NeoAI 1.0: Machine learning-based paradigm for prediction of neonatal and infant risk of death

In this paper, we give a short introduction to machine learning and survey its applications in radiology. We focused on six categories of applications in radiology: medical image segmentation, registration, computer aided detection and diagnosis, brain function or activity analysis and neurological disease diagnosis from fMR images, content-based image retrieval systems for CT or MRI images, and text analysis of radiology reports using natural language processing (NLP) and natural language understanding (NLU). This survey shows that machine learning plays a key role in many radiology applications. Machine learning identifies complex patterns automatically and helps radiologists make intelligent decisions on radiology data such as conventional radiographs, CT, MRI, and PET images and radiology reports. In many applications, the performance of machine learning-based automatic detection and diagnosis systems has shown to be comparable to that of a well-trained and experienced radiologist. Technology development in machine learning and radiology will benefit from each other in the long run. Key contributions and common characteristics of machine learning techniques in radiology are discussed. We also discuss the problem of translating machine learning applications to the radiology clinical setting, including advantages and potential barriers. Copyright © 2012. Published by Elsevier B.V.

Background Neonatal sepsis is a significant cause of morbidity and mortality in neonates. Appropriate clinical diagnosis and empirical treatment in a given setting is crucial as pathogens of bacterial sepsis and antibiotic sensitivity pattern can considerably vary in different settings. This study was conducted at Bugando Medical Centre (BMC), Tanzania to determine the prevalence of neonatal sepsis, predictors of positive blood culture, deaths and antimicrobial susceptibility, thus providing essential information to formulate a policy for management of neonatal sepsis. Methods This was a prospective cross sectional study involving 300 neonates admitted at BMC neonatal unit between March and November 2009. Standard data collection form was used to collect all demographic data and clinical characteristics of neonates. Blood culture was done on Brain Heart Infusion broth followed by identification of isolates using conventional methods and testing for their susceptibility to antimicrobial agents using the disc diffusion method. Results Among 770 neonates admitted during the study period; 300 (38.9%) neonates were diagnosed to have neonatal sepsis by WHO criteria. Of 300 neonates with clinical neonatal sepsis 121(40%) and 179(60%) had early and late onset sepsis respectively. Positive blood culture was found in 57 (47.1%) and 92 (51.4%) among neonates with early and late onset neonatal sepsis respectively (p = 0.466). Predictors of positive blood culture in both early and late onset neonatal sepsis were inability to feed, lethargy, cyanosis, meconium stained liquor, premature rupture of the membrane and convulsion. About 49% of gram negatives isolates were resistant to third generation cephalosporins and 28% of Staphylococcus aureus were found to be Methicillin resistant Staphylococcus aureus (MRSA). Deaths occurred in 57 (19%) of neonates. Factors that predicted deaths were positive blood culture (p = 0.0001), gram negative sepsis (p = 0.0001) and infection with ESBL (p = 0.008) or MRSA (p = 0.008) isolates. Conclusion Our findings suggest that lethargy, convulsion, inability to feed, cyanosis, PROM and meconium stained liquor are significantly associated with positive blood culture in both early and late onset disease. Mortality and morbidity on neonatal sepsis is high at our setting and is significantly contributed by positive blood culture with multi-resistant gram negative bacteria.