Nationwide trends and predictors of inpatient mortality in 83884 transjugular intrahepatic portosystemic shunt

AIM: To evaluate and validate the national trends and predictors of in-patient mortality of transjugular intrahepatic portosystemic shunt (TIPS) in 15 years.METHODS: Using the National Inpatient Sample which is a part of Health Cost and Utilization Project, we identified a discharge-weighted national estimate of 83884 TIPS procedures performed in the United States from 1998 to 2012 using international classification of diseases-9 procedural code 39.1. The demographic, hospital and co-morbility data were analyzed using a multivariant analysis. Using multi-nominal logistic regression analysis, we determined predictive factors related to increases in-hospital mortality. Comorbidity measures are in accordance to the Comorbidity Software designed by the Agency for Healthcare Research and Quality.RESULTS: Overall, 12.3% of patients died during hospitalization with downward trend in-hospital mortality with the mean length of stay of 10.8 ± 13.1 d. Notable, African American patients (OR = 1.809 vs Caucasian patients, P < 0.001), transferred patients (OR = 1.347 vs non-transferred, P < 0.001), emergency admissions (OR = 3.032 vs elective cases, P < 0.001), patients in the Northeast region (OR = 1.449 vs West, P < 0.001) had significantly higher odds of in-hospital mortality. Number of diagnoses and number of procedures showed positive correlations with in-hospital death (OR = 1.249 per one increase in number of procedures). Patients diagnosed with acute respiratory failure (OR = 8.246), acute kidney failure (OR = 4.359), hepatic encephalopathy (OR = 2.217) and esophageal variceal bleeding (OR = 2.187) were at considerably higher odds of in-hospital death compared with ascites (OR = 0.136, P < 0.001). Comorbidity measures with the highest odds of in-hospital death were fluid and electrolyte disorders (OR = 2.823), coagulopathy (OR = 2.016), and lymphoma (OR = 1.842).CONCLUSION: The overall mortality of the TIPS procedure is steadily decreasing, though the length of stay has remained relatively constant. Specific patient ethnicity, location, transfer status, primary diagnosis and comorbidities correlate with increased odds of TIPS in-hospital death.

Pushing Multiple Convertible Constrains into Frequent Itemsets Mining

Constraint pushing techniques have been developed for mining frequent patterns and association rules. How ever, multiple constraints cannot be handled with existing techniques in frequent pattern mining. In this paper, a new algorithm MCFMC （mining complete set of frequent itemsets with multiple constraints） is introduced. The algorithm takes advantage of the fact that a convertible constraint can be pushed into mining algorithm to reduce mining research spaces. By using a sample database, the algorithm develops techniques which select an optimal method based on a sample database to convert multiple constraints into multiple convert ible constraints, disjoined by conjunction and/or, and then partition these constraints into two parts. One part is pushed deep inside the mining process to reduce the research spaces for frequent itemsets, the other part that cannot be pushed in algorithm is used to filter the complete set of frequent itemsets and get the final result. Results from our detailed experi ment show the feasibility and effectiveness of the algorithm.

LuoJiaAI:A cloud-based artificial intelligence platform for remote sensing image interpretation

The rapid processing,analysis,and mining of remote-sensing big data based on intelligent interpretation technology using remote-sensing cloud computing platforms(RS-CCPs)have recently become a new trend.The existing RS-CCPs mainly focus on developing and optimizing high-performance data storage and intelligent computing for common visual representation,which ignores remote sensing data characteristics such as large image size,large-scale change,multiple data channels,and geographic knowledge embedding,thus impairing computational efficiency and accuracy.We construct a LuoJiaAI platform composed of a standard large-scale sample database(LuoJiaSET)and a dedicated deep learning framework(LuoJiaNET)to achieve state-of-the-art performance on five typical remote sensing interpretation tasks,including scene classification,object detection,land-use classification,change detection,and multi-view 3D reconstruction.The details of the LuoJiaAI application experiment can be found at the white paper for LuoJiaAI industrial application.In addition,LuoJiaAI is an open-source RS-CCP that supports the latest Open Geospatial Consortium(OGC)standards for better developing and sharing Earth Artificial Intelligence(AI)algorithms and products on benchmark datasets.LuoJiaAI narrows the gap between the sample database and deep learning frameworks through a user-friendly data-framework collaboration mechanism,showing great potential in high-precision remote sensing mapping applications.