Intelligent evaluation system after anterior cruciate ligament reconstruction based on artificial intelligence

Abstract

The invention discloses an anterior cruciate ligament reconstruction postoperative intelligent evaluation system based on artificial intelligence. The system comprises a data acquisition module used for collecting clinical information and MRI image data of a patient; the ligament graft maturity evaluation module is used for automatically measuring and calculating the maturity of the anterior cruciate ligament graft; the tendon-bone interface healing evaluation module is used for automatically measuring bone tunnel diameter changes and tunnel surrounding edema changes. Accurate evaluation and personalized suggestions are provided for postoperative recovery of anterior cruciate ligament reconstruction patients.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image processing, and in particular to an intelligent evaluation system for anterior cruciate ligament reconstruction surgery based on artificial intelligence. Background Art

[0002] The anterior cruciate ligament (ACL) is one of the most important ligaments in the knee joint, primarily responsible for limiting the anterior translation of the tibia relative to the femur and rotational stability. ACL injuries are very common in sports trauma, especially in high-intensity sports. ACL reconstruction (ACLR) is the main treatment to restore knee stability and function.

[0003] After ACL reconstruction, assessment of graft maturity and healing of the tendon-bone interface is essential to guide rehabilitation training. The graft undergoes a series of biological processes after surgery, including bone integration and ligamentization, until its properties approach those of the native ligament. This process is divided into several stages: the early stage (approximately 4 weeks), the proliferation stage (4 to 12 weeks), and the final ligamentization stage (more than 1 year). Graft maturity and healing of the tendon-bone interface directly affect the development of the rehabilitation plan and the time to return to sports. Returning to sports too early may increase the risk of graft failure, so the rehabilitation plan must take into account the biological characteristics of the graft at different stages and avoid excessive loading.

[0004] Traditional assessment methods mainly rely on the clinical observation experience of clinicians and manual measurement of imaging. This method has many limitations, including the subjectivity of the assessment, the complexity of the operation, and the lack of accuracy of the assessment results, which can easily lead to misjudgment or delay the rehabilitation process. This subjective assessment method not only relies on the doctor's experience level, but is also affected by human factors, resulting in inaccurate guidance during postoperative rehabilitation. Summary of the invention

[0005] In order to overcome the technical problems of strong subjectivity, complex operation and lack of precise rehabilitation guidance in the postoperative evaluation of anterior cruciate ligament reconstruction, the present invention proposes an intelligent postoperative evaluation system for anterior cruciate ligament reconstruction based on artificial intelligence, which provides more accurate and reliable postoperative evaluation results, thereby better guiding the patient's rehabilitation process, optimizing the rehabilitation plan and reducing the risk of graft failure.

[0006] To achieve the above-mentioned purpose, the present invention provides an intelligent evaluation system for anterior cruciate ligament reconstruction based on artificial intelligence, comprising a data acquisition module, a ligament graft maturity evaluation module and a tendon-bone interface healing evaluation module;

[0007] The data acquisition module is used to collect clinical information and MRI image data of patients;

[0008] The ligament graft maturity assessment module is used to automatically measure and calculate the maturity of the anterior cruciate ligament graft;

[0009] The tendon-bone interface healing assessment module is used to automatically measure changes in bone tunnel diameter and edema around the tunnel;

[0010] The data acquisition module is connected to the ligament graft maturity assessment module and the tendon-bone interface healing assessment module respectively.

[0011] Preferably, the data acquisition module includes:

[0012] Clinical information collection unit: used to retrieve basic patient information and surgical information;

[0013] MRI data acquisition unit: used to retrieve the patient's follow-up knee joint MRI data.

[0014] Preferably, the ligament graft maturity assessment module comprises:

[0015] A ligament graft and background segmentation unit: used for segmenting the anterior cruciate ligament graft and the background in the knee joint MRI data by an artificial intelligence automatic recognition method;

[0016] A signal measuring unit: used to measure the signal strength of the background and the anterior cruciate ligament graft respectively;

[0017] A signal-to-noise ratio calculation unit is used to calculate the signal-to-noise ratio of the anterior cruciate ligament graft at various locations, and to evaluate the maturity of the anterior cruciate ligament graft based on the signal-to-noise ratio.

[0018] Preferably, the process of the signal measuring unit measuring the background signal includes:

[0019] Select the sagittal position of T2 fat-suppressed MRI sequence of knee joint;

[0020] Automatically selecting a circle with a diameter of a first preset distance from the segmented background through artificial intelligence and measuring its signal strength as the background signal;

[0021] The method for measuring the signal strength of the anterior cruciate ligament graft is:

[0022] The anterior cruciate ligament graft is segmented based on a convolutional neural network to obtain the distal end, middle end and proximal end of the anterior cruciate ligament graft, and circles with a diameter of a first preset distance are automatically selected at the distal end, middle end and proximal end of the anterior cruciate ligament graft and their signal strengths are measured.

[0023] Preferably, the process of calculating the signal-to-noise ratio of the anterior cruciate ligament graft at different locations by the signal-to-noise ratio calculation unit includes:

[0024] The signal at any part of the distal, middle or proximal end of the anterior cruciate ligament graft is subtracted from the contrast signal and divided by the background signal to obtain a signal-to-noise ratio;

[0025] Among them, a circle with a diameter of a first preset distance is automatically selected at a preset position and the signal strength is measured, and the measurement result is used as the comparison signal. The preset position includes quadriceps tendon tissue and posterior cruciate ligament tissue.

[0026] Preferably, the tendon-bone interface healing assessment module comprises:

[0027] Bone tunnel change detection unit: used to detect bone tunnel changes;

[0028] Edema change detection unit around the tunnel: used to detect edema changes around the bone tunnel.

[0029] Preferably, the process of the bone tunnel change detection unit detecting the bone tunnel change includes:

[0030] Axial and sagittal views of T2 fat-suppressed MRI sequence of knee joint were selected;

[0031] Determining the entry level of the ligament graft at the distal femur and the proximal tibia according to the sagittal position of the T2 fat-suppressed sequence, and locating the axial position of the femoral tunnel and the tibial tunnel;

[0032] Based on a convolutional neural network, the tunnel diameters of the femoral tunnel and the tibial tunnel are measured at the T2 axial level respectively;

[0033] The intraoperative tunnel diameter is retrieved through a data acquisition module, and the ratio of the tunnel diameter to the intraoperative tunnel diameter is measured to determine the change of the bone tunnel.

[0034] Preferably, the process of the tunnel periphery edema change detection unit detecting the edema change around the bone tunnel comprises:

[0035] In the T2 fat suppression sequence at the same level of the femoral tunnel and the tibial tunnel, circles with a diameter of a second preset distance are drawn with the centers of the femoral tunnel and the tibial tunnel as the centers, and the signal intensity is measured and used as the signal intensity around the bone tunnel;

[0036] Draw a circle with a diameter of the second preset distance in the middle femoral region without bone edema, measure the bone marrow signal intensity, and use the bone marrow signal intensity as a reference value;

[0037] The change in bone tunnel edema is determined by the ratio of the signal around the bone tunnel to the reference value.

[0038] Preferably, the system further comprises a rehabilitation suggestion module, wherein the rehabilitation suggestion module comprises:

[0039] Rehabilitation database: used to conduct literature search to study the relationship between the maturity of ligament grafts and rehabilitation training after anterior cruciate ligament reconstruction, rehabilitation methods and movements, and rehabilitation guidelines;

[0040] Early warning unit: used to provide rehabilitation strategies for patients based on the evaluation results of ligament graft maturity and tendon-bone interface healing in combination with the rehabilitation database.

[0041] Preferably, the rehabilitation database can also optimize and update the database content in combination with the patient's rehabilitation results and evaluation results.

[0042] Compared with the prior art, the present invention has the following advantages and technical effects:

[0043] The present invention automatically collects and analyzes patient data and uses a convolutional neural network model to objectively evaluate the maturity of the graft and the healing of the tendon-bone interface. This method reduces the impact of human error and provides more accurate and reliable postoperative evaluation results, thereby better guiding the patient's rehabilitation process, optimizing the rehabilitation plan and reducing the risk of graft failure. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0045] Figure 1 This is a schematic diagram of the structure of an intelligent evaluation system for anterior cruciate ligament reconstruction based on artificial intelligence according to an embodiment of the present invention;

[0046] Figure 2 This is a schematic diagram of ligament graft maturity assessment according to an embodiment of the present invention;

[0047] Figure 3 This is a schematic diagram of tendon-bone healing assessment according to an embodiment of the present invention;

[0048] Figure 4 This is a conceptual diagram of a convolutional neural network algorithm and data processing according to an embodiment of the present invention. DETAILED DESCRIPTION

[0049] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0050] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0051] The present invention proposes an intelligent evaluation system for anterior cruciate ligament reconstruction based on artificial intelligence. Figure 1 , including data acquisition module, ligament graft maturity assessment module and tendon-bone interface healing assessment module;

[0052] The data acquisition module is used to collect the patient's clinical information and MRI image data;

[0053] The ligament graft maturity assessment module is used to automatically measure and calculate the maturity of the anterior cruciate ligament graft;

[0054] The tendon-bone interface healing assessment module is used to automatically measure the changes in bone tunnel diameter and edema around the tunnel;

[0055] The data acquisition module is connected with the ligament graft maturity assessment module and the tendon-bone interface healing assessment module respectively.

[0056] The present invention uses a convolutional neural network model to objectively evaluate the maturity of the graft and the healing of the tendon-bone interface by automatically collecting and analyzing patient data. This method reduces the impact of human error and provides more accurate and reliable postoperative evaluation results, thereby better guiding the patient's rehabilitation process, optimizing the rehabilitation plan and reducing the risk of graft failure.

[0057] Furthermore, the data acquisition module includes:

[0058] Clinical information collection unit: used to retrieve basic patient information and surgical information;

[0059] MRI data acquisition unit: used to retrieve the patient's follow-up knee joint MRI data.

[0060] Specifically, the data acquisition module is used to retrieve basic patient information, surgical information, and follow-up knee joint MRI data from the hospital's case system and imaging system;

[0061] In this embodiment, the knee joint MRI data includes but is not limited to T2 fat-suppressed sagittal and coronal views, PD fat-suppressed sequence sagittal and coronal views, and the like.

[0062] Furthermore, the ligament graft maturity assessment module includes:

[0063] A ligament graft and background segmentation unit: used for segmenting the anterior cruciate ligament graft and the background in the knee joint MRI data by an artificial intelligence automatic recognition method;

[0064] A signal measuring unit: used to measure the signal strength of the background and the anterior cruciate ligament graft respectively;

[0065] A signal-to-noise ratio calculation unit is used to calculate the signal-to-noise ratio of the anterior cruciate ligament graft at various locations, and to evaluate the maturity of the anterior cruciate ligament graft based on the signal-to-noise ratio.

[0066] Specifically, the method for measuring the background signal is:

[0067] Select the sagittal position of T2 fat-suppressed MRI sequence of knee joint;

[0068] Automatically selecting a circle with a diameter of a first preset distance from the segmented background through artificial intelligence and measuring its signal strength as a background signal;

[0069] In this embodiment, the first preset distance is 3 mm, and the background signal automatically selects a circle with a diameter of 3 mm through artificial intelligence and measures its signal strength.

[0070] The signal intensity of the anterior cruciate ligament graft is measured by:

[0071] The anterior cruciate ligament graft is segmented based on a convolutional neural network to obtain the distal end, middle end and proximal end of the anterior cruciate ligament graft, and a circle with a diameter of a first preset distance is automatically selected at the distal end, middle end and proximal end of the anterior cruciate ligament graft and its signal strength is measured.

[0072] Furthermore, the signal-to-noise ratio of the anterior cruciate ligament graft at different locations was calculated, e.g. Figure 2 ,include:

[0073] The signal-to-noise ratio was obtained by subtracting the contrast signal from the signal at any site of the distal, middle, or proximal part of the anterior cruciate ligament graft and dividing it by the background signal;

[0074] Among them, a circle with a diameter of a first preset distance is automatically selected at a preset position and the signal strength is measured, and the measurement result is used as the comparison signal. The preset position includes quadriceps tendon tissue and posterior cruciate ligament tissue.

[0075] Specifically, the anterior cruciate ligament graft signal is automatically identified and segmented by artificial intelligence, and a circle with a diameter of 3 mm is automatically selected at the distal, middle and proximal ends of the ligament graft and its signal intensity (SI) is measured.

[0076] The contrast signal is automatically identified through artificial intelligence in the quadriceps tendon, and a circle with a diameter of 3 mm is automatically selected at the distal end of the quadriceps tendon and its signal strength is measured as the contrast signal; the contrast signal source includes but is not limited to the quadriceps tendon, posterior cruciate ligament and other normal tendon tissues.

[0077] In this embodiment, sagittal images of the knee joint MRI T2 fat-suppressed sequence are used;

[0078] Select a layer where the ligament graft and the distal end of the quadriceps tendon can be fully seen; the anterior cruciate ligament graft is automatically segmented into the distal, middle and proximal ends by a convolutional neural network, and marked with a circle of 3 mm in diameter, marked as L1, L2 and L3 respectively; the convolutional neural network training process is as follows Figure 4 As shown;

[0079] The distal end of the quadriceps tendon was automatically identified by convolutional neural network and annotated with a circle of 3 mm in diameter, labeled as C1;

[0080] The background is identified by a convolutional neural network and marked with a circle with a diameter of 3 mm, marked as B1;

[0081] The signal measurement unit calculates the signal intensity (SI) in the above-mentioned marked areas respectively;

[0082] The calculation method of the signal-to-noise ratio of the ligament graft is (taking the distal end as an example): distal ligament graft SNQ = (ligament graft distal SI - quadriceps tendon distal SI) / background SI.

[0083] In this embodiment, the artificial intelligence methods for automatically detecting anterior cruciate ligament grafts, quadriceps tendons and background are all based on convolutional neural networks, including but not limited to convolutional neural networks such as YOLO and U-Net.

[0084] Furthermore, the tendon-bone interface healing assessment module includes:

[0085] Bone tunnel change detection unit: used to detect bone tunnel changes;

[0086] Edema change detection unit around the tunnel: used to detect edema changes around the bone tunnel.

[0087] Specifically, bone tunnel changes are detected, e.g. Figure 3 ,include:

[0088] Axial and sagittal views of T2 fat-suppressed MRI sequence of knee joint were selected;

[0089] Determining the entry level of the ligament graft at the distal femur and the proximal tibia according to the sagittal position of the T2 fat-suppressed sequence, and locating the axial position of the femoral tunnel and the tibial tunnel;

[0090] Based on a convolutional neural network, the tunnel diameters of the femoral tunnel and the tibial tunnel are measured at the T2 axial level respectively;

[0091] The intraoperative tunnel diameter is retrieved through a data acquisition module, and the ratio of the tunnel diameter to the intraoperative tunnel diameter is measured to determine the change of the bone tunnel.

[0092] Detect edema changes around the bone tunnel, including:

[0093] In a T2 fat-suppressed sequence at the same level of the femoral tunnel and the tibial tunnel, a circle with a diameter of a second preset distance is drawn with the center of the femoral tunnel and the tibial tunnel as the center, and the signal intensity is measured and used as the bone tunnel surrounding signal;

[0094] Draw a circle with a diameter of the second preset distance in the middle femoral region without bone edema, measure the bone marrow signal intensity, and use the bone marrow signal intensity as a reference value;

[0095] The change in bone tunnel edema is determined by the ratio of the signal around the bone tunnel to the reference value.

[0096] Specifically, in this embodiment, the second preset distance is 10 mm.

[0097] In this embodiment, a convolutional neural network is introduced to automatically process the ligament graft maturity assessment and tendon-bone interface healing assessment; since the tasks involved in this task are all basic deep learning model training methods, no special personalized instructions are given. The model training method is: first, the data is labeled, and then the image is preprocessed as necessary, including resizing, normalization, etc., and then the model is trained by selecting the corresponding convolutional neural network, and finally the performance of the model is evaluated by indicators such as mean absolute error and mean square error.

[0098] Furthermore, the system also includes a rehabilitation suggestion module, which includes:

[0099] Rehabilitation database: used to conduct literature search to study the relationship between the maturity of ligament grafts and rehabilitation training after anterior cruciate ligament reconstruction, rehabilitation methods and movements, and rehabilitation guidelines;

[0100] Early warning unit: used to provide rehabilitation strategies for patients based on the evaluation results of ligament graft maturity and tendon-bone interface healing in combination with the rehabilitation database.

[0101] The rehabilitation database can also optimize and update the database content based on the patient's rehabilitation results and evaluation results.

[0102] Specifically, the workflow of the rehabilitation advice module is as follows:

[0103] First, in the process of building the rehabilitation database, a systematic literature search was conducted to conduct in-depth research on the relationship between the maturity of ligament grafts after anterior cruciate ligament reconstruction and rehabilitation training, current rehabilitation methods and movements, and domestically and internationally recognized rehabilitation guidelines;

[0104] The early warning unit provides patients with personalized rehabilitation strategies based on the assessment results of ligament graft maturity and tendon-bone interface healing combined with the rehabilitation database;

[0105] The rehabilitation database can also optimize and update the database content based on the patient's rehabilitation results and evaluation results.

[0106] The present invention is based on artificial intelligence and uses convolutional neural networks to provide objective and accurate assessments of graft maturity and tendon-bone interface healing, effectively formulate rehabilitation plans, improve the accuracy and consistency of assessments, and provide personalized rehabilitation advice, thereby reducing the burden on doctors and optimizing the rehabilitation process.

[0107] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. An intelligent assessment system for anterior cruciate ligament reconstruction based on artificial intelligence, characterized in that: It includes data acquisition module, ligament graft maturity assessment module and tendon-bone interface healing assessment module; The data acquisition module is used to collect clinical information and MRI image data of patients; The ligament graft maturity assessment module is used to automatically measure and calculate the maturity of the anterior cruciate ligament graft; The tendon-bone interface healing assessment module is used to automatically measure changes in bone tunnel diameter and edema around the tunnel; The data acquisition module is connected to the ligament graft maturity assessment module and the tendon-bone interface healing assessment module respectively.

2. The intelligent assessment system for anterior cruciate ligament reconstruction based on artificial intelligence according to claim 1, characterized in that: The data acquisition module comprises: Clinical information collection unit: used to retrieve basic patient information and surgical information; MRI data acquisition unit: used to retrieve the patient's follow-up knee joint MRI data.

3. The intelligent assessment system for anterior cruciate ligament reconstruction based on artificial intelligence according to claim 2, characterized in that: The ligament graft maturity assessment module includes: A ligament graft and background segmentation unit: used for segmenting the anterior cruciate ligament graft and the background in the knee joint MRI data by an artificial intelligence automatic recognition method; A signal measuring unit: used to measure the signal strength of the background and the anterior cruciate ligament graft respectively; A signal-to-noise ratio calculation unit is used to calculate the signal-to-noise ratio of the anterior cruciate ligament graft at various locations, and to evaluate the maturity of the anterior cruciate ligament graft based on the signal-to-noise ratio.

4. The intelligent assessment system for anterior cruciate ligament reconstruction based on artificial intelligence according to claim 3, characterized in that: The process of measuring the background signal by the signal measuring unit includes: Select the sagittal position of T2 fat-suppressed MRI sequence of knee joint; Automatically selecting a circle with a diameter of a first preset distance from the segmented background through artificial intelligence and measuring its signal strength as the background signal; The method for measuring the signal strength of the anterior cruciate ligament graft is: The anterior cruciate ligament graft is segmented based on a convolutional neural network to obtain the distal end, middle end and proximal end of the anterior cruciate ligament graft, and circles with a diameter of a first preset distance are automatically selected at the distal end, middle end and proximal end of the anterior cruciate ligament graft and their signal strengths are measured.

5. The intelligent assessment system for anterior cruciate ligament reconstruction based on artificial intelligence according to claim 4, characterized in that: The process of the signal-to-noise ratio calculation unit calculating the signal-to-noise ratio of the anterior cruciate ligament graft at different locations includes: The signal at any part of the distal, middle or proximal end of the anterior cruciate ligament graft is subtracted from the contrast signal and divided by the background signal to obtain a signal-to-noise ratio; Among them, a circle with a diameter of a first preset distance is automatically selected at a preset position and the signal strength is measured, and the measurement result is used as the comparison signal. The preset position includes quadriceps tendon tissue and posterior cruciate ligament tissue.

6. The intelligent assessment system for anterior cruciate ligament reconstruction based on artificial intelligence according to claim 1, characterized in that: The tendon-bone interface healing assessment module includes: Bone tunnel change detection unit: used to detect bone tunnel changes; Edema change detection unit around the tunnel: used to detect edema changes around the bone tunnel.

7. The intelligent assessment system for anterior cruciate ligament reconstruction based on artificial intelligence according to claim 6, characterized in that: The process of the bone tunnel change detection unit detecting the bone tunnel change includes: Axial and sagittal views of T2 fat-suppressed MRI sequence of knee joint were selected; Determining the entry level of the ligament graft at the distal femur and the proximal tibia according to the sagittal position of the T2 fat-suppressed sequence, and locating the axial position of the femoral tunnel and the tibial tunnel; Based on a convolutional neural network, the tunnel diameters of the femoral tunnel and the tibial tunnel are measured at the T2 axial level respectively; The intraoperative tunnel diameter is retrieved through a data acquisition module, and the ratio of the tunnel diameter to the intraoperative tunnel diameter is measured to determine the change of the bone tunnel.

8. The intelligent assessment system for anterior cruciate ligament reconstruction based on artificial intelligence according to claim 7, characterized in that: The process of the tunnel periphery edema change detection unit detecting the edema change around the bone tunnel comprises: In the T2 fat suppression sequence at the same level of the femoral tunnel and the tibial tunnel, circles with a diameter of a second preset distance are drawn with the centers of the femoral tunnel and the tibial tunnel as the centers, and the signal intensity is measured and used as the signal intensity around the bone tunnel; Draw a circle with a diameter of the second preset distance in the middle femoral region without bone edema, measure the bone marrow signal intensity, and use the bone marrow signal intensity as a reference value; The change in bone tunnel edema is determined by the ratio of the signal around the bone tunnel to the reference value.

9. The intelligent assessment system for anterior cruciate ligament reconstruction based on artificial intelligence according to claim 1, characterized in that: The system further comprises a rehabilitation suggestion module, wherein the rehabilitation suggestion module comprises: Rehabilitation database: used to conduct literature search to study the relationship between the maturity of ligament grafts and rehabilitation training after anterior cruciate ligament reconstruction, rehabilitation methods and movements, and rehabilitation guidelines; Early warning unit: used to provide rehabilitation strategies for patients based on the evaluation results of ligament graft maturity and tendon-bone interface healing in combination with the rehabilitation database.

10. The intelligent assessment system for anterior cruciate ligament reconstruction based on artificial intelligence according to claim 9, characterized in that: The rehabilitation database can also optimize and update the database content in combination with the patient's rehabilitation results and evaluation results.

Images