Imaging method for extrahepatic cholangiocarcinoma

Abstract

The invention relates to the technical field of magnetic resonance imaging, and discloses a thin-layer oblique coronal focusing view diffusion weighted imaging method of a magnetic resonance system, which comprises the following steps of: extracting a three-dimensional space walking path of a target bile duct based on magnetic resonance pancreaticobiliary duct water imaging and a cross section T2 weighted image, performing optimal plane fitting, and determining a personalized oblique coronal scanning plane; in the plane, diffusion weighted imaging is carried out by adopting a small visual field set according to bile duct projection, a layer thickness not higher than 3.0 mm and a low b value not higher than 100 seconds / square millimeters and a high b value not lower than 800 seconds / square millimeters, and echo time is set to a minimum value allowed by a system. Through cooperation of personalized plane positioning, thin-layer high-b-value imaging and intelligent quality control, the problems that conventional magnetic resonance imaging is incomplete in display of an extrahepatic cholangiocarcinoma lesion anatomical structure, low in signal-to-noise ratio, multiple in motion artifacts and difficult to accurately evaluate the longitudinal infiltration range of a tumor are solved.

Description

Technical Field

[0001] This invention relates to the field of magnetic resonance imaging technology, and to an imaging method for extrahepatic bile duct carcinoma, specifically to a thin-slice oblique coronal focusing field diffusion-weighted imaging method for a magnetic resonance system. Background Technology

[0002] Extrahepatic cholangiocarcinoma is a highly malignant tumor originating in the perihepatic and distal bile ducts, accounting for 90% of all cholangiocarcinomas. Its epidemiological characteristics show significant regional differences, with a higher incidence (approximately 6-7 per 100,000) in East Asia, including China. Extrahepatic cholangiocarcinoma has an insidious onset, often presenting with no obvious symptoms initially. Most patients are diagnosed at an advanced stage, making surgical resection ineffective. Therefore, comprehensive and reliable imaging diagnostic methods are crucial in the diagnosis and treatment of extrahepatic cholangiocarcinoma.

[0003] Cholangiocarcinoma exhibits multidimensional invasive growth characteristics, making accurate assessment of tumor lesions and resectableness challenging. Current routine MRI diagnostic assessments still face technical bottlenecks: Imaging evaluation of extrahepatic cholangiocarcinoma primarily relies on conventional axial diffusion-weighted imaging (DWI), magnetic resonance cholangiopancreatography (MRCP), and contrast-enhanced MRI. However, 1) conventional T1 / T2-weighted sequences are insufficiently accurate in displaying tumor boundaries and the extent of invasion; 2) while MRCP can visually show the degree of bile duct invasion and stenosis, it is insufficient in assessing the lesion itself and its surrounding anatomical relationships, limiting its value in determining resectableness; 3) conventional transverse DWI cannot fully display the entire anatomical structure of the bile duct, leading to an underestimation of the extent of invasion. In clinical scanning, oblique coronal thin-slice scanning and imaging along the course of the bile duct can more clearly show the relationship between the tumor and the anatomical structures of the bile duct. However, traditional EPI diffusion imaging in oblique coronal thin-slice imaging often suffers from insufficient gradient performance, resulting in distorted coronal diffusion images and low signal-to-noise ratios, failing to meet clinical diagnostic requirements. Summary of the Invention

[0004] To overcome the aforementioned deficiencies of the prior art, this invention provides a thin-slice oblique coronal focusing field diffusion-weighted imaging method for magnetic resonance imaging systems. This method integrates personalized oblique coronal scanning plane planning, thin-slice high b-value imaging supported by an ultra-high gradient system, a collaborative quality control mechanism for respiratory navigation triggering and real-time quality assessment, and registration and fusion display of multimodal images to solve the problems mentioned in the background art.

[0005] This invention solves the above-mentioned technical problems through the following technical solution: a thin-slice oblique coronal focusing field diffusion-weighted imaging method for magnetic resonance imaging systems, comprising the following steps:

[0006] Step 1: Based on the pre-acquired magnetic resonance pancreaticobiliary hydrops image and transverse T2-weighted image, determine the three-dimensional spatial path of the target bile duct, and perform optimal plane fitting on the three-dimensional spatial path to determine a personalized oblique coronal scanning plane for subsequent scanning.

[0007] The three-dimensional spatial path can be determined by manual delineation or automatic tracking. The optimal plane fitting adopts a plane fitting algorithm based on the least squares method. Its goal is to find a plane that minimizes the sum of the squares of the perpendicular distances from all points on the three-dimensional spatial path to the plane. The normal vector of the plane corresponds to the eigenvector corresponding to the smallest eigenvalue of the covariance matrix of the path point set. Furthermore, the geometric parameters of the fitted plane are converted into rotation angles and center offsets that can be recognized by the magnetic resonance scanning system and locked as a fixed scanning plane for subsequent imaging.

[0008] Step 2: In the personalized oblique coronal scanning plane, imaging is performed using a small field-of-view excitation diffusion-weighted imaging sequence, i.e., diffusion-weighted imaging;

[0009] The small field of view is set based on the projection of the target bile duct onto the personalized oblique coronal scanning plane. Specifically, the center of the diffusion imaging field of view in the head-to-toe direction is the center of the bile duct centerline in the head-to-toe direction. The upper edge of the diffusion imaging field of view is the top of the diaphragm, and the lower edge is the lower end of the bile duct centerline extending downwards by 10-20 mm. The diffusion imaging field of view perpendicular to the head-to-toe direction is centered on the bile duct centerline and extends more than 100 mm to both sides to cover the bile duct and its surrounding soft tissue. The rectangular area formed in this way is defined as the excitation and acquisition range of the small field of view excitation diffusion-weighted imaging sequence. Imaging employs thin-slice acquisition with a slice thickness of less than 3.0 mm and an echo time set to the minimum allowed by the system. Simultaneously, diffusion sensitivity factor b values ​​of both low b-value (not higher than 100 sec / mm²) and high b-value (not lower than 800 sec / mm²) are acquired. This step is performed on an ultra-high gradient magnetic resonance system equipped with a gradient field strength of 200 mT / m and a gradient switching rate of 200 T / m / s. The performance of this system allows the echo time to be shortened to less than 70 milliseconds under the aforementioned thin-slice high b-value imaging conditions, thereby ensuring the image signal-to-noise ratio.

[0010] Step 3: During the imaging process, each data acquisition is triggered based on the respiratory navigation signal during the end-expiratory stabilization period; the quality of the acquired data is evaluated in real time, and if the data does not meet the preset quality standards, it is not adopted, and supplementary acquisition is triggered in subsequent respiratory cycles until complete data that meets all preset quality requirements is obtained;

[0011] The end-expiratory stabilization period is defined as the plateau period after the lowest point of the respiratory motion curve; real-time quality assessment includes parallel signal strength stability assessment and motion artifact assessment: the former compares the deviation of the signal strength of the current and historical k-space center, and the latter judges artifacts by calculating the image entropy or edge gradient index of the preview image; data is only accepted if it passes both assessments simultaneously;

[0012] Step 4: Reconstruct the data obtained in Step 3, generate personalized oblique coronal scanning plane low b-value diffusion-weighted image and high b-value diffusion-weighted image, spatially register the low b-value diffusion-weighted image and high b-value diffusion-weighted image with the magnetic resonance pancreatobiliary duct hydrops image source image and the transverse T1-weighted enhanced image of the same examination, and output the registered multi-sequence fusion display image.

[0013] It should be noted that the reconstruction generates a corresponding apparent diffusion coefficient map; spatial registration uses rigid or non-rigid registration algorithms; the fusion display supports multi-view side-by-side browsing, or overlays high b-value diffusion-weighted images onto anatomical images in a pseudo-color semi-transparent manner, and can automatically measure and output the longitudinal span distance of the lesion along the bile duct direction.

[0014] Furthermore, in step one, determining the three-dimensional spatial path of the target bile duct includes:

[0015] Based on the signal intensity difference between the bile duct and surrounding tissues in the magnetic resonance pancreatobiliary duct hydrops source image and the transverse T2-weighted image, an image processing algorithm for extracting the centerline of blood vessels or tubular structures is used to automatically identify and extract the centerline of the target bile duct lumen to form a three-dimensional spatial path.

[0016] Optionally, the image processing algorithm specifically includes a segmentation step that combines region growing and level set modeling, as well as a subsequent skeletonization extraction step, and finally outputs a continuous centerline path composed of a series of discrete three-dimensional coordinate points.

[0017] Furthermore, in step one, performing optimal planar fitting on the three-dimensional spatial path includes:

[0018] The point set constituting the three-dimensional spatial travel path is fitted with a least-squares plane. The personalized oblique coronal scanning plane is determined by solving the plane equation that minimizes the sum of the squared distances from the point set to the fitted plane.

[0019] The normal vector of the plane equation corresponds to the eigenvector corresponding to the smallest eigenvalue of the point set covariance matrix.

[0020] It should be noted that the core of the fitting process is to construct the covariance matrix of the point set and calculate its eigenvalues ​​and eigenvectors. By finding the plane that minimizes the projection distortion, we can ensure that the plane can maximize the coverage of the projection area of ​​the target bile duct and its adjacent suspicious lesions.

[0021] Furthermore, in step one, determining the personalized oblique coronal scanning plane also includes:

[0022] The geometric parameters of the plane equation obtained by the optimal plane fitting are converted into rotation angles and center offsets that can be recognized by the magnetic resonance scanning system by calling the application programming interface of the magnetic resonance scanning system or using its built-in positioning conversion tool. Based on these parameters, the personalized oblique coronal scanning plane of all subsequent imaging sequences is locked.

[0023] The transformation includes: calculating the azimuth and tilt angles between the plane normal vector and the device's standard coordinate axes as rotation angles; simultaneously calculating the signed distance to the origin of the coordinate system based on the plane equations; and determining the scan center offset by combining this with a preset field of view center. This process achieves a direct mapping from mathematical geometric parameters to scan control parameters.

[0024] Furthermore, in step two, the imaging field of diffusion-weighted imaging is set based on the projection of the target bile duct onto the personalized oblique coronal scanning plane, specifically as follows:

[0025] Using the projection of the three-dimensional spatial path onto the personalized oblique coronal scanning plane as the center line, the center of the diffusion imaging field of view in the head-to-foot direction is the center of the bile duct center line in the head-to-foot direction. The upper edge of the diffusion imaging field of view is the top of the diaphragm, and the lower edge is the lower end of the bile duct center line, extending downwards by 10-20 mm. The diffusion imaging field of view perpendicular to the head-to-foot direction is centered on the bile duct center line, extending more than 100 mm to both sides to cover the bile duct and its surrounding soft tissue. The rectangular area formed by this is defined as the excitation and acquisition range of the small field of view excitation diffusion-weighted imaging sequence.

[0026] This technique precisely focuses the imaging range on the target anatomical area, effectively reducing signal interference from irrelevant tissues and distortion artifacts caused by a large field of view, thereby improving image resolution and contrast.

[0027] Furthermore, in step two, the acquisition layer thickness is set to no more than 3.0 mm, and low diffusion sensitivity factor diffusion images with a sensitivity of no more than 100 seconds / mm² and high diffusion sensitivity factor diffusion images with a sensitivity of no less than 800 seconds / mm² are acquired simultaneously, which are low b-value diffusion-weighted images and high b-value diffusion-weighted images, respectively.

[0028] Furthermore, in step one, the geometric parameters of the plane equation are converted into rotation angles and center offsets that the scanning system can recognize, specifically including:

[0029] Calculate the azimuth and tilt angles between the plane normal vector and the equipment's standard coordinate axes, and use these as rotation angles;

[0030] The signed distance from the origin to the plane is calculated based on the plane equation, and the scan center offset is determined by combining the preset field of view center.

[0031] The parameters are automatically input into the diffusion imaging sequence, ensuring that the acquisition orientation is strictly fixed to the personalized oblique coronal scanning plane.

[0032] Furthermore, in step three, the real-time evaluation of the quality of the collected data includes:

[0033] Signal strength stability assessment: The signal strength of the k-space center region acquired in this acquisition is compared with the corresponding signal strength of several previous acquisitions. If the deviation exceeds the tolerance range determined based on the pre-scan signal-to-noise ratio analysis, the data is considered unstable.

[0034] Motion artifact evaluation: A preview image is generated from the original k-space data acquired in this study. By calculating the edge gradient along the phase encoding direction in the preview image or using the image entropy index, it is determined whether there is obvious blurring or artifacts caused by motion.

[0035] The two assessments are performed in parallel, forming a "double-insurance" judgment mechanism for data quality. Only when a single acquisition of data passes both the stability and artifact assessments is it marked as valid and used for final image reconstruction, ensuring the consistency and clarity of the output image.

[0036] Furthermore, in step four, based on the low b-value diffusion-weighted image and the high b-value diffusion-weighted image, a corresponding apparent diffusion coefficient map is generated. The apparent diffusion coefficient map, the low b-value diffusion-weighted image, and the high b-value diffusion-weighted image are then incorporated into a multi-sequence fusion display image for comparative display. The apparent diffusion coefficient map provides quantitative information on tissue diffusion capacity, complementing the qualitative high-contrast information provided by the low b-value diffusion-weighted image and the high b-value image, thus jointly enhancing the ability to detect and qualitatively diagnose lesions.

[0037] Furthermore, steps one, two, three, and four are performed consecutively within the same examination sequence, and the source images from MRI pancreaticobiliary hydrops, transverse T2-weighted images, and transverse T1-weighted enhanced images are all acquired within the same MRI examination session as diffusion-weighted imaging. This integrated workflow design ensures that all image data share the same spatial reference and physiological state, avoiding registration errors and interpretation difficulties caused by patient displacement or changes in state between multiple examinations, thus improving the efficiency and reliability of the diagnostic process.

[0038] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0039] 1. This invention achieves high spatial resolution, high signal-to-noise ratio, and controllable deformation diffusion-weighted image acquisition along the bile duct's course by integrating personalized oblique coronal scanning plane planning, thin-slice low and high b-value imaging supported by an ultra-high gradient system, respiratory navigation gating, and real-time quality assessment. Specifically, it automatically extracts the three-dimensional spatial path of the bile duct based on magnetic resonance pancreaticobiliary hydrops imaging and T2-weighted images, and determines the personalized oblique coronal scanning plane through optimal plane fitting. Within the personalized oblique coronal scanning plane, it employs small field-of-view excitation, slice thickness not exceeding 3.0 mm, and low diffusion sensitivity diffusion images not exceeding 100 sec / mm², as well as high diffusion sensitivity diffusion images not less than 800 sec / mm², and shortens the echo time to the minimum allowed by the system. Relying on a gradient field strength of 200 mT / m or higher, the echo time can be shortened to below 70 milliseconds. Simultaneously, acquisition is triggered during the end-expiratory stabilization period, and data is screened in real time through signal intensity stability and motion artifact evaluation mechanisms to ensure image quality. This collaborative technology link effectively solves the technical bottlenecks of conventional transverse DWI, which cannot fully display the entire bile duct, and the low signal-to-noise ratio and significant respiratory motion artifacts in oblique coronal thin-slice imaging.

[0040] 2. This invention further achieves precise overlay and linked analysis of functional imaging and anatomical images through spatial registration and fusion display of multi-sequence images. Based on the generation of low and high b-value diffusion-weighted images and apparent diffusion coefficient maps, non-rigid or rigid registration algorithms are used to align them with MRCP images and enhanced T1WI images, and output them in a pseudo-color fusion, multi-view synchronous manner. Simultaneously, the longitudinal span distance of the lesion along the bile duct projection direction is automatically detected and output. This integrated output mode allows for a clear and intuitive presentation of the infiltration range of bile duct tumors and their relationship with surrounding structures, effectively solving the clinical problems of difficult comparison of traditional multi-sequence images, inaccurate spatial localization of lesions, and reliance on subjective judgment for preoperative assessment. It provides reliable imaging evidence for the accurate staging and resectability assessment of extrahepatic bile duct cancer.

[0041] 3. This invention constructs a fully automated, integrated scanning and post-processing workflow, achieving fully automated execution of the entire chain from bile duct path extraction, plane fitting, parameter setting to image acquisition and fusion. This workflow integrates the calculation and locking of personalized oblique coronal scanning planes, the triggering and quality control of small-field-of-view thin-slice low-b-value imaging, and the registration and measurement of multi-sequence images into a single examination protocol, eliminating the need for manual operator intervention in plane adjustments or repositioning between sequences. This not only significantly shortens examination time but also ensures a high degree of consistency and repeatability of imaging schemes and output results across different cases and operators, effectively solving the problems of cumbersome traditional multi-sequence MRI examination procedures, reliance on experience, and high variability of results.

[0042] 4. This invention further achieves comprehensive diagnostic and quantitative assessment capabilities beyond a single imaging modality by deeply integrating low- and high-b-value DWI functional information with detailed anatomical information. It spatially aligns and fuses low- and high-b-value images clearly showing the longitudinal invasion range of the tumor, ADC maps reflecting tissue diffusion characteristics, MRCP images demonstrating biliary anatomy, and enhanced T1WI showing the relationship between blood vessels and soft tissues. It also automatically extracts the spatial span of the lesion, providing physicians with a "one-stop" assessment view that combines macroscopic anatomical relationships with microscopic functional characteristics. This simultaneous presentation and cross-validation of multi-dimensional information effectively solves the clinical problems of low diagnostic efficiency, subjective errors, and the easy omission of small lesions caused by fragmented information and the need for physicians to manually fuse different image sequences in their minds. Attached Figure Description

[0043] Figure 1 This is a flowchart illustrating the principle of the present invention;

[0044] Figure 2 This is a comparative diagram showing the echo time required to achieve the same high b-value diffusion-weighted imaging under two magnetic resonance systems with different gradient performance.

[0045] Figure 3 This is a schematic diagram of the location of a personalized oblique coronal scanning plane based on the bile duct course planning. Figure 1 ;

[0046] Figure 4 This is a schematic diagram of the location of a personalized oblique coronal scanning plane based on the bile duct course planning. Figure 2 ;

[0047] Figure 5 The image shown is a standard coronal enhanced T1WI image of Case 1 in the example.

[0048] Figure 6 The image shown is a magnetic resonance pancreatobiliary duct hydrops source image of Case 1 in the embodiment.

[0049] Figure 7 The image shown is a thin-layer, b-value-weighted oblique coronal image obtained using the method of the present invention in Case 1 of this embodiment.

[0050] Figure 8 The image shown is from Case 1 in this embodiment, obtained using the method of the present invention.

[0051] Figure 9 The image shown is a standard enhanced T1WI image of Case 2 in this embodiment;

[0052] Figure 10 The image is a magnetic resonance pancreatobiliary duct hydrops source image of Case 2 in the example;

[0053] Figure 11The image shown is a diffraction-weighted thin-layer image of the oblique coronal region obtained using the method of the present invention in Case 2 of this embodiment. Detailed Implementation

[0054] The above-mentioned and other technical features and advantages of the present invention will be described in more detail below with reference to the accompanying drawings.

[0055] This embodiment provides a technical solution: a thin-slice oblique coronal focusing field diffusion-weighted imaging method for magnetic resonance imaging systems, comprising the following steps:

[0056] Step 1: Based on the pre-acquired magnetic resonance pancreaticobiliary hydrops image and transverse T2-weighted image, determine the three-dimensional spatial path of the target bile duct, perform optimal plane fitting on the three-dimensional spatial path, and determine a personalized oblique coronal scanning plane for subsequent scanning.

[0057] This step aims to determine an optimal scanning plane (i.e., a personalized oblique coronal scanning plane) based on the actual three-dimensional anatomical course of the bile ducts in each patient, laying a spatial reference for subsequent high-resolution diffusion imaging. The technical essence lies in: first, acquiring and processing image data reflecting the anatomy of the bile ducts to construct a geometric model representing the spatial course of the bile ducts (i.e., a three-dimensional spatial path curve); then, calculating the optimal imaging plane based on this geometric model using specific mathematical optimization methods. This two-stage transformation—"from image to model, and then from model to plane"—is the geometric basis for achieving subsequent precise "along-duct" imaging, effectively overcoming the fundamental limitation of conventional fixed-azimuth scanning in adapting to the complex spatial course of individual bile ducts. The specific implementation process is as follows:

[0058] S101: Acquire and load the pre-acquired magnetic resonance pancreaticobiliary duct hydrops source image and transverse T2-weighted image of the same patient in the same magnetic resonance examination as reference images, which can clearly show the anatomical structure of the extrahepatic bile duct.

[0059] S102: On an image workstation, based on reference images, determine the three-dimensional spatial center path of the target bile duct (such as the common hepatic duct, common bile duct, or hilar bile duct). Methods for determining the three-dimensional spatial center path include, but are not limited to, the following two:

[0060] Manual outlining method: On an image workstation with multi-plane reconstruction capabilities, the operator uses the 3D marking tool of its medical image processing software to manually set a series of marker points along the center of the target bile duct lumen; then, the software's built-in algorithm connects and smooths these marker points to generate a continuous 3D spatial path curve.

[0061] Automatic tracking method: Based on the signal intensity difference between the bile duct and surrounding tissues in the reference image, image processing algorithms known in the art for extracting the centerline of blood vessels or tubular structures (e.g., segmentation algorithms based on region growing and level set models combined with skeletonization algorithms) are used to automatically identify and extract the centerline of the target bile duct lumen, such as... Figure 3 and Figure 4 As shown, a three-dimensional spatial path curve is formed. The automatic tracking method improves the standardization and repeatability of the solution, reduces operator differences, and is key to achieving online automated processing.

[0062] S103: Planning a Personalized Oblique Coronal Scan Plane: Invoking or executing an optimal plane fitting (e.g., a plane fitting based on the least squares principle), using the three-dimensional spatial path curve as input. The goal of the optimal plane fitting is to calculate a plane that minimizes the sum of the projected distances from all points on the three-dimensional spatial path curve to that plane (i.e., minimizes projection distortion), and that the plane spatially encompasses the target bile duct and the projection area of ​​its adjacent suspected lesions to the maximum extent. The plane calculated by the optimal plane fitting that satisfies the above conditions is defined as the personalized oblique coronal scan plane required for this scan. The spatial orientation of the personalized oblique coronal scan plane (defined by its normal vector in the MRI coordinate system) is uniquely determined by the specific orientation of the bile duct.

[0063] It is important to note that the personalized oblique coronal scanning plane is not a standard anatomical plane, but rather the optimal imaging perspective "tailor-made" to display a specific bile duct, ensuring that the target structure is displayed in the most complete manner within a single plane.

[0064] The aforementioned obtained three-dimensional spatial path curve of the target bile duct consists of a series of discrete three-dimensional coordinate points. The structure consists of i = 1, 2, ..., n.

[0065] In this step, the goal of optimal plane fitting is to find an optimal two-dimensional plane that minimizes the sum of the squares of the perpendicular distances from all points on the three-dimensional spatial path curve to the optimal plane fitting. This ensures that the three-dimensional spatial path is the "best approximation plane" of the bile duct path in space, thus accommodating its projection most completely and with minimal distortion.

[0066] The input to optimal plane fitting is the set of all points on the path curve in three-dimensional space. The core calculations and outputs are as follows:

[0067] S1031: Assume the equation of the plane to be found is , where (A, B, C) is the unit normal vector of the plane, which determines the spatial orientation of the plane, and D is a constant term;

[0068] S1032: The algorithm minimizes all data points. Sum of squared perpendicular distances to the plane

[0069] To solve for the plane parameters A, B, C, and D;

[0070] S1033: By constructing the covariance matrix or solving for the eigenvalues, the optimal plane parameters that minimize S can be determined; specifically, the normal vector n(A, B, C) of the optimal plane corresponds to the data point set. The eigenvector corresponding to the smallest eigenvalue of the covariance matrix ensures the objectivity and optimality of the transformation from a three-dimensional curve to a two-dimensional plane, eliminating the subjective bias of manual selection.

[0071] S1034: The final output is the geometric parameters of the optimal plane, expressed as a plane equation in the coordinate system of the magnetic resonance scanning device. The plane equation is converted into a rotation angle and offset that can be recognized by the magnetic resonance scanning device. This plane is then determined as the personalized oblique coronal scanning plane for all subsequent sequence scans.

[0072] The conversion is achieved in the following way:

[0073] Determining the rotation angle of the personalized oblique coronal scanning plane: The unit normal vector n(A, B, C) obtained from the plane equation coefficients defines the spatial orientation of the plane; by calculating the angle between the normal vector n(A, B, C) and the normal vector of the reference plane (such as the transverse or coronal plane) in the standard patient coordinate system, the required rotation angle (including azimuth and tilt angles) for scanning is solved.

[0074] Determining the position of the personalized oblique coronal scanning plane: The constant term D in the plane equation and the unit normal vector n(A, B, C) together determine the position of the personalized oblique coronal scanning plane. This is achieved using the formula: Offset = The signed distance from the plane to the origin of the coordinate system is calculated, and combined with the preset center of the scanning field of view, the center offset of the scan can be determined.

[0075] It should be noted that in practice, the "standard patient coordinate system" is usually aligned with the inherent coordinate system of the MRI scanner, and its definition (such as origin and axis) can be clearly found in the technical documentation of the MRI scanner. The process of converting the plane equation into scanning parameters can be achieved by calling the application programming interface (API) provided by the scanning system or using its preset positioning transformation tool. Specifically, after inputting the calculated unit normal vector n and offset, the system's internal software will automatically calculate and fill the parameter input boxes for "Orientation" and "Position" in the scanning protocol based on its inherent geometric model. This step serves as a bridge connecting virtual planning and physical scanning, achieving seamless integration from software algorithm results to hardware control commands.

[0076] Those skilled in the art will know that although the operating interfaces and parameter naming of devices from different manufacturers differ, they all provide corresponding input items for defining the spatial position and orientation of the scanning plane, and the input of the above conversion results is direct.

[0077] Ultimately, the calculated rotation angle and center offset are used as key positioning parameters and directly input into the MRI scan control interface. Subsequently, when executing the scanning sequences in steps two and three, the system will automatically and strictly lock all data acquisition planes within the individually defined oblique coronal scanning plane based on these parameters, thus ensuring the consistency between subsequent imaging and the initial planning in anatomical space. This "one-time planning, full-process locking" mode fundamentally guarantees a strict spatial correspondence between functional imaging and the reference anatomical image, laying the foundation for subsequent precise registration and fusion.

[0078] The technical essence of "planning and determining a personalized oblique coronal scanning plane that can fully display its entire course" in step one is: first, acquiring and processing image data reflecting the anatomy of the bile duct to construct a geometric model (i.e., a three-dimensional spatial path curve) representing the spatial course of the bile duct, and then calculating the optimal imaging plane based on the geometric model through a specific mathematical optimization method.

[0079] Specifically, "planning" corresponds to the construction process of the three-dimensional spatial path curve (S102) below, while "determining" corresponds to the process of determining the final personalized oblique coronal scanning plane based on the three-dimensional spatial path curve through optimal plane fitting calculation (S103). This progressive processing logic, "from image data to geometric model, and then from geometric model to imaging plane," ensures the "personalization" of the determined plane and the technical effect of "fully displaying its entire path." This core technical feature transforms the imaging plane from "device-centric" to "target anatomical structure-centric," which is the primary prerequisite for this method to obtain high-quality, targeted images.

[0080] Step 2: In the personalized oblique coronal scanning plane, imaging is performed using a small field-of-view excitation diffusion-weighted imaging sequence, which is diffusion-weighted imaging; low diffusion sensitivity factor diffusion images with a value not higher than 100 seconds / mm² and high diffusion sensitivity factor diffusion images with a value not lower than 800 seconds / mm² are acquired, which are respectively low b-value diffusion-weighted images and high b-value diffusion-weighted images.

[0081] One of the core aspects of this step is fully utilizing the hardware performance of the aforementioned ultra-high gradient magnetic resonance imaging system. Its ultra-high gradient field strength and switching rate significantly shorten the echo time (TE) of the diffusion-weighted imaging sequence. The shortened echo time (TE) means that the signal is acquired before undergoing severe transverse relaxation (T2) attenuation, thus significantly reducing signal loss and improving the image signal-to-noise ratio. This is a prerequisite for successfully implementing high b-value (≥800 s / mm²) diffusion-weighted imaging in oblique coronal positions with thin slices (not exceeding 3.0 mm) while maintaining the diagnostically required signal-to-noise ratio. Therefore, the ultra-high gradient system is not simply a hardware upgrade, but rather the physical basis for simultaneously meeting the three challenging imaging conditions of "thin slice," "high b-value," and "oblique coronal position" in this scheme; all four are indispensable.

[0082] This step, based on the personalized oblique coronal scanning plane determined in step one, configures and performs a specially optimized diffusion-weighted imaging scan, aiming to obtain high-resolution, high signal-to-noise ratio, and deformation-controlled image data on this specific plane.

[0083] S201: Set the imaging field of view. On the personalized oblique coronal scanning plane, enable the small field of view excitation imaging mode. The range of the imaging field of view is determined based on the projection of the target bile duct: using the projection of the three-dimensional spatial path curve obtained in step one onto the personalized oblique coronal scanning plane as the center line, extend this center line 10-20 mm to each side in the slice direction, extend it more than 100 mm to each side in the frequency encoding direction, and in the cephalopelvic direction, extend the coverage area from the top of the diaphragm to the lower end of the bile duct center line downwards by 10-20 mm. This covers the bile duct and its surrounding soft tissue, and the resulting rectangular area is defined as the small field of view excitation and acquisition range. This ensures that the imaging target is focused and minimizes distortion artifacts from irrelevant areas. Small field of view technology not only improves spatial resolution, but more importantly, it significantly reduces image geometric distortion caused by an excessively large field of view, which is crucial for bile duct imaging that requires precise anatomical visualization.

[0084] S202: Configure key imaging parameters. The main parameter settings are as follows:

[0085] The acquisition slice thickness is set to no more than 3.0 mm. This range matches the physiological thickness of the extrahepatic bile duct wall, which can effectively reduce the volume effect and make the boundaries between the bile duct wall and the bile in the lumen and the tumor outside the lumen clearly distinguishable on the image. This thin slice setting is the key to realizing "visualizing the bile duct wall", which improves the imaging from showing the obstruction of the lumen to showing the lesion of the duct wall itself.

[0086] Diffusion Sensitivity Factor: Two b values ​​are set, with the high b value being no less than 800 sec / mm² (usually 800 sec / mm² or 1000 sec / mm²) to adequately suppress background tissue signals and highlight tumor areas where water molecule diffusion is restricted; the high b value ensures high contrast between the tumor and surrounding benign tissues (such as inflammatory edema), significantly improving the identification of tumor boundaries.

[0087] Determining the echo time: To achieve the aforementioned thin-slice, high b-value imaging, a sufficiently short echo time must be used to avoid excessive signal attenuation. On a magnetic resonance imaging (MRI) device equipped with a gradient field strength of 200 mT / m and a gradient switching rate of 200 T / m / s, based on the device's physical performance, and according to the selected slice thickness, field of view, b-value, and minimum available gradient pulse width, the shortest achievable echo time (TE) is automatically calculated using standard sequence timing equations. Under this ultra-high gradient MRI system and the above parameter configuration, the echo time (TE) can typically be shortened to below 70 milliseconds, preferably between 50 and 70 milliseconds. This calculation is a direct coupling result of the device's hardware performance and sequence parameters, ensuring the signal-to-noise ratio and image fidelity of the imaging. A short echo time (TE) is the direct technical guarantee that this method can maintain a sufficient signal-to-noise ratio at high b-values, and its implementation is highly dependent on specific ultra-high gradient hardware.

[0088] S203: Set auxiliary control parameters: Use single-shot excitation planar echo imaging reading mode and trigger acquisition based on the respiratory navigation signal in step three. All the above parameters (field of view, slice thickness, b-value, calculated echo time) are configured and locked in the scan sequence at once.

[0089] S204: Perform scan: Based on the above configuration, start the scan sequence on the personalized oblique coronal scanning plane determined in step one, and begin to acquire the original k-space data of diffusion-weighted imaging. The original k-space data is the output of step two and will be used for image reconstruction in step four.

[0090] like Figure 2As shown in the figure, a magnetic resonance imaging (MRI) device with a gradient field strength of 45 mT / m and a gradient switching rate of 45 T / m / s is compared with an MRI device with a gradient field strength of 200 mT / m and a gradient switching rate of 200 T / m / s. The figure demonstrates that as the gradient intensity increases, the echo time (TE) of diffusion imaging decreases. A shorter TE reduces the transverse relaxation (T2) attenuation of the MRI signal, thereby improving the image signal-to-noise ratio (SNR). This ensures that high-resolution thin-slice imaging still maintains a sufficient SNR to guarantee excellent image quality for clinical diagnosis. This comparison intuitively reveals the decisive role of the ultra-high gradient system in achieving the core imaging parameter combination of this invention, emphasizing the indivisibility and hardware dependence of the solution.

[0091] Step 3: During the imaging process, each data acquisition is triggered based on the respiratory navigation signal during the end-expiratory stabilization period; after each acquisition, the quality of the acquired data is evaluated in real time. If the data does not meet the preset quality standards, it will not be adopted, and supplementary acquisition will be triggered in the subsequent respiratory cycle until complete data that meets all preset quality requirements is obtained.

[0092] This step, during the execution of the sequence configured in step two, synchronously triggers data acquisition via respiratory navigation and performs real-time quality assessment on each acquired data point. This iterative approach ensures that all ultimately acquired data meets preset quality standards. This dynamic quality control loop addresses the respiratory motion artifact problem in upper abdominal imaging by upgrading the passive "gated acquisition" to an active "assessment-reacquisition" intelligent iterative process, ensuring the uniformity and high quality of the data used for final reconstruction.

[0093] S301: Set up respiratory navigation and determine the trigger phase: Before the scan begins, a navigation strip is set in the top region of the diaphragm for real-time monitoring of respiratory motion. By observing the navigation signal, the end-expiratory steady period is defined as: the relatively stable phase before the respiratory motion curve transitions from the expiratory phase to the inspiratory phase, typically the plateau period after the curve reaches its lowest point, lasting longer than 100 milliseconds and with a motion amplitude change less than a preset threshold (e.g., 3 mm). This phase is set as the window for allowing trigger acquisition. Selecting the end-expiratory steady period for triggering maximizes the use of the moment when the abdominal motion amplitude is minimal and the position is most stable during the respiratory cycle, reducing motion ambiguity from the source.

[0094] S302: Perform respiratory-triggered data acquisition and real-time quality assessment: After the scan begins, continuously monitor the navigation signal. When respiratory motion enters the end-expiratory stabilization phase as defined above, immediately trigger a diffusion-weighted imaging data acquisition. After each acquisition, perform real-time analysis on the newly acquired raw k-space data to assess its quality. Assessment methods include:

[0095] Signal strength stability assessment: The signal strength of the central region of the k-space acquired this time (e.g., the central 10% of the phase encoding line) is compared with the corresponding signal strength of several previously acquired data. If the deviation exceeds the tolerance range determined based on historical data or pre-scan signal-to-noise ratio analysis (e.g., deviation from the mean ±20%), it is considered unstable. This assessment can effectively identify abnormal signals caused by instantaneous physiological fluctuations (such as blood flow, intestinal peristalsis) or equipment instability, ensuring signal consistency between multiple acquisitions.

[0096] Motion artifact evaluation: A Fast Fourier Transform (FFT) is performed on the raw k-space data acquired in this study to generate a preview image. The presence of significant blurring or artifacts caused by motion is determined by calculating the edge gradient along the phase encoding direction in the preview image or by using the image entropy index. If the index exceeds a preset threshold, motion artifacts are considered to be present. This method directly detects artifacts in the image space, exhibiting high sensitivity and the ability to capture minute irregular movements that navigation signals may not have completely filtered out.

[0097] S303: Based on the evaluation results, make decisions and perform supplementary data collection, establishing the following decision-making logic:

[0098] If the data collected this time passes the above two evaluations, it will be marked as "valid" and stored in the data buffer;

[0099] If the data collected fails any evaluation, it is marked as "invalid" and not adopted. Subsequently, the scan control logic will automatically insert a supplementary acquisition to replace the invalid acquisition when the trigger condition is met in the next respiratory cycle. This automatic decision-making logic realizes closed-loop feedback control of the acquisition process, requiring no manual intervention and ensuring the automation of the process and the reliability of the results.

[0100] S304: Repeat the above process until each required diffusion-weighted image (e.g., each b-value, each averaging count) has obtained a preset number of valid acquisition data points. These data will be used as input for image reconstruction in step four. Ultimately, all "valid data" used for reconstruction is obtained under ideal physiological conditions and has passed strict quality control, which provides data assurance for generating high-quality images without motion artifacts and with stable contrast.

[0101] Step 4: Reconstruct the data obtained in Step 3 to generate low b-value diffusion-weighted images and high b-value diffusion-weighted images of the personalized oblique coronal scanning plane; Spatially register the low b-value diffusion-weighted images, high b-value diffusion-weighted images and the magnetic resonance pancreaticobiliary hydrops image source image of the same examination, and output the registered multi-sequence fusion display image.

[0102] This step aims to integrate the unique functional images generated by this invention with traditional anatomical images to create a standardized imaging dataset that facilitates comprehensive clinical assessment. Its core value lies in precisely correlating and visualizing highly sensitive functional information (tumor invasion) with high-resolution anatomical information (bile ducts, vascular pathways) within a unified spatial framework, directly serving clinical decision-making.

[0103] S401: Reconstructing Specific Functional Images: For all valid raw k-space data obtained in Step 3, image reconstruction and post-processing (such as distortion correction) are performed to generate low b-value (≤100 s / mm²) diffusion-weighted images, high b-value (≥800 s / mm²) diffusion-weighted images, and corresponding apparent diffusion coefficient maps located on the personalized oblique coronal scanning plane from Step 1. The apparent diffusion coefficient map provides quantitative information on water molecule diffusion and can be cross-validated with the qualitative high signal of the low b-value diffusion-weighted images and high b-value diffusion-weighted images, increasing diagnostic specificity.

[0104] S402: Perform spatial registration of multiple image sequences: Using the low b-value diffusion-weighted image and high b-value diffusion-weighted image obtained in the previous step as references, a rigid registration algorithm based on similarity measures such as mutual information or normalized cross-correlation is used to spatially align the following images obtained in the same inspection with them:

[0105] 1. Maximum density projection map or source image of the source image of magnetic resonance pancreatobiliary hydrops imaging;

[0106] 2. Transverse T1-weighted enhanced image (resampled to a personalized oblique coronal scan plane via 3D interpolation).

[0107] The goal of registration is to eliminate anatomical positional deviations caused by slight patient movement between different scan sequences, ensuring consistent spatial coordinates for the same anatomical structure (such as bile ducts or blood vessels) in different images. Since all sequences share the same personalized scanning plane determined in step one as a reference or resampling target, the registration problem is greatly simplified, and the accuracy and robustness of registration are improved.

[0108] S403: Generate a fused display dataset: Combine the registered multi-sequence images into a single multimodal image dataset that can be viewed synchronously and interactively. Standard fusion and output methods include, but are not limited to:

[0109] Multi-view side-by-side display: Within the same screen window, low b-value diffusion-weighted images, high b-value diffusion-weighted images, apparent diffusion coefficient maps, registered MRCP images, and transverse T1-weighted enhanced images are displayed side-by-side, and the slice positions of all views can be moved synchronously; linked browsing allows doctors to intuitively compare the performance of the same location on different sequences and quickly make a comprehensive judgment.

[0110] Image fusion and overlay display: Low b-value diffusion-weighted images and high b-value diffusion-weighted images are translucently overlaid on the corresponding transverse T1-weighted enhanced images or MRCP images with a specific color spectrum (such as red) to form a fused image, which can intuitively show the relationship between the functional abnormal area and the anatomical structure; the fused overlay image makes it clear at a glance "where the lesion is", which is especially helpful in showing the spatial relationship between the tumor and key blood vessels and bile duct branches.

[0111] Coordinate information output: For continuous regions with abnormally high signal on low b-value diffusion-weighted images and high b-value diffusion-weighted images, the spatial range parameters of these regions in the personalized oblique coronal scanning plane coordinate system are automatically calculated and output, such as the position coordinates of the starting and ending points and the longitudinal span distance along the bile duct projection direction.

[0112] S404: Output the above fused display dataset and related coordinate information: This dataset is the final technical output of this imaging method, which can be used by doctors to perform visual comparison analysis and measurement to assist in completing diagnostic assessment.

[0113] To verify the clinical efficacy of this invention, two cases of extrahepatic cholangiocarcinoma imaging using this method are presented below. The final delivery of this invention is not a single image, but an integrated, interactive diagnostic solution containing quantitative information, directly addressing the core clinical needs of preoperative assessment of cholangiocarcinoma: "localization, characterization, quantification, and relationship determination."

[0114] Case 1: Common bile duct adenocarcinoma;

[0115] Patient information: Male, 61 years old;

[0116] Imaging findings:

[0117] Conventional coronal enhanced T1WI images (e.g.) Figure 5 (As shown) No obvious space-occupying lesions were observed;

[0118] MRCP images (such as) Figure 6 As shown in the image, the common bile duct is irregularly narrowed (as indicated by the white arrow), but the tumor itself and its surrounding infiltration are not clearly visible.

[0119] Images obtained using this method, such as 2 mm thin-slice diffusion-weighted imaging (b = 800 s / mm²) in the oblique coronal position (e.g.) Figure 7 and Figure 8 As shown in the image, the tumor lesion running along the common bile duct and its longitudinal infiltration extent (white arrow) are clearly displayed, significantly improving diagnostic confidence. Simultaneously, the image clearly shows a positive enlarged lymph node adjacent to the lesion. Figure 8 (White arrow) The lymph node was confirmed to be a metastatic lymph node during surgery.

[0120] Case 2: Hilar bile duct adenocarcinoma

[0121] Patient information: Female, 61 years old;

[0122] Imaging findings:

[0123] Conventional enhanced T1WI images (e.g.) Figure 9 (As shown) The extent of the lesion in the hilum of the liver is not clearly displayed (white arrow).

[0124] MRCP images (such as) Figure 10 The image shows a filling defect in the porta hepatis and dilatation of the bile duct due to obstruction, but it does not show enough of the tumor itself and its anatomical details.

[0125] Diffusion-weighted imaging (b=800 s / mm²) of oblique coronal 2.0 mm thin slice with small field of view (e.g.) was obtained using this method. Figure 11 As shown in the image, the extent of the tumor lesion in the hilum of the liver is clearly and completely displayed (white arrow), providing accurate imaging evidence for preoperative assessment.

[0126] The above embodiments and cases demonstrate that the thin-slice oblique coronal diffusion-weighted imaging method of the magnetic resonance imaging system provided by this invention can effectively overcome the limitations of conventional MRI sequences in assessing extrahepatic cholangiocarcinoma. Through personalized planar localization, thin-slice high b-value imaging, respiratory navigation gating, and iterative quality monitoring, this invention obtains oblique coronal diffusion-weighted images with high signal-to-noise ratio, low distortion, and high spatial resolution. These images clearly and continuously display the longitudinal infiltration range of cholangiocarcinoma along the bile duct and effectively complement conventional sequences such as MRCP and contrast-enhanced scans, thus providing more comprehensive and reliable imaging support for accurate preoperative staging, surgical planning, and prognostic assessment.

[0127] The above description is merely a preferred embodiment of the present invention and is illustrative rather than restrictive. Those skilled in the art will understand that many changes, modifications, and even equivalents can be made within the spirit and scope defined by the claims of the present invention, all of which will fall within the protection scope of the present invention.

Claims

1. A method of thin slice oblique coronal position focusing view diffusion weighted imaging of a magnetic resonance system, characterized by: The method comprises the following steps: Step 1: based on the pre-acquired magnetic resonance pancreaticobiliary water imaging source image and the transverse T2 weighted image, the three-dimensional spatial running path of the target bile duct is determined, and the three-dimensional spatial running path is optimally planar fitted to determine a personalized oblique coronal scanning plane for subsequent scanning; Step 2: in the personalized oblique coronal scanning plane, a small field of view excitation diffusion weighted imaging sequence is used for imaging, that is, diffusion weighted imaging; Step 3: during the imaging process, each data acquisition is triggered based on the respiratory navigation signal at the end-expiratory stable period; the quality of the collected data is evaluated in real time, if the data does not meet the preset quality standard, it will not be adopted, and supplementary acquisition will be triggered in the subsequent respiratory cycle until complete data meeting all preset quality requirements are obtained; Step 4: reconstruct the data obtained in step 3 to generate low-b-value diffusion weighted images and high-b-value diffusion weighted images of the personalized oblique coronal scanning plane, and perform spatial registration on the low-b-value diffusion weighted images, the high-b-value diffusion weighted images, and the magnetic resonance pancreaticobiliary water imaging source image and the transverse T1 weighted enhanced image of the same examination, and output the registered multi-sequence fusion display image.

2. The method of claim 1, wherein: In step 1, determining the three-dimensional spatial running path of the target bile duct comprises: Based on the signal intensity difference between the bile duct and the surrounding tissue in the magnetic resonance pancreaticobiliary water imaging source image and the transverse T2 weighted image, an image processing algorithm for extracting the center line of the blood vessel or tubular structure is used to automatically identify and extract the center line of the target bile duct lumen to form the three-dimensional spatial running path.

3. The method of claim 1, wherein: In step 1, the optimal planar fitting of the three-dimensional spatial running path comprises: Performing least squares planar fitting on the point set constituting the three-dimensional spatial running path, and determining the personalized oblique coronal scanning plane by solving the plane equation that minimizes the sum of the squares of the distances from the point set to the fitting plane; Wherein, the normal vector of the plane equation corresponds to the eigenvector corresponding to the smallest eigenvalue of the point set covariance matrix.

4. The method of claim 1 or 3, wherein: In step 1, determining the personalized oblique coronal scanning plane further comprises: Converting the geometric parameters of the plane equation obtained by optimal planar fitting into rotation angles and center offsets recognizable by the magnetic resonance scanning system by calling the application programming interface of the magnetic resonance scanning system or using its built-in positioning conversion tool, and locking the personalized oblique coronal scanning plane of all subsequent imaging sequences according to the parameters.

5. The method of claim 1, wherein: In step 2, the imaging field of view of the diffusion weighted imaging is set according to the projection of the target bile duct on the personalized oblique coronal scanning plane, specifically: The projection of the three-dimensional space walking path on the personalized oblique coronal scanning plane is taken as the center line, the center of the diffusion imaging field of view in the head-foot direction is taken as the center of the bile duct center line in the head-foot direction, the upper edge of the diffusion imaging field of view is taken as the top of the diaphragm, and the lower edge is taken as the lower end of the bile duct center line extending downward by 10-20 mm; the diffusion imaging expands the center line by 10-20 mm on both sides in the layer direction, and expands by more than 100 mm on both sides in the frequency encoding direction, so as to cover the bile duct and the surrounding soft tissue. The rectangular area formed thereby is defined as the excitation and acquisition range of the small field-of-view excitation diffusion weighted imaging sequence.

6. The method of claim 5, wherein: the step of acquiring the diffusion weighted images comprises: acquiring a low b-value diffusion weighted image with a diffusion sensitivity factor of not more than 100 s / mm2 and a high b-value diffusion weighted image with a diffusion sensitivity factor of not less than 800 s / mm2.

7. The method of claim 4, wherein: the step of acquiring the diffusion weighted images comprises: acquiring a low b-value diffusion weighted image with a diffusion sensitivity factor of not more than 100 s / mm2 and a high b-value diffusion weighted image with a diffusion sensitivity factor of not less than 800 s / mm2.

8. The method of claim 1, wherein: the step of real-time evaluating the quality of the acquired data comprises: signal intensity stability evaluation: comparing the signal intensity of the k-space center region of the current acquisition with the corresponding signal intensity of several accepted acquisitions, if the deviation exceeds the tolerance range determined based on the pre-scan signal-to-noise ratio analysis, the data is considered unstable; motion artifact evaluation: generating a preview image from the raw k-space data of the current acquisition, and judging whether there are obvious blurring or artifacts caused by motion by calculating the edge gradient along the phase encoding direction of the preview image or using the image entropy index.

9. The method of claim 1, wherein: the step of generating the apparent diffusion coefficient map based on the low b-value diffusion weighted image and the high b-value diffusion weighted image comprises: generating the apparent diffusion coefficient map based on the low b-value diffusion weighted image and the high b-value diffusion weighted image; and the step of displaying the apparent diffusion coefficient map, the low b-value diffusion weighted image and the high b-value diffusion weighted image in the multi-sequence fusion display image comprises: displaying the apparent diffusion coefficient map, the low b-value diffusion weighted image and the high b-value diffusion weighted image in the multi-sequence fusion display image.

10. The method of claim 1, wherein: the step of generating the apparent diffusion coefficient map based on the low b-value diffusion weighted image and the high b-value diffusion weighted image comprises: generating the apparent diffusion coefficient map based on the low b-value diffusion weighted image and the high b-value diffusion weighted image; and the step of displaying the apparent diffusion coefficient map, the low b-value diffusion weighted image and the high b-value diffusion weighted image in the multi-sequence fusion display image comprises: displaying the apparent diffusion coefficient map, the low b-value diffusion weighted image and the high b-value diffusion weighted image in the multi-sequence fusion display image. ​ ​ ​ ​ ​ ​ ​ ​ ​ Steps one, two, three and four are performed consecutively in the same examination sequence, and the MR pancreaticobiliary water imaging source images, the transverse T2-weighted images and the transverse Tl -weighted enhanced images are all acquired in the same MR examination session as the diffusion-weighted imaging.

Images