Pneumonia is a lung infection that causes 15% of childhood mortality (under 5 years old), over 800,000 children under five every year, around 2,200 every day, all over the world. This pathology is ...mainly caused by viruses or bacteria. X-rays imaging analysis is one of the most used methods for pneumonia diagnosis. These clinical images can be analyzed using machine learning methods such as convolutional neural networks (CNN), which learn to extract critical features for the classification. However, the usability of these systems is limited in medicine due to the lack of interpretability, because of these models cannot be used to generate an understandable explanation (from a human-based perspective), about how they have reached those results. Another problem that difficults the impact of this technology is the limited amount of labeled data in many medicine domains. The main contributions of this work are two fold: the first one is the design of a new explainable artificial intelligence (XAI) technique based on combining the individual heatmaps obtained from each model in the ensemble. This allows to overcome the explainability and interpretability problems of the CNN “black boxes”, highlighting those areas of the image which are more relevant to generate the classification. The second one is the development of new ensemble deep learning models to classify chest X-rays that allow highly competitive results using small datasets for training. We tested our ensemble model using a small dataset of pediatric X-rays (950 samples of children between one month and 16 years old) with low quality and anatomical variability (which represents one of the biggest challenges addressed in this work). We also tested other strategies such as single CNNs trained from scratch and transfer learning using CheXNet. Our results show that our ensemble model clearly outperforms these strategies obtaining highly competitive results. Finally we confirmed the robustness of our approach using another pneumonia diagnosis dataset (Kermany et al., 2018).
•New CNN models designed, with pediatric X-rays to maximize the AUC, and the TPR.•The design of new AI explainable algorithms, based on heatmaps, applied to medicine.•The use of a CNN and XAI approaches for detecting Pneumonia in child.•The models were validated by comparing them with a state of the art model and dataset.
Logistic regression (LR) is the most common prediction model in medicine. In recent years, supervised machine learning (ML) methods have gained popularity. However, there are many concerns about ML ...utility for small sample sizes. In this study, we aim to compare the performance of 7 algorithms in the prediction of 1-year mortality and clinical progression to AIDS in a small cohort of infants living with HIV from South Africa and Mozambique. The data set (n = 100) was randomly split into 70% training and 30% validation set. Seven algorithms (LR, Random Forest (RF), Support Vector Machine (SVM), K-Nearest Neighbor (KNN), Naïve Bayes (NB), Artificial Neural Network (ANN), and Elastic Net) were compared. The variables included as predictors were the same across the models including sociodemographic, virologic, immunologic, and maternal status features. For each of the models, a parameter tuning was performed to select the best-performing hyperparameters using 5 times repeated 10-fold cross-validation. A confusion-matrix was built to assess their accuracy, sensitivity, and specificity. RF ranked as the best algorithm in terms of accuracy (82,8%), sensitivity (78%), and AUC (0,73). Regarding specificity and sensitivity, RF showed better performance than the other algorithms in the external validation and the highest AUC. LR showed lower performance compared with RF, SVM, or KNN. The outcome of children living with perinatally acquired HIV can be predicted with considerable accuracy using ML algorithms. Better models would benefit less specialized staff in limited resources countries to improve prompt referral in case of high-risk clinical progression.
Celotno besedilo
Dostopno za:
DOBA, IZUM, KILJ, NUK, PILJ, PNG, SAZU, SIK, UILJ, UKNU, UL, UM, UPUK
Objectives
To describe the etiology of community‐acquired pneumonia (CAP) in hospitalized children in Spain and analyze the predictors of the etiology.
Hypothesis
The different etiological groups of ...pediatric CAP are associated with different clinical, radiographic, and analytical data.
Design
Observational, multicenter, and prospective study.
Patient selection
This study included children aged 1 month to 17 years with CAP, who were hospitalized between April 2012 and May 2019.
Methods
An extensive microbiological workup was performed. The clinical, radiographic, and analytical parameters were analyzed for three etiological groups.
Results
Among the 495 children included, at least one causative pathogen was identified in 262 (52.9%): pathogenic viruses in 155/262 (59.2%); atypical bacteria (AB), mainly Mycoplasma pneumonia, in 84/262 (32.1%); and typical bacteria (TyB) in 40/262 (15.3%). Consolidation was observed in 89/138 (64.5%) patients with viral CAP, 74/84 (88.1%) with CAP caused by AB, and 40/40 (100%) with CAP caused by TyB. Para‐pneumonic pleural effusion (PPE) was observed in 112/495 (22.6%) patients, of which 61/112 (54.5%) presented a likely causative pathogen: viruses in 12/61 (19.7%); AB in 23/61 (37.7%); and TyB in 26/61 (42.6%). Viral etiology was significantly frequent in young patients and in those with low oxygen saturation, wheezing, no consolidation, and high lymphocyte counts. CAP patients with AB as the etiological agent had a significantly longer and less serious course as compared to those with other causative pathogens.
Conclusions
Viruses and M. pneumoniae are the main causes of pediatric CAP in Spain. Wheezing, young age, and no consolidation on radiographs are indicative of viral etiology. Viruses and AB can also cause PPE. Since only a few cases can be directly attributed to TyB, the indications for antibiotics must be carefully considered in each patient.
Background
The chemokine CC motif ligand 1 (CCL1) participates in immune cell recruitment and, as other chemokines, is also involved in nociceptive processing. In contrast with previous reports ...indicating its participation in allodynia and cold hypernociception when spinally administered, its ability to evoke heat thermal analgesia, mediated by circulating leukocytes and endocannabinoids, after systemic administration has recently been reported.
Objectives
Aiming to explore the role played by CCL1 on spinal nociception, we study here the effect of its intrathecal administration on thermal nociception in mice.
Methods
Behavioral nociceptive assays, immunohistochemical experiments, white cell blood depletion procedures and qRT‐PCR experiments were performed.
Results
The intrathecal administration of CCL1 (0.3–30 ng) produced analgesia as measured by the unilateral hot plate test. This effect peaked 1 h after injection, was prevented by the CCR8 antagonist R243 and was accompanied by a reduction of c‐Fos expression in spinal neurons. Whereas blood leukocyte depletion did not modify it, analgesia was abolished by the microglial inhibitor minocycline, but not the astroglial inhibitor aminoadipate. Furthermore, antinociception remained unmodified by the coadministration of cannabinoid type 1 or 2 receptors antagonists. However, it was reversed by naloxone but not by selective blockade of mu‐ or delta‐opioid receptors. The inhibitory effect induced by the selective kappa‐opioid receptor antagonist, nor‐binaltorphimine, and by an anti‐dynorphin A 1‐17 antibody indicates that analgesia evoked by spinal CCL1 is mediated by endogenous dynorphins acting on kappa‐opioid receptors.
Conclusions
Endogenous dynorphin and microglia behave as key players in heat thermal analgesia evoked by spinal CCL1 in mice.
Introduction
HIV infection causes pathological changes in the natural killer (NK) cell compartment that can be only partially restored by antiretroviral therapy (ART). We investigated NK cells ...phenotype and function in children with perinatally acquired HIV (PHIV) and long‐term viral control (five years) due to effective ART in a multicentre cross‐sectional European study (CARMA, EPIICAL consortium). The impact of age at ART start and viral reservoir was also evaluated.
Methods
Peripheral blood mononuclear cells (PBMCs) from 40 PHIV who started ART within two years of life (early treated patients (ET), ≤6 months; late treated patients (LT), > 6 months), with at least five years of HIV‐1 suppression (<40 HIV copies/mL), were collected between November 2017 and August 2018. NK phenotype and function were analysed by flow cytometry and transcriptional profile of PBMCs by RNA‐Seq. HIV‐1 DNA was measured by real‐time polymerase chain reaction (Data were analysed by Spearman correlation plots and multivariable Poisson regression model (adjusted for baseline %CD4 and RNA HIV viral load and for age at ART start as an interaction term, either ET or LT) to explore the association between NK cell parameters and HIV reservoir modulated by age at ART start.
Results
A significantly higher frequency of CD56neg NK cells was found in LT compared with ET. We further found in LT a positive correlation of CD56neg NK cells with HIV‐1 DNA. LT also displayed increased expression of the NKG2D and NKp46 activating receptors and perforin compared with ET. Moreover, CD107a+ and IFN‐γ+ frequencies in non‐stimulated NK were associated with HIV‐1 DNA in LT patients. Finally, RNA‐Seq analysis showed in LT an up‐regulation of genes related to NK‐activating pathways and susceptibility to apoptosis compared with ET.
Conclusions
We show that early initiation of ART during infancy preserves the NK compartment and is associated with lower HIV‐1 reservoir. Such condition persists over adolescence due to long‐term viral control achieved through effective ART.
The size of the latent HIV reservoir is associated with the timing of therapeutic interventions and overall health of the immune system. Here, we demonstrate that T cell phenotypic signatures ...associate with viral reservoir size in a cohort of HIV vertically infected children and young adults under durable viral control, and who initiated anti-retroviral therapy (ART) <2 years old. Flow cytometry was used to measure expression of immune activation (IA), immune checkpoint (ICP) markers, and intracellular cytokine production after stimulation with GAG peptides in CD4 and CD8 T cells from cross-sectional peripheral blood samples. We also evaluated the expression of 96 genes in sort-purified total CD4 and CD8 T cells along with HIV-specific CD4 and CD8 T cells using a multiplexed RT-PCR approach. As a measure of HIV reservoir, total HIV-DNA quantification by real-time PCR was performed. Poisson regression modeling for predicting reservoir size using phenotypic markers revealed a signature that featured frequencies of PD-1+CD4 T cells, TIGIT+CD4 T cells and HIV-specific (CD40L+) CD4 T cells as important predictors and it also shows that time of ART initiation strongly affects their association with HIV-DNA. Further, gene expression analysis showed that the frequencies of PD-1+CD4 T cells associated with a CD4 T cell molecular profile skewed toward an exhausted Th1 profile. Our data provide a link between immune checkpoint molecules and HIV persistence in a pediatric cohort as has been demonstrated in adults. Frequencies of PD-1+ and TIGIT+CD4 T cells along with the frequency of HIV-specific CD4 T cells could be associated with the mechanism of viral persistence and may provide insight into potential targets for therapeutic intervention.
Celotno besedilo
Dostopno za:
DOBA, IZUM, KILJ, NUK, PILJ, PNG, SAZU, SIK, UILJ, UKNU, UL, UM, UPUK
Hydrocephalic hyh mutant mice undergo a programmed loss of the neuroepithelium/ependyma followed by a reaction of periventricular astrocytes, which form a new cell layer covering the denuded ...ventricular surface. We present a comparative morphological and functional study of the newly formed layer of astrocytes and the multiciliated ependyma of hyh mice. Transmission electron microscopy, immunocytochemistry for junction proteins (N-cadherin, connexin 43) and proteins involved in permeability (aquaporin 4) and endocytosis (caveolin-1, EEA1) were used. Horseradish peroxidase (HRP) and lanthanum nitrate were used to trace the intracellular and paracellular transport routes. The astrocyte layer shares several cytological features with the normal multiciliated ependyma, such as numerous microvilli projected into the ventricle, extensive cell–cell interdigitations and connexin 43-based gap junctions, suggesting that these astrocytes are coupled to play an unknown function as a cell layer. The ependyma and the astrocyte layers also share transport properties: (1) high expression of aquaporin 4, caveolin-1 and the endosome marker EEA1; (2) internalization into endocytic vesicles and early endosomes of HRP injected into the ventricle; (3) and a similar paracellular route of molecules moving between CSF, the subependymal neuropile and the pericapillary space, as shown by lanthanum nitrate and HRP. A parallel analysis performed in human hydrocephalic foetuses indicated that a similar phenomenon would occur in humans. We suggest that in foetal-onset hydrocephalus, the astrocyte assembly at the denuded ventricular walls functions as a CSF–brain barrier involved in water and solute transport, thus contributing to re-establish lost functions at the brain parenchyma–CSF interphase.
Long-COVID is a new emerging syndrome worldwide that is characterized by the persistence of unresolved signs and symptoms of COVID-19 more than 4 weeks after the infection and even after more than 12 ...weeks. The underlying mechanisms for Long-COVID are still undefined, but a sustained inflammatory response caused by the persistence of SARS-CoV-2 in organ and tissue sanctuaries or resemblance with an autoimmune disease are within the most considered hypotheses. In this study, we analyzed the usefulness of several demographic, clinical, and immunological parameters as diagnostic biomarkers of Long-COVID in one cohort of Spanish individuals who presented signs and symptoms of this syndrome after 49 weeks post-infection, in comparison with individuals who recovered completely in the first 12 weeks after the infection. We determined that individuals with Long-COVID showed significantly increased levels of functional memory cells with high antiviral cytotoxic activity such as CD8
TEMRA cells, CD8
TCRγδ
cells, and NK cells with CD56
CD57
NKG2C
phenotype. The persistence of these long-lasting cytotoxic populations was supported by enhanced levels of CD4
Tregs and the expression of the exhaustion marker PD-1 on the surface of CD3
T lymphocytes. With the use of these immune parameters and significant clinical features such as lethargy, pleuritic chest pain, and dermatological injuries, as well as demographic factors such as female gender and O
blood type, a Random Forest algorithm predicted the assignment of the participants in the Long-COVID group with 100% accuracy. The definition of the most accurate diagnostic biomarkers could be helpful to detect the development of Long-COVID and to improve the clinical management of these patients.
Objective
Late presenters (LP) for HIV care are associated with higher morbidity and mortality rates. Our aim was to describe the characteristics associated with LP among adolescents in Spain. ...Identification of particular features may help in the design of strategies for improvement.
Methods
Late‐presenting adolescents diagnosed at 12–19 years of age and enrolled in the Spanish paediatric and adult HIV/AIDS cohorts (CoRIS‐CoRISpe) from 2004 to 2019 were selected. LP were defined as those presenting with CD4 count <350 cells/mm3 or an AIDS‐defining event in the 6 months following HIV diagnosis. Confirmed low CD4 count in the next 3 months and before antiretroviral treatment initiation defined confirmed LP (cLP).
Results
Of 410 adolescents newly diagnosed with HIV, 303 (73.9%) had available data for assessing late presentation. Of these, 34.7% were LP and 23.7% were cLP. The median CD4 count for cLP was 235 cells/mm3 (interquartile range 122–285). In a multivariable analysis, adolescents at the highest risk of late presentation were early adolescents (age 12–14 years; odds ratio OR 6.50; 95% confidence interval CI 2.61–18.2), middle adolescents (age 15–17 years; OR 1.85; 95% CI 0.92–3.59), and adolescents born abroad (OR 1.71; 95% CI 0.97–3.00), particularly those of African origin (OR 3.08; 95% CI 1.38–6.79).
Conclusions
One‐quarter of adolescents presented late for HIV care in Spain. Early adolescents, middle adolescents, and those born abroad presented a sevenfold, twofold, and twofold higher risk of being cLP, respectively. Enhancing the awareness of HIV risk and the access to care, especially for younger and foreign adolescents, could help reduce late presentation and tackle the adolescent HIV epidemic.
PDF
