| Part Ⅰ:Evaluation of Cortical Bone Abnormalities in CKD Patients with Dual-Echo UTE SequencePurpose:Chronic kidney disease(CKD)has a significant negative impact on bone health.This study was used to explore the value of dual-echo UTE sequences in quantitative assessment of cortical bone abnormalities in patients with CKD.Materials and Methods:95 CKD patients(12 in stage 2,23 in stage 3,20 in stage 4,and 40 in stage 5)and 31 healthy volunteers were recruited.All subjects underwent calf tibia magnetic resonance dual-echo UTE scan,and Image J post-processing software was used to obtain cortical bone area and porosity index(PI)histogram parameters.The PI value is defined as the ratio of the image signal intensity of the longer to the shortest echo time of the dual-echo UTE,and the PI histogram parameters include the mean,median,minimum,maximum,standard deviation,skewness,and kurtosis.An independent-samples T test or Mann-Whitney U test was used to compare differences between the healthy and CKD groups.One-way ANOVA or Kruskal-Wallis H test was used to assess the differences among all five groups(control group+four CKD stages).A linear regression model was used to calculate the linear trend P(Ptrend)among the five groups.Spearman correlation analysis was used to evaluate the relationship between cortical bone parameters and five groups,and Pearson correlation analysis was used to evaluate the correlation between PI-mean and cortical bone area.Results:The cortical bone area in the CKD group was significant smaller than that in the control group(181.7 mm2 vs.235.7 mm2,P<0.001).Compared with control group,the PI histogram parameters of CKD group had a significant increase in mean(20.20%vs.18.58%,P=0.001),median(18.09%vs.16.92%,P=0.017),minimum(5.89%vs.5.30%,P=0.010),standard deviation(0.091 vs.0.074,P=0.001)and skewness(1.78 vs.1.41,P=0.023);the maximum(P=0.059)and kurtosis(P=0.149)trended to increase,but the differences were not statistically significant.With CKD stages increased,the cortical bone area(Ptrend<0.001)tended to decrease,while the mean(Ptrend<0.001),median(Ptrend<0.001),and standard deviation(Ptrend=0.002)of PI tended to increase.PI-mean was negatively correlated with cortical bone area(r=-3.470,P<0.001).Conclusion:This study found that with the increase of CKD severity,the cortical bone area decreased,while the mean,median and standard deviation of cortical bone PI histogram parameters increased.Cortical bone area was negatively correlated with PI-mean.This study demonstrates the feasibility of dual-echo UTE for quantitative assessment of cortical bone abnormalities in CKD patients.Part Ⅱ: Mediation Effect Analysis among CKD Stages,Bone Metabolism Markers and Porosity IndexPurpose: The mechanism of increased cortical bone porosity caused by chronic kidney disease(CKD)is still unclear.Although a variety of bone metabolism markers can reflect CKD bone turnover to a certain extent,their application is still controversial.This study aimed to evaluate the mediation effect of multiple bone metabolism markers in the effect of increased CKD stage on cortical bone porosity using mediation effect analysis.Materials and Methods: 95 CKD patients(12 in stage 2,23 in stage 3,20 in stage 4,and 40 in stage 5).All subjects underwent calf tibia magnetic resonance dual-echo UTE scan to obtain cortical bone porosity index(PI).The PI value is defined as the ratio of the image signal intensity of the longer to the shortest echo time of the dual-echo UTE.Cortical bone PI values were obtained using Image J processing software.Serum corrected calcium(c Ca),phosphorus,intact parathyroid hormone(i PTH),25-hydroxyvitamin D(25(OH)D),β-Cross Laps(β-CTX),osteocalcin(OC),total procollagen type I amino-terminal propeptide(T-P1NP),and lumbar bone mineral density(BMD)were measured within one week after the MRI.One-way ANOVA or Kruskal-Wallis H test was used to assess the differences among four groups.A linear regression model was used to calculate the linear trend P(Ptrend)among four groups.Partial correlation analysis was performed to address associations between PI,e GFR and the above indicators.Multiple linear regression models were used to assess the association between CKD stages and PI value.Then,a separate exploratory mediation effect analysis was carried out to explore the impact of CKD stages and mediators on the PI value.Results: There were significant differences in PI value,phosphorus,i PTH,β-CTX,OC and T-P1 NP among the 4 groups(all P < 0.001),and with the increase of CKD stage,the above indicators showed an increasing trend(all Ptrend < 0.001).There were no significant differences in c Ca(P = 0.230),25(OH)D(P = 0.297),lumbar spine a BMD(P = 0.293)and T-score(P = 0.348)among the four groups.Partial correlation analysis showed that the correlations between PI value and various bone metabolism markers were statistically significant(phosphorus: r = 0.269,P = 0.009;i PTH: r = 0.381,P < 0.001;β-CTX: r = 0.375,P < 0.001;OC: r = 0.357,P < 0.001;T-P1NP: r = 0.327,P = 0.001).Likewise,there were statistically significant correlations between e GFR and various bone metabolism markers(phosphorus: r =-0.428,P < 0.001;i PTH: r =-0.499,P < 0.001;β-CTX: r =-0.470,P < 0.001;OC: r =-0.538,P < 0.001;T-P1NP: r =-0.512,P < 0.001;a BMD: r = 0.227,P = 0.029;BMD T-score: r = 0.215,P = 0.039).Statistically significant mediators were i PTH and β-CTX.The mediating effect analysis shows that,one increase in CKD stage was associated with an increase of 58.87 pg/m L(95% CI: 38.00,79.73)in i PTH,and associated with an increase of 0.545%(95% CI: 0.096,1.027)in PI;after controlling for the effect of CKD stages on PI,1 pg/m L increase in i PTH was associated with an increase of 0.005%(95% CI: 0.001,0.009)in PI;the increased i PTH mediated 34.4% of the total effect of CKD stages on PI.Similarly,one increase in CKD stage was associated with an increase of 0.47 ng/m L(95% CI: 0.30,0.63)in β-CTX,and associated with an increase of 0.575%(95% CI: 0.099,1.058)in PI;after controlling for the effect of CKD stages on PI,1 ng/m L increase in β-CTX was associated with an increase of 0.549%(95% CI: 0.027,1.071)in PI;the increased β-CTX mediated 30.8% of the total effect of CKD stages on PI.Conclusion: This study found that more severe CKD was associated with higher PI value.The association of CKD stages and PI mediated 34.4% and 30.8% of the total effect by increased i PTH and β-CTX,respectively.This study provides a new idea to monitor bone health in patients with CKD,and reveals the internal mechanism of bone deterioration caused by CKD to some extent.Part Ⅲ: Correlation Analysis of Multi-Echo UTE-T2* Fitting Parameters with Porosity Index and Cortical Bone AreaPurpose: Compared with the dual-echo UTE,the multi-echo UTE multi-component T2* fitting analysis can further obtain the fractions and T2* values of pore water,bound water and fat.This study aimed to investigate the consistency and stability of different analytical methods of UTE in assessing cortical bone composition.Materials and Methods: 16 CKD patients(4 in stage 3,4 in stage 4,and 8 in stage 5)and 10 healthy volunteers were recruited.All subjects underwent calf tibia magnetic resonance UTE scan,and obtained 7 sets of dual-echo UTE images.One set was used to calculate the cortical bone porosity index(PI).The PI value is defined as the ratio of the image signal intensity of the longer to the shortest echo time of the dual-echo UTE.UTE images of 14 echo in 7 sets were used for the construction of bi-and tri-component T2* fitting models.Cortical bone PI values were obtained using Image J processing software.Bi-component and tri-component fitting analysis were performed using MATLAB software to obtain cortical bone bound water fraction(Frac BW),bound water T2* value(T2*BW),pore water fraction(Frac PW),pore water T2* value(T2*PW),fat fraction(Frac Fat)and fat T2* value(T2*Fat).An independent-samples T test or Mann-Whitney U test was used to compare differences between the healthy and CKD groups.Pearson correlation analysis evaluated the correlation between cortical bone area,PI value and UTE-T2* fitting parameters.Results: Compared with the control group,the cortical bone area was significantly decreased(151.0 mm2 vs.230.2 mm2,P < 0.001),and the PI value was significantly increased(21.54% vs.18.96%,P = 0.004)in the CKD group.In the bi-component fitting analysis,compared with the control group,the CKD group had a significantly lower Frac BW(66.53% vs.80.65%,P < 0.001),a significantly higher Frac PW(33.47% vs.19.35%,P < 0.001)and T2*PW(3.50 ms vs.2.58 ms,P < 0.001),and there was no significant difference in T2*BW between the two groups.In the tri-component fitting analysis,compared with the control group,the CKD group had a significantly higher Frac PW(40.40% vs.33.63%,P = 0.023)and T2*PW(3.00 ms vs.1.84 ms,P = 0.001),and there was no significant difference in Frac BW,T2*BW,Frac Fat and T2*Fat between the two groups.The correlation analysis of bi-component fitting indicated that cortical bone PI value was negatively correlated with Frac BW(r =-0.634,P < 0.001),positively correlated with Frac PW(r = 0.634,P < 0.001)and T2*PW(r = 0.405,P = 0.040),but not significantly correlated with T2*BW.In the tri-component fitting,cortical bone PI value was negatively correlated with Frac BW(r =-0.432,= 0.027),positively correlated with Frac PW(r = 0.663,P < 0.001)and T2*PW(r = 0.401,P = 0.042),but not significantly correlated with T2*BW,Frac Fat and T2*Fat.In the bi-component fitting,cortical bone area was positively correlated with Frac BW(r = 0.438,P = 0.025),negatively correlated with Frac PW(r =-0.438,P = 0.025)and T2*PW(r =-0.534,P = 0.005),but not significantly correlated with T2*BW.In the tri-component fitting,cortical bone area value was positively correlated with Frac BW(r = 0.401,P = 0.042),negatively correlated with Frac PW(r =-0.505,P = 0.009)and T2*PW(r =-0.493,P = 0.010),but not significantly correlated with T2*BW,Frac Fat and T2*Fat.Conclusion: The cortical bone pore water fraction and pore water T2* value calculated by UTE-T2* fitting were different between CKD patients and healthy people.The PI value and cortical bone area have a good correlation with the pore water parameters obtained by UTE-T2* fitting.This indicates that different analysis methods using UTE sequences have high consistency and stability in evaluating cortical bone abnormalities. |