| Objective:To explore the diagnostic value of PI-RADS v2.1 combined with PSAD,(f/t)/PSAD,PSAMR and other PSA derived parameters and r ADC for gray zone prostate cancer,and to find the best combined diagnostic model,provide new ideas for accurate preoperative diagnosis of gray zone prostate cancer.Methods: The relevant clinical and MRI image data of patients who underwent prostate mp-MRI for the first time and had PSA values between 4 and 10 ng/m L from January 2018 to June 2021 in our hospital were retrospectively analyzed.The images were scored according to the PI-RADS v2.1 scoring criteria;PSA derived parameters(PSAD,(f/t)/PSAD,and PSAMR)were calculated according to the formula;ADC values were obtained by delineating the region of interest along the lesion boundary at the largest level of the lesion,measuring the ADC values of the obturator internus muscle,and calculating r ADC values.According to the pathological results,the patients were divided into non-prostate cancer group and prostate cancer group.Statistical analysis was performed using SPSS and Med Calc software.Univariate analysis was used to correlate the indicators with statistically significant differences between groups and ROC curves were plotted;binary logistic regression was used to construct a combined prediction model of PI-RADS v2.1 with each PSA derived parameter and r ADC value,respectively,and ROC curves were plotted.AUC values,sensitivity and specificity were calculated for each combined predictive model.The area under the curve of the ROC curve was compared with the Delong method.P <0.05 was considered statistically significant.Results:A total of 388 patients were included in the study,103 in the PCa group with a mean age of 68 ± 11 years and 285 in the non-PCa group with a mean age of 69± 12 years.The results of PI-RADS v2.1 score showed: score 1 in 23 patients(5.9%);score 2 in 84 patients(21.6%);score 3 in 193 patients(49.7%);score 4 in 61 patients(15.7);and score 5 in 27 patients(7.0%).(1)Univariate analysis showed that PSAD,(f/t)/PSAD,PSAMR,PI-RADS v2.1 and r ADC values were different between prostate cancer group and non-prostate cancer group(P < 0.05).Spearman correlation analysis showed that PSAD,PSAMR and PI-RADS v2.1 scores were significantly positively correlated with prostate cancer(r = 0.247,0.232,0.461);(f/t)/PSAD and r ADC were significantly negatively correlated with prostate cancer(r =-0.226,-0.311,-0.619);further ROC curve analysis showed that PSAD,(f/t)/PSAD,PSAMR,PI-RADS v2.1 and r ADC were 0.159,1.184,5.527,3.5 and 0.826.Among them,AUC values were the largest for r ADC values,followed by AUC values for PI-RADS v2.1(P < 0.05).(2)Four combined models of PI-RADS v2.1 combined with PSAD model,PI-RADS v2.1 combined(f/t)/PSAD model,PI-RADS v2.1 combined with PSAMR model and PI-RADS v2.1 combined with r ADC value were established.The results showed that the optimal cutoff values were 0.401,0.238,0.321 and 0.183,the Youden index was 0.499,0.461,0.485 and 0.663,the AUC values were 0.820,0.804,0.816 and 0.902,the sensitivities were 61.17%,65.05%,62.14% and 92.23%,and the specificities were 88.77%,81.05%,86.32% and 74.04%.The AUC values of the four combined models were larger than those of PI-RADS v2.1,and the differences were statistically significant,with the largest area under the AUC curve of the PI-RADS v2.1 combined with r ADC value model(P < 0.05);there was no significant difference in AUC values between the other three models.Conclusion:PI-RADS v2.1 combined with PSAD,(f/t)/PSAD,PSAMR and r ADC values can improve the diagnostic efficacy of PI-RADS v2.1 for gray zone prostate cancer.PI-RADS v2.1 combined with r ADC values has the best diagnostic efficacy. |
PDF Full Text Request