Elsevier

Clinical Radiology

Volume 73, Issue 9, September 2018, Pages 810-817
Clinical Radiology

Prebiopsy biparametric MRI: differences of PI-RADS version 2 in patients with different PSA levels

https://doi.org/10.1016/j.crad.2018.05.007Get rights and content

Highlights

  • bpMRI is useful for detecting clinically significant prostate cancer (csPCa).

  • bpMRI could detect most csPCas in patients with PSA>10 ng/dL.

  • csPCa could be missed by bpMRI in patients with PSA ≤ 10 ng/mL.

Aim

To validate the diagnostic accuracy of Prostate Imaging-Reporting and Data System (PI-RADS) version 2 in detecting clinically significant prostate cancer (csPCa, Gleason score ≥7) on prebiopsy biparametric MRI (bpMRI) in patients with different prostate-specific antigen (PSA) levels.

Materials and methods

This retrospective study included 184 patients who underwent prebiopsy bpMRI followed by transrectal ultrasonography-guided biopsy between June 2015 and February 2017. Reader 1 performed a combination of systematic and targeted biopsy with cognitive fusion after reviewing bpMRI and reader 2 reviewed the bpMRIs retrospectively. PI-RADS categories 4 and 5 were considered positive, and the results of the biopsy were considered the reference standard. Diagnostic performance of PI-RADS of bpMRI was evaluated in two PSA groups with a PSA cut-off level of 10 ng/ml and compared to PSA and the PSA density using receiver operating characteristics (ROC) curve analysis.

Results

csPCa was diagnosed in 24 of 123 patients (19.5%) and 26 of 61 patients (42.6%) in the low and high PSA groups, respectively. A PI-RADS v2 category by either readers 1 or 2 had a significantly better performance to detect csPCa than PSA in both PSA groups. In the high PSA group, only one csPCa was missed by reader 2, but none by reader 1. In the low PSA group, readers 1 and 2 were unable to detect seven and five of the 24 csPCas, respectively.

Conclusion

Prebiopsy bpMRI has good performance for detecting csPCa in the high PSA group but may miss small-volume csPCa in the low PSA group.

Introduction

Magnetic resonance imaging (MRI) of the prostate performed before biopsy, so-called prebiopsy MRI, is useful to avoid unnecessary prostate biopsy and detect clinically significant prostate cancer (csPCa) more often than the standard protocol without a prebiopsy MRI.1, 2 This approach may be useful due to advances in MRI techniques and the widespread use of targeted biopsy. Prebiopsy MRI is usually performed with multiparametric MRI (mpMRI) and includes T2-weighted imaging (WI), diffusion-weighted imaging (DWI), and dynamic contrast-enhanced MRI (DCE-MRI); however, mpMRI requires an injection of contrast material and long acquisition time for DCE-MRI. Additionally, the role of DCE-MRI has been reduced in prostate imaging reporting and data system version 2 (PI-RADS v2) compared to PI-RADS v1.3 Several studies have suggested that the clinical usefulness and good diagnostic performance of biparametric MRI (bpMRI) is consistent with T2WI and DWI.4, 5, 6, 7, 8, 9 Therefore, there is a demand for an adjustment of PI-RADS v2 to make it applicable to bpMRI.10

Prostate specific antigen (PSA) is a tumour marker that is widely used as a screening tool to detect prostate cancer (PCa). The prevalence of PCa is different according to PSA levels: 2.5–4, 4–10, and >10 ng/ml.11, 12 Approximately one-third of men with a PSA level of ≤10 ng/ml and two-thirds of men with a PSA level >10 ng/ml have PCa.11, 12, 13 In two studies performed in Asia, the positive biopsy rate ranged between 9 and 19% in men with a PSA level of 4–10 ng/ml.14, 15 A low positive predictive value for PSA leads to unnecessary biopsies in many men. A combination of the PSA level and prebiopsy MRI has been emphasised to enhance risk modelling of PCa and reduce unnecessary biopsies.16, 17, 18, 19 Considering the different risks of PCa according to the PSA level, risk stratification according to PI-RADS categories may be dependent on PSA levels.

The purpose of this study was to validate the diagnostic accuracy of PI-RADS v2 in detecting csPCa (Gleason score [GS] ≥7 [3+4]) on prebiopsy bpMRI in patients with different PSA levels.

Section snippets

Patients

The Institutional Review Board approved this retrospective study and waived the requirement for informed consent. Patients who underwent transrectal ultrasonography (TRUS) biopsies of their prostate gland between June 2015 and February 2017 were eligible for this study. A total of 184 patients (mean age 66.9±8.4) who underwent prostate MRI within 6 months before biopsy were enrolled in this study with a mean interval of 19.1 days (range, 0–148 days) between TRUS biopsy and MRI. Serum PSA was

Patients

The low and high PSA groups included 123 and 61 patients with 6±1.9 and 94.1±370.7 ng/ml mean PSA, respectively. TRUS biopsy-diagnosed csPCa was found in 24 of 123 patients (19.5%) and 26 of 61 patients (42.6%) in the low and high PSA groups, respectively. Seventy of 184 patients had a previous biopsy history, and the proportion of biopsy-naive patients was not different between the two PSA groups (p=1.000). MRI was performed using a 1.5 T unit in 74 of 123 and 30 of 61 patients in the low and

Discussion

PI-RADS v2 category on prebiopsy bpMRI was useful to detect csPCa in this study; however, the results showed some differences between patients with PSA ≤10 ng/ml and PSA >10 ng/ml. Sensitivity and NPV for PI-RADS v2 were higher in the high PSA group compared to the low PSA group. Only one csPCa and no csPCa were missed by two readers in the high PSA group. By contrast, seven and five csPCa were missed by readers 1 and 2 in the low PSA group, possibly due to the small size of the tumours.

References (25)

  • M. Scialpi et al.

    Biparametric versus multiparametric MRI with non-endorectal coil at 3T in the detection and localization of prostate cancer

    Anticancer Res

    (2017)
  • I. Jambor et al.

    Novel biparametric MRI and targeted biopsy improves risk stratification in men with a clinical suspicion of prostate cancer (IMPROD Trial)

    J Magn Reson Imaging

    (2017)
  • Cited by (0)

    View full text