Biology subjects

Truhn, D.

Publications and source records attributed to Truhn, D..

7 recordsLinked to original sources

bioRxiv2026

Oncodevelopmental plasticity of the skeleton in myeloid neoplasms

Myelofibrosis in patients with myeloproliferative neoplasms (MPNs) is traditionally characterized by bone marrow fibrosis and osteosclerosis, with de novo bone formation commonly attributed to impaired osteoclast-mediated resorption. Here, we challenge this paradigm by demonstrating that a solitary clonal driver mutation simultaneously induces pathological bone formation and resorption, with osteosclerosis acting to conceal localized and active bone destruction rather than inhibiting it. Through population analysis; clinical imaging; patient-derived multi-tissue sequencing; murine models and organ-on-a-chip systems, we demonstrate that spatial and ontogeny-dependent remodeling in mesoderm- and neural crest-derived bones is mechanistically interconnected via a previously unidentified osteochondral stromal injury program. Neural crest-derived stromal cells suppress osteogenic programs and undergo injury-induced lineage plasticity with ectopic chondrogenesis, mirroring pathological remodeling in mesoderm-derived growth plate regions. This shared injury response promotes osteoclastogenesis and is mediated by a conserved Thrombospondin 1+ (THBS1+) stromal population that links fibrotic remodeling to bone loss. Combined pharmacological inhibition of THBS1 and JAK signaling reduces myeloproliferation, halts fibrosis progression, and restores two developmentally distinct bones, establishing THBS1 as a unifying therapeutic target in myelofibrosis.

cancer biology↗

bioRxiv2024

Counterfactual Diffusion Models for Mechanistic Explainability of Artificial Intelligence Models in Pathology

Deep learning can extract predictive and prognostic biomarkers from histopathology whole slide images. However, explainable artificial intelligence approaches widely used in digital pathology, such as attention heatmaps and class activation mapping, offer only limited interpretability regarding the features captured by classifiers. Here, we present MoPaDi (Morphing histoPathology Diffusion), a framework for generating counterfactual explanations for histopathology images that reveal which morphological or style features drive classifier predictions. MoPaDi combines diffusion autoencoders with task-specific multiple instance learning classifiers to manipulate images and flip predictions by modifying relevant features. We evaluated the framework on multiple datasets spanning colorectal, breast, liver, and lung cancers, including tissue type, cancer subtype, and biomarker (microsatellite instability) classification tasks. We assessed counterfactual explanations through quantitative analyses, pathologists evaluations, and independent foundation model-based classifiers. We found that MoPaDi was able to generate realistic counterfactual histopathology images, enabling pathologists to identify morphological features associated with the change in model predictions. Unlike conventional reviews of highly attended regions typical in digital pathology, MoPaDi explanations enabled pathologists to directly identify morphological features driving the classifiers predictions from a limited number of top-contributing tiles. Consistent with the literature, our biomarker classifier associated high microsatellite instability with mucinous differentiation, glandular patterns, and lymphocytic infiltration. Furthermore, MoPaDi revealed that changes in classifier predictions were mainly driven by morphological alterations rather than staining differences. Overall, MoPaDi is a practical framework for counterfactual explanations in computational pathology that reveals model-specific drivers of classification and increases trust in deep learning models.

bioinformatics↗

bioRxiv2022

Adversarial attacks and adversarial robustness in computational pathology

Artificial Intelligence (AI) can support diagnostic workflows in oncology by aiding diagnosis and providing biomarkers. AI applications are therefore expected to evolve from academic prototypes to commercial products in the coming years. However, AI applications are vulnerable to adversarial attacks, such as malicious interference with test data aiming to cause misclassifications. Therefore, it is essential for the use of AI-based diagnostic devices to secure them against such attacks before widespread use. Unfortunately, no resistant systems exist in computational pathology so far. To address this problem, we investigate the susceptibility of convolutional neural networks (CNNs) to multiple types of white- and black-box attacks. We demonstrate that both attacks can easily confuse CNNs in clinically relevant pathology tasks and impair classification performance. Classical adversarially robust training and dual batch normalization (DBN) are possible mitigation strategies but require precise knowledge of the type of attack used in the inference. We demonstrate that vision transformers (ViTs) perform equally well compared to CNNs at baseline and are orders of magnitude more robust to different types of white-box and black-box attacks. At a mechanistic level, we show that this is associated with a more robust latent representation of clinically relevant categories in ViTs compared to CNNs. Our results are in line with previous theoretical studies. We show that ViTs are robust learners in computational pathology. This implies that large-scale rollout of AI models in computational pathology should rely on ViTs rather than CNN-based classifiers to provide inherent protection against adversaries.

bioinformatics↗

bioRxiv2021

DeepMed: A unified, modular pipeline for end-to-end deep learning in computational pathology

The interpretation of digitized histopathology images has been transformed thanks to artificial intelligence (AI). End-to-end AI algorithms can infer high-level features directly from raw image data, extending the capabilities of human experts. In particular, AI can predict tumor subtypes, genetic mutations and gene expression directly from hematoxylin and eosin (H&E) stained pathology slides. However, existing end-to-end AI workflows are poorly standardized and not easily adaptable to new tasks. Here, we introduce DeepMed, a Python library for predicting any high-level attribute directly from histopathological whole slide images alone, or from images coupled with additional meta-data (https://github.com/KatherLab/deepmed). Unlike earlier computational pipelines, DeepMed is highly developer-friendly: its structure is modular and separates preprocessing, training, deployment, statistics, and visualization in such a way that any one of these processes can be altered without affecting the others. Also, DeepMed scales easily from local use on laptop computers to multi-GPU clusters in cloud computing services and therefore can be used for teaching, prototyping and for large-scale applications. Finally, DeepMed is user-friendly and allows researchers to easily test multiple hypotheses in a single dataset (via cross-validation) or in multiple datasets (via external validation). Here, we demonstrate and document DeepMeds abilities to predict molecular alterations, histopathological subtypes and molecular features from routine histopathology images, using a large benchmark dataset which we release publicly. In summary, DeepMed is a fully integrated and broadly applicable end-to-end AI pipeline for the biomedical research community.

cancer biology↗

bioRxiv2021

Swarm learning for decentralized artificial intelligence in cancer histopathology

Artificial Intelligence (AI) can extract clinically actionable information from medical image data. In cancer histopathology, AI can be used to predict the presence of molecular alterations directly from routine histopathology slides. However, training robust AI systems requires large datasets whose collection faces practical, ethical and legal obstacles. These obstacles could be overcome with swarm learning (SL) where partners jointly train AI models, while avoiding data transfer and monopolistic data governance. Here, for the first time, we demonstrate the successful use of SL in large, multicentric datasets of gigapixel histopathology images comprising over 5000 patients. We show that AI models trained using Swarm Learning can predict BRAF mutational status and microsatellite instability (MSI) directly from hematoxylin and eosin (H&E)-stained pathology slides of colorectal cancer (CRC). We trained AI models on three patient cohorts from Northern Ireland, Germany and the United States of America and validated the prediction performance in two independent datasets from the United Kingdom using SL-based AI models. Our data show that SL enables us to train AI models which outperform most locally trained models and perform on par with models which are centrally trained on the merged datasets. In addition, we show that SL-based AI models are data efficient and maintain a robust performance even if only subsets of local datasets are used for training. In the future, SL can be used to train distributed AI models for any histopathology image analysis tasks, overcoming the need for data transfer and without requiring institutions to give up control of the final AI model.

bioinformatics↗

bioRxiv2021

Benchmarking artificial intelligence methods for end-to-end computational pathology

Artificial intelligence (AI) can extract subtle visual information from digitized histopathology slides and yield scientific insight on genotype-phenotype interactions as well as clinically actionable recommendations. Classical weakly supervised pipelines use an end-to-end approach with residual neural networks (ResNets), modern convolutional neural networks such as EfficientNet, or non-convolutional architectures such as vision transformers (ViT). In addition, multiple-instance learning (MIL) and clustering-constrained attention MIL (CLAM) are being used for pathology image analysis. However, it is unclear how these different approaches perform relative to each other. Here, we implement and systematically compare all five methods in six clinically relevant end-to-end prediction tasks using data from N=4848 patients with rigorous external validation. We show that histological tumor subtyping of renal cell carcinoma is an easy task which approaches successfully solved with an area under the receiver operating curve (AUROC) of above 0.9 without any significant differences between approaches. In contrast, we report significant performance differences for mutation prediction in colorectal, gastric and bladder cancer. Weakly supervised ResNet-and ViT-based workflows significantly outperformed other methods, in particular MIL and CLAM for mutation prediction. As a reason for this higher performance we identify the ability of ResNet and ViT to assign high prediction scores to highly informative image regions with plausible histopathological image features. We make all source codes publicly available at https://github.com/KatherLab/HIA, allowing easy application of all methods on any end-to-end problem in computational pathology.

bioinformatics↗

bioRxiv2019

Breaking Medical Data Sharing Boundaries by EmployingArtificial Radiographs

Artificial intelligence (AI) has the potential to change medicine fundamentally. Here, expert knowledge provided by AI can enhance diagnosis by comprehensive and user independent integration of multiple image features. Unfortunately, existing algorithms often stay behind expectations, as databases used for training are usually too small, incomplete, and heterogeneous in quality. Additionally, data protection constitutes a serious obstacle to data sharing. We propose to use generative models (GM) to produce high-resolution artificial radiographs, which are free of personal identifying information. Blinded analyses by computer vision and radiology experts proved the high similarity of artificial and real radiographs. The combination of multiple GM improves the performance of computer vision algorithms and the integration of artificial data into patient data repositories can compensate for underrepresented disease entities. Furthermore, the low computational effort of our method complies with existing IT infrastructure in hospitals and thus facilitates its dissemination. We envision that our approach could lead to scalable databases of anonymous medical images enabling standardized radiomic analyses at multiple sites.

bioinformatics↗