Value of histogram analysis of quantitative parameters in dynamic contrast-enhanced magnetic resonance imaging in diagnosis of prostate cancer

Jiayi YU,Dajing GUO,Yindeng LUO,Xiaojing HE,Qian LING,Wei WU,Weijuan CHEN,Shanwei BAI
DOI: https://doi.org/10.16016/j.1000-5404.201708017
2017-01-01
Abstract:Objective To investigate the value of histogram analysis of the quantitative parameters obtained by dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) in the diagnosis of prostate cancer (PCa).Methods Thirty-eight patients with pathologically confirmed PCa and 18 patients with benign prostatic hyperplasia (BPH) were enrolled in this study along with 16 patients with prostatitis diagnosed by radiographic presentations,laboratory findings and follow-up observation.All the patients underwent conventional pelvic MRI,diffusion-weighted imaging (DWI) and DCE-MRI examinations between October 2016 and June 2017,and histogram analysis was performed using a third-party post-processing software (Omni Kinetic,OK).Regions of interest (ROIs) were drawn on the largest central level of the lesions,and the software automatically analyzed the mean,5%,10%,25%,50%,75%,90% and 95% percentiles of DCE-MRI quantitative parameters (Ktrans,Kep,and Ve) in the ROIs.Each of the parameters was compared between PCa and prostatitis in the peripheral zone and between PCa and BPH in the central gland zone to analyze their respective diagnostic efficacy.Results In the peripheral zone,the mean and percentiles (50%,75%,90% and 95%) of Ktrans and Kep in PCa were significantly higher than those in prostatitis (P <0.05);the 5%,10%,25% and 50% percentiles of Ve in PCa were significantly smaller than those in prostatitis (P < 0.05).In the central gland zone,the mean and percentiles of Ktrans,Kep and Ve all showed no significant differences between PCa and BPH.In PCa in the peripheral zone,the area under the curves (AUC) of high percentiles of Ktrans and Kep were greater those of the means,with the AUCs of the low percentiles being the smallest.The diagnostic efficacy of 90% Kep was the highest for PCa in the peripheral zone with an AUC of 0.932,a sensitivity of 78.3%,and a specificity of 92.9%.The percentiles of Ve did not show any diagnostic efficacy for PCa in the peripheral zone.Conclusion Histogram analysis of the quantitative parameters of DCE-MRI has a clinical value in the differential diagnosis between peripheral PCa and prostatitis.The high percentiles of Ktrans and Kep have better diagnostic efficiencies than the mean and low percentiles of Ktrans and Kep.
What problem does this paper attempt to address?