| Object DWI sequence related parameters differentiated low grade group from high grade group well,but an obvious difference was observed among different fit model and definition method of ROI.So we aimed to explore which parameter and protocol was the best to differentiate low grade group and high grade group.Methods Forty-six patients with pathology proved of glioma were scanned with a 3.0T MR scanner before operation.The protocol of examination included routine sequence,MR spectroscopy,DWI sequence with multi-b value,DSC-perfusion sequence and routine enhancement sequence.The multi-b value DWI sequence was fitted with a multi-model protocol which includes mono-exponention model,bi-exponention model,strectch-exponention model and tri-exponention model.Three different draw methods of hot spot method,two-dimension method and three-dimension method were also include in the protocol.All related parameters of the protocol were compared between low grade group and high grade group,then further comparing among II grade,III grade and IV grade were performed.Finally the most valuable parameter and protocol were selected.Results By comparing the related parameters of spectroscopy between low grade and high grade group,no obvious difference were observed in NAA/Cr,Cho/Cr and Cho/NAA(P=0.412,P=0.148 and P=0.306 respectively).By comparing the related parameters of DSC-perfusion,obvious difference were observed in rCBF-+CMax,rCBF-rCBFMax,rCBV-+CMax,rCBV-rCBVMax,MTT-+CMax and MTT-MTTMax(P=0.042,P=0.030,P=0.002,P=0.001,P=0.001 and P=0.000 respectively).By comparing the related parameters of mono-exponention model for hot spot ROI method,obvious difference were observed in ADC-ADCmin(P=0.001),and no obvious difference in ADC-+CMax(P=0.273);By comparing the related parameters of mono-exponention model for two-dimension ROI method,obvious difference were observed in ADC-P10-2D(P=0.035),and no obvious difference in other parameters;By comparing the related parameters of mono-exponention model for three-dimension ROI method,obvious difference were observed in ADC-mean-3D,ADC-P5-3D and ADC-P10-3D(P=0.041,P=0.009 and P=0.005 respectively),and no obvious difference in other parameters.By comparing the related parameters of bi-exponention model for Hot spot ROI method,obvious difference were observed in D-Dmin and f-fMax(P=0.000 and P=0.002 respectively),and no obvious difference in other parameters;By comparing the related parameters of bi-exponention model for two-dimension ROI method,obvious difference were observed in D-P5-2D,D-P10-2D,f-mean-2D,f-median-2D,f-P90-2D and f-skewness-2D(P=0.039,P=0.022,P=0.001,P=0.000,P=0.016 and P=0.014 respectively),and no obvious difference in other parameters;By comparing the related parameters of bi-exponention model for three-dimension ROI method,obvious difference were observed in D-mean-3D,D-P5-3D,D-P10-3D,f-mean-3,f-median-3D and f-P90-3D(P=0.025,P=0.002,P=0.002,P=0.000,P=0.000 and P=0.043 respectively),and no obvious difference in other parameters.By comparing the related parameters of stretch-exponention model for hot spot ROI method,obvious difference were observed in DDC-DDCmin and ?-?min(P=0.001 and P=0.000 respectively),and no obvious difference in other parameters;By comparing the related parameters of stretch-exponention model for two-dimension ROI method,obvious difference were observed in DDC-P5-2D,DDC-P10-2D,a-mean-2D and a-median-2D(P=0.033,P=0.017,P=0.009 and P=0.004 respectively),and no obvious difference in other parameters;By comparing the related parameters of stretch-exponention model for three-dimension ROI method,obvious difference were observed in DDC-mean-3D,DDC-median-3D,DDC-P5-3D,DDC-P10-3D,a-mean-3D,a-median-3 D,a-P5-3D and a-P10-3D(P=0.067,P=0.047,P=0.006,P=0.005,P=0.005,P=0.006,P=0.033 and P=0.014 respectively),and no obvious difference in other parameters.By comparing the related parameters of tri-exponention model for hot spot ROI method,obvious difference were observed in Ff-+CMax,Ds-Dsmin and Ff-FfMax(P=0.010,P=0.005 and P=0.007 respectively),and no obvious difference in other parameters;By comparing the related parameters of tri-exponention model for two-dimension ROI method,obvious difference were observed in Ds-P5-2D,Ds-P 10-2D,Ff-mean-2D,Ff-median-2D and Ff-skewness-2D(P=0.030,P=0.009,P=0.007,P=0.015 and P=0.003 respectively),and no obvious difference in other parameters;By comparing the related parameters of tri-exponention model for three-dimension ROI method,obvious difference were observed in Ds-P5-3D,Ds-P10-3D,Ff-mean-3D,Ff-median-3D,Ff-skewness-3D,Fp-P90-3D and Fp-P95-3D(P=0.017,P=0.014,P=0.036,P=0.003,P=0.013,P=0.016 and P=0.017 respectively),and no obvious difference in other parameters.By using contralateral normal white matter region to normalize all parameters,for hot spot ROI method,normalized parameters were superior to non-normalized parameters,but it was opposite for two-dimension ROI method and three-dimension ROI method.By comparing the reliability of related parameters,two-dimension ROI method and three-dimension ROI method were superior to hot spot ROI method,and three-dimension ROI method was superior to two-dimension ROI method;For DSC-perfusion related parameters,rCBV-+CMax and rCBV-rCBVMax had lager ICC(ICC=0.994 and ICC=0.994 respectively)than other parameters;For mono-exponention model related parameters,ADC-P10-3D and ADC-P5-3D had lager ICC(ICC=0.994 and ICC=0.982 respectively)than other parameters;For bi-exponention model related parameters,f-mean-3D and f-median-2D had lager ICC(ICC=0.999 and ICC=0.998 respectively)than other parameters;For stretch-exponention model related parameters,a-median-3D and a-median-2D had lager ICC(ICC=0.997 and ICC=0.997 respectively)than other parameters;For tri-exponention model related parameters,Ff-median-3D and Ff-median-2D had lager ICC(ICC=0.997 and ICC=0.997 respectively)than other parameters.By exploring the correlation between related parameters and grade of WHO,rCBV-rCBVMax had the maximum correlation coefficient(r=0.695)for DSC-perfusion related parameters,ADC-P5-3D had the maximum correlation coefficient(r=-0.456)for mono-exponention model related parameters,f-fMax had the maximum correlation coefficient(r=0.507)for bi-exponention model related parameters,a-amin had the maximum correlation coefficient(r=-0.607)for stretch-exponention model related parameters,and Ff-+CMax had the maximum correlation coefficient(r=0.437)for tri-exponention model related parameters.By exploring the correlation between DWI sequence related parameters and DSC-perfusion related parameters,the maximum correlation coefficient was observed between f-median-2D and rCBV-rCBVMax(r=0.604);By exploring the correlation among diffusion related parameters,the maximum correlation coefficient was observed between ADC Vs DDC(r=0.968)and D Vs Ds(r=0.966).By comparing the related parameters of DSC-perfusion among grade II,grade III and grade IV,obvious difference were observed in rCBV-+CMax,rCBV-rCBVMax,MTT-+CMax and MTT-MTTMax(P=0.006,P=0.002,P=0.001 and P=0.000 respectively),and no obvious difference in other parameters;By comparing the related parameters of mono-expenention model among grade II,grade III and grade IV,obvious difference were observed in ADC-ADCmin,ADC-P10-2D,ADC-P5-3D and ADC-P10-3D(P=0.003,P=0.046,P=0.009 and P=0.015 respectively),and no obvious difference in other parameters;By comparing the related parameters of bi-expenention model among grade II,grade III and grade IV,obvious difference were observed in D-Dmin,f-fMax,D-P10-2D,f-mean-2D,f-median-2D,f-skewness-2D,D-P10-3D,f-mean-3D and f-median-3D(P=0.002,P=0.003,P=0.049,P=0.002,P=0.001,P=0.032,P=0.002,P=0.005,P=0.002 and P=0.000 respectively),and no obvious difference in other parameters;By comparing the related parameters of stretch-expenention model among grade ?,grade ? and grade ?,obvious difference were observed in DDC-DDCmin,a-amin,DDC-P10-2D,a-mean-2D,a-median-2D,DDC-P5-3D,DDC-P10-3D,a-mean-3D and a-median-3D(P=0.005,P=0.000,P=0.045,P=0.021,P=0.012,P=0.006,P=0.013,P=0.016 and P=0.025 respectively),and no obvious difference in other parameters;By comparing the related parameters of tri-expenention model among grade II,grade III and grade IV,obvious difference were observed in Ff-+CMax,Ds-Dsmin,Ff-FfMax,Ds-P10-2D,Ff-mean-2D,Ff-median-2D,Ff-skewness-2D,Ds-P5-3,Ds-P10-3D and Ff-median-3D(P=0.014,P=0.018,P=0.021,P=0.020,P=0.026,P=0.013,P=0.016,P=0.029,Ds-P5-3D,P=0.045 and P=0.011 respectively),and no obvious difference in other parameters.For multiple comparison of related parameters,obvious difference were observed in grade II Vs grade II and grade II Vs grade IV,but no obvious difference in and grade III Vs grade IV.Conclusions First,different fit model of DWI sequence with multi-b value have different valuable parameters,we should select suitable parameters based on fit model,bi-exponention model maybe the optimal fit model for differentiating between low grade and high-grade brain gliomas.second,different drawing method of ROI have different valuable parameters,we should select suitable parameters based on different ROI method,two-dimension method maybe the optimal method for differentiating between low grade and high-grade brain gliomas.Third,considering reliability of all parameters,related parameters of three-dimension method have the maximum ICC,then the related parameters of two-dimension method,and the related parameters of hot spot method with the minimum ICC.Fourth,f value has the maximum correlation coefficient with related parameters of DSC-perfusion,which reflects the perfusion information of glioma.Fifth,about related parameters of diffusion,ADC value and DDC value has the maximum correlation coefficient,and D value and Ds value has the maximum correlation coefficient also.Sixth,related parameters of extreme region are superior to obvious enhancement region.Seventh,optimal parameters can be a good distinction between grade ? and grade ?,grade ? and grade ? gliomas,but it is difficult to identify grade ? and grade ?.Finally,tri-exponention fitting model related parameters is of great value in preoperative grading of glioma. |
PDF Full Text Request