摘要
目的提出一种二阶差分质量加权的区域迭代相位提取(SOD-RIPE)Dixon水脂分离方法,以求在磁场不均匀的区域获得更好的水脂分离效果。方法利用Dixon的信号模型,对非对称正反相位图像矩阵进行正相位的角度消除,得到J1和J2,利用余弦定理得到可能的水信号强度(B,S)和正反相位差(P1,P2)。二阶差分计算出J2的相位质量图R,并对正反相位差进行加权,得到(Pv,Pu)。对Pv和Pu使用区域迭代相位提取方法进行迭代选择,得到Pf。Pf带入Dixon信号模型,利用最小二乘法,得到水脂分离图(W,F)。结果对philips ingenia Ⅱ 3.0T磁共振扫描的1000对正反相位图像进行水脂分离。SOD-RIPE算法在全身各部位和稳定性测试中均有比自动生长法和RIPE更好的分离度和稳定性,与飞利浦mDixon-XD算法性能相当。结论 SOD-RIPE具有较高的水脂分离度和稳定性,可作为通用的Dixon水脂分离方法。
Objective To propose a regional iterative phase extraction Dixon method based on second order difference quality weighting(SOD-RIPE) for improving water-lipid separation in heterogeneous magnetic field.Method The in-phase angle of the asymmetric in-phase and opposite phase image matrix was eliminated using the Dixon’s signal model to obtain J1 and J2,from which the possible water signal magnitude(B, S) and in-phase and opposite phase difference(P1, P2) was derived using the Cosine law.The phase quality map R of J2 was calculated using the second-order difference, and the phase difference was weighted to obtain(Pv, Pu).The Pv and Pu were iteratively selected to obtain the Pf.The water-fat separation diagram(W, F) was obtained using the least square method to bring the Pf into the Dixon signal model.Results Water-lipid separation was performed using 1000 pairs of in-and opposite-phase images on a Philips in Genia Ⅱ 3.0T magnetic resonance scanner.The SOD-RIPE algorithm achieved better separation and stability than the automatic growth method and RIPE in all the parts of the body and in the stability test, and had a similar performance to mDixon-XD algorithm.Conclusion SOD-RIPE can achieve robust water-fat separation with a good stability and can be used as a general Dixon water-fat separation method.
作者
侯召瑞
王瑞
葛洪
居胜红
HOU Zhaorui;WANG Rui;GE Hong;JU Shenghong(Department of Radiology,Zhongda Hospital Affiliated to Southeast University,Nanjing 210009,China)
出处
《南方医科大学学报》
CAS
CSCD
北大核心
2021年第2期305-309,共5页
Journal of Southern Medical University
关键词
质量图
可靠度
磁共振成像
DIXON
quality map
reliability
magnetic resonance imaging
Dixon