44197 research outputs found
Sort by
Lab-scale Machine Learning: Tales of the good, the bad and the average
No full textMachine Learning and Artificial Intelligence are presented as the fix-all for current day problems. Also in research it is experiencing a golden age. However, before a Machine Learning model can be created, an enormous quantity of training data needs to be generated. This stands in stark contrast to general academic and industrial lab-scale data sets resulting from research projects. The latter give rise to small or even extremely small data sets (< 50 samples). This makes many of us wonder: "Is it possible to train an ML model with 30 samples instead of 30.000.000?" Using some simple regression models, I'll show that these can be successful in creating a suitable model in the (very) small data regime. Real life use-cases considering adhesive coatings, solvable inks, and spray-coating are discussed. I'll present a strategy to always obtain the best model and highlight caveats and ways to deal with them. [1] "A machine learning approach for the design of hyperbranched polymeric dispersing agents based on aliphatic polyesters for radiation curable inks"
Pathophysiological mechanisms underlying early brain injury and delayed cerebral ischemia in the aftermath of aneurysmal subarachnoid hemorrhage: a comprehensive analysis
No full textEarly brain injury (EBI) and delayed cerebral ischemia (DCI) are pivotal contributors to morbidity and mortality following aneurysmal subarachnoid hemorrhage (aSAH). Despite advances that have reduced mortality and incidence, aSAH remains a significant public health concern due to its early onset, leading to prolonged periods of diminished quality of life for affected individuals. EBI mechanisms, including endothelial dysfunction, blood-brain barrier disruption, cerebral edema, neuro-inflammation, cortical spreading depolarizations, and oxidative damage, trigger cell death and apoptosis, setting the stage for DCI development in later clinical phases. DCI arises not only from large-vessel vasospasm, but also from other complex pathophysiological processes, including thrombo-inflammation, neuro-inflammation, microcirculatory dysfunction, and glycocalyx disruption. Recognizing and understanding these mechanisms is essential, as early interventions could potentially reduce long-term disability in this population. This comprehensive review offers an in-depth analysis of these pathophysiological mechanisms. As our understanding of these processes continues to evolve, further research is crucial to improving outcomes and reducing the long-term impact of aSAH.The authors wish to thank Dr. Wencke Renette. The figures were made using Biorender (https://www.biorender.com
An elevated glycosylated haemoglobin level is associated with a higher risk of penile prosthesis infection: systematic review and meta-analysis
No full textcitation ID: qdaf077.231 Objectives: The impact of preoperative glycometabolic profile of diabetic patients on the risk of penile prosthesis (PP) infection remains uncertain. Current data are conflicting but the studies to date have been limited by retrospective design and by including heterogenous cohorts of patients (both diabetic and non-diabetic). Hence, the available evidence on the impact of diabetes mellitus (DM) and glycemic control on infection rate after PP implantation was systematically reviewed. Methods: A comprehensive Medline, Embase, and Cochrane search was performed including the keywords penile prosthe-sis and diabetes mellitus. English-language articles between January 1st, 1969, up to May 31st, 2024 were included. Primary outcome was the PP infection rate in patients with DM. Secondary endpoints included the contribution of gly-cometabolic control on PP infection rate. Random-effect model was uniformly applied. Robust meta-analytical techniques were employed to control for heterogeneity and unde-tected bias including sensitivity analyses and regression linear adjusted models where appropriate. The protocol of this study (CRD42024557982) was published on PROSPERO. Results: Out of 182 retrieved articles, 9 were included in the study (3 prospective and 6 retrospective) accounting for 5493 subjects with a mean age of 59.7 years, and a mean follow-up of 29.4 months. Overall, a PP infection rate of up to 7% was observed. The PP infection rate increased according to baseline HbA1c levels and was confirmed in multivariate analysis adjusting for age and trial duration (B = 7.8 ± 0.2%; p < 0.0001). PP infection rate was almost 3-times higher when trials with a mean HbA1c greater than 8% were compared to the rest of the sample (9.1 (7.5;11.0) vs. 3.8(3.2;13.5)%; Q = 43.18; p < 0.0001). Conclusions: The present study suggests a significantly increased risk of PP infection for patients with DM and pre-operative HbA1c greater than 8%. A multidisciplinary approach to optimise preoperative glycometabolic control in patients with DM may be the key element for successful PP implant outcomes. However, better quality studies are needed to better guide clinical practice and preoperative optimisation. Conflicts of Interest: nil
Clinical predictors for restrictive allograft syndrome: A nested case-control study
No full textRisk factors for restrictive allograft syndrome (RAS), a severe phenotype of chronic lung allograft dysfunction (CLAD) after lung transplantation, are currently not well known. In this retrospective nested case-control-study, we analyzed 69 patients with RAS and 69 matched non-CLAD controls to identify clinical risk factors for RAS. Patients with RAS demonstrated overall higher blood eosinophils (P = .02), increased bronchoalveolar eosinophils (P < .001) and lymphocytes (P = .03), and higher incidence of infections, particularly Pseudomonas species infection (P = .003), invasive fungal disease (P < .001, mainly due to Aspergillus species), SARS-CoV-2 (P < .001), and cytomegalovirus infection (P = .04), compared with non-CLAD controls. Antihuman leukocyte antigen (anti-HLA) antibodies, especially persistent donor-specific antibodies (P < 0.001), specifically targeting HLA-DQ and HLA-DR loci, and antibody-mediated rejection (P < .001), were strongly associated with later RAS. Histopathologic lung injury patterns on transbronchial biopsy (P < .001), and persistent chest computed tomography opacities in absence of pulmonary dysfunction (P < .001) were identified as early indicators of later RAS. Proactive detection and management of these risk factors could help mitigate future decline in allograft function and reduce progression to clinical RAS. Future studies should explore early treatment strategies targeting these modifiable factors to preserve allograft function and improve long-term outcomes for lung transplant recipients.Clinical Fellowship of the European Respiratory Society; Researc
BioTIME 2.0: Expanding and Improving a Database of Biodiversity Time Series
No full textMotivationHere, we make available a second version of the BioTIME database, which compiles records of abundance estimates for species in sample events of ecological assemblages through time. The updated version expands version 1.0 of the database by doubling the number of studies and includes substantial additional curation to the taxonomic accuracy of the records, as well as the metadata. Moreover, we now provide an R package (BioTIMEr) to facilitate use of the database.Main Types of Variables IncludedThe database is composed of one main data table containing the abundance records and 11 metadata tables. The data are organised in a hierarchy of scales where 11,989,233 records are nested in 1,603,067 sample events, from 553,253 sampling locations, which are nested in 708 studies. A study is defined as a sampling methodology applied to an assemblage for a minimum of 2 years.Spatial Location and GrainSampling locations in BioTIME are distributed across the planet, including marine, terrestrial and freshwater realms. Spatial grain size and extent vary across studies depending on sampling methodology. We recommend gridding of sampling locations into areas of consistent size.Time Period and GrainThe earliest time series in BioTIME start in 1874, and the most recent records are from 2023. Temporal grain and duration vary across studies. We recommend doing sample-level rarefaction to ensure consistent sampling effort through time before calculating any diversity metric.Major Taxa and Level of MeasurementThe database includes any eukaryotic taxa, with a combined total of 56,400 taxa.Software Formatcsv and. SQL.H2020 European Research Counci
Biomorphoelasticity alone: limitations in modeling post-burn contraction and hypertrophy without finite strains
No full textWe present a continuum hypothesis-based two-dimensional biomorphoelastic model describing post-burn scar hypertrophy and contraction. The model is based on morphoelasticity for permanent deformations and combined with a chemical-biological model that incorporates cellular densities, collagen density, and the concentration of chemoattractants. We perform a sensitivity analysis for the independent parameters of the model and focus on the effects on the features of the post-burn dermal thickness given a low myofibroblast apoptosis rate. We conclude that the most sensitive parameters are the equilibrium collagen concentration, the signaling molecule secretion rate and the cell force constant, and link these results to stability constraints. Next, we observe a relationship between the simulated contraction and hypertrophy and show the effects for significant variations in the myofibroblast apoptosis rate (high/low). Our ultimate goal is to optimize post-burn treatments, by developing models that predict with a high degree of certainty. We consider the presented model and sensitivity analysis to be a step toward their construction.The authors are grateful for the fnancial support
from the Dutch Burns Foundation under projects 17.105 and 22.10
Accelerating Open Data Integration of Real-World Health Data Silos
No full textMedical information collected during routine healthcare, or real-world data (RWD), can facilitate medical research projects and allow the generation of real-world evidence. While large amounts of RWD are generated on a daily basis, they remain locked in autonomous and heterogeneous health data silos, inaccessible to medical data analysts. To make RWD accessible, we need to apply open data integration techniques, a topic well-researched in data
management. Open data integration refers to the pay-as-you-go approach to data integration to cope with the volume, variety, velocity and veracity of data in the increasingly common data lake settings. Data integration tasks are the individual steps needed to integrate data. Due to the inherent challenges of the healthcare domain, many automations do not get applied by medical informaticians to RWD. In this thesis, we aim to investigate this gap and
show how the open data integration of real-world health data silos can be accelerated.
We first investigate the gap between data management and medical informatics in a literature review, quantifying which health data integration tasks lack automations developed in data management. Apart from identifying duplicate patients in multiple datasets, which is well researched in medical informatics, we conclude that all data integration tasks could
benefit from data integration approaches developed in data management.
Next, we approach data integration from two parallel perspectives. From a data management perspective, we survey the discovery of approximate functional dependencies (AFDs), a multi-column data profiling approach that detects a strong relationship between sets of attributes of a relation. Based on this comparison, we recommend AFD measures to efficiently discover AFDs in RWD. From a medical informatics perspective, we give a retrospective of a project where we developed a health data integration platform together with three partner hospitals. In particular, we discuss our development approach and lessons learned regarding the complex landscape of real-world health data silos and its associated stakeholders.
The learnings from these two perspectives guide our following contribution. We tackle schema matching, a data integration task that lacks automation in the healthcare domain. We approach name-based schema matching, where we aim to identify semantic correspondences between two schemas based on names and descriptions of schema elements. We use a large language model (LLM) and focus on comparing the impact that the amount
of information put into a single prompt has on the matching quality. We do not use any instances, i.e. actual data values, to increase the applicability of our approach on sensitive RWD. After our initial development based on public data, we validate our approach on private RWD schemas obtained from four Belgian hospitals. We show that our LLM-based schema matching approach returns high-quality correspondences and give practical considerations for future users. By example of a name-based schema matcher developed with the healthcare domain in mind, we illustrate how open data integration techniques developed in data management accelerate unlocking real-world health data silos
Spatiotemporal Gait and Fatigability in Multiple Sclerosis: Overground vs Treadmill Walking
No full textAchtergrond: Multiple sclerose (MS) is een neurodegeneratieve aandoening die vaak leidt tot gangproblemen, verergerd door vermoeibaarheid. Er is weinig bekend over hoe langdurig wandelen onder vaste snelheid het looppatroon beïnvloedt.
Methode: Negen personen met MS (pwMS) en achttien gezonde controles (HC) (29–64 jaar) voerden een 6-minuten wandeltest (6MWT) uit onder drie condities: overground, en loopband bij 50% en 80% van de maximale snelheid (T25FW). Gaitkenmerken (cadans, double support, single limb support (SLS), stapduur (SD), paslengte (SL) en zwaaifase) werden gemeten met sensoren (overground) of het GRAIL-systeem (loopband). Een lineair mixed model analyseerde groeps-, conditie- en interactie-effecten; procentuele veranderingen van minuut één tot zes werden berekend.
Resultaten: SLS en SD veranderden meer bij overground wandelen dan bij beide loopbandcondities. SL veranderde minder overground dan bij de trage loopbandconditie. Gemiddelde T25FW-snelheden waren 2,13 m/s (pwMS) en 2,10 m/s (HC); 6MWT-snelheden 1,63 m/s (pwMS) en 1,72 m/s (HC), zonder groepsverschillen.
Conclusie: De wandelconditie beïnvloedde de gangaanpassing: overground wandelen leidde tot grotere veranderingen in SLS en SD, kleinere veranderingen in SL en tot een grotere wandelafstand dan op de loopband
