New advances of the androgen receptor in prostate cancer: report from the 1st International Androgen Receptor Symposium.

The androgen receptor (AR) is a crucial player in various aspects of male reproduction and has been associated with the development and progression of prostate cancer (PCa). Therefore, the protein is the linchpin of current PCa therapies. Despite great research efforts, the AR signaling pathway has still not been deciphered, and the emergence of resistance is still the biggest problem in PCa treatment. To discuss the latest developments in AR research, the "1st International Androgen Receptor Symposium" offered a forum for the exchange of clinical and scientific innovations around the role of the AR in prostate cancer (PCa) and to stimulate new collaborative interactions among leading scientists from basic, translational, and clinical research. The symposium included three sessions covering preclinical studies, prognostic and diagnostic biomarkers, and ongoing prostate cancer clinical trials. In addition, a panel discussion about the future direction of androgen deprivation therapy and anti-AR therapy in PCa was conducted. Therefore, the newest insights and developments in therapeutic strategies and biomarkers are discussed in this report.

The Role of Multiparametric MRI (mpMRI) in the Prediction of Adverse Prostate Cancer Pathology in Radical Prostatectomy Specimen.

INTRODUCTION: Prostate cancer (PCa) risk stratification is essential in guiding therapeutic decision. Multiparametric magnetic resonance tomography (mpMRI) holds promise in the prediction of adverse pathologies (AP) after prostatectomy (RP). This study aims to identify clinical and imaging markers in the prediction of adverse pathology. METHODS: Patients with PCa, diagnosed by targeted biopsy after mpMRI and undergoing RP, were included. The predictive accuracy of mpMRI for extraprostatic extension (ECE), seminal vesicle infiltration (SVI), and lymph node positivity was calculated from the final histopathology. RESULTS: 846 patients were involved. Independent risk parameters include imaging findings such as ECE (OR 3.12), SVI (OR 2.55), and PI-RADS scoring (4: OR 2.01 and 5: OR 4.34). mpMRI parameters such as ECE, SVI, and lymph node metastases showed a high prognostic accuracy (73.28% vs. 95.35% vs. 93.38%) with moderate sensitivity compared to the final histopathology. The ROC analysis of our combined scoring system (D'Amico classification, PSA density, and MRI risk factors) improves the prediction of adverse pathology (AUC: 0.73 vs. 0.69). CONCLUSION: Our study supports the use of mpMRI for comprehensive pretreatment risk assessment in PCa. Due to the high accuracy of factors like ECE, SVI, and PI-RADS scoring, utilizing mpMRI data enabled accurate prediction of unfavorable pathology after RP.

MRI-Targeted, Systematic, and Combined Biopsy for Prostate Cancer Diagnosis.

BACKGROUND: The use of 12-core systematic prostate biopsy is associated with diagnostic inaccuracy that contributes to both overdiagnosis and underdiagnosis of prostate cancer. Biopsies performed with magnetic resonance imaging (MRI) targeting may reduce the misclassification of prostate cancer in men with MRI-visible lesions. METHODS: Men with MRI-visible prostate lesions underwent both MRI-targeted and systematic biopsy. The primary outcome was cancer detection according to grade group (i.e., a clustering of Gleason grades). Grade group 1 refers to clinically insignificant disease; grade group 2 or higher, cancer with favorable intermediate risk or worse; and grade group 3 or higher, cancer with unfavorable intermediate risk or worse. Among the men who underwent subsequent radical prostatectomy, upgrading and downgrading of grade group from biopsy to whole-mount histopathological analysis of surgical specimens were recorded. Secondary outcomes were the detection of cancers of grade group 2 or higher and grade group 3 or higher, cancer detection stratified by previous biopsy status, and grade reclassification between biopsy and radical prostatectomy. RESULTS: A total of 2103 men underwent both biopsy methods; cancer was diagnosed in 1312 (62.4%) by a combination of the two methods (combined biopsy), and 404 (19.2%) underwent radical prostatectomy. Cancer detection rates on MRI-targeted biopsy were significantly lower than on systematic biopsy for grade group 1 cancers and significantly higher for grade groups 3 through 5 (P<0.01 for all comparisons). Combined biopsy led to cancer diagnoses in 208 more men (9.9%) than with either method alone and to upgrading to a higher grade group in 458 men (21.8%). However, if only MRI-target biopsies had been performed, 8.8% of clinically significant cancers (grade group ≥3) would have been misclassified. Among the 404 men who underwent subsequent radical prostatectomy, combined biopsy was associated with the fewest upgrades to grade group 3 or higher on histopathological analysis of surgical specimens (3.5%), as compared with MRI-targeted biopsy (8.7%) and systematic biopsy (16.8%). CONCLUSIONS: Among patients with MRI-visible lesions, combined biopsy led to more detection of all prostate cancers. However, MRI-targeted biopsy alone underestimated the histologic grade of some tumors. After radical prostatectomy, upgrades to grade group 3 or higher on histopathological analysis were substantially lower after combined biopsy. (Funded by the National Institutes of Health and others; Trio Study ClinicalTrials.gov number, NCT00102544.).

Why Does Magnetic Resonance Imaging-Targeted Biopsy Miss Clinically Significant Cancer?

PURPOSE: Multiple studies demonstrate magnetic resonance imaging (MRI)-targeted biopsy detects more clinically significant cancer than systematic biopsy; however, some clinically significant cancers are detected by systematic biopsy only. While these events are rare, we sought to perform a retrospective analysis of these cases to ascertain the reasons that MRI-targeted biopsy missed clinically significant cancer which was subsequently detected on systematic prostate biopsy. MATERIALS AND METHODS: Patients were enrolled in a prospective study comparing cancer detection rates by transrectal MRI-targeted fusion biopsy and systematic 12-core biopsy. Patients with an elevated prostate specific antigen (PSA), abnormal digital rectal examination, or imaging findings concerning for prostate cancer underwent prostate MRI and subsequent MRI-targeted and systematic biopsy in the same setting. The subset of patients with grade group (GG) ≥3 cancer found on systematic biopsy and GG ≤2 cancer (or no cancer) on MRI-targeted biopsy was classified as MRI-targeted biopsy misses. A retrospective analysis of the MRI and MRI-targeted biopsy real-time screen captures determined the cause of MRI-targeted biopsy miss. Multivariable logistic regression analysis compared baseline characteristics of patients with MRI-targeted biopsy misses to GG-matched patients whose clinically significant cancer was detected by MRI-targeted biopsy. RESULTS: Over the study period of 2007 to 2019, 2,103 patients met study inclusion criteria and underwent combined MRI-targeted and systematic prostate biopsies. A total of 41 (1.9%) men were classified as MRI-targeted biopsy misses. Most MRI-targeted biopsy misses were due to errors in lesion targeting (21, 51.2%), followed by MRI-invisible lesions (17, 40.5%) and MRI lesions missed by the radiologist (3, 7.1%). On logistic regression analysis, lower Prostate Imaging-Reporting and Data System (PI-RADSTM) score was associated with having clinically significant cancer missed on MRI-targeted biopsy. CONCLUSIONS: While uncommon, most MRI-targeted biopsy misses are due to errors in lesion targeting, which highlights the importance of accurate co-registration and targeting when using software-based fusion platforms. Additionally, some patients will harbor MRI-invisible lesions which are untargetable by MRI-targeted platforms. The presence of a low PI-RADS score despite a high PSA is suggestive of harboring an MRI-invisible lesion.

Prognostic Features of Biochemical Recurrence of Prostate Cancer Following Radical Prostatectomy Based on Multiparametric MRI and Immunohistochemistry Analysis of MRI-guided Biopsy Specimens.

Background Although prostate MRI is routinely used for the detection and staging of localized prostate cancer, imaging-based assessment and targeted molecular sampling for risk stratification are an active area of research. Purpose To evaluate features of preoperative MRI and MRI-guided biopsy immunohistochemistry (IHC) findings associated with biochemical recurrence (BCR) of prostate cancer after surgery. Materials and Methods In this retrospective case-control study, patients underwent multiparametric MRI before MRI-guided biopsy followed by radical prostatectomy between 2008 and 2016. Lesions were retrospectively scored with the Prostate Imaging Reporting and Data System (PI-RADS) (version 2) by radiologists who were blinded to the clinical-pathologic results. The IHC staining, including stains for the ETS-related gene, phosphatase and tensin homolog, androgen receptor, prostate specific antigen, and p53, was performed with targeted biopsy specimens of the index lesion (highest suspicion at MRI and pathologic grade) and scored by pathologists who were blinded to clinical-pathologic outcomes. Cox proportional hazards regression analysis was used to evaluate associations with recurrence-free survival (RFS). Results The median RFS was 31.7 months (range, 1-101 months) for 39 patients (median age, 62 years; age range, 47-76 years) without BCR and 14.6 months (range, 1-61 months) for 40 patients (median age, 59 years; age range, 47-73 years) with BCR. MRI features that showed a significant relationship with the RFS interval included an index lesion with a PI-RADS score of 5 (hazard ratio [HR], 2.10; 95% CI: 1.05, 4.21; P = .04); index lesion burden, defined as ratio of index lesion volume to prostate volume (HR, 1.55; 95% CI: 1.2, 2.1; P = .003); and suspicion of extraprostatic extension (EPE) (HR, 2.18; 95% CI: 1.1, 4.2; P = .02). Presurgical multivariable analysis indicated that suspicion of EPE at MRI (adjusted HR, 2.19; 95% CI: 1.1, 4.3; P = .02) and p53 stain intensity (adjusted HR, 2.22; 95% CI: 1.0, 4.7; P = .04) were significantly associated with RFS. Conclusion MRI features, including Prostate Imaging Reporting and Data System score, index lesion burden, extraprostatic extension, and preoperative guided biopsy p53 immunohistochemistry stain intensity are associated with biochemical relapse of prostate cancer after surgery. © RSNA, 2021 Online supplemental material is available for this article. See also the editorial by Costa in this issue.

High throughput assessment of biomarkers in tissue microarrays using artificial intelligence: PTEN loss as a proof-of-principle in multi-center prostate cancer cohorts.

Phosphatase and tensin homolog (PTEN) loss is associated with adverse outcomes in prostate cancer and has clinical potential as a prognostic biomarker. The objective of this work was to develop an artificial intelligence (AI) system for automated detection and localization of PTEN loss on immunohistochemically (IHC) stained sections. PTEN loss was assessed using IHC in two prostate tissue microarrays (TMA) (internal cohort, n = 272 and external cohort, n = 129 patients). TMA cores were visually scored for PTEN loss by pathologists and, if present, spatially annotated. Cores from each patient within the internal TMA cohort were split into 90% cross-validation (N = 2048) and 10% hold-out testing (N = 224) sets. ResNet-101 architecture was used to train core-based classification using a multi-resolution ensemble approach (×5, ×10, and ×20). For spatial annotations, single resolution pixel-based classification was trained from patches extracted at ×20 resolution, interpolated to ×40 resolution, and applied in a sliding-window fashion. A final AI-based prediction model was created from combining multi-resolution and pixel-based models. Performance was evaluated in 428 cores of external cohort. From both cohorts, a total of 2700 cores were studied, with a frequency of PTEN loss of 14.5% in internal (180/1239) and external 13.5% (43/319) cancer cores. The final AI-based prediction of PTEN status demonstrated 98.1% accuracy (95.0% sensitivity, 98.4% specificity; median dice score = 0.811) in internal cohort cross-validation set and 99.1% accuracy (100% sensitivity, 99.0% specificity; median dice score = 0.804) in internal cohort test set. Overall core-based classification in the external cohort was significantly improved in the external cohort (area under the curve = 0.964, 90.6% sensitivity, 95.7% specificity) when further trained (fine-tuned) using 15% of cohort data (19/124 patients). These results demonstrate a robust and fully automated method for detection and localization of PTEN loss in prostate cancer tissue samples. AI-based algorithms have potential to streamline sample assessment in research and clinical laboratories.

Magnetic Resonance Imaging-Targeted and Systematic Biopsy for Detection of Grade Progression in Patients on Active Surveillance for Prostate Cancer.

PURPOSE: Active surveillance for patients with low and intermediate risk prostate cancers is becoming a more utilized option in recent years. However, the use of magnetic resonance imaging and imaging-targeted biopsy for monitoring grade progression has been poorly studied in this population. We aim to define the utility of magnetic resonance imaging-targeted biopsy and systematic biopsy in an active surveillance population. MATERIALS AND METHODS: Between July 2007 and January 2020, patients with diagnosed prostate cancer who elected active surveillance were monitored with prostate magnetic resonance imaging, imaging-targeted biopsy and standard systematic biopsy. Patients were eligible for surveillance if diagnosed with any volume Gleason grade 1 disease and select Gleason grade 2 disease. Grade progression (Gleason grade 1 to ≥2 disease and Gleason grade 2 to ≥3 disease) for each biopsy modality was measured at 2 years, 4 years and 6+ years. RESULTS: In total, 369 patients had both magnetic resonance imaging-targeted and systematic biopsy and were surveilled for at least 1 year. At 2 years, systematic biopsy, magnetic resonance imaging-targeted biopsy and combined biopsy (systematic+imaging-targeted) detected grade progression in 44 patients (15.9%), 73 patients (26.4%) and 90 patients (32.5%), respectively. Magnetic resonance imaging-targeted biopsy detected more cancer grade progression compared to systematic biopsy in both the low and intermediate risk populations (p <0.001). Of all 90 grade progressions at the 2-year time point 46 (51.1%) were found by magnetic resonance imaging-targeted biopsy alone and missed by systematic biopsy. CONCLUSIONS: Magnetic resonance imaging-targeted biopsy detected significantly more grade progressions in our active surveillance cohort compared to systematic biopsy at 2 years. Our results provide compelling evidence that prostate magnetic resonance imaging and imaging-targeted biopsy should be included in contemporary active surveillance protocols.

The Risk of Prostate Cancer Progression in Active Surveillance Patients with Bilateral Disease Detected by Combined Magnetic Resonance Imaging-Fusion and Systematic Biopsy.

PURPOSE: We sought to evaluate whether bilateral prostate cancer detected at active surveillance (AS) enrollment is associated with progression to Grade Group (GG) ≥2 and to compare the efficacy of combined targeted biopsy plus systematic biopsy (Cbx) vs systematic biopsy (Sbx) or targeted biopsy alone to detect bilateral disease. MATERIALS AND METHODS: A prospectively maintained database of patients referred to our institution from 2007-2020 was queried. The study cohort included all AS patients with GG1 on confirmatory Cbx and followup of at least 1 year. Cox proportional hazard analysis identified baseline characteristics associated with progression to ≥GG2 at any point throughout followup. RESULTS: Of 579 patients referred, 103 patients had GG1 on Cbx and were included in the study; 49/103 (47.6%) patients progressed to ≥GG2, with 30/72 (41.7%) patients with unilateral disease progressing and 19/31 (61.3%) patients with bilateral disease progressing. Median time to progression was 68 months vs 52 months for unilateral and bilateral disease, respectively (p=0.006). Both prostate specific antigen density (HR 1.72, p=0.005) and presence of bilateral disease (HR 2.21, p=0.012) on confirmatory biopsy were associated with AS progression. At time of progression, GG and risk group were significantly higher in patients with bilateral versus unilateral disease. Cbx detected 16% more patients with bilateral disease than Sbx alone. CONCLUSIONS: Bilateral disease and prostate specific antigen density at confirmatory Cbx conferred greater risk of earlier AS progression. Cbx was superior to Sbx for identifying bilateral disease. AS risk-stratification protocols may benefit from including presence of bilateral disease and should use Cbx to detect bilateral disease.

Making a case "for" focal therapy of the prostate in intermediate risk prostate cancer: current perspective and ongoing trials.

Focal therapy is growing as an alternative management options for men with clinically localized prostate cancer. Parallel to the increasing popularity of active surveillance (AS) as a treatment for low-risk disease, there has been an increased interest towards providing focal therapy for patients with intermediate-risk disease. Focal therapy can act as a logical "middle ground" in patients who seek treatment while minimizing potential side effects of definitive whole-gland treatment. The aim of the current review is to define the rationale of focal therapy in patients with intermediate-risk prostate cancer and highlight the importance of patient selection in focal therapy candidacy.

PI-RADS® Category as a Predictor of Progression to Unfavorable Risk Prostate Cancer in Men on Active Surveillance.

PURPOSE: We identified baseline imaging and clinical characteristics of patients that may improve risk stratification among patients being evaluated for active surveillance. MATERIALS AND METHODS: From July 2007 to January 2020 patients referred to our institution for prostate cancer were evaluated and those who remained on active surveillance were identified. Men underwent multiparametric magnetic resonance imaging upon entry into our active surveillance protocol during which baseline demographic and imaging data were documented. Patients were then followed and outcomes, specifically progression to Gleason Grade Group (GG)3 or greater disease, were recorded. RESULTS: Of the men placed on active surveillance 344 had at least 1 PI-RADS score documented. For those with an index lesion PI-RADS category of 5, 33% (17/51) had progression to GG3 or greater on active surveillance with a median time to progression of 31 months. When comparing the progression-free survival times and progression rates in each category, PI-RADS category was found to be associated with progression to GG3 or greater on active surveillance (p <0.01). On univariable analysis factors associated with progression included an index lesion PI-RADS category of 5, prostate specific antigen density and the size of the largest lesion. On multivariable analysis only PI-RADS category of 5 and prostate specific antigen density were associated with progression on active surveillance. CONCLUSIONS: PI-RADS lesion categories at baseline multiparametric magnetic resonance imaging during active surveillance enrollment can be used to predict cancer progression to GG3 or greater on active surveillance. This information, along with other clinical data, can better assist urologists in identifying and managing patients appropriate for active surveillance.

Deep-Learning-Based Artificial Intelligence for PI-RADS Classification to Assist Multiparametric Prostate MRI Interpretation: A Development Study.

BACKGROUND: The Prostate Imaging Reporting and Data System (PI-RADS) provides guidelines for risk stratification of lesions detected on multiparametric MRI (mpMRI) of the prostate but suffers from high intra/interreader variability. PURPOSE: To develop an artificial intelligence (AI) solution for PI-RADS classification and compare its performance with an expert radiologist using targeted biopsy results. STUDY TYPE: Retrospective study including data from our institution and the publicly available ProstateX dataset. POPULATION: In all, 687 patients who underwent mpMRI of the prostate and had one or more detectable lesions (PI-RADS score >1) according to PI-RADSv2. FIELD STRENGTH/SEQUENCE: T2 -weighted, diffusion-weighted imaging (DWI; five evenly spaced b values between b = 0-750 s/mm2 ) for apparent diffusion coefficient (ADC) mapping, high b-value DWI (b = 1500 or 2000 s/mm2 ), and dynamic contrast-enhanced T1 -weighted series were obtained at 3.0T. ASSESSMENT: PI-RADS lesions were segmented by a radiologist. Bounding boxes around the T2 /ADC/high-b value segmentations were stacked and saved as JPEGs. These images were used to train a convolutional neural network (CNN). The PI-RADS scores obtained by the CNN were compared with radiologist scores. The cancer detection rate was measured from a subset of patients who underwent biopsy. STATISTICAL TESTS: Agreement between the AI and the radiologist-driven PI-RADS scores was assessed using a kappa score, and differences between categorical variables were assessed with a Wald test. RESULTS: For the 1034 detection lesions, the kappa score for the AI system vs. the expert radiologist was moderate, at 0.40. However, there was no significant difference in the rates of detection of clinically significant cancer for any PI-RADS score in 86 patients undergoing targeted biopsy (P = 0.4-0.6). DATA CONCLUSION: We developed an AI system for assignment of a PI-RADS score on segmented lesions on mpMRI with moderate agreement with an expert radiologist and a similar ability to detect clinically significant cancer. LEVEL OF EVIDENCE: 4 TECHNICAL EFFICACY STAGE: 2.

Prospective Evaluation of PI-RADS Version 2.1 for Prostate Cancer Detection.

OBJECTIVE. The purpose of this study was to prospectively evaluate Prostate Imaging Reporting and Data and System version 2.1 (PI-RADSv2.1), which was released in March 2019 to update version 2.0, for prostate cancer detection with transrectal ultrasound-MRI fusion biopsy and 12-core systematic biopsy. SUBJECTS AND METHODS. This prospective study included 110 consecutively registered patients who underwent multiparametric MRI evaluated with PI-RADSv2.1 criteria followed by fusion biopsy and systematic biopsy between April and September 2019. Lesion-based cancer detection rates (CDRs) were calculated for prostate cancer (Gleason grade group, > 0) and clinically significant prostate cancer (Gleason grade group, > 1). RESULTS. A total of 171 lesions (median size, 1.1 cm) in 110 patients were detected and evaluated with PI-RADSv2.1. In 16 patients no lesion was detected, and only systematic biopsy was performed. Lesions were categorized as follows: PI-RADS category 1, 1 lesion; PI-RADS category 2, 34 lesions; PI-RADS category 3, 54 lesions; PI-RADS category 4, 52 lesions; and PI-RADS category 5, 30 lesions. Histopathologic analysis revealed prostate cancer in 74 of 171 (43.3%) lesions and clinically significant prostate cancer in 57 of 171 (33.3%) lesions. The CDRs of prostate cancer for PI-RADS 2, 3, 4, and 5 lesions were 20.0%, 24.1%, 51.9%, and 90.0%. The CDRs of clinically significant prostate cancer for PI-RADS 1, 2, 3, 4, and 5 lesions were 0%, 5.7%, 14.8%, 44.2%, and 80.0%. In 16 patients with normal multiparametric MRI findings (PI-RADS 1), the CDRs were 50.0% for PCa and 18.8% for clinically significant prostate cancer. CONCLUSION. This investigation yielded CDRs assessed with prospectively assigned PI-RADSv2.1 scores. CDRs increased with higher PI-RADSv2.1 scores. These results can be compared with previously published outcomes derived with PI-RADS version 2.0.

Prospective Evaluation of 18F-DCFPyL PET/CT in Detection of High-Risk Localized Prostate Cancer: Comparison With mpMRI.

OBJECTIVE. The purpose of this study was to assess the utility of PET with (2S)-2-[[(1S)-1-carboxy-5-[(6-(18F)fluoranylpyridine-3-carbonyl)amino]pentyl]carbamoylamino]pentanedioic acid (18F-DCFPyL), a prostate-specific membrane antigen (PSMA)-targeted radiotracer, in the detection of high-risk localized prostate cancer as compared with multiparametric MRI (mpMRI). SUBJECTS AND METHODS. This HIPAA-compliant prospective study included 26 consecutive patients with localized high-risk prostate cancer (median age, 69.5 years [range, 53-81 years]; median prostate-specific antigen [PSA] level, 18.88 ng/mL [range, 1.03-20.00 ng/mL]) imaged with 18F-DCFPyL PET/CT and mpMRI. Images from PET/CT and mpMRI were evaluated separately, and suspicious areas underwent targeted biopsy. Lesion-based sensitivity and tumor detection rate were compared for PSMA PET and mpMRI. Standardized uptake value (SUV) and PSMA PET parameters were correlated with histopathology score, and uptake in tumor was compared with that in nonmalignant tissue. On a patient level, SUV and PSMA tumor volume were correlated with PSA density. RESULTS. Forty-four tumors (one in Gleason grade [GG] group 1, 12 in GG group 2, seven in GG group 3, nine in GG group 4, and 15 in GG group 5) were identified at histopathology. Sensitivity and tumor detection rate of 18F-DCFPyL PET/CT and mpMRI were similar (PET/CT, 90.9% and 80%; mpMRI, 86.4% and 88.4%; p = 0.58/0.17). Total lesion PSMA and PSMA tumor volume showed a relationship with GG (τ = 0.27 and p = 0.08, τ = 0.30 and p = 0.06, respectively). Maximum SUV in tumor was significantly higher than that in nonmalignant tissue (p < 0.05). Tumor burden density moderately correlated with PSA density (r = 0.47, p = 0.01). Five true-positive tumors identified on 18F-DCFPyL PET/CT were not identified on mpMRI. CONCLUSION. In patients with high-risk prostate cancer, 18F-DCFPyL PET/CT is highly sensitive in detecting intraprostatic tumors and can detect tumors missed on mpMRI. Measured uptake is significantly higher in tumor tissue, and PSMA-derived tumor burden is associated with severity of disease.

Data Augmentation and Transfer Learning to Improve Generalizability of an Automated Prostate Segmentation Model.

OBJECTIVE. Deep learning applications in radiology often suffer from overfitting, limiting generalization to external centers. The objective of this study was to develop a high-quality prostate segmentation model capable of maintaining a high degree of performance across multiple independent datasets using transfer learning and data augmentation. MATERIALS AND METHODS. A retrospective cohort of 648 patients who underwent prostate MRI between February 2015 and November 2018 at a single center was used for training and validation. A deep learning approach combining 2D and 3D architecture was used for training, which incorporated transfer learning. A data augmentation strategy was used that was specific to the deformations, intensity, and alterations in image quality seen on radiology images. Five independent datasets, four of which were from outside centers, were used for testing, which was conducted with and without fine-tuning of the original model. The Dice similarity coefficient was used to evaluate model performance. RESULTS. When prostate segmentation models utilizing transfer learning were applied to the internal validation cohort, the mean Dice similarity coefficient was 93.1 for whole prostate and 89.0 for transition zone segmentations. When the models were applied to multiple test set cohorts, the improvement in performance achieved using data augmentation alone was 2.2% for the whole prostate models and 3.0% for the transition zone segmentation models. However, the best test-set results were obtained with models fine-tuned on test center data with mean Dice similarity coefficients of 91.5 for whole prostate segmentation and 89.7 for transition zone segmentation. CONCLUSION. Transfer learning allowed for the development of a high-performing prostate segmentation model, and data augmentation and fine-tuning approaches improved performance of a prostate segmentation model when applied to datasets from external centers.

Multicenter Multireader Evaluation of an Artificial Intelligence-Based Attention Mapping System for the Detection of Prostate Cancer With Multiparametric MRI.

OBJECTIVE. The purpose of this study was to evaluate in a multicenter dataset the performance of an artificial intelligence (AI) detection system with attention mapping compared with multiparametric MRI (mpMRI) interpretation in the detection of prostate cancer. MATERIALS AND METHODS. MRI examinations from five institutions were included in this study and were evaluated by nine readers. In the first round, readers evaluated mpMRI studies using the Prostate Imaging Reporting and Data System version 2. After 4 weeks, images were again presented to readers along with the AI-based detection system output. Readers accepted or rejected lesions within four AI-generated attention map boxes. Additional lesions outside of boxes were excluded from detection and categorization. The performances of readers using the mpMRI-only and AI-assisted approaches were compared. RESULTS. The study population included 152 case patients and 84 control patients with 274 pathologically proven cancer lesions. The lesion-based AUC was 74.9% for MRI and 77.5% for AI with no significant difference (p = 0.095). The sensitivity for overall detection of cancer lesions was higher for AI than for mpMRI but did not reach statistical significance (57.4% vs 53.6%, p = 0.073). However, for transition zone lesions, sensitivity was higher for AI than for MRI (61.8% vs 50.8%, p = 0.001). Reading time was longer for AI than for MRI (4.66 vs 4.03 minutes, p < 0.001). There was moderate interreader agreement for AI and MRI with no significant difference (58.7% vs 58.5%, p = 0.966). CONCLUSION. Overall sensitivity was only minimally improved by use of the AI system. Significant improvement was achieved, however, in the detection of transition zone lesions with use of the AI system at the cost of a mean of 40 seconds of additional reading time.

A Grading System for the Assessment of Risk of Extraprostatic Extension of Prostate Cancer at Multiparametric MRI.

Purpose To evaluate MRI features associated with pathologically defined extraprostatic extension (EPE) of prostate cancer and to propose an MRI grading system for pathologic EPE. Materials and Methods In this prospective study, consecutive male study participants underwent preoperative 3.0-T MRI from June 2007 to March 2017 followed by robotic-assisted laparoscopic radical prostatectomy. An MRI-based EPE grading system was defined as follows: curvilinear contact length of 1.5 cm or capsular bulge and irregularity were grade 1, both features were grade 2, and frank capsular breach were grade 3. Multivariable logistic regression and decision curve analyses were performed to compare the MRI grade model and clinical parameters (prostate-specific antigen, Gleason score) for pathologic EPE prediction by using the area under the receiver operating characteristic curve (AUC) value. Results Among 553 study participants, the mean age was 60 years ± 8 (standard deviation); the median prostate-specific antigen value was 6.3 ng/mL. A total of 125 of 553 (22%) participants had pathologic EPE at radical prostatectomy. Detection of pathologic EPE, defined as number of pathologic EPEs divided by number of participants with individual MRI features, was as follows: curvilinear contact length, 88 of 208 (42%); capsular bulge and irregularity, 78 of 175 (45%); and EPE visible at MRI, 37 of 56 (66%). For MRI, grades 1, 2, and 3 for detection of pathologic EPE were 18 of 74 (24%), 39 of 102 (38%), and 37 of 56 (66%), respectively. Clinical features plus the MRI-based EPE grading system (prostate-specific antigen, International Society of Urological Pathology stage, MRI grade) predicted pathologic EPE better than did MRI grade alone (AUC, 0.81 vs 0.77, respectively; P < .001). Conclusion Higher MRI-based extraprostatic extension (EPE) grading categories were associated with a greater risk of pathologic EPE. Clinical features plus MRI grading had the highest diagnostic performance for prediction of pathologic EPE. © RSNA, 2019 Online supplemental material is available for this article. See also the editorial by Eberhardt in this issue.

Predicting Gleason Group Progression for Men on Prostate Cancer Active Surveillance: Role of a Negative Confirmatory Magnetic Resonance Imaging-Ultrasound Fusion Biopsy.

PURPOSE: Active surveillance has gained acceptance as an alternative to definitive therapy in many men with prostate cancer. Confirmatory biopsies to assess the appropriateness of active surveillance are routinely performed and negative biopsies are regarded as a favorable prognostic indicator. We sought to determine the prognostic implications of negative multiparametric magnetic resonance imaging-transrectal ultrasound guided fusion biopsy consisting of extended sextant, systematic biopsy plus multiparametric magnetic resonance imaging guided targeted biopsy of suspicious lesions on magnetic resonance imaging. MATERIALS AND METHODS: All patients referred with Gleason Grade Group 1 or 2 prostate cancer based on systematic biopsy performed elsewhere underwent confirmatory fusion biopsy. Patients who continued on active surveillance after a positive or a negative fusion biopsy were followed. The baseline characteristics of the biopsy negative and positive cases were compared. Cox regression analysis was used to determine the prognostic significance of a negative fusion biopsy. Kaplan-Meier survival curves were used to estimate Grade Group progression with time. RESULTS: Of the 542 patients referred with Grade Group 1 (466) or Grade Group 2 (76) cancer 111 (20.5%) had a negative fusion biopsy. A total of 60 vs 122 patients with a negative vs a positive fusion biopsy were followed on active surveillance with a median time to Grade Group progression of 74.3 and 44.6 months, respectively (p <0.01). Negative fusion biopsy was associated with a reduced risk of Grade Group progression (HR 0.41, 95% CI 0.22-0.77, p <0.01). CONCLUSIONS: A negative confirmatory fusion biopsy confers a favorable prognosis for Grade Group progression. These results can be used when counseling patients about the risk of progression and for planning future followup and biopsies in patients on active surveillance.

Intra- and interreader reproducibility of PI-RADSv2: A multireader study.

BACKGROUND: The Prostate Imaging Reporting and Data System version 2 (PI-RADSv2) has been in use since 2015; while interreader reproducibility has been studied, there has been a paucity of studies investigating the intrareader reproducibility of PI-RADSv2. PURPOSE: To evaluate both intra- and interreader reproducibility of PI-RADSv2 in the assessment of intraprostatic lesions using multiparametric magnetic resonance imaging (mpMRI). STUDY TYPE: Retrospective. POPULATION/SUBJECTS: In all, 102 consecutive biopsy-naïve patients who underwent prostate MRI and subsequent MR/transrectal ultrasonography (MR/TRUS)-guided biopsy. FIELD STRENGTH/SEQUENCES: Prostate mpMRI at 3T using endorectal with phased array surface coils (TW MRI, DW MRI with ADC maps and b2000 DW MRI, DCE MRI). ASSESSMENT: Previously detected and biopsied lesions were scored by four readers from four different institutions using PI-RADSv2. Readers scored lesions during two readout rounds with a 4-week washout period. STATISTICAL TESTS: Kappa (κ) statistics and specific agreement (Po ) were calculated to quantify intra- and interreader reproducibility of PI-RADSv2 scoring. Lesion measurement agreement was calculated using the intraclass correlation coefficient (ICC). RESULTS: Overall intrareader reproducibility was moderate to substantial (κ = 0.43-0.67, Po = 0.60-0.77), while overall interreader reproducibility was poor to moderate (κ = 0.24, Po = 46). Readers with more experience showed greater interreader reproducibility than readers with intermediate experience in the whole prostate (P = 0.026) and peripheral zone (P = 0.002). Sequence-specific interreader agreement for all readers was similar to the overall PI-RADSv2 score, with κ = 0.24, 0.24, and 0.23 and Po = 0.47, 0.44, and 0.54 in T2 -weighted, diffusion-weighted imaging (DWI), and dynamic contrast-enhanced (DCE), respectively. Overall intrareader and interreader ICC for lesion measurement was 0.82 and 0.71, respectively. DATA CONCLUSION: PI-RADSv2 provides moderate intrareader reproducibility, poor interreader reproducibility, and moderate interreader lesion measurement reproducibility. These findings suggest a need for more standardized reader training in prostate MRI. LEVEL OF EVIDENCE: 2 Technical Efficacy: Stage 2.

Use of multiparametric magnetic resonance imaging and fusion-guided biopsies to properly select and follow African-American men on active surveillance.

OBJECTIVES: To determine the rate of Gleason Grade Group (GGG) upgrading in African-American (AA) men with a prior diagnosis of low-grade prostate cancer (GGG 1 or GGG 2) on 12-core systematic biopsy (SB) after multiparametric magnetic resonance imaging (mpMRI) and fusion biopsy (FB); and whether AA men who continued active surveillance (AS) after mpMRI and FB fared differently than a predominantly Caucasian (non-AA) population. PATIENTS AND METHODS: A database of men who had undergone mpMRI and FB was queried to determine rates of upgrading by FB amongst men deemed to be AS candidates based on SB prior to referral. After FB, Kaplan-Meier curves were generated for AA men and non-AA men who then elected AS. The time to GGG upgrading and time continuing AS were compared using the log-rank test. RESULTS: AA men referred with GGG 1 disease on previous SB were upgraded to GGG ≥3 by FB more often than non-AA men, 22.2% vs 12.7% (P = 0.01). A total of 32 AA men and 258 non-AA men then continued AS, with a median (interquartile range) follow-up of 39.19 (24.24-56.41) months. The median time to progression was 59.7 and 60.5 months, respectively (P = 0.26). The median time continuing AS was 61.9 months and not reached, respectively (P = 0.80). CONCLUSIONS: AA men were more likely to be upgraded from GGG 1 on SB to GGG ≥3 on initial FB; however, AA and non-AA men on AS subsequently progressed at similar rates following mpMRI and FB. A greater tendency for SB to underestimate tumour grade in AA men may explain prior studies that have shown AA men to be at higher risk of progression during AS.

Comparison of Elastic and Rigid Registration during Magnetic Resonance Imaging/Ultrasound Fusion-Guided Prostate Biopsy: A Multi-Operator Phantom Study.

PURPOSE: The relative value of rigid or elastic registration during magnetic resonance imaging/ultrasound fusion guided prostate biopsy has been poorly studied. We compared registration errors (the distance between a region of interest and fiducial markers) between rigid and elastic registration during fusion guided prostate biopsy using a prostate phantom model. MATERIALS AND METHODS: Four gold fiducial markers visible on magnetic resonance imaging and ultrasound were placed throughout 1 phantom prostate model. The phantom underwent magnetic resonance imaging and the fiducial markers were labeled as regions of interest. An experienced user and a novice user of fusion guided prostate biopsy targeted regions of interest and then the corresponding fiducial markers on ultrasound after rigid and then elastic registration. Registration errors were compared. RESULTS: A total of 224 registration error measurements were recorded. Overall elastic registration did not provide significantly improved registration error over rigid registration (mean ± SD 4.87 ± 3.50 vs 4.11 ± 2.09 mm, p = 0.05). However, lesions near the edge of the phantom showed increased registration errors when using elastic registration (5.70 ± 3.43 vs 3.23 ± 1.68 mm, p = 0.03). Compared to the novice user the experienced user reported decreased registration error with rigid registration (3.25 ± 1.49 vs 4.98 ± 2.10 mm, p <0.01) and elastic registration (3.94 ± 2.61 vs 6.07 ± 4.16 mm, p <0.01). CONCLUSIONS: We found no difference in registration errors between rigid and elastic registration overall but rigid registration decreased the registration error of targets near the prostate edge. Additionally, operator experience reduced registration errors regardless of the registration method. Therefore, elastic registration algorithms cannot serve as a replacement for attention to detail during the registration process and anatomical landmarks indicating accurate registration when beginning the procedure and before targeting each region of interest.