What do you do with PI-RADS-3?

In this issue of CUAJ, a study by Lim et al showed that biopsy of Prostate Imaging and Data Reporting System (PI-RADS)-3 lesions yielded a low rate of clinically significant prostate cancer (CSPCa) of 10%.¹ The authors examined the use of magnetic resonance imaging (MRI) and targeted biopsy in the specific populations of men on active surveillance and men with a prior negative biopsy who have a PI-RADS-3 lesion on MRI. In addition to these indications, we are increasingly using MRI in patients with an elevated risk of CSPCa who are biopsy-naive. This is reflected in recent guideline recommendations, including from Cancer Care Ontario.²

Currently, it depends on individual or institutional practice standards whether to observe these lesions or biopsy with targeting. We should consider the variability in how radiologists interpret PI-RADS-3 lesions despite the specific imaging criteria that are defined in the PI-RADS v2 guideline. ³ We have to be wary of using the PI-RADS-3 category as a catch-all for lesions that may be difficult to categorize.

This may be illustrated to some extent in the current study by Lim et al. Of the 397 patients with PI-RADS-3 lesions who underwent expert, blinded review, 137 (35%) were reclassified into different PI-RADS categories. Therefore, seeking a second opinion or review may be a reasonable first approach to a PI-RADS-3 finding if there is elevated clinical suspicion for prostate cancer.

PI-RADS-3 lesions should be considered together with other factors, including prostate-specific antigen density (PSAD) and possibly size of lesion >14 mm. PSAD ≥0.15 has been shown to be a useful predictor of CSPCa in PI-RADS-3 lesions.³ A blanket approach of avoiding biopsy in all PI-RADS-3 lesions will avoidably miss some cases of CSPCa.

With the expanded indications for MRI, we are facing higher demands on our MRI resources. One strategy that we have incorporated at our center is adopting biparametric MRI.⁴ This is done without intravenous gadolinium, which benefits our patients, reduces costs, and speeds up scan times significantly. The only downside is the elimination of the dynamic contrast enhanced (DCE) images. However, if we combine clinical and imaging markers to improve our biopsy decisions, we may compensate for loss of DCE. This could be an important part of moving forward in Canada, given the expected increase in demand for MRI as provincial guidelines begin to adopt an MRI-before-biopsy strategy.

This study helps us in our approach to PI-RADS-3 lesions; however, consider that 10–30% of these lesions may contain CSPCa and 35% may be reclassified with a second review. Incorporate clinical stage, PSAD, and lesion size in the biopsy decision. Further research is needed to look for a predictive model when using MRI in biopsy-naive patients. We need to continue to examine ways of incorporating pre-biopsy MRI strategies to address resource constraints. One such strategy could involve biparametric MRI. As MRI techniques, interpretation, and PI-RADS guidelines improve, we may see fewer PI-RADS-3 lesions being reported. For now, clinical judgement, clinical markers, and imaging characteristics should be used to guide the biopsy vs. observe decisions in these patients.