Study design
This retrospective study was approved by the local ethics committee. Between October 2019 and December 2021, patients who received prostatic artery embolization and prior MRI were included in this study. All patients had severe LUTS due to BPH, refractory to medical treatment, and had been seen by an experienced urologist. Decisions for PAE were made in consensus between patients, urologists, and radiologists. Patients were informed about the intervention at least 24h before treatment and written informed consent was present from all patients. Prostate cancer was excluded prior to PAE by MRI and biopsy (in case of suspicious MRI or clinical indication). All interventions were performed by an interventional radiologist with 6 years of experience in prostatic artery embolization (*blinded*).
Pre-interventional MR scans were analysed retrospectively. The Prostate Imaging Reporting and Data System (PIRADS) classification, prostate volume, intravesical prostatic protrusion (IPP), prostatic urethra angle (PUA), presence of AdBPH, and pattern of hyperplasia according to the Wasserman classification and modified BPH classification with only three different patterns, classified by the predominance of the distribution of hyperplasia (preurethral, retrourethral, biurethral) were assessed (Tab. 1). Urological examination took place before the intervention. Current BPH medication, presence of urine catheter, IPSS, and Qmax were defined as clinical variables. PAE was defined as technically successful, if both sides were embolized, partially successful, if only one side was embolized, and unsuccessful, if embolization of the prostatic arteries failed. Patients received follow-up MRI without the application of contrast media and urological examination three months and one year after embolization. Outcome parameters in MRI included prostate volume, IPP, and PUA, IPSS, BPH medication, and presence of urine catheter were defined as clinical outcome parameters. Finally, MRI and clinical parameters were compared pre- and post-embolization.
Imaging Acquisition
All mpMRI scans were conducted on 3T MRI scanners (Magnetom Prisma, Siemens Healthineers, Forchheim, Germany) using a 60-channel phased-array surface coil. MRI parameters were chosen according to international recommendations and contained T2-weighted turbo spin echo (TSE) sequences in 3 planes (T2WI), diffusion-weighted imaging (DWI), and T1-weigehted imaging [26]. If there was no suspicion for cancer after the acquisition of non-enhanced images, we acquired a MRA with intravenous contrast media (3D T1 FLASH, TE 1.24, TR 3.7ms, Slice thickness 0.9mm, FOV 350mm, 20 ml CM dose, 2 ml/s insertion rate). The field of view was placed over the lower abdominal aorta and iliac vessels and involved PA and entire prostate tissue. For follow-up MRI, we acquired T2-weighted turbo spin echo (TSE) sequences in three planes (T2WI), diffusion-weighted imaging (DWI) and T1-weighted imaging. Prostate volume was measured by software volumetric (DynaCAD, Philips Healthcare).
Prostatic artery embolization
Prostatic artery embolization was conducted in all seventy-one patients. All PAE were performed using a therapeutic angiographic unit with a digital flat-panel detector system (Allura Xper FD20, Phillips Healthcare, Best, The Netherlands) equipped with cone beam CT option. First, the right common femoral artery (CFA) was punctured in seldinger technique and 5F-sheath was inserted. Probing of left internal iliac artery was conducted using a 5F-RIM, 5F-SIM-1, and a hydrophilic guidewire. Next, DSA in an angulated series (LAO 30°, CRAN 10°) or CBCT (using 3D road map) was performed to identify the origin of the left prostatic artery (PA). Afterwards, a microcatheter (Direxion, Bern-Shape, 2.7/2.4 Fr, Boston Scientific, Marlborough, MA, USA) was coaxially inserted and probing of left the PA was performed using a microwire (Fathom 0.016’’). CBCT was executed applying 5 ml of diluted contrast (Imeron 400/NaCl, 50:50) at 0.2 ml/s to check embolization position and exclude collateral vessels. If collaterals were observed to penis, bladder or rectum, these branches were occluded temporarily using Gelfoam. Microcatheter was placed distal in wedge position. Embolization was conducted using 250µm-particles (Embozene Microspheres, Varian Medical Systems, Paolo Alto, CA) and subsequent 350-500µm-Contour-particles (Boston Scientific, Natick, Massachusetts) until full stasis in the vessel was achieved. Embolization was performed subsequently on the right side in the same way. In case of insufficient probing/catheter positioning (e.g., due to vessel stenosis) or if protective embolization of collateral vessels was unfeasible on one or both sides, prostate embolization was not conducted, respectively. After completing embolization all extraneous material was eliminated, and the puncture side was closed using 6F Angioseal.
Statistical Analysis
Statistics were performed using IBM SPSS® Statistics (Version 27, IBM Corp). P-values < 0.05 were defined as statistically significant. Descriptive statistics included mean, median, standard deviation, and interquartile ranges. Wilcoxon signed rank test was performed to check for statistically significant differences in outcome parameters between baseline and follow-up. For the prediction of volume change and IPSS reduction due to treatment, multiple linear regression was performed including different input parameters (age, successful embolization, volume, IPP, PUA, AdBPH, IPSS, urinary retention, type of hyperplasia). The predictive value of the resulting model with nine predictors was assessed by reporting the adjusted r² value. For comparison of outcome for different patterns of hyperplasia and for different volume groups, Kruskal-Wallis test was performed and, and to compare results depending on the presence of AdBPH and the effect of bilateral embolization, Wilcoxon rank sum test was conducted.