Prostate Cancer Screening with PSA and MRI Followed by Targeted Biopsy Only.

BACKGROUND: Screening for prostate cancer is burdened by a high rate of overdiagnosis. The most appropriate algorithm for population-based screening is unknown. METHODS: We invited 37,887 men who were 50 to 60 years of age to undergo regular prostate-specific antigen (PSA) screening. Participants with a PSA level of 3 ng per milliliter or higher underwent magnetic resonance imaging (MRI) of the prostate; one third of the participants were randomly assigned to a reference group that underwent systematic biopsy as well as targeted biopsy of suspicious lesions shown on MRI. The remaining participants were assigned to the experimental group and underwent MRI-targeted biopsy only. The primary outcome was clinically insignificant prostate cancer, defined as a Gleason score of 3+3. The secondary outcome was clinically significant prostate cancer, defined as a Gleason score of at least 3+4. Safety was also assessed. RESULTS: Of the men who were invited to undergo screening, 17,980 (47%) participated in the trial. A total of 66 of the 11,986 participants in the experimental group (0.6%) received a diagnosis of clinically insignificant prostate cancer, as compared with 72 of 5994 participants (1.2%) in the reference group, a difference of -0.7 percentage points (95% confidence interval [CI], -1.0 to -0.4; relative risk, 0.46; 95% CI, 0.33 to 0.64; P<0.001). The relative risk of clinically significant prostate cancer in the experimental group as compared with the reference group was 0.81 (95% CI, 0.60 to 1.1). Clinically significant cancer that was detected only by systematic biopsy was diagnosed in 10 participants in the reference group; all cases were of intermediate risk and involved mainly low-volume disease that was managed with active surveillance. Serious adverse events were rare (<0.1%) in the two groups. CONCLUSIONS: The avoidance of systematic biopsy in favor of MRI-directed targeted biopsy for screening and early detection in persons with elevated PSA levels reduced the risk of overdiagnosis by half at the cost of delaying detection of intermediate-risk tumors in a small proportion of patients. (Funded by Karin and Christer Johansson's Foundation and others; GÖTEBORG-2 ISRCTN Registry number, ISRCTN94604465.).

Transcriptome-wide prediction of prostate cancer gene expression from histopathology images using co-expression-based convolutional neural networks.

MOTIVATION: Molecular phenotyping by gene expression profiling is central in contemporary cancer research and in molecular diagnostics but remains resource intense to implement. Changes in gene expression occurring in tumours cause morphological changes in tissue, which can be observed on the microscopic level. The relationship between morphological patterns and some of the molecular phenotypes can be exploited to predict molecular phenotypes from routine haematoxylin and eosin-stained whole slide images (WSIs) using convolutional neural networks (CNNs). In this study, we propose a new, computationally efficient approach to model relationships between morphology and gene expression. RESULTS: We conducted the first transcriptome-wide analysis in prostate cancer, using CNNs to predict bulk RNA-sequencing estimates from WSIs for 370 patients from the TCGA PRAD study. Out of 15 586 protein coding transcripts, 6618 had predicted expression significantly associated with RNA-seq estimates (FDR-adjusted P-value <1×10-4) in a cross-validation and 5419 (81.9%) of these associations were subsequently validated in a held-out test set. We furthermore predicted the prognostic cell-cycle progression score directly from WSIs. These findings suggest that contemporary computer vision models offer an inexpensive and scalable solution for prediction of gene expression phenotypes directly from WSIs, providing opportunity for cost-effective large-scale research studies and molecular diagnostics. AVAILABILITY AND IMPLEMENTATION: A self-contained example is available from http://github.com/phiwei/prostate_coexpression. Model predictions and metrics are available from doi.org/10.5281/zenodo.4739097. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Interobserver reproducibility of cribriform cancer in prostate needle biopsies and validation of International Society of Urological Pathology criteria.

AIMS: There is strong evidence that cribriform morphology indicates a worse prognosis of prostatic adenocarcinoma. Our aim was to investigate its interobserver reproducibility in prostate needle biopsies. METHODS AND RESULTS: A panel of nine prostate pathology experts from five continents independently reviewed 304 digitised biopsies for cribriform cancer according to recent International Society of Urological Pathology criteria. The biopsies were collected from a series of 702 biopsies that were reviewed by one of the panellists for enrichment of high-grade cancer and potentially cribriform structures. A 2/3 consensus diagnosis of cribriform and noncribriform cancer was reached in 90% (272/304) of the biopsies with a mean kappa value of 0.56 (95% confidence interval 0.52-0.61). The prevalence of consensus cribriform cancers was estimated to 4%, 12%, 21%, and 20% of Gleason scores 7 (3 + 4), 7 (4 + 3), 8, and 9-10, respectively. More than two cribriform structures per level or a largest cribriform mass with ≥9 lumina or a diameter of ≥0.5 mm predicted a consensus diagnosis of cribriform cancer in 88% (70/80), 84% (87/103), and 90% (56/62), respectively, and noncribriform cancer in 3% (2/80), 5% (5/103), and 2% (1/62), respectively (all P < 0.01). CONCLUSION: Cribriform prostate cancer was seen in a minority of needle biopsies with high-grade cancer. Stringent diagnostic criteria enabled the identification of cribriform patterns and the generation of a large set of consensus cases for standardisation.

Ductal and acinar components of mixed prostatic adenocarcinoma frequently have a common clonal origin.

BACKGROUND: Ductal adenocarcinoma (DA) is an aggressive subtype of prostate cancer. It is most commonly seen in mixed tumors together with conventional acinar adenocarcinoma (AA). The genetic profile of DA and its clonal origin is not fully characterized. OBJECTIVE: To investigate whether DA represents a distinct genetic subtype and to investigate the somatic relationship between the ductal and acinar components of mixed cancers. DESIGN, SETTING, AND PARTICIPANTS: In 17 radical prostatectomy specimens ductal and acinar tumor components from the same tumor foci were dissected. DNA was extracted and genomic sequencing performed. After exclusion of two cases with low cell yield, 15 paired samples remained for analysis. RESULTS: In 12 of 15 cases a common somatic denominator was identified, while three cases had clonally separate components. In DA, TMPRSS2-ERG gene fusions were detected in 47% (7/15), clonal FOXA1 alterations in 33% (5/15) and SPOP alterations in 27% (4/15) of cases. In one case KIAA1549-BRAF fusion was identified. Genome doubling events, resulting in an increased ploidy, were identified in the DA in 53% (8/15) of cases, but not seen in any AA. PTEN and CTNNB1 alterations were enriched in DA (6/15) but not seen in any AA. No cancers showed microsatellite instability or high tumor mutation burden. CONCLUSIONS: Ductal and acinar prostate adenocarcinoma components of mixed tumors most often share the same origin and are clonally related. DA components in mixed tumor often exhibit genome doubling events resulting in aneuploidy, consistent with the aggressive nature of high grade prostate cancer.

OpenPhi: an interface to access Philips iSyntax whole slide images for computational pathology.

SUMMARY: Digital pathology enables applying computational methods, such as deep learning, in pathology for improved diagnostics and prognostics, but lack of interoperability between whole slide image formats of different scanner vendors is a challenge for algorithm developers. We present OpenPhi-Open PatHology Interface, an Application Programming Interface for seamless access to the iSyntax format used by the Philips Ultra Fast Scanner, the first digital pathology scanner approved by the United States Food and Drug Administration. OpenPhi is extensible and easily interfaced with existing vendor-neutral applications. AVAILABILITY AND IMPLEMENTATION: OpenPhi is implemented in Python and is available as open-source under the MIT license at: https://gitlab.com/BioimageInformaticsGroup/openphi. The Philips Software Development Kit is required and available at: https://www.openpathology.philips.com. OpenPhi version 1.1.1 is additionally provided as Supplementary Data. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Primary tumour PSMA intensity is an independent prognostic biomarker for biochemical recurrence-free survival following radical prostatectomy.

PURPOSE: The prognostic value of PSMA intensity on PSMA PET/CT due to underlying biology and subsequent clinical implications is an emerging topic of interest. We sought to investigate whether primary tumour PSMA PET intensity contributes to pre- and post-operative prediction of oncological outcomes following radical prostatectomy. METHODS: We performed a retrospective cohort study of 848 men who underwent all of multiparametric MRI (mpMRI), transperineal prostate biopsy, and 68 Ga-PSMA PET/CT prior to radical prostatectomy. PSMA intensity, quantified as maximum standard uptake value (SUVmax), and other clinical variables were considered relative to post-operative biochemical recurrence-free survival (BRFS) using Cox regression and Kaplan-Meier analysis. RESULTS: After a median follow-up of 41 months, 219 events occurred; the estimated 3-year BRFS was 79% and the 5-year BRFS was 70%. Increasing PSMA intensity was associated with less favourable BRFS overall (Log rank p < 0.001), and within subgroups of Gleason score category (Log rank p < 0.03). PSMA intensity was significantly associated with shorter time to biochemical recurrence, after adjusting for pre-operative (HR per 5-unit SUVmax increase = 1.15) and post-operative (HR per 5-unit SUVmax increase = 1.10) parameters. CONCLUSION: These results in a large series of patients confirm PSMA intensity to be a novel, independent prognostic factor for BRFS.

Oncological and urinary outcomes following low-dose-rate brachytherapy with a median follow-up of 11.8 years.

OBJECTIVES: To examine the long-term oncological outcomes and urological morbidity of low-dose-rate prostate brachytherapy (LDRBT) monotherapy using live intraoperative dosimetry planning and an automated needle navigation delivery system for the treatment of men with low and intermediate-risk prostate cancer. PATIENTS AND METHODS: A prospective database of 400 consecutive patients who underwent LDRBT between July 2003 and June 2015 was retrospectively reviewed to assess urinary side-effects and biochemical progression, based on the Phoenix definition and also a definition of a prostate-specific antigen (PSA) level of ≥0.2 µg/L. RESULTS: Minimum patient follow-up was 5.5 years. The median follow-up of the entire cohort was 11.8 years. The median (range) PSA level was 6.1 (0.9-17) µg/L and the median Gleason score was 3 + 4. The biochemical relapse-free survival (RFS; freedom from biochemical recurrence) based on the Phoenix definition was 85.8% (343/400). The RFS using a 'surgical' definition of a PSA level of <0.2 µg/L was 71% (284/400). Of the 297 men followed for ≥10 years, prostate cancer-specific survival (PCSS) was 98% (291/297). Post-LDRBT urethral stricture developed in 11 men (2.8%, 11/400). For men with ≥10 years of follow-up, 22 men (7.4%, 22/297) required a pad for either stress or urge urinary incontinence (UI). UI was identified in only 2.2% (one of 46) of men who had a bladder neck incision (BNI) before LDRBT. CONCLUSION: LDRBT is associated with excellent PCSS, with a median follow-up of 11.8 years. The risk of post-implantation urethral stricture and UI is low and a pre-implantation BNI for management of bladder outflow obstruction does not increase the risk of UI or urethral stricture.

Artificial intelligence assistance significantly improves Gleason grading of prostate biopsies by pathologists.

The Gleason score is the most important prognostic marker for prostate cancer patients, but it suffers from significant observer variability. Artificial intelligence (AI) systems based on deep learning can achieve pathologist-level performance at Gleason grading. However, the performance of such systems can degrade in the presence of artifacts, foreign tissue, or other anomalies. Pathologists integrating their expertise with feedback from an AI system could result in a synergy that outperforms both the individual pathologist and the system. Despite the hype around AI assistance, existing literature on this topic within the pathology domain is limited. We investigated the value of AI assistance for grading prostate biopsies. A panel of 14 observers graded 160 biopsies with and without AI assistance. Using AI, the agreement of the panel with an expert reference standard increased significantly (quadratically weighted Cohen's kappa, 0.799 vs. 0.872; p = 0.019). On an external validation set of 87 cases, the panel showed a significant increase in agreement with a panel of international experts in prostate pathology (quadratically weighted Cohen's kappa, 0.733 vs. 0.786; p = 0.003). In both experiments, on a group-level, AI-assisted pathologists outperformed the unassisted pathologists and the standalone AI system. Our results show the potential of AI systems for Gleason grading, but more importantly, show the benefits of pathologist-AI synergy.

Histological comparison between predictive value of preoperative 3-T multiparametric MRI and 68 Ga-PSMA PET/CT scan for pathological outcomes at radical prostatectomy and pelvic lymph node dissection for prostate cancer.

OBJECTIVE: To evaluate the ability of preoperative multiparametric magnetic resonance imaging (mpMRI) and a gallium-68 prostate-specific membrane antigen positron emission tomography/computed tomography (68 Ga-PSMA PET/CT) scan to predict pathological outcomes and also identify a group of men with a <5% risk of histological pelvic lymph node metastasis (LNM) at pelvic lymph node dissection (PLND) performed during a robot-assisted laparoscopic radical prostatectomy (RALP) for prostate cancer. We then aimed to compare these results to known risk calculators for LNM, including the Cancer of the Prostate Risk Assessment (CAPRA) score, Memorial Sloan Kettering Cancer Centre (MSKCC) and Briganti nomograms. PATIENTS AND METHODS: Between July 2014 and September 2019 only men who had both a preoperative mpMRI and staging 68 Ga-PSMA PET/CT at our institution followed by a RALP with PLND referred to a single specialist uropathology laboratory were considered for inclusion. The data were collected retrospectively prior to February 2019 and in a prospective manner thereafter. A model was built to allocate probabilities of the men with a negative 68 Ga-PSMA PET/CT scan having a <5% risk of histologically LNM at RALP based on the preoperative radiological staging. RESULTS: A total of 233 consecutive men met the inclusion criteria of which 58 men (24.9%) had a LNM identified on PLND histology. The median (range) International Society of Urological Pathology (ISUP) Grade was 5 (1-5) and the median (range) prostate-specific antigen level was 7.4 (1.5-72) ng/mL. The median (range) number of resected lymph nodes was 16 (1-53) and the median (range) number of positive nodes identified on histology was 2 (1-22). Seminal vesicle invasion on mpMRI was more common in node-positive men than in the absence of LNM (31% vs 12%). The maximum standardised uptake value of the primary tumour on 68 Ga-PSMA PET/CT was higher in men with LNM (median 9.2 vs 7.2, P = 0.02). Suspected LNM were identified in 42/233 (18.0%) men with 68 Ga-PSMA PET/CT compared with 22/233 (9.4%) men with mpMRI (P = 0.023). The positive and negative predictive value for 68 Ga-PSMA PET/CT was 66.7% and 84.3% respectively, compared to 59.1% and 78.7% for mpMRI. A predictive model showed only two men (4.2%) with a negative preoperative 68 Ga-PSMA PET/CT would be positive for a histological LNM if they are ISUP Grade < 5 and Prostate Imaging-Reporting and Data System (PI-RADS) <5; or ISUP Grade 5 with PI-RADS < 4. An inspection of three additional variables: CAPRA score, MSKCC and Briganti nomograms did not improve the predictive probability for this group. However, of the 61 men with ISUP Grade 4-5 malignancy and also a PI-RADS 5 mpMRI, 20 (32.8%) men had a microscopic LNM despite a negative preoperative 68 Ga-PSMA PET/CT. CONCLUSION: Preoperative 68 Ga-PSMA/PET CT was more sensitive in identifying histological pelvic LNM than 3-T mpMRI. Men with a negative 68 Ga-PSMA PET/CT have a lower risk of LNM than predicted with CAPRA scores or MSKCC and Briganti nomograms. We identified that the combination of a negative preoperative 68 Ga-PSMA PET/CT, ISUP biopsy Grade <5 and PI-RADS <5 prostate mpMRI, or an ISUP Grade 5 with PI-RADS <4 on mpMRI was associated with a <5% risk of a LNM. The addition of CAPRA scores, MSKCC and Briganti nomograms did not improve the predictive probability within this model. Conversely, men with ISUP Grade 4-5 malignancy associated with a PI-RADS 5 prostate mpMRI had a >30% risk of microscopic LNM despite a negative preoperative 68 Ga-PSMA PET/CT and this high-risk group would appear suitable for an extended PLND at the time of a radical prostatectomy.

Temporal changes in cause-specific death in men with localised prostate cancer treated with radical prostatectomy: a population-based, nationwide study.

BACKGROUND AND OBJECTIVE: Changes in diagnostic work-up, histopathological assessment, and treatment of men with prostate cancer during the last 20 years have affected the prognosis. The objective was to investigate the risk of prostate cancer death in men with clinically localised prostate cancer treated with radical prostatectomy in Sweden in 2000-2010. METHODS: Population-based, nationwide, study on men with clinically localised prostate cancer treated with radical prostatectomy in the period 2000-2010. Cox regression analyses were used to assess differences in risk of prostate cancer death according to calendar period for diagnosis and stratified on risk category. RESULTS: The study included 19 330 men with a median follow-up of 12.4 years. Men diagnosed in 2007-2008 and 2009-2010 had a significantly lower risk of prostate cancer death compared to men diagnosed in 2000-2002. The reduced risk of prostate cancer death was restricted to men with intermediate-risk prostate cancer with no differences observed in men with low- or high-risk prostate cancer. CONCLUSION: During the study period, the risk of prostate cancer death decreased in the total population of men with localised prostate cancer treated with radical prostatectomy. The decrease was restricted to men with intermediate-risk prostate cancer.

Nuclear and stromal expression of Manic fringe in renal cell carcinoma.

Renal cell carcinoma (RCC) is the most common type of kidney cancer and has the highest mortality rate among genitourinary cancers. Despite the advances in molecular targeted therapies to treat RCC, the inevitable emergence of resistance has delineated the need to uncover biomarkers to prospectively identify patient response to treatment and more accurately predict patient prognosis. Fringe is a fucose specific ß1, 3N-acetylglucosaminyltransferase that modifies the Notch receptors. Given the link between its function and aberrant Notch activation in RCC, Fringe may be implicated in this disease. The Fringe homologs comprise of Lunatic fringe (LFng), Manic fringe (MFng) and Radical fringe (RFng). MFng has been reported to play a role in cancer. MFng is also essential in the development of B cells. However, the expression profile and clinical significance of MFng, and its association with B cells in RCC are unknown. CD20 is a clinically employed biomarker for B cells. This pilot study aimed to determine if MFng protein expression can be utilized as a prospective biomarker for therapeutics and prognosis in RCC, as well as to determine its association with CD20+ B cells. Analysis of publicly available MFng gene expression datasets on The Cancer Genome Atlas Netlwork (TCGA) identified MFng gene expression to be up-regulated in Kidney Clear Cell Renal Carcinoma (KIRC) patients. However there was no significant association between the patient survival probability and the level of MFng expression in this cohort. Immunohistochemistry performed on a tissue microarray containing cores from 64 patients revealed an elevated MFng protein expression in the epithelial and stromal tissues of RCC compared to the normal kidney, suggesting a possible role in tumorigenesis. Our study describes for the first time to our knowledge, the protein expression of MFng in the nuclear compartment of normal kidney and RCC, implicating a prospective involvement in gene transcription. At the cellular level, cytoplasmic MFng was also abundant in the normal kidney and RCC. However, MFng protein expression in the malignant epithelial and stromal tissue of RCC had no positive correlation with the patients' overall survival, progression-free survival and time to metastasis, as well as the gender, age, tumor stage and RCC subtype, indicating that MFng may not be an appropriate prognostic marker. The association between CD20+ B cells and epithelial MFng was found to approach borderline insignificance. Nonetheless, these preliminary findings may provide valuable information on the suitability of MFng as a potential therapeutic molecular marker for RCC, thus warrants further investigation using a larger cohort.

Perivascular Neuropilin-1 expression is an independent marker of improved survival in renal cell carcinoma.

Renal cell carcinoma (RCC) treatment has improved in the last decade with the introduction of drugs targeting tumor angiogenesis. However, the 5-year survival of metastatic disease is still only 10-15%. Here, we explored the prognostic significance of compartment-specific expression of Neuropilin 1 (NRP1), a co-receptor for vascular endothelial growth factor (VEGF). NRP1 expression was analyzed in RCC tumor vessels, in perivascular tumor cells, and generally in the tumor cell compartment. Moreover, complex formation between NRP1 and the main VEGF receptor, VEGFR2, was determined. Two RCC tissue microarrays were used; a discovery cohort consisting of 64 patients and a validation cohort of 314 patients. VEGFR2/NRP1 complex formation in cis (on the same cell) and trans (between cells) configurations was determined by in situ proximity ligation assay (PLA), and NRP1 protein expression in three compartments (endothelial cells, perivascular tumor cells, and general tumor cell expression) was determined by immunofluorescent staining. Expression of NRP1 in perivascular tumor cells was explored as a marker for RCC survival in the two RCC cohorts. Results were further validated using a publicly available gene expression dataset of clear cell RCC (ccRCC). We found that VEGFR2/NRP1 trans complexes were detected in 75% of the patient samples. The presence of trans VEGFR2/NRP1 complexes or perivascular NRP1 expression was associated with a reduced tumor vessel density and size. When exploring NRP1 as a biomarker for RCC prognosis, perivascular NRP1 and general tumor cell NRP1 protein expression correlated with improved survival in the two independent cohorts, and significant results were obtained also at the mRNA level using the publicly available ccRCC gene expression dataset. Only perivascular NRP1 expression remained significant in multivariable analysis. Our work shows that perivascular NRP1 expression is an independent marker of improved survival in RCC patients, and reduces tumor vascularization by forming complexes in trans with VEGFR2 in the tumor endothelium. © 2019 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of Pathological Society of Great Britain and Ireland.

Cribriform prostate cancer: Morphologic criteria enabling a diagnosis, based on survey of experts.

Among four sub-patterns of Gleason grade 4 prostate cancer, voluminous evidence supports that the cribriform pattern holds an unfavorable prognostic impact, as compared with poorly-formed, fused, or glomeruloid. The International Society of Urological Pathology (ISUP) recommends specifying whether invasive grade 4 cancer is cribriform. Recently, ISUP experts published a consensus definition of cribriform pattern highlighting criteria that distinguish it from mimickers. The current study aimed to analyze morphologic features separately to identify those that define the essence of the cribriform pattern. Thirty-two selected photomicrographs were classified by 12 urologic pathologists as: definitely cribriform cancer, probably cribriform, unsure, probably not cribriform, or definitely not cribriform. Consensus was defined as 9/12 agree or disagree, with ≤1 strongly supporting the opposite choice. Final consensus was achieved in 21 of 32 cases. Generalized estimating equation (GEE) model with logit link was fitted to estimate effect of multiple morphologic predictors. Fisher exact test was used for categorical findings. Presence of intervening stroma precluded calling cribriform cancer (p = 0.006). Mucin presence detracted (p = 0.003) from willingness to call cribriform cancer (only 3 cases had mucin). Lumen number was associated with cribriform consensus (p = 0.0006), and all consensus cases had ≥9 lumens. Predominant papillary pattern or an irregular outer boundary detracted (p = NS). Invasive cribriform carcinoma should have absence of intervening stroma, and usually neither papillary pattern, irregular outer boundary, nor very few lumens. Setting the criteria for cribriform will help prevent over- or undercalling this important finding.

[Molecular pathology of urogenital tumors : Recommendations from the 2019 International Society of Urological Pathology (ISUP) Consensus Conference]. / Molekularpathologie bei urologischen Tumoren : Empfehlungen der Konsenskonferenz der Internationalen Gesellschaft für Uropathologie (ISUP) 2019.

Comprehensive understanding of molecular principles in cancer and the diversification of oncological therapy promise individual therapeutic concepts, which have not yet found their way into urogenital cancer therapy. In March 2019 the International Society of Urogenital Pathology (ISUP) therefore held a consensus conference on recommendations for molecular diagnostics of genitourinary tumors, which were published in five separate manuscripts and are summarized in this article.In preparation for the conference, a comprehensive survey of current practices for molecular testing of urogenital tumors was carried out by members of the ISUP. At the conference, the results and the corresponding background information were presented by five working groups and recommendations for action for diagnostics were developed. An agreement between 66% of the conference participants was defined as consensus.

Artificial intelligence for diagnosis and grading of prostate cancer in biopsies: a population-based, diagnostic study.

BACKGROUND: An increasing volume of prostate biopsies and a worldwide shortage of urological pathologists puts a strain on pathology departments. Additionally, the high intra-observer and inter-observer variability in grading can result in overtreatment and undertreatment of prostate cancer. To alleviate these problems, we aimed to develop an artificial intelligence (AI) system with clinically acceptable accuracy for prostate cancer detection, localisation, and Gleason grading. METHODS: We digitised 6682 slides from needle core biopsies from 976 randomly selected participants aged 50-69 in the Swedish prospective and population-based STHLM3 diagnostic study done between May 28, 2012, and Dec 30, 2014 (ISRCTN84445406), and another 271 from 93 men from outside the study. The resulting images were used to train deep neural networks for assessment of prostate biopsies. The networks were evaluated by predicting the presence, extent, and Gleason grade of malignant tissue for an independent test dataset comprising 1631 biopsies from 246 men from STHLM3 and an external validation dataset of 330 biopsies from 73 men. We also evaluated grading performance on 87 biopsies individually graded by 23 experienced urological pathologists from the International Society of Urological Pathology. We assessed discriminatory performance by receiver operating characteristics and tumour extent predictions by correlating predicted cancer length against measurements by the reporting pathologist. We quantified the concordance between grades assigned by the AI system and the expert urological pathologists using Cohen's kappa. FINDINGS: The AI achieved an area under the receiver operating characteristics curve of 0·997 (95% CI 0·994-0·999) for distinguishing between benign (n=910) and malignant (n=721) biopsy cores on the independent test dataset and 0·986 (0·972-0·996) on the external validation dataset (benign n=108, malignant n=222). The correlation between cancer length predicted by the AI and assigned by the reporting pathologist was 0·96 (95% CI 0·95-0·97) for the independent test dataset and 0·87 (0·84-0·90) for the external validation dataset. For assigning Gleason grades, the AI achieved a mean pairwise kappa of 0·62, which was within the range of the corresponding values for the expert pathologists (0·60-0·73). INTERPRETATION: An AI system can be trained to detect and grade cancer in prostate needle biopsy samples at a ranking comparable to that of international experts in prostate pathology. Clinical application could reduce pathology workload by reducing the assessment of benign biopsies and by automating the task of measuring cancer length in positive biopsy cores. An AI system with expert-level grading performance might contribute a second opinion, aid in standardising grading, and provide pathology expertise in parts of the world where it does not exist. FUNDING: Swedish Research Council, Swedish Cancer Society, Swedish eScience Research Center, EIT Health.

Perineural invasion by prostate adenocarcinoma in needle biopsies predicts bone metastasis: Ten year data from the TROG 03.04 RADAR Trial.

AIMS: Perineural invasion (PNI) by prostatic adenocarcinoma is debated as a prognostic parameter. This study investigates the prognostic predictive value of PNI in a series of patients with locally advanced prostate cancer treated with radiotherapy and androgen deprivation using 10 years outcome data from the TROG 03.04 RADAR trial. METHODS: Diagnostic prostate biopsies from 976 patients were reviewed and the presence of PNI noted. Patients were followed for 10 years according to the trial protocol or until death. The primary endpoint for the study was time to bone metastasis. Secondary endpoints included time to soft tissue metastasis, transition to castration resistance, prostate cancer-specific mortality and all-cause mortality. RESULTS: PNI was detected in 449 cases (46%), with 234 cases (24%) having PNI in more than one core. The presence of PNI was significantly associated with higher ISUP grade, clinical T staging category, National Comprehensive Cancer Network risk group, and percent positive biopsy cores. The cumulative probability of bone metastases according to PNI status was significant over the 10 years follow-up interval of the study (log-rank test P < 0.0001). PNI was associated with all endpoints on univariable analysis. After adjusting for baseline clinicopathological and treatment factors, bone metastasis was the only endpoint in which PNI retained its prognostic significance (hazard ratio 1.42, 95% confidence interval 1.05-1.92, P = 0.021). CONCLUSIONS: The association between PNI and the development of bone metastases supports the inclusion of this parameter as a component of the routine histology report. Further this association suggests that evaluation of PNI may assist in selecting those patients who should be monitored more closely during follow-up.

The natural history of untreated muscle-invasive bladder cancer.

OBJECTIVE: To describe the natural history of untreated muscle-invasive bladder cancer (MIBC) and compare the oncological outcomes of treated and untreated patients. PATIENTS AND METHODS: We utilised a database encompassing all patients with newly diagnosed bladder cancer in Stockholm, Sweden between 1995 and 1996. The median follow-up for survivors was 14.4 years. Overall, 538 patients were diagnosed with bladder cancer of whom 126 had clinically localised MIBC. Patients were divided into two groups: those who received radical cystectomy or radiation therapy, and those who did not receive any form of treatment. Multivariable Cox or competing-risks regressions were adopted to predict metastasis, overall survival (OS), and cancer-specific mortality (CSM), when appropriate. Analyses were adjusted for age at diagnosis, sex, tumour stage, clinical N stage, and treatment. RESULTS: In all, 64 (51%) patients did not receive any definitive local treatment. In the untreated group, the median (interquartile range) age at diagnosis was 79 (63-83) vs 69 (63-74) years in the treated group (P < 0.001). Overall, 109 patients died during follow-up. At 6 months after diagnosis, 38% of the untreated patients had developed metastatic disease and 41% had CSM. The 5-year OS rate for untreated and treated patients was 5% (95% confidence interval [CI] 1, 12%) vs 48% (95% CI 36, 60%), respectively. Patients not receiving any treatment had a 5-year cumulative incidence of CSM of 86% (95% CI 75, 94%) vs 48% (95% CI 36, 60%) for treated patients. Untreated patients had a higher risk of progression to metastatic disease (hazard ratio [HR] 2.40, 95% CI 1.28, 4.51; P = 0.006), death from any cause (HR 2.63, 95% CI 1.65, 4.19; P < 0.001) and CSM (subdistribution HR 2.02, 95% CI 1.24, 3.30; P = 0.004). CONCLUSIONS: Untreated patients with MIBC are at very high risk of near-term CSM. These findings may help balance the risks vs benefits of integrating curative intent therapy particularly in older patients with MIBC.

Somatic alterations detected in diagnostic prostate biopsies provide an inadequate representation of multifocal prostate cancer.

BACKGROUND: The majority of clinical prostate cancers are multifocal with morphological and molecular heterogeneity. Adequate tissue representation is crucial for the clinical utility of multigene panel sequencing of core needle biopsies. The aim of this study was to evaluate the genomic heterogeneity in multifocal prostate cancer and to analyze how representative preoperative biopsies are of spatially separated tumor foci. METHODS: We analyzed at least 2 tumor foci and 1 to 3 preoperative biopsy cores from 11 patients. Diagnostic biopsies, as well as fresh frozen and formalin-fixed paraffin-embedded samples, from major tumor foci of radical prostatectomy specimens were macrodissected for the enrichment of tumor tissue. DNA was extracted and sequenced. We analyzed structural alterations, mutations, and copy number variations and compared the genomic profiles of tumor foci with those of preoperative biopsies. RESULTS: Alterations were rarely shared between foci, indicating a high degree of genomic heterogeneity. In 8 of 11 men at least 1 tumor focus was represented by the biopsies defined as harboring at least 1 common clonal somatic event. In only one case, somatic alterations from two spatially separate tumors were identified in the biopsies. Of the mutations and structural variants detected in fresh frozen or formalin-fixed paraffin-embedded prostatectomy material, only an average of 19% (range 0-44) and 55% (range 0-100), respectively, were found in preoperative biopsies where a common somatic origin was established. CONCLUSIONS: Multifocal prostate cancer is a somatically heterogeneous disease in which systematic needle biopsies do not provide sufficient molecular representation of the somatic alterations detected in spatially distinct tumor areas. Targeted biopsies, directed at separate tumor foci, could potentially improve tissue representation of these heterogeneous foci in preoperative biopsies.

Prostate cancer induces C/EBPß expression in surrounding epithelial cells which relates to tumor aggressiveness and patient outcome.

BACKGROUND: Implantation of rat prostate cancer cells into the normal rat prostate results in tumor-stimulating adaptations in the tumor-bearing organ. Similar changes are seen in prostate cancer patients and they are related to outcome. One gene previously found to be upregulated in the non-malignant part of tumor-bearing prostate lobe in rats was the transcription factor CCAAT/enhancer-binding protein-ß (C/EBPß). METHODS: To explore this further, we examined C/EBPß expression by quantitative RT-PCR, immunohistochemistry, and Western blot in normal rat prostate tissue surrounding slow-growing non-metastatic Dunning G, rapidly growing poorly metastatic (AT-1), and rapidly growing highly metastatic (MatLyLu) rat prostate tumors-and also by immunohistochemistry in a tissue microarray (TMA) from prostate cancer patients managed by watchful waiting. RESULTS: In rats, C/EBPß mRNA expression was upregulated in the surrounding tumor-bearing prostate lobe. In tumors and in the surrounding non-malignant prostate tissue, C/EBPß was detected by immunohistochemistry in some epithelial cells and in infiltrating macrophages. The magnitude of glandular epithelial C/EBPß expression in the tumor-bearing prostates was associated with tumor size, distance to the tumor, and metastatic capacity. In prostate cancer patients, high expression of C/EBPß in glandular epithelial cells in the surrounding tumor-bearing tissue was associated with accumulation of M1 macrophages (iNOS+) and favorable outcome. High expression of C/EBPß in tumor epithelial cells was associated with high Gleason score, high tumor cell proliferation, metastases, and poor outcome. CONCLUSIONS: This study suggest that the expression of C/EBP-beta, a transcription factor mediating multiple biological effects, is differentially expressed both in the benign parts of the tumor-bearing prostate and in prostate tumors, and that alterations in this may be related to patient outcome.

Immunoreactivity for prostate specific antigen and Ki67 differentiates subgroups of prostate cancer related to outcome.

Based on gene-expression profiles, prostate tumors can be subdivided into subtypes with different aggressiveness and response to treatment. We investigated if similar clinically relevant subgroups can be identified simply by the combination of two immunohistochemistry markers: one for tumor cell differentiation (prostate specific antigen, PSA) and one for proliferation (Ki67). This was analyzed in men with prostate cancer diagnosed at transurethral resection of the prostate 1975-1991 (n = 331) where the majority was managed by watchful waiting. Ki67 and PSA immunoreactivity was related to outcome and to tumor characteristics previously associated with prognosis. Increased Ki67 and decreased PSA were associated with poor outcome, and they provided independent prognostic information from Gleason score. A combinatory score for PSA and Ki67 immunoreactivity was produced using the median PSA and Ki67 levels as cut-off (for Ki67 the upper quartile was also evaluated) for differentiation into subgroups. Patients with PSA low/Ki67 high tumors showed higher Gleason score, more advanced tumor stage, and higher risk of prostate cancer death compared to other patients. Their tumor epithelial cells were often ERG positive and expressed higher levels of ErbB2, phosphorylated epidermal growth factor receptor (pEGF-R) and protein kinase B (pAkt), and their tumor stroma showed a reactive response with type 2 macrophage infiltration, high density of blood vessels and hyaluronic acid, and with reduced levels of caveolin-1, androgen receptors, and mast cells. In contrast, men with PSA high/Ki67 low tumors were characterized by low Gleason score, and the most favorable outcome amongst PSA/Ki67-defined subgroups. Men with PSA low/Ki67 low tumors showed clinical and tumor characteristics intermediate of the two groups above. A combinatory PSA/Ki67 immunoreactivity score identifies subgroups of prostate cancers with different epithelial and stroma phenotypes and highly different outcome but the clinical usefulness of this approach needs to be validated in other cohorts.