Application Of Nomogram Based On Multiparametric Magnetic Resonance Imaging In Prostate Cancer Prediction

Background and Objective Multiparametric magnetic resonance imaging（mp-MRI）is becoming a better choice than single imaging methods of assessing prostate cancer.The purpose of this study was to investigate the diagnostic value of various biological indicators for PCa,and to establish a nomogram model for predicting PCa,which provides a basis for clinical prediction of PCa.Materials and Methods A clear pathology of prostate biopsy or radical prostatectomy was used as a reference standard.89 cases of pathologically confirmed prostate cases in Maanshan People’s Hospital were retrospectively analyzed,including 48 benign lesions and 41 cases of PCa.Antigen（PSA）and conventional MRI,diffusion-weighted imaging（DWI）and dynamic enhanced magnetic resonance imaging（DCE-MRI）,observation of lesion-related MR images,determination of PSA,prostate specific antigen density（PSAD）,apparent diffusion coefficient（ADC）Slow component(ADC_slow),ADC fast component(ADC_fast),standard ADC(ADC_standard),perfusion fraction（f）,K^trans,K_ep,V_e,prostate imaging report,and data system version 2（PI-RADSv2）score.Statistical analysis was performed using R software.Measurement data were expressed as mean±standard deviation（x±s）,and single factor comparison was performed by t test.The diagnostic threshold was determined to use a receiver operator characteristic curve（ROC）.The diagnostic value of the above biological indicators for PCa was analyzed by using two-class logistic regression（acceptance standard 0.05,exclusion criterion 0.10）,and a predictive PCa nomogram model was established.Results The PCa group was compared with the benign prostatic lesion group,with variables PSA（t=-5.29,P<0.001）,PSAD（t=-7.33,P<0.001）,ADC_slow（t=7.02,P<0.001）,ADC_standard（t=9.04,P<0.001）,K^trans（t=-10.34,P<0.001）,K_ep（t=-7.92,P<0.001）,V_e（t=-2.99,P=0.004）,and PI-RADS v2（x²=54.09,P<0.001）The difference was statistically significant.There was no significant difference between ADC_fast and f groups（P>0.05）.All of the above parameters were used as the field under the receiver operating characteristic curve（ROCAUC）:0.788,0.894,0.863,0.621,0.907,0.418,0.938,0.874,0.706,0.885;PSA,PSAD,ADC_slow,ADC_fast,ADC_standard,K^trans,K_ep,V_e,PI-RADS v2 sensitivity and specificity were 73.17%,81.25%,85.37%,79.17%,75.61%,93.75%,63.41%,66.67%,80.49%,93.75%,95.12%,83.33%,78.05%,83.33%,82.93%,56.25%,100%,77.08%.Multivariate logistic regression analysis:K^trans（OR=999.99,95%CI:11.163-999.99,P=0.015）,PI-RADS v2 score（OR=7.69,95%CI:1.719-34.398,P=0.008）for PCa The independent predictor,the concordance index（C-index）was 0.978.Conclusion:1.PSAD,ADC_slow,ADC_standard,K^trans,K_ep,V_e,PI-RADS v2 have diagnostic significance for PCa,of which Ktrans has the greatest diagnostic value;.2.K^transrans and PI-RADS v2 are independent factors for predicting PCa.The nomogram model provides a better and more accurate method to predict PCa.