基于压缩感知的3D mDixon Quant对腰椎椎体脂肪含量的定量研究

3D mDixon Quant based on compressed SENSE for quantitative study of lumbar vertebral body fat content

目的探讨敏感度编码(sensitivity encoded,SENSE)和压缩感知(compressed SENSE,CS)不同加速因子(acceleration factor,AF)对3D mDixon Quant技术腰椎椎体脂肪定量的影响,并分析腰椎椎体脂肪含量(fat fraction,FF)与年龄、性别和体质量指数(body mass index,BMI)的关系。材料与方法2020年1~7月在大连医科大学附属第一医院招募健康志愿者96名,其中男性45名,女性51名,年龄16~79岁,平均年龄(43.85±17.98)岁。将志愿者按照年龄段分为3组,青年组(<40岁)、中年组(40~60岁)、老年组(>60岁)。在Philips Ingenia CX 3.0 T MRI设备上采用3D mDixon Quant序列并使用不同加速技术(SENSE和CS)结合不同加速因子(acceleration factor,AF)(SENSE AF=2、4)和(CS AF=2、3、4、5、6、7、8)对志愿者行全腰椎扫描。两名放射科医师在Philips ISP(IntelliSpace Portall Version7)工作站上采用双盲法测量腰椎椎体(L1~L5)FF值,并对不同AF时图像信噪比(signal noise ratio,SNR)、对比噪声比(contrast noise ratio,CNR)及扫描时间进行比较。结果两名观察者测量结果具有良好的一致性(ICC值均>0.75)。不同AF时3D mDixon Quant序列测量腰椎椎体FF值无明显统计学差异(P=0.653),但不同AF时SNR和CNR有明显统计学差异(P=0.001、0.006)。其中,CS3和CS7组、CS4和CS7组、CS4和CS8组间SNR和CNR差异有统计学意义(P<0.05)。CS6时扫描时间较SENSE 2时缩短了60.66%,但FF值、SNR和CNR较其他组(S2、S4、CS2、CS3、CS4、CS5)而言无明显统计学差异(P>0.05)。腰椎椎体FF值在3个年龄组间有明显统计学差异(F=20.876,P<0.01)。其中,青年组FF值分别小于中年组与老年组,差异有统计学意义(P<0.01),而中年组与老年组间差异无统计学意义(P=0.086)。青年组中,男性腰椎椎体FF值高于女性(P<0.05),而在中年组和老年组,男性FF值略低于女性,但无统计学差异(P>0.05)。男性腰椎椎体FF值与BMI呈中等正相关(r=0.634,P<0.01),而女性FF值与BMI无相关性(r=0.207,P=0.146)。结论3D mDixon Quant序列结合CS技术评估腰椎椎体脂肪含量是可靠的,且选择CS的AF最大为6时可在保证图像质量的同时大大缩短扫描时间,实现脂肪含量的快速、精准测量。年龄、性别和BMI等因素均对腰椎椎体脂肪含量产生影响,在对不同个体进行腰椎椎体脂肪含量分析和评估时应综合考虑这些影响因素。

Objective:To investigate the effects of different acceleration factors(AF)on lumbar vertebra fat quantification using 3D mDixon Quant sequence accelerated by sensitivity coding(SENSE)and compressed SENSE(CS-SENSE),and analyze the relationship among lumbar body fat fraction(FF),age,sex and body mass index(BMI).Materials and Methods:From January to July 2020,96 healthy volunteers were recruited,including 45 males and 51 females,from 16 to 79 years old,with an average age of 43.85±17.98 years.Volunteers were divided into three groups according to age:young group(<40 years old),middle-aged group(40—60 years old),and elderly group(>60 years old).All volunteers were imaged at Philips Ingenia CX 3.0 T MRI with a fat quantification sequence(3D mDixon Quant)for the entire lumbar spine,and the sequence was accelerated at different folds with either SENSE(acceleration factor S=2,4)or CS-SENSE(acceleration factor CS=2,3,4,5,6,7,8).FF,signal-noise-ratio(SNR),and contrast-noise-ration(CNR)were measured for the L1—L5 vertebra by two radiologists independently with a double-blind method,measurement carried out on a IntelliSpace Portal workstation(ISP version 7.0;Philips Healthcare,Best,the Netherlands).The measured parameters were compared among different AFs.Results:The measurement results of the two observers showed good agreement(ICC value>0.75).There was no statistically significant difference in the FF values measured by 3D mDixon Quant sequence with different AF(P=0.653),but there were statistically significant differences in SNR and CNR with different AF(P=0.001,0.006).Among them,there were statistically significant differences in SNR and CNR between CS3 and CS7 groups,CS4 and CS7 groups,and CS4 and CS8 groups(P<0.05).The scanning time of CS6 was 60.66%shorter than that of SENSE 2,but there were no statistically significant differences in FF values,SNR and CNR values compared with those of other groups(S2,S4,CS2,CS3,CS4 and CS5)(P>0.05).FF values were significantly different between the three age groups(F=20.876,P<0.01).Among them,FF of the young group was significantly lower(P<0.01)than that of the middle-aged and the elderly groups,while the difference between the middle-aged and the elderly groups was not statistically significant(P=0.086).In the young group,the lumbar vertebra FF of the male was higher than that of the female(P<0.05),while in the middle-aged group and the elderly group,the FF of the male was slightly lower than that of the female,albeit not significantly(P>0.05).The lumbar vertebrae FF in men was moderately positively correlated with BMI(r=0.634,P<0.01),while in females such correlation was not observed(r=0.207,P=0.146).Conclusions:The 3D mDixon Quant sequence combined with CS technology is reliable for assessing the lumbar vertebral body fat content.When the maximum AF of CS was selected to 6,the image quality and the measurement accuracy were maintained while the imaging time dramatically reduced.Age,gender,and BMI are all factors that may affect the lumbar vertebra fat content.These factors should be considered in the analysis and evaluation of lumbar vertebra fat content in different individuals.