Classification of COVID-19 and Influenza Patients Using Deep Learning.

Coronavirus (COVID-19) is a deadly virus that initially starts with flu-like symptoms. COVID-19 emerged in China and quickly spread around the globe, resulting in the coronavirus epidemic of 2019-22. As this virus is very similar to influenza in its early stages, its accurate detection is challenging. Several techniques for detecting the virus in its early stages are being developed. Deep learning techniques are a handy tool for detecting various diseases. For the classification of COVID-19 and influenza, we proposed tailored deep learning models. A publicly available dataset of X-ray images was used to develop proposed models. According to test results, deep learning models can accurately diagnose normal, influenza, and COVID-19 cases. Our proposed long short-term memory (LSTM) technique outperformed the CNN model in the evaluation phase on chest X-ray images, achieving 98% accuracy.

Predicting the short-term success of human influenza virus variants with machine learning.

Seasonal influenza viruses are constantly changing and produce a different set of circulating strains each season. Small genetic changes can accumulate over time and result in antigenically different viruses; this may prevent the body's immune system from recognizing those viruses. Due to rapid mutations, in particular, in the haemagglutinin (HA) gene, seasonal influenza vaccines must be updated frequently. This requires choosing strains to include in the updates to maximize the vaccines' benefits, according to estimates of which strains will be circulating in upcoming seasons. This is a challenging prediction task. In this paper, we use longitudinally sampled phylogenetic trees based on HA sequences from human influenza viruses, together with counts of epitope site polymorphisms in HA, to predict which influenza virus strains are likely to be successful. We extract small groups of taxa (subtrees) and use a suite of features of these subtrees as key inputs to the machine learning tools. Using a range of training and testing strategies, including training on H3N2 and testing on H1N1, we find that successful prediction of future expansion of small subtrees is possible from these data, with accuracies of 0.71-0.85 and a classifier 'area under the curve' 0.75-0.9.

Host susceptibility to severe influenza A virus infection.

Most people exposed to a new flu virus do not notice any symptoms. A small minority develops critical illness. Some of this extremely broad variation in susceptibility is explained by the size of the initial inoculum or the influenza exposure history of the individual; some is explained by generic host factors, such as frailty, that decrease resilience following any systemic insult. Some demographic factors (pregnancy, obesity, and advanced age) appear to confer a more specific susceptibility to severe illness following infection with influenza viruses. As with other infectious diseases, a substantial component of susceptibility is determined by host genetics. Several genetic susceptibility variants have now been reported with varying levels of evidence. Susceptible hosts may have impaired intracellular controls of viral replication (e.g. IFITM3, TMPRS22 variants), defective interferon responses (e.g. GLDC, IRF7/9 variants), or defects in cell-mediated immunity with increased baseline levels of systemic inflammation (obesity, pregnancy, advanced age). These mechanisms may explain the prolonged viral replication reported in critically ill patients with influenza: patients with life-threatening disease are, by definition, abnormal hosts. Understanding these molecular mechanisms of susceptibility may in the future enable the design of host-directed therapies to promote resilience.

Analytical evaluation of the microarray-based FluChip-8G Influenza A+B Assay.

BACKGROUND: Influenza causes a significant annual disease burden, with characterization of the infecting virus important in clinical and public health settings. Rapid immunoassays are fast but insensitive, whereas real-time RT-PCR is sensitive but susceptible to genetic mutations and often requires multiple serial assays. The FluChip-8G Influenza A+B Assay provides type and subtype/lineage identification of influenza A and B, including non-seasonal A viruses, in a single microarray-based assay with same day turnaround time. OBJECTIVE: To evaluate key analytical performance characteristics of the FluChip-8G Influenza A+B Assay. STUDY DESIGN: Analytical sensitivity, cross-reactivity, and multi-site reproducibility were evaluated. RESULTS: The limit of detection (LOD) for the FluChip-8G influenza A+B Assay ranged from 5.8 × 102-1.5 × 105 genome copies/mL, with most samples ∼2 × 103 genome copies/mL (∼160 genome copies/reaction). Fifty two (52) additional strains were correctly identified near the LOD, demonstrating robust reactivity. Two variant viruses (H1N1v and H3N2v) resulted in dual identification as both "non-seasonal influenza A" and A/H1N1pdm09. No reproducible cross-reactivity was observed for the 34 organisms tested, however, challenges with internal control inhibition due to crude growth matrix were observed. Lastly, samples tested near the LOD showed high reproducibility (97.0% (95% CI 94.7-98.7)) regardless of operator, site, reagent lot, or testing day. CONCLUSION: The FluChip-8G Influenza A+B Assay is an effective new method for detecting and identifying both seasonal and non-seasonal influenza viruses, as revealed by good sensitivity and robust reactivity to 52 unique strains of influenza virus. In addition, the lack of cross-reactivity to non-influenza pathogens and high lab-to-lab reproducibility highlight the analytical performance of the assay as an alternative to real-time RT-PCR and sequencing-based assays. Clinical validation of the technology in a multi-site clinical study is the subject of a separate investigation.

Acute Influenza Infection Presenting with Cardiac Tamponade: A Case Report and Review of Literature.

INTRODUCTION: Cardiac complications associated with influenza infection can occur either via a direct effect of the virus on the heart or through exacerbation of preexisting cardiovascular disease. We present a case of a 57-year-old man with acute influenza infection complicated by pericardial effusion and cardiac tamponade. CASE PRESENTATION: A 57-year-old white man presented to the Emergency Department with sudden onset of severe, nonexertional, retrosternal, pressure-like chest pain for a few hours and with fever and muscle aches for 2 days. The patient was initially admitted because of suspected acute coronary syndrome. The next morning, he complained of acute-onset shortness of breath and had hypotension and tachycardia. On examination, his peripheral extremities were cold and heart sounds were distant. Pulsus paradoxus was 20 mmHg. The electrocardiogram showed low-voltage QRS complex with electrical alternans. An urgently performed bedside echocardiogram showed moderate pericardial effusion with a small right ventricular cavity with diastolic collapse. Emergent pericardiocentesis was performed, with removal of 250 mL of fluid from the pericardial space. The patient's hemodynamic status immediately improved. Analyses of pericardial fluid demonstrated no bacteria, acid-fast bacilli, or malignant cells. The result of a rapid influenza diagnostic test with polymerase chain reaction was positive for influenza A virus, with other viral panels yielding normal results. The patient was treated with oseltamivir for 5 days. DISCUSSION: Pericardial involvement is a rare and perhaps underreported complication of influenza infection. Early recognition of cardiac symptoms and appropriate diagnostic workup in a patient presenting with influenza-like symptoms is important to avoid life-threatening complications.

Are we missing respiratory viral infections in infants and children? Comparison of a hospital-based quality management system with standard of care.

OBJECTIVES: Hospital-based surveillance of influenza and acute respiratory infections relies on International Classification of Diseases (ICD) codes and hospital laboratory reports (Standard-of-Care). It is unclear how many cases are missed with either method, i.e. remain undiagnosed/coded as influenza and other respiratory virus infections. Various influenza-like illness (ILI) definitions co-exist with little guidance on how to use them. We compared the diagnostic accuracy of standard surveillance methods with a prospective quality management (QM) programme at a Berlin children's hospital with the Robert Koch Institute. METHODS: Independent from routine care, all patients fulfilling pre-defined ILI-criteria (QM-ILI) participated in the QM programme. A separate QM team conducted standardized clinical assessments and collected nasopharyngeal specimens for blinded real-time quantitative PCR for influenza A/B viruses, respiratory syncytial virus, adenovirus, rhinovirus and human metapneumovirus. RESULTS: Among 6073 individuals with ILI qualifying for the QM programme, only 8.7% (528/6073) would have undergone virus diagnostics during Standard-of-Care. Surveillance based on ICD codes would have missed 61% (359/587) of influenza diagnoses. Of baseline ICD codes, 53.2% (2811/5282) were non-specific, most commonly J06 ('acute upper respiratory infection'). Comparison of stakeholder case definitions revealed that QM-ILI and the WHO ILI case definition showed the highest overall sensitivities (84%-97% and 45%-68%, respectively) and the CDC ILI definition had the highest sensitivity for influenza infections (36%, 95% CI 31.4-40.8 for influenza A and 48%, 95% CI 40.5-54.7 for influenza B). CONCLUSIONS: Disease-burden estimates and surveillance should account for the underreporting of cases in routine care. Future studies should explore the effect of ILI screening and surveillance in various age groups and settings. Diagnostic algorithms should be based on the WHO ILI case definition combined with targeted testing.

Estimating severity of influenza epidemics from severe acute respiratory infections (SARI) in intensive care units.

BACKGROUND: While influenza-like-illness (ILI) surveillance is well-organized at primary care level in Europe, few data are available on more severe cases. With retrospective data from intensive care units (ICU) we aim to fill this current knowledge gap. Using multiple parameters proposed by the World Health Organization we estimate the burden of severe acute respiratory infections (SARI) in the ICU and how this varies between influenza epidemics. METHODS: We analyzed weekly ICU admissions in the Netherlands (2007-2016) from the National Intensive Care Evaluation (NICE) quality registry (100% coverage of adult ICUs in 2016; population size 14 million) to calculate SARI incidence, SARI peak levels, ICU SARI mortality, SARI mean Acute Physiology and Chronic Health Evaluation (APACHE) IV score, and the ICU SARI/ILI ratio. These parameters were calculated both yearly and per separate influenza epidemic (defined epidemic weeks). A SARI syndrome was defined as admission diagnosis being any of six pneumonia or pulmonary sepsis codes in the APACHE IV prognostic model. Influenza epidemic periods were retrieved from primary care sentinel influenza surveillance data. RESULTS: Annually, an average of 13% of medical admissions to adult ICUs were for a SARI but varied widely between weeks (minimum 5% to maximum 25% per week). Admissions for bacterial pneumonia (59%) and pulmonary sepsis (25%) contributed most to ICU SARI. Between the eight different influenza epidemics under study, the value of each of the severity parameters varied. Per parameter the minimum and maximum of those eight values were as follows: ICU SARI incidence 558-2400 cumulated admissions nationwide, rate 0.40-1.71/10,000 inhabitants; average APACHE score 71-78; ICU SARI mortality 13-20%; ICU SARI/ILI ratio 8-17 cases per 1000 expected medically attended ILI in primary care); peak-incidence 101-188 ICU SARI admissions in highest-incidence week, rate 0.07-0.13/10,000 population). CONCLUSIONS: In the ICU there is great variation between the yearly influenza epidemic periods in terms of different influenza severity parameters. The parameters also complement each other by reflecting different aspects of severity. Prospective syndromic ICU SARI surveillance, as proposed by the World Health Organization, thereby would provide insight into the severity of ongoing influenza epidemics, which differ from season to season.

Association between IFITM3 rs12252 polymorphism and influenza susceptibility and severity: A meta-analysis.

Intrinsic host susceptibility to viral infections plays a major role in determining infection severity in different individuals. In human influenza virus infections, multiple genetic association studies have identified specific human gene variants that might contribute to enhanced susceptibility or resistance to influenza. Recent studies suggested, the rs12252 T > C polymorphism in the interferon-inducible transmembrane protein 3 (IFITM3) gene might be associated with susceptibility to severe influenza. However, the studies reported conflicting and inconclusive results. To resolve the controversy, we conducted a systematic meta-analysis to evaluate the role of the IFITM3 rs12252 polymorphism in influenza susceptibility and severity, including twelve studies published before February 19, 2018 with a total 16,263 subjects (1836 influenza cases and 14,427 controls). Odds ratios (OR) and 95% confidence intervals were used to assess the strength of the association. Our results indicated increased risk of both severe and mild influenza in subjects carrying the IFITM3 rs12252 polymorphism in the allele contrast C vs. T: OR (severe) = 1.69, 95% CI = 1.23-2.33, P = 0.001, and OR (mild) = 1.46, 95% CI = 1.13-1.87, P = 0.004. Similar results were obtained in the homozygote comparison and dominant model. Stratified analyses by ethnicity revealed increased risk of severe influenza in both the White and East Asian populations, but significant association with mild influenza was found only in the White population. Overall, our meta-analysis suggests a significant association between the IFITM3 rs12252 polymorphism and the risk of influenza in both the White and East Asian populations.

Assessment of 12 influenza-like illness case definitions using Department of Defense Global, Laboratory-based Influenza Surveillance Program data, 2011-2014.

Despite the growth in influenza surveillance programs, standardization of a globally accepted influenza-like illness (ILI) case definition remains difficult. With 2011-2014 Department of Defense Global, Laboratory-based Influenza Surveillance Program (DISP) data, 12 case definitions were evaluated using a combination of ILI case definitions from the Centers for Disease Control and Prevention, World Health Organization, and the DISP. The sensitivity, specificity, positive and negative predictive values, and odds ratios for each case definition were calculated. Additionally, area under the curve (AUC) was calculated for a receiver operating characteristic (ROC) curve to compare the case definitions. Between 2 October 2011 and 27 September 2014, 52.3% (5,575 of 10,662) of respiratory specimens submitted met the inclusion criteria. The case definition for the DISP had a sensitivity of 54.6% and specificity of 63.7%. Case definitions should be selected according to the objectives of the surveillance system and resources available. Sensitive case definitions capture a larger proportion of cases but at the cost of testing more specimens. Definitions with higher specificity result in fewer false positives but may miss more cases.

Multi-center evaluation of the cobas® Liat® Influenza A/B & RSV assay for rapid point of care diagnosis.

Point of Care Testing (POCT) provides the capability for rapid laboratory test results in patient care environments where a traditional clinical laboratory is not available. POCTs have shorter turn-around times (TATs), they may be performed by non-laboratory personnel, and the need for transport time is eliminated. The Food and Drug Administration (FDA) recently granted Clinical Laboratory Improvements Amendment (CLIA) waiver status to the cobas® Influenza A/B & RSV assay, a rapid, accurate point-of-care test for Influenza and respiratory syncytial virus (RSV) performed on the Liat® System. The performance characteristics of this test were determined though a multi-site study consisting of different point of care testing environments. Prospectively collected Nasopharyngeal (NP) swabs from 1361 patients seen at 8 primary care clinics and 4 emergency departments (EDs) and 295 retrospectively identified specimens were tested for Influenza A/B and RSV on the cobas® Liat® platform. Performance characteristics were determined through comparison to ProFlu+, a laboratory-based PCR test for Influenza A/B and RSV (reference test). Discordant specimens were adjudicated following bi-directional sequencing. The cobas® Influenza A/B and RSV assay showed sensitivities of 99.6%, 99.3%, and 96.8% for Influenza A, Influenza B, and RSV, respectively as determined from percent positive agreement (PPA) following comparison to the reference test. Sequencing confirmed cobas® Influenza A/B and RSV results in 49.2% of reference test discordant specimens, while crossing threshold data suggest increased sensitivity compared to the reference test. The cobas® Influenza A/B and RSV assay was found to be a rapid, sensitive POCT for the detection of these viruses, and provides laboratory-quality PCR-based diagnostic results in point of care settings.

Characterization of Seasonal Influenza Virus Type and Subtypes Isolated from Influenza Like Illness Cases of 2012.

Background Seasonal influenza is one of the increasing public health burdens in Nepal. Objective The objective of this study was to isolate and characterize the influenza virus type and subtypes of Nepal. Method A total of 1536 throat swab specimens were collected from January to December 2012. Total ribonucleic acid was extracted using Qiagen viral nucleic acid extraction kit and polymerase chain reaction assay was performed following the US; CDC Real-time PCR protocol. Ten percent of positive specimens were inoculated onto Madin-Darby Canine Kidney cells. Isolates were characterized by using reference ferret antisera. Result Of the total specimens (n=1536), influenza virus type A was detected in 196 (22%) cases; of which 194 (99%) were influenza A (H1N1) pdm09 and 2 (1 %) were influenza A/H3 subtype. Influenza B was detected in 684 (76.9%) cases. Influenza A (H1N1) pdm09, A/H3 and influenza B virus were antigenically similar to the recommended influenza virus vaccine candidate of the year 2012. Although sporadic cases of influenza were observed throughout the year, peak was observed during July to November. Conclusion Similar to other tropical countries, A (H1N1) pdm09, A/H3 and influenza B viruses were co-circulated in Nepal.

Rates of neutropenia in adults with influenza A or B: a retrospective analysis of hospitalised patients in South East Queensland during 2015.

Neutropenia in adult patients is often attributed to intercurrent viral infections; however, there are limited data describing the frequency or natural history of this phenomenon. We examined all patients presenting to three large hospitals in the Metro South region of South East Queensland with laboratory-confirmed influenza A or B throughout the 2015 influenza season (January-October). Four hundred and thirty-six patients were studied and 15.3% of this cohort were neutropenic (absolute neutrophil count <2.0 × 109 /L) with no identifiable cause other than the influenza. Importantly, the majority of cases were mild, with absolute neutrophil count remaining >1.0 × 109 /L. The incidence of neutropenia was significantly higher in association with influenza B than influenza A (18.3% vs 10.3%). We conclude that mild, transient neutropenia is common among patients with influenza infection and advise that it should not cause alarm or invite specific investigation unless severe or prolonged.

[Influenza and community acquired pneumonia in German primary care]. / Influenza und ambulant erworbene Pneumonie in hausärztlich tätigen Arztpraxen in Deutschland.

Influenza and community-acquired pneumonia (CAP) impose a considerable annual burden on the German primary care system. Yet there is a lack of epidemiological data from the country's outpatient sector on groups at risk as well as on the complications of these diseases.The Robert Koch Institute (RKI) initiated the study to identify population groups at increased risk for influenza or CAP as well as related comorbidities and sequelae. We present the methodology of the study and the descriptive analysis of the patients.ICD-10-based data was collected in 89 primary health care practices between January 2012 and April 2015 using a data extraction tool developed on behalf of the RKI. Case-based anonymized information was recorded for all patients in whom influenza, CAP or other acute respiratory infections (ARI) were diagnosed. For each patient information on all diagnoses including the date were retrospectively and prospectively collected (each for six months) as well as age, sex and influenza vaccination.Data on 156,803 patients with ARI was collected, of them 7909 patients with influenza (within influenza waves) and 8528 patients with CAP diagnoses. Influenza diagnoses showed a strong seasonal pattern and captured annual influenza waves in Germany. Of the influenza cases 1.6 % had a following diagnosis of CAP within 30 days. Age-specific prevalence of chronic diseases such as asthma and diabetes was significantly higher in the study population as compared to the German population.The developed tool delivers in a standardized fashion ICD-10-coded epidemiological data on population-based burden of influenza and CAP in Germany. As the descriptive analysis showed, the collected dataset is a reliable and solid basis for the further investigations of the study questions.

Assessment of ICD-9-based case definitions for influenza-like illness surveillance.

Population-based surveillance of influenza routinely relies on administrative medical encounter databases and ICD-9 codes. However, an assessment of the ICD-9 codes used for the Department of Defense (DoD) influenza-like illness (ILI) case definition has not been conducted since 2007. As coding practices may have changed over time, this analysis was done to determine the sensitivity, specificity, and positive predictive value (PPV) of the current ILI case definition and three alternative case definitions for the 2014-2015 influenza season. Influenza laboratory tests conducted on specimens from DoD beneficiaries during the 2014-2015 season were matched to ambulatory and inpatient medical encounters. The current DoD ILI case definition had high sensitivity (92%) but low specificity (30%) and moderate PPV (63%). A more specific ILI case definition utilizing only codes with greater than 75% influenza positivity for the matched laboratory test had high specificity (96%) and PPV (96%) and moderate sensitivity (62%). The current ILI case definition is sufficient for broad, sensitive population-based surveillance; however, an alternative case definition may be more appropriate when there is a need to maximize specificity.

An infectious disease/fever screening radar system which stratifies higher-risk patients within ten seconds using a neural network and the fuzzy grouping method.

OBJECTIVES: To classify higher-risk influenza patients within 10 s, we developed an infectious disease and fever screening radar system. METHODS: The system screens infected patients based on vital signs, i.e., respiration rate measured by a radar, heart rate by a finger-tip photo-reflector, and facial temperature by a thermography. The system segregates subjects into higher-risk influenza (HR-I) group, lower-risk influenza (LR-I) group, and non-influenza (Non-I) group using a neural network and fuzzy clustering method (FCM). We conducted influenza screening for 35 seasonal influenza patients and 48 normal control subjects at the Japan Self-Defense Force Central Hospital. Pulse oximetry oxygen saturation (SpO2) was measured as a reference. RESULTS: The system classified 17 subjects into HR-I group, 26 into LR-I group, and 40 into Non-I group. Ten out of the 17 HR-I subjects indicated SpO2 <96%, whereas only two out of the 26 LR-I subjects showed SpO2 <96%. The chi-squared test revealed a significant difference in the ratio of subjects showed SpO2 <96% between HR-I and LR-I group (p < 0.001). There were zero and nine normal control subjects in HR-I and LR-I groups, respectively, and there was one influenza patient in Non-I group. CONCLUSIONS: The combination of neural network and FCM achieved efficient detection of higher-risk influenza patients who indicated SpO2 96% within 10 s.

Characterization of pandemic influenza immune memory signature after vaccination or infection.

The magnitude, quality, and maintenance of immunological memory after infection or vaccination must be considered for future design of effective influenza vaccines. In 2009, the influenza pandemic produced disease that ranged from mild to severe, even fatal, illness in infected healthy adults and led to vaccination of a portion of the population with the adjuvanted, inactivated influenza A(H1N1)pdm09 vaccine. Here, we have proposed a multiparameter quantitative and qualitative approach to comparing adaptive immune memory to influenza 1 year after mild or severe infection or vaccination. One year after antigen encounter, severely ill subjects maintained high levels of humoral and polyfunctional effector/memory CD4⁺ T cells responses, while mildly ill and vaccinated subjects retained strong cellular immunity, as indicated by high levels of mucosal homing and degranulation markers on IFN-γ⁺ antigen-specific T cells. A principal component analysis distinguished 3 distinct clusters of individuals. The first group comprised vaccinated and mildly ill subjects, while clusters 2 and 3 included mainly infected individuals. Each cluster had immune memory profiles that differed in magnitude and quality. These data provide evidence that there are substantial similarities between the antiinfluenza response that mildly ill and vaccinated individuals develop and that this immune memory signature is different from that seen in severely ill individuals.