Psa Density Improves The Rate Of Prostate Cancer Detection In Chinese Men With A PSA Between2.5-10.0and10.1-20.0ng/ml: A Multicenter Study

Objective: Prostate specific antigen (PSA) density (PSAD) has been confirmedas a tool to improve the prostate cancer detection rate in PSA gray zone. But the resultfrom populations in industrialized countries may lead to sizable bias when applied toChinese men directly. So we used multicenter data to assess the efficiencies of PSADin the diagnosis of prostate cancer in Chinese men with a PSA between2.5-10.0and10.1-20.0ng/ml.Background: Prostate cancer is most common cancer and the second frequentcause of cancer-related death in western countries men. Since Wang MC et al. firstlyseparated and purified prostate specific antigen (PSA) from prostate tissue in1979,PSA has been widely used in clinic as a tumor marker for prostate cancer. However,whether PSA testing for early screening of prostate cancer could reduce the prostatecancer mortality has been controversial for years. Convincing evidence indicates thatan elevated PSA is not specific to prostate cancer. Many sorts of benign diseases oroperation may also cause the increase of serum PSA level, such as prostatitis, urinaryretention, benign prostatic hyperplasia, prostate massage, digital rectal examination and prostate biopsy. In order to improve the diagnostic efficiency of prostate cancer,especially in the PSA gray zone, researchers had found some derivates of PSA, suchas PSA velocity (PSAV), percent free PSA (f/t PSA) and transition zone PSA density(TZPSAD), could increase the positive rate of prostate biopsy in different degrees andreduce unnecessary biopsies. PSA density (PSAD) also proved to be effective toimprove the prostate cancer detection rate. However, there are significant differencesin the characteristics of prostate cancer between Chinese men and other races. From aseries of our previous studies, we found that:1.The distributions of baseline PSA andPSA velocity in young men without prostate cancer among Chinese,African-American and Caucasian-American men. Young Chinese men have a lowerbaseline PSA and higher PSA velocity as compare to young African-American andCaucasian-American men (The Prostate2012;72:173-180).2.Percent free PSA doesnot improve the effectiveness for prostate cancer detection in Chinese men with aPSA of2.5-10.0or10.1-20.0ng/ml (Medical Oncology2014;31:925).3.Usingtransition zone PSA density can improve the efficiency of PSA in diagnosis ofprostate cancer and decreases the unnecessary prostatic biopsy in men with a PSA ofboth4.0-10.0or10.1-20.0ng/ml in Chinese men (Urologic Oncology2013;31:744-748). The incidence of prostate cancer in our country has beenincreasing in recent years, but to date it is still much lower than in western countries.It is reasonable that PSA gray zone in Chinese men is higher than the traditional PSAgray zone. But reports as to PSAD can improve the diagnostic rate of prostate cancerare mostly based on populations in western countries, whether the result can beapplied directly to our population requires further validation. To investigate theusefulness of PSAD for detecting prostate cancer in Chinese men, we analyzed theefficiencies of PSAD in the diagnosis of prostate cancer in Chinese men with a PSAof2.5-10.0ng/ml (traditional PSA gray zone) and PSA high above this range (PSA10.1-20.0ng/ml)(we hypothesized it was Chinese PSA “gray zone”) using multicenterdata. Materials and Methods: A total of461men with a PSA of2.5-20.0ng/mlwho had undergone trans-rectal ultrasound guided systematic prostatic biopsy (≥10cores) for diagnosis of prostate cancer from Guangzhou First People’s Hospital(Guangzhou FPH) and Zhujiang Hospital were included in the study. All of these menhad a documentation of PSA and prostate volume (PV) measurements. The patientswere divided into two groups as prostate cancer and non-prostate cancer, and weredivided into two groups as PSA2.5-10.0and10.1-20.0ng/ml. Statistical analyses wereperformed using SPSS (version18.0, Chicago, IL). Receiver operating characteristic(ROC) curve was used to evaluate and compare the efficiency of PSA and PSAD inthe diagnosis of prostate cancer. Binary logistic regression analysis was used to verifywhether PSAD was an independent predictor of positive biopsy based on clinicalvariables.Results:For men with a PSA of2.5-20.0ng/ml, prostate cancer was diagnosed in87(26.3%) of the331men in the Guangzhou FPH cohort and31(23.8%) of the130menin the Zhujiang Hospital cohort (P=0.589). In men with a PSA of2.5-20.0ng/ml, thedifferences in PSA, PV and PSAD for patients with and without prostate cancer in theGuangzhou FPH and Zhujiang Hospital groups were not statistical significant (all Pvalues>0.05). Only age at time of biopsy for both cohorts was significant different(both P values <0.05). Prostate cancer detection rates increased with increasingPSAD in both cohorts at PSA levels of2.5-10.0,10.1-20.0and2.5-20.0ng/ml (all Pvalues <0.001).In men with a PSA of2.5-10.0,10.1-20.0, and2.5-20.0ng/ml, the areas under theROC curve (AUCs) for PSAD, when considered as a continuous variables, forpredicting prostate cancer were all higher than those for PSA for the Guangzhou FPHcohort, Zhujiang Hospital cohort and whole cohort.When considering all461men,118(25.6%) were diagnosed prostate cancer.When broken down by PSA level,45(19.9%) of the226men with a PSA of 2.5-10.0ng/ml and73(31.1%) of the235men with a PSA of10.1-20.0ng/ml werediagnosed with prostate cancer, respectively (P=0.006). For the whole cohort, theAUCs for PSA and PSAD for predicting prostate cancer in men with a PSA of2.5-10.0ng/ml were0.628and0.698, respectively. When the optimal PSA cut-off of8.2ng/ml was determined, an efficiency of39.0%found, with correspondingsensitivity and specificity55.6%and70.2%, respectively. For the optimal PSADcut-off of0.15ng/ml/ml, the efficiency was41.7%with a corresponding sensitivityand specificity of64.4%and64.6%, respectively. The AUCs for PSA and PSAD forpredicting prostate cancer in men with a PSA of10.1-20.0ng/ml were0.558and0.722,respectively. For men with a PSA in this rang, optimal PSA cut-off of13.6ng/ml wasfound, yielding a total efficiency of30.3%with a corresponding sensitivity andspecificity of63.0%and48.1%, respectively. An optimum PSAD cut-off of0.33ng/ml/ml was determined, with an efficiency of49.9%and a correspondingsensitivity and specificity of60.3%and82.7%, respectively. The AUCs for PSA andPSAD for predicting prostate cancer in men with a PSA of2.5-20.0ng/ml were0.616and0.720, respectively. When an optimal PSA cut-off of11.2ng/ml was chosen, thetotal efficiency was35.7%, with a corresponding sensitivity and specificity of55.9%and63.8%, respectively. For the ideal PSAD cut-off of0.22ng/ml/ml, the totalefficiency was46.8%, with a corresponding sensitivity and specificity were64.4%and72.6%, respectively.Had a PSA of11.2ng/ml been chosen as the biopsy cut-off, only55.9%(66/118)of the men with prostate cancer would have been diagnosed, with a positive biopsyrate of34.7%(66/190). However, if a PSAD of0.22ng/ml/ml had been chosen as thebiopsy cut-off,64.4%(76/118) of the men with prostate cancer would have beendiagnosed, with a positive rate of44.7%(76/170), saving at least20men from anunnecessary prostate biopsy.On univariate analysis, the risk of prostate cancer was associated with PSAD. Onmultivariate analysis, PSAD (as either a continuous or categorical variable) was foundto be an independent predictor of prostate cancer in men with a of PSA2.5-20.0ng/ml(all odds ratios>1, all P values≤0.01). Conclusion: PSA was not the ideal predictor of prostate cancer for men with aPSA of2.5-10.0and10.1-20.0ng/ml. Using PSAD may improve the efficiency ofPSA in diagnosing prostate cancer and decrease unnecessary prostatic biopsies inChinese men with a PSA of2.5-20.0ng/ml.