A5073713636- AI in cancer detection
- Radiomics and Machine Learning in Medical Imaging
- Prostate Cancer Diagnosis and Treatment
- Cell Image Analysis Techniques
- Molecular Biology Techniques and Applications
- Digital Imaging for Blood Diseases
- Advanced X-ray and CT Imaging
- Prostate Cancer Treatment and Research
- Gene expression and cancer classification
- MRI in cancer diagnosis
- Colorectal Cancer Screening and Detection
- Medical Image Segmentation Techniques
- Single-cell and spatial transcriptomics
- Cancer Genomics and Diagnostics
- Cancer, Lipids, and Metabolism
- Arsenic contamination and mitigation
- Chromium effects and bioremediation
- Renal and Vascular Pathologies
- Systemic Sclerosis and Related Diseases
- Breast Cancer Treatment Studies
- Microbial Community Ecology and Physiology
- Medical Imaging Techniques and Applications
- Immunotherapy and Immune Responses
- Machine Learning in Healthcare
- Soil and Unsaturated Flow
University Hospital of Geneva
2023-2025
Emory University
2022-2025
Georgia Institute of Technology
2022-2025
SIB Swiss Institute of Bioinformatics
2020-2024
Hôpital Beau-Séjour
2024
University of Lausanne
2019-2023
Case Western Reserve University
2013-2023
University of Bern
2023
The Netherlands Cancer Institute
2023
Maastricht University Medical Centre
2023
Background: Deep learning (DL) is a representation approach ideally suited for image analysis challenges in digital pathology (DP). The variety of tasks the context DP includes detection and counting (e.g., mitotic events), segmentation nuclei), tissue classification cancerous vs. non-cancerous). Unfortunately, issues with slide preparation, variations staining scanning across sites, vendor platforms, as well biological variance, such presentation different grades disease, make these...
Digital pathology (DP), referring to the digitization of tissue slides, is beginning change landscape clinical diagnostic workflows and has engendered active research within area computational pathology. One challenges in DP presence artefacts batch effects, unintentionally introduced during both routine slide preparation (eg, staining, folding) blurriness, variations contrast hue). Manual review glass digital slides laborious, qualitative, subject intra- inter-reader variability. Therefore,...
The application of deep learning for automated segmentation (delineation boundaries) histologic primitives (structures) from whole slide images can facilitate the establishment novel protocols kidney biopsy assessment. Here, we developed and validated networks structures on biopsies nephrectomies. For development, examined 125 Minimal Change Disease collected across 29 NEPTUNE enrolling centers along with 459 stained Hematoxylin & Eosin (125), Periodic Acid Schiff Silver (102), Trichrome...
Early-stage estrogen receptor-positive (ER+) breast cancer (BCa) is the most common type of BCa in United States. One critical question with these tumors identifying which patients will receive added benefit from adjuvant chemotherapy. Nuclear pleomorphism (variance nuclear shape and morphology) an important constituent grading schemes, ER+ cases, grade highly correlated disease outcome. This study aimed to investigate whether quantitative computer-extracted image features orientation on...
Identification of patients with early stage non-small cell lung cancer (NSCLC) high risk recurrence could help identify who would receive additional benefit from adjuvant therapy. In this work, we present a computational histomorphometric image classifier using nuclear orientation, texture, shape, and tumor architecture to predict disease in NSCLC digitized H&E tissue microarray (TMA) slides. Using retrospective cohort (Cohort #1, n = 70), constructed supervised classification model...
Over 26 million people worldwide suffer from heart failure annually. When the cause of cannot be identified, endomyocardial biopsy (EMB) represents gold-standard for evaluation disease. However, manual EMB interpretation has high inter-rater variability. Deep convolutional neural networks (CNNs) have been successfully applied to detect cancer, diabetic retinopathy, and dermatologic lesions images. In this study, we develop a CNN classifier clinical H&E stained whole-slide images total 209...
Gene-expression companion diagnostic tests, such as the Oncotype DX test, assess risk of early stage Estrogen receptor (ER) positive (+) breast cancers, and guide clinicians in decision whether or not to use chemotherapy. However, these tests are typically expensive, time consuming, tissue-destructive.In this paper, we evaluate ability computer-extracted nuclear morphology features from routine hematoxylin eosin (H&E) stained images 178 ER+ cancer patients predict corresponding categories...
Abstract Early stage estrogen receptor positive (ER+) breast cancer (BCa) treatment is based on the presumed aggressiveness and likelihood of recurrence. Oncotype DX (ODX) other gene expression tests have allowed for distinguishing more aggressive ER+ BCa requiring adjuvant chemotherapy from less cancers benefiting hormonal therapy alone. However these are expensive, tissue destructive require specialized facilities. Interestingly grade has been shown to be correlated with ODX risk score....
Bile acids, which are synthesized from cholesterol by the liver, chemically transformed along intestinal tract gut microbiota, and products of these transformations signal through host receptors, affecting overall health. These include bile acid deconjugation, oxidation, 7α-dehydroxylation. An understanding biogeography in is critical because deconjugation a prerequisite for 7α-dehydroxylation most microorganisms harbor transformation capacity. Here, we used coupled metabolomic metaproteomic...
Prostate cancer treatment planning is largely dependent upon examination of core-needle biopsies. The microscopic architecture the prostate glands forms basis for prognostic grading by pathologists. Interpretation these convoluted three-dimensional (3D) glandular structures via visual inspection a limited number two-dimensional (2D) histology sections often unreliable, which contributes to under- and overtreatment patients. To improve risk assessment decisions, we have developed workflow...
Soil microbiomes harbour unparalleled functional and phylogenetic diversity. However, extracting isolates with a targeted function from complex is not straightforward, particularly if the associated phenotype does lend itself to high-throughput screening. Here, we tackle methylation of arsenic (As) in anoxic soils. As was proposed be catalysed by sulfate-reducing bacteria. date, there are no available anaerobic capable methylation, whether or otherwise. The isolation such microorganism has...
Abstract The treatment and management of early stage estrogen receptor positive (ER+) breast cancer is hindered by the difficulty in identifying patients who require adjuvant chemotherapy contrast to those that will respond hormonal therapy. To distinguish between more less aggressive tumors, which a fundamental criterion for selection an appropriate plan, Oncotype DX (ODX) other gene expression tests are typically employed. While informative, these expensive, tissue destructive, specialized...
Intratumoural heterogeneity has been previously shown to be related clonal evolution and genetic instability associated with tumour progression. Phenotypically, it is reflected in the diversity of appearance morphology within cell populations. Computer-extracted features relating cellular on routine tissue images might correlate outcome. This study investigated prognostic ability computer-extracted (CellDiv) from haematoxylin eosin (H&E)-stained histology non-small lung carcinomas...
Abstract Purpose: Between 30%–40% of patients with prostate cancer experience disease recurrence following radical prostatectomy. Existing clinical models for risk prediction do not account population-based variation in the tumor phenotype, despite recent evidence suggesting presence a unique, more aggressive phenotype African American (AA) patients. We investigated capacity digitally measured, population-specific phenotypes intratumoral stroma to create improved Experimental Design: This...
Abstract Aim Allograft rejection is a serious concern in heart transplant medicine. Though endomyocardial biopsy with histological grading the diagnostic standard for rejection, poor inter-pathologist agreement creates significant clinical uncertainty. The aim of this investigation to demonstrate that cellular grades generated via computational analysis are on-par those provided by expert pathologists Methods and results study cohort consisted 2472 slides originating from three major US...
Abstract Background Image-based machine learning tools hold great promise for clinical applications in pathology research. However, the ideal end-users of these computational (e.g., pathologists and biological scientists) often lack programming experience required setup use which rely on command line interfaces. Methods We have developed Histo-Cloud , a tool segmentation whole slide images (WSIs) that has an easy-to-use graphical user interface. This runs state-of-the-art convolutional...
Abstract Prostate cancer treatment decisions rely heavily on subjective visual interpretation [assigning Gleason patterns or International Society of Urological Pathology (ISUP) grade groups] limited numbers two‐dimensional (2D) histology sections. Under this paradigm, interobserver variance is high, with ISUP grades not correlating well outcome for individual patients, and contributes to the over‐ undertreatment patients. Recent studies have demonstrated improved prognostication prostate...
Deep learning (DL) has recently been successfully applied to a number of image analysis problems. However, DL approaches tend be inefficient for segmentation on large data, such as high-resolution digital pathology slide images. For example, typical breast biopsy images scanned at 40 magnification contain billions pixels, which usually only small percentage belong the class interest. naïve deep scheme, parsing through and interrogating all pixels would represent hundreds if not thousands...
Neoadjuvant chemotherapy (NAC) is routinely used to treat breast tumors before surgery reduce tumor size and improve outcome. However, no current clinical or imaging metrics can effectively predict treatment which NAC recipients will achieve pathological complete response (pCR), the absence of residual invasive disease in lymph nodes following surgical resection. In this work, we developed applied a convolu- tional neural network (CNN) pCR from pre-treatment dynamic contrast-enhanced...
Inconsistencies in the preparation of histology slides and whole-slide images (WSIs) may lead to challenges with subsequent image analysis machine learning approaches for interrogating WSI. These variabilities are especially pronounced multicenter cohorts, where batch effects (i.e. systematic technical artifacts unrelated biological variability) introduce biases algorithms. To date, manual quality control (QC) has been de facto standard dataset curation, but remains highly subjective is too...