The comprehensive analysis of 339 men found that the combination of prostate-specific antigen density (PSAD) and the Prostate Imaging Reporting and Data System (PI-RADS) improves the performance of the PI-RADS protocol in prostate cancer diagnosis in biopsy-naïve patients and can help in the risk decision-making process before prostate biopsy.

Following the methodology of existing study, the population was stratified into four PSAD risk groups: < 0.10, 0.10–0.15, 0.15–0.20, and > 0.20 ng/mL/mL, with csPCa detection rates of 26.3%, 46.9%, 53.3%, and 70.6% of the total population, respectively [12].

In the PI-RADS 3 MRI group, where diagnostic uncertainty is common, the likelihood of clinically significant prostate cancer (csPCa) exhibited a notable escalation with higher PSAD levels. The percentage of csPCa in this group increased from 13.5% for low-risk PSAD < 0.10 to 54.5% for high-risk PSAD 0.2 and higher, demonstrating a clear association between elevated PSAD and an increased risk of significant malignancy.

In the low-risk PSAD group (< 0.10 ng/mL/mL) consisting of 52 men, seven men had diagnosed csPCa from biopsy, of which six were ISUP 2 and one ISUP 3, putting them in intermediate–low-risk group with 13.5% chance of presence of csPCa, these men may avoid immediate biopsy only with an adequate monitoring as part of shared decision-making [4, 13]. The high-risk PSAD group (> 0.20 ng/mL/mL) have an increased risk of csPCa of 54.5%, so a biopsy should be performed.

Within the MRI group classified as PI-RADS 4–5, the occurrence of clinically significant prostate cancer (csPCa) consistently showed higher rates across all prostate-specific antigen density (PSAD) categories. The highest proportion, reaching 73.0%, was observed in the high-risk PSAD (> 0.20 ng/mL/mL) subgroup. According to European Association of Urology (EAU) recommendations, individuals falling into this category should undergo targeted biopsy, either alone or in combination with systematic biopsy. Conversely, in the low-risk PSAD group (< 0.10 ng/mL/mL), 36.8% of individuals were diagnosed with csPCa, indicating that biopsies should be performed in men with low PSAD levels if MRI scans are positive.

When comparing our results with those of a similar study with a cohort of 3006 patients, men with PI-RADS 3 scores and low-risk PSAD (< 0.10 ng/mL/mL) had a lower risk (4%) of significant disease than in our study, suggesting that biopsies could have been avoided altogether, because our risk was higher—13%—we suggest avoiding biopsy only under close monitoring. The high-risk PSAD group (> 0.20 ng/mL/mL) had an increased risk of csPCa of 29%; in our study, the risk was even higher at 54%, so a biopsy should be done anyway [11].

According to various studies, PSAD does not influence biopsy decision in PI-RADS 4–5 category due to the higher prevalence of csPCa in these patients, so it was not found to be a useful tool to prevent biopsy in these patients; our results are consistent with these studies [11, 14].

Logistic regression affirmed the independent contributions of both PI-RADS category and PSAD levels to the likelihood of detecting csPCa. This statistical confirmation underscores the importance of integrating advanced imaging and biomarker data for a comprehensive risk assessment. The ROC analysis further emphasized the superior predictive performance of the combined model, where the AUC for PI-RADS category and PSAD (0.756) surpassed the individual predictors (PSA AUC 0.627, PI-RADS AUC 0.689, PSAD AUC 0.708.) The statistical significance of the differences in AUC values adds robustness to the argument for the synergistic value of PSAD and PI-RADS in predicting csPCa (Fig. 3). According to the results, the optimal PSAD cut-off of 0.10 ng/mL/mL is recommended in men with equivocal MRI results (PI-RADS 3) to determine the need for biopsy.

ROC curves for the biopsy-naive group of patients (N = 339) based on csPCa cases detected by combined biopsy from logistic regression models, years 2021–2023. Predictors are listed in the legend

Our findings are consistent with the results of several other studies that investigated the predictive value of PSAD in conjunction with MRI outcomes [1, 12, 15]. In the biopsy-naïve subset of the entire cohort comprising 526 patients, the incorporation of PI-RADS with PSAD resulted in an increased area under the curve (AUC) (PI-RADS score AUC = 0.799 vs. PI-RADS score + PSAD AUC = 0.830, difference in AUC = 0.031, 95% CI 0.012, 0.050, p = 0.002). However, such an effect was not observed in the active surveillance group (PI-RADS score AUC = 0.762 vs. PI-RADS score + PSAD AUC = 0.778, difference in AUC = 0.016, 95% CI 0.040, 0.071, p = 0.579) [12]. In a cohort of 372 men of comparable size, logistic regression models were formulated. The combination of the PI-RADS score and PSAD (AUC = 0.80) demonstrated a significantly superior performance compared to both PSA (AUC = 0.75, p < 0.01) and the Prostate Cancer Prevention Trial (PCPT) risk calculator (AUC = 0.76, p < 0.01). This combination substantially enhanced the risk discrimination for clinically significant prostate cancer on biopsy [1]. Within a cohort of 833 biopsy-naïve men, the optimal diagnostic performance for detecting csPCa was achieved through the combination of the PI-RADS score and PSAD. The AUC for the combined PI-RADS score with PSAD was 0.942 (95% CI 0.924–0.957). Notably, the diagnostic accuracy surpassed that of any individual clinical variable, including the PI-RADS score (p < 0.001) [15].

The study presents certain limitations that should be acknowledged. Despite efforts to involve experienced radiologists, the interobserver variability in MRI reporting may impact result consistency. The classification of lesions as PI-RADS category 3 introduces uncertainty in decision-making, requiring careful consideration of the balance between overtesting and detecting clinically significant cancer. Exclusion criteria, including patients under active surveillance and those with missing essential data, may introduce selection bias. The study’s single-center nature limits generalizability, and the relatively short follow-up until May 2023 restricts the assessment of long-term outcomes. Additionally, the absence of external validation in diverse cohorts and geographic regions emphasizes the need for cautious interpretation and potential adjustments when applying the combined PSAD and PI-RADS diagnostic approach.

Furthermore, an acknowledged challenge within the field is the interobserver variability in MRI reporting, a factor that could potentially introduce biases. The reliance on subjective interpretations, even by experienced radiologists, underscores the need for ongoing efforts to standardize reporting practices and mitigate potential variations in assessments. As the field advances, collaborative initiatives and standardized reporting protocols will be essential to enhance the robustness and reliability of future studies in prostate imaging.