| Part ? To evaluate the characteristics of DKI quantitative parameters in different zones of benign prostatic hyperplasiaObjective:To investigate the characteristics of diffusion kurtosis imaging(DKI)post-processing quantitative parameters in different prostatic zones of benign prostatic hyperplasia(BPH)[transition zone(BPH-TZ)and peripheral zone(BPH-PZ)],and to evaluate the difference and degree in DKI quantitative parameters between the two zones.Materials and Methods:The imaging,pathology and clinical data of patients with BPH confirmed by pathology in our hospital from March 2017 to December 2017 underwent conventional multi-parameter MRI(Mp-MRI)and DKI were analyzed retrospectively.The DKI(b=0,700 and 1400 s/mm2)were attained by applying diffusion gradients in independent 15 different directions for each b value.Transrectal ultrasound(TRUS)guided system biopsy(10+X),partial patients combined with TRUS-MRI fusion targeted biopsy pathology as reference,the pathological results were obtained in all cases.MRI images of prostate in patients who met the inclusion criteria were post-processed.The region of interested(ROI)setting used a three-dimensional full-volume method to manually delineate on each consecutive T2W imaging in the BPH-TZ and BPH-PZ,respectively.Then the ROI were copied from T2W imaging to the same anatomical location in DKI images through image registration,which provide by the Python-based post-processing software.In the two zones,the K,D(×10-3mm2/s)and FA value of DKI post-processing parameters were obtained by voxel-by-voxel fitting of the kurtosis model.A paired sample T-test was used to determine whether there was a significant difference in these parameters between BPH-TZ and BPH-PZ.At the same time,the receiver operating characteristic(ROC)curve analysis was used to evaluate the degree of difference about the DKI post-processing parameters between BPH-TZ and BPH-PZ.The results were expressed by the area under the ROC curve(AUC).The Z test was used to judge whether the AUCs of these parameters were significantly different.Results:The final 75 patients were enrolled in the study,aged 52 to 86 years,mean(69.96±8.21)years old,PSA 9.12(6.72?13.28)ng/mL.The K value,D value and FA value of the DKI post-processing parameters were 0.94±0.08,1.49±0.12 and 0.29±0.07 in BPH-TZ,respectively,and 0.81±0.15,1.61±0.23 and 0.27±0.04 in BPH-PZ,respectively.Statistical analysis showed that the post-processing parameters of DKI were significantly different between BPH-TZ and BPH-PZ.The D value of TZ was lower than PZ(P<0.001),while the K value FA value of BPH-TZ were significantly higher(P<0.001 and P=0.013).ROC curve analysis showed that the AUC of K value,D value and FA value were 0.763[95%CI,0.681-0.845],0.667[95%CI,0.567-0.758],and 0.577[95%CI,0.482-0.672],respectively.When comparing AUC,the AUC of K value was significantly higher than D value and FA value(Z value was 3.490,3.111,and P value was 0.001,0.002,respectively).Conclusions:In summary,the post-processing quantitative parameters of DKI have certain characteristics in BPH-TZ and BPH-PZ tissues.D,K and FA values were significant different between BPH-TZ and BPH-PZ tissues.the D value of BPH-TZ tissue was lower than that of BPH-PZ tissue,while the K value and FA value were higher than BPH-PZ tissue,and the K value was the most different between the two zones.Part ? Preliminary application of DKI quantitative parameters in differentiating prostate cancer and benign prostatic hyperplasiaObjective:To investigate the application value of diffusion kurtosis imaging(DKI)post-processing quantitative parameters in the differential diagnosis of prostate cancer(PCa)and benign prostatic hyperplasia transition zone(BPH-TZ)or peripheral zone(BPH-PZ),and compared the diagnostic performance with Diffusion-weighted imaging ADC values.Materials and Methods:The imaging,pathology and clinical data were analyzed retrospectively from March 2017 to December 2017 in our hospital with patients who has underwent conventional multi-parametric MRI(Mp-MRI)and DKI examination.The traditional DWI took 9 b-values,and each b-value applied three orthogonal diffusion gradients,and the DKI(b=0,700 and 1400 s/mm2)were attained by applying diffusion gradients in independent 15 different directions for each b value.Transrectal ultrasound(TRUS)guided system biopsy(10+X),partial patients combined with TRUS-MRI fusion targeted biopsy pathology as reference,pathological results were obtained in the included cases.According to the pathological results,the patients were divided into PCa group and BPH group.MRI images of prostate in patients who met the inclusion criteria were post-processed.The region of interested(ROI)setting used a three-dimensional full-volume method to manually delineated on each consecutive T2W imaging in the PCa zones of PCa group and BPH-TZ or BPH-PZ of BPH group,respectively.Then the ROI were copied from T2W imaging to the same anatomical location in DWI and DKI images through image registration,which provide by the Python-based post-processing software.The ADC value was obtained by monoexponential fitting of DWI images with five of these b-values(b=0,200,500,1000,2000 s/mm2),and the K,D and FA value of DKI post-processing quantitative parameters were obtained by voxel-by-voxel fitting with the kurtosis model.First,whether the DKI and DWI post-processing quantitative parameters were differed between PCa and BPH-TZ or BPH-PZ were determined by independent sample T-test.The receiver operating characteristic(ROC)curve was then used to evaluate the diagnostic efficacy in distinguishing PCa and BPH-TZ or BPH-PZ of each quantitative parameter individually and DKI post-processing quantitative parameters combined with a logistic regression.The Z-test was used to determine whether their area under the curve(AUC)was different.Pearson correlation analysis was used to investigate the correlation between D value and K value or D value and ADC value in PCa,BPH-TZ or BPH-PZ tissues.Results:The final 116 patients were enrolled in the study,including 41 patients with PCa and 75 patients with BPH.The results showed that compared with BPH-TZ or BPH-PZ,the K value and FA value were higher in PCa tissues(K value 1.25±0.23 vs 0.94±0.08,0.81±0.15,all P<0.05;FA value 0.33±0.06 vs 0.29±0.07,0.27±0.04,all P<0.05),while D value(×10-3mm2/s)and ADC value(×10-3mm2/s)were lower than BPH-TZ or BPH-PZ tissues(D value 1.19±0.24 vs 1.49±0.12,1.61±0.23,all P<0.05;ADC value 0.88±0.31 vs 1.27±0.20,1.44±0.35,all P<0.05).The results of ROC analysis showed that the K value had a higher AUC value than the FA value(0.914[95%CI,0.848-0.958]vs 0.668[95%CI,0.574-0.753],P<0.001)when identifying PCa and BPH-TZ,but there was no significant difference when compared that with the AUC value of D or ADC value(0.914[95%CI,0.848-0.958]vs 0.855[95%CI,0.778-0.914],0.853[95%CI,0.776-0.912],P=0.104,0.106,respectively).The K value presented a higher AUC value than the D,FA and ADC values when identifying PCa and BPH-PZ tissues(0.946[95%CI,0.887-0.979]vs 0.889[95%CI,0.817-0.940],0.787[95%CI,0.702-0.858],and 0.880[95%CI,0.807-0.933],P=0.026,<0.001,and 0.030,respectively).However,when the DKI post-processing parameters were combined to distinguish PCa and BPH-TZ or BPH-PZ,there was no significant increase in AUC values.Pearson correlation analysis showed that there was a significant negative correlation between D and K values in all three tissues(PCa,BPH-TZ and BPH-PZ tissues)(r=-0.802,-0.796 and-0.776,all P<0.001),and a significant positive correlation was also showed between D and ADC values in all three tissues(r=0.774,0.674 and 0.624,all P<0.001).Conclusions:In summary,the post-processing quantitative parameters of DKI can identify PCa and BPH-TZ or BPH-PZ tissues.The K value of DKI has more advantages in recognizing PCa and BPH-PZ tissue than the ADC value of traditional DWI.But the advantages of identifying PCa and BPH-TZ tissues are not clear.The FA value obtained by the DKI scanning parameters(using 15 different directions of diffusion-sensitive gradient fields for each b value)has no advantage over the other quantitative parameters for the differential diagnosis of PCa.There is a significant negative correlation between the D value and the K value,and a significant positive correlation between the D value and the ADC value. |
PDF Full Text Request