Sex and gender differences significantly influence health outcomes, yet they remain insufficiently represented in clinical datasets and digital health infrastructures. This paper synthesizes existing literature and current interoperability frameworks, including national initiatives such as the German Medical Informatics Initiative (MII) and emerging European developments like the European Health Data Space (EHDS). Three major challenges are identified: semantic limitations in representing sex and gender concepts, structural gaps in interoperability frameworks and data models, and operational deficiencies in data collection and reuse. These limitations propagate into downstream applications, particularly artificial intelligence systems, leading to biased models and reduced generalizability. The paper discusses how extensions of interoperability standards such as HL7 FHIR, enrichment of core datasets, and improved metadata and governance frameworks can support more inclusive biomedical technologies. Addressing these challenges is essential for ensuring equitable, robust, and clinically meaningful data-driven innovation.
Registration of longitudinal Computed Tomography (CT) scans is critical for monitoring treatment responses, yet its precision is often compromised by inconsistencies in patient positioning between sessions. Historically, the medical imaging community has relied purely on post-hoc ("postscan") image registration to correct these inconsistencies [1]. Whether utilizing rigid or deformable models, post-hoc registration is a reactive tool applied to data that is inherently flawed at acquisition. This work proposes a paradigm shift: addressing variability at its source through pre-scan patient pose alignment. By capturing the patient’s current pose with an integrated RGB-D camera prior to the follow-up CT scan, discrepancies relative to a baseline CT scan can be registered and visualized. This paper outlines an articulated registration pipeline designed to overcome the domain gap between occluded RGB-D surface data and volumetric CT data. The torso serves as a rigid anchor to align the modalities, after which individual limbs undergo local point-to-plane Iterative Closest Point refinement. The spatial corrections are then projected onto the original 2D RGB image, providing intuitive visual guidance for clinical staff. Validated on an anthropomorphic phantom, our pipeline demonstrates highly stable torso anchoring (average centroid error over 10 iterations: 2.69 cm) and consistent limb alignment (mean surface error: 1.61-1.84 cm). This proof-of-concept establishes a foundation for proactive patient alignment in longitudinal imaging.
Self-driving labs have become a widespread paradigm for automating research and development processes across disciplines such as biology, chemistry, and pharmaceutical research. Such systems combine robotic hardware and software to iteratively optimize experimental outcomes in a closed loop. Noise inherent in sensors and robotic actuation not only affects the speed at which the optimal result is found, but also which optimization strategies perform best. We present a simulation-based methodology to analyze selfdriving lab loops and to select the optimal acquisition function and hyperparameters for Bayesian optimization by incorporating known uncertainty in realistic simulated system behavior. Our results indicate that optimizer selection informed by noise characteristics can significantly reduce the number of required iterations.
This work presents an open-source dynamic model of a pressure swing adsorption (PSA)-based medical oxygen concentrator, enabling a systematic investigation of its process behavior. The model describes the coupled interaction of the compressor, adsorption columns, valves, and auxiliary volumes, and is based on a one-dimensional isothermal formulation including mass transfer, adsorption kinetics, and pressure drop. The governing equations are discretized using a finite volume approach and numerically solved in MATLAB. Simulation results are compared with experimental data from a commercial oxygen concentrator. The model captures the main characteristics of the PSA process. Deviations from experimental data are mainly observed during transient phases, while the model accurately reproduces the cyclic steady-state behavior. The developed model provides a basis for the analysis of operating parameters and for the assessment of the energy efficiency of PSA-based oxygen concentrators.
Physiological closed-loop control systems (PCLCS) are safety critical and require rigorous evaluation. Yet, in silico testing is often limited by study-specific setups with restricted reproducibility, traceability, and reusability. This work presents a modular, scenario-driven framework for traceable and reproducible testing of PCLCS. The framework decouples models, controllers, and testing scenarios, enabling structured and modular composition of experiments from reusable components. Parameter sets can be derived from literature ranges or fitted from data to provide patient-specific models (digital twins). Experiments are encoded in a structured, version-controllable format for advanced traceability. We demonstrate the framework using a closedloop mechanical ventilation controller for neonates and a neonatal patient model. Two experiments are conducted: setpoint tracking with and without spontaneous breathing, and replication of a previous in vivo experiment using a subjectspecific digital twin. Results illustrate evaluation across physiological variations with fully reproducible execution. This approach provides a foundation for rigorous and transparent testing of PCLCS.
