heat sink effect on the frozen area by continuously delivering heat. The stronger the blood flow perfusion, the more the ice ball expansion is hindered, the more the boundary is more likely to deviate away from the blood vessel, and the more obvious the fluctuations are in the steady state phase. The mechanism of heat conduction: The thermal conductivity of local tissue determines the efficiency of cold energy transfer from the probe to the surrounding area. Abnormal thermal conductivity (such as the thermal conductivity of fibrotic tissue being lower than that of normal liver tissue) will lead to irregular ice ball boundaries, nonlinear changes in expansion speed, and abnormal boundary curvature distribution.

[0009] Based on this, the present invention constructs a technical chain of "feature extraction - state inversion - spatial distribution output", realizing cross-domain mapping from the geometric shape of the ice hockey puck to the physical state of the organization.

[0010] The technical solution of the present invention is as follows: A method for identifying tissue state based on the spatiotemporal evolution characteristics of intraoperative hockey puck, characterized by the following steps: S1. During the cryoablation of the target tissue, acquire time-series medical image data containing the ice ball region, and acquire the cryoablation process parameters corresponding to the cryoablation process; S2. Extract the spatiotemporal evolution features of ice hockey based on the time-series medical image data. The spatiotemporal evolution features of ice hockey include at least two of the following: ice hockey boundary features, directional expansion features, and expansion dynamics features. S3. Based on the spatiotemporal evolution characteristics of the ice hockey puck and the freezing process parameters, identify the tissue state parameters of the target tissue. The tissue state parameters include at least a first tissue state parameter characterizing the degree of local perfusion influence and a second tissue state parameter characterizing the local heat conduction capacity. S4. Output the spatial distribution results of the first tissue state parameter and / or the second tissue state parameter.

[0011] In some embodiments, the freezing process parameters include at least one of refrigerant flow rate, freezing duration, probe temperature, and pipeline pressure; the ice puck boundary characteristics include at least one of ice puck boundary position, ice puck volume, ice puck surface area, ice puck major-minor axis ratio, sphericity, boundary curvature, boundary irregularity, and boundary fractal dimension; the directional expansion characteristics include at least one of ice puck expansion radius in different angular directions, expansion direction offset, symmetry index, and expansion difference in each direction; the expansion dynamic characteristics include at least one of ice puck expansion velocity, expansion acceleration, expansion velocity change rate, time required to reach a preset radius, steady-state maintenance fluctuation characteristics, and retraction velocity.

[0012] In some implementations, when extracting the spatiotemporal evolution features of the ice hockey puck in step S2, the ice hockey puck boundary at multiple time points is determined based on the time-series medical image data, and the directional expansion features and the expansion dynamic features are extracted based on the spatial changes of the ice hockey puck boundary at multiple time points; the ice hockey puck boundary is determined by segmentation based on CT value thresholds, segmentation based on region growing, or segmentation based on deep learning.

[0013] In some implementations, in step S3, the tissue state parameters are identified by a tissue state identification model. The tissue state identification model is used to establish the correspondence between the spatiotemporal evolution characteristics of the ice hockey puck and the first and second tissue state parameters. The tissue state identification model is an inversion model based on biological heat transfer constraints, a machine learning model based on physical constraints, a finite element inverse problem solving model, an ensemble Kalman filter model, or a combination thereof.

[0014] In some embodiments, the spatial distribution results include at least one of the following: local perfusion heterogeneity distribution map, local heat conduction heterogeneity distribution map, abnormal freezing response area, undercooling risk area, and vascular impact risk direction; the spatial distribution results output in step S4 are at least one of the following: two-dimensional distribution map, three-dimensional distribution map, zonal risk level map, or directional risk indication.

[0015] In some implementations, based on the first tissue state parameter and the second tissue state parameter, abnormal expansion areas, perfusion enhancement areas, heat conduction obstruction areas, or potential ablation blind areas in the target tissue are marked and output; when the first tissue state parameter is higher than a preset threshold, the corresponding area is marked as a perfusion enhancement area or a vascular influence area; when the second tissue state parameter is lower than or higher than a preset threshold, the corresponding area is marked as a heat conduction obstruction area or a heat conduction abnormal area; when the first tissue state parameter and the second tissue state parameter jointly indicate that the freezing of the area is difficult to fully reach the preset threshold, the area is marked as a potential undercooling risk area or a potential ablation blind area.

[0016] In some embodiments, the time-series medical image data is CT image data, MRI image data, ultrasound image data, or a combination thereof; the acquisition time interval of the time-series medical image data in step S1 is dynamically adjusted according to the puck expansion speed. When the puck expansion speed exceeds a preset threshold, the acquisition time interval is shortened; when the puck expansion speed is below the preset threshold, the acquisition time interval is extended.

[0017] In some implementations, when identifying tissue state parameters in step S3, prior anatomical information of the target tissue is also introduced. The prior anatomical information includes at least one of a vascular distribution map and a tissue type distribution map. The spatial resolution of the first tissue state parameter and the second tissue state parameter is higher than the spatial resolution of the time-series medical image data, which is achieved through interpolation or super-resolution reconstruction of the tissue state identification model.

[0018] Another aspect of the present invention discloses a tissue state identification system based on the spatiotemporal evolution characteristics of intraoperative hockey puck, comprising: The data acquisition module is used to acquire time-series medical image data including the ice puck area and the corresponding freezing process parameters during the cryoablation of the target tissue; The feature extraction module is used to extract the spatiotemporal evolution features of ice hockey based on the time-series medical image data. The spatiotemporal evolution features of ice hockey include at least two of the following: ice hockey boundary features, directional expansion features, and expansion dynamics features. The state recognition module is used to identify the tissue state parameters of the target tissue based on the spatiotemporal evolution characteristics of the ice hockey puck and the freezing process parameters. The tissue state parameters include at least a first tissue state parameter characterizing the degree of local perfusion influence and a second tissue state parameter characterizing the local heat conduction capacity. The output module is used to output the spatial distribution results of the first tissue state parameter and / or the second tissue state parameter.

[0019] Furthermore, the output module is further used to output at least one of the following: local perfusion heterogeneity distribution map, local heat conduction heterogeneity distribution map, abnormal freezing response area, undercooling risk area, and vascular influence risk direction; the state recognition module integrates a tissue state recognition model, which is an inversion model based on biological heat transfer constraints, a machine learning model based on physical constraints, a finite element inverse problem solving model, an ensemble Kalman filter model, or a combination thereof; the data acquisition module includes at least one of CT acquisition unit, MRI acquisition unit, and ultrasound acquisition unit, as well as a process parameter acquisition unit for acquiring at least one of probe temperature, refrigerant flow rate, freezing duration, and pipeline pressure; the system also includes a feedback control module for generating cryoablation parameter adjustment suggestions based on the first tissue state parameter and / or the second tissue state parameter, the adjustment suggestions including at least one of probe position adjustment, refrigerant flow rate adjustment, and freezing duration adjustment.

[0020] Compared with the prior art, the present invention has at least the following beneficial effects: 1. This invention elevates the ice puck from a traditional "coverage boundary display result" to a "tissue state identification signal," thus changing the way ice pucks are used in intraoperative cryoablation; 2. This invention can identify local perfusion heterogeneity and thermal conductivity heterogeneity that are difficult to obtain through direct observation of conventional images or single-point temperature measurement; 3. This invention can output abnormal freezing response areas, undercooling risk areas, or vascular impact risk directions based on the spatiotemporal evolution differences of ice hockey, providing a more objective basis for intraoperative judgment; 4. This invention can obtain richer intraoperative tissue status information on the basis of existing cryoablation procedures without relying on additional invasive detection, and has good clinical application value; 5. This invention can serve as the upstream information basis for subsequent treatment decisions, risk alerts, and ablation strategy optimization; 6. This invention optimizes intraoperative resource utilization efficiency while ensuring information density by dynamically adjusting the image acquisition interval; 7. This invention incorporates prior anatomical information, which improves the anatomical localization accuracy and clinical interpretability of state recognition. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the overall process of the method of the present invention;

[0022] Figure 2 This is a schematic diagram of the ice hockey time-series medical image data acquisition and processing process in this invention;

[0023] Figure 3 This is a schematic diagram of the process for extracting spatiotemporal evolution features of ice hockey in this invention;

[0024] Figure 4 This is a schematic diagram illustrating the input-output relationship of the tissue state recognition model in this invention;

[0025] Figure 5 This is a schematic diagram of the output of the spatial distribution results of tissue state in this invention;

[0026] Figure 6 This is a block diagram of the system structure of the present invention. Detailed Implementation

[0027] The present invention will be further described in detail below with reference to specific embodiments. It should be understood that the following embodiments are only used to explain the present invention and are not intended to limit the scope of protection of the present invention.

[0028] Example 1

[0029] like Figure 1 As shown, in this embodiment, CT is used as the intraoperative medical imaging method to acquire time-series CT image data including the ice hockey puck area during the cryoablation of the target tissue.

[0030] The system includes at least: Cryoablation probe; Refrigerant supply and refrigeration control unit; CT image acquisition unit; Process parameter acquisition unit; Data processing and organizational status identification unit.

[0031] The process parameter acquisition unit is used to acquire at least one of the following: refrigerant flow rate, freezing duration, probe temperature, and pipeline pressure.

[0032] After cryoablation begins, the CT image acquisition unit acquires images of the hockey puck area at preset time intervals, forming a time-series CT image sequence. For example... Figure 2 As shown, the time interval can be any value from 0.5 seconds to 30 seconds, preferably any value from 1 second to 10 seconds. In a preferred embodiment, the acquisition time interval can be dynamically adjusted according to the puck's expansion speed: when the puck's expansion speed exceeds a preset threshold (e.g., 5 mm / min), the acquisition time interval is shortened to 1-3 seconds; when the puck's expansion speed is below the preset threshold (e.g., 1 mm / min), the acquisition time interval is extended to 10-30 seconds.

[0033] The data processing and tissue state recognition unit first performs hockey puck segmentation on the time-series CT image sequence to obtain hockey puck boundaries at multiple time points.

[0034] The specific methods for dividing a hockey puck include: (1) Threshold-based segmentation In CT images, the hockey puck appears as a low-density area (CT value approximately -20 to -50 HU, lower than the surrounding soft tissue). A CT value threshold range (e.g., -60 HU to 0 HU) is set, and noise is removed using morphological opening and closing operations to extract the hockey puck region.

[0035] (2) Segmentation based on region growing Using the probe tip or the center of a known ice puck as the seed point, a gray-scale similarity threshold and morphological constraints are set to perform region growth and obtain the ice puck boundary.

[0036] (3) Deep learning-based segmentation We employ network architectures such as U-Net and nnU-Net, and train them using historically annotated intraoperative CT images of ice hockey pucks to achieve automatic segmentation. The training data should include samples of different tissue types, different freezing stages, and different puck morphologies to enhance generalization ability.

[0037] After the segmentation is completed, the ice hockey puck boundaries are smoothed and topologically corrected to ensure the continuity and rationality of the boundaries.

[0038] Subsequently, as Figure 3 As shown, the spatiotemporal evolution features of the ice puck are extracted based on the positional changes of the ice puck boundary at multiple time points.

[0039] The extracted hockey puck boundary features may include: Ice hockey boundary position; Ice puck volume; Ice hockey puck surface area; The ratio of the length to the short axis of an ice hockey puck; sphericity of the ice hockey puck; Ice hockey boundary curvature; Ice hockey boundary irregularity; Fractal dimension of ice hockey boundary.

[0040] The extracted directional extension features may include: Ice puck expansion radius R(θ,t) in different angular directions: Establish a polar coordinate system with the probe position as the origin, divide the 360° azimuth angle into N sectors (N≥8, preferably N=16 or 32), and calculate the ice puck expansion radius R(θ) in each sector direction for each time t. i , t); Extended direction offset: Define the centroid offset vector d(t) = C(t) - P, where C(t) is the coordinate of the puck's centroid at time t, P is the coordinate of the probe position, the offset direction is the direction of d(t), and the offset amount is |d(t)|. Symmetry index in the left-right, front-back, or up-down directions: For a left-right axis of symmetry (such as the x-axis), define the symmetry index S. x (t) = 1 - Σ|R(θi ,t) - R(π-θ i ,t)| / ΣR(θ i ,t), with a value range of [0,1], and the closer it is to 1, the more symmetrical it is; Difference in each direction: Define difference D(t) = σ[R(θ)] i ,t)] / μ[R(θ i ,t)], where σ is the standard deviation and μ is the mean. The larger D(t) is, the more uneven the expansion in each direction.

[0041] The extracted extended dynamic features may include: The velocity of the ice puck's expansion is: v(t) = [R(t) - R(t-Δt)] / Δt, which can be used to calculate the velocity of change of the total volume and the velocity of change of the radius in each direction. Extended acceleration of a hockey puck: a(t) = [v(t) - v(t-Δt)] / Δt; Rate of change of puck expansion velocity; Time required to reach the preset radius: for each direction θ i Record R(θ) i The moment when the preset threshold R0 is first reached (θ) i This time reflects the speed of the freezing response in that direction; Steady-state fluctuation characteristics during the freezing maintenance phase: During the freezing maintenance phase (when the ice puck volume change rate is below a preset threshold), calculate the coefficient of variation CV = σ / μ for the volume or boundary position, and the fluctuation frequency characteristics (analyze the dominant frequency component through Fourier transform). The retraction speed of the ice puck after cooling is stopped or during the switching phase: v_retract = -[R(t) - R(t-Δt)] / Δt.

[0042] After obtaining the spatiotemporal evolution characteristics of the ice hockey puck, the data processing and tissue state recognition unit inputs them together with the freezing process parameters into the tissue state recognition model to identify the tissue state parameters of the target tissue.

[0043] In this embodiment, the first tissue state parameter is used to characterize the degree of local perfusion influence. This parameter reflects the extent to which blood flow in a certain area affects the expansion of the ice puck, directional deviation, boundary deformation, or steady-state fluctuations. The second tissue state parameter is used to characterize local heat conduction capacity. This parameter reflects the ability of the tissue body in a certain area to transfer and diffuse cold.

[0044] For example: When the puck's expansion velocity in a certain direction is consistently lower than in other directions, and the steady-state fluctuation in the corresponding direction is significant, it can be determined that the first tissue state parameter in that direction is higher, indicating a stronger impact on local blood perfusion. When the boundary of a hockey puck in a certain area is irregular, the boundary curvature changes greatly, and the expansion speed shows obvious nonlinear changes, it can be determined that the second tissue state parameter of the area is abnormal, indicating that there is a difference in local heat conduction capacity. When the hockey puck boundary shows a continuous deflection or stagnation in a certain direction at multiple moments, that direction can be marked as the direction of vascular impact risk.

[0045] During the state recognition process, prior anatomical information of the target tissue can also be introduced, such as vascular distribution maps and tissue type distribution maps obtained from preoperative enhanced CT or MRI, to improve the accuracy of state recognition and the precision of anatomical localization.

[0046] The output module can output the first and second tissue state parameters in the form of two-dimensional distribution maps, three-dimensional distribution maps, zonal risk level maps, or directional risk indicators. The spatial resolution of the first and second tissue state parameters can be achieved through model interpolation or super-resolution reconstruction, resulting in a higher spatial resolution than the original time-series medical image data, thereby providing more refined spatial distribution information of tissue state.

[0047] like Figure 5 As shown, the output results may include: Localized perfusion heterogeneity distribution map; Local heat conduction heterogeneity distribution diagram; Anomaly zone in freezing response; Low-temperature risk areas; Blood vessels influence the direction of risk.

[0048] Example 2 like Figure 4 As shown, the organizational status identification model in step S3 can be implemented in the following way.

[0049] The spatiotemporal evolution characteristics of hockey pucks, freezing process parameters, and corresponding tissue state labels from historical cases were used together to train the model. During model training, heat transfer physics constraints were introduced to ensure that the model satisfies the basic heat transfer laws in cryoablation. After training, the model was used to predict the first and second tissue state parameters of the current case.

[0050] Physical constraints can be introduced in the following ways: Add a physical residual term to the loss function to penalize predictions that violate the heat transfer laws; Embedding a Physical Information Neural Network (PINN) structure into the model architecture allows the heat transfer equation to be used as a hard or soft constraint. By using physics-inspired feature engineering, dimensionless numbers related to heat transfer (such as Péclet number and Biot number) are used as model inputs.

[0051] Tissue status labels for training data can be obtained through post-hoc histopathological perfusion assessments (such as CD31 immunohistochemical staining density) and / or thermal conductivity measurements, for supervised or semi-supervised training.

[0052] The above describes a machine learning model based on physical constraints. Of course, tissue state recognition models can also be inversion models based on biological heat transfer constraints, finite element inverse problem solving models, ensemble Kalman filter models, or effective combinations of their outputs.

[0053] Example 3 In some embodiments, the present invention not only outputs a first tissue state parameter and a second tissue state parameter, but also further marks abnormal regions or risk regions based on the parameters.

[0054] For example: 1. When the first tissue state parameter is higher than the preset threshold, the corresponding area is marked as the perfusion enhancement area or the vascular influence area; 2. When the second tissue state parameter is lower or higher than the preset threshold, the corresponding area is marked as a heat conduction obstruction area or a heat conduction abnormal area; 3. When the first tissue state parameter and the second tissue state parameter jointly indicate that the freezing of the area is difficult to fully reach the preset threshold, the area is marked as a potential under-cooling risk area or a potential ablation blind area; 4. When the difference in expansion radius and expansion speed in different directions exceeds a preset threshold, the corresponding direction will be marked as an abnormal expansion direction.

[0055] The marking results can be overlaid on the original image or output as a separate layer, making it easier for doctors to intuitively identify risk areas.

[0056] Example 4 Although the present invention preferably uses CT image data to obtain time-series medical image data of ice hockey, the present invention is not limited to CT images.

[0057] In some implementations, the time-series medical imaging data may also be MRI imaging data, ultrasound imaging data, or a combination thereof. The hockey puck segmentation method and feature extraction method can be adjusted according to imaging characteristics under different imaging modalities. CT images: high spatial resolution (typically 0.5-1.25 mm), clear hockey puck boundaries, suitable for extracting precise boundary features and directional extension features; MRI imaging: good soft tissue contrast, can simultaneously observe edema and blood flow changes in the tissues surrounding the hockey puck, and is suitable for combining with functional imaging information; Ultrasound imaging: good real-time performance (frame rate up to 10-30 fps), no radiation, can capture the dynamic process of ice puck formation, and is suitable for extracting extended dynamic features.

[0058] Multimodal fusion strategy: After spatiotemporal registration of features extracted from two or more modalities, they are input into the same tissue state recognition model, or input into multiple sub-models and then fused to improve the robustness and accuracy of state recognition.

[0059] Example 5 like Figure 6 As shown, the present invention also provides a tissue state recognition system based on the spatiotemporal evolution characteristics of intraoperative ice hockey puck.

[0060] The system includes: The data acquisition module is used to acquire time-series medical imaging data and freezing process parameters; The feature extraction module is used to extract ice hockey puck boundary features, directional expansion features, and expansion dynamics features from time-series medical image data. The status recognition module is used to identify the first organizational status parameter and the second organizational status parameter through the organizational status recognition model; The output module is used to output the spatial distribution results of the tissue status and the marking results of abnormal areas or risk directions.

[0061] Specifically, the data acquisition module includes at least one of a CT acquisition unit, an MRI acquisition unit, and an ultrasound acquisition unit, as well as a process parameter acquisition unit for acquiring at least one of probe temperature, refrigerant flow rate, freezing duration, and pipeline pressure.

[0062] Specifically, the system further includes a feedback control module, used to generate cryoablation parameter adjustment suggestions based on the first tissue state parameter and / or the second tissue state parameter, the adjustment suggestions including: 1. When the first tissue state parameter in a certain direction is significantly higher than that in other directions, it indicates that there is strong vascular heat sink in that direction. It is recommended to adjust the probe position to concentrate the cooling effect on the area, or increase the probe coverage in that direction. 2. When the overall ice puck expansion speed is lower than expected and the second tissue state parameter is low, it indicates that the tissue has weak thermal conductivity. It is recommended to appropriately reduce the refrigerant flow rate to avoid overcooling of adjacent normal tissue, or extend the freezing time to ensure the melting effect. 3. When potential ablation blind spots or areas of undercooling risk are identified, it is recommended to extend the freezing duration or adopt an intermittent freezing strategy to take advantage of the reduced perfusion effect following vascular injury.

[0063] This feedback control module realizes a closed loop from state recognition to treatment strategy optimization, improving the accuracy and safety of cryoablation treatment.

[0064] The "spatiotemporal evolution characteristics of ice hockey" referred to in this specification are a set of characteristics formed by changes in the shape of ice hockey at multiple moments, including boundary characteristics, directional expansion characteristics, expansion dynamics characteristics, and combinations thereof.

[0065] It should be noted that the "first tissue state parameter" referred to in this specification refers to a parameter used to characterize the degree of influence of local blood perfusion on puck expansion.

[0066] It should be noted that the "second tissue state parameter" referred to in this specification refers to a parameter used to characterize the heat conduction capacity of a local tissue.

[0067] It should be noted that the "spatial distribution results" referred to in this specification refer to the results of location-related expression of the first organizational state parameter and / or the second organizational state parameter within the target organizational range, including two-dimensional distribution maps, three-dimensional distribution maps, zonal risk level maps, directional risk warnings, and combinations thereof.

[0068] It should be noted that "controlled freezing stimulation" as used in this manual refers to freezing applied to the target tissue under known and recordable freezing parameters (refrigerant type, flow rate, pressure, probe temperature, etc.).

[0069] It should be noted that the "spatiotemporal response signal" referred to in this specification refers to the set of morphological characteristics of the ice balls formed inside the target tissue under controlled freezing stimulation, which change with time and space.

[0070] It should be noted that the term "inversion" in this manual refers to the mathematical process of inferring the cause (organizational state parameters) of the result from the observation results (spatiotemporal evolution characteristics of ice hockey).

[0071] Of course, the above are just typical examples of the present invention. In addition, the present invention may have many other specific embodiments. All technical solutions formed by equivalent substitution or equivalent transformation fall within the scope of protection claimed by the present invention.

Claims

1. A method for identifying tissue state based on the spatiotemporal evolution characteristics of intraoperative hockey puck, characterized in that, The steps include the following: S1. During the cryoablation of the target tissue, acquire time-series medical image data containing the ice ball region, and acquire the cryoablation process parameters corresponding to the cryoablation process; S2. Extract the spatiotemporal evolution features of ice hockey based on the time-series medical image data. The spatiotemporal evolution features of ice hockey include at least two of the following: ice hockey boundary features, directional expansion features, and expansion dynamics features. S3. Based on the spatiotemporal evolution characteristics of the ice hockey puck and the freezing process parameters, identify the tissue state parameters of the target tissue. The tissue state parameters include at least a first tissue state parameter characterizing the degree of local perfusion influence and a second tissue state parameter characterizing the local heat conduction capacity. S4. Output the spatial distribution results of the first tissue state parameter and / or the second tissue state parameter.

2. The tissue state identification method according to claim 1, characterized in that, The freezing process parameters include at least one of refrigerant flow rate, freezing duration, probe temperature, and pipeline pressure; the ice puck boundary characteristics include at least one of ice puck boundary position, ice puck volume, ice puck surface area, ice puck major-minor axis ratio, sphericity, boundary curvature, boundary irregularity, and boundary fractal dimension; the directional expansion characteristics include at least one of ice puck expansion radius in different angular directions, expansion direction offset, symmetry index, and expansion difference in each direction; the expansion dynamic characteristics include at least one of ice puck expansion velocity, expansion acceleration, expansion velocity change rate, time required to reach the preset radius, steady-state maintenance fluctuation characteristics, and retraction velocity.

3. The tissue state identification method according to claim 1 or 2, characterized in that, When extracting the spatiotemporal evolution features of the ice hockey puck in step S2, the ice hockey puck boundary at multiple time points is determined based on the time-series medical image data, and the directional expansion features and expansion dynamic features are extracted based on the spatial changes of the ice hockey puck boundary at multiple time points; the ice hockey puck boundary is determined by segmentation based on CT value threshold, segmentation based on region growing, or segmentation based on deep learning.

4. The tissue state identification method according to claim 3, characterized in that, In step S3, the tissue state parameters are identified by the tissue state identification model. The tissue state identification model is used to establish the correspondence between the spatiotemporal evolution characteristics of the ice hockey puck and the first tissue state parameters and the second tissue state parameters. The tissue state identification model is an inversion model based on biological heat transfer constraints, a machine learning model based on physical constraints, a finite element inverse problem solving model, an ensemble Kalman filter model, or a combination thereof.

5. The tissue state identification method according to claim 4, characterized in that, The spatial distribution results include at least one of the following: local perfusion heterogeneity distribution map, local heat conduction heterogeneity distribution map, abnormal freezing response area, undercooling risk area, and vascular impact risk direction; the spatial distribution results output in step S4 are at least one of the following: two-dimensional distribution map, three-dimensional distribution map, zonal risk level map, or directional risk indication.

6. The tissue state identification method according to claim 5, characterized in that, Also includes: Based on the first tissue state parameter and the second tissue state parameter, abnormal expansion areas, perfusion enhancement areas, heat conduction obstruction areas, or potential ablation blind areas in the target tissue are marked and output; when the first tissue state parameter is higher than a preset threshold, the corresponding area is marked as a perfusion enhancement area or a vascular influence area; when the second tissue state parameter is lower than or higher than a preset threshold, the corresponding area is marked as a heat conduction obstruction area or a heat conduction abnormal area; when the first tissue state parameter and the second tissue state parameter jointly indicate that the freezing of the area is difficult to fully reach the preset threshold, the area is marked as a potential undercooling risk area or a potential ablation blind area.

7. The tissue state identification method according to claim 6, characterized in that, The time-series medical imaging data is CT imaging data, MRI imaging data, ultrasound imaging data, or a combination thereof; the acquisition time interval of the time-series medical imaging data in step S1 is dynamically adjusted according to the puck expansion speed. When the puck expansion speed exceeds a preset threshold, the acquisition time interval is shortened; when the puck expansion speed is below the preset threshold, the acquisition time interval is extended.

8. The tissue state identification method according to claim 7, characterized in that, In step S3, when identifying tissue state parameters, prior anatomical information of the target tissue is also introduced. The prior anatomical information includes at least one of a vascular distribution map and a tissue type distribution map. The spatial resolution of the first tissue state parameter and the second tissue state parameter is higher than the spatial resolution of the time-series medical image data, which is achieved through interpolation or super-resolution reconstruction of the tissue state identification model.

9. A tissue state recognition system based on the spatiotemporal evolution characteristics of intraoperative hockey puck, characterized in that, include: The data acquisition module is used to acquire time-series medical image data including the ice puck area and the corresponding freezing process parameters during the cryoablation of the target tissue; The feature extraction module is used to extract the spatiotemporal evolution features of ice hockey based on the time-series medical image data. The spatiotemporal evolution features of ice hockey include at least two of the following: ice hockey boundary features, directional expansion features, and expansion dynamics features. The state recognition module is used to identify the tissue state parameters of the target tissue based on the spatiotemporal evolution characteristics of the ice hockey puck and the freezing process parameters. The tissue state parameters include at least a first tissue state parameter characterizing the degree of local perfusion influence and a second tissue state parameter characterizing the local heat conduction capacity. The output module is used to output the spatial distribution results of the first tissue state parameter and / or the second tissue state parameter.

10. The organization status identification system according to claim 9, characterized in that, The output module is further used to output at least one of the following: local perfusion heterogeneity distribution map, local heat conduction heterogeneity distribution map, abnormal freezing response area, undercooling risk area, and vascular impact risk direction; the state recognition module integrates a tissue state recognition model, which is an inversion model based on biological heat transfer constraints, a machine learning model based on physical constraints, a finite element inverse problem solving model, an ensemble Kalman filter model, or a combination thereof; the data acquisition module includes at least one of CT acquisition unit, MRI acquisition unit, and ultrasound acquisition unit, as well as a process parameter acquisition unit for acquiring at least one of probe temperature, refrigerant flow rate, freezing duration, and pipeline pressure; the system also includes a feedback control module for generating cryoablation parameter adjustment suggestions based on the first tissue state parameter and / or the second tissue state parameter, the adjustment suggestions including at least one of probe position adjustment, refrigerant flow rate adjustment, and freezing duration adjustment.

Images