A5041778310- Radiomics and Machine Learning in Medical Imaging
- MRI in cancer diagnosis
- Advanced MRI Techniques and Applications
- Epigenetics and DNA Methylation
- Medical Imaging Techniques and Applications
- RNA modifications and cancer
- Biomedical Text Mining and Ontologies
- Machine Learning in Healthcare
- Data Quality and Management
- Scientific Computing and Data Management
- Meta-analysis and systematic reviews
- Digital Radiography and Breast Imaging
- Research Data Management Practices
- Electronic Health Records Systems
- Advanced Neuroimaging Techniques and Applications
- AI in cancer detection
- Genetic Syndromes and Imprinting
- COVID-19 and healthcare impacts
- Soil and Unsaturated Flow
- Clinical Reasoning and Diagnostic Skills
- Ethics in Clinical Research
- Lanthanide and Transition Metal Complexes
- Artificial Intelligence in Healthcare and Education
- Lung Cancer Diagnosis and Treatment
- Long-Term Effects of COVID-19
Universitätsklinikum Erlangen
2019-2025
Friedrich-Alexander-Universität Erlangen-Nürnberg
2019-2025
Klinikum rechts der Isar
2024
Technical University of Munich
2024
Research Institute of Radiology
2022
The COVID-19 pandemic has caused strains on health systems worldwide disrupting routine hospital services for all non-COVID patients. Within this retrospective study, we analyzed inpatient admissions across 18 German university hospitals during the 2020 lockdown period compared to 2018. Patients admitted between January 1st and May 31st, corresponding periods in 2018 2019 were included study. Data derived from electronic records collected using data integration center infrastructure...
To take full advantage of decision support, machine learning, and patient-level prediction models, it is important that models are not only created, but also deployed in a clinical setting. The KETOS platform demonstrated this work implements tool for researchers allowing them to perform statistical analyses deploy resulting secure environment.The proposed system uses Docker virtualization provide with reproducible data analysis development environments, accessible via Jupyter Notebook,...
Abstract Appropriate reference intervals are essential when using laboratory test results to guide medical decisions. Conventional approaches for the establishment of rely on large samples from healthy and homogenous populations. However, this approach is associated with substantial financial logistic challenges, subject ethical restrictions in children, limited older individuals due high prevalence chronic morbidities medication. We implemented an indirect method interval estimation, which...
To evaluate whether neural networks can distinguish between seropositive RA, seronegative and PsA based on inflammatory patterns from hand MRIs to test how psoriasis patients with subclinical inflammation fit into such patterns.ResNet were utilized compare RA vs PsA, respect MRI data. Results T1 coronal, T2 coronal axial fat-suppressed contrast-enhanced (CE), sequences used. The performance of trained was analysed by the area under receiver operating characteristics curve (AUROC) without...
COVID-19 has challenged the healthcare systems worldwide. To quickly identify successful diagnostic and therapeutic approaches large data sharing are inevitable. Though organizational clinical abundant, many of them available only in isolated silos largely inaccessible to external researchers. overcome tackle this challenge university medicine network (comprising all 36 German hospitals) been founded April 2020 coordinate action plans, strategies collaborative research activities. 13...
Introduction: Data quality (DQ) is an important prerequisite for secondary use of electronic health record (EHR) data in clinical research, particularly with regards to progressing towards a learning system, one the MIRACUM consortium's goals. Following successful integration i2b2 research repository MIRACUM, we present standardized and generic DQ framework.
Many research initiatives aim at using data from electronic health records (EHRs) in observational studies. Participating sites of the German Medical Informatics Initiative (MII) established integration centers to integrate EHR within repositories support local and federated analyses. To address concerns regarding possible quality (DQ) issues hospital routine compared with specifically collected for scientific purposes, we have previously presented a assessment (DQA) tool providing...
Abstract The objective of this IRB approved retrospective study was to apply deep learning identify magnetic resonance imaging (MRI) artifacts on maximum intensity projections (MIP) the breast, which were derived from diffusion weighted (DWI) protocols. dataset consisted 1309 clinically indicated breast MRI examinations 1158 individuals (median age [IQR]: 50 years [16.75 years]) acquired between March 2017 and June 2020, in a DWI sequence with high b-value equal 1500 s/mm 2 acquired. From...
Abstract Background Breast magnetic resonance imaging (MRI) protocols often include T2-weighted fat-saturated (T2w-FS) sequences, which support tissue characterization but significantly increase scan time. This study aims to evaluate whether a 2D-U-Net neural network can generate virtual T2w-FS (VirtuT2w) images from routine multiparametric breast MRI images. Methods IRB-approved, retrospective included 914 examinations January 2017 June 2020. The dataset was divided into training ( n =...
ABSTRACT Background Virtual contrast-enhanced (vCE) imaging techniques are an emerging topic of research in breast MRI. Purpose To investigate how different combinations T1-weighted (T1w), T2-weighted (T2w), and diffusion-weighted (DWI) impact the performance vCE Materials Methods The IRB-approved, retrospective study included 1064 multiparametric MRI scans (age:52±12 years) obtained from 2017-2020 (single site, two 3T MRI). Eleven independent neural networks were trained to derive images...
Abstract Background The current gap between the availability of routine imaging data and its provisioning for medical research hinders utilization radiological information secondary purposes. To address this, German Medical Informatics Initiative (MII) has established frameworks harmonizing integrating clinical across institutions, including integration into repositories, which can be expanded to data. Objectives This project aims this by developing a large-scale processing pipeline extract,...
The use of electronic health records for clinical research offers access to large-scale real-world data, but it requires the accurate transformation data across repositories. In this study, we evaluate quality and completeness in three repositories (DWH, FHIR, TriNetX) at Erlangen University Hospital. Key elements (diagnosis, procedure, laboratory codes) were analyzed, alongside a specific question. Our results show good overall consistency, discrepancies arise due differences code systems,...
Data quality assessments (DQA) are necessary to ensure valid research results. Despite the growing availability of tools relevance for DQA in R language, a systematic comparison their functionalities is missing. Therefore, we review packages related data (DQ) and assess scope against DQ framework observational health studies. Based on search, screened more than 140 Comprehensive Archive Network. From these, selected which target at least three four dimensions (integrity, completeness,...
Abstract Background The identification of patient cohorts for recruiting patients into clinical trials requires an evaluation study-specific inclusion and exclusion criteria. These criteria are specified depending on corresponding facts. Some these facts may not be present in the source systems need to calculated either advance or at cohort query runtime (so-called feasibility query). Objectives We use Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM) as repository...
Integrative bioinformatics is an emerging field in the big data era, offering a steadily increasing number of algorithms and analysis tools. However, for researchers experimental life sciences it often difficult to follow properly apply bioinformatical methods order unravel complexity systemic effects omics data. Here, we present integrative pipeline decipher crucial biological insights from global transcriptome profiling validate innovative therapeutics. It available as web application...
To assess the change in inpatient radiotherapy related to COVID-19 lockdown measures during first wave of pandemic 2020.
To automatically detect MRI artifacts on dynamic contrast-enhanced (DCE) maximum intensity projections (MIPs) of the breast using deep learning.Women who underwent clinically indicated between October 2015 and December 2019 were included in this IRB-approved retrospective study. We employed two convolutional neural network architectures (ResNet DenseNet) to presence DCE MIPs left right breasts. Networks trained images acquired up including year 2018 a 5-fold cross-validation (CV). Ensemble...
Abstract Background Breast magnetic resonance imaging (MRI) protocols often include T2-weighted fat-saturated (T2w-FS) sequences, which are vital for tissue characterization but significantly increase scan time. Purpose This study aims to evaluate whether a 2D-U-Net neural network can generate virtual T2w-FS images from routine multiparametric breast MRI sequences. Materials and Methods IRB approved, retrospective included n=914 examinations performed between January 2017 June 2020. The...
With the growing impact of observational research studies, there is also a focus on data quality (DQ). As opposed to experimental study designs, studies are performed using mostly collected in non-research context (secondary use). Depending number elements be analyzed, DQ reports stored within networks can grow very large. They might cumbersome read and important information could overseen quickly. To address this issue, assessment (DQA) tool with graphical user interface (GUI) was developed...
The identification of biomarker signatures is important for cancer diagnosis and prognosis. However, the detection clinical reliable influenced by limited data availability, which may restrict statistical power. Moreover, methods integration large sample cohorts signature are limited. We present a step-by-step computational protocol functional gene expression analysis diagnostic prognostic combining meta-analysis with machine learning survival analysis. novelty toolbox lies in its all-in-one...
Abstract The purpose of this feasibility study is to investigate if latent diffusion models (LDMs) are capable generate contrast enhanced (CE) MRI-derived subtraction maximum intensity projections (MIPs) the breast, which conditioned by lesions. We trained an LDM with n = 2832 CE-MIPs breast MRI examinations 1966 patients (median age: 50 years) acquired between years 2015 and 2020. was subsequently 756 segmented lesions from 407 examinations, indicating their location BI-RADS scores. By...
Introduction Prenatal androgen exposure has important organizing effects on brain development and therefore future behavior. Previous research shown, that the ratio between index finger (2D) ring (4D) (2D:4D) could function as a marker of prenatal effects, with relatively shorter 2D indicating higher exposure. 2D:4D is associated status-seeking competitive behavior but also altruism. Therefore, should be related to academic success. Methods We examined both hands, well difference variables...