Construction Of A Predictive Model For Clinically Significant Prostate Cancer And Comparison With ¹⁸F-PSMA PET/CT:Avoiding Unnecessary Prostate Biopsies

Background:Prostate cancer（PCa）is a common malignancy in men with high morbidity and mortality.Prostate biopsy remains the standard method for the diagnosis of PCa.As an invasive procedure,complications of biopsy such as pain,bleeding,lower urinary tract symptoms,and infection are frequent.There is a growing trend to reduce unnecessary prostate biopsies.The increasing popularity,different functional imaging modalities,and increased standardization of prostate MRI have increased its role in the detection,localization,and staging of prostate cancer.PI-RADS score was an independent predictor of total prostate cancer and clinically significant prostate cancer and performed well as a predictor of biopsy outcome.Besides,Multiple studies have pointed out that PSAD is one of the important predictors of cs PCa.Therefore,the combination of MRI with other clinical parameters and biomarkers such as PSAD may be considered to improve the diagnostic performance of cs PCa and guide the decision on whether to proceed with prostate biopsy.Objective:The study was to construct a clinically significant prostate cancer prediction model combining mp MRI,biochemical and clinical parameters,to conduct external verification and compare its diagnostic efficacy with PET/CT,so as to provide clinicians with a higher basis for diagnosis and treatment decisions,thus reducing unnecessary prostate biopsies.Methods:The clinical data of 847 patients who underwent prostate biopsy at the First Hospital of Shanxi Medical University from January 2018 to June 2021 were retrospectively analyzed.Univariate and multivariate analyses were performed to identify predictors of clinically significant prostate cancer.The prediction model was established by binary Logistic regression and the line chart was drawn.The diagnostic performance of each predictor and model was compared by receiver operating characteristic（ROC）curve,and the optimal cut-off value of the model was judged by You-den index.The comparison cohort included 72 patients from institution 2 who underwent¹⁸F-PSMA PET/CT between May 2019 and January 2022.The probability of disease for each person was calculated by the model and compared with ¹⁸F-PSMA PET/CT.The diagnostic performances were compared via sensitivity,specificity,positive predictive value（PPV）,negative predictive value（NPV）and area under receiver operating characteristic curve（AUC）.Their net benefit and reduction in the proportion of unnecessary prostate biopsies were determined by decision curve analysis.Results:Univariate analysis showed that age,total prostate-specific antigen（TPSA）,free ratio total prostate-specific antigen（F/T）,prostate volume（PV）,prostate-specific antigen density（PSAD）,digital rectal examination（DRE）and PI-Rad SV2.1 score were significantly correlated with clinical prostate cancer（P<0.05）.After multi-factor analysis,age,prostate-specific antigen density（PSAD）,and PI-RADS v2.1 score were included as predictors of the model.The area under the receiver operating characteristic curve（AUC）for diagnosing cs PCa about the model was 0.938（95%CI:0.922-0.955）.Under the optimal cut-off value,and the sensitivity and specificity were 0.925 and 0.825.In cohort 2,sensitivity,specificity,PPV,NPV of the model were 91.84%、78.26%、90.00%and81.81%;¹⁸F-PSMA PET/CT were 97.96%、52.17%、81.36%and 92.31%.The model had a significantly higher AUC compared with ¹⁸F-PSMA PET/CT:0.936（95%CI 0.879-0.993）versus 0.751（95%CI 0.613-0.888）.Decision curve analysis showed that the clear net benefit of the model was higher than ¹⁸F-PSMA PET/CT at lower risk thresholds,resulting in a significant reduction in unnecessary prostate cancer.Conclusion:This study developed a model with age,PSAD and PI-RADS v2.1 score.The model may have better diagnostic performance than 18F-PSMA PET/CT for predicting cs PCa and could distinctly reduce unnecessary prostate biopsies.