A population-level digital histologic biomarker for enhanced prognosis of invasive breast cancer

Breast cancer is a heterogeneous disease with variable survival outcomes. Pathologists grade the microscopic appearance of breast tissue using the Nottingham criteria, which are qualitative and do not account for noncancerous elements within the tumor microenvironment. Here we present the Histomic Prognostic Signature (HiPS), a comprehensive, interpretable scoring of the survival risk incurred by breast tumor microenvironment morphology. HiPS uses deep learning to accurately map cellular and tissue structures to measure epithelial, stromal, immune, and spatial interaction features. It was developed using a population-level cohort from the Cancer Prevention Study-II and validated using data from three independent cohorts, including the Prostate, Lung, Colorectal, and Ovarian Cancer trial, Cancer Prevention Study-3, and The Cancer Genome Atlas. HiPS consistently outperformed pathologists in predicting survival outcomes, independent of tumor-node-metastasis stage and pertinent variables. This was largely driven by stromal and immune features. In conclusion, HiPS is a robustly validated biomarker to support pathologists and improve patient prognosis.

Pitfalls in machine learning-based assessment of tumor-infiltrating lymphocytes in breast cancer: A report of the International Immuno-Oncology Biomarker Working Group on Breast Cancer

The clinical significance of the tumor-immune interaction in breast cancer is now established, and tumor-infiltrating lymphocytes (TILs) have emerged as predictive and prognostic biomarkers for patients with triple-negative (estrogen receptor, progesterone receptor, and HER2-negative) breast cancer and HER2-positive breast cancer. How computational assessments of TILs might complement manual TIL assessment in trial and daily practices is currently debated. Recent efforts to use machine learning (ML) to automatically evaluate TILs have shown promising results. We review state-of-the-art approaches and identify pitfalls and challenges of automated TIL evaluation by studying the root cause of ML discordances in comparison to manual TIL quantification. We categorize our findings into four main topics: (1) technical slide issues, (2) ML and image analysis aspects, (3) data challenges, and (4) validation issues. The main reason for discordant assessments is the inclusion of false-positive areas or cells identified by performance on certain tissue patterns or design choices in the computational implementation. To aid the adoption of ML for TIL assessment, we provide an in-depth discussion of ML and image analysis, including validation issues that need to be considered before reliable computational reporting of TILs can be incorporated into the trial and routine clinical management of patients with triple-negative breast cancer. © 2023 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

A Deep Learning Approach for Histology-Based Nucleus Segmentation and Tumor Microenvironment Characterization

Microscopic examination of pathology slides is essential to disease diagnosis and biomedical research. However, traditional manual examination of tissue slides is laborious and subjective. Tumor whole-slide image (WSI) scanning is becoming part of routine clinical procedures and produces massive data that capture tumor histologic details at high resolution. Furthermore, the rapid development of deep learning algorithms has significantly increased the efficiency and accuracy of pathology image analysis. In light of this progress, digital pathology is fast becoming a powerful tool to assist pathologists. Studying tumor tissue and its surrounding microenvironment provides critical insight into tumor initiation, progression, metastasis, and potential therapeutic targets. Nucleus segmentation and classification are critical to pathology image analysis, especially in characterizing and quantifying the tumor microenvironment (TME). Computational algorithms have been developed for nucleus segmentation and TME quantification within image patches. However, existing algorithms are computationally intensive and time consuming for WSI analysis. This study presents Histology-based Detection using Yolo (HD-Yolo), a new method that significantly accelerates nucleus segmentation and TME quantification. We demonstrate that HD-Yolo outperforms existing WSI analysis methods in nucleus detection, classification accuracy, and computation time. We validated the advantages of the system on 3 different tissue types: lung cancer, liver cancer, and breast cancer. For breast cancer, nucleus features by HD-Yolo were more prognostically significant than both the estrogen receptor status by immunohistochemistry and the progesterone receptor status by immunohistochemistry. The WSI analysis pipeline and a real-time nucleus segmentation viewer are available at https://github.com/impromptuRong/hd_wsi.

A population-level computational histologic signature for invasive breast cancer prognosis

Breast cancer is a heterogeneous disease with variable survival outcomes. Pathologists grade the microscopic appearance of breast tissue using the Nottingham criteria, which is qualitative and does not account for non-cancerous elements within the tumor microenvironment (TME). We present the Histomic Prognostic Signature (HiPS), a comprehensive, interpretable scoring of the survival risk incurred by breast TME morphology. HiPS uses deep learning to accurately map cellular and tissue structures in order to measure epithelial, stromal, immune, and spatial interaction features. It was developed using a population-level cohort from the Cancer Prevention Study (CPS)-II and validated using data from three independent cohorts, including the PLCO trial, CPS-3, and The Cancer Genome Atlas. HiPS consistently outperformed pathologists' performance in predicting survival outcomes, independent of TNM stage and pertinent variables. This was largely driven by stromal and immune features. In conclusion, HiPS is a robustly validated biomarker to support pathologists and improve prognosis.

Pilot study to evaluate tools to collect pathologist annotations for validating machine learning algorithms

Purpose: Validation of artificial intelligence (AI) algorithms in digital pathology with a reference standard is necessary before widespread clinical use, but few examples focus on creating a reference standard based on pathologist annotations. This work assesses the results of a pilot study that collects density estimates of stromal tumor-infiltrating lymphocytes (sTILs) in breast cancer biopsy specimens. This work will inform the creation of a validation dataset for the evaluation of AI algorithms fit for a regulatory purpose. Approach: Collaborators and crowdsourced pathologists contributed glass slides, digital images, and annotations. Here, "annotations" refer to any marks, segmentations, measurements, or labels a pathologist adds to a report, image, region of interest (ROI), or biological feature. Pathologists estimated sTILs density in 640 ROIs from hematoxylin and eosin stained slides of 64 patients via two modalities: an optical light microscope and two digital image viewing platforms. Results: The pilot study generated 7373 sTILs density estimates from 29 pathologists. Analysis of annotations found the variability of density estimates per ROI increases with the mean; the root mean square differences were 4.46, 14.25, and 26.25 as the mean density ranged from 0% to 10%, 11% to 40%, and 41% to 100%, respectively. The pilot study informs three areas of improvement for future work: technical workflows, annotation platforms, and agreement analysis methods. Upgrades to the workflows and platforms will improve operability and increase annotation speed and consistency. Conclusions: Exploratory data analysis demonstrates the need to develop new statistical approaches for agreement. The pilot study dataset and analysis methods are publicly available to allow community feedback. The development and results of the validation dataset will be publicly available to serve as an instructive tool that can be replicated by developers and researchers.

NuCLS: A scalable crowdsourcing approach and dataset for nucleus classification and segmentation in breast cancer

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.

Explainable nucleus classification using Decision Tree Approximation of Learned Embeddings

MOTIVATION

Nucleus detection, segmentation and classification are 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 nonintuitive to pathologists.

RESULTS

In this article, 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 toward realizing the promise of computational pathology in computer-aided diagnosis and discovery of morphologic biomarkers.

AVAILABILITY AND IMPLEMENTATION

Relevant code can be found at github.com/CancerDataScience/NuCLS.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

A Pathologist-Annotated Dataset for Validating Artificial Intelligence: A Project Description and Pilot Study

Purpose:

Validating artificial intelligence algorithms for clinical use in medical images is a challenging endeavor due to a lack of standard reference data (ground truth). This topic typically occupies a small portion of the discussion in research papers since most of the efforts are focused on developing novel algorithms. In this work, we present a collaboration to create a validation dataset of pathologist annotations for algorithms that process whole slide images. We focus on data collection and evaluation of algorithm performance in the context of estimating the density of stromal tumor-infiltrating lymphocytes (sTILs) in breast cancer.

Methods:

We digitized 64 glass slides of hematoxylin- and eosin-stained invasive ductal carcinoma core biopsies prepared at a single clinical site. A collaborating pathologist selected 10 regions of interest (ROIs) per slide for evaluation. We created training materials and workflows to crowdsource pathologist image annotations on two modes: an optical microscope and two digital platforms. The microscope platform allows the same ROIs to be evaluated in both modes. The workflows collect the ROI type, a decision on whether the ROI is appropriate for estimating the density of sTILs, and if appropriate, the sTIL density value for that ROI.

Results:

In total, 19 pathologists made 1645 ROI evaluations during a data collection event and the following 2 weeks. The pilot study yielded an abundant number of cases with nominal sTIL infiltration. Furthermore, we found that the sTIL densities are correlated within a case, and there is notable pathologist variability. Consequently, we outline plans to improve our ROI and case sampling methods. We also outline statistical methods to account for ROI correlations within a case and pathologist variability when validating an algorithm.

Conclusion:

We have built workflows for efficient data collection and tested them in a pilot study. As we prepare for pivotal studies, we will investigate methods to use the dataset as an external validation tool for algorithms. We will also consider what it will take for the dataset to be fit for a regulatory purpose: study size, patient population, and pathologist training and qualifications. To this end, we will elicit feedback from the Food and Drug Administration via the Medical Device Development Tool program and from the broader digital pathology and AI community. Ultimately, we intend to share the dataset, statistical methods, and lessons learned.

Image Analysis Pipeline for Renal Allograft Evaluation and Fibrosis Quantification

Introduction

Digital pathology improves the standardization and reproducibility of kidney biopsy specimen assessment. We developed a pipeline allowing the analysis of many images without requiring human preprocessing and illustrate its use with a simple algorithm for quantification of interstitial fibrosis on a large dataset of kidney allograft biopsy specimens.

Methods

Masson trichrome-stained images from kidney allograft biopsy specimens were used to train and validate a glomeruli detection algorithm using a VGG19 convolutional neural network and an automatic cortical region of interest (ROI) selection algorithm including cortical regions containing all predicted glomeruli. A positive-pixel count algorithm was used to quantify interstitial fibrosis on the ROIs and the association between automatic fibrosis and pathologist evaluation, estimated glomerular filtration rate (GFR) and allograft survival was assessed.

Results

The glomeruli detection (F1 score of 0.87) and ROIs selection (F1 score 0.83 [SD 0.13]) algorithms displayed high accuracy. The correlation between the automatic fibrosis quantification on manually and automatically selected ROIs was high (r = 1.00 [0.99-1.00]). Automatic fibrosis quantification was only moderately correlated with pathologists' assessment and was not significantly associated with eGFR or allograft survival.

Conclusion

This pipeline can automatically and accurately detect glomeruli and select cortical ROIs that can easily be used to develop, validate, and apply image analysis algorithms.

Impact of manual removal of the placenta and intrauterine cleaning during elective cesarean delivery on maternal infectious morbidity and blood loss

OBJECTIVE

to assess the impact of manual removal of the placenta and intrauterine cleaning during elective cesarean delivery on maternal infectious morbidity and blood loss.

METHODS

This prospective multicenter trial was conducted on 436 pregnant women at term who were intended for elective cesarean delivery and allocated into four groups; group 1 (n = 110) who underwent manual removal of the placenta with intrauterine cleaning, group 2 (n = 106) who underwent manual removal of the placenta without intrauterine cleaning, group 3 (n = 108) who underwent spontaneous placental delivery with intrauterine cleaning and group 4 (n = 112) who underwent spontaneous placental delivery without intrauterine cleaning. Maternal operative blood loss, the rate of endometritis and surgical site infections (SSIs) was assessed and recorded.

RESULTS

There was no significant difference between the four groups regarding drop of hemoglobin concentration, drop of hematocrit value, re-operation, re-admission to hospital, duration of hospital stay, the rate of endometritis and SSIs as well as maternal acceptability in terms of overall discomfort, overall satisfaction with delivery and recommendation to other women (p > .05). Intrauterine cleaning was associated with a significantly shorter duration of discharge of lochia and rapid return to daily activity compared to non-intrauterine cleaning (p < .05).

CONCLUSION

Manual removal of the placenta and intrauterine cleaning have no deleterious impact on maternal blood loss and infectious morbidity after elective cesarean section. Also, intrauterine cleaning was associated with faster cessation of lochia and faster return to daily activity.

Artificial intelligence and algorithmic computational pathology: an introduction with renal allograft examples

Whole slide imaging, which is an important technique in the field of digital pathology, has recently been the subject of increased interest and avenues for utilisation, and with more widespread whole slide image (WSI) utilisation, there will also be increased interest in and implementation of image analysis (IA) techniques. IA includes artificial intelligence (AI) and targeted or hypothesis-driven algorithms. In the overall pathology field, the number of citations related to these topics has increased in recent years. Renal pathology is one anatomical pathology subspecialty that has utilised WSIs and IA algorithms; it can be argued that renal transplant pathology could be particularly suited for whole slide imaging and IA, as renal transplant pathology is frequently classified by use of the semiquantitative Banff classification of renal allograft pathology. Hypothesis-driven/targeted algorithms have been used in the past for the assessment of a variety of features in the kidney (e.g. interstitial fibrosis, tubular atrophy, inflammation); in recent years, the amount of research has particularly increased in the area of AI/machine learning for the identification of glomeruli, for histological segmentation, and for other applications. Deep learning is the form of machine learning that is most often used for such AI approaches to the 'big data' of pathology WSIs, and deep learning methods such as artificial neural networks (ANNs)/convolutional neural networks (CNNs) are utilised. Unsupervised and supervised AI algorithms can be employed to accomplish image or semantic classification. In this review, AI and other IA algorithms applied to WSIs are discussed, and examples from renal pathology are covered, with an emphasis on renal transplant pathology.

Interactive Classification of Whole-Slide Imaging Data for Cancer Researchers

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.

Abstract PO-129: High expression of MKK3 correlates with poor clinical outcomes in African American breast cancer patients

African American women experience a 2-fold higher incidence of triple-negative breast cancer (TNBC) and are 40% more likely to die from breast cancer than women of other ethnicities. Previous studies have revealed a higher frequency of inactivating mutations in multiple tumor suppressor genes in African American breast cancer patients as compared to White patients. However, the molecular bases for the survival disparity in breast cancer remain unclear, and no race-specific therapeutically actionable targets have been proposed. To address this clinical need, we took the computational systems biology approach to uncover new potentially druggable genes that contribute in poor clinical outcomes of African American breast cancer patients. Through a comprehensive analysis of gene expression and survival data determined for The Cancer Genome Atlas (TCGA) breast cancer patient cohort, we found that more than 30% of all protein-coding genes are differentially expressed in White and African American breast cancer patients. The cluster analysis of co- regulated genes combined with the pathway enrichment analysis revealed a strong association of genes upregulated in African American patients with major oncogenic processes including cell cycle regulation, immunodeficiency, and oxidative phosphorylation. We have further identified more than 30 genes whose overexpression correlates with worsened clinical outcomes of African American breast cancer patients but not White patients. Among those genes, the overexpression of mitogen-activated protein kinase kinase 3 (MKK3) has one of the most dramatic and race-specific impacts on the survival of African American TNBC patients. We found that MKK3 can promote the TNBC tumorigenesis in African American patients in part by activating the master regulator MYC through protein- protein interaction. Together, this study uncovered new genetic features and molecular mechanisms of breast cancer tumorigenesis. As one example, we discovered new functional connectivity between poor clinical outcomes of African American TNBC patients and activation of the MYC transcriptional program through MKK3 signaling. These findings may open new opportunities for clinical investigations to reduce the survival disparity in breast cancer.

Citation Format: Xuan Yang, Mohamed Amgad, Lee A.D. Cooper, Yuhong Du, Haian Fu, Andrey A. Ivanov. High expression of MKK3 correlates with poor clinical outcomes in African American breast cancer patients [abstract]. In: Proceedings of the AACR Virtual Conference: Thirteenth AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2020 Oct 2-4. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2020;29(12 Suppl):Abstract nr PO-129.

Early Development of a Machine Learning Approach to Quantify MYC Immunohistochemical Staining in Lymphoma

Newer data suggest that double expression of MYC and BCL2 proteins (DE) evaluated by quantitative immunohistochemistry (qIHC) may be a powerful marker of worse prognosis in diffuse large B cell lymphoma (DLBCL). Testing for DE status, defined as &gt;40% MYC+ and &gt;50% BCL2+ tumor cells, is recommended in the WHO 2016 classification and clinical trials are using DE scoring to assign therapy arms. However, other data suggest that significant variability in manual DE scoring diminishes the predictive value. Error sources include high interobserver variability (IOV) associated with field choice, discrimination of tumor immunoreactivity from adjacent non-neoplastic cells, cell-to-cell variability in staining intensity, crush artifacts and necrosis. Thus, there is a need for standardized, reproducible approaches for DE scoring by qIHC. To address this need, we have begun developing a novel machine-learning approach to analyze IHC digital pathology whole-slide images, focusing initially on MYC IHC.

Digital whole-slide images (400x) of 22 DLBCL cases were uploaded to a web-based annotation platform. Using all cases, one annotator created 138 regions of interest (ROIs) containing approximately 200 nucleated cells and representing a variety of tissue types. Eight pathologists were assigned the same 10 ROIs in which to annotate all nuclei from which ground-truth seed nucleus labels (location, classification) were created for a validation set. Nuclei were classified as “tumor-positive”, “tumor-negative”, “non-tumor-positive”, “non-tumor-negative”, or “unknown”. This generated a set of 15,792 annotations with 1974 +/- 272 (Avg+/-STD) labels/annotator. Agglomerative hierarchical clustering afforded the creation of 2299 ground-truth seed locations. A maximum diameter of 3 mm/cluster was set by visual inspection of annotations. Of these seed locations, 1041 (45%) were detected by 8/8 annotators and, on average, 6/8 agreed on class. 302 +/- 72 (Avg+/-STD) “tumor positive” labels per annotator generated 382 seeds locations, 178 (47%) of which were detected by 8/8 annotators, with an average of 7.5/8 agreeing on class. 286 +/- 168 (Avg+/-STD) “tumor-negative” labels per annotator generated 336 seeds, 195 (58%) of which were detected by 8/8 annotators, with an average of 5/8 agreeing on class. Among all classes, the “tumor-positive” label displayed best overall label agreement whereas the “tumor-negative“ label yielded similar localization rate, but lower class agreement. These promising early findings provide a novel basis for quantifying IOV and utilizing multi-observer agreement to create a ground-truth validation set for a supervised machine learning approach to qIHC. Future efforts will make use of these data to optimize the validation set by rationally determining the number of additional annotations required, optimizing the number of annotators per ROI required, devising an adaptive approach to nuclear clustering based on nuclear density, and utilizing the additional 31,422 annotations in hand from all annotators as a robust algorithm training set.

High expression of MKK3 is associated with worse clinical outcomes in African American breast cancer patients

BACKGROUND

African American women experience a twofold higher incidence of triple-negative breast cancer (TNBC) and are 40% more likely to die from breast cancer than women of other ethnicities. However, the molecular bases for the survival disparity in breast cancer remain unclear, and no race-specific therapeutic targets have been proposed. To address this knowledge gap, we performed a systematic analysis of the relationship between gene mRNA expression and clinical outcomes determined for The Cancer Genome Atlas (TCGA) breast cancer patient cohort.

METHODS

The systematic differential analysis of mRNA expression integrated with the analysis of clinical outcomes was performed for 1055 samples from the breast invasive carcinoma TCGA PanCancer cohorts. A deep learning fully-convolutional model was used to determine the association between gene expression and tumor features based on breast cancer patient histopathological images.

RESULTS

We found that more than 30% of all protein-coding genes are differentially expressed in White and African American breast cancer patients. We have determined a set of 32 genes whose overexpression in African American patients strongly correlates with decreased survival of African American but not White breast cancer patients. Among those genes, the overexpression of mitogen-activated protein kinase kinase 3 (MKK3) has one of the most dramatic and race-specific negative impacts on the survival of African American patients, specifically with triple-negative breast cancer. We found that MKK3 can promote the TNBC tumorigenesis in African American patients in part by activating of the epithelial-to-mesenchymal transition induced by master regulator MYC.

CONCLUSIONS

The poor clinical outcomes in African American women with breast cancer can be associated with the abnormal elevation of individual gene expression. Such genes, including those identified and prioritized in this study, could represent new targets for therapeutic intervention. A strong correlation between MKK3 overexpression, activation of its binding partner and major oncogene MYC, and worsened clinical outcomes suggests the MKK3-MYC protein-protein interaction as a new promising target to reduce racial disparity in breast cancer survival.

Pitfalls in assessing stromal tumor infiltrating lymphocytes (sTILs) in breast cancer

Stromal tumor-infiltrating lymphocytes (sTILs) are important prognostic and predictive biomarkers in triple-negative (TNBC) and HER2-positive breast cancer. Incorporating sTILs into clinical practice necessitates reproducible assessment. Previously developed standardized scoring guidelines have been widely embraced by the clinical and research communities. We evaluated sources of variability in sTIL assessment by pathologists in three previous sTIL ring studies. We identify common challenges and evaluate impact of discrepancies on outcome estimates in early TNBC using a newly-developed prognostic tool. Discordant sTIL assessment is driven by heterogeneity in lymphocyte distribution. Additional factors include: technical slide-related issues; scoring outside the tumor boundary; tumors with minimal assessable stroma; including lymphocytes associated with other structures; and including other inflammatory cells. Small variations in sTIL assessment modestly alter risk estimation in early TNBC but have the potential to affect treatment selection if cutpoints are employed. Scoring and averaging multiple areas, as well as use of reference images, improve consistency of sTIL evaluation. Moreover, to assist in avoiding the pitfalls identified in this analysis, we developed an educational resource available at www.tilsinbreastcancer.org/pitfalls.

Report on computational assessment of Tumor Infiltrating Lymphocytes from the International Immuno-Oncology Biomarker Working Group

Assessment of tumor-infiltrating lymphocytes (TILs) is increasingly recognized as an integral part of the prognostic workflow in triple-negative (TNBC) and HER2-positive breast cancer, as well as many other solid tumors. This recognition has come about thanks to standardized visual reporting guidelines, which helped to reduce inter-reader variability. Now, there are ripe opportunities to employ computational methods that extract spatio-morphologic predictive features, enabling computer-aided diagnostics. We detail the benefits of computational TILs assessment, the readiness of TILs scoring for computational assessment, and outline considerations for overcoming key barriers to clinical translation in this arena. Specifically, we discuss: 1. ensuring computational workflows closely capture visual guidelines and standards; 2. challenges and thoughts standards for assessment of algorithms including training, preanalytical, analytical, and clinical validation; 3. perspectives on how to realize the potential of machine learning models and to overcome the perceptual and practical limits of visual scoring.

Machine-based detection and classification for bone marrow aspirate differential counts: initial development focusing on nonneoplastic cells

Bone marrow aspirate (BMA) differential cell counts (DCCs) are critical for the classification of hematologic disorders. While manual counts are considered the gold standard, they are labor intensive, time consuming, and subject to bias. A reliable automated counter has yet to be developed, largely due to the inherent complexity of bone marrow specimens. Digital pathology imaging coupled with machine learning algorithms represents a highly promising emerging technology for this purpose. Yet, training datasets for BMA cellular constituents, critical for building and validating machine learning algorithms, are lacking. Herein, we report our experience creating and employing such datasets to develop a machine learning algorithm to detect and classify BMA cells. Utilizing a web-based system that we developed for annotating and managing digital pathology images, over 10,000 cells from scanned whole slide images of BMA smears were manually annotated, including all classes that comprise the standard clinical DCC. We implemented a two-stage, detection and classification approach that allows design flexibility and improved classification accuracy. In a sixfold cross-validation, our algorithms achieved high overall accuracy in detection (0.959 ± 0.008 precision-recall AUC) and classification (0.982 ± 0.03 ROC AUC) using nonneoplastic samples. Testing on a small set of acute myeloid leukemia and multiple myeloma samples demonstrated similar detection and classification performance. In summary, our algorithms showed promising early results and represent an important initial step in the effort to devise a reliable, objective method to automate DCCs. With further development to include formal clinical validation, such a system has the potential to assist in disease diagnosis and prognosis, and significantly impact clinical practice.

Structured crowdsourcing enables convolutional segmentation of histology images

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.

Abstract 2436: Systematic computational analysis of histologic-genomic associations in triple-negative infiltrating ductal carcinomas of the breast

Background: Triple-negative breast cancer (TNBC) is characterized by rapid progression and lack of therapeutic targets. There is a pressing need for in-depth characterization of the biological correlates (and potential prognostic biomarkers) in TNBC. We performed a systematic analysis of genomic correlates of histologic markers of tumor aggressiveness and immune infiltration in 125 infiltrating ductal TNBC patients from The Cancer Genome Atlas.

Methods: Fully-convolutional networks of the VGG16-FCN8 architecture were trained to classify various tissue regions in H&E slides (0.956 AUC on unseen slides), and morphologic descriptors of tumor aggressiveness, invasion, and immune infiltration were extracted from predictions. Expression levels of 17,052 genes were entered into sample purity-adjusted linear regression models, and a Gene Set Enrichment Analysis was performed using the model coefficients to find gene set-level associations.

Results: Significant associations are summarized in Table 1. Expression of mTOR pathway genes (especially HIGD1C and PDE6H) is positively associated with large, dense tumor nests with a smooth tumor-stroma boundary. Boundary complexity is also positively associated with the oxidant stress response of NFE2L2. In contrast, some genes downregulated by the TGF-β pathway (including FGF6, PSG2 and CHRNG) are associated with a large tumor nest phenotype. The cell cycle regulator E2F1 (through PRM1, KRT72, and DBF4) is associated with dense immune infiltration, a known marker of good prognosis, and also small tumor nest size.

Conclusion: mTOR and NFE2L2-mediated mechanisms are significantly associated with features of tumor aggressiveness in TNBC, while E2F and some TGF-β targets are associated with morphological markers of less aggressive tumors. Further research is needed to elucidate the biological basis of these associations and their potential significance in therapeutic targeting of TNBC.

Table 1:Significant histologic-genomic associations in infiltrating-ductal TNBC.Histologic phenotype (feature description)Enriched gene set (MSigDB Oncogenic Signatures)Gene set descriptionNES(P-value, FDR)Leading-edge genesTumor nest size (Mean tumor nest area)TGFB_UP.V1_DNGenes down-regulated by TGFB11.55(p&lt;0.001, FDR=0.036)FGF6; PSG2; CHRNGTumor-stroma interface (non-)complexity (mean solidity of tumor nests)NFE2L2.V2Genes upregulated with knockout of NFE2L2 gene (involved in oxidant stress response and inflammation)-1.51(p&lt;0.001, FDR=0.027)DEFB119; CNTNAP5; SCGB1D1MTOR_UP.V1_DNGenes downregulated by everolimus (an mTOR inhibitor)1.43(p=0.0017, FDR= 0.094)HIGD1CSmall, solitary tumor nestsE2F1_UP.V1_UPGenes up-regulated when E2F1 is over-expressed (cell cycle regulation)1.48(p&lt;0.001, FDR&lt;0.001)PRM1MTOR_UP.V1_DNGenes downregulated by everolimus (an mTOR inhibitor)-1.50(p=0.0027, FDR= 0.079)PDE6H; HIGD1CSmall tumor nests with abundant surrounding immune infiltration (a spatial descriptor meant to capture immune success)E2F1_UP.V1_UPGenes up-regulated when E2F1 is over-expressed.(cell cycle regulation)1.55(p&lt;0.001, FDR= 0.0019)PRM1; KRT72; DBF4

Citation Format: Mohamed Amgad, Uday Kurkure, Habiba Elfandy, Hagar H. Khallaf, David A. Gutman, Carlos S. Moreno, Michael Barnes, Lee A. Cooper. Systematic computational analysis of histologic-genomic associations in triple-negative infiltrating ductal carcinomas of the breast [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 2436.

Abstract P5-07-01: Computational scoring of tumor infiltrating lymphocytes in triple-negative breast cancer

Background: Stromal Tumor Infiltrating Lymphocytes (sTIL) are an established prognostic feature in triple-negative breast cancer, yet manual assessment or visual estimation of sTILs with conventional light microscopy may be subject to inter-pathologist variability. Recently published guidelines by the International TIL Working Group help address inter-pathologist variability, yet there remains a need for more objective and quantitative computational sTIL scoring.

Methods: Our study used 120 triple-negative breast cancer slides (one slide per patient). A deep-learning based image analysis workflow is used to perform segmentation and classification of tissue regions and cells on the digital whole slide image. We used 14 annotated slides to train and validate the deep learning model, and to obtain image segmentation and classification accuracy statistics. Non-training slides were used to evaluate the concordance of manual (m-sTIL) and computationally derived (c-sTIL) scores. To generate data to create the model we manually annotated tissue regions in FFPE H&E stained digital slides, including: tumor, stroma, and necrosis. Initial classification of cell nuclei was performed using a semi-automated image analysis method, and then manually corrected to generate ground truth for tumor, stroma (fibroblasts), and lymphocytes. All annotations were performed by a trained research fellow and reviewed by a board-certified pathologist. Corresponding region and nucleus-level annotations were combined to train and validate a fully-convolutional neural network that jointly classifies tissue regions and cell nuclei. Tissue region segmentation accuracy was assessed by the Dice coefficient to measure degree of overlap between predicted tissue regions and ground truth annotations. Cell classification accuracy was assessed using area under curve (AUC). Two board-certified pathologists independently generated an m-sTIL score for all slides according to clinical guidelines, and discrepancies between pathologists were resolved by consensus. c-sTIL scores were calculated as the percentage of classified stromal areas occupied by nuclei classified as lymphocytic infiltrates.

Results: Tissue region segmentation was accurate for both stroma (0.77 Dice) and tumor (0.83 Dice) regions, and accurate overall (0.78 Dice). Cell classification was highly accurate for lymphocytes (0.89 AUC), tumor cells (0.90 AUC), stromal cells (0.78 AUC), and overall (0.89 AUC, micro average). Inter observer spearman correlation between the m-sTIL scores of our two pathologists was 0.66 (p &lt; 0.001). By comparison, the correlation between c-sTIL and consensus m-sTIL was higher at 0.73 (p &lt; 0.001). Dichotomizing at a threshold sTIL score of 10%, c-sTIL scoring identifies low-sTIL patients with an accuracy of 85%. High- and Low- sTIL score patient groups show clear separation on a Kaplan-Meier curve for both c-sTIL and m-sTIL scoring approaches.

Conclusions: Our pipeline quantifies stromal TILs with high concordance with manual pathologist scores, and sheds light on the ability of computational approaches in standardizing diagnostic pathology workflows. Future work will investigate how other computationally driven histology biomarkers can predict outcomes and help prognosticate breast cancer patients.

Citation Format: Amgad M, Sarkar A, Srinivas C, Redman R, Ratra S, Bechert CJ, Calhoun BC, Mrazeck K, Kurkure U, Cooper LA, Barnes M. Computational scoring of tumor infiltrating lymphocytes in triple-negative breast cancer [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P5-07-01.

Joint Region and Nucleus Segmentation for Characterization of Tumor Infiltrating Lymphocytes in Breast Cancer

Histologic assessment of stromal tumor infiltrating lymphocytes (sTIL) as a surrogate of the host immune response has been shown to be prognostic and potentially chemo-predictive in triple-negative and HER2-positive breast cancers. The current practice of manual assessment is prone to intra- and inter-observer variability. Furthermore, the inter-play of sTILs, tumor cells, other microenvironment mediators, their spatial relationships, quantity, and other image-based features have yet to be determined exhaustively and systemically. Towards analysis of these aspects, we developed a deep learning based method for joint region-level and nucleus-level segmentation and classification of breast cancer H&E tissue whole slide images. Our proposed method simultaneously identifies tumor, fibroblast, and lymphocyte nuclei, along with key histologic region compartments including tumor and stroma. We also show how the resultant segmentation masks can be combined with seeding approaches to yield accurate nucleus classifications. Furthermore, we outline a simple workflow for calibrating computational scores to human scores for consistency. The pipeline identifies key compartments with high accuracy (Dice= overall: 0.78, tumor: 0.83, and fibroblasts: 0.77). ROC AUC for nucleus classification is high at 0.89 (micro-average), 0.89 (lymphocytes), 0.90 (tumor), and 0.78 (fibroblasts). Spearman correlation between computational sTIL and pathologist consensus is high (R=0.73, p<0.001) and is higher than inter-pathologist correlation (R=0.66, p<0.001). Both manual and computational sTIL scores successfully stratify patients by clinical progression outcomes.

Publisher Correction: Management of Low and Intermediate Risk Adult Rhabdomyosarcoma: A Pooled Survival Analysis of 553 Patients

A correction to this article has been published and is linked from the HTML and PDF versions of this paper. The error has been fixed in the paper.

Management of Low and Intermediate Risk Adult Rhabdomyosarcoma: A Pooled Survival Analysis of 553 Patients

This is the second-largest retrospective analysis addressing the controversy of whether adult rhabdomyosarcoma (RMS) should be treated with chemotherapy regimens adopted from pediatric RMS protocols or adult soft-tissue sarcoma protocols. A comprehensive database search identified 553 adults with primary non-metastatic RMS. Increasing age, intermediate-risk disease, no chemotherapy use, anthacycline-based and poor chemotherapy response were significant predictors of poor overall and progression-free survival. In contrast, combined cyclophosphamide-based, cyclophosphamide + anthracycline-based, or cyclophosphamide + ifosfamide + anthracycline-based regimens significantly improved outcomes. Intermediate-risk disease was a significant predictor of poor chemotherapy response. Overall survival of clinical group-III patients was significantly improved if they underwent delayed complete resection. Non-parameningeal clinical group-I patients had the best local control, which was not affected by additional adjuvant radiotherapy. This study highlights the superiority of chemotherapy regimens –adapted from pediatric protocols- compared to anthracycline-based regimens. There is lack of data to support the routine use of adjuvant radiotherapy for non-parameningeal group-I patients. Nonetheless, intensive local therapy should be always considered for those at high risk for local recurrence, including intermediate-risk disease, advanced IRS stage, large tumors or narrow surgical margins. Although practically difficult (due to tumor’s rarity), there is a pressing need for high quality randomized controlled trials to provide further guidance.

Interactive phenotyping of large-scale histology imaging data with HistomicsML

Whole-slide imaging of histologic sections captures tissue microenvironments and cytologic details in expansive high-resolution images. These images can be mined to extract quantitative features that describe tissues, yielding measurements for hundreds of millions of histologic objects. A central challenge in utilizing this data is enabling investigators to train and evaluate classification rules for identifying objects related to processes like angiogenesis or immune response. In this paper we describe HistomicsML, an interactive machine-learning system for digital pathology imaging datasets. This framework uses active learning to direct user feedback, making classifier training efficient and scalable in datasets containing 10⁸+ histologic objects. We demonstrate how this system can be used to phenotype microvascular structures in gliomas to predict survival, and to explore the molecular pathways associated with these phenotypes. Our approach enables researchers to unlock phenotypic information from digital pathology datasets to investigate prognostic image biomarkers and genotype-phenotype associations.

Molecular underpinnings of corneal angiogenesis: advances over the past decade

The cornea is maintained in an avascular state by maintaining an environment whereby anti-angiogenic factors take the upper hand over factors promoting angiogenesis. Many of the common pathologies affecting the cornea involve the disruption of such equilibrium and the shift towards new vessel formation, leading to corneal opacity and eventually-vision loss. Therefore it is of paramount importance that the molecular underpinnings of corneal neovascularization (CNV) be clearly understood, in order to develop better targeted treatments. This article is a review of the literature on the recent discoveries regarding pro-angiogenic factors of the cornea (such as vascular endothelial growth factors, fibroblast growth factor and matrix metalloproteinases) and anti-angiogenic factors of the cornea (such as endostatins and neostatins). Further, we review the molecular underpinnings of lymphangiogenesis, a process now known to be almost separate from (yet related to) hemangiogenesis.

Host immune cellular reactions in corneal neovascularization

Corneal neovascularization (CNV) is a global important cause of visual impairment. The immune mechanisms leading to corneal heme- and lymphangiogenesis have been extensively studied over the past years as more attempts were made to develop better prophylactic and therapeutic measures. This article aims to discuss immune cells of particular relevance to CNV, with a focus on macrophages, Th17 cells, dendritic cells and the underlying immunology of common pathologies involving neovascularization of the cornea. Hopefully, a thorough understanding of these topics would propel the efforts to halt the detrimental effects of CNV.

Extending Ripley's K-Function to Quantify Aggregation in 2-D Grayscale Images

In this work, we describe the extension of Ripley's K-function to allow for overlapping events at very high event densities. We show that problematic edge effects introduce significant bias to the function at very high densities and small radii, and propose a simple correction method that successfully restores the function's centralization. Using simulations of homogeneous Poisson distributions of events, as well as simulations of event clustering under different conditions, we investigate various aspects of the function, including its shape-dependence and correspondence between true cluster radius and radius at which the K-function is maximized. Furthermore, we validate the utility of the function in quantifying clustering in 2-D grayscale images using three modalities: (i) Simulations of particle clustering; (ii) Experimental co-expression of soluble and diffuse protein at varying ratios; (iii) Quantifying chromatin clustering in the nuclei of wt and crwn1 crwn2 mutant Arabidopsis plant cells, using a previously-published image dataset. Overall, our work shows that Ripley's K-function is a valid abstract statistical measure whose utility extends beyond the quantification of clustering of non-overlapping events. Potential benefits of this work include the quantification of protein and chromatin aggregation in fluorescent microscopic images. Furthermore, this function has the potential to become one of various abstract texture descriptors that are utilized in computer-assisted diagnostics in anatomic pathology and diagnostic radiology.

Medical Student Research: An Integrated Mixed-Methods Systematic Review and Meta-Analysis

Importance

Despite the rapidly declining number of physician-investigators, there is no consistent structure within medical education so far for involving medical students in research.

Objective

To conduct an integrated mixed-methods systematic review and meta-analysis of published studies about medical students' participation in research, and to evaluate the evidence in order to guide policy decision-making regarding this issue.

Evidence Review

We followed the PRISMA statement guidelines during the preparation of this review and meta-analysis. We searched various databases as well as the bibliographies of the included studies between March 2012 and September 2013. We identified all relevant quantitative and qualitative studies assessing the effect of medical student participation in research, without restrictions regarding study design or publication date. Prespecified outcome-specific quality criteria were used to judge the admission of each quantitative outcome into the meta-analysis. Initial screening of titles and abstracts resulted in the retrieval of 256 articles for full-text assessment. Eventually, 79 articles were included in our study, including eight qualitative studies. An integrated approach was used to combine quantitative and qualitative studies into a single synthesis. Once all included studies were identified, a data-driven thematic analysis was performed.

Findings and Conclusions

Medical student participation in research is associated with improved short- and long- term scientific productivity, more informed career choices and improved knowledge about-, interest in- and attitudes towards research. Financial worries, gender, having a higher degree (MSc or PhD) before matriculation and perceived competitiveness of the residency of choice are among the factors that affect the engagement of medical students in research and/or their scientific productivity. Intercalated BSc degrees, mandatory graduation theses and curricular research components may help in standardizing research education during medical school.

Clinical correlates of common corneal neovascular diseases: a literature review

A large subset of corneal pathologies involves the formation of new blood and lymph vessels (neovascularization), leading to compromised visual acuity. This article aims to review the clinical causes and presentations of corneal neovascularization (CNV) by examining the mechanisms behind common CNV-related corneal pathologies, with a particular focus on herpes simplex stromal keratitis, contact lenses-induced keratitis and CNV secondary to keratoplasty. Moreover, we reviewed CNV in the context of different types of corneal transplantation and keratoprosthesis, and summarized the most relevant treatments available so far.

Development of an Arabic version of the National Eye Institute Visual Function Questionnaire as a tool to study eye diseases patients in Egypt

AIM

To develop and test an Arabic version of the National Eye Institute Visual Function Questionnaire-25 (NEI-VFQ-25).

METHODS

NEI-VFQ-25 was translated into Arabic according to WHO translation guidelines. We enrolled adult consenting patients with bilateral chronic eye diseases who presented to 14 hospitals across Egypt from October to December 2012, and documented their clinical findings. Psychometric properties were then tested using STATA.

RESULTS

We recruited 379 patients, whose mean age was (54.5±15)y. Of 46.2% were males, 227 had cataract, 31 had glaucoma, 23 had retinal detachment, 37 had diabetic retinopathy, and 61 had miscellaneous visual defects. Non-response rate and the floor and ceiling numbers of the Arabic version (ARB-VFQ-25) were calculated. Internal consistency was high in all subscales (except general health), with Cronbach-α ranging from 0.702-0.911. Test-retest reliability was high (intraclass correlation coefficient 0.79).

CONCLUSION

ARB-VFQ-25 is a reliable and valid tool for assessing visual functions of Arabic speaking patients. However, some questions had high non-response rates and should be substituted by available alternatives. Our results support the importance of including self-reported visual functions as part of routine ophthalmologic examination.

Integrating web 2.0 in clinical research education in a developing country

The use of Web 2.0 tools in education and health care has received heavy attention over the past years. Over two consecutive years, Children's Cancer Hospital - Egypt 57357 (CCHE 57357), in collaboration with Egyptian universities, student bodies, and NGOs, conducted a summer course that supports undergraduate medical students to cross the gap between clinical practice and clinical research. This time, there was a greater emphasis on reaching out to the students using social media and other Web 2.0 tools, which were heavily used in the course, including Google Drive, Facebook, Twitter, YouTube, Mendeley, Google Hangout, Live Streaming, Research Electronic Data Capture (REDCap), and Dropbox. We wanted to investigate the usefulness of integrating Web 2.0 technologies into formal educational courses and modules. The evaluation survey was filled in by 156 respondents, 134 of whom were course candidates (response rate = 94.4 %) and 22 of whom were course coordinators (response rate = 81.5 %). The course participants came from 14 different universities throughout Egypt. Students' feedback was positive and supported the integration of Web 2.0 tools in academic courses and modules. Google Drive, Facebook, and Dropbox were found to be most useful.

Cancer education for medical students in developing countries: where do we stand and how to improve?

BACKGROUND

This article is a review of the literature regarding the state of oncology education for medical students in developing countries, and possible solutions to the problems at hand.

METHODS

Ovid MEDLINE, PubMed, ERIC, The Cochrane CENTRAL Register of Controlled Trials (CENTRAL) and Google Scholar were searched using the terms oncology, undergraduate, cancer, education and teaching.

RESULTS

The search resulted in 40 relevant articles in total. Ten articles showed that there is a lack of adequate knowledge in the scientific, clinical and psychological aspects of oncology and palliative care amongst students and physicians in developing countries. Eight articles describe the relevance and usefulness of summer schools, workshops and trainings. The rest of them discuss possible methods of addressing the issue, the most important of which is the inclusion of a clinical oncology rotation in the undergraduate syllabus.

CONCLUSION

Graduated physicians and medical students are a long way from reaching the standard knowledge and skills required in oncology. Thus, there is a pressing need to reform the undergraduate medical curricula in developing countries in order to increase cancer awareness for better graduated future physicians.