Toshio Naito

A comprehensive screening method using machine learning and many factors (biological characteristics, Helicobacter pylori infection status, endoscopic findings and blood test results), accumulated daily as data in hospitals, could improve the accuracy of screening to classify patients at high or low risk of developing gastric cancer. We used XGBoost, a classification method known for achieving numerous winning solutions in data analysis competitions, to capture nonlinear relations among many input variables and outcomes using the boosting approach to machine learning. Longitudinal and comprehensive medical check-up data were collected from 25,942 participants who underwent multiple endoscopies from 2006 to 2017 at a single facility in Japan. The participants were classified into a case group (y = 1) or a control group (y = 0) if gastric cancer was or was not detected, respectively, during a 122-month period. Among 1,431 total participants (89 cases and 1,342 controls), 1,144 (80%) were randomly selected for use in training 10 classification models; the remaining 287 (20%) were used to evaluate the models. The results showed that XGBoost outperformed logistic regression and showed the highest area under the curve value (0.899). Accumulating more data in the facility and performing further analyses including other input variables may help expand the clinical utility.

Detection of dysmorphic cells in peripheral blood (PB) smears is essential in diagnostic screening of hematological diseases. Myelodysplastic syndromes (MDS) are hematopoietic neoplasms characterized by dysplastic and ineffective hematopoiesis, which diagnosis is mainly based on morphological findings of PB and bone marrow. We developed an automated diagnostic support system of MDS by combining an automated blood cell image-recognition system using a deep learning system (DLS) powered by convolutional neural networks (CNNs) with a decision-making system using extreme gradient boosting (XGBoost). The DLS of blood cell image-recognition has been trained using datasets consisting of 695,030 blood cell images taken from 3,261 PB smears including hematopoietic malignancies. The DLS simultaneously classified 17 blood cell types and 97 morphological features of such cells with >93.5% sensitivity and >96.0% specificity. The automated MDS diagnostic system successfully differentiated MDS from aplastic anemia (AA) with high accuracy; 96.2% of sensitivity and 100% of specificity (AUC 0.990). This is the first CNN-based automated initial diagnostic system for MDS using PB smears, which is applicable to develop new automated diagnostic systems for various hematological disorders.

. Here we conducted a genome-wide association study (GWAS) involving 2,393 cases of COVID-19 in a cohort of Japanese individuals collected during the initial waves of the pandemic, with 3,289 unaffected controls. We identified a variant on chromosome 5 at 5q35 (rs60200309-A), close to the dedicator of cytokinesis 2 gene (DOCK2), which was associated with severe COVID-19 in patients less than 65 years of age. This risk allele was prevalent in East Asian individuals but rare in Europeans, highlighting the value of genome-wide association studies in non-European populations. RNA-sequencing analysis of 473 bulk peripheral blood samples identified decreased expression of DOCK2 associated with the risk allele in these younger patients. DOCK2 expression was suppressed in patients with severe cases of COVID-19. Single-cell RNA-sequencing analysis (n = 61 individuals) identified cell-type-specific downregulation of DOCK2 and a COVID-19-specific decreasing effect of the risk allele on DOCK2 expression in non-classical monocytes. Immunohistochemistry of lung specimens from patients with severe COVID-19 pneumonia showed suppressed DOCK2 expression. Moreover, inhibition of DOCK2 function with CPYPP increased the severity of pneumonia in a Syrian hamster model of SARS-CoV-2 infection, characterized by weight loss, lung oedema, enhanced viral loads, impaired macrophage recruitment and dysregulated type I interferon responses. We conclude that DOCK2 has an important role in the host immune response to SARS-CoV-2 infection and the development of severe COVID-19, and could be further explored as a potential biomarker and/or therapeutic target.

BACKGROUND: Time to antibiotic administration is a key element in sepsis care; however, it is difficult to implement sepsis care bundles. Additionally, sepsis is different from other emergent conditions including acute coronary syndrome, stroke, or trauma. We aimed to describe the association between time to antibiotic administration and outcomes in patients with severe sepsis and septic shock in Japan. METHODS: This prospective observational study enrolled 1184 adult patients diagnosed with severe sepsis based on the Sepsis-2 criteria and admitted to 59 intensive care units (ICUs) in Japan between January 1, 2016, and March 31, 2017, as the sepsis cohort of the Focused Outcomes Research in Emergency Care in Acute Respiratory Distress Syndrome, Sepsis and Trauma (FORECAST) study. We compared the characteristics and in-hospital mortality of patients administered with antibiotics at varying durations after sepsis recognition, i.e., 0-60, 61-120, 121-180, 181-240, 241-360, and 361-1440 min, and estimated the impact of antibiotic timing on risk-adjusted in-hospital mortality using the generalized estimating equation model (GEE) with an exchangeable, within-group correlation matrix, with "hospital" as the grouping variable. RESULTS: Data from 1124 patients in 54 hospitals were used for analyses. Of these, 30.5% and 73.9% received antibiotics within 1 h and 3 h, respectively. Overall, the median time to antibiotic administration was 102 min [interquartile range (IQR), 55-189]. Compared with patients diagnosed in the emergency department [90 min (IQR, 48-164 min)], time to antibiotic administration was shortest in patients diagnosed in ICUs [60 min (39-180 min)] and longest in patients transferred from wards [120 min (62-226)]. Overall crude mortality was 23.4%, where patients in the 0-60 min group had the highest mortality (28.0%) and a risk-adjusted mortality rate [28.7% (95% CI 23.3-34.1%)], whereas those in the 61-120 min group had the lowest mortality (20.2%) and risk-adjusted mortality rates [21.6% (95% CI 16.5-26.6%)]. Differences in mortality were noted only between the 0-60 min and 61-120 min groups. CONCLUSIONS: We could not find any association between earlier antibiotic administration and reduction in in-hospital mortality in patients with severe sepsis.