Accessing reliable clinical knowledge quickly is an everyday challenge for clinicians. Large Language Models (LLMs) can assist healthcare professionals by providing this knowledge, but their responses often deviate from expert consensus or are not up to date necessitating reliable validation and possible correction. To address this, we introduce MedBlock-Bot, an interactive Streamlit-based system integrating a blockchain-enabled Retrieval-Augmented Generation (RAG) framework for expert-driven assessment and immutable feedback storage within a permissioned consortium network. Unlike traditional feedback mechanisms that may be altered or lost, MedBlock-Bot employs smart contracts to securely store and verify any feedback, ensuring transparency and auditability. We evaluated the system using three opensource LLMs-BioMistral, HippoMistral, and LLaMa 3.1-on clinical guideline interpretation for neonates with hypoplastic left heart syndrome. Human experts assessed model responses based on accuracy and relevance, revealing variations in adherence to the guideline knowledge. Additionally, deploying the blockchain component in a local permissioned environment (Ganache) ensured efficient transaction processing and tamperproof feedback retrieval without gas cost concerns. Our results demonstrate the integration of blockchain for LLM feedback review enhancing trust, accountability, and structured knowledge retention. Clinicians can access past expert assessments for validation, while developers can leverage this feedback for potential model refinement. Taking the long-term impact into account this approach targets towards a reliable and dynamic representation of clinical knowledge and consensus. Open-Source Code: https://github.com/yaseen28/MedBlock-Bot

An automatic classification of the malignancy of lung nodules in computed tomography (CT) scans can support early detection of lung cancer, which is crucial for the treatment success. The novel photon-counting CT (PCCT) technology enables high image quality with a low radiation dose and provides additional spectral information. This research focuses on whether PCCT scans offer a benefit in the automatic classification of lung nodules. Establishing a dataset of PCCT images poses several challenges, such as the extraction of annotations or the data imbalance.

We propose a coordinate regression driven weakly supervised segmentation approach comprising a novel Maha-lanobis distance based loss function and a unimodality facilitating regularization term. We furthermore explore its use for the segmentation of difficult structures such as the locus coeruleus (LC) that are characterized by low inter-rater agreements of delineations due to substantial rater bias. The LC has been attracting increasing interest due to its role in the pathogenesis of neurodegenerative diseases such as Parkinson's Disease (PD) and Alzheimer's Disease (AD), but its in vivo analysis using Magnetic Resonance Imaging (MRI) is challenging. Thorough evaluation yielded promising results. Although the mask similarity is lower compared to a fully supervised approach, the intra-class correlation (ICC) of the extracted features suggests a good agreement.

The development and validation of Clinical Decision Support Models (CDSM) based on Bayesian networks (BN) is commonly done in a collaborative work between medical researchers providing the domain expertise and computer scientists developing the decision support model. Although modern tools provide facilities for data-driven model generation, domain experts are required to validate the accuracy of the learned model and to provide expert knowledge for fine-tuning it while computer scientists are needed to integrate this knowledge in the learned model (hybrid modeling approach). This generally time-expensive procedure hampers CDSM generation and updating. To address this problem, we developed a novel interactive visual approach allowing medical researchers with less knowledge in CDSM to develop and validate BNs based on domain specific data mainly independently and thus, diminishing the need for an additional computer scientist. In this context, we abstracted and simplified the common workflow in BN development as well as adjusted the workflow to medical experts' needs. We demonstrate our visual approach with data of endometrial cancer patients and evaluated it with six medical researchers who are domain experts in the gynecological field.

The identification of interesting patterns and relationships is essential to exploratory data analysis. This becomes increasingly difficult in high dimensional datasets. While dimensionality reduction techniques can be utilized to reduce the analysis space, these may unintentionally bury key dimensions within a larger grouping and obfuscate meaningful patterns. With this work we introduce DimLift, a novel visual analysis method for creating and interacting with dimensional bundles. Generated through an iterative dimensionality reduction or user-driven approach, dimensional bundles are expressive groups of dimensions that contribute similarly to the variance of a dataset. Interactive exploration and reconstruction methods via a layered parallel coordinates plot allow users to lift interesting and subtle relationships to the surface, even in complex scenarios of missing and mixed data types. We exemplify the power of this technique in an expert case study on clinical cohort data alongside two additional case examples from nutrition and ecology.

The Dual Analysis framework is a powerful enabling technology for the exploration of high dimensional quantitative data by treating data dimensions as first-class objects that can be explored in tandem with data values. In this article, we extend the Dual Analysis framework through the joint treatment of quantitative (numerical) and qualitative (categorical) dimensions. Computing common measures for all dimensions allows us to visualize both quantitative and qualitative dimensions in the same view. This enables a natural joint treatment of mixed data during interactive visual exploration and analysis. Several measures of variation for nominal qualitative data can also be applied to ordinal qualitative and quantitative data. For example, instead of measuring variability from a mean or median, other measures assess inter-data variation or average variation from a mode. In this work, we demonstrate how these measures can be integrated into the Dual Analysis framework to explore and generate hypotheses about high-dimensional mixed data. A medical case study using clinical routine data of patients suffering from Cerebral Small Vessel Disease (CSVD), conducted with a senior neurologist and a medical student, shows that a joint Dual Analysis approach for quantitative and qualitative data can rapidly lead to new insights based on which new hypotheses may be generated.

Abstract Simulations and measurements of blood and airflow inside the human circulatory and respiratory system play an increasingly important role in personalized medicine for prevention, diagnosis and treatment of diseases. This survey focuses on three main application areas. (1) Computational fluid dynamics (CFD) simulations of blood flow in cerebral aneurysms assist in predicting the outcome of this pathologic process and of therapeutic interventions. (2) CFD simulations of nasal airflow allow for investigating the effects of obstructions and deformities and provide therapy decision support. (3) 4D phase‐contrast (4D PC) magnetic resonance imaging of aortic haemodynamics supports the diagnosis of various vascular and valve pathologies as well as their treatment. An investigation of the complex and often dynamic simulation and measurement data requires the coupling of sophisticated visualization, interaction and data analysis techniques. In this paper, we survey the large body of work that has been conducted within this realm. We extend previous surveys by incorporating nasal airflow, addressing the joint investigation of blood flow and vessel wall properties and providing a more fine‐granular taxonomy of the existing techniques. From the survey, we extract major research trends and identify open problems and future challenges. The survey is intended for researchers interested in medical flow but also more general, in the combined visualization of physiology and anatomy, the extraction of features from flow field data and feature‐based visualization, the visual comparison of different simulation results and the interactive visual analysis of the flow field and derived characteristics.

We present a Cerebral Aneurysm Vortex Classification (CAVOCLA) that allows to classify blood flow in cerebral aneurysms. Medical studies assume a strong relation between the progression and rupture of aneurysms and flow patterns. To understand how flow patterns impact the vessel morphology, they are manually classified according to predefined classes. However, manual classifications are time-consuming and exhibit a high inter-observer variability. In contrast, our approach is more objective and faster than manual methods. The classification of integral lines, representing steady or unsteady blood flow, is based on a mapping of the aneurysm surface to a hemisphere by calculating polar-based coordinates. The lines are clustered and for each cluster a representative is calculated. Then, the polar-based coordinates are transformed to the representative as basis for the classification. Classes are based on the flow complexity. The classification results are presented by a detail-on-demand approach using a visual transition from the representative over an enclosing surface to the associated lines. Based on seven representative datasets, we conduct an informal interview with five domain experts to evaluate the system. They confirmed that CAVOCLA allows for a robust classification of intra-aneurysmal flow patterns. The detail-on-demand visualization enables an efficient exploration and interpretation of flow patterns.

Epidemiological population-based studies collect hundreds of socio-demographic, lifestyle-related, and health related variables for thousands of individuals in order to better characterize health and disease in a defined population. To understand the relations between the variables in the study data, we employ Bayesian Networks, as they not only represent associations between variables but also assign probabilities to these associations. The probabilistic associations allow us to draw inference for unknown events in question, based on the provided evidence. In our work, we induce a Bayesian Network from the data of the population-based epidemiological Study of Health in Pomerania (SHIP), to identify variables related to the outcome "fatty liver". We report on Bayesian Network structure learning, identification of variables associated with the outcome, and the strong associations identified among the variables.

Abstract It is generally accepted practice that each cancer patient case should be discussed in a clinical expert meeting, the so‐called tumor board. A central role in finding the best therapy options for patients with solid tumors plays the Tumor, lymph Node, and Metastasis staging (TNM staging). Correctness of TNM staging has a significant impact on the therapy choice and hence on the patient's post‐therapeutic quality of life or even survival. If inconsistencies in the TNM staging occur, possible explanations and solutions must be found based on the complex patient records, which takes the costly time of (multiple) physicians. We propose a more efficient visual analysis component, which supports a physician in verifying the given TNM staging before forwarding it to the tumor board. Our component comprises a Bayesian network model of the TNM staging process. Using information from the patient records and Bayesian inference, the models computes a patient‐specific TNM staging, which is then explored and compared to the given staging by means of a graph‐based visualization. Our component is implemented in a research prototype that supports an understanding of the model computations, allows for a fast identification of important influencing factors, and facilitates a quick detection of differences between two TNM stagings. We evaluated our component with five physicians, each studying 20 cases of laryngeal cancer.

Computational fluid dynamic (CFD) simulations of blood flow provide new insights into the hemodynamics of vascular pathologies such as cerebral aneurysms. Understanding the relations between hemodynamics and aneurysm initiation, progression, and risk of rupture is crucial in diagnosis and treatment. Recent studies link the existence of vortices in the blood flow pattern to aneurysm rupture and report observations of embedded vortices -a larger vortex encloses a smaller one flowing in the opposite direction -whose implications are unclear. We present a clustering-based approach for the visual analysis of vortical flow in simulated cerebral aneurysm hemodynamics. We show how embedded vortices develop at saddle-node bifurcations on vortex core lines and convey the participating flow at full manifestation of the vortex by a fast and smart grouping of streamlines and the visualization of group representatives. The grouping result may be refined based on spectral clustering generating a more detailed visualization of the flow pattern, especially further off the core lines. We aim at supporting CFD engineers researching the biological implications of embedded vortices.

Epidemiological population studies impose information about a set of subjects (a cohort) to characterize disease-specific risk factors. Cohort studies comprise heterogenous variables describing the medical condition as well as demographic and lifestyle factors and, more recently, medical image data. We propose an Interactive Visual Analysis (IVA) approach that enables epidemiologists to rapidly investigate the entire data pool for hypothesis validation and generation. We incorporate image data, which involves shape-based object detection and the derivation of attributes describing the object shape. The concurrent investigation of image-based and non-image data is realized in a web-based multiple coordinated view system, comprising standard views from information visualization and epidemiological data representations such as pivot tables. The views are equipped with brushing facilities and augmented by 3D shape renderings of the segmented objects, e.g., each bar in a histogram is overlaid with a mean shape of the associated subgroup of the cohort. We integrate an overview visualization, clustering of variables and object shape for data-driven subgroup definition and statistical key figures for measuring the association between variables. We demonstrate the IVA approach by validating and generating hypotheses related to lower back pain as part of a qualitative evaluation.

Understanding the hemodynamics of blood flow in vascular pathologies such as intracranial aneurysms is essential for both their diagnosis and treatment. Computational fluid dynamics (CFD) simulations of blood flow based on patient-individual data are performed to better understand aneurysm initiation and progression and more recently, for predicting treatment success. In virtual stenting, a flow-diverting mesh tube (stent) is modeled inside the reconstructed vasculature and integrated in the simulation. We focus on steady-state simulation and the resulting complex multiparameter data. The blood flow pattern captured therein is assumed to be related to the success of stenting. It is often visualized by a dense and cluttered set of streamlines.We present a fully automatic approach for reducing visual clutter and exposing characteristic flow structures by clustering streamlines and computing cluster representatives. While individual clustering techniques have been applied before to streamlines in 3D flow fields, we contribute a general quantitative and a domain-specific qualitative evaluation of three state-of-the-art techniques. We show that clustering based on streamline geometry as well as on domain-specific streamline attributes contributes to comparing and evaluating different virtual stenting strategies. With our work, we aim at supporting CFD engineers and interventional neuroradiologists.

In Toponomics, the function protein pattern in cells or tissue (the toponome) is imaged and analyzed for applications in toxicology, new drug development and patient-drug-interaction. The most advanced imaging technique is robot-driven multi-parameter fluorescence microscopy. This technique is capable of co-mapping hundreds of proteins and their distribution and assembly in protein clusters across a cell or tissue sample by running cycles of fluorescence tagging with monoclonal antibodies or other affinity reagents, imaging, and bleaching in situ. The imaging results in complex multi-parameter data composed of one slice or a 3D volume per affinity reagent. Biologists are particularly interested in the localization of co-occurring proteins, the frequency of co-occurrence and the distribution of co-occurring proteins across the cell. We present an interactive visual analysis approach for the evaluation of multi-parameter fluorescence microscopy data in toponomics. Multiple, linked views facilitate the definition of features by brushing multiple dimensions. The feature specification result is linked to all views establishing a focus+context visualization in 3D. In a new attribute view, we integrate techniques from graph visualization. Each node in the graph represents an affinity reagent while each edge represents two co-occurring affinity reagent bindings. The graph visualization is enhanced by glyphs which encode specific properties of the binding. The graph view is equipped with brushing facilities. By brushing in the spatial and attribute domain, the biologist achieves a better understanding of the function protein patterns of a cell. Furthermore, an interactive table view is integrated which summarizes unique fluorescence patterns. We discuss our approach with respect to a cell probe containing lymphocytes and a prostate tissue section.

Abstract Cardiovascular imaging with current multislice spiral computed tomography (MSCT) technology enables a non‐invasive evaluation of the coronary arteries. Contrast‐enhanced MSCT angiography with high spatial resolution allows for a segmentation of the coronary artery tree. We present an automatically adapted transfer function (TF) specification to highlight pathologic changes of the vessel wall based on the segmentation result of the coronary artery tree. The TFs are combined with common visualization techniques, such as multiplanar reformation and direct volume rendering for the evaluation of coronary arteries in MSCT image data. The presented TF‐based mapping of CT values in Hounsfield Units (HU) to color and opacity leads to a different color coding for different plaque types. To account for varying HU values of the vessel lumen caused by the contrast medium, the TFs are adapted to each dataset by local histogram analysis. We describe an informal evaluation with three board‐certified radiologists which indicates that the represented visualizations guide the user's attention to pathologic changes of the vessel wall as well as provide an overview about spatial variations.

Dynamic contrast-enhanced image data (perfusion data) are used to characterize regional tissue perfusion. Perfusion data consist of a sequence of images, acquired after a contrast agent bolus is applied. Perfusion data are used for diagnostic purposes in oncology, ischemic stroke assessment or myocardial ischemia. The diagnostic evaluation of perfusion data is challenging, since the data is complex and exhibits various artifacts, e.g., motion artifacts. We provide an overview on existing methods to analyze, and visualize CT and MR perfusion data. The integrated visualization of several 2D parameter maps, the 3D visualization of parameter volumes and exploration techniques are discussed. An essential aspect in the diagnosis of perfusion data is the correlation between perfusion data and derived time-intensity curves as well as with other image data, in particular with high resolution morphologic image data. We discuss visualization support with respect to the three major application areas: ischemic stroke diagnosis, breast tumor diagnosis and the diagnosis of coronary heart disease.

Abstract In this paper we present a new approach to the interactive visual analysis of time‐dependent scientific data – both from measurements as well as from computational simulation – by visualizing a scalar function over time for each of tenthousands or even millions of sample points. In order to cope with overdrawing and cluttering, we introduce a new four‐level method of focus+context visualization. Based on a setting of coordinated, multiple views (with linking and brushing), we integrate three different kinds of focus and also the context in every single view. Per data item we use three values (from the unit interval each) to represent to which degree the data item is part of the respective focus level. We present a color compositing scheme which is capable of expressing all three values in a meaningful way, taking semantics and their relations amongst each other (in the context of our multiple linked view setup) into account. Furthermore, we present additional image‐based postprocessing methods to enhance the visualization of large sets of function graphs, including a texture‐based technique based on line integral convolution (LIC). We also propose advanced brushing techniques which are specific to the time‐dependent nature of the data (in order to brush patterns over time more efficiently). We demonstrate the usefulness of the new approach in the context of medical perfusion data.

Current magnetic resonance imaging (MRI) technology allows the determination of patient-individual coronary tree structure, detection of infarctions, and assessment of myocardial perfusion. Joint inspection of these three aspects yields valuable information for therapy planning, e.g., through classification of myocardium into healthy tissue, regions showing a reversible hypoperfusion, and infarction with additional information on the corresponding supplying artery. Standard imaging protocols normally provide image data with different orientations, resolutions and coverages for each of the three aspects, which makes a direct comparison of analysis results difficult. The purpose of this work is to develop methods for the alignment and combined analysis of these images. The proposed approach is applied to 21 datasets of healthy and diseased patients from the clinical routine. The evaluation shows that, despite limitations due to typical MRI artifacts, combined inspection is feasible and can yield clinically useful information.

Expressive and efficient visualizations of complex vascular structures are essential for medical applications, such as diagnosis and therapy planning. A variety of techniques has been developed which provide smooth high-quality visualizations of vascular structures based on rather simple model assumptions. For diagnostic applications, these model assumptions and the resulting deviations from the actual vessel surface are not acceptable. We present a model-free approach which employs the binary result of a prior vessel segmentation as input. Instead of directly converting the segmentation result into a surface, we compute a point cloud which is adaptively refined at thin structures, where aliasing effects are particularly obvious and artifacts may occur. The point cloud is transformed into a surface representation by means of MPU Implicits, which provide a smooth piecewise quadratic approximation. Our method has been applied to a variety of datasets including pathologic cases. The generated visualizations are considerably more accurate than model-based approaches. Compared to other model-free approaches, our method produces smoother results.

Perfusion data are dynamic medical image data which characterize the regional blood flow in human tissue. These data bear a great potential in medical diagnosis, since diseases can be better distinguished and detected at an earlier stage compared to static image data. The wide-spread use of perfusion data is hampered by the lack of efficient evaluation methods. For each voxel, a time-intensity curve characterizes the enhancement of a contrast agent. Parameters derived from these curves characterize the perfusion and have to be integrated for diagnosis. The diagnostic evaluation of this multi-field data is challenging and time-consuming due to its complexity. For the visual analysis of such datasets, feature-based approaches allow to reduce the amount of data and direct the user to suspicious areas. We present an interactive visual analysis approach for the evaluation of perfusion data. For this purpose, we integrate statistical methods and interactive feature specification. Correlation analysis and Principal Component Analysis (PCA) are applied for dimensionreduction and to achieve a better understanding of the inter-parameter relations. Multiple, linked views facilitate the definition of features by brushing multiple dimensions. The specification result is linked to all views establishing a focus+context style of visualization in 3D. We discuss our approach with respect to clinical datasets from the three major application areas: ischemic stroke diagnosis, breast tumor diagnosis, as well as the diagnosis of the coronary heart disease (CHD). It turns out that the significance of perfusion parameters strongly depends on the individual patient, scanning parameters, and data pre-processing.

We present static and dynamic techniques to visualize perfusion data and to relate perfusion data to morphologic image data. In particular, we describe the integrated analysis of MRI myocardial perfusion data with CT coronary angiographies depicting the morphology. We refined the Bull’s-Eye Plot, a wide-spread and accepted analysis tool in cardiac diagnosis, to show aggregated information of perfusion data at rest and under stress. The correlation between regions of the myocard with reduced perfusion and 3d renditions of the coronary vessels can be explored within a synchronized visualization of both. With our research, we attempt to improve the diagnosis of early stage coronary artery disease.

We present a method for visualizing vasculature based on clinical computed tomography or magnetic resonance data. The vessel skeleton as well as the diameter information per voxel serve as input. Our method adheres to these data, while producing smooth transitions at branchings and closed, rounded ends by means of convolution surfaces. We examine the filter design with respect to irritating bulges, unwanted blending and the correct visualization of the vessel diameter. The method has been applied to a large variety of anatomic trees. We discuss the validation of the method by means of a comparison to other visualization methods. Surface distance measures are carried out to perform a quantitative validation. Furthermore, we present the evaluation of the method which has been accomplished on the basis of a survey by 11 radiologists and surgeons.