A5069504905- AI in cancer detection
- Radiomics and Machine Learning in Medical Imaging
- Molecular Biology Techniques and Applications
- Cancer-related molecular mechanisms research
- Cell Image Analysis Techniques
- Gene expression and cancer classification
- Prostate Cancer Diagnosis and Treatment
- Cancer, Lipids, and Metabolism
- Prostate Cancer Treatment and Research
- Genetics, Bioinformatics, and Biomedical Research
- Artificial Intelligence in Healthcare and Education
- RNA Research and Splicing
- Medical Image Segmentation Techniques
- Autopsy Techniques and Outcomes
- Cancer Genomics and Diagnostics
- 3D Printing in Biomedical Research
- Medical Imaging and Analysis
- Digital Imaging for Blood Diseases
- Cellular Mechanics and Interactions
- Urologic and reproductive health conditions
- 3D Shape Modeling and Analysis
- Advanced Fluorescence Microscopy Techniques
- Artificial Intelligence in Healthcare
- Electrospun Nanofibers in Biomedical Applications
- Image Processing Techniques and Applications
Karolinska Institutet
2020-2025
Science for Life Laboratory
2025
Svenska Örtmedicinska Institute
2023
Tampere University
2013-2022
HTW Berlin - University of Applied Sciences
2017
<h3>Importance</h3> Application of deep learning algorithms to whole-slide pathology images can potentially improve diagnostic accuracy and efficiency. <h3>Objective</h3> Assess the performance automated at detecting metastases in hematoxylin eosin–stained tissue sections lymph nodes women with breast cancer compare it pathologists' diagnoses a setting. <h3>Design, Setting, Participants</h3> Researcher challenge competition (CAMELYON16) develop solutions for node (November 2015-November...
Abstract Artificial intelligence (AI) has shown promise for diagnosing prostate cancer in biopsies. However, results have been limited to individual studies, lacking validation multinational settings. Competitions be accelerators medical imaging innovations, but their impact is hindered by lack of reproducibility and independent validation. With this mind, we organized the PANDA challenge—the largest histopathology competition date, joined 1,290 developers—to catalyze development...
Automatic Non-rigid Histological Image Registration (ANHIR) challenge was organized to compare the performance of image registration algorithms on several kinds microscopy histology images in a fair and independent manner. We have assembled 8 datasets, containing 355 with 18 different stains, resulting 481 pairs be registered. accuracy evaluated using manually placed landmarks. In total, 256 teams registered for challenge, 10 submitted results, 6 participated workshop. Here, we present...
Abstract Unreliable predictions can occur when an artificial intelligence (AI) system is presented with data it has not been exposed to during training. We demonstrate the use of conformal prediction detect unreliable predictions, using histopathological diagnosis and grading prostate biopsies as example. digitized 7788 from 1192 men in STHLM3 diagnostic study, used for training, 3059 676 testing. With prediction, 1 794 (0.1%) incorrect cancer (compared 14 errors [2%] without prediction)...
The alignment of tissue between histopathological whole-slide-images (WSI) is crucial for research and clinical applications. Advances in computing, deep learning, availability large WSI datasets have revolutionised analysis. Therefore, the current state-of-the-art registration unclear. To address this, we conducted ACROBAT challenge, based on largest dataset to date, including 4,212 WSIs from 1,152 breast cancer patients. challenge objective was align that stained with routine diagnostic...
Abstract Castration-resistant prostate cancers (CRPC) that arise after the failure of androgen-blocking therapies cause most deaths from cancer, intensifying need to fully understand CRPC pathophysiology. In this study, we characterized transcriptomic differences between untreated cancer and locally recurrent CRPC. Here, report identification 145 previously unannotated intergenic long noncoding RNA transcripts (lncRNA) or isoforms are associated with Of one third these were specific for...
Abstract Molecular profiling is central in cancer precision medicine but remains costly and based on tumor average profiles. Morphologic patterns observable histopathology sections from tumors are determined by the underlying molecular phenotype therefore have potential to be exploited for prediction of phenotypes. We report here first transcriptome-wide expression–morphology (EMO) analysis breast cancer, where individual deep convolutional neural networks were optimized validated mRNA...
Orientation and the degree of isotropy are important in many biological systems such as sarcomeres cardiomyocytes other fibrillar structures cytoskeleton. Image based analysis is often limited to qualitative evaluation by human experts, hampering throughput, repeatability reliability analyses. Software tools not readily available for this purpose existing methods typically rely at least partly on manual operation. We developed CytoSpectre, an automated tool spectral analysis, allowing...
Abstract The International Society of Urological Pathology (ISUP) hosts a reference image database supervised by experts with the purpose establishing an international standard in prostate cancer grading. Here, we aimed to identify areas grading difficulties and compare results those obtained from artificial intelligence system trained In series 87 needle biopsies cancers selected include problematic cases, failed reach 2/3 consensus 41.4% (36/87). Among non-consensus weighted kappa was 0.77...
Background and objective Image-based artificial intelligence (AI) methods have shown high accuracy in prostate cancer (PCa) detection. Their impact on patient outcomes cost effectiveness comparison to human pathologists remains unknown. Our aim was evaluate the cost-effectiveness of AI-assisted pathology for PCa diagnosis Sweden. Methods We modeled quadrennial prostate-specific antigen (PSA) screening men between ages 50 74 yr over a lifetime horizon using health care perspective. Men with...
Abstract Digital pathology has led to a demand for automated detection of regions interest, such as cancerous tissue, from scanned whole slide images. With accurate methods using image analysis and machine learning, significant speed‐up, savings in costs through increased throughput histological assessment could be achieved. This article describes learning approach tissue Our method is based on feature engineering supervised with random forest model. The features extracted the images include...
Adipose tissue is an attractive stem cell source for soft and bone engineering applications therapies. The adipose-derived stromal/stem cells (ASCs) have a multilineage differentiation capacity that regulated through extracellular signals. cellular events related to adhesion cytoskeleton been suggested as central regulators of fate decision. However, the detailed knowledge these molecular mechanisms in human ASCs remains limited. This study examined significance focal kinase (FAK),...
Abstract Motivation Digital pathology enables new approaches that expand beyond storage, visualization or analysis of histological samples in digital format. One novel opportunity is 3D histology, where a three-dimensional reconstruction the sample formed computationally based on serial tissue sections. This allows examining architecture 3D, for example, diagnostic purposes. Importantly, histology joint mapping cellular morphology with spatially resolved omics data true context at...
Molecular phenotyping by gene expression profiling is central in contemporary cancer research and molecular diagnostics but remains resource intense to implement. Changes occurring tumours cause morphological changes tissue, which can be observed on the microscopic level. The relationship between patterns some of phenotypes exploited predict from routine haematoxylin eosin-stained whole slide images (WSIs) using convolutional neural networks (CNNs). In this study, we propose a new,...
Abstract The presence of perineural invasion (PNI) by carcinoma in prostate biopsies has been shown to be associated with poor prognosis. assessment and quantification PNI are, however, labor intensive. To aid pathologists this task, we developed an artificial intelligence (AI) algorithm based on deep neural networks. We collected, digitized, pixel-wise annotated the findings each approximately 80,000 biopsy cores from 7406 men who underwent a screening trial between 2012 2014. In total, 485...
The analysis of FFPE tissue sections stained with haematoxylin and eosin (H&E) or immunohistochemistry (IHC) is essential for the pathologic assessment surgically resected breast cancer specimens. IHC staining has been broadly adopted into diagnostic guidelines routine workflows to assess status several established biomarkers, including ER, PGR, HER2 KI67. Biomarker can also be facilitated by computational pathology image methods, which have made numerous substantial advances recently, often...
Prostate cancer is a common type in men, yet some of its traits are still under-explored. One reason for this high molecular and morphological heterogeneity. The purpose study was to develop method gain new insights into the connection between changes underlying patterns. We used artificial intelligence (AI) analyze morphology seven hematoxylin eosin (H&E)-stained prostatectomy slides from patient with multi-focal prostate cancer. also paired spatially resolved expression thousands genes...
Virtual reality (VR) enables data visualization in an immersive and engaging manner, it can be used for creating ways to explore scientific data. Here, we use VR of 3D histology data, a novel interface digital pathology aid cancer research.Our contribution includes modeling whole organ embedded objects interest, fusing the models with associated quantitative features full resolution serial section patches, implementing virtual application. Our application is multi-scale nature, covering two...
None of the authors have any conflicts interest.
Polydimethylsiloxane (PDMS) is widely used in dynamic biological microfluidic applications. As a highly hydrophobic material, native PDMS does not support cell attachment and culture, especially conditions. Previous covalent coating methods use glutaraldehyde (GA) which, however, cytotoxic. This paper introduces novel simple method for binding collagen type I covalently on using ascorbic acid (AA) as cross-linker instead of GA. We compare the against physisorption GA cross-linker-based...