Network graph representation of COVID-19 scientific publications to aid knowledge discovery.

INTRODUCTION: Numerous scientific journal articles related to COVID-19 have been rapidly published, making navigation and understanding of relationships difficult. METHODS: A graph network was constructed from the publicly available COVID-19 Open Research Dataset (CORD-19) of COVID-19-related publications using an engine leveraging medical knowledge bases to identify discrete medical concepts and an open-source tool (Gephi) to visualise the network. RESULTS: The network shows connections between diseases, medications and procedures identified from the title and abstract of 195 958 COVID-19-related publications (CORD-19 Dataset). Connections between terms with few publications, those unconnected to the main network and those irrelevant were not displayed. Nodes were coloured by knowledge base and the size of the node related to the number of publications containing the term. The data set and visualisations were made publicly accessible via a webtool. CONCLUSION: Knowledge management approaches (text mining and graph networks) can effectively allow rapid navigation and exploration of entity inter-relationships to improve understanding of diseases such as COVID-19.

Application of automated pathology reporting concepts to radiology reports.

OBJECTIVE: This study was designed to extend the concept of automated pathology reporting to radiology reports to find central nervous system (CNS) neoplasms that may currently go undetected. METHODS: Existing E-Path software was modified to account for the structure and language of radiology reports. Logic was added to allow registries to configure whether they want only new reports or if they also want history, metastatic, and/or previously known reports. Five hospital registries and 3 central registries participated. Three quality-control (QC) studies were conducted with fine-tuning taking place between the studies. The first QC study included random samples of 1,500 reports from 3 data sources. The second and third QC studies each included 1 random sample from 2 different data sources. RESULTS: The software was able to extract reportable CNS neoplasms with a high degree of specificity and sensitivity at 99% and 100% respectively, using the original set of coding rules. This rule set was favored by our hospital registries. Participating population-based registries preferred to receive only positive-new cases. The specificity and sensitivity for this category was 96% and 94% respectively. One hospital registry compared the cases found by the software to their registry database and found 13 additional CNS neoplasm cases in a 10-month period which represented an increase of 18%. CONCLUSION: Automated radiology reporting is a promising method of mining a previously untapped data source to find cases of CNS neoplasms that may be missed by conventional techniques.