O2 supplementation disambiguation in clinical narratives to support retrospective COVID-19 studies.

BACKGROUND: Oxygen saturation, a key indicator of COVID-19 severity, poses challenges, especially in cases of silent hypoxemia. Electronic health records (EHRs) often contain supplemental oxygen information within clinical narratives. Streamlining patient identification based on oxygen levels is crucial for COVID-19 research, underscoring the need for automated classifiers in discharge summaries to ease the manual review burden on physicians. METHOD: We analysed text lines extracted from anonymised COVID-19 patient discharge summaries in German to perform a binary classification task, differentiating patients who received oxygen supplementation and those who did not. Various machine learning (ML) algorithms, including classical ML to deep learning (DL) models, were compared. Classifier decisions were explained using Local Interpretable Model-agnostic Explanations (LIME), which visualize the model decisions. RESULT: Classical ML to DL models achieved comparable performance in classification, with an F-measure varying between 0.942 and 0.955, whereas the classical ML approaches were faster. Visualisation of embedding representation of input data reveals notable variations in the encoding patterns between classic and DL encoders. Furthermore, LIME explanations provide insights into the most relevant features at token level that contribute to these observed differences. CONCLUSION: Despite a general tendency towards deep learning, these use cases show that classical approaches yield comparable results at lower computational cost. Model prediction explanations using LIME in textual and visual layouts provided a qualitative explanation for the model performance.

Primary sarcopenia is associated with elevated spontaneous NET formation.

Introduction: Sarcopenia is a frequent complication of liver cirrhosis, but it can also occur independently as a result of any underlying cause. The immune system plays an important role in the pathogenesis of both liver cirrhosis and sarcopenia. Neutrophil function, including neutrophil extracellular trap (NET) formation, is linked to chronic inflammation; however, it has not been extensively studied in patients with sarcopenia. Here, we aim to study if main neutrophil functions, such as phagocytosis, reactive oxygen species (ROS) production, and NET formation, are altered in patients with sarcopenia with or without liver cirrhosis. Methods: Neutrophils from 92 patients (52 patients with liver cirrhosis and sarcopenia, 25 patients with liver cirrhosis without sarcopenia, and 15 patients with sarcopenia without liver cirrhosis) and 10 healthy controls were isolated and stimulated with heat-inactivated E. coli (250 bacteria/cell), phorbol 12-myristate 13-acetate (PMA) (100 nM), or incubation medium in duplicates for 2 h at 37°C. Cells were fixed with paraformaldehyde and stained with 4',6-diamidino-2-phenylindole (DAPI). Pictures of 10 random fields of vision per slide were taken with an Olympus BX51 fluorescence microscope (Olympus, Shinjuku, Tokyo, Japan) at 600x total magnification. The DNA Area and NETosis Analysis (DANA) algorithm was used to quantify the percentage of NET formation per patient. Phagocytosis and ROS production were assessed with the PhagotestTM kit and PhagoburstTM kit (Glycotope, Heidelberg, Germany) in 92 patients and 21 healthy controls, respectively. Results: Spontaneous NET formation was significantly elevated in patients with only sarcopenia compared to patients with cirrhosis and sarcopenia (p = 0.008) and healthy controls (p = 0.039). NET formation in response to PMA was significantly decreased in patients with cirrhosis (p = 0.007), cirrhosis and sarcopenia (p < 0.001), and sarcopenia (p = 0.002) compared to healthy controls. There was no significant difference in NET formation in response to E. coli between the groups. The DANA algorithm was successfully optimized and validated for assessment of clinical samples. There were no significant changes in neutrophil phagocytosis between patients' groups compared to healthy controls. A significantly lower percentage of neutrophils produced ROS in response to N-formylmethionine-leucyl-phenylalanine (fMLF) in patients compared to healthy controls. Discussion: Spontaneous NET formation might contribute to chronic inflammation and sarcopenia pathogenesis. This, however, does not result in the impairment of the NET formation function of neutrophils in response to a bacterial stimulus and, therefore, cannot be not linked with the increased risk of bacterial infections neither in sarcopenia nor in cirrhosis. The semi-automated NET formation analysis can be successfully implemented to analyze the vast amount of data generated within clinical studies. This approach opens up the possibilities to develop an NET formation-based biomarker in different diseases including sarcopenia and implement NET formation analysis into clinical settings. Phagocytosis and ROS production were not affected in patients with sarcopenia. Further research is needed to explore the mechanism of NET formation in patients with sarcopenia and its potential as a biomarker in sarcopenia.

Corrigendum: Application of a multispecies probiotic reduces gastro-intestinal discomfort and induces microbial changes after colonoscopy.

[This corrects the article DOI: 10.3389/fonc.2022.1078315.].