A Machine Learning Approach to First Pass Reperfusion in Mechanical Thrombectomy: Prediction and Feature Analysis.

INTRODUCTION: Novel machine learning (ML) methods are being investigated across medicine for their predictive capabilities while boasting increased adaptability and generalizability. In our study, we compare logistic regression with machine learning for feature importance analysis and prediction in first-pass reperfusion. METHODS: We retrospectively identified cases of ischemic stroke treated with mechanical thrombectomy (MT) at our institution from 2012-2018. Significant variables used in predictive modeling were demographic characteristics, medical history, admission NIHSS, and stroke characteristics. Outcome was binarized TICI on first pass (0-2a vs 2b-3). Shapley feature importance plots were used to identify variables that strongly affected outcomes. RESULTS: Accuracy for the Random Forest and SVM models were 67.1% compared to 65.8% for the logistic regression model. Brier score was lower for the Random Forest model (0.329 vs 0.342) indicating better predictive capability. Other supervised learning models performed worse than the logistic regression model, with accuracy of 56.2% for Naïve Bayes and 61.6% for XGBoost. Shapley plots for the Random Forest model showed use of aspiration, hyperlipidemia, hypertension, use of stent retriever, and time between symptom onset and catheterization as the top five predictors of first pass reperfusion. CONCLUSION: Use of machine learning models, such as Random Forest, for the study of MT outcomes, is more accurate than logistic regression for our dataset, and identifies new factors that contribute to achieving first pass reperfusion. The benefits of machine learning, such as improved predictive capabilities, integration of new data, and generalizability, establish ML as the preferred model for studying outcomes in stroke.

Inositol polyphosphate phosphatases in human disease.

Phosphoinositide signalling molecules interact with a plethora of effector proteins to regulate cell proliferation and survival, vesicular trafficking, metabolism, actin dynamics and many other cellular functions. The generation of specific phosphoinositide species is achieved by the activity of phosphoinositide kinases and phosphatases, which phosphorylate and dephosphorylate, respectively, the inositol headgroup of phosphoinositide molecules. The phosphoinositide phosphatases can be classified as 3-, 4- and 5-phosphatases based on their specificity for dephosphorylating phosphates from specific positions on the inositol head group. The SAC phosphatases show less specificity for the position of the phosphate on the inositol ring. The phosphoinositide phosphatases regulate PI3K/Akt signalling, insulin signalling, endocytosis, vesicle trafficking, cell migration, proliferation and apoptosis. Mouse knockout models of several of the phosphoinositide phosphatases have revealed significant physiological roles for these enzymes, including the regulation of embryonic development, fertility, neurological function, the immune system and insulin sensitivity. Importantly, several phosphoinositide phosphatases have been directly associated with a range of human diseases. Genetic mutations in the 5-phosphatase INPP5E are causative of the ciliopathy syndromes Joubert and MORM, and mutations in the 5-phosphatase OCRL result in Lowe's syndrome and Dent 2 disease. Additionally, polymorphisms in the 5-phosphatase SHIP2 confer diabetes susceptibility in specific populations, whereas reduced protein expression of SHIP1 is reported in several human leukaemias. The 4-phosphatase, INPP4B, has recently been identified as a tumour suppressor in human breast and prostate cancer. Mutations in one SAC phosphatase, SAC3/FIG4, results in the degenerative neuropathy, Charcot-Marie-Tooth disease. Indeed, an understanding of the precise functions of phosphoinositide phosphatases is not only important in the context of normal human physiology, but to reveal the mechanisms by which these enzyme families are implicated in an increasing repertoire of human diseases.

A public health initiative to increase annual influenza immunization among hospital health care personnel: the San Diego Hospital Influenza Immunization Partnership.

BACKGROUND: A public health department-supported intervention to increase influenza immunization among hospital-based health care practitioners (HCPs) in San Diego County took place between 2005 and 2008. The study included all major hospitals in the county, with a population of approximately 3.5 million. METHODS: Information on hospital activities was collected from before, during and after initiative activities. Vaccination status and demographics were collected directly from HCP using hospital-based and random-dialed telephone surveys. RESULTS: Between 2006 and 2008, hospitals increased promotion activities and reported increases in vaccination rates. Based on the random-dialed surveys, HCP influenza vaccination coverage rates did not increase significantly. Vaccination rates were significantly higher in HCPs who reported that employers provided free vaccination and those who believed that their employers mandated influenza vaccination. CONCLUSIONS: This local public health initiative and concurrent state legislation were effective in increasing employer efforts to promote influenza vaccination; however, population-based surveys of HCPs did not show significant increases in influenza vaccination. Overall, this study suggests that public health leadership, intensive employer promotion activities, and state-required declinations alone were not sufficient to significantly increase HCP influenza vaccination. Policymakers and employers should consider mandates to achieve optimal influenza vaccination among HCPs.