A method for real-time processing and analysis of magnetic resonance scan data

Abstract

The application provides a real-time processing and analysis method for magnetic resonance scanning data, comprising: capturing the change process of intravascular contrast agent concentration and magnetic resonance signal intensity with time and the spatial distribution of magnetic field uniformity in real time, and identifying the signal attenuation amplitude and signal intensity decline rate caused by T2 star effect; for the determined arterial phase time starting point, combining the signal enhancement amplitude caused by T1 effect and the signal attenuation amplitude caused by T2 star effect, continuously monitoring the peak arrival time and peak duration in the signal intensity change process, and evaluating the optimal duration of the acquisition time window; according to the evaluated optimal duration of the acquisition time window, generating a trigger signal when the T1 effect is dominant and the T2 star effect has not yet significantly affected the signal intensity, activating the image acquisition of the scanning system, and obtaining pure arterial phase images.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a method for real-time processing and analysis of magnetic resonance scanning data. Background Technology

[0002] In magnetic resonance angiography (MRA), contrast-enhanced MRA (CE-MRA) has long been the preferred diagnostic method for vascular lesions in the head, neck, chest, abdomen, and limbs due to its radiation-free nature and excellent soft tissue contrast. Clinicians ideally desire pure arterial phase images, where arteries appear bright with high signal intensity while veins maintain extremely low signal intensity, allowing for clear separation of arteries and veins and facilitating accurate assessment of vascular stenosis, aneurysms, and vascular malformations. However, in practice, doctors often find that bright arteries suddenly darken or even disappear completely within seconds, with the peak signal intensity becoming the darkest. The root cause of this phenomenon lies in the dual mechanism of action of gadolinium contrast agents. Gadolinium is a paramagnetic substance; when its concentration in blood vessels is low to moderate, it primarily shortens the T1 relaxation time of blood, resulting in a significant signal enhancement in T1-weighted sequences, manifesting as extremely bright vascular images – the desired positive effect. However, when the contrast agent concentration increases further and exceeds a certain critical value, paramagnetic gadolinium ions generate extremely strong magnetic field inhomogeneities locally, causing rapid proton dephase and a sharp shortening of the T2 relaxation time, ultimately resulting in severe signal loss, and the blood vessels appear black instead. This signal quenching phenomenon at high concentrations is very common clinically, especially when using double or even triple dose rapid injection protocols, and is almost unavoidable. Typical examples include the sudden "disappearance" of the main renal artery in renal artery MRA at the moment the contrast agent peak is reached, the appearance of a ring-shaped black border in the cavernous segment of the internal carotid artery in carotid artery MRA, and the "black blood" appearance of the main pulmonary artery in pulmonary artery MRA. These phenomena can be mistaken for the end of the arterial phase or misjudged as severe stenosis, directly leading to diagnostic errors. At the same time, even without obvious signal quenching, another more common problem is the extremely short first-pass time window of the artery. Normally, after the contrast agent is injected into the elbow vein, it needs to pass through the superior vena cava, right heart, pulmonary circulation, left heart, and then into the systemic artery. This process varies greatly among individuals, influenced by factors such as cardiac function, blood pressure, respiratory status, and injection rate. The actual time to reach peak arterial contrast can fluctuate between 15 and 35 seconds. The ideal concentration window for achieving pure arterial contrast is often only 3-8 seconds. If it is too early, the artery has not yet fully enhanced; if it is too late, the vein has already begun to contrast. Once these few seconds are missed, the image enters the arteriovenous overlap period, where both arteries and veins appear bright white simultaneously. Doctors cannot distinguish which is which, especially in densely vascularized areas such as the head, neck, and abdomen, significantly reducing diagnostic value.

[0003] To address these issues, various technological attempts have emerged over the past two decades, including low-dose bolus injections combined with fixed delay times, real-time fluoride triggering technology, and post-hoc selection of the optimal phase after multi-phase dynamic acquisition. However, these methods all have fundamental limitations. They cannot detect the true intravascular gadolinium concentration in real time, nor can they predict the ebb and flow of the T1 enhancement and T2 quenching effects every second. Therefore, even with the most advanced triggering technology, a significant proportion of clinical examinations still require repeated scans, or the final image quality is unsatisfactory, necessitating subsequent CTA or DSA for supplementary diagnosis. Because the dynamic changes in contrast agent concentration within the blood vessel are completely "invisible," doctors can only rely on experience or indirect methods to guess the optimal acquisition time. If the guess is incorrect, it inevitably leads to two disastrous consequences: signal quenching causing vascular darkening, or arteriovenous overlap causing structural confusion. Summary of the Invention

[0004] This invention provides a method for real-time processing and analysis of magnetic resonance scanning data, mainly including: Signal data of the vascular region were acquired by magnetic resonance imaging (MRI) to obtain the changes in contrast agent concentration and signal intensity, as well as the spatial distribution of magnetic field homogeneity. Based on the signal data, identify the signal attenuation amplitude and rate of decline caused by the T2 effect, and assess the correlation strength between contrast agent concentration and magnetic field homogeneity. The arrival time of the first-pass peak concentration and the signal intensity performance were extracted from the changes in contrast agent concentration and signal intensity to analyze the conversion process between the T1 effect and the T2 effect. Based on the signal attenuation rate, a corrected curve is generated to determine the signal strength deviation amplitude at the transition critical point between the T1 and T2 effects. Based on the deviation amplitude, the T1 effect-dominant interval and the T2 effect-dominant interval are divided to identify the arterial phase start point and evaluate the optimal duration of the acquisition time window. A trigger signal is generated based on the optimal duration to activate image acquisition and obtain arterial phase images; By comparing the changes in the arterial phase images with those in the initial signal, the effect identification parameters are updated, thereby optimizing the timing of data acquisition.

[0005] Furthermore, signal data from the vascular region is acquired using magnetic resonance imaging (MRI) sequences to obtain the changes in contrast agent concentration and signal intensity, as well as the spatial distribution of magnetic field homogeneity, including: Raw data of the vascular region were acquired using a multi-echo gradient echo sequence. The local magnetic field offset at each voxel location was calculated based on the phase accumulation difference at different echo times, generating a spatial distribution map of magnetic field inhomogeneity. For the aforementioned spatial distribution map, the T2 star relaxation rate is extracted from the ratio of signal intensity of short echo time to long echo time. If the ratio is lower than a preset threshold, it is determined that there is a high concentration of contrast agent accumulation in the region, and an initial estimate of the contrast agent concentration distribution is obtained. Based on the initial estimate, the signal strength at consecutive time points is differentially calculated using the sliding window method. The signal before injection is used as the baseline, and the instantaneous rate of decrease relative to the baseline signal is calculated to determine the time point at which the T2 effect begins to dominate.

[0006] Furthermore, after obtaining the changes in contrast agent concentration and signal intensity, as well as the spatial distribution of magnetic field homogeneity, the following steps are taken: A spatiotemporal feature matrix is ​​constructed from the signal attenuation amplitude and descent rate. The correlation coefficient is used to calculate the linear correlation between the contrast agent concentration value and the magnetic field offset at each voxel location, and a correlation intensity distribution map is generated. Based on the distribution map, regions with correlation coefficients higher than a preset threshold are identified, and the attenuation start point, peak point, and stable point are determined from the signal time series of these regions to obtain the time window of the T2 effect. For signal data within the time window, if the contrast agent concentration exceeds a preset critical value, the signal amplitude and phase changes are separated, feature vectors are extracted through spectrum analysis, and principal component analysis is used to obtain signal change characteristics.

[0007] Furthermore, the conversion process between the T1 effect and the T2 effect is analyzed, including: By continuously monitoring the signal strength time curve, the instantaneous slope of adjacent time points is calculated, and the moment corresponding to the maximum slope is identified as the arrival time of the first peak concentration. The signal strength value and the estimated concentration value before and after this moment are recorded. When the peak concentration is lower than the preset threshold, the signal enhancement amplitude caused by the T1 effect is extracted, and the signal contrast between the vascular region and the surrounding tissue is calculated. If the peak concentration exceeds the critical value, the T1 effect enhancement rate and the T2 effect decay rate are calculated, and the change in the ratio of the T1 effect enhancement rate to the T2 effect decay rate is monitored to determine the time point of the effect dominance transition and the corresponding concentration value.

[0008] Furthermore, the signal strength deviation amplitude at the transition critical point between the T1 and T2 effects is determined, including: The baseline drift of the signal attenuation rate curve was removed by a fitting method to generate a corrected rate curve. According to the corrected rate curve, when the rate value is positive, it is identified as the T1 effect dominant segment, and when the rate value is negative, it is identified as the T2 effect dominant segment. The zero point where the rate changes from positive to negative can be determined by finding adjacent data points where the sign of the rate value changes. Based on the zero-point position, the signal strength value at the corresponding time is extracted, and the difference between it and the baseline signal strength is calculated to obtain the signal strength deviation amplitude at the effect transition critical point.

[0009] Furthermore, based on the magnitude of the deviation, the T1-effect dominant interval and the T2-effect dominant interval are divided to identify the arterial phase time start point, including: Based on the signal intensity deviation amplitude, determine the lower limit of the concentration in the T1 effect-dominant region and the upper limit of the concentration in the T2 effect-dominant region, and obtain the concentration range boundaries of the two regions. Using the concentration range boundary, time-series data of signal intensity within the T1 effect-dominant region are extracted; The direction of rate change is determined by calculating the first derivative. If the derivative changes from positive to negative, it is marked as the peak time, thus identifying the starting point of the arterial phase.

[0010] Furthermore, the optimal duration of the data acquisition time window was evaluated, including: Starting from the arterial phase time start point, record the signal enhancement amplitude value generated by the T1 effect and the signal attenuation amplitude value generated by the T2 effect. Obtain the signal intensity time series through continuous sampling to determine the time when the signal reaches its peak. Based on the moment when the signal reaches its peak, calculate the continuous time period during which the signal strength remains above a preset percentage. If the signal strength drops by more than the preset percentage, mark it as the end of the peak and obtain the duration of the peak. The duration of the peak value is compared with the pre-determined contrast agent duration over time to determine the duration of the acquisition time window.

[0011] Furthermore, by comparing the changes in the arterial phase images with those in the initial signal, the effect identification parameters are updated to optimize the timing of data acquisition, including: By comparing the average signal intensity of the vascular region in the arterial phase image with the value of the initial signal intensity time curve at the corresponding time, the signal deviation value is calculated, the time when the T2 effect has a significant impact is identified, and the actual effect transition point is obtained. Update the effect identification parameters in the historical acquisition records based on the actual effect transition point; Using the effect identification parameters from the updated historical acquisition records, a mapping table between contrast agent concentration and signal attenuation amplitude was established. The predicted values ​​for different concentration points were refined by interpolation, and a reference curve was generated. Based on the reference curve, the trigger timing and acquisition duration parameters are adjusted to optimize the acquisition timing control.

[0012] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: This invention discloses a real-time processing and analysis method for magnetic resonance imaging (MRI) scan data. Addressing the complex influence of intravascular contrast agent concentration variations and magnetic field homogeneity distribution on signal intensity, it integrates the identification of signal attenuation and enhancement effects, the determination of critical transition points, and the practical challenges of acquiring clean arterial phase images. This invention captures real-time changes in contrast agent concentration and signal intensity over time, analyzes the signal enhancement characteristics dominated by the T1 effect and the signal attenuation characteristics dominated by the T2 effect, accurately extracts the effect transition critical points and corresponding concentration ranges, and, combined with the identification of the arterial phase start point and the evaluation of the optimal acquisition time window, generates a trigger signal to activate image acquisition, obtaining arterial phase images free from venous contamination. Simultaneously, it achieves closed-loop optimization through historical data updates. The core innovation of this invention lies in the combination of dynamic monitoring and adaptive adjustment, ensuring precise signal acquisition timing, significantly improving image quality and diagnostic accuracy, and providing an efficient solution for MRI. Attached Figure Description

[0013] Figure 1 This is a flowchart of a real-time processing and analysis method for magnetic resonance scanning data according to the present invention.

[0014] Figure 2 This is a schematic diagram of a real-time processing and analysis method for magnetic resonance scanning data according to the present invention.

[0015] Figure 3 This is another schematic diagram of a real-time processing and analysis method for magnetic resonance scanning data according to the present invention. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0017] like Figure 1-3 This embodiment of a method for real-time processing and analysis of magnetic resonance scanning data may specifically include: S101. Real-time capture of the changes in intravascular contrast agent concentration and magnetic resonance signal intensity over time, as well as the spatial distribution of magnetic field homogeneity, identifying the signal attenuation amplitude and signal intensity decrease rate caused by the T2 asterisk effect.

[0018] Raw K-space data of the vascular region was acquired using a multi-echo gradient echo sequence. Based on the phase accumulation difference at different echo times, the local magnetic field offset at each voxel location was calculated, yielding a spatial distribution map of magnetic field inhomogeneity. For this magnetic field inhomogeneity spatial distribution map, the T2 star relaxation rate was extracted from the signal intensity ratio of short echo time to long echo time. If the ratio was less than a first preset threshold, it was determined that a high concentration of contrast agent was present in the region, obtaining an initial estimate of the contrast agent concentration distribution. Based on this initial estimate, a sliding window method was used to perform differential calculations on the signal intensity at consecutive time points. The signal acquired before contrast agent injection was used as the baseline signal, and the instantaneous rate of decline relative to the baseline signal was calculated. When the rate exceeded a second preset threshold, the time point at which the T2 star effect began to dominate was determined. A correspondence table between contrast agent concentration and signal attenuation amplitude was established using the reconstructed signal intensity sequence from the K-space data acquired before and after the time point. The concentration critical value corresponding to when the signal attenuation exceeds a third preset threshold was extracted from the table, obtaining a quantitative identification standard for the T2 star effect.

[0019] Specifically, in one implementation, the timing and extent of the T2 asterisk effect are accurately identified by real-time monitoring of magnetic resonance signal intensity fluctuations caused by changes in intravascular contrast agent concentration. In practice, the gradient echo sequence parameters of the magnetic resonance scanner are first configured, setting multiple different echo times, typically three to five echo time points, with the echo time interval increasing in the range of 2 to 5 milliseconds, thereby capturing phase evolution information at different times.

[0020] Specifically, the multi-echo gradient echo sequence acquisition process begins with the application of a radio frequency pulse excitation, continuously acquiring multiple echo signals within each repetition time interval. The K-space data acquisition employs a Cartesian sampling trajectory, applying a readout gradient in the frequency encoding direction and phase encoding gradients of varying intensities in the phase encoding direction. The K-space data corresponding to each echo contains the phase information at that moment; the phase difference at different echo times directly reflects the degree of inhomogeneity of the local magnetic field. By performing phase unwrapping processing on the complex signal at each voxel location, the phase difference between adjacent echoes is calculated, and then divided by the echo time interval to obtain the frequency offset of that voxel. The frequency offset is then converted into a magnetic field offset, with the conversion relationship being magnetic field offset equal to frequency offset divided by the gyromagnetic ratio, forming a spatial distribution map of magnetic field inhomogeneity across the entire imaging field of view.

[0021] In one possible implementation, the spatial distribution map of magnetic field inhomogeneity reflects the degree to which the local magnetic field at each voxel location deviates from the main magnetic field. The greater the deviation, the worse the magnetic field homogeneity at that location, which usually corresponds to a region with a higher contrast agent concentration.

[0022] For example, the extraction of the T2 star relaxation rate is based on the exponential decay of signal intensity with echo time. In practice, for each voxel location, the signal intensity at different echo times is logarithmically transformed, and the slope of the decay curve is obtained through linear fitting. The negative value of this slope is the T2 star relaxation rate. T2 star relaxation includes contributions from both the tissue's inherent T2 relaxation and additional signal attenuation caused by magnetic field inhomogeneity. When the ratio of signal intensity at short echo time to long echo time decreases, it indicates that T2 star relaxation is accelerating and signal attenuation is enhancing. By setting a first preset threshold, when the ratio is below this threshold, it is determined that there is a high concentration of contrast agent accumulation in the region. The initial estimate is obtained based on an empirical formula that establishes an approximately linear relationship between the increase in T2 star relaxation rate and contrast agent concentration; each increase in relaxation rate corresponds to a specific increase in contrast agent concentration.

[0023] Preferably, after obtaining an initial estimate of the contrast agent concentration distribution, a dynamic monitoring method based on the time dimension is further employed. The sliding window method sets a fixed-length time window containing several continuously acquired time points, and the dynamic trend of the signal is evaluated by calculating the rate of change of signal intensity within the window.

[0024] In one embodiment, the sliding window implementation process includes selecting signal intensity data at five consecutive time points, calculating the difference between adjacent time points, smoothing the difference sequence to eliminate noise effects, and obtaining the instantaneous rate of change. The baseline signal is obtained from a steady-state scan before contrast agent injection, typically using the average signal intensity within 10 to 20 seconds before injection as a reference. When the signal decline rate relative to the baseline signal exceeds a second preset threshold, it indicates that the effect of the T2 asterisk effect begins to significantly exceed the T1 enhancement effect; at this point, the corresponding time point is recorded as a critical node for effect transition. For example, in renal artery imaging, when the contrast agent reaches the main renal artery trunk approximately 15 seconds after injection, if the concentration is too high, a rapid signal drop will occur within 2 to 3 seconds, with the rate of drop exceeding 20% ​​per second of the baseline signal intensity. This abrupt change is a hallmark feature of the dominant T2 asterisk effect. Determining this time point is crucial for subsequent imaging quality. This time point allows for retrospective analysis of the time interval during which the T1 enhancement effect dominates, and also allows for prediction of the time range during which the T2 asterisk effect will continue to influence the image. Multiple sets of K-space data are selected before and after the time point, and corresponding image sequences are obtained through Fourier transform reconstruction. The average signal intensity value of the vascular region is extracted from the images. Establishing the correspondence table requires comprehensive consideration of multiple influencing factors. By pairing the estimated contrast agent concentration at different times with the corresponding signal attenuation amplitude to form a data point set, a continuous mapping relationship is established using polynomial fitting or spline interpolation methods. Optionally, a continuous mapping relationship between contrast agent concentration and signal attenuation amplitude can be established using the following formula: S(t) represents the signal attenuation magnitude at time t, C(t) represents the estimated contrast agent concentration at time t, and a k denoted by , k-th order coefficients of the polynomial fitting, and n represents the highest order of the polynomial. This formula establishes a continuous mapping relationship between contrast agent concentration and signal attenuation amplitude using a polynomial fitting method.

[0025] Alternatively, a smooth mapping function between contrast agent concentration and signal attenuation can be constructed using the following formula: F(x) represents the output value of the spline interpolation function, x represents the input contrast agent concentration value, and c i B represents the coefficient of the i-th spline basis function. i (x) represents the i-th spline basis function, and m represents the total number of spline basis functions. This formula uses spline interpolation to construct a smooth mapping function between contrast agent concentration and signal attenuation. When the signal attenuation exceeds a third preset threshold, the corresponding concentration value is retrieved from the relation table; this concentration value is the critical concentration at which the T2 asterisk effect significantly affects image quality. Through the above implementation method, quantitative identification and prediction of the T2 asterisk effect are achieved. This scheme can accurately capture the optimal imaging time window during the first pass of the contrast agent, significantly improving the diagnostic value of vascular imaging.

[0026] S102. Based on the identified signal attenuation amplitude and signal intensity decrease rate, assess the correlation strength between contrast agent concentration and spatial distribution of magnetic field homogeneity, extract the signal change characteristics dominated by the T2 star effect when the contrast agent concentration exceeds the critical value, and obtain the signal attenuation rate curve dominated by the T2 star effect.

[0027] From the signal attenuation amplitude and descent rate, a spatiotemporal feature matrix is ​​constructed according to time series and spatial location. Each row of the matrix represents a voxel location, and each column represents the signal intensity value at different time points. The Pearson correlation coefficient is used to calculate the linear correlation between the contrast agent concentration value and the magnetic field offset at each voxel location, resulting in a correlation intensity distribution map. Based on the correlation intensity distribution map, high correlation regions with correlation coefficients greater than a preset threshold are identified. From the signal time series of these regions, the attenuation start point is determined by the moment when the first derivative changes from a positive or zero value to a negative value, the attenuation peak point is determined by the minimum value of the first derivative, and the attenuation stabilization point is determined by the signal change rate being less than the baseline noise level, thus obtaining the time window of the T2 star effect. For the signal data within the time window, if the contrast agent concentration exceeds a preset critical value, the signal amplitude is separated from the original complex signal, the phase difference between adjacent time points is calculated to obtain the phase change, and the spectral shift is obtained through Fourier transform. The data from these three dimensions are combined to form a feature vector, and the first three principal components are extracted through principal component analysis as the signal change features dominated by the T2 star effect. Using the aforementioned signal change characteristics, the attenuation rate at different time points is nonlinearly fitted, and the fitting parameters are determined by the least squares method to generate the signal attenuation rate curve dominated by the T2 asterisk effect.

[0028] Specifically, in one implementation, the construction of the spatiotemporal feature matrix begins with the organization of the raw magnetic resonance signal data. The matrix adopts a two-dimensional structure, where each row corresponds to the position coordinates of a specific voxel within the imaging field of view, and each column represents the signal measurement value at a specific acquisition time point.

[0029] Specifically, for a typical abdominal vascular imaging involving 10,000 voxels and 50 time sampling points, the constructed matrix has a dimension of 10,000×50, and each element in the matrix stores the absolute value of the signal intensity of that voxel at the corresponding time.

[0030] It should be noted that the Pearson correlation coefficient in this application measures the degree of linear correlation between two variables. The calculation first obtains the estimated contrast agent concentration and magnetic field offset sequences for each voxel location at all time points. Then, the covariance of the two sequences is calculated and divided by the product of their respective standard deviations. The correlation coefficient ranges from -1 to 1, with values ​​close to 1 indicating a strong positive correlation, close to -1 indicating a strong negative correlation, and close to 0 indicating no linear correlation. In magnetic resonance angiography, when contrast agent accumulation leads to local magnetic field inhomogeneity, the two variables show a high positive correlation, with the correlation coefficient typically greater than 0.8.

[0031] Preferably, the correlation strength distribution map is generated using a pseudo-color coding method, which maps the correlation coefficient values ​​to different colors, with red representing high correlation areas and blue representing low correlation areas, making it easy to intuitively identify the spatial distribution characteristics of the T2 asterisk effect.

[0032] For example, the identification of the three key time points is based on the mathematical characteristics of the signal time curve. The attenuation start point is determined by calculating the first derivative of the signal strength with respect to time; attenuation begins when the derivative changes from a positive or zero value to a negative value. In actual processing, the signal curve is smoothed using a five-point smoothing filter to eliminate noise interference, and then the derivative value is calculated using the central difference method. The attenuation peak point corresponds to the extreme position of the second derivative, indicating the moment when the signal decline rate reaches its fastest. The determination of the attenuation stabilization point is based on the comparison between the signal rate of change and the baseline noise; when the signal rate of change at three consecutive time points is less than twice the standard deviation of the baseline noise, the signal is considered to have entered a stable period. These three time points together define the effective time window of the T2 asterisk effect, typically lasting 5 to 10 seconds.

[0033] In one possible implementation, when the contrast agent concentration exceeds a preset threshold, signal feature extraction involves multi-dimensional decomposition of the complex signal. The original magnetic resonance signal contains two components: a real part and an imaginary part. The signal amplitude is obtained by calculating the complex modulus, i.e., the square root of the sum of the squares of the real and imaginary parts. Phase information is extracted from the argument of the complex signal; the phase difference between adjacent time points reflects the degree of frequency shift. The spectral shift is obtained through a Fast Fourier Transform (FFT), converting the time-domain signal to the frequency domain. The shift of the peak frequency relative to the center frequency quantitatively characterizes the degree of inhomogeneity of the local magnetic field.

[0034] Specifically, Principal Component Analysis (PCA) plays a crucial role in feature dimensionality reduction. The eigenvectors consist of three components: signal amplitude sequence, phase change sequence, and spectral shift sequence, each being a one-dimensional array of the same length. PCA first calculates the covariance matrix of the feature data, then solves for the eigenvalues ​​and eigenvectors of the covariance matrix. The eigenvalues ​​are sorted from largest to smallest, and the eigenvectors corresponding to the three largest eigenvalues ​​are selected as the principal component directions. The original feature data is projected onto these three principal component directions to obtain the dimensionality-reduced feature representation, retaining over 95% of the information in the data while eliminating redundancy and noise.

[0035] For example, during carotid artery imaging, when the contrast agent concentration reaches a critical value, the signal amplitude drops by more than 60% from its peak within 2 seconds, the phase shifts by more than 180 degrees in the same time period, and the main peak of the spectrum deviates from the center frequency by more than 50 Hz. These characteristics together constitute the typical performance pattern of the T2 asterisk effect.

[0036] Understandably, nonlinear fitting uses an exponential decay model to describe the change in signal strength over time. The fitting function is defined as signal strength equal to the initial strength multiplied by the product of the natural exponent of time and the decay constant raised to the power of the product, plus the background signal offset. The least squares method iteratively optimizes and solves for three fitting parameters: initial signal strength, decay time constant, and background offset. Furthermore, the fitting process employs a weighted least squares method, assigning different weights to each data point based on its signal-to-noise ratio (SNR). Data points with higher SNR receive greater weights, thus improving the stability and accuracy of the fitting. Iterative optimization uses the Levenberg-Marquardt algorithm, combining the advantages of gradient descent and Gauss-Newton methods, typically converging to the optimal solution within 10 iterations.

[0037] In one embodiment, the obtained signal attenuation rate curve exhibits a characteristic biphasic change pattern: the initial slow decline corresponds to the stage dominated by the T1 effect, the middle rapid decline corresponds to the stage where the T2 asterisk effect is enhanced, and the later trend towards flattening corresponds to the stage where the effect saturates.

[0038] The arrival time of the first-pass peak concentration and the signal intensity performance of the vascular region are obtained from the real-time captured contrast agent concentration changes. The signal enhancement phenomenon and bright display effect of the vascular region caused by the T1 effect when the contrast agent concentration is low are analyzed. The transformation process of the T2 star effect replacing the T1 effect as the dominant factor after the peak contrast agent concentration exceeds the critical threshold is evaluated. The proton dephase acceleration phenomenon caused by local magnetic field inhomogeneity and the signal loss state of the vascular region from bright to dark are identified. The range of signal attenuation caused by the T2 star effect in the highly concentrated contrast agent area is determined.

[0039] By continuously monitoring the signal intensity time curve of the vascular region, the instantaneous slope is calculated by dividing the signal intensity difference between adjacent time points by the time interval. The moment when the slope reaches its maximum value is identified as the arrival time of the first-pass peak concentration. The signal intensity values ​​and estimated contrast agent concentrations at each time point before and after this moment are recorded to obtain a dataset of the correspondence between concentration and signal intensity. Based on this dataset, when the contrast agent concentration is lower than a preset threshold, the signal enhancement amplitude caused by the shortening of the T1 relaxation time is extracted. The signal contrast is calculated by subtracting the signal intensity of the surrounding muscle tissue from the signal intensity of the vascular region and then dividing by the signal intensity of the muscle tissue, thus obtaining a quantitative index of bright vascular display. For this quantitative index, if the peak concentration exceeds a critical threshold, the signal increase per unit time caused by the T1 effect is calculated from the signal time curve as the enhancement rate, and the signal decrease per unit time caused by the T2 asterisk effect is calculated as the attenuation rate. The change in the ratio of the two is monitored. When the ratio changes from greater than 1 to less than 1, the time point of the effect dominance transition and the corresponding concentration value are determined. Using the time points and concentration values, the frequency shift of each voxel is calculated by phase diagrams of different echo times in the gradient echo sequence to track the proton precession frequency difference. The phase accumulation rate is obtained by dividing the phase difference between adjacent echo time points by the time interval to identify the dephase acceleration region. The spatial range of signal attenuation caused by the T2 asterisk effect is determined by the distribution of voxels whose signal intensity drops beyond the average signal intensity threshold before injection.

[0040] Specifically, in one implementation, accurate identification of the first-pass peak concentration arrival time relies on a high temporal resolution dynamic scanning sequence. In practice, a continuous acquisition mode is used, acquiring a complete vascular image every 0.5 seconds to form a time-series dataset. The instantaneous slope is calculated using the three-point difference method. For the signal intensity S(t) at time point t, the slope value is equal to the difference between S(t+1) and S(t-1), divided by twice the time interval. In actual processing, the original signal undergoes Gaussian filtering preprocessing, with the filter kernel width set to three time points to eliminate high-frequency noise interference before slope calculation. The position corresponding to the maximum value in the slope sequence is the first-pass peak arrival time. Data from 10 time points before and after this moment constitute the core part of the concentration-signal intensity correlation dataset. The preset threshold is determined based on statistical analysis of a large amount of clinical data. In the low concentration range, the contrast agent mainly enhances the signal by shortening the T1 relaxation time; the vascular signal intensity in the T1-weighted sequence shows an approximately linear relationship with the contrast agent concentration. The signal enhancement magnitude is quantified by the difference between the peak signal intensity and the baseline signal intensity, which is extracted from a steady-state scan within 20 seconds prior to injection.

[0041] For example, the calculation of signal contrast involves the selection of the target blood vessel and the reference tissue. In abdominal vascular imaging, adjacent muscle tissue is typically chosen as the reference, while in neck vascular imaging, the neck muscles are selected. The contrast calculation formula is the average signal intensity of the vascular region minus the average signal intensity of the muscle tissue, and then divided by the average signal intensity of the muscle tissue, yielding a dimensionless contrast value. When the contrast value exceeds 2.0, the blood vessel exhibits a bright, high signal, forming a clear contrast with the surrounding tissue. As the contrast agent concentration continues to increase, the contrast value begins to decrease, indicating the onset of the T2 asterisk effect.

[0042] Preferably, the dynamic monitoring of quantitative indicators adopts a sliding window method, with the window width set to 5 time points. Each time point is slid forward by one time point, and the contrast value is updated in real time.

[0043] In one possible implementation, the T1 enhancement rate and T2 asterisk effect attenuation rate are calculated based on piecewise linear fitting of the signal-time curve. For the rising phase, the time interval from the baseline level to 90% of the peak signal intensity is selected, and the rising slope is obtained by least squares fitting. This slope is the T1 enhancement rate, expressed as a percentage change in signal intensity per second. For the falling phase, consecutive time points where the signal intensity decreases by more than 10% after the peak are selected, and the absolute value of the falling slope is obtained by linear fitting as the T2 asterisk effect attenuation rate. The ratio of the two rates reflects the relative strength of the two effects; a ratio greater than 1 indicates that the T1 effect is dominant, less than 1 indicates that the T2 asterisk effect is dominant, and a ratio of 1 is the transition threshold. In renal artery imaging, this transition typically occurs within 3 to 5 seconds after contrast agent arrival, corresponding to a concentration approximately 8 to 12 times the plasma concentration.

[0044] Specifically, proton precession frequency differences are tracked using multi-echo phase imaging. The phase map acquired at each echo time contains phase information for all voxels at that moment, and the frequency shift map is obtained by dividing the phase difference between adjacent echoes by the echo time interval. The larger the local magnetic field gradient, the more drastic the spatial variation in proton precession frequency. In regions with high contrast agent concentrations, the frequency difference between adjacent voxels can reach hundreds of hertz, leading to rapid signal dephasing. The phase accumulation rate is obtained by calculating the phase change angle per unit time; when the accumulation rate exceeds 10 degrees per millisecond, the region is considered to have entered a rapid dephasing state.

[0045] For example, in the cavernous sinus segment of the internal carotid artery, due to the tortuous course of the vessel and changes in blood flow velocity, contrast agent tends to accumulate locally, forming ring-shaped high-concentration regions. The phase accumulation rate in these regions can reach over 15 degrees per millisecond, leading to significant signal loss at the vessel edges. The acquisition of the average signal intensity before injection needs to consider physiological fluctuations. In practice, baseline data is continuously collected for 30 seconds before contrast agent injection, and the mean and standard deviation are calculated for all time points. The mean is used as the baseline level, and three times the standard deviation is used as the threshold for determining a significant signal decrease. When the signal intensity of a voxel is lower than the baseline level minus three times the standard deviation, that voxel is marked as a signal attenuation region. Furthermore, the spatial range of signal attenuation is determined using connected component analysis. All voxels marked as signal attenuation undergo three-dimensional connectivity checks, and adjacent attenuating voxels are grouped into the same connected component. The volume and shape characteristics of each connected component reflect the spatial distribution pattern of the T2 asterisk effect.

[0046] S103. Baseline correction and adjustment are performed on the obtained signal attenuation rate curve to identify the signal enhancement characteristics dominated by the T1 effect when the contrast agent concentration is low and the signal attenuation characteristics dominated by the T2 asterisk effect when the concentration is high. The signal intensity deviation amplitude corresponding to the transition critical point of the two effects is extracted.

[0047] The signal attenuation rate curve was baseline-shifted using a polynomial fitting method. The low-frequency trend term was fitted using the least squares method, and the trend term was subtracted from the original curve to obtain the corrected rate curve. Based on the corrected rate curve, a positive rate value was identified as a signal enhancement segment dominated by the T1 effect, and a negative rate value was identified as a signal attenuation segment dominated by the T2 asterisk effect. The zero-point position where the rate value sign changed was determined by finding adjacent data points and performing linear interpolation. The signal intensity value at this zero-point position was extracted, and the difference between this value and the baseline signal intensity acquired before contrast agent injection was calculated to obtain the signal intensity deviation amplitude corresponding to the transition point between the two effects.

[0048] Specifically, in one implementation, baseline drift removal is achieved using third-order polynomial fitting. Specifically, the time axis of the signal attenuation rate curve is used as the independent variable, and the rate value as the dependent variable. The polynomial coefficients are solved using the least squares method. The polynomial expression includes constant terms, linear terms, quadratic terms, and cubic terms. The fitted curve represents the low-frequency drift trend of the signal. Specifically, the low-frequency drift trend of the signal attenuation rate curve is fitted using the following formula: P(t) represents the third-order polynomial fitting function, t represents the time axis independent variable, a0 represents the constant term coefficient, a1 represents the linear term coefficient, a2 represents the quadratic term coefficient, and a3 represents the cubic term coefficient. The corrected rate curve is obtained by subtracting the fitted trend curve values ​​point by point from the original rate curve. The corrected rate curve exhibits a biphasic characteristic. In the initial stage of contrast agent arrival, the T1 relaxation time shortening effect dominates, and the rate value is positive, indicating a continuous increase in signal intensity. As the concentration increases, the T2 asterisk effect gradually strengthens, the rate value begins to decrease and turns negative, indicating that the signal intensity begins to decay.

[0049] For example, the precise determination of the zero point is achieved by scanning the corrected rate curve and finding the position where the rate values ​​of two adjacent data points have opposite signs. Let the time of the previous point be t1 and the rate be r1, and the time of the next point be t2 and the rate be r2. The zero point time t0 is calculated using a linear interpolation formula, which equals t1 minus r1 multiplied by the difference between t2 and t1, and then divided by the difference between r2 and r1. This time is the precise time point at which the T1 effect transitions to the T2 asterisk effect.

[0050] Preferably, the baseline signal intensity is obtained from a steady-state scan before contrast agent injection, and the average signal intensity within 10 seconds before injection is typically selected as the baseline reference value.

[0051] In one possible implementation, the calculation of signal intensity deviation involves two metrics: absolute deviation and relative deviation. The absolute deviation is the difference between the signal intensity at zero point and the baseline signal intensity, reflecting the absolute level of signal enhancement. The relative deviation is the absolute deviation divided by the baseline signal intensity, yielding a dimensionless relative enhancement ratio. In carotid artery imaging, the relative deviation at the transition threshold is typically between 1.5 and 2.5, indicating a signal intensity enhancement of 150% to 250%.

[0052] For example, in renal artery imaging, if the baseline signal intensity is 100 units and the signal intensity reaches 280 units at the zero point, the absolute deviation is 180 units and the relative deviation is 1.8. This deviation provides a quantitative basis for subsequent determination of the optimal acquisition time.

[0053] S104. Based on the extracted signal intensity deviation amplitude, determine the upper and lower limits of contrast agent concentration corresponding to the T1 effect dominance interval and the T2 asterisk effect dominance interval, extract the signal intensity change process over time within the T1 effect dominance interval, and identify the time starting point of the arterial phase.

[0054] Based on the signal intensity deviation amplitude, when the deviation amplitude is lower than the first threshold 1, it corresponds to the lower limit of the concentration in the T1 effect-dominant region; when the deviation amplitude is higher than the second threshold 5, it corresponds to the upper limit of the concentration in the T2 asterisk effect-dominant region, thus obtaining the concentration range boundaries of the two regions. Using the concentration range boundaries, time-series data of signal intensity within the T1 effect-dominant region are extracted. The direction of rate change is determined by the first derivative; if the derivative changes from positive to negative, it is marked as the peak time, thus identifying the arterial phase start point.

[0055] Specifically, in one implementation, the setting of the first threshold and the second threshold is based on statistical analysis of a large amount of clinical data. When the signal intensity deviation is lower than the first threshold, it indicates that the contrast agent concentration has not yet caused a significant T2 asterisk effect, and the T1 effect is dominant. The concentration value corresponding to this threshold is the upper limit of the T1 effect dominance range.

[0056] Specifically, the second threshold is typically set to 1.5 to 2 times the first threshold. When the deviation exceeds this value, the T2 asterisk effect begins to significantly affect the signal strength, and the corresponding concentration value serves as the lower limit of the T2 asterisk effect's dominant range. There is a transition region between the two thresholds, within which the two effects compete with each other.

[0057] For example, the first derivative is calculated using the central difference method. For the signal intensity S(t) at time point t, the derivative value is equal to S(t+1) minus S(t-1) divided by twice the time interval. When the derivative values ​​at three consecutive time points change from positive to negative, the intermediate time point is marked as the peak time. Identifying the arterial phase start point is crucial for acquiring clean arterial images, as this moment marks the peak concentration of the contrast agent within the artery.

[0058] S105. For the determined arterial phase time start point, combine the signal enhancement amplitude caused by the T1 effect and the signal attenuation amplitude caused by the T2 asterisk effect, continuously monitor the peak arrival time and peak duration during the signal intensity change process, and evaluate the optimal duration of the acquisition time window.

[0059] Starting from the arterial phase time start point, the signal enhancement amplitude value generated by the T1 effect and the signal attenuation amplitude value generated by the T2 asterisk effect are recorded. A signal intensity time series is acquired through continuous sampling to determine the moment when the signal reaches its peak. Based on the peak moment, a continuous period of time during which the signal intensity remains above a preset percentage of the peak is calculated. If the signal intensity decreases by more than a preset percentage of the peak, it is marked as the end of the peak, thus obtaining the peak duration. The peak duration is compared with a pre-measured contrast agent transit time from artery to vein, and the smaller of the two values ​​is selected as the duration of the acquisition time window to evaluate and obtain the optimal acquisition duration parameter.

[0060] Specifically, in one implementation, once the arterial phase start point is determined, a high-frequency sampling mode is immediately initiated, with a sampling interval set to 0.2 to 0.5 seconds, continuously recording the signal intensity values ​​of the vascular region. The signal enhancement amplitude caused by the T1 effect is obtained by subtracting the baseline signal intensity from the current signal intensity, and the signal attenuation amplitude caused by the T2 asterisk effect is obtained by subtracting the current signal intensity from the peak signal intensity. Real-time recording of these two amplitude values ​​provides a data basis for subsequent judgment.

[0061] Specifically, a sliding window detection method is used to determine the peak moment. A window with five consecutive sampling points is set. When the signal intensity at the center of the window is greater than that at the two points before and after it, and the difference exceeds the noise level, it is marked as a local peak. Multiple local peaks may occur during the entire acquisition process, and the point with the highest signal intensity is selected as the global peak moment. This moment usually occurs within 15 to 25 seconds after contrast agent injection, but the specific time varies depending on the location of the blood vessel and the patient's cardiac function.

[0062] It should be noted that the choice of the preset percentage directly affects the calculation of the peak duration. In clinical practice, this percentage is usually set to 70% to 80% of the peak value. The peak period is considered to have ended when the signal intensity first drops below 75% of the peak value. The time interval from the peak moment to the peak end moment is the peak duration, which is approximately 3 to 5 seconds in head and neck vascular imaging and can reach 5 to 8 seconds in abdominal vascular imaging.

[0063] Preferably, the contrast agent transit time is pre-determined using a small-dose test bolus injection method. Before the formal scan, 1 to 2 ml of diluted contrast agent is injected, and the time interval from the onset of arterial enhancement to venous visualization is recorded by dynamically monitoring signal changes in the target vessel. This time reflects the complete process of the contrast agent flowing from the arterial system through the capillary bed into the venous system.

[0064] For example, in renal artery imaging, if the peak duration is 4 seconds and the measured arterial-to-venous transit time is 6 seconds, then 4 seconds is chosen as the duration of the acquisition time window. This choice ensures that there is a sufficient concentration of contrast agent in the blood vessel throughout the acquisition period, while avoiding venous contamination.

[0065] In one possible implementation, the optimal acquisition duration parameter also needs to consider the temporal resolution of the scan sequence. If a single 3D acquisition takes 3 seconds, and the evaluated optimal duration is 4 seconds, then one complete acquisition plus one partial acquisition can be performed, or the acquisition parameters can be adjusted to shorten the single acquisition time to 2 seconds, and two complete acquisitions can be performed.

[0066] S106. Based on the optimal duration of the acquisition time window as assessed, a trigger signal is generated when the T1 effect dominates and the T2 asterisk effect has not yet significantly affected the signal strength, thereby activating the image acquisition of the scanning system and obtaining a clean arterial phase image.

[0067] Based on the optimal duration of the acquisition time window, when the signal rise rate caused by the T1 effect is greater than the signal fall rate caused by the T2 asterisk effect, and the ratio of the two exceeds a preset threshold, a trigger signal is generated and sent to the magnetic resonance scanner. The trigger signal activates a three-dimensional gradient echo sequence, and K-space data acquisition is performed within the optimal duration after triggering to reconstruct a clean arterial phase image.

[0068] Specifically, in one implementation, the trigger signal is generated based on a rate ratio monitored in real time. The signal rise rate caused by the T1 effect is obtained by dividing the signal strength difference between two consecutive time points by the time interval, and the signal fall rate caused by the T2 asterisk effect is calculated using the same method. When the ratio of the rise rate to the fall rate exceeds 1.5, it indicates that the T1 effect still dominates, and a trigger signal is generated at this point.

[0069] Specifically, the trigger signal is in the form of a digital pulse and is transmitted to the gradient system and radio frequency system through the scanner's control interface. Upon arrival of the trigger signal, a pre-set three-dimensional gradient echo sequence is immediately initiated. The sequence parameters include a repetition time of 3 to 5 milliseconds, an echo time of 1 to 2 milliseconds, and a flip angle of 15 to 25 degrees.

[0070] It should be noted that the K-space data acquisition adopts a center-first acquisition sequence, first acquiring low-frequency data from the center of the K-space, and then acquiring high-frequency data from the periphery. This acquisition sequence ensures that contrast information is acquired when the contrast agent concentration is highest. The entire acquisition process is completed within the optimal duration, typically 3 to 6 seconds. After acquisition, an arterial phase image is obtained through three-dimensional Fourier transform reconstruction. In the image, arteries show high signal intensity, while veins are not yet visualized, achieving pure arterial imaging.

[0071] S107. Compare and analyze the obtained pure arterial phase image with the initial magnetic resonance signal intensity change process, update the historical acquisition record of T2 star effect recognition, extract the signal attenuation amplitude corresponding to different contrast agent concentrations in the next scan, and realize closed-loop optimization of acquisition timing control to continuously suppress venous contamination.

[0072] By comparing the average signal intensity of the vascular region in the pure arterial phase image with the signal value at the corresponding moment in the initial magnetic resonance signal intensity time curve, the signal deviation value at that time point is calculated to identify the precise moment when the T2 effect begins to have a significant impact, thus obtaining the actual effect transition point. Based on the actual effect transition point, the T2 effect identification parameters in the historical acquisition records are updated, including the effect occurrence time, corresponding concentration value, and signal decay rate. Data is fused by assigning higher weights to new data and lower weights to old data to obtain an updated effect identification parameter set. Using the updated effect identification parameter set, a mapping relationship table between contrast agent concentration and signal decay amplitude is established. The predicted decay amplitude values ​​corresponding to different concentration points are refined using linear interpolation to obtain the reference curve for the next scan. Based on the reference curve, the trigger threshold and acquisition duration parameters for the next scan are adjusted. Acquisition is terminated early when the predicted venous signal intensity exceeds three times the standard deviation of the baseline signal measured before contrast agent injection, achieving closed-loop optimization of acquisition timing control to continuously suppress venous contamination.

[0073] Specifically, the comparative analysis of the pure arterial phase image and the initial magnetic resonance signal intensity time curve employs a multi-step processing workflow. In practice, the target vessel region is extracted from the arterial phase image, and the average signal intensity value of all voxels within that region is calculated as the measured value. Simultaneously, the theoretical predicted value for the corresponding time series is extracted from the initial signal intensity time curve. The two are subtracted point by point to obtain the signal deviation sequence, where the deviation value reflects the degree of difference between the actual imaging process and the expected result. When the deviation value abruptly changes from positive to negative, it indicates that the T2 asterisk effect begins to exceed the expected level; this moment is the actual effect transition point.

[0074] It should be noted that the historical data acquisition records are organized using structured data storage. Each record contains multiple fields, including patient identifier, scan time, contrast agent type, injection rate, effect onset time, corresponding concentration value, and signal attenuation rate. Records are arranged chronologically to form a time-series database. When a new scan is completed, the extracted actual effect transition point data is added to the database as a new record. The database capacity is set to store the most recent 100 scan records; if the capacity is exceeded, the oldest record is deleted using a first-in, first-out (FIFO) principle.

[0075] For example, the weighting strategy is determined based on the timeliness and relevance of the data. New data within one week is assigned a high weight of 0.7 to 0.9; data from one week to one month is assigned a medium weight of 0.3 to 0.7; and older data older than one month is assigned a low weight of 0.1 to 0.3. Furthermore, if the new and old data come from the same patient, the weight is increased by 0.1; if the same type of contrast agent is used, the weight is increased by 0.05. This multi-factor weighting achieves both the rational use of historical experience and sufficient consideration of new information. When calculating the weighted average, each parameter value is multiplied by its corresponding weight, summed, and then divided by the total weights to obtain the merged updated value.

[0076] Preferably, the effect identification parameter set includes three core parameters: effect onset time, corresponding concentration value, and signal decay rate. The effect onset time records the time interval from contrast agent injection to the onset of the T2 asterisk effect, with an accuracy of 0.1 seconds. The corresponding concentration value represents the critical contrast agent concentration that triggers the effect transition, expressed in millimoles per liter. The signal decay rate quantifies how quickly the signal decreases due to the T2 asterisk effect, expressed as a percentage per second.

[0077] In one possible implementation, the mapping table between contrast agent concentration and signal attenuation amplitude is established based on statistical analysis of a large amount of historical data. The concentration range from 0 to 20 mmol / L is divided into 200 equally spaced points, with each concentration point corresponding to an attenuation amplitude value. Initially, only some concentration points have measured data, and the attenuation amplitudes of the remaining points are obtained through linear interpolation. The linear interpolation method is as follows: for the point to be interpolated, find the nearest known data points on both sides, and calculate the corresponding attenuation amplitude based on the linear proportionality of the concentrations. As the number of scans increases, the number of measured data points gradually increases, and the accuracy of the mapping relationship continuously improves.

[0078] Specifically, the application of reference curves involves both prediction and decision-making. Before the next scan, a curve under similar conditions is selected from a mapping table as a reference based on the patient's physiological parameters such as weight and heart rate. During the scan, the real-time monitored signal intensity is compared with the reference curve to calculate the degree of deviation. When the difference between the measured value and the predicted value exceeds 15%, a parameter adjustment mechanism is triggered.

[0079] For example, in carotid artery imaging, if the reference curve predicts that the T2 asterisk effect will appear 18 seconds after injection, but actual monitoring shows that the effect signs appear at 16 seconds, the system will immediately reduce the trigger threshold by 20% and shorten the acquisition time by 2 seconds to avoid venous contamination.

[0080] Understandably, the prediction of venous signal intensity is based on the time difference between arterial and venous circulation. Typically, it takes 5 to 8 seconds for the contrast agent to travel from the artery to the vein. The rise time and intensity of the venous signal are estimated based on the trend of arterial signal changes. The baseline signal standard deviation is calculated using continuous sampling over the first 30 seconds before contrast agent injection, reflecting the inherent noise level of the signal. When the predicted venous signal intensity exceeds three times the standard deviation, the vein is considered to be beginning to visualize, and acquisition should be stopped immediately. Furthermore, closed-loop optimization relies on dynamic parameter adjustment. After each scan, the system automatically analyzes image quality, including arterial clarity, venous contamination level, and signal-to-noise ratio. Based on the analysis results, the trigger threshold and acquisition duration parameters are fine-tuned. If venous contamination is detected, the acquisition duration of the next scan is shortened by 0.5 seconds; if the artery is not sufficiently visualized, the trigger threshold is reduced by 10%. Through this iterative optimization, the system gradually converges to a parameter combination suitable for specific scanning conditions.

[0081] In one embodiment, after 5 to 10 scans of closed-loop optimization, the vein contamination rate, defined as the ratio of vein signal intensity to arterial signal intensity, VPR=Vs / As, where Vs is the vein signal intensity and As is the arterial signal intensity, decreased from an initial 0.3 to below 0.05. The arterial display quality score, based on a 5-point scale for image sharpness, increased from 3.2 to 4.5, achieving continuous improvement in imaging quality.

[0082] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A method for real-time processing and analysis of magnetic resonance scanning data, characterized in that, include: Signal data of the vascular region were acquired by magnetic resonance imaging (MRI) to obtain the changes in contrast agent concentration and signal intensity, as well as the spatial distribution of magnetic field homogeneity. Based on the signal data, identify the signal attenuation amplitude and rate of decline caused by the T2 effect, and assess the correlation strength between contrast agent concentration and magnetic field homogeneity. The arrival time of the first-pass peak concentration and the signal intensity performance were extracted from the changes in contrast agent concentration and signal intensity to analyze the conversion process between the T1 effect and the T2 effect. Based on the signal attenuation rate, a corrected curve is generated to determine the signal strength deviation amplitude at the transition critical point between the T1 and T2 effects. Based on the deviation amplitude, the T1 effect-dominant interval and the T2 effect-dominant interval are divided to identify the arterial phase start point and evaluate the optimal duration of the acquisition time window. A trigger signal is generated based on the optimal duration to activate image acquisition and obtain arterial phase images; By comparing the changes in the arterial phase images with those in the initial signal, the effect identification parameters are updated, thereby optimizing the timing of data acquisition.

2. The method for real-time processing and analysis of magnetic resonance scanning data as described in claim 1, characterized in that, The acquisition of signal data from the vascular region via magnetic resonance imaging (MRI) sequences, obtaining the changes in contrast agent concentration and signal intensity, and the spatial distribution of magnetic field homogeneity, includes: Raw data of the vascular region were acquired using a multi-echo gradient echo sequence. The local magnetic field offset at each voxel location was calculated based on the phase accumulation difference at different echo times, generating a spatial distribution map of magnetic field inhomogeneity. For the aforementioned spatial distribution map, the T2 star relaxation rate is extracted from the ratio of signal intensity of short echo time to long echo time. If the ratio is lower than a preset threshold, it is determined that there is a high concentration of contrast agent accumulation in the region, and an initial estimate of the contrast agent concentration distribution is obtained. Based on the initial estimate, the signal strength at consecutive time points is differentially calculated using the sliding window method. The signal before injection is used as the baseline, and the instantaneous rate of decrease relative to the baseline signal is calculated to determine the time point at which the T2 effect begins to dominate.

3. The method for real-time processing and analysis of magnetic resonance scanning data as described in claim 1, characterized in that, After acquiring the changes in contrast agent concentration and signal intensity, as well as the spatial distribution of magnetic field homogeneity, the process includes: A spatiotemporal feature matrix is ​​constructed from the signal attenuation amplitude and descent rate. The correlation coefficient is used to calculate the linear correlation between the contrast agent concentration value and the magnetic field offset at each voxel location, and a correlation intensity distribution map is generated. Based on the distribution map, regions with correlation coefficients higher than a preset threshold are identified, and the attenuation start point, peak point, and stable point are determined from the signal time series of these regions to obtain the time window of the T2 effect. For signal data within the time window, if the contrast agent concentration exceeds a preset critical value, the signal amplitude and phase changes are separated, feature vectors are extracted through spectrum analysis, and principal component analysis is used to obtain signal change characteristics.

4. The method for real-time processing and analysis of magnetic resonance scanning data as described in claim 1, characterized in that, The analysis of the conversion process between the T1 effect and the T2 effect includes: By continuously monitoring the signal strength time curve, the instantaneous slope of adjacent time points is calculated, and the moment corresponding to the maximum slope is identified as the arrival time of the first peak concentration. The signal strength value and the estimated concentration value before and after this moment are recorded. When the peak concentration is lower than the preset threshold, the signal enhancement amplitude caused by the T1 effect is extracted, and the signal contrast between the vascular region and the surrounding tissue is calculated. If the peak concentration exceeds the critical value, the T1 effect enhancement rate and the T2 effect decay rate are calculated, and the change in the ratio of the T1 effect enhancement rate to the T2 effect decay rate is monitored to determine the time point of the effect dominance transition and the corresponding concentration value.

5. The method for real-time processing and analysis of magnetic resonance scanning data as described in claim 1, characterized in that, The signal strength deviation amplitude used to determine the transition critical point between the T1 and T2 effects includes: The baseline drift of the signal attenuation rate curve was removed by a fitting method to generate a corrected rate curve. According to the corrected rate curve, when the rate value is positive, it is identified as the T1 effect dominant segment, and when the rate value is negative, it is identified as the T2 effect dominant segment. The zero point where the rate changes from positive to negative can be determined by finding adjacent data points where the sign of the rate value changes. Based on the zero-point position, the signal strength value at the corresponding time is extracted, and the difference between it and the baseline signal strength is calculated to obtain the signal strength deviation amplitude at the effect transition critical point.

6. The method for real-time processing and analysis of magnetic resonance scanning data as described in claim 1, characterized in that, The step of dividing the T1-effect dominant interval and the T2-effect dominant interval according to the deviation amplitude, and identifying the arterial phase time start point, includes: Based on the signal intensity deviation amplitude, determine the lower limit of the concentration in the T1 effect-dominant region and the upper limit of the concentration in the T2 effect-dominant region, and obtain the concentration range boundaries of the two regions. Using the concentration range boundary, time-series data of signal intensity within the T1 effect-dominant region are extracted; The direction of rate change is determined by calculating the first derivative. If the derivative changes from positive to negative, it is marked as the peak time, thus identifying the starting point of the arterial phase.

7. The method for real-time processing and analysis of magnetic resonance scanning data as described in claim 1, characterized in that, The optimal duration of the evaluation acquisition time window includes: Starting from the arterial phase time start point, record the signal enhancement amplitude value generated by the T1 effect and the signal attenuation amplitude value generated by the T2 effect. Obtain the signal intensity time series through continuous sampling to determine the time when the signal reaches its peak. Based on the moment when the signal reaches its peak, calculate the continuous time period during which the signal strength remains above a preset percentage. If the signal strength drops by more than the preset percentage, mark it as the end of the peak and obtain the duration of the peak. The duration of the peak value is compared with the pre-determined contrast agent duration over time to determine the duration of the acquisition time window.

8. The method for real-time processing and analysis of magnetic resonance scanning data as described in claim 1, characterized in that, The optimization of acquisition timing control by comparing the arterial phase image with the initial signal changes and updating the effect identification parameters includes: By comparing the average signal intensity of the vascular region in the arterial phase image with the value of the initial signal intensity time curve at the corresponding time, the signal deviation value is calculated, the time when the T2 effect has a significant impact is identified, and the actual effect transition point is obtained. Update the effect identification parameters in the historical acquisition records based on the actual effect transition point; Using the effect identification parameters from the updated historical acquisition records, a mapping table between contrast agent concentration and signal attenuation amplitude was established. The predicted values ​​for different concentration points were refined by interpolation, and a reference curve was generated. Based on the reference curve, the trigger timing and acquisition duration parameters are adjusted to optimize the acquisition timing control.

Images