Ultrasound contrast-enhanced quantitative analysis method, system and computer-readable storage medium for dynamic follow-up of tumor local radiotherapy and chemotherapy effects

Abstract

The present invention provides an ultrasound contrast quantitative analysis method, system and computer-readable medium for dynamic follow-up of the effects of local radiotherapy and chemotherapy on tumors, and relates to the field of medical imaging technology. The ultrasound contrast quantitative analysis method described in the present invention first collects continuous dynamic contrast images starting from the injection of ultrasound contrast agent, performs lesion annotation based on respiratory dynamic compensation after preprocessing the dynamic contrast images, and finally obtains the time-intensity curve TIC by fitting calculation. Preprocessing eliminates errors caused by equipment, human factors and other factors, and the lesion annotation based on respiratory dynamic compensation offsets the interference of the patient's respiratory movement on the image. Finally, the TIC curve is obtained by fitting calculation, which effectively improves the accuracy and repeatability of quantitative analysis and provides a basis for non-invasive and sensitive evaluation of the effects of local radiotherapy and chemotherapy on tumors.

Description

Technical Field

[0001] The present invention relates to the field of medical imaging technology, and in particular to an ultrasound contrast quantitative analysis method, system and computer-readable storage medium for dynamically following up the effects of local radiotherapy and chemotherapy on tumors. Background Art

[0002] Radiotherapy (hereinafter referred to as "radiotherapy") and chemotherapy (hereinafter referred to as "chemotherapy") are the main treatment methods for patients with advanced solid tumors, which can increase the local control rate of tumors. Timely evaluation of treatment effects and adjustment of treatment plans during treatment can help prolong patient survival and improve the quality of life. However, radiotherapy and chemotherapy are adjuvant or local palliative treatments, and their efficacy is difficult to accurately evaluate. At present, the most commonly used imaging method for evaluating the efficacy of radiotherapy and chemotherapy for solid tumors is the RECIST standard based on CT / MRI. However, because CT is radioactive, it cannot accurately show the size, blood supply, and relationship with adjacent blood vessels of solid tumors after radiotherapy and chemotherapy, which may underestimate the effect of radiotherapy. MRI is time-consuming and cumbersome to operate, so it is difficult to become a routine method for evaluating radiotherapy and chemotherapy for solid tumors. In clinical practice, after radiotherapy and chemotherapy for solid tumors, the internal microcirculation blood perfusion is often reduced first, followed by the appearance of tumor size, while the RECIST standard can only evaluate the efficacy based on the size of the tumor. Therefore, it is very important to explore an imaging method that can be sensitive, accurate, non-invasive, and quantitatively follow up the efficacy of cancer before and after radiotherapy and chemotherapy.

[0003] Ultrasound contrast imaging is an imaging technique that can display the microcirculatory blood perfusion in tumors in real time. Its contrast phase can be divided into arterial phase, venous phase, delayed phase and late vascular phase. Ultrasound contrast quantitative analysis technology is a new method to objectively evaluate the intensity of contrast agents and microcirculatory blood perfusion. By dynamically analyzing the ultrasound contrast enhancement-dissipation video in the lesion, a time-intensity curve (TIC) that objectively reflects the ultrasound contrast enhancement and disappearance performance in the region of interest (ROI) is obtained. By analyzing the quantitative / semi-quantitative characteristics of the TIC curve, such as peak intensity (PE), rise time (RT), time to peak (TTP), mean transit time (mTT) and area under the TIC curve (AUC), the changes in microcirculatory blood perfusion in the tumor before and after tumor treatment can be sensitively and accurately quantified. However, the current interpretation of ultrasound contrast images in ultrasound contrast quantitative analysis is affected by the doctor's personal experience, differences in instruments and processing software from different manufacturers, differences in contrast agent configuration and injection, differences in individual patient conditions, and differences in ROI selection, making ultrasound contrast quantitative analysis lack accuracy and repeatability.

[0004] Chinese patent CN110969618A discloses a method for quantitative analysis of liver tumor angiogenesis based on dynamic ultrasound contrast-enhanced imaging based on liver tumor angiogenesis patterns. This method uses a twin convolutional network with spatial feature recalibration to track the lesion area to eliminate respiratory motion interference, extract pixel-level perfusion patterns, combine perfusion phases to quantify pixel-level perfusion pattern differences and generate node graphs, characterize intratumor perfusion differences based on nodes, and quantitatively analyze and extract quantitative features that reflect the heterogeneity and infiltrativeness of intratumor angiogenesis. Although this prior art has partially optimized the quantitative analysis of ultrasound contrast-enhanced imaging, its quantitative analysis results are mainly expressed through cluster analysis, which enhances the visualization effect of heterogeneity analysis, but still cannot solve the defects of poor accuracy and repeatability of ultrasound contrast-enhanced imaging quantitative analysis. Summary of the invention

[0005] In order to improve the accuracy and repeatability of ultrasound contrast quantitative analysis methods, the present invention provides a new ultrasound contrast quantitative analysis method for dynamic follow-up of tumor local radiotherapy and chemotherapy effects. Preprocessing eliminates errors caused by equipment, human factors and other factors, and lesion annotation based on respiratory dynamic compensation offsets the interference of patient respiratory movement on the image, and finally obtains the TIC curve through fitting calculation. By comparing the changes in TIC quantitative parameters in the lesion before and after radiotherapy and chemotherapy, a highly accurate, repeatable, non-invasive and sensitive tumor local radiotherapy and chemotherapy effect evaluation method can be obtained.

[0006] In order to achieve the above-mentioned object of the invention, the present invention provides a method for quantitative analysis of ultrasound contrast imaging for dynamic follow-up of the effect of local radiotherapy and chemotherapy on cancer, comprising the following steps:

[0007] (1) Acquisition of continuous dynamic angiography images;

[0008] (2) preprocessing at least two of the pixel size, pixel intensity, signal noise and image frame rate of the dynamic contrast image obtained in step (1);

[0009] (3) Perform lesion marking based on respiratory dynamic compensation on the preprocessed dynamic contrast images;

[0010] (4) The echo power corresponding to a single pixel in the lesion marking area was calculated to obtain the original discrete data of the time-average echo power. A complex Gaussian function was used to construct an in vivo pharmacokinetic perfusion model of the contrast agent. The average echo power of the lesion marking area before perfusion, the maximum echo power, and the time when the contrast agent arrived at the lesion marking area calculated using the actual echo power data were used as boundary conditions of the in vivo pharmacokinetic perfusion model of the contrast agent. The original discrete data of the time-average echo power were fitted to obtain the time-intensity curve of ultrasound contrast imaging in the lesion marking area.

[0011] Preferably, the image acquisition time of step (1) is more than 2 minutes.

[0012] Preferably, in the preprocessing described in step (2), the preprocessing method of pixel size includes mean downsampling and / or linear interpolation sampling, the preprocessing method of pixel intensity includes normalization processing, the preprocessing method of signal noise includes filtering noise reduction, and the preprocessing method of image frame rate includes resampling to a fixed frame rate.

[0013] Preferably, the preprocessing described in step (2) also includes difference replacement of invalid angiography frames.

[0014] Preferably, the lesion labeling method based on respiratory dynamic compensation in step (3) is selected from a semi-automatic labeling method and a fully automatic labeling method, the semi-automatic labeling method includes labeling with the Kanade-Lucas-Tomasi dynamic tracking method followed by manual correction, and the fully automatic labeling method includes an ultrasound image segmentation network based on deep learning.

[0015] Preferably, in step (4), the fitting of the original discrete data of the time-average echo power to a time-intensity curve having the smallest mean square error with the original discrete data.

[0016] The present invention also provides an ultrasound contrast quantitative analysis system for dynamically following up the effects of local radiotherapy and chemotherapy for cancer, comprising an image acquisition module, an image preprocessing module, a lesion marking auxiliary module and a quantitative analysis module;

[0017] The image acquisition module acquires continuous dynamic angiography images;

[0018] The image preprocessing module is used to preprocess at least two of the pixel size, pixel intensity, signal noise and image frame rate of the acquired dynamic contrast image;

[0019] The lesion marking auxiliary module is used to mark the lesions based on respiratory dynamic compensation on the pre-processed dynamic angiography image;

[0020] The quantitative analysis module is used to implement the following calculations:

[0021] The echo power corresponding to a single pixel in the lesion marked area was calculated to obtain the original discrete data of time-average echo power. The complex Gaussian function was used to construct the in vivo pharmacokinetic perfusion model of contrast agent. The average echo power of the lesion marked area before perfusion, the maximum echo power and the time when the contrast agent arrived at the lesion marked area calculated from the actual echo power data were used as the boundary conditions of the in vivo pharmacokinetic perfusion model of contrast agent to fit the original discrete data of time-average echo power, and the time-intensity curve of ultrasound contrast imaging in the lesion marked area was obtained.

[0022] Preferably, the image preprocessing module includes at least two of the following modules:

[0023] A pixel size pre-processing module, including a computer element capable of performing mean downsampling and / or linear interpolation sampling;

[0024] A pixel intensity pre-processing module including a computer component capable of performing normalization;

[0025] A signal noise preprocessing module, including computer components that can perform filtering and noise reduction;

[0026] An image frame rate preprocessing module, including a computer element capable of resampling to a fixed frame rate;

[0027] Preferably, it also includes an invalid contrast frame processing module, including a computer component that can perform difference replacement of invalid contrast frames.

[0028] Preferably, the lesion marking auxiliary module includes a semi-automatic marking module and / or a fully automatic marking module;

[0029] The semi-automatic labeling module includes a Kanade-Lucas-Tomasi dynamic tracking method labeling module and an auxiliary manual correction module;

[0030] The fully automatic labeling module includes an ultrasound image segmentation network module based on deep learning.

[0031] The present invention also provides a computer-readable storage medium storing a computer program, characterized in that when the computer program is executed by a processor, the steps of the ultrasound contrast quantitative analysis method as described in the above technical solution are implemented, or the functions of the ultrasound contrast quantitative analysis system as described in the above technical solution are implemented.

[0032] Compared with the prior art, the present invention has the following beneficial effects:

[0033] The present invention provides a new ultrasound contrast imaging quantitative analysis method for dynamic follow-up of the effect of local radiotherapy and chemotherapy on tumors. The method pre-processes the pixel size, pixel intensity, signal noise and image frame rate of the collected images in a standardized manner, eliminates the accuracy influencing factors from different ultrasound equipment, different contrast agents and the like, and offsets the interference of the patient's respiratory movement on the image through semi-automatic or fully automated lesion annotation based on respiratory dynamic compensation, and finally obtains the TIC curve through fitting calculation. The method eliminates the differences in ROI annotation of ultrasound contrast imaging images in different periods, different equipment, different contrast agents and different doctors, and provides highly accurate and repeatable quantitative analysis parameters for clinical comparison of perfusion parameter changes of solid tumors in radiotherapy and chemotherapy. The method can accurately and quantitatively analyze the contrast agent attention situation in the lesion area (ROI), thereby providing a highly accurate, repeatable, non-invasive and sensitive evaluation method for clinical evaluation of the early efficacy of radiotherapy and chemotherapy of solid tumors. DETAILED DESCRIPTION

[0034] The present invention provides a method for quantitative analysis of ultrasound contrast imaging for dynamic follow-up of the effects of local radiotherapy and chemotherapy on cancer:

[0035] The present invention first collects continuous dynamic contrast images from the start of ultrasound contrast agent injection. For the convenience of comparison, the probe should be set at the same position during multiple collections. Preferably, the present invention collects continuous dynamic contrast images of at least 2 minutes of the patient's lesion area to be observed during collection, so that the collected dynamic contrast images include images from the start of contrast agent injection to the exit of the contrast agent from the lesion site as much as possible. The original image is exported, and the image format can be JPG, AVI and DICOM.

[0036] After the dynamic contrast image is obtained, at least two of the pixel size, pixel intensity, signal noise and image frame rate of the dynamic contrast image are preprocessed. The purpose of preprocessing the dynamic contrast image is to eliminate errors such as equipment, contrast agent, and human operation, and improve the accuracy of quantitative analysis.

[0037] Preferably, the pixel size preprocessing method includes mean downsampling and / or linear interpolation sampling; the pixel intensity preprocessing method includes normalization. Preprocessing the image pixel size and intensity can reduce the image resolution and overall image intensity difference problems caused by the device and contrast agent concentration. Preferably, the signal noise preprocessing method includes filtering noise reduction, and filtering noise reduction can use common filters such as Gaussian filters to reduce the noise of the image; the image frame rate preprocessing method includes resampling to a fixed frame rate. The signal noise and image frame rate of the image are preprocessed to cope with the situation where the time intervals of each frame of images from different patients (devices) are different, so that the images collected by different devices are adjusted in time to be consistent with the actual angiography observation time. In order to further improve the accuracy of ultrasound contrast quantitative analysis, the present invention preferably also performs difference replacement on invalid contrast frames caused by force majeure factors such as acoustic shadow, probe movement, lesion loss, etc. in ultrasound contrast images, reducing information in the original data that is irrelevant to actual blood perfusion; in some specific embodiments of the present invention, after deleting the invalid contrast frames, a linear interpolation replacement method is used to ensure the temporal stability of the frame number after deletion, ensuring that subsequent time-related quantitative parameters all reflect the actual time. The present invention performs multi-faceted preprocessing on the collected ultrasound contrast images to avoid the image data used for quantitative analysis from being affected by error factors such as different patients and different equipment, and to reduce the influence of quantitative parameters on factors irrelevant to the actual focus of the lesion, thereby improving the accuracy of quantitative analysis and the repeatability of the results.

[0038] After obtaining the pre-processed dynamic contrast image, the lesion annotation based on respiratory dynamic compensation is performed. In the present invention, the lesion annotation method based on respiratory dynamic compensation is selected from a semi-automatic annotation method and a fully automatic annotation method. In the present invention, the semi-automatic annotation method can be annotated by the Kanade-Lucas-Tomasi dynamic tracking method and then manually corrected; in some specific embodiments of the present invention, after an operator annotates the lesion at any time on the pre-processed dynamic contrast image, the Kanade-Lucas-Tomasi dynamic tracking method identifies the image feature points in the annotated lesion area frame by frame, and tracks the lesion annotated area according to the movement of the feature points, offsets the image fluctuation caused by the patient's respiratory movement during the acquisition process, and manually corrects the unsatisfactory part of the Kanade-Lucas-Tomasi dynamic tracking method annotation result to ensure that the lesion area on each frame of the image is completely and accurately annotated. In the present invention, the semi-automatic labeling method can be a region growing method or a fast marching method based on the selection point and the stop condition; the fully automatic labeling method can be an ultrasound image segmentation network based on deep learning, specifically, a neural segmentation network such as U-Net can be trained on some of the labeled ultrasound images, and this network can be used to assist the lesion segmentation of subsequent images and the lesion tracking at the frame-by-frame level. The lesion region described in the present invention is the region of interest (ROI) in the ultrasound image analysis. Usually, the ROI is labeled by the operator. The semi-automatic or fully automatic labeling method saves the cost of manual labeling one by one, improves the repeatability of lesion labeling, and improves the contrast of ultrasound angiography collected at different times.

[0039] After obtaining the dynamic contrast image of the lesion annotation, the average echo power corresponding to the pixel intensity in the lesion annotation area is calculated to obtain the time-average echo power original discrete data; the contrast agent in vivo pharmacokinetic perfusion model is constructed according to the blood perfusion time period of the pre-processed dynamic contrast image, and the average echo power of the lesion annotation area before perfusion, the maximum echo power and the time when the contrast agent reaches the lesion annotation area calculated by the actual echo power data are used as the boundary conditions of the contrast agent in vivo pharmacokinetic perfusion model, and the time-average echo power original discrete data is fitted to obtain the time-intensity curve of ultrasound contrast in the lesion annotation area. In the present invention, the fitting of the time-average echo power original discrete data is preferably performed to minimize the mean square error between the fitted time-intensity curve and the original discrete data.

[0040] By analyzing and calculating the time-intensity curve TIC obtained by the ultrasound contrast quantitative analysis method of the present invention, quantitative characteristics such as peak intensity (PE), rise time (RT), time to peak (TTP), mean transit time (mTT) and area under the TIC curve (AUC) can be obtained. By comparing the quantitative characteristic parameters of ultrasound contrast before and after radiotherapy and chemotherapy, combined with statistical methods and actual clinical events of patients, non-invasive and sensitive evaluation of the early local treatment effect of radiotherapy and chemotherapy can be achieved.

[0041] The present invention also provides an ultrasound contrast quantitative analysis system for dynamically following up the effects of local radiotherapy and chemotherapy for cancer, comprising an image acquisition module, an image preprocessing module, a lesion marking auxiliary module and a quantitative analysis module.

[0042] The image acquisition module of the present invention acquires continuous dynamic contrast images starting from the injection of ultrasound contrast agent. Preferably, the acquisition time is preferably more than 2 minutes. The present invention exports the acquired continuous dynamic contrast images to the image preprocessing module, and the image format can be JPG, AVI and DICOM.

[0043] The image preprocessing module of the present invention is used to preprocess at least two of the pixel size, pixel intensity, signal noise and image frame rate of the collected dynamic angiography images. Preferably, the image preprocessing module of the present invention includes at least two of the following modules: a pixel size preprocessing module, including a computer element capable of performing mean downsampling and / or linear interpolation sampling; a pixel intensity preprocessing module, including a computer element capable of performing normalization processing; a signal noise preprocessing module, including a computer element capable of performing filtering and noise reduction; an image frame rate preprocessing module, including a computer element capable of resampling to a fixed frame rate; further preferably, an invalid angiography frame processing module is also included, including a computer element capable of performing difference replacement of invalid angiography frames.

[0044] The lesion annotation auxiliary module of the present invention is used to perform lesion annotation based on respiratory dynamic compensation on the pre-processed dynamic contrast image. Preferably, the lesion annotation auxiliary module includes a semi-automatic annotation module and / or a fully automatic annotation module. In some specific embodiments of the present invention, the semi-automatic annotation module includes a Kanade-Lucas-Tomasi dynamic tracking annotation module and a manual correction auxiliary module; the fully automatic annotation module can be an artificial intelligence annotation module. Usually, the ROI is annotated by the operator. The semi-automatic or fully automatic annotation method saves the cost of manual annotation one by one, improves the repeatability of lesion annotation, and improves the contrast of ultrasound angiography collected at different times.

[0045] The quantitative analysis module of the present invention is used to realize the following calculations: calculating the echo power corresponding to a single pixel in the lesion marking area to obtain the original discrete data of the time-average echo power; using a complex Gaussian function to construct a pharmacokinetic perfusion model of the contrast agent in vivo, using the average echo power of the lesion marking area before perfusion, the maximum echo power and the time when the contrast agent reaches the lesion marking area calculated using the actual echo power data as the boundary conditions of the pharmacokinetic perfusion model of the contrast agent in vivo to fit the original discrete data of the time-average echo power, and obtain the time-intensity curve of ultrasound contrast imaging of the lesion marking area.

[0046] The present invention also provides a computer-readable storage medium storing a computer program, characterized in that when the computer program is executed by a processor, the steps of the ultrasound contrast quantitative analysis method as described in the above technical solution are implemented, or the functions of the ultrasound contrast quantitative analysis system as described in the above technical solution are implemented.

[0047] The ultrasound contrast quantitative analysis method, system and computer-readable storage medium provided by the present invention can be used for evaluating the efficacy of tumor radiotherapy and chemotherapy, especially for dynamic evaluation of the efficacy of radiotherapy and chemotherapy for solid tumors. In some embodiments of the present invention, the ultrasound contrast quantitative analysis method, system and computer-readable storage medium of the present invention are used in evaluating the efficacy of radiotherapy and chemotherapy for pancreatic cancer.

[0048] The technical solutions provided by the present invention are described in detail below in conjunction with the embodiments, but they should not be construed as limiting the protection scope of the present invention.

[0049] Example 1 Evaluation of the effect of radiotherapy and chemotherapy on patients with locally advanced pancreatic cancer

[0050] A. Obtaining pancreatic cancer ultrasound contrast-enhanced dynamic images: Perform ultrasound contrast-enhanced imaging on patients with locally advanced pancreatic cancer. Start with contrast agent injection, observe the lesion, and collect at least 2 minutes of continuous dynamic video, and export the original DICOM images compressed in JPEG format.

[0051] B. Dynamic contrast image preprocessing: In view of the problem that the contrast images collected from different ultrasound equipment sources and contrast agent concentrations have image resolution and overall image intensity differences caused by the equipment and contrast agent concentration, the image pixel size is averaged downsampled or linearly interpolated upsampled, and the pixel intensity is normalized. Reduce the influence of quantitative parameters between different patients on factors unrelated to the actual perfusion of the lesion. Use common filters such as Gaussian filtering to reduce the noise in the image signal. And according to the experimental design, the frame rate of different contrast images is resampled to a fixed frame rate to cope with the situation that the time interval between each frame of images from different patients (equipment) is different, ensuring that the images from different devices are consistent with the actual time in time. Afterwards, invalid contrast frames caused by force majeure such as acoustic shadows, probe movement, and lesion loss during ultrasound contrast examination are manually replaced and deleted to reduce information unrelated to actual perfusion in the original data. At the same time, the time stability of the image after the frame number is deleted is ensured by linear interpolation replacement. Ensure that subsequent time-related quantitative parameters reflect the actual time.

[0052] C. Lesion delineation and respiratory motion compensation: The operator uses medical image annotation software to annotate the region of interest, i.e., the lesion area. Dynamic data containing multiple frames of images does not need to be annotated frame by frame. After the operator annotates the lesion at any time in the angiography sequence, the Kanade-Lucas-Tomasi dynamic tracking method is used for the pixel features in the annotated area to identify the image feature points in the annotated area frame by frame, and the annotated area is tracked according to the movement of the feature points to offset the image fluctuations caused by the patient's respiratory movement during the examination, reduce the cost of manual annotation, and improve the repeatability of lesion annotation. For unsatisfactory tracking results, manual correction methods are used to ensure that the lesions on each frame of the image are completely and accurately identified.

[0053] D. Time-intensity curve (TIC) fitting: Taking dynamic frames as units, according to the principle of ultrasound imaging, the average echo power corresponding to the pixel intensity in the lesion area marked in each frame of the image is calculated to obtain the original discrete data of time-average echo power. According to the time periods included in the contrast image, such as the perfusion segment, the withdrawal segment, and the perfusion-withdrawal segment, a suitable Gaussian model is selected to simulate the pharmacokinetic perfusion of the contrast agent in the human body, and the original discrete echo power data is fitted to obtain a continuous time-intensity curve (TIC). Based on the actual discrete echo power data, the average echo power of the lesion site before perfusion, the maximum echo power during the contrast process, and the approximate range of the time when the contrast agent reaches the lesion site are calculated. This range is used as the boundary condition of the Gaussian model, and the time-intensity curve (TIC) is iteratively calculated until the mean square error between the fitting curve TIC and the original discrete data is minimized.

[0054] E. Evaluate the effect of pancreatic cancer after chemoradiotherapy: By comparing the TIC curves and related ultrasound contrast quantitative parameters before and after treatment, evaluate the changes in microcirculatory blood perfusion in pancreatic cancer tumors, thereby providing objective, quantitative, and dynamic indicators for clinical reflection of efficacy. Pancreatic ultrasound contrast imaging is performed on pancreatic cancer patients who receive chemoradiotherapy before, after, and during postoperative follow-up, and contrast images are collected according to step A above; the above preprocessing is performed on different contrast images to obtain the time-intensity curve reflecting the perfusion of ultrasound contrast agents, and quantitative and semi-quantitative parameters related to microcirculatory blood perfusion in the lesions of patients before and after chemoradiotherapy are obtained through fitting calculations. By comparing the changes in these quantitative parameters before and after chemoradiotherapy, an imaging method is provided for clinical non-invasive and sensitive evaluation of the efficacy of chemoradiotherapy for locally advanced pancreatic cancer.

[0055] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A quantitative analysis method of ultrasound contrast imaging for dynamic follow-up of the effect of local radiotherapy and chemotherapy on tumors. It is characterized in that The following steps are involved: (1) Acquisition of continuous dynamic angiography images; (2) preprocessing at least two of the pixel size, pixel intensity, signal noise and image frame rate of the dynamic contrast image obtained in step (1); (3) Perform lesion marking based on respiratory dynamic compensation on the preprocessed dynamic contrast images; (4) The echo power corresponding to a single pixel in the lesion marking area is calculated to obtain the original discrete data of the time-average echo power. A complex Gaussian function is used to construct an in vivo pharmacokinetic perfusion model of the contrast agent. The average echo power of the lesion marking area before perfusion, the maximum echo power, and the time when the contrast agent reaches the lesion marking area calculated using the actual echo power data are used as boundary conditions of the in vivo pharmacokinetic perfusion model of the contrast agent to fit the original discrete data of the time-average echo power, and the time-intensity curve of ultrasound contrast imaging in the lesion marking area is obtained.

2. The ultrasound contrast-enhanced quantitative analysis method according to claim 1, It is characterized in that The image acquisition time of step (1) is more than 2 minutes.

3. The ultrasound contrast-enhanced quantitative analysis method according to claim 1, It is characterized in that In the preprocessing described in step (2), the preprocessing method of pixel size includes mean downsampling and / or linear interpolation sampling, the preprocessing method of pixel intensity includes normalization processing, the preprocessing method of signal noise includes filtering noise reduction, and the preprocessing method of image frame rate includes resampling to a fixed frame rate.

4. The ultrasound contrast-enhanced quantitative analysis method according to claim 1 or 3, It is characterized in that The preprocessing described in step (2) also includes difference replacement of invalid angiography frames.

5. The ultrasound contrast-enhanced quantitative analysis method according to claim 1, It is characterized in that The lesion labeling method based on respiratory dynamic compensation in step (3) is selected from a semi-automatic labeling method and a fully automatic labeling method. The semi-automatic labeling method includes labeling with the Kanade-Lucas-Tomasi dynamic tracking method and then performing manual correction. The fully automatic labeling method includes an ultrasound image segmentation network based on deep learning.

6. The ultrasound contrast-enhanced quantitative analysis method according to claim 1, It is characterized in that The step (4) is to fit the original discrete data of the time-average echo power to minimize the mean square error between the fitted time-intensity curve and the original discrete data.

7. A quantitative analysis system for ultrasound contrast imaging for dynamic follow-up of the effects of local radiotherapy and chemotherapy on cancer. It is characterized in that It includes image acquisition module, image preprocessing module, lesion annotation auxiliary module and quantitative analysis module. The image acquisition module acquires continuous dynamic angiography images; The image preprocessing module is used to preprocess at least two of the pixel size, pixel intensity, signal noise and image frame rate of the acquired dynamic contrast image; The lesion marking auxiliary module is used to mark the lesions based on respiratory dynamic compensation on the pre-processed dynamic angiography image; The quantitative analysis module is used to implement the following calculations: The echo power corresponding to a single pixel in the lesion marked area was calculated to obtain the original discrete data of time-flat wave power. The complex Gaussian function was used to construct the in vivo pharmacokinetic perfusion model of contrast agent. The average echo power, maximum echo power and the time when the contrast agent arrived at the lesion marked area before perfusion calculated from the actual echo power data were used as the boundary conditions of the in vivo pharmacokinetic perfusion model of contrast agent to fit the original discrete data of time-average echo power, and the time-intensity curve of ultrasound contrast imaging in the lesion marked area was obtained.

8. The ultrasound contrast quantitative analysis system according to claim 7, It is characterized in that The image preprocessing module comprises: A pixel size pre-processing module, including a computer element capable of performing mean downsampling and / or linear interpolation sampling; A pixel intensity pre-processing module including a computer component capable of performing normalization; A signal noise preprocessing module, including computer components that can perform filtering and noise reduction; An image frame rate preprocessing module, including a computer element capable of resampling to a fixed frame rate; Also included is an invalid contrast frame processing module, which includes a computer component capable of performing difference replacement of invalid contrast frames.

9. The ultrasound contrast quantitative analysis system according to claim 7, It is characterized in that The lesion marking auxiliary module includes a semi-automatic marking module and / or a fully automatic marking module. The semi-automatic labeling module includes a Kanade-Lucas-Tomasi dynamic tracking method labeling module and an auxiliary manual correction module; The fully automatic labeling module includes an ultrasound image segmentation network module based on deep learning.

10. A computer-readable storage medium having stored thereon a computer program, It is characterized in that When the computer program is executed by a processor, the steps of the ultrasound contrast quantitative analysis method according to any one of claims 1 to 6 are implemented, or the functions of the ultrasound contrast quantitative analysis system according to any one of claims 7 to 9 are implemented.