Rapid dynamic PET quantitative parameter imaging method and system based on linear regression model, terminal and storage medium

Abstract

The invention relates to the technical field of image processing, and discloses a rapid dynamic PET quantitative parameter imaging method and system based on a linear regression model, a terminal and a storage medium, and the method comprises the steps: constructing a fitting data set and a test data set; constructing a standard Patlak analysis model, and selecting a descending aorta slice dynamic signal as an artery input function; establishing a linear regression model between the early integral value and the late mean value by using the fitting data set; extracting a late mean value of the artery input function from the test data set, and estimating an early integral value corresponding to the late mean value in the test data set by using the optimal linear fitting parameter of the linear regression model; and combining the estimated early integral value with a standard Patlak analysis model to calculate each voxel to obtain a voxel-level Patlak quantitative parameter image. According to the method, parameter images with higher quality are obtained by using fewer dynamic scanning sequences, and short-time dynamic PET quantitative parameter imaging is realized.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and in particular to a fast dynamic PET quantitative parameter imaging method, system, terminal and computer-readable storage medium based on a linear regression model. Background Art

[0002] Positron emission tomography (PET) is an advanced medical imaging technology that can visualize physiological and pathological states by detecting the distribution of radioactive tracers in the body. The advantages of PET are its high sensitivity and good spatial resolution. In addition, compared with other imaging modalities, PET can provide information about metabolic activity. However, dynamic PET scans usually last more than 60 minutes, and such long scans may cause discomfort or anxiety to patients.

[0003] Currently, most methods for shortening dynamic PET scan duration rely on deep learning. However, deep learning models are often viewed as "black boxes," meaning it's difficult to clearly understand how they generate outputs from input data. Existing technologies struggle to understand the internal mechanisms and decision-making processes of deep learning models, potentially impacting clinical reliability. These models involve complex network structures and hyperparameter settings, making parameter adjustment complex. They often rely on large amounts of training data to build effective models, and in some cases, obtaining sufficient high-quality data can be challenging.

[0004] Therefore, the existing technology still needs to be improved and developed. Summary of the Invention

[0005] The main purpose of the present invention is to provide a fast dynamic PET quantitative parameter imaging method, system, terminal and computer-readable storage medium based on a linear regression model, aiming to solve the problem in the prior art that dynamic PET scanning lasts too long and high-quality parameter images cannot be obtained in a short time.

[0006] To achieve the above object, the present invention provides a rapid dynamic PET quantitative parameter imaging method based on a linear regression model, wherein the rapid dynamic PET quantitative parameter imaging method based on a linear regression model comprises the following steps:

[0007] Constructing a data set and drawing a region of interest in a dynamic PET image, dividing the dynamic sequence in the data set into a fitting data set and a test data set according to a preset ratio;

[0008] A standard Patlak analysis model is constructed according to the fitting data set, and based on the standard Patlak analysis model, a dynamic signal of a descending aorta slice is selected as an arterial input function;

[0009] Obtaining an early integral value and a late mean value of the arterial input function, and establishing a linear regression model between the early integral value and the late mean value using the fitting data set;

[0010] extracting a late mean value of the arterial input function from the test data set, and estimating an early integral value corresponding to the late mean value in the test data set using the best linear fit parameter of the linear regression model;

[0011] The estimated early integral value is combined with the standard Patlak analysis model, and calculation is performed on each voxel of the three-dimensional human body to obtain a voxel-level Patlak quantitative parameter image.

[0012] Optionally, in the fast dynamic PET quantitative parameter imaging method based on a linear regression model, the region of interest represents a specific three-dimensional region selected in the dynamic PET image, which is used to analyze and quantify the distribution and dynamic changes of the radioactive tracer in the specific three-dimensional region.

[0013] Optionally, in the fast dynamic PET quantitative parameter imaging method based on a linear regression model, the standard Patlak analysis model is used for pharmacokinetic parameter imaging of irreversible tracers.

[0014] Optionally, in the fast dynamic PET quantitative parameter imaging method based on a linear regression model, the expression of the standard Patlak analysis model is:

[0015]

[0016] Among them, C T (t) and C P (t) represent the tissue time-activity curve and arterial input function, t represents the time, t * represents the moment when the tracer reaches a steady state, represents the area under the plasma concentration-time curve, K i represents the slope and V represents the intercept.

[0017] Optionally, the method for rapid dynamic PET quantitative parameter imaging based on a linear regression model, wherein the step of obtaining the early integral value and the late mean value of the arterial input function and establishing a linear regression model between the early integral value and the late mean value using the fitting data set, specifically includes:

[0018] Get the early integral value S of the arterial input function early and late mean M late ;

[0019] Modify the expression of the standard Patlak analysis model to:

[0020]

[0021] Among them, S early represents the early integral value of the arterial input function, and t1 represents the starting time of the scan;

[0022] Early integral value S early The expression is:

[0023]

[0024] A linear regression model is established between the early integral value and the late mean value, and the expression is:

[0025] S early =k·M late +b; (4)

[0026] Among them, M late represents the late mean of the arterial input function, k and b are the slope and intercept parameters of the linear regression model, respectively.

[0027] Optionally, in the fast dynamic PET quantitative parameter imaging method based on the linear regression model, the early integral value represents the integral value of the AIF curve before a set frame node, and the late mean represents the mean value of the AIF curve after a set frame node.

[0028] Optionally, in the rapid dynamic PET quantitative parameter imaging method based on a linear regression model, the preset ratio is 7:3.

[0029] In addition, to achieve the above-mentioned purpose, the present invention further provides a fast dynamic PET quantitative parameter imaging system based on a linear regression model, wherein the fast dynamic PET quantitative parameter imaging system based on a linear regression model comprises:

[0030] A data set construction module is used to construct a data set and draw a region of interest in a dynamic PET image, and divide the dynamic sequence in the data set into a fitting data set and a test data set according to a preset ratio;

[0031] a signal selection module, configured to construct a standard Patlak analysis model according to the fitted data set, and select a descending aorta slice dynamic signal as an arterial input function based on the standard Patlak analysis model;

[0032] a model building module, configured to obtain an early integral value and a late mean value of the arterial input function, and to establish a linear regression model between the early integral value and the late mean value using the fitting data set;

[0033] a numerical estimation module, configured to extract a late mean value of the arterial input function from the test data set, and estimate an early integral value corresponding to the late mean value in the test data set using the best linear fitting parameter of the linear regression model;

[0034] The image generation module is used to combine the estimated early integral value with the standard Patlak analysis model, calculate each voxel of the three-dimensional human body, and obtain a voxel-level Patlak quantitative parameter image.

[0035] In addition, to achieve the above-mentioned purpose, the present invention also provides a terminal, wherein the terminal includes: a memory, a processor, and a fast dynamic PET quantitative parameter imaging program based on a linear regression model stored in the memory and runnable on the processor, wherein the fast dynamic PET quantitative parameter imaging program based on a linear regression model, when executed by the processor, implements the steps of the fast dynamic PET quantitative parameter imaging method based on a linear regression model as described above.

[0036] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a rapid dynamic PET quantitative parameter imaging program based on a linear regression model, and when the rapid dynamic PET quantitative parameter imaging program based on a linear regression model is executed by a processor, the steps of the rapid dynamic PET quantitative parameter imaging method based on a linear regression model as described above are implemented.

[0037] In the present invention, a data set is constructed and a region of interest is drawn in a dynamic PET image. The dynamic sequence in the data set is divided into a fitting data set and a test data set according to a preset ratio. A standard Patlak analysis model is constructed based on the fitting data set. Based on the standard Patlak analysis model, the dynamic signal of the descending aorta slice is selected as the arterial input function. The early integral value and late mean of the arterial input function are obtained, and a linear regression model is established between the early integral value and the late mean using the fitting data set. The late mean of the arterial input function is extracted from the test data set, and the early integral value corresponding to the late mean in the test data set is estimated using the optimal linear fitting parameters of the linear regression model. The estimated early integral value is combined with the standard Patlak analysis model, and calculations are performed on each voxel of the three-dimensional human body to obtain a voxel-level Patlak quantitative parametric image. The present invention obtains higher-quality parametric images using fewer dynamic scanning sequences, realizes short-term dynamic PET quantitative parametric imaging, and provides a new solution for improving imaging efficiency and accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a flow chart of a preferred embodiment of the fast dynamic PET quantitative parameter imaging method based on the linear regression model of the present invention;

[0039] Figure 2 It is a technical circuit diagram of the fitting process and the testing process in a preferred embodiment of the rapid dynamic PET quantitative parameter imaging method based on the linear regression model of the present invention;

[0040] Figure 3 Schematic diagram of a 60-min standard parameter image in a preferred embodiment of the rapid dynamic PET quantitative parameter imaging method based on a linear regression model of the present invention;

[0041] Figure 4 Schematic diagram of an unprocessed 30-min short-time parameter image in a preferred embodiment of the fast dynamic PET quantitative parameter imaging method based on a linear regression model of the present invention;

[0042] Figure 5 1 is a schematic diagram of the restoration results in a preferred embodiment of the rapid dynamic PET quantitative parameter imaging method based on the linear regression model of the present invention;

[0043] Figure 6 This is a structural diagram of a preferred embodiment of a fast dynamic PET quantitative parameter imaging system based on a linear regression model of the present invention;

[0044] Figure 7 FIG. 4 is a structural diagram of a preferred embodiment of the terminal of the present invention. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical solutions and advantages of the present invention more clear and distinct, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0046] The present invention establishes a linear regression model between the early integral and the late mean of the arterial input function (AIF), thereby effectively estimating the early integral value of the input function required for parameter fitting calculations in the pharmacokinetic map model, realizing the slope fitting correction of the Patlak analysis in short-time dynamic scanning, and realizing short-time dynamic PET quantitative parameter imaging. Ultimately, higher-quality parametric images are obtained using fewer dynamic scanning sequences, providing a new solution for improving imaging efficiency and accuracy.

[0047] The fast dynamic PET quantitative parameter imaging method based on the linear regression model described in the preferred embodiment of the present invention is as follows: Figure 1 and Figure 2As shown, the rapid dynamic PET quantitative parameter imaging method based on the linear regression model includes the following steps:

[0048] Step S10: constructing a data set and drawing a region of interest in the dynamic PET image, and dividing the dynamic sequence in the data set into a fitting data set and a test data set according to a preset ratio.

[0049] Specifically, a data set is constructed, and the dynamic sequences in the data set are divided into a fitting data set and a test data set according to a preset ratio (e.g., 7:3). For example, the present invention selects data from 105 patients with obvious relevant symptoms or characteristics, and randomly selects 74 dynamic sequences (about 70%) as the fitting data set for Figure 2 The remaining 31 dynamic sequences (about 30%) were used as test data sets for the implementation of the fitting process. Figure 2 Implementation of the testing process.

[0050] Furthermore, the present invention manually outlines volumes of interest (i.e., regions of interest, VOI refers to a specific three-dimensional region selected in the image for further analysis and quantification of the distribution and dynamic changes of radioactive tracers in the region. VOI is a region that is outlined and determined when the present invention performs parameter imaging effect analysis on lesions at specific locations; the fitting dataset and the test dataset are both dynamic PET images, and VOI is outlined and determined by finding the location of the lesion in these images). Among them, pathological lesions are divided into primary tumors (PT), lymph node metastasis (LNM), and other lesions (OL). VOIs are drawn on at least three consecutive dynamic PET image slices, and TAC (Time-Activity Curve) is extracted from the dynamic sequence. The proposed model was evaluated using 31 dynamic sequences of the test dataset, which contains 53 lesions (25 PTs, 15 LNMs, and 13 OLs). The VOIs of these lesions were used to evaluate K i The quality of the parametric images was assessed qualitatively and quantitatively.

[0051] Step S20: constructing a standard Patlak analysis model according to the fitting data set, and selecting the descending aorta slice dynamic signal as the arterial input function based on the standard Patlak analysis model.

[0052] Specifically, the standard Patlak analysis model is used for irreversible tracers (such as 18The goal of Patlak parameter imaging is to generate the parameter K i The graph of the parameter V represents the net glucose metabolic flux and the distribution volume of free glucose molecules, respectively. These two parameters enable researchers to better understand the physiological state of tissues and their response to drugs or metabolites. The expression of the standard Patlak analysis model is:

[0053]

[0054] Among them, C T (t) and C P (t) represents the tissue time-activity curve (TAC) and arterial input function, t represents the time, t * represents the moment when the tracer reaches a steady state, represents the area under the plasma concentration-time curve, K i represents the slope and V represents the intercept.

[0055] In the standard Patlak analysis, the segment corresponding to the last 30-40 minutes of the dynamic PET scan is approximately linear, and this linear region is used to fit the parameter K i (slope) and V (intercept). By calculating voxel by voxel, a parametric image can be generated and spatially registered with the original PET image. The parameter K obtained by Patlak analysis using the complete 1 hour data i It is considered the gold standard for evaluating the performance of other methods. The general method for obtaining the AIF (arterial input function) is to extract dynamic signals from large blood vessels in PET images. In this paper, the dynamic signals of descending aorta (DA) slices (with distinct vascular structures) are selected as the AIF. The manually drawn DA region of interest is strictly limited to the vascular area, excluding the vessel wall and extravascular area.

[0056] Step S30: Obtain the early integral value and the late mean value of the arterial input function, and establish a linear regression model between the early integral value and the late mean value using the fitting data set.

[0057] Specifically, it can be seen from the standard Patlak analysis model that the complete dynamic sequence of AIF is very important for parameter estimation. Therefore, in the study of realizing short-time parametric imaging, how to use the short-time dynamic sequence to obtain the complete AIF integral value is the key to parameter fitting calculation. In order to shorten the dynamic scanning time to 30 minutes, the present invention selects a 15-frame (30-minute) scanning protocol from the 43rd to 66th frames in the late scanning period (during which the tracer concentration reaches a stable state) to conduct short-time parametric imaging research. There are a total of 10 schemes available for evaluation: frames 43-57, frames 44-58, ..., frames 52-66. To simplify the description, the present invention uses the starting frame number of each scheme to represent the corresponding frame scheme. For example, the 43rd frame node represents the scanning protocol for frames 43-57, the 44th frame node represents the scanning protocol for frames 44-58, and so on. is the integral value of the 15-frame scanning protocol, expressed as S early It represents the early integral value of AIF, t1 represents the starting time of the 15-frame scan, and the average tracer concentration of these 15 frames Cp (arterial input function, which represents the curve of tracer activity concentration changing with time) is expressed as M late It represents the late mean of AIF.

[0058] The AIF is the TAC curve extracted from the aortic region at that location. The AIF early integral value is the integral value of the AIF curve before a set frame node, and the AIF late mean value is the mean value of the AIF curve after a set frame node. The linear relationship between the two forms the linear regression model constructed in this invention. A strong positive correlation exists between the AIF early integral value and the AIF late mean value across multiple dynamic sequences from the same scan batch.

[0059] Get the early integral value S of the arterial input function early and late mean M late , modify the expression of the standard Patlak analysis model to:

[0060]

[0061] Among them, S early represents the early integral value of the arterial input function, and t1 represents the starting time of the scan;

[0062] Early integral value S early The expression is:

[0063]

[0064] It should be noted that the points Starting from the end of t1min, when there is no early dynamic sequence, S earlyThe value of is unknown. From formula (2), we can see that this will seriously affect the K in Patlak analysis. i Parameter calculation. Therefore, a linear regression model is established between the early integral value and the late mean value, and the expression is:

[0065] S early =k·M late +b; (4)

[0066] Among them, M late represents the late mean of the arterial input function, k and b are the slope and intercept parameters of the linear regression model, respectively.

[0067] Using linear regression model to calculate the late In the case of S early , thereby improving K in short-time scanning i To ensure parameter accuracy, we constructed a linear regression model between the early integral value and the late mean of the AIF using the dynamic series of the fitting dataset. Next, for the dynamic series in the test dataset, we used this linear regression model and the late mean of the AIF in the test dataset to estimate the early integral value of the AIF. This estimate is crucial for correcting the parameter fit in the Patlak analysis.

[0068] Figure 2 The fitting process shown shows how to use the fitting data set to construct a linear regression model between the early integral and late mean of AIF. The specific steps are: ① Randomly select 74 dynamic sequences (about 70%) from the 105 patient data included in the present invention as the fitting data set. ② Use the TAC of the descending aorta as the arterial input function AIF in the Patlak analysis for parameter fitting calculation. ③ Extract the AIFs of the complete 1-hour scanning protocol from the 74 fitting data sets. ④ Calculate the S of each frame node from 43 to 52. early and its corresponding M late For example, when the frame node is 43, S early Indicates the AIF integral value from 0 to the end of the 42nd frame scan, M late Indicates the average value of AIF for frames 43-57. When the frame node is 52, S early Indicates the AIF integral value from 0 to the end of the 51st frame scan, M late Represents the average value of AIF for frames 52-66. ⑤ For each frame node of frames 43-52, construct S using the fitting set early With M late The linear regression model between , is formula (4).

[0069] Table 1 shows the S of each frame node in frames 43-52. early With Mlate The linear fitting parameter results of each frame can be used to obtain the dynamic whole-body parameter image of each patient during the test.

[0070] Table 1: Parameters k and b of the linear regression model and S for each frame node from 43 to 52 early With M late The correlation coefficient r

[0071] Frame Node 43 44 45 46 47 k 22.97 26.01 29.02 32.03 34.97 b -1226.99 3355.08 7641.32 12349.29 7399.71 r 0.91 0.92 0.93 0.93 0.93 Frame Node 48 49 50 51 52 k 37.83 40.64 43.21 45.94 48.55 b 22588.32 28365.90 36255.12 42424.34 49403.17 r 0.94 0.94 0.94 0.94 0.94

[0072] The correlation coefficient, r, is a statistic that measures the degree of linear correlation between two variables (the initial integral value and the subsequent mean value of the AIF). A larger r indicates a stronger linear correlation between the two variables. Since the purpose of this study is to predict one variable using another, a larger absolute value of r indicates a more reliable prediction.

[0073] Step S40: extracting the late mean of the arterial input function from the test data set, and estimating the early integral value corresponding to the late mean in the test data set using the best linear fitting parameter of the linear regression model.

[0074] Specifically, Figure 2 The test process shown shows how to use the test data set and the established linear regression model to estimate the early integral value of AIF. 31 dynamic sequences (about 30%) were randomly selected from the 105 patient data included in the present invention as the test data set. For the AIF of the test data set, the late mean M of each frame node from 43 to 52 was calculated. late , using the fitted parameters of the constructed optimal linear regression model, estimate the late mean M of the test set late The corresponding early integral value.

[0075] Step S50: combining the estimated early integral value with the standard Patlak analysis model, performing calculations on each voxel of the three-dimensional human body, and obtaining a voxel-level Patlak quantitative parameter image.

[0076] Specifically, the estimated S early Combine the value with equation (2) to calculate the C corresponding to each frame time t in equation (2) T (t), C P (t), and Therefore, according to equation (2), the parameter K of a certain voxel can be fitted using the least square method. i And parameter V, this is the Patlak analysis process, which is calculated for each voxel of the three-dimensional human body, and finally a voxel-level Patlak quantitative parameter image can be obtained.

[0077] The ultimate goal of the present invention is to shorten the conventional 1-hour dynamic scan to 30 minutes and still obtain reliable Patlak quantitative parameter images. The Patlak quantitative parameter image is an image used in positron emission tomography (PET). It is based on the Patlak graphical analysis method, which is a fast linear graphical analysis technology. Through the Patlak model, a parameter image of slope and intercept can be generated, where the slope can represent the net transfer rate or inflow constant. The Patlak quantitative parameter image helps doctors interpret PET images more accurately and provide information about tissue metabolism and function, thereby supporting diagnosis, treatment response monitoring and radiotherapy planning.

[0078] Conventional dynamic scanning often lasts for 1 hour or more. Such long scanning time increases the burden on patients and is prone to artifacts. The purpose of this invention is to shorten the duration of dynamic scanning to 30 minutes, and to attempt to complete the parametric imaging of Patlak analysis using only the dynamic sequence of the last 30 minutes of dynamic scanning; then, by comparing equations (1) and (2), it can be seen that for short-term Patlak analysis, in order to approach the effect of standard Patlak analysis, the missing S early Therefore, the present invention establishes a linear relationship between the early integral and the late mean of AIF, so that the early integral value S can be estimated when only the late mean is known. early , thereby completing the Patlak analysis calculation of equation (2) and ultimately achieving fast parameter imaging.

[0079] Figure 3-Figure 5 All images in the article use a unified numerical range to map different colors to ensure comparability of the results. The details are as follows:

[0080] like Figure 3 As shown, Figure 3 The standard 1-hour Patlak parameter image results calculated according to equation (1) are shown. Figure 3 The image serves as a reference benchmark for subsequent comparison.

[0081] like Figure 4 As shown, Figure 4 The Patlak parameter image is generated based on the data of the last 30 minutes of dynamic scanning. Figure 3 compared to, Figure 4 The parameter values are significantly higher, the noise points are particularly obvious, and even the shapes of some lesion areas are different from Figure 3 There is a large deviation in the reference results.

[0082] like Figure 5 As shown, Figure 5The results of the 30-minute Patlak parameter image after processing using the method proposed in this invention are presented. Figure 4 , Figure 5 The noise points are significantly reduced, and the parameter values are closer to Figure 3 In addition, Figure 5 The lesion shape in the image is well restored and is consistent with the reference image. Figure 3 More consistent.

[0083] The present invention provides a new short-time parametric imaging technology, which finds the linear relationship between the early integral and the late mean of the AIF, obtains the complete AIF integral, and realizes high-resolution rapid parametric imaging; constructs a linear regression model between the early integral and the late mean at each frame node in the late AIF to complete the estimation of the complete AIF integral; and determines the optimal linear regression model based on the parametric image generated by the linear regression model at each frame node in the late dynamic scanning period to realize optimal short-time parametric imaging.

[0084] Compared with existing technologies, the present invention offers the following advantages: By constructing a linear regression model between the early integral and the late mean at each frame node in the late AIF range of 43-52, the present invention estimates the complete AIF value. Furthermore, the optimal linear regression model fitting parameters are determined based on the parametric image results at each frame node, and short-term parametric imaging is achieved using Patlak analysis. This invention effectively reduces the duration of dynamic scanning while maintaining the detailed features of the parametric image.

[0085] Furthermore, if Figure 6 As shown, based on the above-mentioned fast dynamic PET quantitative parameter imaging method based on the linear regression model, the present invention also provides a fast dynamic PET quantitative parameter imaging system based on the linear regression model, wherein the fast dynamic PET quantitative parameter imaging system based on the linear regression model includes:

[0086] A data set construction module 51 is used to construct a data set and draw a region of interest in a dynamic PET image, and divide the dynamic sequence in the data set into a fitting data set and a test data set according to a preset ratio;

[0087] a signal selection module 52 for constructing a standard Patlak analysis model according to the fitting data set, and selecting a descending aorta slice dynamic signal as an arterial input function based on the standard Patlak analysis model;

[0088] a model building module 53 for obtaining an early integral value and a late mean value of the arterial input function, and establishing a linear regression model between the early integral value and the late mean value using the fitting data set;

[0089] a numerical estimation module 54 for extracting the late mean of the arterial input function from the test data set, and estimating the early integral value corresponding to the late mean in the test data set using the best linear fitting parameters of the linear regression model;

[0090] The image generation module 55 is used to combine the estimated early integral value with the standard Patlak analysis model, calculate each voxel of the three-dimensional human body, and obtain a voxel-level Patlak quantitative parameter image.

[0091] Furthermore, if Figure 7 As shown, based on the above-mentioned fast dynamic PET quantitative parameter imaging method and system based on the linear regression model, the present invention also provides a terminal, which includes a processor 10, a memory 20 and a display 30. Figure 7 Only some of the components of the terminal are shown, but it should be understood that implementation of all of the shown components is not required, and more or fewer components may be implemented instead.

[0092] In some embodiments, the memory 20 may be an internal storage unit of the terminal, such as a hard drive or memory of the terminal. In other embodiments, the memory 20 may also be an external storage device of the terminal, such as a plug-in hard drive, a Smart Media Card (SMC), a Secure Digital (SD) card, a flash memory card, etc. equipped with the terminal. Furthermore, the memory 20 may include both the internal storage unit of the terminal and an external storage device. The memory 20 is used to store application software installed on the terminal and various types of data, such as program code of the terminal. The memory 20 may also be used to temporarily store data that has been output or is about to be output. In one embodiment, the memory 20 stores a fast dynamic PET quantitative parameter imaging program 40 based on a linear regression model. The fast dynamic PET quantitative parameter imaging program 40 based on a linear regression model can be executed by the processor 10, thereby implementing the fast dynamic PET quantitative parameter imaging method based on a linear regression model described in the present application.

[0093] In some embodiments, the processor 10 can be a central processing unit (CPU), a microprocessor, or other data processing chip, used to run the program code or process data stored in the memory 20, such as executing the fast dynamic PET quantitative parameter imaging method based on the linear regression model.

[0094] In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. The display 30 is used to display information on the terminal and to display a visual user interface. The processor 10, memory 20, and display 30 of the terminal communicate with each other via a system bus.

[0095] In one embodiment, when the processor 10 executes the linear regression model-based rapid dynamic PET quantitative parameter imaging program 40 in the memory 20 , the steps of the linear regression model-based rapid dynamic PET quantitative parameter imaging method are implemented.

[0096] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a rapid dynamic PET quantitative parameter imaging program based on a linear regression model, and when the rapid dynamic PET quantitative parameter imaging program based on a linear regression model is executed by a processor, the steps of the rapid dynamic PET quantitative parameter imaging method based on a linear regression model as described above are implemented.

[0097] In summary, the present invention provides a method, system, terminal, and computer-readable storage medium for rapid dynamic PET quantitative parametric imaging based on a linear regression model. The method comprises: constructing a data set and drawing a region of interest in a dynamic PET image, dividing the dynamic sequence in the data set into a fitting data set and a test data set according to a preset ratio; constructing a standard Patlak analysis model based on the fitting data set, selecting a dynamic signal of a descending aorta slice as an arterial input function based on the standard Patlak analysis model; obtaining an early integral value and a late mean of the arterial input function, and establishing a linear regression model between the early integral value and the late mean using the fitting data set; extracting the late mean of the arterial input function from the test data set, and estimating the early integral value corresponding to the late mean in the test data set using the optimal linear fitting parameters of the linear regression model; combining the estimated early integral value with the standard Patlak analysis model, and calculating each voxel of the three-dimensional human body to obtain a voxel-level Patlak quantitative parametric image. The present invention achieves higher-quality parametric images using fewer dynamic scanning sequences, realizing short-term dynamic PET quantitative parametric imaging and providing a new solution for improving imaging efficiency and accuracy.

[0098] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or terminal comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or terminal. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or terminal comprising the element.

[0099] Of course, those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing related hardware (such as a processor, controller, etc.) through a computer program. The program can be stored in a computer-readable storage medium that can be read by a computer. When the program is executed, it can include the processes in the above-described method embodiments. The computer-readable storage medium can be a memory, a magnetic disk, an optical disk, etc.

[0100] It should be understood that the application of the present invention is not limited to the above examples. For those skilled in the art, improvements or changes can be made based on the above description. All these improvements and changes should fall within the scope of protection of the claims attached to the present invention.

Claims

1. A fast dynamic PET quantitative parameter imaging method based on a linear regression model, characterized in that: The rapid dynamic PET quantitative parameter imaging method based on the linear regression model includes: Constructing a data set and drawing a region of interest in a dynamic PET image, dividing the dynamic sequence in the data set into a fitting data set and a test data set according to a preset ratio; A standard Patlak analysis model is constructed according to the fitting data set, and based on the standard Patlak analysis model, a dynamic signal of a descending aorta slice is selected as an arterial input function; Obtaining an early integral value and a late mean value of the arterial input function, and establishing a linear regression model between the early integral value and the late mean value using the fitting data set; extracting a late mean value of the arterial input function from the test data set, and estimating an early integral value corresponding to the late mean value in the test data set using the best linear fit parameter of the linear regression model; The estimated early integral value is combined with the standard Patlak analysis model, and calculation is performed on each voxel of the three-dimensional human body to obtain a voxel-level Patlak quantitative parameter image.

2. The rapid dynamic PET quantitative parameter imaging method based on the linear regression model according to claim 1, characterized in that: The region of interest represents a specific three-dimensional region selected in a dynamic PET image, and is used to analyze and quantify the distribution and dynamic changes of the radioactive tracer in the specific three-dimensional region.

3. The rapid dynamic PET quantitative parameter imaging method based on the linear regression model according to claim 1, characterized in that: The standard Patlak analysis model was used to image the pharmacokinetic parameters of irreversible tracers.

4. The rapid dynamic PET quantitative parameter imaging method based on the linear regression model according to claim 1 or 3, characterized in that: The expression of the standard Patlak analysis model is: Among them, C T (t) and C P (t) represent the tissue time-activity curve and arterial input function, t represents the time, t * represents the moment when the tracer reaches a steady state, represents the area under the plasma concentration-time curve, K i represents the slope and V represents the intercept.

5. The rapid dynamic PET quantitative parameter imaging method based on the linear regression model according to claim 1, characterized in that: The step of obtaining the early integral value and the late mean value of the arterial input function and establishing a linear regression model between the early integral value and the late mean value using the fitting data set specifically includes: Get the early integral value S of the arterial input function early and late mean M late ; Modify the expression of the standard Patlak analysis model to: Among them, S early represents the early integral value of the arterial input function, and t1 represents the starting time of the scan; Early integral value S early The expression is: A linear regression model is established between the early integral value and the late mean value, and the expression is: S early =k·M late +b; (4) Among them, M late represents the late mean of the arterial input function, k and b are the slope and intercept parameters of the linear regression model, respectively.

6. The rapid dynamic PET quantitative parameter imaging method based on the linear regression model according to claim 1, characterized in that: The early integral value represents the integral value of the AIF curve before a certain frame node, and the late mean value represents the mean value of the AIF curve after a certain frame node.

7. The rapid dynamic PET quantitative parameter imaging method based on the linear regression model according to claim 1, characterized in that: The preset ratio is 7:

3.

8. A fast dynamic PET quantitative parameter imaging system based on a linear regression model, characterized in that: The fast dynamic PET quantitative parameter imaging system based on the linear regression model includes: A data set construction module is used to construct a data set and draw a region of interest in a dynamic PET image, and divide the dynamic sequence in the data set into a fitting data set and a test data set according to a preset ratio; a signal selection module, configured to construct a standard Patlak analysis model according to the fitted data set, and select a descending aorta slice dynamic signal as an arterial input function based on the standard Patlak analysis model; a model building module, configured to obtain an early integral value and a late mean value of the arterial input function, and to establish a linear regression model between the early integral value and the late mean value using the fitting data set; a numerical estimation module, configured to extract a late mean value of the arterial input function from the test data set, and estimate an early integral value corresponding to the late mean value in the test data set using the best linear fitting parameter of the linear regression model; The image generation module is used to combine the estimated early integral value with the standard Patlak analysis model, calculate each voxel of the three-dimensional human body, and obtain a voxel-level Patlak quantitative parameter image.

9. A terminal, characterized in that: The terminal includes: a memory, a processor, and a fast dynamic PET quantitative parameter imaging program based on a linear regression model stored in the memory and executable on the processor. When the fast dynamic PET quantitative parameter imaging program based on a linear regression model is executed by the processor, the steps of the fast dynamic PET quantitative parameter imaging method based on a linear regression model are implemented.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a rapid dynamic PET quantitative parameter imaging program based on a linear regression model. When the rapid dynamic PET quantitative parameter imaging program based on a linear regression model is executed by a processor, the steps of the rapid dynamic PET quantitative parameter imaging method based on a linear regression model as described in any one of claims 1 to 7 are implemented.

Images