Bartlomiej Piechowski-Jozwiak

This manuscript presents a proof-of-concept for a generalizable strategy, the full algorithm, designed to estimate disease risk using real-world clinical tabular data systems, such as electronic health records (EHR) or claims databases. By integrating classic statistical methods and modern artificial intelligence techniques, this strategy automates the production of a disease prediction model that comprehensively reflects the dynamics contained within the underlying data system. Specifically, the full algorithm parses through every facet of the data (e.g., encounters, diagnoses, procedures, medications, labs, chief complaints, flowsheets, vital signs, demographics, etc.), selects which factors to retain as predictor variables by evaluating the data empirically against statistical criteria, structures and formats the retained data into time-series, trains a neural network-based prediction model, then subsequently applies this model to current patients to generate risk estimates. A distinguishing feature of the proposed strategy is that it produces a self-adaptive prediction system, capable of evolving the prediction mechanism in response to changes within the data: as newly collected data expand/modify the dataset organically, the prediction mechanism automatically evolves to reflect these changes. Moreover, the full algorithm operates without the need for a-priori data curation and aims to harness all informative risk and protective factors within the real-world data. This stands in contrast to traditional approaches, which often rely on highly curated datasets and domain expertise to build static prediction models based solely on well-known risk factors. As a proof-of-concept, we codified the full algorithm and tasked it with estimating 12-month risk of initial stroke or myocardial infarction using our hospital’s real-world EHR. A 66-month pseudo-prospective validation was conducted using records from 558,105 patients spanning April 2015 to September 2023, totalling 3,424,060 patient-months. Area under the receiver operating characteristic curve (AUROC) values ranged from .830 to .909, with an improving trend over time. Odds ratios describing model precision for patients 1–100 and 101–200 (when ranked by estimated risk) ranged from 15.3 to 48.1 and 7.2 to 45.0, respectively, with both groups showing improving trends over time. Findings suggest the feasibility of developing high-performing disease risk calculators in the proposed manner.

Stroke is a life-threatening medical condition that could lead to mortality or significant sensorimotor deficits. Various machine learning techniques have been successfully used to detect and predict stroke-related outcomes. Considering the diversity in the type of clinical modalities involved during management of patients with stroke, such as medical images, bio-signals, and clinical data, multimodal machine learning has become increasingly popular. Thus, we conducted a systematic literature review to understand the current status of state-of-the-art multimodal machine learning methods for stroke prognosis and diagnosis. Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines during literature search and selection, our results show that the most dominant techniques are related to the fusion paradigm, specifically early, joint and late fusion. We discuss opportunities to leverage other multimodal learning paradigms, such as multimodal translation and alignment, which are generally less explored. We also discuss the scale of datasets and types of modalities used to develop existing models, highlighting opportunities for the creation of more diverse multimodal datasets. Finally, we present ongoing challenges and provide a set of recommendations to drive the next generation of multimodal learning methods for improved prognosis and diagnosis of patients with stroke.

Stroke is a life-threatening medical condition and is a leading cause of disability. Cerebral ischemia is characterized by a distinct inflammatory response starting with the production of various cytokines and other inflammation-related agents. Progranulin (PGRN), a multifunctional protein, is critical in diverse physiological reactions, such as cell proliferation, inflammation, wound healing, and nervous system development. A mature PGRN is anti-inflammatory, while granulin, its derivative, conversely induces pro-inflammatory cytokine expression. PGRN is significantly involved in the brain tissue and its damage, for example, improving mood and cognitive disorders caused by cerebral ischemia. It may also have protective effects against nerve and spinal cord injuries by inhibiting neuroinflammatory response and apoptosis or it may be related to the proliferation, accumulation, differentiation, and activation of microglia. PGRN is a neurotrophic factor in the central nervous system. It may increase post-stroke neurogenesis of the subventricular zone (SVZ), which is particularly important in improving long-term brain function following cerebral ischemia. The neurogenesis enhanced via PGRN in the ischemic brain SVZ may be attributed to the induction of PI3K/AKT and MAPK/ERK signaling routes. PGRN can also promote the proliferation of neural stem/progenitor cells through PI3K/AKT signaling pathway. PGRN increases hippocampal neurogenesis, reducing anxiety and impaired spatial learning post-cerebral ischemia. PGRN alleviates cerebral ischemia/reperfusion injury by reducing endoplasmic reticulum stress and suppressing the NF-κB signaling pathway. PGRN can be introduced as a potent neuroprotective agent capable of improving post-ischemia neuronal actions, mainly by reducing and elevating the inflammatory and anti-inflammatory cytokines. Graphical Abstract Expression, storage, cleavage, and function of progranulin (PGRN) in the pathogenesis of ischemic stroke.

Omega-3 fatty acids constitute a group of fatty acids with anti-inflammatory and preventive effects against various diseases. Studies in animal models have demonstrated the preventive and therapeutic effects of omega-3 fatty acids after spinal cord injury (SCI) in reducing inflammatory reactions and promoting neuroregeneration. However, studies on the efficacy of omega-3 fatty acids in treatment and prevention after SCI seem to be questionable. This study evaluates potential reasons for omega-3 fatty acid therapy oversight in populations after SCI. Therefore, some of the reasons could cover heterogeneous patient groups in size, level of injury, quality of life assessment, time since injury, no single standardised dose, various follow-up durations and metabolic changes, often insufficient to record. Due to the difficulty of collecting cases for the study, especially in the acute phase after SCI, multicenter, coordinated studies are needed to establish the effects of omega-3 fatty acids on treatment, recovery, and disease prevention in patients after SCI. Although the present results of such studies are still inconclusive, the failure to exploit the potential properties of omega-3 fatty acids in the treatment of patients with SCI solely due to methodological difficulties should be considered a potential waste.

Desmoteplase is a bat (Desmodus rotundus) saliva-derived fibrinolytic enzyme resembling a urokinase and tissue plasminogen activator. It is highly dependent on fibrin and has some neuroprotective attributes. Intravenous administration of desmoteplase is safe and well tolerated in healthy subjects. Plasma fibrinolytic activity is linearly related to its blood concentration, its terminal elimination half-life ranges from 3.8 to 4.92 h (50 vs. 90 μg/kg dose). Administration of desmoteplase leads to transitory derangement of fibrinogen, D-dimer, alpha2-antiplasmin, and plasmin and antiplasmin complex which normalize within 4–12 h. It does not alter a prothrombin test, international normalized ratio, activated partial thromboplastin time, and prothrombin fragment 1.2. Desmoteplase was tested in myocardial infarction and pulmonary embolism and showed promising results versus alteplase. In ischemic stroke trials, desmoteplase was linked to increased rates of symptomatic intracranial hemorrhages and case fatality. However, data from “The desmoteplase in Acute Ischemic Stroke” Trials, DIAS-3 and DIAS-J, suggest that the drug is well tolerated and its safety profile is comparable to placebo. Desmoteplase is theoretically a superior thrombolytic because of high fibrin specificity, no activation of beta-amyloid, and lack of neurotoxicity. It was associated with better outcomes in patients with significant stenosis or occlusion of a proximal precerebral vessels. However, DIAS-4 was stopped as it might have not reached its primary endpoint. Due to its promising properties, desmoteplase may be added into treatment of ischemic stroke with extension of the time window and special emphasis on patients presenting outside the 4.5-h thrombolysis window, with wake-up strokes and strokes of unknown onset.

SUMMARY: The mechanisms and phenotype of ischemic stroke associated with coronavirus disease 2019 (COVID-19) remain uncertain. A retrospective study was conducted in patients with COVID-19 presenting with ischemic stroke from March 1 to May 25, 2020, and cases with large-vessel occlusion were identified. To provide baseline institutional stroke data within and outside the COVID-19 pandemic, all consecutive ischemic stroke and TIA admissions (COVID and non-COVID) to the hospital during a 10-week period from March 1 to May 10, 2020, were collected and compared with data from the same time period in 2019. Among 20 patients with COVID-19 and acute ischemic stroke, 15 (75%) had large-vessel occlusion. These patients were young (mean age, 46.5 years), male (93%), without major burden of traditional cardiovascular risk factors, and had a severe stroke presentation. Large-vessel occlusions were observed in multiple vessels (40%), uncommonly affected vessels, and atypical locations with a large thrombus burden. Systemic thrombosis separate from large-vessel occlusion was not uncommon (26%). At short-term follow-up, stroke etiology remained undetermined in 46% of patients and functional outcome was poor. The above findings raise the possibility of stroke related to mechanisms induced by the COVID-19 infection itself, including a hypercoagulable state and/or endothelial damage. In addition, they document the severe presentation and poor outcomes of large-vessel occlusion in COVID-19 ischemic stroke.

Introduction Worldwide, 2 million patients aged 18–50 years suffer a stroke each year, and this number is increasing. Knowledge about global distribution of risk factors and aetiologies, and information about prognosis and optimal secondary prevention in young stroke patients are limited. This limits evidence-based treatment and hampers the provision of appropriate information regarding the causes of stroke, risk factors and prognosis of young stroke patients. Methods and analysis The Global Outcome Assessment Life-long after stroke in young adults (GOAL) initiative aims to perform a global individual patient data meta-analysis with existing data from young stroke cohorts worldwide. All patients aged 18–50 years with ischaemic stroke or intracerebral haemorrhage will be included. Outcomes will be the distribution of stroke aetiology and (vascular) risk factors, functional outcome after stroke, risk of recurrent vascular events and death and finally the use of secondary prevention. Subgroup analyses will be made based on age, gender, aetiology, ethnicity and climate of residence. Ethics and dissemination Ethical approval for the GOAL study has already been obtained from the Medical Review Ethics Committee region Arnhem-Nijmegen. Additionally and when necessary, approval will also be obtained from national or local institutional review boards in the participating centres. When needed, a standardised data transfer agreement will be provided for participating centres. We plan dissemination of our results in peer-reviewed international scientific journals and through conference presentations. We expect that the results of this unique study will lead to better understanding of worldwide differences in risk factors, causes and outcome of young stroke patients.

Background: Cardioembolism in paroxysmal atrial fibrillation (PAF) is a preventable cause of transient ischaemic attack (TIA) or ischaemic stroke; however, due to its transient nature, a short-duration Holter monitor may miss a significant proportion of events. Methods: We conducted an open-label randomised controlled trial of cardiac monitoring after a TIA or ischaemic stroke comparing a 14-day ECG monitoring patch (Zio^® Patch, iRhythm Technologies) with short-duration Holter monitoring for the detection of PAF. The primary outcome was the detection of one or more episodes of ECG-documented PAF lasting at least 30 s within 90 days in each of the study arms. A budget impact analysis from the healthcare perspective was performed. Results: From February 2016 through February 2017, 43 (76.8%) of the 56 patients assigned to the patch-based monitoring group and 47 (78.3%) of the 60 patients assigned to short-duration Holter monitoring group had successful monitor placement with 90 days of follow-up. Of the 26 protocol failures between the two groups, 23 (88.5%) were due to patient refusal for outpatient short-duration ECG monitor placement, whilst only 1 (3.8%) was due unsuccessful ZioPatch placement. The rate of detection of PAF at 90 days was 16.3% in the patch-based monitoring group (seven patients) compared to 2.1% in the short-duration Holter monitoring group (1 patient), with an odds ratio of 8.9 (95% CI 1.1–76.0; P = 0.026). An economic model demonstrated that implementation of the Zio Patch service would result in 10.8 more strokes avoided per year compared to current practice with Holter monitoring with an associated yearly saving in direct medical costs of £113,630, increasing to £162,491 over 5 years. Conclusions: Early, prolonged, patch-based monitoring after an index stroke or TIA is superior to short-duration Holter monitoring in the detection of PAF and likely cost-effective for preventing recurrent strokes. Trial registrationhttp://www.isrctn.com. Unique identifier: ISRCTN 50253271. Registered 21 January 2016