Blood transcriptomic discrimination of bacterial and viral infections in the emergency department: a multi-cohort observational validation study

Blood calprotectin as a biomarker for infection and sepsis - the prospective CASCADE trial

BACKGROUND

Early in the host-response to infection, neutrophils release calprotectin, triggering several immune signalling cascades. In acute infection management, identifying infected patients and stratifying these by risk of deterioration into sepsis, are crucial tasks. Recruiting a heterogenous population of patients with suspected infections from the emergency department, early in the care-path, the CASCADE trial aimed to evaluate the accuracy of blood calprotectin for detecting bacterial infections, estimating disease severity, and predicting clinical deterioration.

METHODS

In a prospective, observational trial from February 2021 to August 2022, 395 patients (n = 194 clinically suspected infection; n = 201 controls) were enrolled. Blood samples were collected at enrolment. The accuracy of calprotectin to identify bacterial infections, and to predict and identify sepsis and mortality was analysed. These endpoints were determined by a panel of experts.

RESULTS

The Area Under the Receiver Operating Characteristic (AUROC) of calprotectin for detecting bacterial infections was 0.90. For sepsis within 72 h, calprotectin's AUROC was 0.83. For 30-day mortality it was 0.78. In patients with diabetes, calprotectin had an AUROC of 0.94 for identifying bacterial infection.

CONCLUSIONS

Calprotectin showed notable accuracy for all endpoints. Using calprotectin in the emergency department could improve diagnosis and management of severe infections, in combination with current biomarkers.

CLINICAL TRIAL REGISTRATION NUMBER

DRKS00020521.

Toward a more generalizable blood RNA signature for bacterial and viral infections

Host-response profiles can discriminate different infections. A new 8-gene blood RNA signature to discriminate bacterial and viral infections extends our focus hitherto on the case mix from the US and Europe to include that of low- and middle-income countries.¹ Challenges remain.

Sepsis biomarkers and diagnostic tools with a focus on machine learning

Over the last years, there have been advances in the use of data-driven techniques to improve the definition, early recognition, subtypes characterisation, prognostication and treatment personalisation of sepsis. Some of those involve the discovery or evaluation of biomarkers or digital signatures of sepsis or sepsis sub-phenotypes. It is hoped that their identification may improve timeliness and accuracy of diagnosis, suggest physiological pathways and therapeutic targets, inform targeted recruitment into clinical trials, and optimise clinical management. Given the complexities of the sepsis response, panels of biomarkers or models combining biomarkers and clinical data are necessary, as well as specific data analysis methods, which broadly fall under the scope of machine learning. This narrative review gives a brief overview of the main machine learning techniques (mainly in the realms of supervised and unsupervised methods) and published applications that have been used to create sepsis diagnostic tools and identify biomarkers.

Enhanced red blood cell distribution width to platelet ratio is a predictor of mortality in patients with sepsis: a propensity score matching analysis based on the MIMIC-IV database

OBJECTIVE

To explore the association between dynamic changes in red blood cell distribution width to platelet count ratio (RPR) during hospitalisation and short-term mortality in patients with sepsis.

DESIGN

A retrospective cohort study using propensity score matching.

SETTING

Intensive care units (ICUs) of Beth Israel Deaconess Medical Center.

PARTICIPANTS

A total of 8731 adult patients with sepsis were included in the study. The patients were identified from the ICU of the Medical Information Mart for Intensive Care database. The observed group included patients who experienced an increase in RPR of more than 30% during the first week of ICU admission, whereas the control group included the rest.

MAIN OUTCOME AND MEASURE

Using propensity score matching, a matched control group was created. The primary outcome was 28-day mortality, and the length of hospital stay and in-hospital mortality were the secondary outcomes.

RESULTS

The difference was evident in 28-day mortality between the two groups (85.8% vs 74.5%, p<0.001, Kaplan-Meier analysis, and HR=1.896, 95% CI=1.659 to 2.168, p<0.001, Cox regression). In the secondary outcomes, there was a significant difference in in-hospital mortality (p<0.001). In addition, the study discovered that the observed groups had a significantly longer hospital stay (p<0.001). Meanwhile, the results of subgroup analyses were consistent with those of the primary analyses.

CONCLUSIONS

In patients with sepsis, a significantly increased RPR is positively associated with the short-term death rate. Continuous RPR monitoring could be a valuable measure for predicting short-term mortality in patients with sepsis.

Evaluation of a Multivalent Transcriptomic Metric for Diagnosing Surgical Sepsis and Estimating Mortality Among Critically Ill Patients

Importance

Rapid and accurate discrimination of sepsis and its potential severity currently require multiple assays with slow processing times that are often inconclusive in discerning sepsis from sterile inflammation.

Objective

To analyze a whole-blood, multivalent, host-messenger RNA expression metric for estimating the likelihood of bacterial infection and 30-day mortality and compare performance of the metric with that of other diagnostic and prognostic biomarkers and clinical parameters.

Design, Setting, and Participants

This prospective diagnostic and prognostic study was performed in the surgical intensive care unit (ICU) of a single, academic health science center. The analysis included 200 critically ill adult patients admitted with suspected sepsis (cohort A) or those at high risk for developing sepsis (cohort B) between July 1, 2020, and July 30, 2021.

Exposures

Whole-blood sample measurements of a custom 29-messenger RNA transcriptomic metric classifier for likelihood of bacterial infection (IMX-BVN-3) or 30-day mortality (severity) (IMX-SEV-3) in a clinical-diagnostic laboratory setting using an analysis platform (510[k]-cleared nCounter FLEX; NanoString, Inc), compared with measurement of procalcitonin and interleukin 6 (IL-6) plasma levels, and maximum 24-hour sequential organ failure assessment (SOFA) scores.

Main Outcomes and Measures

Estimated sepsis and 30-day mortality performance.

Results

Among the 200 patients included (124 men [62.0%] and 76 women [38.0%]; median age, 62.5 [IQR, 47.0-72.0] years), the IMX-BVN-3 bacterial infection classifier had an area under the receiver operating characteristics curve (AUROC) of 0.84 (95% CI, 0.77-0.90) for discriminating bacterial infection at ICU admission, similar to procalcitonin (0.85 [95% CI, 0.79-0.90]; P = .79) and significantly better than IL-6 (0.67 [95% CI, 0.58-0.75]; P < .001). For estimating 30-day mortality, the IMX-SEV-3 metric had an AUROC of 0.81 (95% CI, 0.66-0.95), which was significantly better than IL-6 levels (0.57 [95% CI, 0.37-0.77]; P = .006), marginally better than procalcitonin levels (0.65 [95% CI, 0.50-0.79]; P = .06), and similar to the SOFA score (0.76 [95% CI, 0.62-0.91]; P = .48). Combining IMX-BVN-3 and IMX-SEV-3 with procalcitonin or IL-6 levels or SOFA scores did not significantly improve performance. Among patients with sepsis, IMX-BVN-3 scores decreased over time, reflecting the resolution of sepsis. In 11 individuals at high risk (cohort B) who subsequently developed sepsis during their hospital course, IMX-BVN-3 bacterial infection scores did not decline over time and peaked on the day of documented infection.

Conclusions and Relevance

In this diagnostic and prognostic study, a novel, multivalent, transcriptomic metric accurately estimated the presence of bacterial infection and risk for 30-day mortality in patients admitted to a surgical ICU. The performance of this single transcriptomic metric was equivalent to or better than multiple alternative diagnostic and prognostic metrics when measured at admission and provided additional information when measured over time.

Systematic review of host genomic biomarkers of invasive bacterial disease: Distinguishing bacterial from non-bacterial causes of acute febrile illness

Background

Infectious diseases play a significant role in the global burden of disease. The gold standard for the diagnosis of bacterial infection, bacterial culture, can lead to diagnostic delays and inappropriate antibiotic use. The advent of high- throughput technologies has led to the discovery of host-based genomic biomarkers of infection, capable of differentiating bacterial from other causes of infection, but few have achieved validation for use in a clinical setting.

Methods

A systematic review was performed. PubMed/Ovid Medline, Ovid Embase and Scopus databases were searched for relevant studies from inception up to 30/03/2022 with forward and backward citation searching of key references. Studies assessing the diagnostic performance of human host genomic biomarkers of bacterial infection were included. Study selection and assessment of quality were conducted by two independent reviewers. A meta-analysis was undertaken using a diagnostic random-effects model. The review was registered with PROSPERO (ID: CRD42021208462).

Findings

Seventy-two studies evaluating the performance of 116 biomarkers in 16,216 patients were included. Forty-six studies examined TB-specific biomarker performance and twenty-four studies assessed biomarker performance in a paediatric population. The results of pooled sensitivity, specificity, negative and positive likelihood ratio, and diagnostic odds ratio of genomic biomarkers of bacterial infection were 0.80 (95% CI 0.78 to 0.82), 0.86 (95% CI 0.84 to 0.88), 0.18 (95% CI 0.16 to 0.21), 5.5 (95% CI 4.9 to 6.3), 30.1 (95% CI 24 to 37), respectively. Significant between-study heterogeneity (I² 77%) was present.

Interpretation

Host derived genomic biomarkers show significant potential for clinical use as diagnostic tests of bacterial infection however, further validation and attention to test platform is warranted before clinical implementation can be achieved.

Funding

No funding received.

Bacteremia detection from complete blood count and differential leukocyte count with machine learning: complementary and competitive with C-reactive protein and procalcitonin tests

Background

Biomarkers, such as leukocyte count, C-reactive protein (CRP), and procalcitonin (PCT), have been commonly used to predict the occurrence of life-threatening bacteremia and provide prognostic information, given the need for prompt intervention. However, such diagnosis methods require much time and money. Therefore, we propose a method with a high prediction capability using machine learning (ML) models based on complete blood count (CBC) and differential leukocyte count (DC) and compare its performance with traditional CRP or PCT biomarker methods and those of models incorporating CRP or PCT biomarkers.

Methods

We collected 366,586 daily blood culture (BC) results, of which 350,775 (93.2%), 308,803 (82.1%), and 23,912 (6.4%) cases were issued CBC/DC (CBC/DC group), CRP with CBC/DC (CRP&CBC/DC group), and PCT with CBC/DC (PCT&CBC/DC group), respectively. For the ML methods, conventional logistic regression and random forest models were selected, trained, applied, and validated for each group. Fivefold validation and prediction capability were also evaluated and reported.

Results

Overall, the ML methods, such as the random forest model, demonstrated promising performances. When trained with CBC/DC data, it achieved an area under the ROC curve (AUC) of 0.802, which is superior to the prediction conventionally made with CRP/PCT levels (0.699/0.731). Upon evaluating the performance enhanced by incorporating CRP or PCT biomarkers, it reported no substantial AUC increase with the addition of either CRP or PCT to CBC/DC data, which suggests the predicting power and applicability of using only CBC/DC data. Moreover, it showed competitive prognostic capability compared to the PCT test with similar all-cause in-hospital mortality (45.10% vs. 47.40%) and overall median survival time (27 vs. 25 days).

Conclusions

The ML models using only CBC/DC data yielded more accurate bacteremia predictions compared to those by methods using CRP and PCT data and reached similar prognostic performance as by PCT data. Thus, such models are potentially complementary and competitive with traditional CRP and PCT biomarkers for conducting and guiding antibiotic usage.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12879-022-07223-7.

Systematic comparison of published host gene expression signatures for bacterial/viral discrimination

Background

Measuring host gene expression is a promising diagnostic strategy to discriminate bacterial and viral infections. Multiple signatures of varying size, complexity, and target populations have been described. However, there is little information to indicate how the performance of various published signatures compare to one another.

Methods

This systematic comparison of host gene expression signatures evaluated the performance of 28 signatures, validating them in 4589 subjects from 51 publicly available datasets. Thirteen COVID-specific datasets with 1416 subjects were included in a separate analysis. Individual signature performance was evaluated using the area under the receiving operating characteristic curve (AUC) value. Overall signature performance was evaluated using median AUCs and accuracies.

Results

Signature performance varied widely, with median AUCs ranging from 0.55 to 0.96 for bacterial classification and 0.69–0.97 for viral classification. Signature size varied (1–398 genes), with smaller signatures generally performing more poorly (P < 0.04). Viral infection was easier to diagnose than bacterial infection (84% vs. 79% overall accuracy, respectively; P < .001). Host gene expression classifiers performed more poorly in some pediatric populations (3 months–1 year and 2–11 years) compared to the adult population for both bacterial infection (73% and 70% vs. 82%, respectively; P < .001) and viral infection (80% and 79% vs. 88%, respectively; P < .001). We did not observe classification differences based on illness severity as defined by ICU admission for bacterial or viral infections. The median AUC across all signatures for COVID-19 classification was 0.80 compared to 0.83 for viral classification in the same datasets.

Conclusions

In this systematic comparison of 28 host gene expression signatures, we observed differences based on a signature’s size and characteristics of the validation population, including age and infection type. However, populations used for signature discovery did not impact performance, underscoring the redundancy among many of these signatures. Furthermore, differential performance in specific populations may only be observable through this type of large-scale validation.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13073-022-01025-x.

Pediatric sepsis biomarkers for prognostic and predictive enrichment

Sepsis is a major public health problem in children throughout the world. Given that the treatment guidelines emphasize early recognition, there is interest in developing biomarkers of sepsis, and most attention is focused on diagnostic biomarkers. While there is a need for ongoing discovery and development of diagnostic biomarkers for sepsis, this review will focus on less well-known applications of sepsis biomarkers. Among patients with sepsis, the biomarkers can give information regarding the risk of poor outcome from sepsis, risk of sepsis-related organ dysfunction, and subgroups of patients with sepsis who share underlying biological features potentially amenable to targeted therapeutics. These types of biomarkers, beyond the traditional concept of diagnosis, address the important concepts of prognostic and predictive enrichment, which are key components of bringing the promise of precision medicine to the bedside of children with sepsis.

A host transcriptomic signature for identification of respiratory viral infections in the community

BACKGROUND

Fever-7 is a test evaluating host mRNA expression levels of IFI27, JUP, LAX, HK3, TNIP1, GPAA1 and CTSB in blood able to detect viral infections. This test has been validated mostly in hospital settings. Here we have evaluated Fever-7 to identify the presence of respiratory viral infections in a Community Health Center.

METHODS

A prospective study was conducted in the "Servicio de Urgencias de Atención Primaria" in Salamanca, Spain. Patients with clinical signs of respiratory infection and at least one point in the National Early Warning Score were recruited. Fever-7 mRNAs were profiled on a Nanostring nCounter® SPRINT instrument from blood collected upon patient enrolment. Viral diagnosis was performed on nasopharyngeal aspirates (NPAs) using the Biofire-RP2 panel.

RESULTS

A respiratory virus was detected in the NPAs of 66 of the 100 patients enrolled. Median National Early Warning Score was 7 in the group with no virus detected and 6.5 in the group with a respiratory viral infection (P > .05). The Fever-7 score yielded an overall AUC of 0.81 to predict a positive viral syndromic test. The optimal operating point for the Fever-7 score yielded a sensitivity of 82% with a specificity of 71%. Multivariate analysis showed that Fever-7 was a robust marker of viral infection independently of age, sex, major comorbidities and disease severity at presentation (OR [CI95%], 3.73 [2.14-6.51], P < .001).

CONCLUSIONS

Fever-7 is a promising host immune mRNA signature for the early identification of a respiratory viral infection in the community.

Blood transcriptional biomarkers of acute viral infection for detection of pre-symptomatic SARS-CoV-2 infection: a nested, case-control diagnostic accuracy study

BACKGROUND

We hypothesised that host-response biomarkers of viral infections might contribute to early identification of individuals infected with SARS-CoV-2, which is critical to breaking the chains of transmission. We aimed to evaluate the diagnostic accuracy of existing candidate whole-blood transcriptomic signatures for viral infection to predict positivity of nasopharyngeal SARS-CoV-2 PCR testing.

METHODS

We did a nested case-control diagnostic accuracy study among a prospective cohort of health-care workers (aged ≥18 years) at St Bartholomew's Hospital (London, UK) undergoing weekly blood and nasopharyngeal swab sampling for whole-blood RNA sequencing and SARS-CoV-2 PCR testing, when fit to attend work. We identified candidate blood transcriptomic signatures for viral infection through a systematic literature search. We searched MEDLINE for articles published between database inception and Oct 12, 2020, using comprehensive MeSH and keyword terms for "viral infection", "transcriptome", "biomarker", and "blood". We reconstructed signature scores in blood RNA sequencing data and evaluated their diagnostic accuracy for contemporaneous SARS-CoV-2 infection, compared with the gold standard of SARS-CoV-2 PCR testing, by quantifying the area under the receiver operating characteristic curve (AUROC), sensitivities, and specificities at a standardised Z score of at least 2 based on the distribution of signature scores in test-negative controls. We used pairwise DeLong tests compared with the most discriminating signature to identify the subset of best performing biomarkers. We evaluated associations between signature expression, viral load (using PCR cycle thresholds), and symptom status visually and using Spearman rank correlation. The primary outcome was the AUROC for discriminating between samples from participants who tested negative throughout the study (test-negative controls) and samples from participants with PCR-confirmed SARS-CoV-2 infection (test-positive participants) during their first week of PCR positivity.

FINDINGS

We identified 20 candidate blood transcriptomic signatures of viral infection from 18 studies and evaluated their accuracy among 169 blood RNA samples from 96 participants over 24 weeks. Participants were recruited between March 23 and March 31, 2020. 114 samples were from 41 participants with SARS-CoV-2 infection, and 55 samples were from 55 test-negative controls. The median age of participants was 36 years (IQR 27-47) and 69 (72%) of 96 were women. Signatures had little overlap of component genes, but were mostly correlated as components of type I interferon responses. A single blood transcript for IFI27 provided the highest accuracy for discriminating between test-negative controls and test-positive individuals at the time of their first positive SARS-CoV-2 PCR result, with AUROC of 0·95 (95% CI 0·91-0·99), sensitivity 0·84 (0·70-0·93), and specificity 0·95 (0·85-0·98) at a predefined threshold (Z score >2). The transcript performed equally well in individuals with and without symptoms. Three other candidate signatures (including two to 48 transcripts) had statistically equivalent discrimination to IFI27 (AUROCs 0·91-0·95).

INTERPRETATION

Our findings support further urgent evaluation and development of blood IFI27 transcripts as a biomarker for early phase SARS-CoV-2 infection for screening individuals at high risk of infection, such as contacts of index cases, to facilitate early case isolation and early use of antiviral treatments as they emerge.

FUNDING

Barts Charity, Wellcome Trust, and National Institute of Health Research.

Discovery and validation of a three-gene signature to distinguish COVID-19 and other viral infections in emergency infectious disease presentations: a case-control and observational cohort study

Background

Emergency admissions for infection often lack initial diagnostic certainty. COVID-19 has highlighted a need for novel diagnostic approaches to indicate likelihood of viral infection in a pandemic setting. We aimed to derive and validate a blood transcriptional signature to detect viral infections, including COVID-19, among adults with suspected infection who presented to the emergency department.

Methods

Individuals (aged ≥18 years) presenting with suspected infection to an emergency department at a major teaching hospital in the UK were prospectively recruited as part of the Bioresource in Adult Infectious Diseases (BioAID) discovery cohort. Whole-blood RNA sequencing was done on samples from participants with subsequently confirmed viral, bacterial, or no infection diagnoses. Differentially expressed host genes that met additional filtering criteria were subjected to feature selection to derive the most parsimonious discriminating signature. We validated the signature via RT-qPCR in a prospective validation cohort of participants who presented to an emergency department with undifferentiated fever, and a second case-control validation cohort of emergency department participants with PCR-positive COVID-19 or bacterial infection. We assessed signature performance by calculating the area under receiver operating characteristic curves (AUROCs), sensitivities, and specificities.

Findings

A three-gene transcript signature, comprising HERC6, IGF1R, and NAGK, was derived from the discovery cohort of 56 participants with bacterial infections and 27 with viral infections. In the validation cohort of 200 participants, the signature differentiated bacterial from viral infections with an AUROC of 0·976 (95% CI 0·919−1·000), sensitivity of 97·3% (85·8−99·9), and specificity of 100% (63·1−100). The AUROC for C-reactive protein (CRP) was 0·833 (0·694−0·944) and for leukocyte count was 0·938 (0·840−0·986). The signature achieved higher net benefit in decision curve analysis than either CRP or leukocyte count for discriminating viral infections from all other infections. In the second validation analysis, which included SARS-CoV-2-positive participants, the signature discriminated 35 bacterial infections from 34 SARS-CoV-2-positive COVID-19 infections with AUROC of 0·953 (0·893−0·992), sensitivity 88·6%, and specificity of 94·1%.

Interpretation

This novel three-gene signature discriminates viral infections, including COVID-19, from other emergency infection presentations in adults, outperforming both leukocyte count and CRP, thus potentially providing substantial clinical utility in managing acute presentations with infection.

Funding

National Institute for Health Research, Medical Research Council, Wellcome Trust, and EU-FP7.

New Microbiological Techniques for the Diagnosis of Bacterial Infections and Sepsis in ICU Including Point of Care

Purpose of Review

The aim of this article is to review current and emerging microbiological techniques that support the rapid diagnosis of bacterial infections in critically ill patients, including their performance, strengths and pitfalls, as well as available data evaluating their clinical impact.

Recent Findings

Bacterial infections and sepsis are responsible for significant morbidity and mortality in patients admitted to the intensive care unit and their management is further complicated by the increase in the global burden of antimicrobial resistance. In this setting, new diagnostic methods able to overcome the limits of traditional microbiology in terms of turn-around time and accuracy are highly warranted. We discuss the following broad themes: optimisation of existing culture-based methodologies, rapid antigen detection, nucleic acid detection (including multiplex PCR assays and microarrays), sepsis biomarkers, novel methods of pathogen detection (e.g. T2 magnetic resonance) and susceptibility testing (e.g. morphokinetic cellular analysis) and the application of direct metagenomics on clinical samples. The assessment of the host response through new “omics” technologies might also aid in early diagnosis of infections, as well as define non-infectious inflammatory states.

Summary

Despite being a promising field, there is still scarce evidence about the real-life impact of these assays on patient management. A common finding of available studies is that the performance of rapid diagnostic strategies highly depends on whether they are integrated within active antimicrobial stewardship programs. Assessing the impact of these emerging diagnostic methods through patient-centred clinical outcomes is a complex challenge for which large and well-designed studies are awaited.

Exclusion of bacterial co-infection in COVID-19 using baseline inflammatory markers and their response to antibiotics

BACKGROUND

COVID-19 is infrequently complicated by bacterial co-infection, but antibiotic prescriptions are common. We used community-acquired pneumonia (CAP) as a benchmark to define the processes that occur in bacterial pulmonary infections, testing the hypothesis that baseline inflammatory markers and their response to antibiotic therapy could distinguish bacterial co-infection from COVID-19.

METHODS

Retrospective cohort study of CAP (lobar consolidation on chest radiograph) and COVID-19 (PCR detection of SARS-CoV-2) patients admitted to Royal Free Hospital (RFH) and Barnet Hospital (BH), serving as independent discovery and validation cohorts. All CAP and >90% COVID-19 patients received antibiotics on hospital admission.

RESULTS

We identified 106 CAP and 619 COVID-19 patients at RFH. Compared with COVID-19, CAP was characterized by elevated baseline white cell count (WCC) [median 12.48 (IQR 8.2-15.3) versus 6.78 (IQR 5.2-9.5) ×106 cells/mL, P < 0.0001], C-reactive protein (CRP) [median 133.5 (IQR 65-221) versus 86.0 (IQR 42-160) mg/L, P < 0.0001], and greater reduction in CRP 48-72 h into admission [median ΔCRP -33 (IQR -112 to +3.5) versus +14 (IQR -15.5 to +70.5) mg/L, P < 0.0001]. These observations were recapitulated in the independent validation cohort at BH (169 CAP and 181 COVID-19 patients). A multivariate logistic regression model incorporating WCC and ΔCRP discriminated CAP from COVID-19 with AUC 0.88 (95% CI 0.83-0.94). Baseline WCC >8.2 × 106 cells/mL or falling CRP identified 94% of CAP cases, and excluded bacterial co-infection in 46% of COVID-19 patients.

CONCLUSIONS

We propose that in COVID-19, absence of both elevated baseline WCC and antibiotic-related decrease in CRP can exclude bacterial co-infection and facilitate antibiotic stewardship efforts.

Predicting tissue-specific gene expression from whole blood transcriptome

Complex diseases are mediated via transcriptional dysregulation in multiple tissues. Thus, knowing an individual's tissue-specific gene expression can provide critical information about her health. Unfortunately, for most tissues, the transcriptome cannot be obtained without invasive procedures. Could we, however, infer an individual's tissue-specific expression from her whole blood transcriptome? Here, we rigorously address this question. We find that an individual's whole blood transcriptome can significantly predict tissue-specific expression levels for ~60% of the genes on average across 32 tissues, with up to 81% of the genes in skeletal muscle. The tissue-specific expression inferred from the blood transcriptome is almost as good as the actual measured tissue expression in predicting disease state for six different complex disorders, including hypertension and type 2 diabetes, substantially surpassing the blood transcriptome. The code for tissue-specific gene expression prediction, TEEBoT, is provided, enabling others to study its potential translational value in other indications.

Artificial Intelligence in Differential Diagnostics of Meningitis: A Nationwide Study

Differential diagnosis between bacterial and viral meningitis is crucial. In our study, to differentiate bacterial vs. viral meningitis, three machine learning (ML) algorithms (multiple logistic regression (MLR), random forest (RF), and naïve-Bayes (NB)) were applied for the two age groups (0-14 and >14 years) of patients with meningitis by both conventional (culture) and molecular (PCR) methods. Cerebrospinal fluid (CSF) neutrophils, CSF lymphocytes, neutrophil-to-lymphocyte ratio (NLR), blood albumin, blood C-reactive protein (CRP), glucose, blood soluble urokinase-type plasminogen activator receptor (suPAR), and CSF lymphocytes-to-blood CRP ratio (LCR) were used as predictors for the ML algorithms. The performance of the ML algorithms was evaluated through a cross-validation procedure, and optimal predictions of the type of meningitis were above 95% for viral and 78% for bacterial meningitis. Overall, MLR and RF yielded the best performance when using CSF neutrophils, CSF lymphocytes, NLR, albumin, glucose, gender, and CRP. Also, our results reconfirm the high diagnostic accuracy of NLR in the differential diagnosis between bacterial and viral meningitis.

Correction to: Blood transcriptomic discrimination of bacterial and viral infections in the emergency department: a multi-cohort observational validation study

An amendment to this paper has been published and can be accessed via the original article.