When the two strategies for avoiding biopsies were compared, the index lesion volume strategy resulted in a significantly greater proportion of missed sPCas than the PSAd strategy (16% vs. 7%). Furthermore, two men had GG 4 PCa detected. Both of these men had PSAd ≥ 0.15 ng/mL2 and would have been recommended for biopsies, but only one had an index lesion volume ≥ 0.5 mL (Table 1). Consequently, one man with high-grade GG 4 disease would have been missed if the lesion volume strategy alone had been used to recommend biopsies.

The PSAd strategy had a higher NPV than the index lesion volume strategy for ruling out sPCa (93% vs. 83%). Consequently, the PSAs strategy is better at avoiding under-diagnosing of sPCa. However, the PPV of both strategies was low (42% vs. 19%).

The BIDOC study by Boesen et al. found that a prebiopsy bpMRI in biopsy naïve men had a low NPV for any PCa (72%) for a modified PI-RADS score of 3 or higher, but a high NPV for sPCa (97%) [5]. Boesen et al. defined sPCa as any core with high-grade PCa (GG ≥ 3) or a maximum cancerous core length greater than 50% of GG 2 PCa [5]. If the definition of sPCa was changed to any core with PCa GG ≥ 2, the NPV decreased to 93%.

Interestingly, a study of 141 men with PI-RADS 3 lesions on mpMRI was done to investigate whether PI-RADS 3 lesions changed over time [18]. Overall, 77% of men with PI-RADS 3 lesions exhibited a change from PI-RADS 3 to either PI-RADS 2 or 4 within the first year and 15% of all the patients harbored sPCa [18]. This observation suggests an additional strategy to the two evaluated in our study. Instead of immediate prostate biopsies, a repeat confirmatory mpMRI of men with PI-RADS 3 index lesions 1 year after the initial bpMRI may be beneficial, to discover whether any lesions have changed.

A recent study by Kortenbach et al. investigated 200 biopsy-naïve men with clinical suspicion of PCa who underwent a prebiopsy bpMRI and had PSAd measured, with a 2-year clinical follow-up [20]. The same PSAd strategy was applied in that study as we used in the present study: If a man had a PI-RADS score of 3 and a calculated PSAd < 0.15 ng/mL2, no biopsies were performed. If the man had a PSAd ≥ 0.15 ng/mL2 then he had targeted and systematic biopsies. Among 109 men with a PI-RADS score of 1–3, 11 men (10%) had a PI-RADS 3 lesion, distributed as four men with PSAd < 0.15 ng/mL2 and seven men with a PSAd ≥ 0.15 ng/mL2. The biopsies from the men with PSAd ≥ 0.15 ng/mL2 revealed that five men had sPCa and two men had no cancer [19]. The four men with PSAd < 0.15 ng/mL2 had no biopsies initially and no further biopsies during the follow-up surveillance period, which involved monitoring PSA levels, follow-up multiparametric MRI and digital rectal examinations [19]. The number of men with PI-RADS 3 lesions in the study by Kortenbach et al. is limited, but no man in that study who had a PI-RADS 3 lesion and PSAd < 0.15 ng/mL2 had sPCa detected within the follow-up period. This observation validates the results of our present study in which men with PI-RADS 3 lesions and PSAd < 0.15 ng/mL2 had a maximum GG of 3. However, in the study by Kortenbach et al., men with PI-RADS 3 lesions and PSAd < 0.15 ng/mL2 had no biopsies, which could have detected sPCas that were missed [19].

Studies have reported that tumors < 0.5 mL are unlikely to become clinically significant during a man’s life span and therefore do not warrant treatment [20, 21]. These studies support the index lesion volume strategy; however, approximately 20% of lesions are ≥ 0.5 mL and are more likely to develop to into sPCa. In addition, the approach recommended by Scialpi et al. uses MRI to measure lesion volumes and not prostatectomy specimens [12].

Both of the strategies compared in the present study could reduce unnecessary biopsies and increase the diagnostic yield of sPCa [7, 8, 20]. Neither strategy incurs additional costs. The PSAd strategy would not prolong reading times because we already use PSAd in our daily practice; therefore, the volume of every prostate is measured using MRI. We already measure lesion size, although not in three planes; however, this could be accommodated without substantially increasing reading times.

The main limitation of our study was the small study population. We would need a larger sample to reach more rigorous conclusions. However, the results we have warrant larger studies to validate these findings.

A second limitation was our study’s retrospective design. Unfortunately, bpMRI results were not read using the particular method recommended as part of the index lesion volume strategy. The retrospective design also means that this study is a sub analysis based on another study with a different purpose. Furthermore, although our study population had targeted biopsies, as recommended by Scialpi et al., they also had systematic biopsies, which were not recommended [12].

A third limitation is that we did not have prostatectomy-specimens as our reference test, which is a perfect gold standard. Our database was limited to the combined results of systematic and targeted biopsies. Thus, the final pathological results are unknown and sPCa could have been missed. This also means that we are unable to assess the true false negative rate.

Finally, no inter-reader assessments were performed as all bpMRI readings were reported by a single experienced urogenital radiologist. The performance of less experienced readers may affect the results as well as image quality and disease prevalence. When making clinical decisions based on MRI findings each institution should know their own test performance statistics.

In conclusion, the PSAd strategy performed better than the lesion volume strategy to rule out sPCa in men with equivocal PI-RADS category 3 index lesions, but larger studies are needed to support this conclusion.