External Validation of a Digital Pathology-based Multimodal Artificial Intelligence Architecture in the NRG/RTOG 9902 Phase 3 Trial.

BACKGROUND: Accurate risk stratification is critical to guide management decisions in localized prostate cancer (PCa). Previously, we had developed and validated a multimodal artificial intelligence (MMAI) model generated from digital histopathology and clinical features. Here, we externally validate this model on men with high-risk or locally advanced PCa treated and followed as part of a phase 3 randomized control trial. OBJECTIVE: To externally validate the MMAI model on men with high-risk or locally advanced PCa treated and followed as part of a phase 3 randomized control trial. DESIGN, SETTING, AND PARTICIPANTS: Our validation cohort included 318 localized high-risk PCa patients from NRG/RTOG 9902 with available histopathology (337 [85%] of the 397 patients enrolled into the trial had available slides, of which 19 [5.6%] failed due to poor image quality). OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS: Two previously locked prognostic MMAI models were validated for their intended endpoint: distant metastasis (DM) and PCa-specific mortality (PCSM). Individual clinical factors and the number of National Comprehensive Cancer Network (NCCN) high-risk features served as comparators. Subdistribution hazard ratio (sHR) was reported per standard deviation increase of the score with corresponding 95% confidence interval (CI) using Fine-Gray or Cox proportional hazards models. RESULTS AND LIMITATIONS: The DM and PCSM MMAI algorithms were significantly and independently associated with the risk of DM (sHR [95% CI] = 2.33 [1.60-3.38], p < 0.001) and PCSM, respectively (sHR [95% CI] = 3.54 [2.38-5.28], p < 0.001) when compared against other prognostic clinical factors and NCCN high-risk features. The lower 75% of patients by DM MMAI had estimated 5- and 10-yr DM rates of 4% and 7%, and the highest quartile had average 5- and 10-yr DM rates of 19% and 32%, respectively (p < 0.001). Similar results were observed for the PCSM MMAI algorithm. CONCLUSIONS: We externally validated the prognostic ability of MMAI models previously developed among men with localized high-risk disease. MMAI prognostic models further risk stratify beyond the clinical and pathological variables for DM and PCSM in a population of men already at a high risk for disease progression. This study provides evidence for consistent validation of our deep learning MMAI models to improve prognostication and enable more informed decision-making for patient care. PATIENT SUMMARY: This paper presents a novel approach using images from pathology slides along with clinical variables to validate artificial intelligence (computer-generated) prognostic models. When implemented, clinicians can offer a more personalized and tailored prognostic discussion for men with localized prostate cancer.

Prostate biopsy histopathologic features correlate with a commercial gene expression assay's reclassification of patient NCCN risk category.

BACKGROUND: For biopsies with Gleason 3 + 3 = 6 or 3 + 4 = 7 prostate cancer, the Genomic Prostate Score (GPS; OncotypeDx) is designed to predict severe pathology at prostatectomy, and, in some cases, recommends reclassification of the National Comprehensive Cancer Network (NCCN) risk category. We hypothesized that certain histopathologic features that were not considered in the original design of the assay actually would be associated with the NCCN risk category change indicated by GPS testing. METHODS: For patients with recommended NCCN risk category change, the biopsy cores used for GPS were re-reviewed for stromal reaction, chronic inflammation, and tumor nuclear polarization. RESULTS: Of 520 patients from May 2011 to December 2018, GPS testing suggested NCCN risk reclassification in 131 (25%); 127 of these slides were available. Of these, the NCCN risk category increased from intermediate to high in 8, low to intermediate in 15, very low to low in 1, and decreased from intermediate to low in 32, and low to very low in 71. Biopsies with NCCN risk increase were associated with moderate or severe stromal reaction (p < .001) and chronic inflammation (p < .001); biopsies with NCCN risk decrease were associated with absence of these features. In Gleason 3 + 3 = 6 cases (n = 93), presence of nuclear polarization was associated with NCCN risk decrease and its absence with increase (p < .001). CONCLUSIONS: Moderate or severe stromal reaction, chronic inflammation, and lack of nuclear polarization in Gleason score 3 + 3 = 6 tumors were each associated with an increase in NCCN risk category indicated by GPS and vice versa. Our results suggest that GPS captures histologic features associated with aggressiveness that are not routinely assessed in standard histopathologic assessments, and that consideration of such histologic features may improve upon current tumor grading approaches.

Decipher identifies men with otherwise clinically favorable-intermediate risk disease who may not be good candidates for active surveillance.

BACKGROUND: We aimed to validate Decipher to predict adverse pathology (AP) at radical prostatectomy (RP) in men with National Comprehensive Cancer Network (NCCN) favorable-intermediate risk (F-IR) prostate cancer (PCa), and to better select F-IR candidates for active surveillance (AS). METHODS: In all, 647 patients diagnosed with NCCN very low/low risk (VL/LR) or F-IR prostate cancer were identified from a multi-institutional PCa biopsy database; all underwent RP with complete postoperative clinicopathological information and Decipher genomic risk scores. The performance of all risk assessment tools was evaluated using logistic regression model for the endpoint of AP, defined as grade group 3-5, pT3b or higher, or lymph node invasion. RESULTS: The median age was 61 years (interquartile range 56-66) for 220 patients with NCCN F-IR disease, 53% classified as low-risk by Cancer of the Prostate Risk Assessment (CAPRA 0-2) and 47% as intermediate-risk (CAPRA 3-5). Decipher classified 79%, 13% and 8% of men as low-, intermediate- and high-risk with 13%, 10%, and 41% rate of AP, respectively. Decipher was an independent predictor of AP with an odds ratio of 1.34 per 0.1 unit increased (p value = 0.002) and remained significant when adjusting by CAPRA. Notably, F-IR with Decipher low or intermediate score did not associate with significantly higher odds of AP compared to VL/LR. CONCLUSIONS: NCCN risk groups, including F-IR, are highly heterogeneous and should be replaced with multivariable risk-stratification. In particular, incorporating Decipher may be useful for safely expanding the use of AS in this patient population.

Management of Inverted Papilloma During Holmium Laser Enucleation of the Prostate.

Inverted papilloma of the prostatic urethra is an especially rare finding. A 75-year-old man with urinary retention wished to proceed with a holmium laser enucleation of the prostate (HoLEP) and was found to have a mass arising from his prostate vs bladder on preoperative imaging. Cystourethroscopy revealed the mass arising from the median lobe of the prostate. After transurethral resection and frozen analysis confirmed the benign pathology of an inverted papilloma, the patient subsequently underwent a successful HoLEP during the same surgical setting. Images of this rare prostatic mass are presented to increase urologist recognition and to assist management during HoLEP.

18F Fluorocholine Dynamic Time-of-Flight PET/MR Imaging in Patients with Newly Diagnosed Intermediate- to High-Risk Prostate Cancer: Initial Clinical-Pathologic Comparisons.

Purpose To investigate the initial clinical value of fluorine 18 (18F) fluorocholine (FCH) dynamic positron emission tomography (PET)/magnetic resonance (MR) imaging by comparing its parameters with clinical-pathologic findings in patients with newly diagnosed intermediate- to high-risk prostate cancer (PCa) who plan to undergo radical prostatectomy. Materials and Methods The institutional review board approved the study protocol, and informed written consent was obtained from all subjects for this HIPAA-compliant study. Twelve men (mean age ± standard deviation, 61.7 years ± 8.4; range, 46-74 years) with untreated intermediate- to high-risk PCa characterized according to Cancer of the Prostate Risk Assessment (CAPRA) underwent preoperative FCH dynamic PET/MR imaging followed by radical prostatectomy between April and November 2015. PET/MR imaging parameters including average and maximum K1 (delivery rate constant) and standardized uptake values (SUVs) and Prostate Imaging Reporting and Data System (PI-RADS) version 2 scores were measured and compared with clinical-pathologic characteristics. For statistical analysis, the Spearman rank correlation and Mann-Whitney U tests were performed. Results Of the PET parameters, maximum SUV of primary tumors showed significant correlations with several clinical-pathologic parameters including serum prostate-specific antigen level (ρ = 0.71, P = .01), pathologic stage (ρ = 0.59, P = .043), and postsurgical CAPRA score (ρ = 0.72, P = .008). The overall PI-RADS score showed significant correlations with pathologic tumor volume (ρ = 0.81, P < .001), percentage of tumor cells with Gleason scores greater than 3 (ρ = 0.59, P = .02), and postsurgical CAPRA score (ρ = 0.58, P = .046). The high-risk postsurgical CAPRA score patient group had a significantly higher maximum SUV than did the intermediate-risk group. Combined PET and MR imaging showed improved sensitivity (88%) for prediction of pathologic extraprostatic extension compared with that with MR imaging (50%) and PET (75%) performed separately. Conclusion Maximum SUVs and PI-RADS scores from FCH PET/MR imaging show good correlation with clinical-pathologic characteristics, such as postsurgical CAPRA score, which are related to prognosis in patients with newly diagnosed intermediate- to high-risk PCa. © RSNA, 2016 Online supplemental material is available for this article.

Patient-specific Meta-analysis of 2 Clinical Validation Studies to Predict Pathologic Outcomes in Prostate Cancer Using the 17-Gene Genomic Prostate Score.

OBJECTIVE: To perform patient-specific meta-analysis (MA) of two independent clinical validation studies of a 17-gene biopsy-based genomic assay as a predictor of favorable pathology at radical prostatectomy. MATERIALS AND METHODS: Patient-specific MA was performed on data from 2 studies (732 patients) using the Genomic Prostate Score (GPS; scale 0-100) together with Cancer of the Prostate Risk Assessment (CAPRA) score or National Comprehensive Cancer Network (NCCN) risk group as predictors of the likelihood of favorable pathology (LFP). Risk profile curves associating GPS with LFP by CAPRA score and NCCN risk group were generated. Decision curves and receiver operating characteristic curves were calculated using patient-specific MA risk estimates. RESULTS: Patient-specific MA-generated risk profiles ensure more precise estimates of LFP with narrower confidence intervals than either study alone. GPS added significant predictive value to each clinical classifier. A model utilizing GPS and CAPRA provided the most risk discrimination. In decision-curve analysis, greater net benefit was shown when combining GPS with each clinical classifier compared with the classifier alone. The area under the receiver operating characteristic curve improved from 0.68 to 0.73 by adding GPS to CAPRA, and 0.64 to 0.70 by adding GPS to NCCN risk group. The proportion of patients with LFP >80% increased from 11% using NCCN risk group alone to 23% using GPS with NCCN. Using GPS with CAPRA identified the highest proportion-31%-of patients with LFP >80%. CONCLUSION: Patient-specific MA provides more precise risk estimates that reflect the complete body of evidence. GPS adds predictive value to 3 widely used clinical classifiers, and identifies a larger proportion of low-risk patients than identified by clinical risk group alone.