The Application Value And Accuracy Of Prostate Imaging Reporting And Data System

Research purposes:The purpose of this study is to evaluate the false positive rate of 4,5 points lesions in the prostate imaging report and data system.To explore the consistency of the scoring results of PI-RADS v2.0 and PI-RADS v2.1 in diagnosis by imaging doctors with different experiences.The comparison is based on PI-RADS v2.1,bp MRI and mp MRI for the detection of clinically significant prostate cancer between observers.Research methods: Retrospective analysis of 226 patients undergoing prostate MRI examination in our hospital from December 2017 to July 2020.All lesions underwent MRI ultrasound fusion targeted biopsy within 2months after MRI examination.Two radiologists retrospectively Locally use PI-RADS v2.0 and PI-RADS v2.1 for identification and evaluation.In PI-RADS v2.1,there were 168 lesions with 4 and 5 points,among which the incidence of clinically significant prostate cancer(defined as Gleason score ≥ 7)was determined.The clinical data first compares categorical variables through t-test,and compares continuous variables through Fisher’s exact test.Multivariate logistic regression model is used to determine factors related to benign pathological results.The Kappa test was used to compare the consistency of the PI-RADS v2.0 and v2.1 scores of two different experienced imaging physicians.And based on PI-RADS v2.1,bp MRI and mp MRI are used to detect clinically significant prostate cancer.Result : Among the 36 PI-RADS 4-point lesions assessed by multi-parameter magnetic resonance imaging cognitive fusion biopsy of the prostate,80.55%(29/36)were prostate cancer,19.45%(7/36)were benign lesions;132 PI-RADS Among the 5-point lesions,93.94%(124/132)were prostate cancer,and 6.06%(8/132)were benign lesions.PI-RADS 4-point lesions showed 61.11%(22/36)as clinically significant prostate cancer;PI-RADS 5-point lesions showed 87.88%(116/132)as clinically significant prostate cancer.In addition,there are no clinically meaningful prostate cancers elsewhere in the 15 benign lesions.In univariate analysis,compared with PI-RADS 4 and 5 lesions with malignant pathological results,PI-RADS 4 and 5 lesions with benign pathological results are significantly related to lower prostate specific antigen(PSA)density,OR = 0.017,P = 0.002.In the multivariate analysis,the factor associated with the benign characteristics was the lower PSA density,OR = 0.016,P = 0.005,and 95% CI [0.453,3.794].The 15 lesions with benign pathological results were re-examined again.53.33%(8/15)of the pathology was stromal benign prostatic hyperplasia,and 46.67%(7/15)of the pathology was inflammation.In PI-RADS v2.0,the level of agreement between two readers with different experiences was moderate(k value 0.51,95% confidence interval 0.43-0.60).In PI-RADS v2.1,the k value of two readers with different experience(k value 0.54,95% confidence interval 0.46-0.63)is slightly higher than PI-RADS v2.0,but the consistency level is also It is medium level.Based on PI-RADS v2.1,the consistency among different readers in bp MRI and mp MRI in all the enrolled cases was moderate.Furthermore,in the PZ and TZ lesions,the consistency of bp MRI and mp MRI was moderate.Conclusion: In PI-RADS v2.1,the incidence of benign lesions with 4or 5 points is very low,and clinical parameters(PSA density)help clinical decision-making.In this study,the consistency of PI-RADS v2.0and PI-RADS v2.1 scores of different imaging physicians was moderate.Based on PI-RADS v2.1,bp MRI and mp MRI have similar levels of consistency among different readers.It is clear that PI-RADS v2.1 can help improve the accuracy of preoperative imaging diagnosis of prostate cancer and reduce the dependence on the experience of the diagnostician.