PI-RADS Offers Prognostic Advantage Over mpMRI Alone for Prostatectomy

Mieszko Kozikowski, MD

Use of the novel prostate imaging reporting and data system (PI-RADS) enhanced the ability to determine prostate cancer stage versus standard multiparametric MRI (mpMRI) alone, offering advantages for deciding the extent of resection needed during an endoscopic radical prostatectomy (ERP), according to retrospective findings from a study presented at the 2017 European Multidisciplinary Meeting on Urological Cancers (EMUC).

In the study, the extent of ERP could have been altered for 70% of patients following a re-evaluation of mpMRI findings using the PI-RADS guideline. Use of the standardized radiology scoring system would have resulted in a less aggressive, neurovascular bundle (NVB)-sparing procedure for 55.5% of men and a more aggressive resection for 17%. The decision to change to a less aggressive surgical approach was associated with a lower rate of positive surgical margins (P = .036).

"We can effectively use MRI and PI-RADS to stage prostate cancer and assistant our decision-making process regarding extent of resection," lead investigator Mieszko Kozikowski, MD, Medical Centre of Postgraduate Education, European Health Centre Otwock, Urology Clinic, Warsaw, Poland, told OncLive at the EMUC meeting. "We are able to change our approach in terms of extensiveness of surgery in 70% of men qualified for surgical treatment. In half of these men, we are able to spare neurovascular bundles that would be normally resected."

PI-RADS is a preoperative workup utilized prior to radical prostatectomy to help expedite and improve findings from mpMRI. The tool scores each lesion on a scale of 1 to 5, with 1 being benign and 5 being highly suspicious of malignancy. The system was developed jointly by the European Society of Urogenital Radiology, American College of Radiology, and AdMeTech Foundation.

The retrospective trial presented at EMUC analyzed data from 110 men at a mean age of 63.7 years (±6.6) who were included in a study using staging with a 3-Tesla mpMRI prior to ERP. The mean PSA of those included in the analysis was 11.8 ng/ml (±15.3), the mean prostate volume was 41.0 (±19.3), the PSA density was 0.34 mg/ml2 (±0.54). The Gleason score (GS) by biopsy was ≤6 for 58, 7 for 41, and ≥8 for 11 patients. In the final pathology, the GS was ≤6 for 40 of patients, 7 for 86, and ≥8 for 28. The pathologic stage was T2 for 61.8% of patients and T3 for the remainder.

NVB-sparing surgery was pursued in patients with stage T1c disease who had GS <7 and serum PSA <10 ng/mL. For those who did not meet these clinical criteria, NVB eligibility was determined using mpMRI, based on the presences or absence of extraprostatic extension (EPE). All imaging reports were compared with the final pathology to evaluated performance.

For the retrospective analysis, the 176 total lesions detected in 104 evaluable participants were assessed using PI-RADS, with scores of 5 (42.7%), 4 (43.6%), 3 (9.1%), and 1 or 2 (4.5%). EPE was found in 61.7% of those with a PI-RADS score of 5 and in 23% of those with a score of 4. In those scored 1 or 2, there was no EPE found, and only 1 patient with a PI-RADS score of 3 had EPE. Based on the limited data from the study, the researchers were unable to definitively recommend a score threshold for EPE. However, lower PI-RADS scores suggested the most changes in surgical decisions (P = .001).

In the original study, which used mpMRI alone, the positive surgical margin (PSM) rate was 14.5%. Of those deemed appropriate for less aggressive surgery using PI-RADS (n = 61), the PSM rate was 8.2%. The PSM rate was 25% for those chosen for a more aggressive resection by PI-RADS (n = 16) and in those with no change indicated (n = 33) the PSM rate was 21.2%.

The overall prostate cancer diagnostic accuracy was higher with the addition of PI-RADS compared with mpMRI alone (area under the curve, 0.737 vs 0.597; P = .025). The sensitivity with the addition of PI-RADS was 69.1% compared with 35.7% for mpMRI alone. The specificity was 73.5% for PI-RADS and 82.4% for mpMRI. For the PI-RADS and mpMRI approaches, respectively, the positive predictive values were 61.7% and 55.6% and the negative predictive values were 79.4% and 67.5%. "PI-RADS can be safely incorporated into preoperative workup without increasing the risk of PSM," the researchers concluded.

These findings compliment other retrospective analyses exploring PI-RADS, with further validation of the system under way. Several prospective trials are currently enrolling participants to further validate the 5-point scale.

Kozikowski M, Powro&#378;nik J, Piotrowicz S, et al. PI-RADS as a tool for estimating prostate cancer stage and extension of resection. Presented at: 2017 EMUC Congress; Barcelona, Spain; November 16-19, 2017. Abstract P009.