Risk-adapted biopsy decision based on prostate magnetic resonance imaging and prostate-specific antigen density for enhanced biopsy avoidance in first prostate cancer diagnostic evaluation.

European Association of Urology

International Society of Urological Pathology

National Institute for Health and Care Excellence

Prostate Imaging‐Reporting and Data System

PSA density

Prostate MRI has been integrated into the diagnostic evaluation for men at risk of having clinically significant cancer in multiple clinical care guidelines. Owing to the low false‐negative rate of prostate MRI accompanying a sensitivity averaging 91%, we can reduce biopsies (by 30%) and indolent cancer detection, while maintaining (or even improving) detection of significant cancers compared to systematic biopsy [1]. Reducing false‐negative MRI results further remains a clinical priority, calling for risk‐adapted strategies for patient biopsy selection [2, 3]. Multivariable risk‐prediction tools incorporating MRI results have been developed [4], but remain unvalidated.

Multiple publications advise the combination of MRI findings and PSA density (PSAD) for defining men who can safely avoid biopsy [5, 6]. In the present analysis, we extract in biopsy‐naïve men, the data relating to prostate MRI results and PSAD values to the likelihood of clinically significant disease. We propose a simple, risk‐adapted advice for biopsy avoidance, based on current urological guideline thresholds. Our analyses are meant to be hypothesis generating, requiring prospective validation.

Using PubMed/Ovid‐MEDLINE literature searches (up to May 2020), studies on biopsy‐naïve men undergoing systematic and where indicated MRI‐directed targeted biopsies at the same time, that provided three or four PSAD risk categories in the low‐, intermediate‐ and high‐risk range were found and related to Prostate Imaging‐Reporting and Data System (PI‐RADS) scores [5, 6, 7, 8, 9]. Studies using a single threshold for PSAD or binary MRI classification (positive/negative) were excluded, as were multivariate studies using PSADs. We also excluded studies that only included selected negative or equivocal MRI studies without prevalence data.

In total, 3006 biopsy‐naïve men in five studies were included (study characteristics Table S1). The overall prevalence of significant disease (International Society of Urological Pathology [ISUP] ≥2 cancers) among studies was 39% (range 28–49%).

The population was stratified into four (or three; rates between brackets) PSAD risk groups (Table 1 [5–10): <0.10, 0.10–0.15, 0.15–0.20, and >0.20 ng/mL/mL, corresponding to 31% (29%), 28%, 16%, and 25% (27%) of the whole population; with ISUP ≥2 cancer detection rates of 11% (14%), 28%, 47%, and 66% (68%), respectively.

Table 1

Risk data table of clinically significant prostate cancer (csPCa), related to PI‐RADS score and PSAD categories in biopsy‐naïve men, clinically suspected of having significant disease.

The population was also stratified into three PI‐RADS category risk groups: PI‐RADS 1–2 (low risk), 3 (intermediate risk ), and 4–5 (high risk), corresponding to 38% (36%), 12% (14%), and 50% (51%) of whole study population, with ISUP ≥2 detection rates of 6% (9%), 16% (22%), and 62% (65%), respectively (Table 1).

MRI‐negative men (PI‐RADS 1–2) with a low‐risk PSAD (<0.10 ng/mL/mL) have a 3% (4%) risk of significant disease (Table 1, Fig. S1a). MRI‐negative men with intermediate–low (0.10–0.15 ng/mL/mL) or intermediate–high risk PSAD (0.15–0.20 ng/mL/mL) have respectively 7% and 8% risk of significant disease. MRI‐negative men with a high‐risk PSAD (>0.20 ng/mL/mL) have 18% (24%) risk of significant disease. These data suggest that MRI‐negative men with low‐risk PSADs can avoid biopsies all together (lower than average population risk [<5%] [10]). Additionally, using the European Association of Urology (EAU)/National Institute for Health and Care Excellence (NICE)/AUA recommendations, MRI‐negative men with a PSAD of 0.10–0.15 or 0.15–0.20 ng/mL/mL may avoid immediate biopsy with an adequate safety‐net of monitoring as part of shared decision‐making. However, MRI‐negative men with a high‐risk PSAD (>0.20 ng/mL/mL) require systemic biopsies, despite the absence of visible MRI‐targets (Fig. S1a).

Men with indeterminate PI‐RADS 3 scores and a low‐risk PSAD (<0.10 ng/mL/mL) have a low 4% (9%) risk of significant disease, and biopsies could be avoided (Table 1, Fig. S1a). Men with PI‐RADS 3 scores and a high‐risk PSAD (>0.20 ng/mL/mL) have an elevated 29% (36%) risk, and targeted and systematic biopsies should be taken.

MRI‐positive men with PI‐RADS 4–5 category scores, irrespective of PSAD‐risk categories, should undergo targeted with or without systematic biopsies (EAU recommendations). Indeed, PI‐RADS 4–5 category scores were found in 189 of 674 (28%) men with low‐risk PSAD values (<0.10 ng/mL/mL); in these men, 31% were found to have significant disease, indicating that biopsies should be taken in men with low PSAD values when MRI scans are positive.

When only considering biopsy at a PSAD of >0.10 ng/mL/mL (without MRI assessments), 31% (28%) of biopsies are avoided at the expense of missing significant disease in 9% (10%) in whole tested population (Table S2). On the other hand, taking the combination of PI‐RADS scores with PSAD risk categories described above for biopsy decisions, a total of 38% (36%) biopsies can be avoided at the lower expense of missing significant cancers in 5% (6%) (Table S2).

Overall, a PSAD of >0.20 ng/mL/mL impacts 68 of 839 (8%) MRI‐negative men by increasing biopsy procedures (Table 1). A PSAD of <0.10 ng/mL/mL impacts 74 of 254 (29%) PI‐RADS 3 category men who could avoid biopsies. PSAD does not impact biopsy decisions in PI‐RADS 4–5 category men. Therefore, in the combined risk strategy of MRI with PSAD thresholds proposed above, the total number of men avoiding biopsies (38%) is unchanged compared to using only MRI risk assessment (38%), but men are more appropriately selected according to individual risk (Fig. S1a).

This analysis has limitations that require careful consideration. We have categorised biopsy‐naïve men into different risk profiles based on their MRI risk categories and PSAD ranges. We are cognisant that categorisation of continuous variables such as PSAD values is not advisable statistically for creating risk‐prediction models, due to the loss of information and power. However, we are not creating a model for biopsy avoidance; instead, we are merely confirming the potential for clinically meaningful inferences as noted by Boesen et al. [5, 6] and Falagario et al. [5, 6]. Further, justification of PSAD categories comes from their common usage in clinical care guidelines (EAU, NICE) making our present findings practically relevant (Fig. 1). We recognise that some data points have small numbers of patients making the estimates of the risk of significant cancers less reliable. We also recognise that these data are applicable for a mean ISUP ≥2 prevalence of 35% (range 28–46%) in biopsy‐naïve men, and would need to be adjusted to other population prevalence’s. Validation of our approach of adapting MRI findings by PSAD values for biopsy decisions is beginning to emerge [11] (Fig. S1b), reinforcing their routine use for biopsy decisions in clinical practice (EAU, NICE), while we wait the validation of MRI‐based multivariate risk prediction tools. The proposed approach is hypothesis generating and in need of further prospective validation in different populations of men at risk of prostate cancer.

Fig. 1

Risk‐adapted matrix table for guidance in biopsy decision, based on PSAD and on MRI risk assessments. The continuum of estimated risks of having a biopsy‐detectable prostate cancer are categorised into low‐ (not elevated), intermediate‐ and high‐risk (elevated). Generally, men at higher risk of having clinically significant prostate cancer require biopsy regardless of how risk is estimated. The MRI risk assessment is categorised into low (PI‐RADS 1 or 2), intermediate (PI‐RADS 3) or high (PI‐RADS 4–5) risk of having a biopsy‐detectable clinically significant prostate cancer. Each cell ascribes a different biopsy action, based on the results of Table 1. This matrix table may help guide biopsy‐decision management, while awaiting the validation of multivariable MRI prediction tools.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not‐for‐profit sectors.

Conflict of Interest

None declared.

Table S1. Study characteristics.

Table S2. Diagnostics metrics for significant prostate cancer detection (ISUP Grade ≥2) at various thresholds of PI‐RADS score and PSAD risk category, related to biopsy avoidance.

Fig. S1. Proportions of clinically significant prostate cancer related to PI‐RADS and PSAD risk category in biopsy‐naïve men.

Click here for additional data file.