Imaging Screening of Catastrophic Neurological Events Using a Software Tool: Preliminary Results

Fernandes, AP; Gomes, A; Veiga, J; Ermida, D; Vardasca, T

http://hdl.handle.net/10400.10/1467

Use this identifier to reference this record.

Name:	Description:	Size:	Format:
Transp proceedings.pdf		320.57 KB	Adobe PDF	Download

Send Feedback

Authors

Abstract(s)

BACKGROUND: In Portugal, as in most countries, the most frequent organ donors are brain-dead donors. To answer the increasing need for transplants, donation programs have been implemented. The goal is to recognize virtually all the possible and potential brain-dead donors admitted to hospitals. The aim of this work was to describe preliminary results of a software application designed to identify devastating neurological injury victims who may progress to brain death and can be possible organ donors. METHODS: This was an observational, longitudinal study with retrospective data collection. The software application is an automatic algorithm based on natural language processing for selected keywords/expressions present in the cranio-encephalic computerized tomography (CE CT) scan reports to identify catastrophic neurological situations, with e-mail notification to the Transplant Coordinator (TC). The first 7 months of this application were analyzed and compared with the standard clinical evaluation methodology. RESULTS: The imaging identification tool showed a sensitivity of 77% and a specificity of 66%; predictive positive value (PPV) was 0.8 and predictive negative value (PNV) was 0.7 for the identification of catastrophic neurological events. CONCLUSION: The methodology proposed in this work seems promising in improving the screening efficiency of critical neurological events.