Various organizations propose different ages at which to initiate and terminate prostate cancer screening. The European Association of Urology (EAU) recommends earlier and more frequent PSA testing for men with risk factors including positive family history, African American race, or a baseline PSA level >1 ng/ml in the 40s [13]. The American Urological Association (AUA) recommends offering routine screening from ages 55 to 69 if normal-risk or younger than age 55 if higher-risk [14]. Similar to the EAU, the AUA does not recommend routine screening in any man with less than a 10–15 year life expectancy.

Evidence for these recommendations comes primarily from the ERSPC and Göteborg trials. The core age group in which the ERSPC demonstrated a survival benefit to PSA screening was ages 55–69 [3]. The Göteborg trial suggests that this benefit extends to men ages 50–54 [15, 16].

Although men <50 years were not included in the major randomized trials, baseline PSA testing in younger men is strongly supported by several observational studies [17–20]. In men younger than 50 years, a baseline PSA above the age-specific median but still within normal range is a robust predictor of future prostate cancer risk; it is a stronger predictor than family history, race, or suspicious digital rectal examination (DRE) [18, 20]. Baseline PSA testing also predicts long-term risk of locally advanced disease, metastases, and disease-specific mortality [20–22]. Thus, it has been suggested that a baseline PSA can be used for risk stratification and to determine frequency of subsequent prostate cancer screening for an individual.

This evidence has been incorporated into select professional guidelines but is still lacking from the guidelines released by the AUA and others that are based more strictly on randomized trials. The EAU suggests baseline PSA testing in men as young as 40 years of age [13]. The National Comprehensive Cancer Network (NCCN) similarly recommends offering baseline PSA testing at age 45 [23].

The optimal screening interval would minimize the overdiagnosis of low-risk disease while still allowing detection of life-threatening tumors at a curable stage. Data from Rotterdam (4 year interval) and Göteborg (2 year interval) ERSPC were used to compare the implications of different screening intervals. Compared to screening every 4 years, a 2-year interval reduced diagnosis of advanced prostate cancer by 43 % but increased diagnosis of low-risk disease by 46 %.

The AUA endorses screening no more often than every 2 years to reduce harms. [14] An alternate approach is to determine the screening frequency by a patient’s individual risk. The Rotterdam section of the ERSPC showed that men with a PSA <1 ng/mL had a very low risk for prostate cancer after 4 and 8 years (0.23 % and 0.49 %, respectively) [24]. Considering both this data and the value of baseline PSA testing, the EAU and NCCN recommend risk-based screening intervals [13, 23]. The EAU suggests follow-up every 2 years if the PSA is >1 ng/mL at age 40 or >2 ng/mL at age 60. Follow-up for men with PSA values below these thresholds can be postponed up to 8 years.

The indication for prostate biopsy in the ERSPC was primarily a PSA cutoff of 3 ng/mL. As this study demonstrated a survival benefit, a PSA threshold of 3 ng/mL can be considered evidence-based [3]. However, such a cutoff may diagnose many low-risk cancers and overlook some higher-risk cancers. PSA predicts risk of cancer in a continuous fashion, with no specific cutoff that can exclude clinically significant cancer [25]. Analysis of data from empiric biopsies in the Prostate Cancer Prevention Trial (PCPT) demonstrated that risk of Gleason ≥7 cancer in men with PSA levels between 2.1 and 3.0 ng/mL and 3.1–4.0 ng/mL was 4.6 % and 6.7 %, respectively.

Age-specific ranges have been suggested as a tool to increase the accuracy of PSA [11]. However, studies have shown mixed results regarding the value of these reference ranges [26]. As such, they are not included in recent guidelines from the EAU and NCCN. However, the AUA suggests considering a higher threshold of 10 ng/ml for biopsy in men over age 70 to reduce harms.

For men with mildly elevated PSA values, many guidelines now suggest further risk stratification with secondary tools such as free PSA, the prostate health index (phi), 4 kallikrein score (4Kscore), prostate cancer gene 3 (PCA3), MRI, and risk calculators. The EAU specifically recommends application of these tools in men with a PSA between 2 and 10 ng/mL to determine need for prostate biopsy [13]. The NCCN offers free PSA, phi and the 4 K score as secondary testing options for men with PSA >3 ng/ml considering initial prostate biopsy. The same three tests as well as PCA3, ConfirmMDx and multiparametric MRI are suggested as potential secondary testing options for men considering repeat biopsy.

The majority of total PSA (tPSA) circulates bound to proteins, with the remainder existing as free PSA (fPSA). Men with prostate cancer have a lower fraction of fPSA, or %fPSA [27]. Many prospective studies have indicated improved specificity and accuracy for detection of prostate cancer and clinically significant cancer (i.e., Gleason ≥7, Epstein criteria) using %fPSA in men with tPSA between 4 and 10 ng/mL [28]. Evidence suggests that its utility even extends to men with tPSA <4 ng/mL [32].

Indications for %fPSA include men with mildly elevated PSA for whom initial biopsy or repeat biopsy is under consideration [33]. As with total PSA, the optimal %fPSA cutoff for clinical use is subject to debate. One study demonstrated that a %fPSA cutoff of 25 % in men with PSA between 4 and 10 ng/mL detected 95 % of cancers while avoiding 20 % of unnecessary biopsies [29].

An isoform of fPSA, termed [−2]proPSA, is elevated in men with prostate cancer and appears more specific for cancer than tPSA [34]. The phi test incorporates tPSA, fPSA, and [-2]proPSA into a formula which predicts the probability of prostate cancer on biopsy. Compared to its individual components, phi has been shown to be more accurate for detection of prostate cancer and clinically significant cancer, defined by Gleason ≥7 or Epstein criteria [30, 31, 35–37]. Using a phi cutoff of 28.6, one study estimated that 30.1 % of unnecessary biopsies could have been avoided in men with PSA values between 4 and 10 ng/mL [30].

By reducing unnecessary biopsies, phi in conjunction with tPSA may be more cost-effective overall [38]. These studies support the application of phi to help patients make more informed decisions about initial or repeat biopsy.

The 4Kscore is an algorithm containing tPSA, fPSA, intact PSA, and human kallikrein 2 along with age, DRE, and prior biopsy results to predict the risk of biopsy-detectable high-grade prostate cancer. Several studies have demonstrated that 4Kscore has high accuracy for detection of clinically significant cancer, comparable to that of phi [39–41]. It has also been shown to predict prostatectomy pathology and metastatic disease [42]. Both phi and 4Kscore are suggested by the EAU and NCCN as reflex testing options for prostate biopsy decisions [13, 23].