An efficient deep learning-based framework for tuberculosis detection using chest X-ray images.

Early diagnosis of tuberculosis (TB) is an essential and challenging task to prevent disease, decrease mortality risk, and stop transmission to other people. The chest X-ray (CXR) is the top choice for lung disease screening in clinics because it is cost-effective and easily accessible in most countries. However, manual screening of CXR images is a heavy burden for radiologists, resulting in a high rate of inter-observer variances. Hence, proposing a cost-effective and accurate computer aided diagnosis (CAD) system for TB diagnosis is challenging for researchers. In this research, we proposed an efficient and straightforward deep learning network called TBXNet, which can accurately classify a large number of TB CXR images. The network is based on five dual convolutions blocks with varying filter sizes of 32, 64, 128, 256 and 512, respectively. The dual convolution blocks are fused with a pre-trained layer in the fusion layer of the network. In addition, the pre-trained layer is utilized for transferring pre-trained knowledge into the fusion layer. The proposed TBXNet has achieved an accuracy of 98.98%, and 99.17% on Dataset A and Dataset B, respectively. Furthermore, the generalizability of the proposed work is validated against Dataset C, which is based on normal, tuberculous, pneumonia, and COVID-19 CXR images. The TBXNet has obtained the highest results in Precision (95.67%), Recall (95.10%), F1-score (95.38%), and Accuracy (95.10%), which is comparatively better than all other state-of-the-art methods.

Higher admission activated partial thromboplastin time, neutrophil-lymphocyte ratio, serum sodium, and anticoagulant use predict in-hospital COVID-19 mortality in people with Diabetes: Findings from Two University Hospitals in the U.K.

AIMS: To create and compare survival models from admission laboratory indices in people hospitalized with coronavirus disease 2019 (COVID-19) with and without diabetes. METHODS: Retrospective observational study of patients with COVID-19 with or without diabetes admitted to Sheffield Teaching Hospitals from 29 February to 01 May 2020. Predictive variables for in-hospital mortality from COVID-19 were explored using Cox proportional hazard models. RESULTS: Out of 505 patients, 156 (30.8%) had diabetes mellitus (DM) of which 143 (91.7%) had type 2 diabetes. There were significantly higher in-hospital COVID-19 deaths in those with DM [DM COVID-19 deaths 54 (34.6%) vs. non-DM COVID-19 deaths 88 (25.2%): P < 0.05]. Activated partial thromboplastin time (APPT) > 24 s without anticoagulants (HR 6.38, 95% CI: 1.07-37.87: P = 0.04), APTT > 24 s with anticoagulants (HR 24.01, 95% CI: 3.63-159.01: P < 0.001), neutrophil-lymphocyte ratio > 8 (HR 6.18, 95% CI: 2.36-16.16: P < 0.001), and sodium > 136 mmol/L (HR 3.27, 95% CI: 1.12-9.56: P = 0.03) at admission, were only associated with in-hospital COVID-19 mortality for those with diabetes. CONCLUSIONS: At admission, elevated APTT with or without anticoagulants, neutrophil-lymphocyte ratio and serum sodium are unique factors that predict in-hospital COVID-19 mortality in patients with diabetes compared to those without. This novel finding may lead to research into haematological and biochemical mechanisms to understand why those with diabetes are more susceptible to poor outcomes when infected with Covid-19, and contribute to identification of those most at risk when admitted to hospital.

A Bittersweet Response to Infection in Diabetes; Targeting Neutrophils to Modify Inflammation and Improve Host Immunity.

Chronic and recurrent infections occur commonly in both type 1 and type 2 diabetes (T1D, T2D) and increase patient morbidity and mortality. Neutrophils are professional phagocytes of the innate immune system that are critical in pathogen handling. Neutrophil responses to infection are dysregulated in diabetes, predominantly mediated by persistent hyperglycaemia; the chief biochemical abnormality in T1D and T2D. Therapeutically enhancing host immunity in diabetes to improve infection resolution is an expanding area of research. Individuals with diabetes are also at an increased risk of severe coronavirus disease 2019 (COVID-19), highlighting the need for re-invigorated and urgent focus on this field. The aim of this review is to explore the breadth of previous literature investigating neutrophil function in both T1D and T2D, in order to understand the complex neutrophil phenotype present in this disease and also to focus on the development of new therapies to improve aberrant neutrophil function in diabetes. Existing literature illustrates a dual neutrophil dysfunction in diabetes. Key pathogen handling mechanisms of neutrophil recruitment, chemotaxis, phagocytosis and intracellular reactive oxygen species (ROS) production are decreased in diabetes, weakening the immune response to infection. However, pro-inflammatory neutrophil pathways, mainly neutrophil extracellular trap (NET) formation, extracellular ROS generation and pro-inflammatory cytokine generation, are significantly upregulated, causing damage to the host and perpetuating inflammation. Reducing these proinflammatory outputs therapeutically is emerging as a credible strategy to improve infection resolution in diabetes, and also more recently COVID-19. Future research needs to drive forward the exploration of novel treatments to improve infection resolution in T1D and T2D to improve patient morbidity and mortality.

A UK nationwide study of people with type 1 diabetes admitted to hospital with COVID-19 infection.

AIMS/HYPOTHESIS: The aim of this work was to describe the clinical characteristics of adults with type 1 diabetes admitted to hospital and the risk factors associated with severe coronavirus disease-2019 (COVID-19) in the UK. METHODS: A retrospective cohort study was performed using data collected through a nationwide audit of people admitted to hospital with diabetes and COVID-19, conducted by the Association of British Clinical Diabetologists from March to October 2020. Prespecified demographic, clinical, medication and laboratory data were collected from the electronic and paper medical record systems of the participating hospitals by local clinicians. The primary outcome of the study, severe COVID-19, was defined as death in hospital and/or admission to the adult intensive care unit (AICU). Logistic regression models were used to generate age-adjusted ORs. RESULTS: Forty UK centres submitted data. The final dataset included 196 adults who were admitted to hospital and had both type 1 diabetes and COVID-19 on admission (male sex 55%, white 70%, with mean [SD] age 62 [19] years, BMI 28.3 [7.3] kg/m2 and last recorded HbA1c 76 [31] mmol/mol [9.1 (5.0)%]). The prevalence of pre-existing microvascular disease and macrovascular disease was 56% and 39%, respectively. The prevalence of diabetic ketoacidosis on admission was 29%. A total of 68 patients (35%) died or were admitted to AICU. The proportions of people that died were 7%, 38% and 38% of those aged <55, 55-74 and ≥75 years, respectively. BMI, serum creatinine levels and having one or more microvascular complications were positively associated with the primary outcome after adjusting for age. CONCLUSIONS/INTERPRETATION: In people with type 1 diabetes and COVID-19 who were admitted to hospital in the UK, higher BMI, poorer renal function and presence of microvascular complications were associated with greater risk of death and/or admission to AICU. Risk of severe COVID-19 is reassuringly very low in people with type 1 diabetes who are under 55 years of age without microvascular or macrovascular disease. IN PEOPLE WITH TYPE 1 DIABETES AND COVID-19 ADMITTED TO HOSPITAL IN THE UK, BMI AND ONE OR MORE MICROVASCULAR COMPLICATIONS HAD A POSITIVE ASSOCIATION AND LOW SERUM CREATINE LEVELS HAD A NEGATIVE ASSOCIATION WITH DEATH/ADMISSION TO INTENSIVE CARE UNIT AFTER ADJUSTING FOR AGE.