Clinical Study On Establishing A Predictive Model Of Clinically Significant Prostate Cancer Based On PI-RADS V2.1

Part 1 Diagnostic value of prostate imaging reporting and data system version 2.1 for clinically significant prostate cancerObjective:The objective of this study was to compare the interobserver agreement between PI-RADS v2.1and v2.0,as well as the diagnostic performance of the two versions for clinically significant prostate cancer(CsPCa)and prostate cancer(PCa).Materials and Methods:The MRI and clinical data of 481 patients who underwent prostate biopsy in our hospital from January 2018 to July 2021 were retrospectively collected.Two radiologists(radiologist 1 and radiologist 2)used PI-RADS v2.1 and v2.0 to score the transition zone and peripheral zone of the 204 randomly selected cases.The score of the index lesion was used as the total score.The agreement between the two radiologists’ scores was analyzed using the weighted Kappa test.The remaining patients were scored by radiologist 1 using both versions.The ROC curve of the index lesion score of the two versions were drawn to evaluate the diagnostic performance for CsPCa and PCa.The area under curve(AUC)was compared using the Delong test.Results:Among the 481 patients enrolled,there were 127 cases of CsPCa,33 cases of clinically insignificant prostate cancer(Cis PCa),and 321 cases of benign prostatic hyperplasia(BPH).The weighted Kappa value of the two radiologists using PI-RADS v2.1 and v2.0 between the transition zone,the peripheral zone,and the total score were 0.794、0.804、0.781 and 0.670、0.754、0.738.The AUC of PI-RADS v2.1 and v2.0 for the diagnosis of CsPCa were 0.896 and0.895,respectively,and there was no significant difference between the AUC(Z=0.160,P=0.873).The AUC for the diagnosis of PCa were 0.851 and 0.838,respectively,and there was no significant difference between the AUC(Z=1.365,P=0.172).Conclusion:PI-RADS v2.1 has improved interobserver agreement and has better diagnostic performance for CsPCa and PCa compared to v2.0.However,there is no significant difference in the diagnostic performance of CsPCa and PCa between PI-RADS v2.1 and v2.0.Part 2 A clinical study on development of PI-RADS v2.1 based model for clinically significant prostate cancerObjective:To explore the predictive value of PI-RADS v2.1 combined with clinical indicators for CsPCa.Establish a prediction model and evaluate its diagnostic performance and clinical benefit by external validation.Materials and Methods:Retrospectively collected 481 patients who underwent MRI and prostate biopsy in our hospital from January 2018 to July 2021.The training cohort(360 cases)was composed of patients from January 2018 to December 2020,and the validation cohort(121 cases)was composed of patients from 2021.The data of PI-RADS v2.1 score(from radiologist 1 in part 1)and relevant clinical indicators of the patients were collected and used for univariate analysis and multivariate Logistic regression analysis.Prediction model was established by combining independent risk factors.The ROC curve,calibration curve and decision curve were used to evaluate the model’s diagnostic performance,degree of calibration and net benefit.External validation was performed to evaluate the reliability of the model.Results:There were 102 cases of CsPCa in the training cohort,and 25 cases of CsPCa in the validation cohort.Univariate analysis of the training cohort showed that there were significantly different in PI-RADS v2.1,age,PSA,PV and PSAD between the CsPCa group and the non-CsPCa group(P<0.01).Multivariate Logistic regression analysis showed that PI-RADS v2.1 score and PSAD were independent risk factors for CsPCa.The AUC of the combined prediction model was 0.909,which was higher than the independent risk factors above.The calibration curves showed that the model was well calibrated in both the training cohort and the validation cohort.The decision curves showed that the model could achieve a net benefit when the risk threshold was greater than 0.05 and 0.02 in the training cohort and validation cohort,respectively.Conclusion:The combined prediction model of CsPCa based on PI-RADS v2.1 and PSAD had good diagnostic performance.External validation showed that the model had good clinical benefits and certain clinical application value.