A man aged 53 years was admitted to our hospital due to general malaise, fever and chills for the past 24 hours. He had a history of chronic alcoholic liver disease. The blood tests showed leucocytosis with neutrophilia, lactic acidosis and acute-phase reactants. The blood cultures were positive for Parvimonas micra, an anaerobic pathogen which is part of the flora of the oral cavity. There was no evidence of abscess formation in either the examination or the imaging tests, but in the work-up that followed, a gastroscopy showed a stenotic oesophageal mass that turned out to be an invasive squamous cell carcinoma.
Kimura disease is a benign rare chronic inflammatory disorder of unknown aetiology. This disease is mainly endemic in Asia, although cases have also been reported in Europe and America. We describe a case in a 34-year-old Chinese man presenting with severe eosinophilia and multiple lymphadenopathy. Since our initial aim was to rule out the diagnosis of lymphoma, and given the limitations of our laboratory, we decided to perform an excision of one of the cervical lymph nodes. The histological diagnosis was consistent with Kimura disease. We review the epidemiology, the aetiology and clinical features of this entity.
We used machine-learning algorithms to evaluate demographic and clinical data in an administrative data set to identify relevant predictors of mortality due to Listeria monocytogenes infection. We used the Spanish Minimum Basic Data Set at Hospitalization (MBDS-H) to estimate the impacts of several predictors on mortality. The MBDS-H is a mandatory registry of clinical discharge reports. Data were coded with International Classification of Diseases, either Ninth or Tenth Revisions, codes. Diagnoses and clinical conditions were defined using recorded data from these codes or a combination of them. We used two different statistical approaches to produce two predictive models. The first was logistic regression, a classic statistical approach that uses data science to preprocess data and measure performance. The second was a random forest algorithm, a strategy based on machine learning and feature selection. We compared the performance of the two models using predictive accuracy and the area under the curve. Between 2001 and 2016, a total of 5603 hospitalized patients were identified as having any clinical form of listeriosis. Most patients were adults (94.9%). Among all hospitalized individuals, there were 2318 women (41.4%). We recorded 301 pregnant women and 287 newborns with listeriosis. The mortality rate was 0.13 patients per 100,000 population. The performance of the model produced by logistic regression after intense preprocessing was similar to that of the model produced by the random forest algorithm. Predictive accuracy was 0.83, and the area under the receiver operating characteristic curve was 0.74 in both models. Sepsis, age, and malignancy were the most relevant features related to mortality. Our combined use of data science, preprocessing, conventional statistics, and machine learning provides insights into mortality due to Listeria-related infection. These methods are not mutually exclusive. The combined use of several methods would allow researchers to better explain results and understand data related to Listeria monocytogenes infection.
is the oropharyngeal pathogen usually associated with Lemierre's syndrome, a pharyngeal infection which evolves to sepsis, septic emboli and thrombophlebitis of the adjacent neck vessels. It is an uncommon causative bacteria of a liver abscess, and an extensive workup should, therefore, be performed in order to rule out potential sources of the infection. This case report describes the workup that led to the diagnosis of a colorectal carcinoma, which was deemed to be the source of the bacteraemia.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.