The accumulation of erythrocytes quantified and visualized by Glycophorin C in carotid atherosclerotic plaque reflects intraplaque hemorrhage and pre-procedural neurological symptoms.

The accumulation of erythrocyte membranes within an atherosclerotic plaque may contribute to the deposition of free cholesterol and thereby the enlargement of the necrotic core. Erythrocyte membranes can be visualized and quantified in the plaque by immunostaining for the erythrocyte marker glycophorin C. Hence, we theorized that the accumulation of erythrocytes quantified by glycophorin C could function as a marker for plaque vulnerability, possibly reflecting intraplaque hemorrhage (IPH), and offering predictive value for pre-procedural neurological symptoms. We employed the CellProfiler-integrated slideToolKit workflow to visualize and quantify glycophorin C, defined as the total plaque area that is positive for glycophorin C, in single slides of culprit lesions obtained from the Athero-Express Biobank of 1819 consecutive asymptomatic and symptomatic patients who underwent carotid endarterectomy. Our assessment included the evaluation of various parameters such as lipid core, calcifications, collagen content, SMC content, and macrophage burden. These parameters were evaluated using a semi-quantitative scoring method, and the resulting data was dichotomized as predefined criteria into categories of no/minor or moderate/heavy staining. In addition, the presence or absence of IPH was also scored. The prevalence of IPH and pre-procedural neurological symptoms were 62.4% and 87.1%, respectively. The amount of glycophorin staining was significantly higher in samples from men compared to samples of women (median 7.15 (IQR:3.37, 13.41) versus median 4.06 (IQR:1.98, 8.32), p < 0.001). Glycophorin C was associated with IPH adjusted for clinical confounders (OR 1.90; 95% CI 1.63, 2.21; p = < 0.001). Glycophorin C was significantly associated with ipsilateral pre-procedural neurological symptoms (OR:1.27, 95%CI:1.06-1.41, p = 0.005). Sex-stratified analysis, showed that this was also the case for men (OR 1.37; 95%CI 1.12, 1.69; p = 0.003), but not for women (OR 1.15; 95%CI 0.77, 1.73; p = 0.27). Glycophorin C was associated with classical features of a vulnerable plaque, such as a larger lipid core, a higher macrophage burden, less calcifications, a lower collagen and SMC content. There were marked sex differences, in men, glycophorin C was associated with calcifications and collagen while these associations were not found in women. To conclude, the accumulation of erythrocytes in atherosclerotic plaque quantified and visualized by glycophorin C was independently associated with the presence of IPH, preprocedural symptoms in men, and with a more vulnerable plaque composition in both men and women. These results strengthen the notion that the accumulation of erythrocytes quantified by glycophorin C can be used as a marker for plaque vulnerability.

Sexual dimorphism in peripheral blood cell characteristics linked to recanalization success of endovascular thrombectomy in acute ischemic stroke.

Endovascular thrombectomy (EVT) success to treat acute ischemic stroke varies with factors like stroke etiology and clot composition, which can differ between sexes. We studied if sex-specific blood cell characteristics (BCCs) are related to recanalization success. We analyzed electronic health records of 333 EVT patients from a single intervention center, and extracted 71 BCCs from the Sapphire flow cytometry analyzer. Through Sparse Partial Least Squares Discriminant Analysis, incorporating cross-validation and stability selection, we identified BCCs associated with successful recanalization (TICI 3) in both sexes. Stroke etiology was considered, while controlling for cardiovascular risk factors. Of the patients, successful recanalization was achieved in 51% of women and 49% of men. 21 of the 71 BCCs showed significant differences between sexes  (pFDR-corrected < 0.05). The female-focused recanalization model had lower error rates than both combined [t(192.4) = 5.9, p < 0.001] and male-only models [t(182.6) = - 15.6, p < 0.001]. In women, successful recanalization and cardioembolism were associated with a higher number of reticulocytes, while unsuccessful recanalization and large artery atherosclerosis (LAA) as cause of stroke were associated with a higher mean corpuscular hemoglobin concentration. In men, unsuccessful recanalization and LAA as cause of stroke were associated with a higher coefficient of variance of lymphocyte complexity of the intracellular structure. Sex-specific BCCs related to recanalization success varied and were linked to stroke etiology. This enhanced understanding may facilitate personalized treatment for acute ischemic stroke.

A cluster of blood-based protein biomarkers associated with decreased cerebral blood flow relates to future cardiovascular events in patients with cardiovascular disease.

Biological processes underlying decreased cerebral blood flow (CBF) in patients with cardiovascular disease (CVD) are largely unknown. We hypothesized that identification of protein clusters associated with lower CBF in patients with CVD may explain underlying processes. In 428 participants (74% cardiovascular diseases; 26% reference participants) from the Heart-Brain Connection Study, we assessed the relationship between 92 plasma proteins from the Olink® cardiovascular III panel and normal-appearing grey matter CBF, using affinity propagation and hierarchical clustering algorithms, and generated a Biomarker Compound Score (BCS). The BCS was related to cardiovascular risk and observed cardiovascular events within 2-year follow-up using Spearman correlation and logistic regression. Thirteen proteins were associated with CBF (ρSpearman range: -0.10 to -0.19, pFDR-corrected <0.05), and formed one cluster. The cluster primarily reflected extracellular matrix organization processes. The BCS was higher in patients with CVD compared to reference participants (pFDR-corrected <0.05) and was associated with cardiovascular risk (ρSpearman 0.42, p < 0.001) and cardiovascular events (OR 2.05, p < 0.01). In conclusion, we identified a cluster of plasma proteins related to CBF, reflecting extracellular matrix organization processes, that is also related to future cardiovascular events in patients with CVD, representing potential targets to preserve CBF and mitigate cardiovascular risk in patients with CVD.

Design, validation and implementation of an automated e-alert for acute kidney injury: 6-month pilot study shows increased awareness.

BACKGROUND: Acute kidney injury (AKI) is defined as a sudden episode of kidney failure but is known to be under-recognized by healthcare professionals. The Kidney Disease Improving Global Outcome (KDIGO) guidelines have formulated criteria to facilitate AKI diagnosis by comparing changes in plasma creatinine measurements (PCr). To improve AKI awareness, we implemented these criteria as an electronic alert (e-alert), in our electronic health record (EHR) system. METHODS: For every new PCr measurement measured in the University Medical Center Utrecht that triggered the e-alert, we provided the physician with actionable insights in the form of a memo, to improve or stabilize kidney function. Since e-alerts qualify for software as a medical device (SaMD), we designed, implemented and validated the e-alert according to the European Union In Vitro Diagnostic Regulation (IVDR). RESULTS: We evaluated the impact of the e-alert using pilot data six months before and after implementation. 2,053 e-alerts of 866 patients were triggered in the before implementation, and 1,970 e-alerts of 853 patients were triggered after implementation. We found improvements in AKI awareness as measured by (1) 2 days PCr follow up (56.6-65.8%, p-value: 0.003), and (2) stop of nephrotoxic medication within 7 days of the e-alert (59.2-63.2%, p-value: 0.002). CONCLUSION: Here, we describe the design and implementation of the e-alert in line with the IVDR, leveraging a multi-disciplinary team consisting of physicians, clinical chemists, data managers and data scientists, and share our firsts results that indicate an improved awareness among treating physicians.

In-vitro and in-silico evidence for oxidative stress as drivers for RDW.

Red blood cell distribution width (RDW) is a biomarker associated with a variety of clinical outcomes. While anemia and subclinical inflammation have been posed as underlying pathophysiology, it is unclear what mechanisms underlie these assocations. Hence, we aimed to unravel the mechanisms in silico using a large clinical dataset and validate our findings in vitro. We retrieved complete blood counts (CBC) from 1,403,663 measurements from the Utrecht Patient Oriented Database, to model RDW using gradient boosting regression. We performed (sex-stratified) analyses in patients with anemia, patients younger/older than 50 and validation across platforms and care settings. We then validated our hypothesis regarding oxidative stress using an in vitro approach. Only percentage microcytic (pMIC) and macrocytic (pMAC) erythrocytes and mean corpuscular volume were most important in modelling RDW (RMSE = 0.40, R2 = 0.96). Subgroup analyses and validation confirmed our findings. In vitro induction of oxidative stress underscored our results, namely increased RDW and decreased erythrocyte volume, yet no vesiculation was observed. We found that erythrocyte size, especially pMIC, is most informative in predicting RDW, but no role for anemia or inflammation. Oxidative stress affecting the size of the erythrocytes may play a role in the association between RDW and clinical outcomes.

Characteristics of peripheral blood cells are independently related to major adverse cardiovascular events after carotid endarterectomy.

Background and aims: Patients who underwent carotid endarterectomy (CEA) still have a residual risk of 13% of developing a major adverse cardiovascular event (MACE) within 3 years. Inflammatory processes leading up to MACE are not fully understood. Therefore, we examined blood cell characteristics (BCCs), possibly reflecting inflammatory processes, in relation to MACE to identify BCCs that may contribute to an increased risk. Methods: We analyzed 75 pretreatment BCCs from the Sapphire analyzer, and clinical data from the Athero-Express biobank in relation to MACE after CEA using Random Survival Forests, and a Generalized Additive Survival Model. To understand biological mechanisms, we related the identified variables to intraplaque hemorrhage (IPH). Results: Of 783 patients, 97 (12%) developed MACE within 3 years after CEA. Red blood cell distribution width (RDW) (HR 1.23 [1.02, 1.68], p = 0.022), CV of lymphocyte size (LACV) (HR 0.78 [0.63, 0.99], p = 0.043), neutrophil complexity of the intracellular structure (NIMN) (HR 0.80 [0.64, 0.98], p = 0.033), mean neutrophil size (NAMN) (HR 0.67 [0.55, 0.83], p < 0.001), mean corpuscular volume (MCV) (HR 1.35 [1.09, 1.66], p = 0.005), eGFR (HR 0.65 [0.52, 0.80], p < 0.001); and HDL-cholesterol (HR 0.62 [0.45, 0.85], p = 0.003) were related to MACE. NAMN was related to IPH (OR 0.83 [0.71-0.98], p = 0.02). Conclusions: This is the first study to present a higher RDW and MCV and lower LACV, NIMN and NAMN as biomarkers reflecting inflammatory processes that may contribute to an increased risk of MACE after CEA.

Routine Lupus Anticoagulant Sensitive aPTT Testing Can Prevent Unnecessary LA Testing.

Even though routine screening of the general hospital population is discouraged, medical laboratories may use a "lupus sensitive" activated partial thromboplastin time test (aPTT) with phospholipid concentrations that are susceptible to inhibition by lupus anticoagulant (LA), to screen for the presence of LA. If deemed necessary, follow-up testing according to ISTH guidelines may be performed. However, LA testing is a laborious and time-consuming effort that is often not readily available due to a lack of automation and/or temporary unavailability of experienced staff. In contrast, the aPTT is a fully automated test that is available 24/7 in almost all medical laboratories and is easily interpreted with the use of reference ranges. In addition to clinical signs, the result of an LA sensitive aPTT may thus be used to lower the suspicion of the presence of LA and reduce costly follow-up testing. In this study, we show that a normal LA sensitive aPTT result may be safely used to refrain from LA testing in the absence of strong clinical suspicion.

Added diagnostic value of routinely measured hematology variables in diagnosing immune checkpoint inhibitor mediated toxicity in the emergency department.

BACKGROUND: Immune checkpoint inhibitors (ICI) show remarkable results in cancer treatment, but at the cost of immune-related adverse events (irAE). irAE can be difficult to differentiate from infections or tumor progression, thereby challenging treatment, especially in the emergency department (ED) where time and clinical information are limited. As infections are traceable in blood, we were interested in the added diagnostic value of routinely measured hematological blood cell characteristics in addition to standard diagnostic practice in the ED to aid irAE assessment. METHODS: Hematological variables routinely measured with our hematological analyzer (Abbott CELL-DYN Sapphire) were retrieved from Utrecht Patient Oriented Database (UPOD) for all patients treated with ICI who visited the ED between 2013 and 2020. To assess the added diagnostic value, we developed and compared two models; a base logistic regression model trained on the preliminary diagnosis at the ED, sex, and gender, and an extended model trained with lasso that also assessed the hematology variables. RESULTS: A total of 413 ED visits were used in this analysis. The extended model showed an improvement in performance (area under the receiver operator characteristic curve) over the base model, 0.79 (95% CI 0.75-0.84), and 0.67 (95% CI 0.60-0.73), respectively. Two standard blood count variables (eosinophil granulocyte count and red blood cell count) and two advanced variables (coefficient of variance of neutrophil depolarization and red blood cell distribution width) were associated with irAE. CONCLUSION: Hematological variables are a valuable and inexpensive aid for irAE diagnosis in the ED. Further exploration of the predictive hematological variables could yield new insights into the pathophysiology underlying irAE and in distinguishing irAE from other inflammatory conditions.

Asking informed consent may lead to significant participation bias and suboptimal cardiovascular risk management in learning healthcare systems.

BACKGROUND: The Utrecht Cardiovascular Cohort - CardioVascular Risk Management (UCC-CVRM) was set up as a learning healthcare system (LHS), aiming at guideline based cardiovascular risk factor measurement in all patients in routine clinical care. However, not all patients provided informed consent, which may lead to participation bias. We aimed to study participation bias in a LHS by assessing differences in and completeness of cardiovascular risk management (CVRM) indicators in electronic health records (EHRs) of consenting, non-consenting, and non-responding patients, using the UCC-CVRM as an example. METHODS: All patients visiting the University Medical Center Utrecht for first time evaluation of a(n) (a)symptomatic vascular disease or condition were invited to participate. Routine care data was collected in the EHR and an informed consent was asked. Differences in patient characteristics were compared between consent groups. We performed multivariable logistic regression to identify determinants of non-consent. We used multinomial regression for an exploratory analysis for the determinants of non-response. Presence of CVRM indicators were compared between consent groups. A waiver (19/641) was obtained from our ethics committee. RESULTS: Out of 5730 patients invited, 2378 were consenting, 1907 non-consenting, and 1445 non-responding. Non-consent was related to young and old age, lower education level, lower BMI, physical activity and haemoglobin levels, higher heartrate, cardiovascular disease history and absence of proteinuria. Non-response increased with young and old age, higher education level, physical activity, HbA1c and decreased with lower levels of haemoglobin, BMI, and systolic blood pressure. Presence of CVRM indicators was 5-30% lower in non-consenting patients and even lower in non-responding patients, compared to consenting patients. Non-consent and non-response varied across specialisms. CONCLUSIONS: A traditional informed consent procedure in a LHS may lead to participation bias and potentially to suboptimal CVRM, which is detrimental for feedback on findings in a LHS. This underlines the importance of reassessing the informed consent procedure in a LHS.

[Eight misconceptions about AI in healthcare]. / Acht misvattingen over AI in de zorg.

It is of paramount importance that healthcare professionals can participate in the academic and societal debate surrounding medical AI. To realise this critical-constructive guidance of AI, it is necessary to be able to distinguish between different types of AI, different applications of AI and to paint the different shades of grey in the current black-and-white debate. This article describes and nuances eight misconceptions that currently dominate the public debate surrounding AI in healthcare. By asking ourselves as healthcare professionals 'what specifically defines our line of work?' we must define what aspects of our occupation we want to have AI either carry out or support, and in what way.

Lipoproteins act as vehicles for lipid antigen delivery and activation of invariant natural killer T cells.

Invariant natural killer T (iNKT) cells act at the interface between lipid metabolism and immunity because of their restriction to lipid antigens presented on CD1d by antigen-presenting cells (APCs). How foreign lipid antigens are delivered to APCs remains elusive. Since lipoproteins routinely bind glycosylceramides structurally similar to lipid antigens, we hypothesized that circulating lipoproteins form complexes with foreign lipid antigens. In this study, we used 2-color fluorescence correlation spectroscopy to show, for the first time to our knowledge, stable complex formation of lipid antigens α-galactosylceramide (αGalCer), isoglobotrihexosylceramide, and OCH, a sphingosine-truncated analog of αGalCer, with VLDL and/or LDL in vitro and in vivo. We demonstrate LDL receptor-mediated (LDLR-mediated) uptake of lipoprotein-αGalCer complexes by APCs, leading to potent complex-mediated activation of iNKT cells in vitro and in vivo. Finally, LDLR-mutant PBMCs of patients with familial hypercholesterolemia showed impaired activation and proliferation of iNKT cells upon stimulation, underscoring the relevance of lipoproteins as a lipid antigen delivery system in humans. Taken together, circulating lipoproteins form complexes with lipid antigens to facilitate their transport and uptake by APCs, leading to enhanced iNKT cell activation. This study thereby reveals a potentially novel mechanism of lipid antigen delivery to APCs and provides further insight into the immunological capacities of circulating lipoproteins.

Optimizing cardiovascular risk assessment and registration in a developing cardiovascular learning health care system: Women benefit most.

Since 2015 we organized a uniform, structured collection of a fixed set of cardiovascular risk factors according the (inter)national guidelines on cardiovascular risk management. We evaluated the current state of a developing cardiovascular towards learning healthcare system-the Utrecht Cardiovascular Cohort Cardiovascular Risk Management (UCC-CVRM)-and its potential effect on guideline adherence in cardiovascular risk management. We conducted a before-after study comparing data from patients included in UCC-CVRM (2015-2018) and patients treated in our center before UCC-CVRM (2013-2015) who would have been eligible for UCC-CVRM using the Utrecht Patient Oriented Database (UPOD). Proportions of cardiovascular risk factor measurement before and after UCC-CVRM initiation were compared, as were proportions of patients that required (change of) blood pressure, lipid, or blood glucose lowering treatment. We estimated the likelihood to miss patients with hypertension, dyslipidemia, and elevated HbA1c before UCC-CVRM for the whole cohort and stratified for sex. In the present study, patients included up to October 2018 (n = 1904) were matched with 7195 UPOD patients with similar age, sex, department of referral and diagnose description. Completeness of risk factor measurement increased, ranging from 0% -77% before to 82%-94% after UCC-CVRM initiation. Before UCC-CVRM, we found more unmeasured risk factors in women compared to men. This sex-gap resolved in UCC-CVRM. The likelihood to miss hypertension, dyslipidemia, and elevated HbA1c was reduced by 67%, 75% and 90%, respectively, after UCC-CVRM initiation. A finding more pronounced in women compared to men. In conclusion, a systematic registration of the cardiovascular risk profile substantially improves guideline adherent assessment and decreases the risk of missing patients with elevated levels with an indication for treatment. The sex-gap disappeared after UCC-CVRM initiation. Thus, an LHS approach contributes to a more inclusive insight into quality of care and prevention of cardiovascular disease (progression).

Bi-allelic variants in NAE1 cause intellectual disability, ischiopubic hypoplasia, stress-mediated lymphopenia and neurodegeneration.

Neddylation has been implicated in various cellular pathways and in the pathophysiology of numerous diseases. We identified four individuals with bi-allelic variants in NAE1, which encodes the neddylation E1 enzyme. Pathogenicity was supported by decreased NAE1 abundance and overlapping clinical and cellular phenotypes. To delineate how cellular consequences of NAE1 deficiency would lead to the clinical phenotype, we focused primarily on the rarest phenotypic features, based on the assumption that these would best reflect the pathophysiology at stake. Two of the rarest features, neuronal loss and lymphopenia worsening during infections, suggest that NAE1 is required during cellular stress caused by infections to protect against cell death. In support, we found that stressing the proteasome system with MG132-requiring upregulation of neddylation to restore proteasomal function and proteasomal stress-led to increased cell death in fibroblasts of individuals with NAE1 genetic variants. Additionally, we found decreased lymphocyte counts after CD3/CD28 stimulation and decreased NF-κB translocation in individuals with NAE1 variants. The rarest phenotypic feature-delayed closure of the ischiopubic rami-correlated with significant downregulation of RUN2X and SOX9 expression in transcriptomic data of fibroblasts. Both genes are involved in the pathophysiology of ischiopubic hypoplasia. Thus, we show that NAE1 plays a major role in (skeletal) development and cellular homeostasis during stress. Our approach suggests that a focus on rare phenotypic features is able to provide significant pathophysiological insights in diseases caused by mutations in genes with pleiotropic effects.

Negation detection in Dutch clinical texts: an evaluation of rule-based and machine learning methods.

When developing models for clinical information retrieval and decision support systems, the discrete outcomes required for training are often missing. These labels need to be extracted from free text in electronic health records. For this extraction process one of the most important contextual properties in clinical text is negation, which indicates the absence of findings. We aimed to improve large scale extraction of labels by comparing three methods for negation detection in Dutch clinical notes. We used the Erasmus Medical Center Dutch Clinical Corpus to compare a rule-based method based on ContextD, a biLSTM model using MedCAT and (finetuned) RoBERTa-based models. We found that both the biLSTM and RoBERTa models consistently outperform the rule-based model in terms of F1 score, precision and recall. In addition, we systematically categorized the classification errors for each model, which can be used to further improve model performance in particular applications. Combining the three models naively was not beneficial in terms of performance. We conclude that the biLSTM and RoBERTa-based models in particular are highly accurate accurate in detecting clinical negations, but that ultimately all three approaches can be viable depending on the use case at hand.

From computer to bedside, involving neonatologists in artificial intelligence models for neonatal medicine.

In recent years, data have become the main driver of medical innovation. With increased availability and decreased price of storage and computing power, the potential for improvement in care is enormous. Many data-driven explorations have started. However, the actual implementation of artificial intelligence in healthcare remains scarce. We describe essential elements during a computer-to-bedside process in a data science project that support the crucial role of the neonatologist. IMPACT: There is a great potential for data science in neonatal medicine. Multidisciplinary teams form the foundation of a data science project. Domain experts will need to play a pivotal role. We need an open learning environment.

Shorter laboratory turnaround time is associated with shorter emergency department length of stay: a retrospective cohort study.

BACKGROUND: A longer emergency department length of stay (EDLOS) is associated with poor outcomes. Shortening EDLOS is difficult, due to its multifactorial nature. A potential way to improve EDLOS is through shorter turnaround times for diagnostic testing. This study aimed to investigate whether a shorter laboratory turnaround time (TAT) and time to testing (TTT) were associated with a shorter EDLOS. METHODS: A retrospective cohort study was performed, including all visits to the emergency department (ED) of an academic teaching hospital from 2017 to 2020 during which a standardized panel of laboratory tests had been ordered. TTT was calculated as the time from arrival in the ED to the ordering of laboratory testing. TAT was calculated as the time from test ordering to the reporting of the results, and was divided into a clinical and a laboratory stage. The outcome was EDLOS in minutes. The effect of TTT and TAT on EDLOS was estimated through a linear regression model. RESULTS: In total, 23,718 ED visits were included in the analysis. Median EDLOS was 199.0 minutes (interquartile range [IQR] 146.0-268.0). Median TTT was 7.0 minutes (IQR 2.0-12.0) and median TAT was 51.1 minutes (IQR 41.1-65.0). Both TTT and TAT were positively associated with EDLOS. The laboratory stage comprised a median of 69% (IQR 59-78%) of total TAT. CONCLUSION: Longer TTT and TAT are independently associated with longer EDLOS. As the laboratory stage predominantly determines TAT, it provides a promising target for interventions to reduce EDLOS and ED crowding.

A machine learning approach using endpoint adjudication committee labels for the identification of sepsis predictors at the emergency department.

Accurate sepsis diagnosis is paramount for treatment decisions, especially at the emergency department (ED). To improve diagnosis, clinical decision support (CDS) tools are being developed with machine learning (ML) algorithms, using a wide range of variable groups. ML models can find patterns in Electronic Health Record (EHR) data that are unseen by the human eye. A prerequisite for a good model is the use of high-quality labels. Sepsis gold-standard labels are hard to define due to a lack of reliable diagnostic tools for sepsis at the ED. Therefore, standard clinical tools, such as clinical prediction scores (e.g. modified early warning score and quick sequential organ failure assessment), and claims-based methods (e.g. ICD-10) are used to generate suboptimal labels. As a consequence, models trained with these "silver" labels result in ill-trained models. In this study, we trained ML models for sepsis diagnosis at the ED with labels of 375 ED visits assigned by an endpoint adjudication committee (EAC) that consisted of 18 independent experts. Our objective was to evaluate which routinely measured variables show diagnostic value for sepsis. We performed univariate testing and trained multiple ML models with 95 routinely measured variables of three variable groups; demographic and vital, laboratory and advanced haematological variables. Apart from known diagnostic variables, we identified added diagnostic value for less conventional variables such as eosinophil count and platelet distribution width. In this explorative study, we show that the use of an EAC together with ML can identify new targets for future sepsis diagnosis research.

A Wolf in Sheep's Clothing: Reuse of Routinely Obtained Laboratory Data in Research.

Electronic health records (EHRs) contain valuable data for reuse in science, quality evaluations, and clinical decision support. Because routinely obtained laboratory data are abundantly present, often numeric, generated by certified laboratories, and stored in a structured way, one may assume that they are immediately fit for (re)use in research. However, behind each test result lies an extensive context of choices and considerations, made by both humans and machines, that introduces hidden patterns in the data. If they are unaware, researchers reusing routine laboratory data may eventually draw incorrect conclusions. In this paper, after discussing health care system characteristics on both the macro and micro level, we introduce the reader to hidden aspects of generating structured routine laboratory data in 4 steps (ordering, preanalysis, analysis, and postanalysis) and explain how each of these steps may interfere with the reuse of routine laboratory data. As researchers reusing these data, we underline the importance of domain knowledge of the health care professional, laboratory specialist, data manager, and patient to turn routine laboratory data into meaningful data sets to help obtain relevant insights that create value for clinical care.

A Network of Serum Proteins Predict the Need for Systemic Immunomodulatory Therapy at Diagnosis in Noninfectious Uveitis.

Purpose: Early identification of patients with noninfectious uveitis requiring steroid-sparing immunomodulatory therapy (IMT) is currently lacking in objective molecular biomarkers. We evaluated the proteomic signature of patients at the onset of disease and associated proteomic clusters with the need for IMT during the course of the disease. Design: Multicenter cohort study. Participants: Two hundred thirty treatment-free patients with active noninfectious uveitis. Methods: We used aptamer-based proteomics (n = 1305 proteins) and a bioinformatic pipeline as a molecular stratification tool to define the serum protein network of a Dutch discovery cohort (n = 78) of patients and healthy control participants and independently validated our results in another Dutch cohort (n = 111) and a United States cohort (n = 67). Multivariate Cox analysis was used to assess the relationship between the protein network and IMT use. Main Outcome Measures: Serum protein levels and use of IMT. Results: Network-based analyses revealed a tightly coexpressed serum cluster (n = 85 proteins) whose concentration was consistently low in healthy control participants (n = 26), but varied among patients with noninfectious uveitis (n = 52). Patients with high levels of the serum cluster at disease onset showed a significantly increased need for IMT during follow-up, independent of anatomic location of uveitis (hazard ratio, 3.42; 95% confidence interval, 1.22-9.5; P = 0.019). The enrichment of neutrophil-associated proteins in the protein cluster led to our finding that the neutrophil count could serve as a clinical proxy for this proteomic signature (correlation: r = 0.57, P = 0.006). In an independent Dutch cohort (n = 111), we confirmed that patients with relatively high neutrophil count at diagnosis (> 5.2 × 109/L) had a significantly increased chance of requiring IMT during follow-up (hazard ratio, 3.2; 95% confidence interval, 1.5-6.8; P = 0.002). We validated these findings in a third cohort of 67 United States patients. Conclusions: A serum protein signature correlating with neutrophil levels was highly predictive for IMT use in noninfectious uveitis. We developed a routinely available tool that may serve as a novel objective biomarker to aid in clinical decision-making for noninfectious uveitis.

A Perspective on a Quality Management System for AI/ML-Based Clinical Decision Support in Hospital Care.

Although many artificial intelligence (AI) and machine learning (ML) based algorithms are being developed by researchers, only a small fraction has been implemented in clinical-decision support (CDS) systems for clinical care. Healthcare organizations experience significant barriers implementing AI/ML models for diagnostic, prognostic, and monitoring purposes. In this perspective, we delve into the numerous and diverse quality control measures and responsibilities that emerge when moving from AI/ML-model development in a research environment to deployment in clinical care. The Sleep-Well Baby project, a ML-based monitoring system, currently being tested at the neonatal intensive care unit of the University Medical Center Utrecht, serves as a use-case illustrating our personal learning journey in this field. We argue that, in addition to quality assurance measures taken by the manufacturer, user responsibilities should be embedded in a quality management system (QMS) that is focused on life-cycle management of AI/ML-CDS models in a medical routine care environment. Furthermore, we highlight the strong similarities between AI/ML-CDS models and in vitro diagnostic devices and propose to use ISO15189, the quality guideline for medical laboratories, as inspiration when building a QMS for AI/ML-CDS usage in the clinic. We finally envision a future in which healthcare institutions run or have access to a medical AI-lab that provides the necessary expertise and quality assurance for AI/ML-CDS implementation and applies a QMS that mimics the ISO15189 used in medical laboratories.