Diffusion kurtosis imaging can be predictive in high-aggressive prostate cancer / 实用放射学杂志

ABSTRACT

Objective To investigate whether diffusion kurtosis imaging(DKI)can be predictive in high-aggressive prostate cancer (PCa).Methods 51 patients with pathologically confirmed PCa underwent preoperative DK-MR imaging(b of 0,700,1 400 and 2 100 s/mm2).Data was post-processed by mono-exponential and non-Gauss DK model,respectively,for quantitation of apparent diffusion coefficient(ADC),apparent non-Gaussian diffusion coefficient(Dapp)and apparent non-Gauss kurtosis coefficient(Kapp). The ability of Dapp,Kapp and ADC for predicting high-aggressive(Gleason score>4+3)PCa was analyzed by ROC regression.Results There were 29 low-aggressive and 33 high-aggressive PCa on pathologic findings.High-aggressive PCa had significantly lower ADC[(0.764 ± 0.114)×10-3mm2/s vs(0.825 ± 0.116)×10-3mm2/s,P=0.004],lower Dapp[(1.212 ± 0.194)×10-3mm2/s vs (1.472 ± 0.297)×10-3mm2/s,P< 0.001],while higher Kapp(1.114 ± 0.177 vs 0.835 ± 0.192,P<0.001)than that for low-aggressive PCa.Dapp and Kapp had significantly higher sensitivity(Dapp:75.3%;Kapp:74.1%),specificity(Dapp:85.4%;Kapp:86.7%)and area under curve(AUC)(Dapp:0.889;Kapp:0.894)than that for ADC(64.1%;76.4%;0.738;P<0.01)in differentiating low-aggressive from high-aggressive PCa.Conclusion DKI can be a reliable way for predicting high-aggressive PCa.