Exploring The Value Of Preoperative Mp-MRI In Assessing The Risk Stratification Of PCA And Differentiating The Benign And Malignant Lesions Of PI-RADS 3 In TZ

Objective: Preoperative multiparameter MRI(mp-MRI)scan was performed in patients with highly suspected prostate cancer(PCA),including diffusion kurtosis imaging(DKI),Introvoxel incoherent motion(IVIM),stretched exponential model and conventional MRI.The parameters of DKI,IVIM and stretched exponential model were analyzed quantitatively.The purpose of this study was to explore the value of preoperative multi parameters MRI in the assessment of prostate cancer risk stratification and differentiation the benign and malignant lesions of PI-RADS 3 in Transition Zone.Materials and Methods: From December 2018 to January 2020,clinically highly suspicious prostate cancer patients in the Second Hospital of Dalian Medical University were collected.Preoperative imaging(including routine MRI,DKI,IVIM and stretched exponential model)were collected by 3.0-T MRI scanner(Discovery 750W;GE Medical Systems).According to the Version 2.1 of PI-RADS(2019,ACR),the image of patients were analyzed and evaluated,and then,lesions of PI-RADS 3,4 and 5 were selected as the target lesions.Outline of target lesions were delineated on platform of Magnetic resonance imaging-Transrectal ultrasound(MRI-TRUS,MIM,America),preoperatively.During operation,make sure that the real-time ultrasound imaging matched with MR imaging before getting the puncture tissue.Hematoxylin eosin staining and basal cell labeling were used to evaluate the Gleason score.The value of related parameters were obtained in the workstation(Advantage Workstation 4.6;GE Medical Systems),such as,DKI related parameters: mean kurtosis(MK),axial kurtosis(Ka),radial kurtosis(Kr),mean diffusivity(MD),axial diffusivity(Da),radial diffusivity(Dr),IVIM related parameters: standard apparent diffusion coefficient(standard ADC),pure molecular diffusion(D),pseudo diffusion coefficient(D*),perfusion fraction(f),distributed diffusion coefficient(DDC)and water molecular diffusion heterogeneity index(?)value of stretched exponential model.SPSS software(version 25.0,IBM,Chicago,IL)was used for statistical analysis.Spearman correlation analysis was used to analyze the correlation between parameters of DKI,IVIM and stretched exponential model and PCA risk stratification.One-Way ANOVA was used to compare the differences of relevant parameters among risk stratification of PCA.For parameters with homogeneous variance,Tukey HSD was used for inter group comparison.For parameters with uneven variance,games Howell was used for inter group comparison,and then Bonferroni correction was carried out for the data.The differences of PI-RADS 3 between benign and malignant groups were compared by T-test and Mann Whitney test.Then,the Receiver Operating Characteristic(ROC)curve was used to analyze the diagnostic efficacy of related parameters on PCA in PI-RADS 3 lesions.Results: Results of part 1 1.A total of 118 patients with PCA were enrolled,including 23 in the low-risk group,37 in the intermediate-risk group and 58 in the high-risk group.2.The correlation analysis between the parameters of FA,MK,Ka,Kr,MD,Da,Dr of DKI,standard ADC,D and D* of IVIM,DDC of stretched exponential model and PCA risk stratification were statistically significant.The values of FA,MK,Ka,Kr,D* were positively correlated with the risk stratification of PCA(rs = 0.675;rs = 0.656;rs = 0.691;rs = 0.719;rs = 0.200),and the values of MD,Da,Dr,standard ADC,D and DDC were negatively correlated with the risk stratification of PCA(rs =-0.777;rs =-0.827;rs =-0.698;rs =-0.753;rs =-0.700;rs =-0.685).3.The parameters of FA,MK,Ka,Kr,MD,Da,Dr of DKI,standard ADC and D of IVIM,DDC of stretched exponential model can effectively differentiate PCA of low-risk group,PCA of intermediate-risk group and PCA of high-risk group.The parameters of FA,MK,Ka,Kr,MD,Da,Dr,standard ADC,D and DDC value of low-risk group were 0.253±0.0548;1.077±0.0870;1.032±0.0869;1.027±0.0705;1.041 ± 0.0443(× 10-3mm2/s);1.360 ± 0.0492(× 10-3mm2/s);1.022 ± 0.0616(×10-3mm2/s);0.756±0.0611(×10-3mm2/s);0.654±0.0440(×10-3mm2/s)? 0.833±0.0391(×10-3mm2/s),respectively,intermediate-risk group were 0.345±0.0545;1.211±0.0994;1.230±0.0951;1.116±0.0791;0.887±0.0645(×10-3mm2/s);1.224±0.0903(×10-3mm2/s);0.845±0.0839(×10-3mm2/s);0.650±0.0736(×10-3mm2/s);0.555 ± 0.0578(× 10-3mm2/s);0.761 ± 0.0394(× 10-3mm2/s),respectively and high-risk group were 0.416±0.0774;1.301±0.0981;1.344±0.1394;1.263±0.1201;0.751 ± 0.1190(× 10-3mm2/s);0.979 ± 0.1425(× 10-3mm2/s);0.694 ± 0.1162(×10-3mm2/s);0.466±0.1308(×10-3mm2/s);0.387±0.1687(×10-3mm2/s);0.524±0.2081(×10-3mm2/s).Results of part 2 1.129 lesions of PI-RADS 3 were included in part 2,106 of which were benign prostatic nodules and 23 were PCA.Among them,13 PCA nodules were Gleason 3+3,6 PCA nodules were Gleason 3+4,and 4 PCA nodules were Gleason 4+3.2.The parameters of FA,MK,Ka,Kr,MD,Da,Dr of DKI,standard ADC and D of IVIM,DDC of stretched exponential model were statistically different between benign and malignant groups in PI-RADS 3 lesions.The parameters of FA?MK?Ka and Kr of benign group(0.1962±0.0477;0.8963±0.1567;0.8421±0.1793;0.8902±0.1435)were lower than PCA group(0.2560±0.0441;1.0701±0.1058;1.0672±0.1112;1.0800±0.1435)in PI-RADS 3 lesions.The parameters of MD?Da?Dr?Standard ADC?D and DDC of benign group(1.2024±0.2057(×10-3mm2/s);1.5621±0.1593(×10-3mm2/s);1.2489±0.1070(×10-3mm2/s);0.8848±0.1116(×10-3mm2/s);0.8374±0.1470(×10-3mm2/s);0.9766±0.1661(×10-3mm2/s))were lower than PCA group(1.0217±0.0982(×10-3mm2/s);1.3657±0.0614(×10-3mm2/s);1.0410±0.0630(×10-3mm2/s);0.7377±0.1036(×10-3mm2/s);0.6440±0.1004(×10-3mm2/s);0.8085±0.1122(×10-3mm2/s))in PI-RADS 3 lesions.3.Among the parameters related to DKI,IVIM and stretched exponential model,Dr had the highest diagnostic efficacy(AUC 0.930,sensitivity 95.7%,specificity 83.0%,CUT-OFF 1.137 × 10-3mm2 / s)for PCA in PI-RADS 3 lesions.Conclusion: 1.The parameters of preoperative mp-MRI(IVIM,stretched exponential model and DKI)have significant correlation with PCA risk stratification,and can effectively differentiate PCA risk stratification groups.2.The parameters of preoperative mp-MRI(IVIM,stretched exponential model and DKI)can effectively distinguish the benign and malignant lesions in PI-RADS 3.