Application Of Magnetic Resonance Imaging In The Detection And Diagnosis Of Prostate Cancer

Purpose To explore the application of biparametric magnetic resonance imaging(bpMRI)in the detection of prostate cancer(PCa)based on whole-mount histopathology.Methods Retrospective analysis was performed on 67 patients with clinically suspected prostate cancer who received preoperative routine MRI examination and underwent radical resection of prostate cancer followed by whole mount section analysis in our hospital from May 2017 to September 2019.The diameter,location and Gleason score of each lesion on whole mount section were recorded in detail.Matching analysis was conducted between bpMRI including T2WI,DWI and ADC,and whole mount section images.When the location(quadrant and plane)and shape size of MR lesions and whole mount section lesions were similar,they were considered to be matched.SPSS 22 software was used for statistical analysis.The non-normal distribution data is represented by the median(quarter spacing),the normal distribution data is represented by the mean±standard deviation,and the count variable is represented by the frequency(percentage).The independent sample t-test was used to compare whether there were statistical differences in the maximum diameter of lesions and Gleason score between the detected group and the missed group.Chi-square test and Fisher test were used to analyze whether there was statistical difference in the detection rate of lesions with different sizes and Gleason scores.P value less than 0.05 was considered statistically significant.Results For the 67 patients included,123 lesions were confirmed by whole mount section,and 94 lesions were matched between imaging and whole mount section.There were 5 false positive lesions(5%)and 29 false negative lesions(23.6%).There were 67 index lesions with a detection rate of 98.5%and 56 non-index lesions with a detection rate of 50%.There were 113 clinically significant lesions with a detection rate of 74.3%,and 10 clinically insignificant lesions with a detection rate of 30%.The detection rate was 92.1%and 84.3%for lesions with diameters larger than 1cm and which with Gleason score>3+3,respectively.Conclusion bpMRI is highly sensitive to the detection of prostate cancer,especially for foci larger than 1cm in diameter or Gleason score>6.Purpose To explore the diagnostic performance of intravoxel incoherent motion diffusion weighted imaging(IVIM-DWI)in the diagnosis and risk stratification of prostate cancer(PCa)based on whole-mount histopathology.Materials and Methods A total of 28 patients with clinically suspected prostate cancer who received preoperative prostate magnetic resonance IVIM-DWI examination in our hospital from May 2018 to December 2019 were collected,and underwent radical resection of prostate cancer followed by whole mount section analysis.According to the clinical risk stratification guidelines for localized prostate cancer,the risk of prostate cancer was classified,including 18 cases in the medium risk group,and 10 cases in the high risk group.All patients underwent a 3.0T MRI scan(Signa Pioneer,GE Medical Systems,USA),including T2WI FS,DWI,and IVIM-DWI.The MR Body Diffusion Toolbox software of Siemens Syngo.via Frontier post-processing workstation was used for post-processing.ROIs of each patient were placed,including prostate cancer(PCa-ROI),noncancerous areas of the transitional zone(tz-ROI)and noncancerous areas of the peripheral zone(pz-ROI).The following parameters were calculated:apparent diffusion coefficient(ADC),signal reference(S0),real molecular diffusion coefficient(D),perfusion-related diffusion coefficient(D*),perfusion fraction(f value),and calculated DWI signal intensities value(C)on b value of 2 000 s/mm2.IBM SPSS Statistics 22 software and MedCalc 19.3 software were used to analyze the data.The independent sample t-test was used to analyze whether there was statistical difference among the parameters between different risk stratifications.Receiver operating characteristic(ROC)curve was used to evaluate the diagnostic efficacy of each parameter for prostate cancer.Results There were statistically significant differences in S0,D,and C values between prostate cancer and non-cancer zone(P<0.001).The area under the ROC curve of D value was the largest among the related parameters of IVIM-DWI,which was 0.979(95%CI 0.917-0.998),and was not statistically significant defferent compared with the area under the ROC curve of ADC value(0.987(95%CI 0.930-1.000)(P=0.304).Gleason score was poorly correlated with IVIM-DWI parameters(P>0.05).The ADC and D value of the intermediate risk group were higher than those of the high-risk group(P=0.024 and 0.031,respectively).The C value of the intermediate risk group was lower than high risk group(P=0.008).Conclusion IVIM-DWI imaging is useful for the diagnosis of prostate cancer and help distinguish between intermediate and high risk prostate cancer groups.Objective To explore the role of MRI-based radiomics in prediction of Gleason score in prostate cancer,especially in predicting low-grade prostate cancer with Gleason score=3+3 or intermediate/high-grade prostate cancer with a Gleason score?3+4.Materials and Methods Retrospective analysis was conducted on the patients who received preoperative routine MRI examination and underwent radical resection of prostate cancer followed by pathological section during May 2017 to September 2019 with clinically suspected prostate cancer in our hospital.MRI images including T2WI FS,DWI,ADC,T2WI,and T1WI were matched with pathological section images for analysis.Identify image-pathological matching lesions.The MRI was transferred to the Siemens Syngo.via Frontier post-processing workstation for post-processing using Radiomics software.Each image-pathology matching lesion on the MRI was delineated and segmented layer by layer,and the radiomics features were extracted from the ROI.The maximum correlation and minimum redundancy(MRMR)method was used to screen the image omics features of the training set.Univariate analysis and multivariate analysis(linear regression and logistic regression)were used to screen out the image omics features and construct the diagnostic model.Verify the performance of the diagnostic model in both the training set and the validation set.The selected indicators included area under receiver operating characteristic curve(ROC_AUC),sensitivity,specificity,positive predictive value and negative predictive value.Results A total of 66 cases and 190 layers of lesions were included in the database.The training set included 52 cases with 151 layers of lesions,and the validation set included 14 cases with 39 layers of lesions.There were no statistically significant differences in age,PSA level and Gleason score between the training set and validation set.The diagnostic efficacy(ROC_AUC)of radiomics models based on T2WI FS,ADC,DWI,T2WI and T1 WI for distinguishing Gleason scores of 3+3 and ?3+4 in the training set were 0.958,0.654,0.604,0.698 and 0.517,respectively.The diagnostic efficacy(ROC_AUC)of Gleason score 3+3 and ?3+4 for validation set were 1,0.698,0.526,0.56,0.804,respectively.Conclusion This study establishes and verifies that the radiomics model based on MRI,especially based on T2WI FS,can predict low-grade prostate cancer and intermediate/high-grade prostate cancer preoperatively and assist in clinical diagnosis and treatment.Prostate cancer(PCa)is increasingly prevalent in China.Multiparametric magnetic resonance imaging(mpMRI),as a non-invasive imaging method,has become an important tool for prostate cancer detection,diagnosis and treatment selection,and patient management during PCa patient follow-up.Based on T2WI,DWI and other functional MRI images,radiomics can be the bridge between physiological activity and imaging.In this review,we will briefly introduce the role of different mpMRI sequences in the detection and diagnosis of prostate cancer and the basic process of radiomics.We will focus on the recent application and exploration of mpMRI-based radiomics in prostate cancer.