@article{Tsanda2026SMRestoration,
author = {A. Tsanda, S. Reiss, K. Scheffler, M. Boberg, and T. Knopp},
title = {Deep learning for restoring MPI system matrices using simulated training data.},
journal = {Physics in Medicine & Biology.},
year = {2026},
volume = {71.},
number = {(9),},
note = {article, openaccess, ml},
doi = {10.1088/1361-6560/ae6016},
abstract = {Objective. Magnetic particle imaging reconstructs tracer distributions using a system matrix (SM) obtained through time-consuming, noise-prone calibration measurements. Methods for addressing imperfections in measured system matrices increasingly rely on deep neural networks, yet curated training data remain scarce. This study evaluates whether physics-based simulated system matrices can be used to train deep learning (DL) models for different SM restoration tasks, i.e. denoising, accelerated calibration, upsampling, and inpainting, that generalize to measured data. Approach. A large dataset of system matrices was generated using an equilibrium magnetization model extended with uniaxial anisotropy. The dataset spans particle, scanner, and calibration parameters for 2D and 3D trajectories, and includes background noise injected from empty-frame measurements. For each restoration task, DL models were compared with classical non-learning baseline methods. Quantitative performance was evaluated on simulated data using peak signal-to-noise ratio (PSNR) and structural similarity index measure (SSIM). For measured data, performance was assessed qualitatively by visual comparison of system matrices and the resulting reconstructions. Main results. The models trained solely on simulated system matrices generalized to measured data across all tasks: for denoising, DnCNN/RDN/SwinIR outperformed discrete cosine transform and soft thresholding baseline by >10 dB PSNR and up to +0.1 SSIM on simulations and led to perceptually better reconstructions of real data; for 2D upsampling, SMRnet exceeded bicubic by ∼ 20 dB PSNR and ∼ 0.08 SSIM at ×2–×4 but these gains did not transfer qualitatively to real measurements. For 3D accelerated calibration, SMRnet matched tricubic in noiseless cases and was more robust under noise, and for 3D inpainting, biharmonic inpainting was superior when noise-free but degraded with noise, while a PConvUNet maintained quality and yielded less blurry reconstructions. Significance. The demonstrated transferability of DL models trained on simulations to real measurements mitigates the data-scarcity problem, which intensifies with model scale. This enables the development of new methods beyond current measurement capabilities and supports pre-training of large models on simulated data.}
}

@article{Backers2026Sepsis,
author = {J. Backes, A. Tsanda, T. Knopp, W. Renz, and E. Schöll},
title = {Combining machine learning and physiological network models for sepsis prediction.},
journal = {Frontiers in Network Physiology.},
year = {2026},
volume = {6.},
note = {article,openaccess,ml},
doi = {10.3389/fnetp.2026.1852577},
url = {https://www.frontiersin.org/journals/network-physiology/articles/10.3389/fnetp.2026.1852577},
abstract = {As the most extreme course of an infectious disease, sepsis poses a serious health threat, with a high mortality rate and frequent long-term consequences for survivors. Despite its enormous burden on global healthcare and ongoing research efforts, early sepsis onset prediction remains challenging due to the complex nature of its pathophysiology. Current approaches face a fundamental trade-off: data-driven machine learning models achieve strong performance but lack interpretability, while biologically inspired models provide mechanistic insights but have limited clinical validation. In this study, we propose the Latent Dynamics Model, a hybrid machine learning approach that integrates a functional model of coupled oscillators representing organ- and immune-cell populations and their interactions. Here, the model parameters encode physiological conditions and allow for an interpretable differentiation between healthy and pathological states. By projecting high-dimensional patient data into the low-dimensional parameter space of the functional model, machine-learned trajectories through this space allow the prediction of critical organ system states and simultaneously offer interpretability beyond plain risk estimates. The proposed method is trained and evaluated on real intensive care patients, achieving competitive AUROC/AUPRC performance on a retrospective MIMIC-IV cohort. Additional qualitative analysis reveals that the learned trajectories exhibit clinically plausible patterns of deterioration, recovery, and stability. We demonstrate that a physiological network model can be embedded within a deep learning architecture without compromising predictive performance while providing an interpretable latent structure for sepsis onset prediction.}
}

@article{Adrian2026NMRRelaxometry,
author = {M. Adrian, K.M. Eckert, M.R. Serial, A. Tsanda, L. Rennpferdt, S. Benders, H.K. Trieu, T. Knopp, I. Smirnova, and A. Penn},
title = {NMR relaxometry probes solvent-polarity-dependent molecular interactions in stimuli-responsive lyogels.},
journal = {Phys. Chem. Chem. Phys..},
year = {2026},
volume = {28.},
pages = {1645-1654},
note = {article},
publisher = {The Royal Society of Chemistry:},
doi = {10.1039/D5CP04032A},
url = {http://dx.doi.org/10.1039/D5CP04032A},
abstract = {Stimuli-responsive gels demonstrate macroscopic changes upon exposure to external stimuli{,} offering potential for the development of adaptive chemical reactors. Early investigations into hydrogels established that crosslinked polymer networks experience reversible volume phase transitions{,} with temperature{,} pH{,} and solvent composition governing swelling and shrinking dynamics. Although hydrogels behavior in aqueous environments has been extensively characterized{,} lyogels that incorporate organic solvents remain comparatively underexplored{,} despite their potential for enhanced chemical compatibility and functional versatility. Here{,} we investigate how solvent polarity and crosslinking density govern the swelling behavior{,} pore formation{,} and molecular-scale dynamics of poly(N-isopropylacrylamide)-based lyogels. Using a combination of swelling measurement{,} scanning electron microscopy{,} and multiscale NMR relaxometry and diffusometry{,} we demonstrate that solvent polarity fundamentally alters lyogel structure and dynamics. Lyogels swollen in a high-polarity solvent exhibits macroporous networks and slower solvent exchange rates{,} whereas a low-polarity solvent induces shrinkage{,} denser microstructures{,} faster solvent exchange rates{,} and stronger surface interactions. These results establish a mechanistic framework linking thermodynamic affinity{,} solvent dynamics{,} and microstructural confinement to macroscopic gel responsiveness. This framework provides guidance for tailoring lyogels in dynamic environments{,} with potential applications in adaptable and tunable chemical reactors.}
}

@article{Merbach2026MRIVelocityTPMS,
author = {T. Merbach, M. Adrian, C. Wigger, S. Iraqi Houssaini, B. Bayer, A. Tsanda, S. Acikgöz, C. Weiland, F. Kexel, D. Herzog, M. Hoffmann, I. Kelbassa, T. Knopp, A. Penn, and M. Schlüter},
title = {Comprehensive study of 3D liquid flow fields in additively manufactured structures for SMART reactors using large-scale vertical magnetic resonance imaging and computational fluid dynamics.},
journal = {Chemical Engineering Journal.},
year = {2026},
volume = {539.},
pages = {176536},
note = {article,openaccess,MRI},
doi = {https://doi.org/10.1016/j.cej.2026.176536},
url = {https://www.sciencedirect.com/science/article/pii/S1385894726039975},
keywords = {Porous media, Magnetic resonance imaging, Computational fluid dynamics, Triply periodic minimal surfaces},
abstract = {Triply Periodic Minimal Surface (TPMS) structures have emerged as a new class of porous materials with variable geometries and favourable transport properties, making them promising for reactor internals in chemical engineering. However, experimental data on internal TPMS flow behaviour are still limited. To address this gap, the flow behaviour in additively manufactured TPMS structures is analysed using three-dimensional Magnetic Resonance Imaging (MRI) velocimetry in a large-bore vertical 3 T MRI system, in cylindrical columns of 38 mm diameter and Reynolds numbers between 50 and 300. Three different TPMS geometries are investigated, and consistency between Computational Fluid Dynamics (CFD) simulations and experimentally measured MRI velocity fields is established through cross-validation. The MRI system provides fully three-dimensional velocity fields with a divergence deviation below 4 %. MRI revealed distinct flow features: the Gyroid TPnS exhibited pronounced channelling, while the Schwarz-Diamond TPSf showed merge-split behaviour, achieving a 46 % increase in lateral mixing compared to the Gyroid TPnS structures. Numerical simulations reproduce the flow features and show agreement with the MRI data. The combined methodology demonstrates the suitability of MRI velocimetry for the experimental validation of CFD simulations and establishes a robust foundation for future studies of heat and mass transfer, as well as reactive flow, in structured reactor systems.}
}

@article{TsandaDoseRobustness2024CT,
author = {A. Tsanda, H. Nickisch, T. Wissel, T. Klinder, T. Knopp, and M. Grass},
title = {Dose robustness of deep learning models for anatomic segmentation of computed tomography images.},
journal = {Journal of Medical Imaging.},
year = {2024},
volume = {11.},
number = {(4),},
pages = {044005},
note = {article},
publisher = {SPIE:},
doi = {10.1117/1.JMI.11.4.044005},
url = {https://doi.org/10.1117/1.JMI.11.4.044005},
keywords = {Low-Dose Computed Tomography, Semantic Segmentation, Denoising, Deep Learning},
abstract = {PurposeThe trend towards lower radiation doses and advances in computed tomography (CT) reconstruction may impair the operation of pretrained segmentation models, giving rise to the problem of estimating the dose robustness of existing segmentation models. Previous studies addressing the issue suffer either from a lack of registered low- and full-dose CT images or from simplified simulations.ApproachWe employed raw data from full-dose acquisitions to simulate low-dose CT scans, avoiding the need to rescan a patient. The accuracy of the simulation is validated using a real CT scan of a phantom. We consider down to 20% reduction of radiation dose, for which we measure deviations of several pretrained segmentation models from the full-dose prediction. In addition, compatibility with existing denoising methods is considered.ResultsThe results reveal the surprising robustness of the TotalSegmentator approach, showing minimal differences at the pixel level even without denoising. Less robust models show good compatibility with the denoising methods, which help to improve robustness in almost all cases. With denoising based on a convolutional neural network (CNN), the median Dice between low- and full-dose data does not fall below 0.9 (12 for the Hausdorff distance) for all but one model. We observe volatile results for labels with effective radii less than 19 mm and improved results for contrasted CT acquisitions.ConclusionThe proposed approach facilitates clinically relevant analysis of dose robustness for human organ segmentation models. The results outline the robustness properties of a diverse set of models. Further studies are needed to identify the robustness of approaches for lesion segmentation and to rank the factors contributing to dose robustness.}
}

@COMMENT{Bibtex file generated on 2026-7-3 with typo3 si_bibtex plugin. Data from https://www.tuhh.de/ibi/people/artyom-tsanda }