Highly accurate detection of the intracranial hemorrhage without delay is a critical clinical issue for the diagnostic decision and treatment in an emergency room. In the context of a study on diagnostic accuracy, there is a tradeoff between sensitivity and specificity. In order to improve sensitivity while preserving specificity, we propose a cascade deep learning model constructed using two convolutional neural networks (CNNs) and dual fully convolutional networks (FCNs). The cascade CNN model is built for identifying bleeding; hereafter the dual FCN is to detect five different subtypes of intracranial hemorrhage and to delineate their lesions. Using a total of 135,974 CT images including 33,391 images labeled as bleeding, each of CNN/FCN models was trained separately on image data preprocessed by two different settings of window level/width. One is a default window (50/100[level/width]) and the other is a stroke window setting (40/40). By combining them, we obtained a better outcome on both binary classification and segmentation of hemorrhagic lesions compared to a single CNN and FCN model. In determining whether it is bleeding or not, there was around 1% improvement in sensitivity (97.91% [± 0.47]) while retaining specificity (98.76% [± 0.10]). For delineation of bleeding lesions, we obtained overall segmentation performance at 80.19% in precision and 82.15% in recall which is 3.44% improvement compared to using a single FCN model.

No abstract

Introduction: We experienced an emergency room (ER) shutdown related to an accidental exposure to a patient with COVID-19 who had not been isolated. To prevent in-hospital transmission of the disease, we subsequently isolated patients with suspected symptoms, relevant radiographic findings, or epidemiology. The SARS-CoV-2 reverse-transcriptase polymerase chain reaction was performed in most patients requiring hospitalization. Universal mask policy and comprehensive use of personal protective equipment (PPE) were implemented. Materials and Methods: The effect of the interventions was analyzed in a 635-bed, tertiary-care hospital in Daegu, South Korea. Results: From the pre-shutdown (February 10 to 25) to the post-shutdown (February 28 to March 16, 2020) period, the mean hourly turnaround time (23:31±6:43 versus 9:27±3:41, P<0.001) was shortened, which increased the proportions of the patients tested (5.8% [N=1,037] versus 64.6% [N=690], P<0.001) and the average number of tests per day (3.8±4.3 versus 24.7±5.0, P<0.001). All 23 patients with COVID-19 in the post-shutdown period were isolated in the ER without any problematic accidental exposure or in-hospital transmission. From the pre- to the post-shutdown period, the median duration of stay in the ER among hospitalized patients (4:30 hr [2:17-9:48] versus 14:33 hr (6:55-24:50), P<0.001), rates of intensive care unit admissions (1.4% versus 2.9%, P=0.023), and mortality (0.9% versus 3.0%, P=0.001) increased. Conclusions: Problematic accidental exposure and in-hospital transmission of COVID-19 can be successfully prevented through active isolation and surveillance policy and comprehensive PPE use despite longer ER stays and the presence of more severely ill patients during an explosive COVID-19 outbreak.

PurposeThe aim of this study is to investigate the changing prevalence rate of Helicobacter pylori infection in children, of different age groups, with recurrent abdominal pain over a 10-year period.MethodsChildren with recurrent abdominal pain who visited the pediatric outpatient clinic at university hospital were screened for H. pylori. Children were divided into 3 age categories of 4-5, 6-11, and 12-16 years. To study the changes in the annual prevalence rates of H. pylori infection, the study period was divided into 3 time periods: 2004-2007, 2008-2010, and 2011-2014. Urea breath test was performed for all children aged 4-16 years, with a cut-off value of 4.0‰ for children aged ≥6 years and 7‰ for children aged <6 years.ResultsA total of 2,530 children (1,191 boys) with a mean age of 10.0±3.0 years (range, 4.0-16.9 years) were included in the study. The total prevalence of H. pylori infection was 7.4% (187/2,530). The prevalence rate of H. pylori infection in children with recurrent abdominal pain was 8.0% (70/873) in 2004-2007, 7.7% (51/666) in 2008-2010, and 6.7% (66/991) in the 2011-2014. Nevertheless, a significant difference was observed in the prevalence rate between children <12 years old and ≥12 years of age (p=0.018).ConclusionThe prevalence of H. pylori infection in Korean children with recurrent abdominal pain was 7.4%, showing no significant decrease in the last 11 years; however, the prevalence rate in children <12 years old was significantly lower than that in those ≥12 years old.

Intussusception is rarely observed in adults. Adult cases account for only 5% of all cases of intussusceptions and almost 1%-5% of bowel obstruction cases. The etiology, presentation and management of intussusception in adults are different from those in children. The clinical presentation in adults often includes nonspecific signs and symptoms, thereby complicating differential diagnosis from other causes of abdominal pain. We report a 29-year-old Asian woman who visited our emergency department with complaints of fever associated with epigastric pain since one day. Abdominal computed tomography demonstrated ileocolic intussusception, and laparoscopic small bowel luminal mass resection was performed. Histopathology report confirmed a 3.5 cm × 2.7 cm submucosal lipoma in the terminal ileum. Sufficient vigilance and appropriate investigations are important for prompt diagnosis and surgical referral of patients to enable favorable outcomes. A computed tomography scan can be a helpful modality in establishing a diagnosis.