Plasma calprotectin in the emergency department: a potential clinical biomarker for patients with infectious diseases.

Increased levels of plasma calprotectin are reported in patients with infectious diseases. However, the clinical usefulness of calprotectin as a biomarker to identify patients with infectious diseases in the emergency department (ED) setting has not been investigated. To study the ability of calprotectin to discriminate patients with acute infectious diseases and dyspnea from patients with other causes of acute dyspnea in the ED setting. Patients aged ≥18 years seeking ED during daytime on weekdays between March 2013 and July 2018, with acute dyspnea, were included. Participants (n = 1287) were triaged according to Medical Emergency Triage and Treatment System-Adult score (METTS-A) or Rapid Emergency Triage and Treatment System (RETTS), and blood samples were collected. The association between calprotectin and other markers of infectious diseases, i.e. biomarkers (CRP, leucocytes) and body temperature, was studied. The predictive value of calprotectin for the outcome of acute infection was evaluated with receiver operating characteristic (ROC) analysis. Univariate cross-sectional regression showed significant associations between calprotectin and leucocytes, CRP and body temperature. Patients with severe infections including pneumonia (n = 119) had significantly higher concentrations of calprotectin compared to patients with heart failure (n = 162) or chronic obstructive pulmonary disease (n = 183). When tested for the outcome of acute infection (n = 109), the area under the ROC curve (AUROC) was for CRP 0.83 and for calprotectin 0.78. Plasma calprotectin identifies infectious diseases in ED patients with acute dyspnea, and the clinical usefulness of Calprotectin in the ED has to be further studied.

Multi-stream LSTM-HMM decoding and histogram equalization for noise robust keyword spotting.

Highly spontaneous, conversational, and potentially emotional and noisy speech is known to be a challenge for today's automatic speech recognition (ASR) systems, which highlights the need for advanced algorithms that improve speech features and models. Histogram Equalization is an efficient method to reduce the mismatch between clean and noisy conditions by normalizing all moments of the probability distribution of the feature vector components. In this article, we propose to combine histogram equalization and multi-condition training for robust keyword detection in noisy speech. To better cope with conversational speaking styles, we show how contextual information can be effectively exploited in a multi-stream ASR framework that dynamically models context-sensitive phoneme estimates generated by a long short-term memory neural network. The proposed techniques are evaluated on the SEMAINE database-a corpus containing emotionally colored conversations with a cognitive system for "Sensitive Artificial Listening".