CN-122017704-A - Water-fat separation magnetic resonance imaging method, medium and equipment

CN122017704ACN 122017704 ACN122017704 ACN 122017704ACN-122017704-A

Abstract

The invention relates to a water-fat separation magnetic resonance imaging method, medium and equipment. S1, collecting images of water and fat in same phase and opposite phase, and calculating to obtain a winding phase diagram. S2, calculating to obtain an unwrapped phase diagram based on a phase unwrapping method of pixel clustering and local surface fitting, and S3, completing water-fat separation according to the true phase obtained by unwrapping. The water-fat separation magnetic resonance imaging method of the invention adopts a two-dimensional phase imaging algorithm based on pixel clustering and local surface fitting to perform phase unwrapping, can still robustly obtain potential real phases under the conditions of serious noise, rapid phase change or non-connected areas, improves the accuracy of water-fat separation, can effectively avoid water-fat exchange phenomenon, generates accurate water images and fat images, meets clinical diagnosis requirements, and provides a new technical scheme for phase-related magnetic resonance imaging application.

Inventors

FENG YANQIU
CHENG SHUYU
ZHANG WENYAN
GE YUWEI
QIAO PENG

Assignees

宁波市第二医院
南方医科大学

Dates

Publication Date: 20260512
Application Date: 20260119

Claims (10)

1. The water-fat separation magnetic resonance imaging method is characterized by comprising the following steps of: S1, collecting images of the same water and grease and opposite phases, and calculating to obtain a winding phase diagram; s2, calculating to obtain an unwrapped phase map based on a phase unwrapping method of pixel clustering and local surface fitting; And S3, completing water-fat separation according to the real phase obtained by unwrapping.
2. The method of claim 1, wherein in step S2, the same-phase and opposite-phase images of the water and fat are acquired based on FSE-Dixon and GRE-Dixon sequences.
3. The method according to claim 1, wherein the step S2 comprises the steps of: S21, constructing a multi-resolution image pyramid based on the acquired same-phase and opposite-phase images; s22, adopting a layered phase unwrapping strategy from the lowest resolution layer to the highest resolution layer of the multi-resolution image pyramid, carrying out phase unwrapping by combining pixel clustering and local surface fitting CLOSE algorithm, and guiding phase unwrapping of the high-resolution layer by an unwrapping result of the low-resolution layer.
4. A method according to claim 3, wherein the construction of the multi-resolution image pyramid in step S21 includes downsampling and upsampling, each using bilinear interpolation, with a scaling factor of 0.5, and the number of layers of the multi-resolution image pyramid is 3.
5. A method according to claim 3, wherein the layered phase unwrapping in step S22 comprises: (a) Performing initial phase unwrapping on the phase to be unwrapped of the current resolution layer by using an up-sampling phase map from an unwrapping result of the low resolution layer; (b) Inputting an initial phase unwrapping result into a CLOSE algorithm, and dividing pixels into an easy unwrapped block and an difficult unwrapped residual pixel according to a local change threshold value of a real phase and a preset minimum block size; (c) And (3) carrying out intra-block unwrapping and inter-block unwrapping on the easy unwrapped blocks and unwrapping on the difficult unwrapped residual pixels in sequence by adopting a region growing local polynomial surface fitting method.
6. The method of claim 5, wherein the initial phase unwrapping in step (a) is calculated as: wherein ; As a result of the initial phase unwrapping, For the k-th layer to be unwrapped phase, n is an integer offset, To upsample the phase, the round () function is used to calculate the nearest integer.
7. The method of claim 5, wherein the minimum block size preset in step (b) is adjusted according to a resolution layer.
8. The method according to claim 1, wherein in the step S3, the water-fat separation is calculated by the formula: Wherein, the As the same phase image, In the form of an anti-phase image, For the true phase after unwrapping, W is the water signal and F is the fat signal.
9. A computer readable storage medium storing computer instructions for causing a processor to perform the method of any one of claims 1-8 for phase unwrapping and water-fat separation imaging.
10. A magnetic resonance imaging apparatus, characterized in that phase unwrapping and water-fat separation imaging are performed using the method of any one of claims 1-8.

Description

Water-fat separation magnetic resonance imaging method, medium and equipment Technical Field The invention belongs to the field of cardiac magnetic resonance imaging, and particularly relates to a water-fat separation magnetic resonance imaging method, medium and equipment. Background Fat is an important constituent of human tissue, and is mainly distributed between abdominal cavity, subcutaneous and muscle fibers, with important physiological functions. In magnetic resonance imaging, fat has a shorter T1 and a longer T2, and a high signal is present on the T1 and T2 weighted images. Due to these characteristics of fat, defects such as masking lesions, reducing enhanced scanning effects, etc., exist in clinical magnetic resonance examinations, which interfere with diagnosis. In addition, there is a clinical need for quantitative measurement of fat content, such as fatty liver degree assessment and the like. Currently, the most clinically significant fat suppression techniques are frequency selective saturation (CHEMICAL SHIFT SELECTIVE presaturation), short-time inversion recovery (short-tau inversion recovery, STIR), and Dixon. In practical application, although the frequency selective saturation technology has the advantages of high selectivity, simplicity, convenience and practicability, the method has high field strength dependence and high requirements on uniformity of a main magnetic field and a radio frequency field. The STIR technique is less dependent on the main magnetic field and the radio frequency field than the frequency selective saturation technique, but has the disadvantages of low image signal to noise ratio, low selectivity of signal suppression, long scanning time, incompatibility with contrast agent enhanced imaging, and the like. The fat inhibition can only be realized in the former two ways, and the Dixon method is used as a water-fat separation technology, so that fat can be quantified while the fat inhibition is realized. The Dixon method is proposed in 1984 by William Thomas Dixon, and by utilizing the chemical displacement difference of water and fat protons and adjusting sequence parameters, images with the included angle of 0 degrees and images with the included angle of pi are respectively collected, namely a water-fat homophase diagram and a water-fat anti-phase diagram, and then the two diagrams are used for calculating a water diagram and a fat diagram. The original two-point Dixon technology causes phase errors caused by the non-uniformity of a main magnetic field, so that the obtained water image and fat image have water and fat displacement. To overcome this drawback, in 1991, glover and Schneider proposed a three-point Dixon method that increased one more measurement (-pi, 0, pi) on the basis of the original two measurements. The phase error is corrected using the redundant information. Since three-point Dixon needs to make three measurements, the scanning time is long and is easy to be interfered by motion artifacts. In fact, the three-point Dixon data contains redundant information, and phase error correction can be realized by only using the collected two-point Dixon data, so that the expanded two-point Dixon method is proposed by Thomas E. Skinner in 1996, and compared with the three-point Dixon method, the expanded two-point Dixon method can reduce data collection time and is widely applied in practice. However, the method has certain defects that the water and fat ratio is close near the boundary of the water and the fat, the signal to noise ratio is low, and the phase error correction is inaccurate due to the uneven magnetic field. The Dixon method is critical in correcting phase errors caused by inhomogeneities in the main magnetic field. The currently developed phase error correction methods mainly include a phase unwrapping algorithm (phase unwrapping), a region growing algorithm (region-growing), a region iterative vector selection algorithm (regional iterative phasor extraction), and the like. Both of these methods have driven the development of Dixon technology. In recent years, the scanning speed is further improved by combining a multi-echo Dixon sequence with a coil parallel acquisition technology, and the fat quantitative measurement precision is further improved because of being based on a multimodal fat model, so that the method has been successfully applied to accurate fat quantitative research of livers. But the Dixon technology still has the problems that (1) the phase winding phenomenon caused by the non-uniform magnetic field is overweight, and the Dixon method is sensitive to the non-uniformity of the main magnetic field depending on the chemical displacement difference of water and fat protons, which may cause a larger phase error and affect the separation effect of the water and the fat. (2) And in the phase correction process of the two-point Dixon method, the water and fat are close in proportion and low in signal-to-noise ratio near the wate