Predictors of Ipsilateral Extraprostatic Extension on Radical Prostatectomy using 3-Tesla Multiparametric MRI and Contemporary Pathologic Features
Prostate cancer risk stratification is integral to the decision algorithm for neurovascular bundle (NVB) preservation during radical prostatectomy (RP). While existing pre-RP risk models are useful for clinical staging, the laterality of extraprostatic extension (EPE) may not be clearly established prior to surgery. Herein, we sought to utilize multiparametric MRI (mpMRI) and contemporary pathologic criteria to determine predictors of ipsilateral EPE at RP.
Among 290 patients who underwent MRI/US fusion targeted biopsy (TB) at a single institution, we performed a retrospective review of 63 patients who subsequently underwent robotic-assisted RP. mpMRI was performed using 3-Tesla MRI scanners using pelvic phased-array coil and scoring was performed by two MRI radiologists using PI-RADS version 2. All patients underwent systematic biopsy (SB) and TB of PI-RADS ≥4 lesions in the same setting. Patients with a PI-RADS ≥4 lesion in the midline were excluded (n=7). A genitourinary pathologist performed RP tumor mapping. For each hemiprostate (n=112), mpMRI and pathologic parameters (including new Grade Grouping (GG) and Gleason pattern 4 (GP4) components) at the time of biopsy were recorded. Associations between covariates with ipsilateral EPE at RP were determined using logistic regression analyses.
On univariate analysis, mpMRI predictors of ipsilateral EPE at RP were the presence of PI-RADS 5 or ≥4 lesion and capsular abutment/bulging (all p≤0.0006). Pathologic predictors (all p≤0.0181) included: higher GG, % overall positive cores, maximum core % involvement, maximum core % GP4, number of cores with cribriform pattern, and presence of perineural invasion on SB+TB; number of positive cores and number of positive sextants on SB. On multivariate analysis, only the presence of PI-RADS 5 lesion (OR 3.827, 95%CI 1.293-11.327, p=0.0154) and maximum core % GP4 (OR 1.027, 95%CI 1.014-1.041, p<0.0001) on SB+TB were independently predictive.</p>
When selecting patients for NVB-sparing RP, balancing the achievement of safe oncologic margins with potential functional morbidity is a challenge. While the utilization of mpMRI in pre-RP staging has been generally limited due to poor sensitivity for detecting microscopic EPE, incorporation of contemporary pathologic criteria after MRI/US fusion biopsy can enhance pre-RP staging. Our data suggests that ipsilateral NVB-sparing may be considered in the absence of a PIRADS 5 lesion and with low volume GP4 involvement on MRI/US fusion biopsy.