The Effect of Different Levels of Prostate-Specific Antigen and Prostate Imaging Report and Data System Scores on Logistic Regression Model to Predict Clinically Significant Prostate Cancer—A Single Center Retrospective Study

Background: At present, there are various clinical regression models for predicting prostate cancer. But what about the diagnostic effectiveness of these models in different parameter ranges, and are the models applicable to everyone? This study aimed to study the influence of different levels of prostate-specific antigen (PSA) and Prostate Imaging Report and Data System version 2 (PI-RADS v2) scores on the regression model to predict clinically significant prostate cancer (csPCa).Methods: This retrospective study screened 251 patients from our hospital, who were divided into different groups. The regression model was established for each group to predict csPCa, and the effects of PSA and PI-RADS scores on each model were analyzed through the diagnostic effects of the model.Results: In patients with lower PSA scores, although the model was less sensitive than PSA, the AUC of the model was much greater. With the rise of PSA, the sensitivity of the model surpassed that of PSA, while the specificity became the opposite, and the AUC gap also gradually decreased. In the group with low PI-RADS score, the sensitivity and specificity of PI-RADS were lower than the model, and the gap was larger. Although the gap between the two gradually decreased with the increase of PI-RADS, the diagnostic efficiency of the model was still slightly larger than that of pure PI-RADS.Conclusion: As the PSA and PI-RADS v2 scores increase, the diagnostic advantages of the regression model will gradually decrease. However, for patients with low levels of PSA and PI-RADS scores,the regression model is less affected by PSA and PI-RADS, and can better utilize its clinical diagnostic advantages.