CN-121987342-A - Ultrasonic and CT image fusion percutaneous puncture real-time path planning system

CN121987342ACN 121987342 ACN121987342 ACN 121987342ACN-121987342-A

Abstract

The invention relates to a percutaneous puncture real-time path planning system for fusion of ultrasonic and CT images, which relates to the technical field of medical image navigation, and the method comprises the steps of extracting bone multi-level characteristics in preoperative CT images and bone interface and acoustic characteristics in intraoperative real-time ultrasonic images, and carrying out unmarked real-time registration and fusion by using a multi-level matching engine, and calculating and displaying a real-time puncture path in a fused image space by integrating the path length, the obstacle avoidance distance and the tissue penetrability cost based on a dynamic confidence evaluation result. The invention realizes the high-precision image fusion without external marks, the adaptive guidance of the dynamic change of the tissue in the operation and the path planning of fusion biomechanical safety consideration.

Inventors

DENG XIANLIANG
WANG YAN
GAO YUAN
LONG ZHILI
DU HUIMIN

Assignees

电子科技大学

Dates

Publication Date: 20260508
Application Date: 20260308

Claims (9)

1. An ultrasound and CT image fused percutaneous puncture real-time path planning system, comprising: the image acquisition module is used for acquiring CT three-dimensional volume images before operation of a patient and real-time two-dimensional ultrasonic image sequences in operation; the feature extraction module is connected with the image acquisition module and is used for extracting a macroscopic three-dimensional geometric model and a mesoscopic attribute feature map of a bone from the CT three-dimensional volume image and extracting a bone interface contour feature and an acoustic attribute feature from the real-time two-dimensional ultrasonic image sequence; The registration fusion module is connected with the feature extraction module and internally provided with a multi-level matching engine, wherein the multi-level matching engine executes first-level matching and second-level matching; The first level matching carries out initial registration according to the skeleton interface contour features and the macroscopic three-dimensional geometric model, and a first real-time pose is output; The second level matching is used for carrying out refined matching on the acoustic attribute characteristics and the mesoscopic attribute characteristic map according to the first real-time pose, and outputting an optimized second real-time pose; the registration fusion module performs space fusion on the real-time two-dimensional ultrasonic image sequence and the CT three-dimensional volume image based on the second real-time pose; And the path planning module is connected with the registration fusion module and is used for calculating a puncture path from a designated needle insertion point on a real-time ultrasonic image plane to a preset target point in a CT three-dimensional space in the fused image space and displaying the puncture path in a superposition manner on the real-time two-dimensional ultrasonic image sequence.
2. The ultrasound and CT image fused percutaneous puncture real-time path planning system of claim 1, wherein the mesoscopic attribute profile is an acoustic impedance profile or an elastic modulus profile generated by physical model inversion of the gray value distribution of the CT three-dimensional volumetric image.
3. The percutaneous puncture real-time path planning system according to claim 1, wherein the second-level matching adopts a feature fusion method based on an attention mechanism, the method allocates a dynamic weight to each feature element in the acoustic attribute feature and mesoscopic attribute feature map, the dynamic weight is calculated according to the spatial coordinates and feature dimensions of the feature elements, and a matching result is determined by weighted feature similarity.
4. The percutaneous puncture real-time path planning system for ultrasound and CT image fusion according to claim 1, further comprising a confidence assessment module connected to the registration fusion module; The inputs of the confidence evaluation module are the residual error of the first-level matching, the similarity score of the second-level matching and the variation of the second real-time pose between continuous frames; the output of the confidence assessment module is a comprehensive confidence score.
5. The ultrasound and CT image fused percutaneous transluminal real-time path planning system of claim 4, wherein the path planning module controls a display state of the puncture path according to the integrated confidence score: When the comprehensive confidence score is greater than or equal to a first threshold, displaying a complete puncture path; when the integrated confidence score is smaller than the first threshold value and larger than or equal to the second threshold value, displaying the puncture path and synchronously displaying the precision warning mark, and when And stopping displaying the puncture path and generating a repositioning prompt when the comprehensive confidence score is smaller than the second threshold value.
6. The percutaneous puncture real-time path planning system for ultrasound and CT image fusion according to claim 4 or 5, further comprising a policy adjustment module connecting the confidence assessment module with the registration fusion module, wherein the policy adjustment module adjusts the iteration number or convergence threshold of the second-level matching in the multi-level matching engine according to the historical trend of the integrated confidence score.
7. The percutaneous puncture real-time path planning system for ultrasound and CT image fusion according to claim 1, wherein the operation of the path planning module to calculate the puncture path comprises: S1, receiving two-dimensional needle point coordinates defined on a current frame of the real-time two-dimensional ultrasonic image sequence; S2, mapping the two-dimensional needle insertion point coordinate to a coordinate system of the CT three-dimensional volume image based on the second real-time pose to obtain a three-dimensional needle insertion point coordinate; S3, calculating a three-dimensional straight line path between the three-dimensional needle point coordinate and the preset target point; And S4, back projecting the three-dimensional straight line path to an imaging plane of the real-time two-dimensional ultrasonic image sequence, and generating a two-dimensional guide line for superposition display.
8. The percutaneous puncture real-time path planning system for ultrasound and CT image fusion according to claim 7, wherein the path planning module further performs obstacle avoidance detection: predefining a three-dimensional space range of at least one dangerous area in the CT three-dimensional volume image; In step S3, calculating the shortest distance between the three-dimensional straight line path and each dangerous area space range in real time; And when any distance is smaller than a preset safety distance, carrying out highlighting warning on a projection area of the corresponding section of the two-dimensional guide line or the dangerous area on the ultrasonic image.
9. The percutaneous puncture real-time path planning system according to claim 1, wherein the feature extraction module calculates the average energy, the frequency spectrum center and the attenuation coefficient of the signal in the depth range corresponding to the bone interface contour feature when extracting the acoustic attribute feature from the real-time two-dimensional ultrasound image sequence, in particular from the original radio frequency signal of the ultrasound device.

Description

Ultrasonic and CT image fusion percutaneous puncture real-time path planning system Technical Field The invention relates to the technical field of medical image navigation, in particular to a percutaneous puncture real-time path planning system for ultrasonic and CT image fusion. Background In percutaneous aspiration procedures, such as nerve block, biopsy, or ablation treatment, image guidance is critical to improving the accuracy and safety of the procedure. Currently, clinical practice is mainly conducted by single-mode images such as ultrasound, computed Tomography (CT) or Magnetic Resonance Imaging (MRI). Ultrasonic imaging has real-time performance, but has limited image resolution, and can not clearly display deep complex anatomical structures, while CT or MRI can provide high-resolution static anatomical images, but can not reflect the dynamic interaction process of the puncture needle and tissues in real time. Therefore, operators need to complete multi-mode information fusion and path planning in the brain, and the operators are highly dependent on personal experience and space imagination, so that inaccurate puncture path planning, increased operation times and long learning curve are easy to cause. To overcome the limitations of single modality, the prior art attempts to fuse preoperative high resolution images with intraoperative real-time images. For example, chinese patent publication No. CN120918797a discloses a dynamic obstacle avoidance navigation method for nerve puncture based on multi-modal image fusion, which performs three-dimensional reconstruction and path planning by fusing preoperative static images and intra-operative ultrasound and using artificial intelligence to segment key structures. However, the method relies on positioning mark points stuck on the surface of a patient to realize a coordinate system in image fusion, the method is invasive, and the mark points can shift or fall off due to movement, sweating or disinfection of the patient, so that the registration accuracy and stability are affected. In addition, although the path planning considers obstacle avoidance, quantitative evaluation of an important biomechanical factor of tissue penetrability is lacking. For another example, the Chinese patent publication No. CN110025379A discloses a real-time navigation system and method for fusing an ultrasonic image and a CT image. The system utilizes an optical positioning sensor and a tracer to carry out space tracking on CT equipment, surgical instruments and ultrasonic probes, and realizes image fusion and instrument navigation. The technology has the defects that the technology depends on an external optical or electromagnetic positioning system, the system is complex to build, the cost is high, and the positioning precision is easily influenced by environmental shielding or electromagnetic interference of an operating room. Meanwhile, the image registration and navigation logic of the device are relatively solidified, and the self-adaptive optimization capability based on the operation process data is lacking. In summary, the prior art has the problems of dependence on in-vitro markers or external positioning equipment, insufficient adaptability to dynamic deformation of tissues in operation, lack of self-learning and optimizing capability of the system, insufficient consideration of biomechanical characteristics of the tissues in path planning and the like. Disclosure of Invention In order to overcome the above-mentioned drawbacks of the prior art, embodiments of the present invention provide a percutaneous puncture real-time path planning system for ultrasound and CT image fusion, so as to solve the problems of complexity and stability of the system caused by the dependence on external markers or external positioning devices in the prior art. In order to achieve the purpose, the invention provides the following technical scheme that the percutaneous puncture real-time path planning system for the fusion of ultrasonic and CT images comprises: The image acquisition module is used for acquiring CT three-dimensional volume images before operation of a patient and real-time two-dimensional ultrasonic image sequences in operation and providing an original multi-mode data source for subsequent processing; The feature extraction module is connected with the image acquisition module and is used for extracting a macroscopic three-dimensional geometric model and a mesoscopic attribute feature map of a bone from the CT three-dimensional volume image, and extracting bone interface contour features and acoustic attribute features from the real-time two-dimensional ultrasonic image sequence so as to convert image information of different modes into unified feature expression which can be matched in a computing way; The registration fusion module is connected with the feature extraction module and internally provided with a multi-level matching engine, wherein the multi-level matching e