Advertisement

Prospective Evaluation of Cancer Detection Rates of the Prostate Imaging Reporting and Data System version 2

Login to Access Video or Poster Abstract: MP08-10
Sources of Funding: The author&[prime]s postdoctoral fellowship is funded by a research grant from the &[Prime]Dr. Mildred Scheel&[Prime] foundation (Bonn, Germany)

Introduction

Prostate Imaging-Reporting and Data System version 2 (PI-RADSv2) was introduced in 2015. The likelihood of harboring clinically significant prostate cancer (CS PCa) on multiparametric MRI (mpMRI) is assessed on a five-point scale. We prospectively evaluated cancer detection rates (CDRs) of PI-RADSv2 scores using the new International Society of Urological Pathology (ISUP) grading group system as the gold standard.

Methods

From May 2015-May 2016, 963 patients underwent prostate mpMRI including T2 weighted (T2W), diffusion weighted, apparent diffusion coefficient maps, high b value (1500-2000s/mm2) and dynamic contrast enhancement sequences. 339/963 patients underwent MRI/US fusion guided biopsy. The highest Gleason score per target lesion was given an ISUP score. Lesion-based CDRs for all PCa and CS PCa (ISUP≥2, ≥Gleason 3+4) were calculated for each PI-RADSv2 score in the entire prostate, peripheral (PZ) and transition zones (TZ).

Results

CDRs for all and CS PCa for each PIRADSv2 score are shown in Figure 1. PI-RADSv2 score 5 had the highest CDRs for all and CS PCa at 87% and 72%, respectively. PI-RADSv2 score 4 had unexpectedly low CDR for both all and CS PCa (39% and 22%, respectively). Specifically, in the PZ, the CDR of T2W PI-RADSv2 score 4 was significantly higher than the CDR of overall PI-RADSv2 score 4 for all PCa (48% vs. 37%, p=0.01) and CS PCa (33% vs 23%, p=0.002) (Figure 2).

Conclusions

CDRs increase with higher PI-RADSv2 scores. CDR of PI-RADSv2 score 4 is low due to a high false positive rate. In the PZ, T2W scores combined with DWI scores, rather than DWI scores alone may improve the CDR for PI-RADSv2 score 4 lesions. Future versions of PI-RADS should take this into account.

Funding

The author&[prime]s postdoctoral fellowship is funded by a research grant from the &[Prime]Dr. Mildred Scheel&[Prime] foundation (Bonn, Germany)

Authors
Sherif Mehralivand
Sandra Bednarova
Joanna Shih
Francesca Mertan
Sonia Gaur
Maria Merino
Bradford Wood
Peter Pinto
Peter Choyke
Baris Turkbey
back to top