A deep-learning workflow to predict upper tract urothelial carcinoma protein-based subtypes from H&E slides supporting the prioritization of patients for molecular testing.

Upper tract urothelial carcinoma (UTUC) is a rare and aggressive, yet understudied, urothelial carcinoma (UC). The more frequent UC of the bladder comprises several molecular subtypes, associated with different targeted therapies and overlapping with protein-based subtypes. However, if and how these findings extend to UTUC remains unclear. Artificial intelligence-based approaches could help elucidate UTUC's biology and extend access to targeted treatments to a wider patient audience. Here, UTUC protein-based subtypes were identified, and a deep-learning (DL) workflow was developed to predict them directly from routine histopathological H&E slides. Protein-based subtypes in a retrospective cohort of 163 invasive tumors were assigned by hierarchical clustering of the immunohistochemical expression of three luminal (FOXA1, GATA3, and CK20) and three basal (CD44, CK5, and CK14) markers. Cluster analysis identified distinctive luminal (N = 80) and basal (N = 42) subtypes. The luminal subtype mostly included pushing, papillary tumors, whereas the basal subtype diffusely infiltrating, non-papillary tumors. DL model building relied on a transfer-learning approach by fine-tuning a pre-trained ResNet50. Classification performance was measured via three-fold repeated cross-validation. A mean area under the receiver operating characteristic curve of 0.83 (95% CI: 0.67-0.99), 0.8 (95% CI: 0.62-0.99), and 0.81 (95% CI: 0.65-0.96) was reached in the three repetitions. High-confidence DL-based predicted subtypes showed significant associations (p < 0.001) with morphological features, i.e. tumor type, histological subtypes, and infiltration type. Furthermore, a significant association was found with programmed cell death ligand 1 (PD-L1) combined positive score (p < 0.001) and FGFR3 mutational status (p = 0.002), with high-confidence basal predictions containing a higher proportion of PD-L1 positive samples and high-confidence luminal predictions a higher proportion of FGFR3-mutated samples. Testing of the DL model on an independent cohort highlighted the importance to accommodate histological subtypes. Taken together, our DL workflow can predict protein-based UTUC subtypes, associated with the presence of targetable alterations, directly from H&E slides.

Enhancer profiling identifies epigenetic markers of endocrine resistance and reveals therapeutic options for metastatic castration-resistant prostate cancer patients.

Androgen Receptor (AR) signaling inhibitors, including enzalutamide, are treatment options for patients with metastatic castration-resistant prostate cancer (mCRPC), but resistance inevitably develops. Using metastatic samples from a prospective phase II clinical trial, we epigenetically profiled enhancer/promoter activities with H3K27ac chromatin immunoprecipitation followed by sequencing, before and after AR-targeted therapy. We identified a distinct subset of H3K27ac-differentially marked regions that associated with treatment responsiveness. These data were successfully validated in mCRPC patient-derived xenograft models (PDX). In silico analyses revealed HDAC3 as a critical factor that can drive resistance to hormonal interventions, which we validated in vitro . Using cell lines and mCRPC PDX tumors in vitro , we identified drug-drug synergy between enzalutamide and the pan-HDAC inhibitor vorinostat, providing therapeutic proof-of-concept. These findings demonstrate rationale for new therapeutic strategies using a combination of AR and HDAC inhibitors to improve patient outcome in advanced stages of mCRPC.

Stratification based on methylation of TBX2 and TBX3 into three molecular grades predicts progression in patients with pTa-bladder cancer.

The potential risk of recurrence and progression in patients with non-muscle-invasive bladder cancer necessitates followup by cystoscopy. The risk of progression to muscle-invasive bladder cancer is estimated based on the European Organisation of Research and Treatment of Cancer score, a combination of several clinicopathological variables. However, pathological assessment is not objective and reproducibility is insufficient. The use of molecular markers could contribute to the estimation of tumor aggressiveness. We recently demonstrated that methylation of GATA2, TBX2, TBX3, and ZIC4 genes could predict progression in Ta tumors. In this study, we aimed to validate the markers in a large patient set using DNA from formalin-fixed and paraffin-embedded tissue. PALGA: the Dutch Pathology Registry was used for patient selection. We included 192 patients with pTaG1/2 bladder cancer of whom 77 experienced progression. Methylation analysis was performed and log-rank analysis was used to calculate the predictive value of each methylation marker for developing progression over time. This analysis showed better progression-free survival in patients with low methylation rates compared with the patients with high methylation rates for all markers (P<0.001) during a followup of ten-years. The combined predictive effect of the methylation markers was analyzed with the Cox-regression method. In this analysis, TBX2, TBX3, and ZIC4 were independent predictors of progression. On the basis of methylation status of TBX2 and TBX3, patients were divided into three new molecular grade groups. Survival analysis showed that only 8% of patients in the low molecular grade group progressed within 5 years. This was 29 and 63% for the intermediate- and high-molecular grade groups. In conclusion, this new molecular-grade based on the combination of TBX2 and TBX3 methylation is an excellent marker for predicting progression to muscle-invasive bladder cancer in patients with primary pTaG1/2 bladder cancer.

ERG immunohistochemistry is not predictive for PSA recurrence, local recurrence or overall survival after radical prostatectomy for prostate cancer.

In prostate cancer genomic rearrangements involving genes encoding ETS transcription factors are commonly present, with androgen-regulated transmembrane protease, serine 2 (TMPRSS2)-v-ets erythroblastosis virus E26 oncogen homologue (ERG) gene fusion occurring in 40-70%. Studies on the predictive value of ERG rearrangement as detected by in-situ hybridization or polymerase chain reaction have resulted in varying outcomes. The objective of this study was to correlate immunohistochemical ERG protein expression with clinico-pathological parameters at radical prostatectomy specimens, and to determine its predictive value for postoperative disease recurrence and progression in a prostate cancer screening cohort. Since androgen receptor is downregulated by ERG in cell lines, we also compared the expression of respective proteins. We selected 481 participants from the European Randomized Study of Screening for Prostate Cancer treated by radical prostatectomy for prostate adenocarcinoma. A tissue microarray was constructed containing representative cores of all prostate cancer specimens as well as 22 xenografts and seven cell lines. Immunohistochemical expression of ERG and androgen receptor was correlated with prostate-specific antigen (PSA), Gleason sum, pT-stage, surgical margins, biochemical recurrence, local recurrence, overall death and disease-specific death. ERG expression was detected in 284 patients (65%). Expression occurred significantly more frequent in patients with PSA ≤10 ng/ml (P=0.024). There was no significant association between ERG and Gleason sum, pT-stage or surgical margin status. PSA (P=0.011), Gleason sum (P=0.003), pT-stage (P=0.001) and surgical margin status (P<0.001) all had independent value for postoperative biochemical recurrence, while positive surgical margin (P=0.021) was the only independent predictor for local recurrence. ERG protein expression did not have prognostic value for the clinical end points in uni- and multivariate analyses. A positive correlation existed between ERG and androgen receptor expression in single tissue cores (P<0.001). In conclusion, immunohistochemical ERG expression has no predictive value for prostate cancer recurrence or progression after radical prostatectomy. Increasing ERG levels are associated with the upregulation of androgen receptor expression in clinical specimens.

Non-prostatic pathology on prostate needle-biopsy - colorectal carcinoid: a case report.

INTRODUCTION: Prostate needle-biopsies are among the most common specimens in routine histopathological practice; in 15% colorectal tissue is also present. Rectal pathology is described to be found in 17% of this coincidentally obtained material. CASE PRESENTATION: We present a case in which colorectal carcinoid was found in the rectal mucosa obtained via transrectal prostate biopsies in a screening program for prostate cancer in a 71-year old Caucasian male. To the best of our knowledge, this was the first time that such a coincidental finding was discovered. Besides a colonoscopy with polypectomy, this coincidental detection remained without any further clinical consequences for this patient until today. CONCLUSION: As there is a considerable chance that abnormalities are found in the rectal tissue of prostate biopsies, it is advisable for all pathologists to include this tissue in the histology evaluation and look for potential irregularities in this simultaneously collected material.