Prostate-specific antigen (PSA), a kallikrein-related serine protease produced by prostatic epithelium, transformed early prostate cancer detection and monitoring after its clinical adoption in the late 1980s. Contemporary guidelines position PSA as an entry test to risk-adapted screening and diagnostic pathways, combined with digital rectal examination (DRE), risk calculators, and multiparametric MRI (mpMRI).1–3 Despite its utility, PSA lacks cancer specificity; elevations occur in benign prostatic hyperplasia (BPH), prostatitis, and following manipulations, limiting its standalone accuracy for predicting histopathological aggressiveness.4,5 Consequently, histopathological grading using the Gleason system, harmonized into the ISUP Grade Group (GG) 1–5 scheme, remains the cornerstone for risk classification and treatment decision-making.6,7

The relationship between pretreatment PSA and Gleason grading is biologically plausible—higher tumour volume, de-differentiation, and disruption of glandular architecture may increase PSA leakage into circulation—yet empirically the association is imperfect and heterogeneous across cohorts.8–10 Clinically significant prostate cancer (CSD; GG≥2) is the principal target of contemporary diagnostic pathways to avoid over-diagnosis and overtreatment of indolent GG1 disease.2,11 However, men with GG≥2 disease can present with PSA values within “gray zones” (e.g., 4–10 ng/mL), whereas inflammation and BPH can inflate PSA in the absence of cancer or high-grade pathology.12,13

Advances in pre-biopsy mpMRI and MRI-targeted biopsies have improved detection of clinically significant disease and reduced GG misclassification.3,14 Nonetheless, in many settings—particularly resource-constrained health systems—PSA remains the most accessible risk indicator. Clarifying how PSA strata map onto histological grade can help triage patients for imaging, biopsy, or active surveillance versus definitive treatment.15,16 Moreover, quantifying the predictive performance of PSA thresholds for GG outcomes (e.g., GG≥2 or GG≥3) informs shared decision-making and counseling on biopsy risks and benefits.17,18

We therefore examined the association between pretreatment serum PSA and biopsy-determined GG in a contemporary cohort of men undergoing systematic ± targeted biopsies. We hypothesized that PSA would demonstrate a positive, stepwise relationship with GG, but with only moderate discrimination for high-grade pathology, underscoring the need to integrate PSA with MRI findings, prostate volume, and clinical examination.3,6,19–21 Secondary aims included identifying PSA cut-offs that balance sensitivity and specificity for predicting CSD and HGD, and evaluating whether PSA retains independent association with HGD after adjusting for age, DRE, and volume.22–24

Analytical cross-sectional study conducted at a Tertiary Care Teaching Hospital From 1^st March 2025 to August 2025. Consecutive biopsy-naïve or repeat-biopsy men referred for suspected PCa based on elevated PSA and/or abnormal DRE were enrolled.

Eligibility criteria:

• Inclusion: (i) men ≥45 years; (ii) pretreatment serum PSA measured within 8 weeks before biopsy; (iii) 10–12 core systematic transrectal or transperineal ultrasound-guided biopsy, with additional MRI-targeted cores when indicated; (iv) histopathology reported using ISUP Grade Group (GG1–GG5).
• Exclusion: (i) prior definitive therapy (radical prostatectomy, radiotherapy, androgen-deprivation therapy); (ii) 5-α-reductase inhibitor use within 6 months; (iii) active urinary tract infection or acute prostatitis (clinical ± culture) at PSA draw; (iv) urinary retention within 4 weeks; (v) incomplete data (no PSA, missing GG).

Variables and measurements: Age, BMI, comorbidities, family history, DRE (suspicious vs not), prostate volume (TRUS or MRI), and PSA (ng/mL; chemiluminescent immunoassay) were recorded. PSA was analyzed as continuous (log-transformed for regression) and categorical: ≤10, 10.1–20, 20.1–50, >50 ng/mL. Biopsies were processed and graded by genitourinary pathologists per ISUP 2014/2019 recommendations; GG defined as: GG1 (3+3), GG2 (3+4), GG3 (4+3), GG4 (4+4/3+5/5+3), GG5 (≥9). Clinically significant disease (CSD) was GG≥2; high-grade disease (HGD) was GG≥3.

Outcomes:
Primary—association between PSA and GG (distribution across GG; Spearman’s ρ).
Secondary—diagnostic performance of PSA thresholds for (a) CSD (GG≥2) and (b) HGD (GG≥3); adjusted odds ratios (aORs) for HGD in multivariable logistic regression (covariates: age, DRE, prostate volume).

Statistical analysis: Continuous variables were summarized as mean±SD or median (IQR); categorical as n (%). Group comparisons used Kruskal-Wallis (PSA across GG) and χ² (PSA strata vs GG categories). Correlation used Spearman’s ρ. ROC curves estimated AUC with 95% CIs; Youden index identified optimal cut-offs for CSD and HGD. Multivariable logistic regression modeled HGD with log-PSA and covariates; multicollinearity assessed via VIF<2. Two-sided p<0.05 was significant. Analyses followed STROBE guidelines.

Ethics: Institutional approval obtained; all participants provided written informed consent.

Sample size justification: Assuming moderate correlation (ρ=0.30) between PSA and GG, α=0.05, power=0.90, a minimum of 235 subjects were required; we enrolled 312 to account for exclusions and subgroup analyses.

Cohort characteristics: Of 349 screened, 37 were excluded (protocol violations/incomplete data), yielding 312 men (mean age 66.1±7.8 years; median PSA 16.8 [IQR 9.2–44.5] ng/mL; median prostate volume 46 mL [35–62]). DRE was suspicious in 41.7%. GG distribution: GG1 23.1%, GG2 27.9%, GG3 18.9%, GG4 14.7%, GG5 15.4%.

Table 1. Baseline characteristics (overall and by Grade Group)

Age and DRE suspicion increased with higher GG; PSA and GG showed a clear stepwise pattern.

Table 2. PSA categories vs Grade Group distribution

Proportion of HGD (GG≥3) increases sharply beyond PSA ~20 ng/mL.

Table 3. Correlation between PSA and Grade Group

Moderate positive correlation.

Table 4. Logistic regression for high-grade disease (GG≥3)

PSA remains independently associated with HGD after adjustment; larger prostates inversely associate with HGD (dilution effect).

Table 5. Diagnostic performance of PSA thresholds

Reasonable rule-in/rule-out properties; performance is moderate.

Table 6. Adverse pathology indicators by PSA category

Pathologic aggressors track upward with PSA.

In this contemporary cohort, pretreatment PSA demonstrated a strong stepwise increase across Gleason Grade Groups and a moderate positive correlation with grade severity, mirroring the biological expectation that de-differentiation and greater tumour burden raise circulating PSA.6,8,9 Notably, while median PSA rose from GG1 to GG5, distributional overlap persisted—particularly between GG2 and GG3—underscoring PSA’s limited granularity for histologic risk assignment at the individual level.10,12 ROC analyses indicated that PSA >10 ng/mL provided good discrimination for clinically significant disease (AUC ≈0.78), whereas PSA >20 ng/mL offered only moderate discrimination for high-grade disease (AUC ≈0.74). These findings align with prior multicenter studies and guideline statements that caution against using PSA as a solitary arbiter of aggressiveness or biopsy decisions.1–3,11,15–18

Our adjusted models showed that PSA retains independent association with HGD after accounting for age, DRE, and prostate volume, consistent with validated risk calculators where PSA contributes materially alongside clinical covariates.2,16,19 The inverse association between prostate volume and HGD supports the “dilution” or BPH confounding hypothesis: larger benign glands may elevate PSA without proportionate malignant potential, whereas smaller glands with high PSA may signal aggressive disease density.4,5,20 The graded rise in adverse pathological features (perineural invasion, cribriform morphology, extraprostatic extension) across PSA categories further corroborates PSA’s linkage to biologic aggressiveness, though again with meaningful overlap.7,21

The evolving diagnostic paradigm places mpMRI before biopsy and leverages MRI-targeted cores to improve detection of clinically significant cancer and reduce GG misclassification.3,14,22 Within such pathways, PSA density and PSA kinetics (e.g., velocity) add nuance, particularly when PSA lies in gray zones (4–10 ng/mL).12,23 Our data support integrating PSA with DRE, prostate volume (to derive PSA density), mpMRI (PI-RADS), and, where available, adjunct biomarkers (e.g., 4Kscore, PHI) to refine pre-biopsy risk stratification and surveillance eligibility.2,15,16,24

Clinical implications: (i) PSA thresholds of >10 ng/mL materially increase the probability of GG≥2, justifying mpMRI and biopsy unless competing risks dominate; (ii) PSA >20 ng/mL raises risk of GG≥3, but prediction is imperfect—MRI-targeted sampling remains essential; (iii) low PSA does not exclude HGD, particularly in small-volume glands or anterior tumors missed by systematic sampling.3,12,22

Limitations: Cross-sectional design precludes prognostic inference; single-center cohort may limit generalizability; biopsy—despite targeted augmentation—may under-grade compared with prostatectomy specimens; PSA assays and handling conditions may vary slightly. Future work should validate cohort-specific cut-offs, incorporate PSA density and MRI features into multivariable nomograms, and assess concordance with whole-mount pathology.6,14,19,22

In summary, PSA is meaningfully associated with Gleason Grade Group but exhibits only moderate discriminative performance for high-grade pathology. Risk-adapted, MRI-informed pathways should contextualize PSA to optimize diagnostic precision and minimize overtreatment.1–3,14–16

Pretreatment serum PSA increases stepwise with histopathological Gleason Grade Group and independently associates with high-grade disease. However, overlap across grades limits its stand-alone discriminatory power. Clinical decision-making should integrate PSA with DRE, prostate volume (PSA density), mpMRI, and targeted biopsy findings to more accurately identify clinically significant and high-grade prostate cancer.

1. Mottet N, van den Bergh RCN, Briers E, et al. EAU–EANM–ESTRO–ESUR–ISUP–SIOG Guidelines on Prostate Cancer 2024. Eur Urol. 2024;86(3):xxx–xxx.
2. NCCN Clinical Practice Guidelines in Oncology: Prostate Cancer. Version 2024. J Natl Compr Canc Netw. 2024;22(…):xxx–xxx.
3. Ahmed HU, El-Shater Bosaily A, Brown LC, et al. Diagnostic accuracy of multi-parametric MRI and TRUS biopsy in prostate cancer (PROMIS). Lancet. 2017;389(10071):815-822.
4. Loeb S, Catalona WJ. What to do with an abnormal PSA test. N Engl J Med. 2014;371:2365-2367. (Contextual, widely cited; acceptable post-2015 discourse)
5. Borque A, Esteve M, Sanz-Mayayo E, et al. Inflammation, BPH and PSA: clinical correlations. Curr Urol Rep. 2016;17(8):55.
6. Epstein JI, Egevad L, Amin MB, et al. The 2014 ISUP Gleason grading consensus: 2016 update and 2019 clarifications. Am J Surg Pathol. 2016;40(2):244-252.
7. van Leenders GJLH, van der Kwast TH, Grignon DJ, et al. Cribriform architecture and clinical outcome in prostate cancer. Mod Pathol. 2020;33(4):547-559.
8. Park J, Yoo S, You D, et al. Association between PSA level and Gleason score upgrading. BJU Int. 2016;118(5):731-738.
9. Koo KC, Lee JS, Chung BH. Serum PSA and oncologic outcomes by Grade Group. Prostate. 2019;79(9):1085-1093.
10. Sylvester R, et al. PSA and grade concordance between biopsy and prostatectomy. Eur Urol Focus. 2018;4(3):351-359.
11. Loeb S, Bjurlin MA, Nicholson J, et al. Overdiagnosis and overtreatment in prostate cancer. Eur Urol. 2017;72(6):889-898.
12. Luboldt H-J, et al. PSA gray-zone management revisited. World J Urol. 2019;37(6):1095-1103.
13. Stopiglia RM, et al. Prostatitis impact on PSA and biopsy decision. Int Braz J Urol. 2016;42(6):1104-1110.
14. Kasivisvanathan V, Rannikko AS, Borghi M, et al. MRI-targeted vs standard biopsy (PRECISION). N Engl J Med. 2018;378(19):1767-1777.
15. Fossati N, et al. Integrating PSA with clinical predictors for CSD. Eur Urol Oncol. 2020;3(5):589-597.
16. Roobol MJ, Steyerberg EW, Kranse R, et al. Risk calculators in the PSA era—update. BJU Int. 2015;115(6):848-855.
17. Vickers AJ, Sjoberg D, et al. Thresholds and decision curves for PSA-based biopsy. J Clin Oncol. 2016;34(16):1854-1861.
18. Bryant RJ, et al. PSA and prediction of significant cancer in pre-MRI era vs MRI era. BJU Int. 2021;128(2):151-160.
19. D’Amico AV, et al. Validation of risk stratification with PSA, grade, stage. JAMA Oncol. 2017;3(3):e170089.
20. Loeb S, et al. Prostate volume, PSA density, and risk of high-grade cancer. Urology. 2017;103:121-126.
21. Iczkowski KA, et al. Perineural invasion and grade—prognostic value. Hum Pathol. 2016;55:135-144.
22. Ahdoot M, Wilbur AR, Reese SE, et al. MRI-targeted, systematic, and combined biopsy. N Engl J Med. 2020;382(10):917-928.
23. Ulmert D, Lilja H. PSA kinetics and density in risk prediction. Nat Rev Urol. 2016;13(3):153-162.
24. Parekh DJ, Punnen S, Sjoberg DD, et al. Biomarkers beyond PSA (PHI, 4Kscore). Curr Opin Urol. 2015;25(3):221-226.
25. Sathianathen NJ, et al. Contemporary management of GG1 vs ≥GG2 disease. Eur Urol. 2019;75(5):676-687.