Swarm Learning for decentralized and confidential clinical machine learning.

Fast and reliable detection of patients with severe and heterogeneous illnesses is a major goal of precision medicine1,2. Patients with leukaemia can be identified using machine learning on the basis of their blood transcriptomes3. However, there is an increasing divide between what is technically possible and what is allowed, because of privacy legislation4,5. Here, to facilitate the integration of any medical data from any data owner worldwide without violating privacy laws, we introduce Swarm Learning-a decentralized machine-learning approach that unites edge computing, blockchain-based peer-to-peer networking and coordination while maintaining confidentiality without the need for a central coordinator, thereby going beyond federated learning. To illustrate the feasibility of using Swarm Learning to develop disease classifiers using distributed data, we chose four use cases of heterogeneous diseases (COVID-19, tuberculosis, leukaemia and lung pathologies). With more than 16,400 blood transcriptomes derived from 127 clinical studies with non-uniform distributions of cases and controls and substantial study biases, as well as more than 95,000 chest X-ray images, we show that Swarm Learning classifiers outperform those developed at individual sites. In addition, Swarm Learning completely fulfils local confidentiality regulations by design. We believe that this approach will notably accelerate the introduction of precision medicine.

qSOFA combined with suPAR for early risk detection and guidance of antibiotic treatment in the emergency department: a randomized controlled trial.

BACKGROUND: Sepsis guidelines suggest immediate start of resuscitation for patients with quick Sequential Organ Failure Assessment (qSOFA) 2 or 3. However, the interpretation of qSOFA 1 remains controversial. We investigated whether measurements of soluble urokinase plasminogen activator receptor (suPAR) may improve risk detection when qSOFA is 1. METHODS: The study had two parts. At the first part, the combination of suPAR with qSOFA was analyzed in a prospective cohort for early risk detection. At the second part, the double-blind, randomized controlled trial (RCT) SUPERIOR evaluated the efficacy of the suPAR-guided medical intervention. SUPERIOR took place between November 2018 and December 2020. Multivariate stepwise Cox regression was used for the prospective cohort, while univariate and multivariate logistic regression was used for the RCT. Consecutive admissions at the emergency department (ED) with suspected infection, qSOFA 1 and suPAR ≥ 12 ng/mL were allocated to single infusion of placebo or meropenem. The primary endpoint was early deterioration, defined as at least one-point increase of admission Sequential Organ Failure Assessment (SOFA) score the first 24 h. RESULTS: Most of the mortality risk was for patients with qSOFA 2 and 3. Taking the hazard ratio (HR) for death of patients with qSOFA = 1 and suPAR < 12 ng/mL as reference, the HR of qSOFA = 1 and suPAR ≥ 12 ng/mL for 28-day mortality was 2.98 (95% CI 2.11-3.96). The prospective RCT was prematurely ended due to pandemia-related ED re-allocations, with 91 patients enrolled: 47 in the placebo and 44 in the meropenem arm. The primary endpoint was met in 40.4% (n = 19) and 15.9% (n = 7), respectively (difference 24.5% [5.9-40.8]; odds ratio 0.14 [0.04-0.50]). One post hoc analysis showed significant median changes of SOFA score after 72 and 96 h equal to 0 and - 1, respectively. CONCLUSIONS: Combining qSOFA 1 with the biomarker suPAR improves its prognostic performance for unfavorable outcome and can help decision for earlier treatment. Trial registration EU Clinical Trials Register (EudraCT, 2018-001008-13) and Clinical-Trials.gov (NCT03717350). Registered 24 October 2018.

Integration of heparin-binding protein and interleukin-6 in the early prediction of respiratory failure and mortality in pneumonia by SARS-CoV-2 (COVID-19).

Recent publications on the probable role of heparin-binding protein (HBP) as a biomarker in sepsis prompted us to investigate its diagnostic and prognostic performance in severe COVID-19. HBP and IL-6 were measured by immunoassays at admission and on day 7 in 178 patients with pneumonia by SARS-CoV-2. Patients were classified into non-sepsis and sepsis as per the Sepsis-3 definitions and were followed up for the development of severe respiratory failure (SRF) and for outcome. Results were confirmed by multivariate analyses. HBP was significantly higher in patients classified as having sepsis and was negatively associated with the oxygenation ratio and positively associated with creatinine and lactate. Logistic regression analysis evidenced admission HBP more than 18 ng/ml and IL-6 more than 30 pg/ml as independent risk factors for the development of SRP. Their integration prognosticated SRF with respective sensitivity, specificity, positive predictive value, and negative predictive 59.1%, 96.3%, 83.9%, and 87.8%. Cox regression analysis evidenced admission HBP more than 35 ng/ml and IL-6 more than 30 pg/ml as independent risk factors for 28-day mortality. Their integration prognosticated 28-day mortality with respective sensitivity, specificity, positive predictive value, and negative predictive value 69.2%, 92.7%, 42.9%, and 97.5%. HBP remained unchanged over-time course. A prediction score of the disposition of patients with COVID-19 is proposed taking into consideration admission levels of IL-6 and HBP. Using different cut-offs, the score may predict the likelihood for SRF and for 28-day outcome.

Plasmacytoid dendritic cells drive acute asthma exacerbations.

BACKGROUND: Although acute exacerbations, mostly triggered by viruses, account for the majority of hospitalizations in asthmatic patients, there is still very little known about the pathophysiologic mechanisms involved. Plasmacytoid dendritic cells (pDCs), prominent cells of antiviral immunity, exhibit proinflammatory or tolerogenic functions depending on the context, yet their involvement in asthma exacerbations remains unexplored. OBJECTIVES: We sought to investigate the role of pDCs in allergic airway inflammation and acute asthma exacerbations. METHODS: Animal models of allergic airway disease (AAD) and virus-induced AAD exacerbations were used to dissect pDC function in vivo and unwind the potential mechanisms involved. Sputum from asthmatic patients with stable disease or acute exacerbations was further studied to determine the presence of pDCs and correlation with inflammation. RESULTS: pDCs were key mediators of the immunoinflammatory cascade that drives asthma exacerbations. In animal models of AAD and rhinovirus-induced AAD exacerbations, pDCs were recruited to the lung during inflammation and migrated to the draining lymph nodes to boost TH2-mediated effector responses. Accordingly, pDC depletion after allergen challenge or during rhinovirus infection abrogated exacerbation of inflammation and disease. Central to this process was IL-25, which was induced by allergen challenge or rhinovirus infection and conditioned pDCs for proinflammatory function. Consistently, in asthmatic patients pDC numbers were markedly increased during exacerbations and correlated with the severity of inflammation and the risk for asthma attacks. CONCLUSIONS: Our studies uncover a previously unsuspected role of pDCs in asthma exacerbations with potential diagnostic and prognostic implications. They also propose the therapeutic targeting of pDCs and IL-25 for the treatment of acute asthma.

COVID­19 vaccination in liver transplant recipients (Review).

Severe acute respiratory syndrome coronavirus 2 (Sars-CoV-2) infection has significantly affected immunocompromised individuals and subsequently, liver transplant recipients (LTRs). Early in the course of pandemic, this vulnerable population was prioritized for vaccination, after obtaining encouraging data about the vaccination benefits on disease severity and mortality. As the published knowledge was mainly supported from studies which were limited to the healthy population, the present review summarizes the data from the literature on coronavirus disease 2019 (COVID-19) vaccination in LTRs and the available vaccination guidelines of international societies. The COVID-19 vaccination of LTRs is strongly recommended as a safe and effective measure in order to prevent severe disease and mortality.

Analysis of inflammatory protein profiles in the circulation of COVID-19 patients identifies patients with severe disease phenotypes.

BACKGROUND: The coronavirus disease (COVID-19) caused by the severe acute respiratory syndrome-coronavirus-2 (SARS-CoV-2) can present with a broad range of clinical manifestations, ranging from asymptomatic to severe multiple organ failure. The severity of the disease can vary depending on factors such as age, sex, ethnicity, and pre-existing medical conditions. Despite multiple efforts to identify reliable prognostic factors and biomarkers, the predictive capacity of these markers for clinical outcomes remains poor. Circulating proteins, which reflect the active mechanisms in an individual, can be easily measured in clinical practice and therefore may be useful as biomarkers for COVID-19 disease severity. In this study, we sought to identify protein biomarkers and endotypes for COVID-19 severity and evaluate their reproducibility in an independent cohort. METHODS: We investigated a cohort of 153 Greek patients with confirmed SARS-CoV-2 infection in which plasma protein levels were measured using the Olink Explore 1536 panel, which consists of 1472 proteins. We compared the protein profiles from severe and moderate COVID-19 patients to identify proteins associated with disease severity. To evaluate the reproducibility of our findings, we compared the protein profiles of 174 patients with comparable COVID-19 severities in a US COVID-19 cohort to identify proteins consistently correlated with COVID-19 severity in both groups. RESULTS: We identified 218 differentially regulated proteins associated with severity, 20 proteins were also replicated in an external cohort which we used for validation. Moreover, we performed unsupervised clustering of patients based on 97 proteins with the highest log2 fold changes in order to identify COVID-19 endotypes. Clustering of patients based on differentially regulated proteins revealed the presence of three clinical endotypes. While endotypes 2 and 3 were enriched for severe COVID-19 patients, endotypes 3 represented the most severe form of the disease. CONCLUSIONS: These results suggest that identified circulating proteins may be useful for identifying COVID-19 patients with worse outcomes, and this potential utility may extend to other populations. TRIAL REGISTRATION: NCT04357366.

COVID-19 in Patients with Melanoma: A Single-Institution Study.

We conducted a single-center, non-interventional retrospective study of melanoma patients with COVID-19 (1 March 2020 until 17 March 2023). The cohort was further divided into three groups according to the periods of SARS-CoV-2 variant dominance in Greece. We recorded demographics, comorbidities, vaccination data, cancer diagnosis/stage, types of systemic melanoma treatments, date of COVID-19 diagnosis and survival. We identified 121 patients. The vast majority (87.6%) had advanced disease (stages III or IV). A total of 80.1% of the patients were receiving immune checkpoint inhibitor-based therapies, 92.5% had asymptomatic/mild COVID-19 and 7.4% had moderate/severe/critical disease, while 83.5% contracted COVID-19 during the third period of the pandemic. Sixteen patients (13.2%) were hospitalized for COVID-19 with a median length of stay of 12 days (range: 1-55 days). Advanced age, heart failure, number of comorbidities (≤1 vs. >1), vaccination status and the time period of the infection correlated with more severe COVID-19, whereas only heart failure and time period were independently correlated with severity. The 30-day mortality rate after COVID-19 was 4.2%. With a median follow-up of 340 days post-COVID-19, 17.4% of patients were deceased. In this cohort of melanoma patients with COVID-19, the 30-day mortality rate was low. There was no association between melanoma stage, treatment receipt and type of treatment with COVID-19 severity.

ESCAPE: An Open-Label Trial of Personalized Immunotherapy in Critically lll COVID-19 Patients.

BACKGROUND: Macrophage activation-like syndrome (MALS) and complex immune dysregulation (CID) often underlie acute respiratory distress (ARDS) in COVID-19. We aimed to investigate the effect of personalized immunotherapy on clinical improvement of critical COVID-19. METHODS: In this open-label prospective trial, 102 patients with ARDS by SARS-CoV-2 were screened for MALS (ferritin >4,420 ng/mL) and CID (ferritin ≤4,420 ng/mL and low human leukocyte antigen (HLA)-DR expression on CD14-monocytes). Patients with MALS or CID with increased aminotransferases received intravenous anakinra; those with CID and normal aminotransferases received tocilizumab. The primary outcome was ≥25% decrease in the Sequential Organ Failure Assessment (SOFA) score and/or 50% increase in the respiratory ratio by day 8; 28-day mortality, change of SOFA score by day 28, serum biomarkers, and cytokine production by mononuclear cells were secondary endpoints. RESULTS: The primary study endpoint was met in 58.3% of anakinra-treated patients and in 33.3% of tocilizumab-treated patients (p: 0.01). Most patients in both groups received dexamethasone as standard of care. No differences were found in secondary outcomes, mortality, and SOFA score changes. Ferritin decreased among anakinra-treated patients; interleukin-6, soluble urokinase plasminogen activator receptor, and HLA-DR expression increased among tocilizumab-treated patients. Survivors by day 28 who received anakinra were distributed to lower severity levels of the WHO clinical progression scale. Greater incidence of secondary infections was found with tocilizumab treatment. CONCLUSION: Immune assessment resulted in favorable anakinra responses among critically ill patients with COVID-19 and features of MALS.

Development and validation of SCOPE score: A clinical score to predict COVID-19 pneumonia progression to severe respiratory failure.

Most patients infected with SARS-CoV-2 (COVID-19) experience mild, non-specific symptoms, but many develop severe symptoms associated with an excessive inflammatory response. Elevated plasma concentrations of soluble urokinase plasminogen activator receptor (suPAR) provide early warning of progression to severe respiratory failure (SRF) or death, but access to suPAR testing may be limited. The Severe COvid Prediction Estimate (SCOPE) score, derived from circulating concentrations of C-reactive protein, D- dimers, interleukin-6, and ferritin among patients not receiving non-invasive or invasive mechanical ventilation during the SAVE-MORE study, offers predictive accuracy for progression to SRF or death within 14 days comparable to that of a suPAR concentration of ≥6 ng/mL (area under receiver operator characteristic curve 0.81 for both). The SCOPE score is validated in two similar independent cohorts. A SCOPE score of 6 or more is an alternative to suPAR for predicting progression to SRF or death within 14 days of hospital admission for pneumonia, and it can be used to guide treatment decisions.

BioFire® FilmArray® Pneumonia Panel for Severe Lower Respiratory Tract Infections: Subgroup Analysis of a Randomized Clinical Trial.

INTRODUCTION: The epidemiology of severe lower respiratory tract infections (LRTI) is constantly changing. We aimed to describe it using the BioFire® FilmArray® Pneumonia plus (PNplus) Panel. METHODS: In a sub-study of the PROGRESS trial, sputum samples of 90 patients with sepsis and LRTI were retrospectively studied. The primary endpoint was the comparative detection rate of pathogens between conventional microbiology and PNplus Panel; secondary endpoints were microbiology and the association with the inflammatory host response. RESULTS: Fifty-six patients with community-acquired pneumonia without risk factors for multidrug-resistant (MDR) pathogens and another 34 patients with risk factors for MDR were studied; median pneumonia severity index (PSI) was 113 (88-135). PNplus detection rate was 72.2% compared to 10% by conventional microbiology (p < 0.001); Streptococcus pneumoniae was the most common pathogen. PSI and procalcitonin were greater among patients with bacterial pathogens than viral pathogens. Median procalcitonin was 0.49 ng/ml and 0.18 ng/ml among patients with ≥ 105 and < 105 copies/ml of detected bacteria, respectively (p = 0.004). Resistance reached 14.4%. CONCLUSION: PNplus detects severe pneumonia pathogens at a greater rate than conventional microbiology. High levels of inflammation accompany bacterial detection. TRIAL REGISTRATION: PROGRESS, ClinicalTrials.gov NCT03333304, 06/11/2017.

An open label trial of anakinra to prevent respiratory failure in COVID-19.

Background: It was studied if early suPAR-guided anakinra treatment can prevent severe respiratory failure (SRF) of COVID-19. Methods: A total of 130 patients with suPAR ≥6 ng/ml were assigned to subcutaneous anakinra 100 mg once daily for 10 days. Primary outcome was SRF incidence by day 14 defined as any respiratory ratio below 150 mmHg necessitating mechanical or non-invasive ventilation. Main secondary outcomes were 30-day mortality and inflammatory mediators; 28-day WHO-CPS was explored. Propensity-matched standard-of care comparators were studied. Results: 22.3% with anakinra treatment and 59.2% comparators (hazard ratio, 0.30; 95% CI, 0.20-0.46) progressed into SRF; 30-day mortality was 11.5% and 22.3% respectively (hazard ratio 0.49; 95% CI 0.25-0.97). Anakinra was associated with decrease in circulating interleukin (IL)-6, sCD163 and sIL2-R; IL-10/IL-6 ratio on day 7 was inversely associated with SOFA score; patients were allocated to less severe WHO-CPS strata. Conclusions: Early suPAR-guided anakinra decreased SRF and restored the pro-/anti-inflammatory balance. Funding: This study was funded by the Hellenic Institute for the Study of Sepsis, Technomar Shipping Inc, Swedish Orphan Biovitrum, and the Horizon 2020 Framework Programme. Clinical trial number: NCT04357366.

People infected with the SARS-CoV-2 virus, which causes COVID-19, can develop severe respiratory failure and require a ventilator to keep breathing, but this does not happen to every infected individual. Measuring a blood protein called suPAR (soluble urokinase plasminogen activator receptor) may help identify patients at the greatest risk of developing severe respiratory failure and requiring a ventilator. Previous investigations have suggested that measuring suPAR can identify pneumonia patients at highest risk for developing respiratory failure. The protein can be measured by taking a blood sample, and its levels provide a snapshot of how the body's immune system is reacting to infection, and of how it may respond to treatment. Anakinra is a drug that forms part of a class of medications called interleukin antagonists. It is commonly prescribed alone or in combination with other medications to reduce pain and swelling associated with rheumatoid arthritis. Kyriazopoulou et al. investigated whether treating COVID-19 patients who had developed pneumonia with anakinra could prevent the use of a ventilator and lower the risk of death. The findings show that treating COVID-19 patients with an injection of 100 milligrams of anakinra for ten days may be an effective approach because the drug combats inflammation. Kyriazopoulou et al. examined various markers of the immune response and discovered that anakinra was able to improve immune function, protecting a significant number of patients from going on a ventilator. The drug was also found to be safe and cause no significant adverse side effects. Administering anakinra decreased of the risk of progression into severe respiratory failure by 70%, and reduced death rates significantly. These results suggest that it may be beneficial to use suPAR as an early biomarker for identifying those individuals at highest risk for severe respiratory failure, and then treat them with anakinra. While the findings are promising, they must be validated in larger studies.

Disease severity-specific neutrophil signatures in blood transcriptomes stratify COVID-19 patients.

BACKGROUND: The SARS-CoV-2 pandemic is currently leading to increasing numbers of COVID-19 patients all over the world. Clinical presentations range from asymptomatic, mild respiratory tract infection, to severe cases with acute respiratory distress syndrome, respiratory failure, and death. Reports on a dysregulated immune system in the severe cases call for a better characterization and understanding of the changes in the immune system. METHODS: In order to dissect COVID-19-driven immune host responses, we performed RNA-seq of whole blood cell transcriptomes and granulocyte preparations from mild and severe COVID-19 patients and analyzed the data using a combination of conventional and data-driven co-expression analysis. Additionally, publicly available data was used to show the distinction from COVID-19 to other diseases. Reverse drug target prediction was used to identify known or novel drug candidates based on finding from data-driven findings. RESULTS: Here, we profiled whole blood transcriptomes of 39 COVID-19 patients and 10 control donors enabling a data-driven stratification based on molecular phenotype. Neutrophil activation-associated signatures were prominently enriched in severe patient groups, which was corroborated in whole blood transcriptomes from an independent second cohort of 30 as well as in granulocyte samples from a third cohort of 16 COVID-19 patients (44 samples). Comparison of COVID-19 blood transcriptomes with those of a collection of over 3100 samples derived from 12 different viral infections, inflammatory diseases, and independent control samples revealed highly specific transcriptome signatures for COVID-19. Further, stratified transcriptomes predicted patient subgroup-specific drug candidates targeting the dysregulated systemic immune response of the host. CONCLUSIONS: Our study provides novel insights in the distinct molecular subgroups or phenotypes that are not simply explained by clinical parameters. We show that whole blood transcriptomes are extremely informative for COVID-19 since they capture granulocytes which are major drivers of disease severity.