Investigation Of Ultrashort Echo Time MRI In The Evaluation Of Osseous Changes In Axial Spondyloarthritis

Objective: To investigate the feasibility and value of new magnetic resonance(MR)techniques with ultrashort echo time,including ultrashort echo time(UTE)and zero echo time(ZTE)MRI,for the evaluation of osseous changes in axial spondyloarthritis(ax Sp A),and to use these techniques to explore the regulations of osseous changes.Methods: Firstly,the qualitative diagnostic and semi-quantitative scoring performance of ZTE MRI,which has an echo time approximately equal to zero and is sensitive to bone MR signals,was evaluated for osseous structural lesions of the sacroiliac joint in ax Sp A.The diagnostic performance of ZTE MRI was calculated and compared with conventional T1-weighted fast spin echo(T1 FSE)using CT as the reference standard,and the differences in scoring of structural lesions among different imaging methods were compared.The inter-method and inter-reader agreement of diagnosis and sum scores were also evaluated.Subsequently,before applying dual-echo UTE MRI,which enables quantitative assessment of cortical bone microarchitecture,to ax Sp A patients,the feasibility of in vivo study of this technique in the proximal femur and the regulations associated with cortical bone measurement parameters of the proximal femur in healthy populations were investigated.Dual-echo UTE MRI was performed on the proximal femur of young and middle-aged healthy volunteers with the same age group and gender composition as the ax Sp A patients,and the PI(porosity index)mapping was obtained by dividing the original images of the two echoes.Cortical bone PI histogram parameters and cortical bone area were measured to compare the differences in measured parameters between different lateral limbs and different genders,and the correlation of measurement parameters with age and body mass index(BMI)was analyzed.The repeatability of the cortical bone parameter measurements was also evaluated.Finally,dual-echo UTE MRI was applied to the quantitative study of proximal femoral cortical bone alterations in ax Sp A patients,comparing the PI and area of proximal femoral cortical bone in ax Sp A patients with gender-age matched healthy volunteers,and analyzing the correlation of cortical bone measurement parameters with demographic characteristics and disease-related characteristics in ax Sp A patients,and identifying the predictor variables among them that could influence cortical bone measurement parameters by multiple linear regression analysis.Results: The diagnostic accuracy of ZTE MRI was higher than that of T1 FSE for osseous structural lesions in sacroiliac joints,and the diagnostic sensitivity of ZTE MRI for erosions and sclerosis and the diagnostic specificity for joint space changes were higher than that of T1 FSE.Also,the diagnostic results of ZTE MRI were in better agreement with CT.For the sum scores of erosions and sclerosis,ZTE MRI and CT were significantly higher than T1 FSE,and ZTE MRI had more consistency with CT than T1 FSE for sum scores.Inter-reader agreement for the diagnosis and sum scores ZTE MRI and CT were similar and both were greater than T1 FSE.The quantitative study of proximal femoral cortical bone in healthy volunteers using dual-echo UTE MRI showed that the PI mean and median of the left cortical bone were greater than that of the right,the PI mean,median,standard deviation and area of the proximal femoral cortical bone were greater in males than in females,and proximal femoral cortical bone area,PI standard deviation,skewness,kurtosis,right proximal femoral cortical bone PI mean and median increased with increasing BMI.The UTE measurement parameters of proximal femoral cortical bone showed excellent inter-reader repeatability.Applying dual-echo UTE MRI to the quantitative assessment of proximal femoral cortical bone alterations in ax Sp A,it was found that ax Sp A patients had larger proximal femoral cortical bone PI and smaller area than controls.Factors associated with cortical bone measurement parameters included gender,BMI,disease duration,onset age,alkaline phosphatase(ALP),enthesitis,hip involvement,radiographic axial spondyloarthritis(r-ax Sp A),Sacroiliac Joint Structural Score(SSS)-backfill,SSS-ankylosis,and the use of nonsteroidal anti-inflammatory drugs(NSAIDs).Gender and BMI were the demographic predictor variables;ALP,SSS-backfill,SSS-ankylosis,and the use of NSAIDs were disease-related predictors;and gender,BMI,and the use of NSAIDs were the significant predictor variables after considering all factors together.Conclusions: ZTE MRI and UTE MRI can realize qualitative,semi-quantitative and quantitative evaluation of local and systemic osseous changes in ax Sp A,with good clinical feasibility and repeatability,providing a new and effective evaluation method for osseous changes in ax Sp A,which is expected to be more applied in clinical practice and related studies in the future.Part Ⅰ: Application of ZTE MRI in Qualitative and Semi-Quantitative Evaluation of Osseous Structural Lesions of Sacroiliac JointsObjective: To verify the clinical feasibility and advantages of ZTE MRI in the assessment of osseous structural lesions of sacroiliac joints caused by ax Sp A compared with conventional T1 FSE,using CT as the reference standard.Methods: Fifty-one patients aged 18 to 60 years with suspicion of sacroiliitis were recruited,and 80 sacroiliac joints(160 bone articular surfaces)from 40 patients were finally included for analysis.All patients underwent CT and MR scans within a one-week interval.Osseous structural lesions of sacroiliac joints were scored independently by two readers.The diagnostic capability of ZTE MRI and T1 FSE were calculated using CT as the reference standard,and were compared by the Mc Nemar test.Diagnosis agreements between MR sequences and CT as well as between readers were investigated using Cohen’s κappa tests.The Friedman M test was performed to compare the sum scores of different imaging methods,and the intraclass correlation coefficient(ICC)and Bland-Altman plots were used to evaluate the inter-method and inter-reader agreement of sum scores.Results: Diagnostic accuracy of ZTE MRI was higher than that of T1 FSE for erosions,sclerosis,and joint space changes(erosions at the joint level: 95.0% vs 81.3%,erosions at the bone level: 94.4% vs 78.1%;sclerosis at the joint level: 86.3% vs 56.3%,sclerosis at the bone level: 90.6% vs 72.5%;joint space changes: 91.3% vs 75.0%).ZTE MRI also improved sensitivity for detection of erosions and sclerosis(erosions at the joint level:98.2% vs 80.0%,erosions at the bone level: 95.8% vs 67.7%;sclerosis at the joint level:89.5% vs 40.4%,sclerosis at the bone level: 84.8% vs 33.3%)as well as specificity for detection of joint space changes(93.0% vs 67.4%).The sum scores of ZTE MRI and CT for erosions and sclerosis were greater than T1 FSE.ZTE MRI had more consistency with CT(diagnosis: κ values,0.669~0.883;sum scores: ICC,0.802~0.941)than T1 FSE(diagnosis: κ values,0.252~0.595;sum scores: ICC,0.215~0.901)for both diagnosis and sum scores.ZTE MRI(diagnosis: κ values,0.611~0.789;sum scores: ICC,0.828~0.922)showed comparable inter-reader agreement to CT(diagnosis: κ values,0.618 ~0.825;sum scores: ICC,0.833~0.938)and better inter-reader agreement than T1 FSE(diagnosis: κ values,0.300~0.706;sum scores: ICC,0.621~0.833).Conclusions: ZTE MRI showed higher accuracy and reliability than conventional T1-weighted MR sequence in the evaluation of osseous structural lesions of sacroiliitis and comparable reliability to CT.ZTE MRI,which has good feasibility and application value,is expected to be incorporated into clinical practice to improve the diagnostic performance.Part Ⅱ: Preliminary Quantitative Study of the Proximal Femoral Cortical Bone using Dual-Echo UTE MRIObjective: In order to realize the in vivo quantitative study of the proximal femoral cortical bone by dual-echo UTE MRI,explore the relevant factors affecting the measurement parameters of proximal femoral cortical bone in a healthy young and middle-aged population,verify the repeatability of dual-echo UTE MRI,and provide the necessary reference for the subsequent application of this technology on ax Sp A.Methods: Fifty-two healthy volunteers aged 18 to 55 years were included.All volunteers underwent dual-echo UTE MR scan of the proximal femur,which allowed simultaneous imaging of both femurs in a single scan.The cortical PI mapping was obtained using image processing software(PI is the ratio of signal intensity between the longer and shorter echo time images,which represents the pore water ratio and can indicate the porosity).ROIs(region of interest)were obtained by two radiologists who independently manually outlined the subject’s proximal femoral cortical bone on dual-echo UTE MRI images.The cortical bone PI histogram parameters(including mean,median,standard deviation,skewness and kurtosis)and cortical bone area were finally obtained.Paired t-test or Wilcoxon signed-rank test was used to compare the left and right proximal femoral cortical bone PI histogram parameters and cortical bone area.To exclude the interference of confounding factors,covariance analysis was used to compare the cortical bone PI histogram parameters and cortical bone area in different genders.Pearson or Spearman correlation analysis was used to determine the correlation of proximal femoral cortical bone measurement parameters with age and BMI.Repeatability of dual-echo UTE MRI was evaluated using ICC and Bland-Altman plots.Results: Participants included 33 males and 19 females,with no statistical difference in age by gender(P=0.586)and a greater BMI in males than in females(23.36±3.12 kg/m2 vs21.02±2.44 kg/m2,P=0.007).The PI mean and median of proximal femoral cortical bone were greater on the left than that on the right(mean: 34.24±3.38% vs 32.23±3.21%,P<0.001;median: 29.42±3.06% vs 27.16±2.63%,P<0.001).Males had larger proximal femoral cortical bone PI mean,median and standard deviation(e.g.,PI mean: left:35.71±2.36% vs 31.68±3.42%,P<0.001;right: 33.27±3.00% vs 30.43±2.80%,P=0.032)and larger proximal femoral cortical bone area(433.21±65.54 mm2 vs 336.22±40.99 mm2,P<0.001)compared to females.Correlation analysis showed that right proximal femoral cortical bone PI mean and median were positively correlated with BMI(PI mean: r=0.516,P<0.001;PI median: r=0.392,P=0.004);cortical bone PI standard deviation,skewness and kurtosis were positively correlated with BMI(standard deviation: r=0.271,P=0.005;skewness: r=0.308,P=0.001;kurtosis: r=0.258,P=0.008);and cortical bone area was positively correlated with BMI(r=0.340,P<0.001).The repeatability of dual-echo UTE MRI parameters was excellent,among which the PI mean was the best(cortical bone PI histogram parameters ICCs: mean: 0.985,median: 0.973,standard deviation: 0.963,skewness: 0.964,kurtosis: 0.955;cortical bone area ICC: 0.951).Conclusions: Dual-echo UTE MRI enabled in vivo quantitative assessment of the proximal femoral cortical bone microarchitecture in healthy young and middle-aged adults.Differences in cortical bone measurement parameters of the proximal femur could be caused by different lateral limbs and different genders,and there was a correlation between BMI and cortical bone measurement parameters.The repeatability of dual-echo UTE MRI parameters was excellent.All these have provided the necessary reference for subsequent studies using dual-echo UTE MRI for the evaluation of cortical bone in ax Sp A patients.Part III: Quantitative Study of Proximal Femoral Cortical Bone Changes in Ax Sp A by Dual-Echo UTE MRIObjective: The aim of this study was to quantitatively evaluate cortical bone loss in ax Sp A patients using dual-echo UTE MRI.To investigate changes of proximal femoral cortical bone microarchitecture in ax Sp A patients and to analyze the factors associated with bone loss in ax Sp A patients in order to identify meaningful predictor variables among them.Methods: A total of 72 ax SpA patients and 52 sex and age matched healthy volunteers were included for analysis.All volunteers underwent dual-echo UTE MR scan of the proximal femur.Cortical bone PI mapping(images with longer echo time divided by images with shorter echo time)was obtained using image processing software.ROIs of proximal femoral cortical bone were outlined by two radiologists independently,and cortical bone PI mean and cortical bone area were measured for subsequent analysis.Demographic characteristics of all subjects and disease-related characteristics of ax Sp A patients(including disease duration,blood test,involvement of other sites,imaging characteristics,and treatments)were collected.The PI as well as the area of proximal femoral cortical bone were compared between ax Sp A patients and healthy volunteers using two-sample t-test or Mann-Whitney U test.Subsequently,Pearson or Spearman correlation analysis and Point-biserial correlation analysis were used to explore the correlation between cortical bone PI as well as area and demographic and disease-related characteristics in ax Sp A patients.Finally,multiple linear regression analysis was performed using cortical bone PI and area as dependent variables to determine which variables were predictive variables affecting the proximal femoral cortical bone UTE measurements.Results: The 72 ax Sp A patients included 47 males and 25 females.There was no statistical difference in the gender composition of the ax Sp A patients and the healthy volunteers(P=0.835),as well as no statistical difference in age(P=0.231)and BMI(P=0.193)between the two groups.The PI of proximal femoral cortical bone was larger in ax Sp A patients on the left(42.82±5.29% vs 34.24±3.38%,P<0.001)and right side(40.44±5.40% vs 32.23±3.21%,P<0.001)than in controls,while the proximal femoral cortical bone area was smaller in ax Sp A patients than in controls(241.56±44.46 mm2 vs397.77±74.28 mm2,P<0.001).Correlation analysis of patients with ax Sp A showed that proximal femoral cortical bone measurements were associated with gender(left PI:r=0.456,right PI: r=0.493,area: r=0.466),BMI(left PI: r=0.341,right PI: r=0.276,area:r=0.267),disease duration(left PI: r=0.238),onset age(left PI: r=-0.294,right PI:r=-0.278),ALP(left PI: r=0.434,right PI: r=0.357),enthesitis(left PI: r=0.234,right PI:r=0.245),hip involvement(right PI: r=0.240),r-ax Sp A(left PI: r=0.287,right PI: r=0.274),SSS-backfill(left PI: r=0.239,right PI: r=0.248),SSS-ankylosis(left PI: r=0.251,right PI:r=0.298),and the use of NSAIDs(left PI: r=0.259,area: r=-0.241).Multiple linear regression analysis demonstrated that the demographic predictor variables were gender and BMI,and the disease-related predictor variables were ALP,SSS-backfill,SSS-ankylosis,and the use of NSAIDs.After considering all factors together,the significant predictor variables obtained were gender,BMI and the use of NSAIDs.Conclusions: Cortical bone loss in ax Sp A patients was evaluated using dual-echo UTE MRI.It was found that ax Sp A could lead to increased porosity and decreased area of proximal femoral cortical bone;and gender,BMI,ALP,SSS-backfill,SSS-ankylosis and the use of NSAIDs were predictor variables affecting cortical bone measurement parameters.This study demonstrated the value of dual-echo UTE MRI for the evaluation of bone loss in ax Sp A patients and provided new methods and ideas for future related studies and clinical practice.