Motivation While deep-learning algorithms have demonstrated outstanding performance in semantic image segmentation tasks, large annotation datasets are needed to create accurate models. Annotation of histology images is challenging due to the effort and experience required to carefully delineate tissue structures, and difficulties related to sharing and markup of whole-slide images. Results We recruited 25 participants, ranging in experience from senior pathologists to medical students, to delineate tissue regions in 151 breast cancer slides using the Digital Slide Archive. Inter-participant discordance was systematically evaluated, revealing low discordance for tumor and stroma, and higher discordance for more subjectively defined or rare tissue classes. Feedback provided by senior participants enabled the generation and curation of 20 000+ annotated tissue regions. Fully convolutional networks trained using these annotations were highly accurate (mean AUC=0.945), and the scale of annotation data provided notable improvements in image classification accuracy. Availability and Implementation Dataset is freely available at: https://goo.gl/cNM4EL. Supplementary information Supplementary data are available at Bioinformatics online.

Systemic metabolic alterations associated with increased consumption of saturated fat and obesity are linked with increased risk of prostate cancer progression and mortality, but the molecular underpinnings of this association are poorly understood. Here, we demonstrate in a murine prostate cancer model, that high-fat diet (HFD) enhances the MYC transcriptional program through metabolic alterations that favour histone H4K20 hypomethylation at the promoter regions of MYC regulated genes, leading to increased cellular proliferation and tumour burden. Saturated fat intake (SFI) is also associated with an enhanced MYC transcriptional signature in prostate cancer patients. The SFI-induced MYC signature independently predicts prostate cancer progression and death. Finally, switching from a high-fat to a low-fat diet, attenuates the MYC transcriptional program in mice. Our findings suggest that in primary prostate cancer, dietary SFI contributes to tumour progression by mimicking MYC over expression, setting the stage for therapeutic approaches involving changes to the diet.

Among prostate cancers containing Gleason pattern 4, cribriform morphology is associated with unfavorable clinicopathologic factors, but its genetic features and association with long-term outcomes are incompletely understood. In this study, genetic, transcriptional, and epigenetic features of invasive cribriform carcinoma (ICC) tumors were compared with non-cribriform Gleason 4 (NC4) in The Cancer Genome Atlas (TCGA) cohort. ICC (n ¼ 164) had distinctive molecular features when compared with NC4 (n ¼ 102). These include: (i) increased somatic copy number variations (SCNV), specifically deletions at 6q, 8p and 10q, which encompassed PTEN and MAP3K7 losses and gains at 3q; (ii) increased SPOP mut and ATM mut ; (iii) enrichment for mTORC1 and MYC pathways by gene expression; and (iv) increased methylation of selected genes. In addition, when compared with the metastatic prostate cancer, ICC clustered more closely to metastatic prostate cancer than NC4. Validation in clinical cohorts and genomically annotated murine models confirmed the association with SPOP mut (n ¼ 38) and PTEN loss (n ¼ 818). The association of ICC with lethal disease was evaluated in the Health Professionals Follow-up Study (HPFS) and Physicians' Health Study (PHS) prospective prostate cancer cohorts (median follow-up, 13.4 years; n ¼ 818). Patients with ICC were more likely to develop lethal cancer [HR, 1.62; 95% confidence interval (CI), 1.05-2.49], independent from Gleason score (GS).Implications: ICC has a distinct molecular phenotype that resembles metastatic prostate cancer and is associated with progression to lethal disease.

No abstract

Background Deep learning enables accurate high-resolution mapping of cells and tissue structures that can serve as the foundation of interpretable machine-learning models for computational pathology. However, generating adequate labels for these structures is a critical barrier, given the time and effort required from pathologists. Results This article describes a novel collaborative framework for engaging crowds of medical students and pathologists to produce quality labels for cell nuclei. We used this approach to produce the NuCLS dataset, containing >220,000 annotations of cell nuclei in breast cancers. This builds on prior work labeling tissue regions to produce an integrated tissue region- and cell-level annotation dataset for training that is the largest such resource for multi-scale analysis of breast cancer histology. This article presents data and analysis results for single and multi-rater annotations from both non-experts and pathologists. We present a novel workflow that uses algorithmic suggestions to collect accurate segmentation data without the need for laborious manual tracing of nuclei. Our results indicate that even noisy algorithmic suggestions do not adversely affect pathologist accuracy and can help non-experts improve annotation quality. We also present a new approach for inferring truth from multiple raters and show that non-experts can produce accurate annotations for visually distinctive classes. Conclusions This study is the most extensive systematic exploration of the large-scale use of wisdom-of-the-crowd approaches to generate data for computational pathology applications.

Purpose: Identifying cancers with high PI3K pathway activity is critical for treatment selection and eligibility into clinical trials of PI3K inhibitors. Assessments of tumor signaling pathway activity need to consider intratumoral heterogeneity and multiple regulatory nodes. Experimental Design: We established a novel, mechanistically informed approach to assessing tumor signaling pathways by quantifying single-cell–level multiplex immunofluorescence using custom algorithms. In a proof-of-concept study, we stained archival formalin-fixed, paraffin-embedded (FFPE) tissue from patients with primary prostate cancer in two prospective cohort studies, the Health Professionals Follow-up Study and the Physicians’ Health Study. PTEN, stathmin, and phospho-S6 were quantified on 14 tissue microarrays as indicators of PI3K activation to derive cell-level PI3K scores. Results: In 1,001 men, 988,254 tumor cells were assessed (median, 743 per tumor; interquartile range, 290–1,377). PI3K scores were higher in tumors with PTEN loss scored by a pathologist, higher Gleason grade, and a new, validated bulk PI3K transcriptional signature. Unsupervised machine-learning approaches resulted in similar clustering. Within-tumor heterogeneity in cell-level PI3K scores was high. During long-term follow-up (median, 15.3 years), rates of progression to metastases and death from prostate cancer were twice as high in the highest quartile of PI3K activation compared with the lowest quartile (hazard ratio, 2.04; 95% confidence interval, 1.13–3.68). Conclusions: Our novel pathway-focused approach to quantifying single-cell–level immunofluorescence in FFPE tissue identifies prostate tumors with PI3K pathway activation that are more aggressive and may respond to pathway inhibitors.

Whole-slide histology images contain information that is valuable for clinical and basic science investigations of cancer but extracting quantitative measurements from these images is challenging for researchers who are not image analysis specialists. In this article, we describe HistomicsML2, a software tool for learn-by-example training of machine learning classifiers for histologic patterns in whole-slide images. This tool improves training efficiency and classifier performance by guiding users to the most informative training examples for labeling and can be used to develop classifiers for prospective application or as a rapid annotation tool that is adaptable to different cancer types. HistomicsML2 runs as a containerized server application that provides web-based user interfaces for classifier training, validation, exporting inference results, and collaborative review, and that can be deployed on GPU servers or cloud platforms. We demonstrate the utility of this tool by using it to classify tumor-infiltrating lymphocytes in breast carcinoma and cutaneous melanoma. Significance: An interactive machine learning tool for analyzing digital pathology images enables cancer researchers to apply this tool to measure histologic patterns for clinical and basic science studies.

Motivation Nucleus detection, segmentation, and classification is fundamental to high-resolution mapping of the tumor microenvironment using whole-slide histopathology images. The growing interest in leveraging the power of deep learning to achieve state-of-the-art performance often comes at the cost of explainability, yet there is general consensus that explainability is critical for trustworthiness and widespread clinical adoption. Unfortunately, current explainability paradigms that rely on pixel saliency heatmaps or superpixel importance scores are not well-suited for nucleus classification. Techniques like Grad-CAM or LIME provide explanations that are indirect, qualitative, and/or non-intuitive to pathologists. Results In this paper, we present techniques to enable scalable nuclear detection, segmentation and explainable classification. First, we show how modifications to the widely used Mask R-CNN architecture, including decoupling the detection and classification tasks, improves accuracy and enables learning from hybrid annotation datasets like NuCLS, which contain mixtures of bounding boxes and segmentation boundaries. Second, we introduce an explainability method called Decision Tree Approximation of Learned Embeddings (DTALE), which provides explanations for classification model behavior globally, as well as for individual nuclear predictions. DTALE explanations are simple, quantitative, and can flexibly use any measurable morphological features that make sense to practicing pathologists, without sacrificing model accuracy. Together, these techniques present a step towards realizing the promise of computational pathology in computer-aided diagnosis and discovery of morphologic biomarkers. Availability Relevant code can be found at github.com/CancerDataScience/NuCLS Supplementary information Supplementary data are available at Bioinformatics online.