Paper: Identifying Pathological Findings in German Radiology Reports Using a Syntacto-semantic Parsing Approach

ACL ID W13-1904
Title Identifying Pathological Findings in German Radiology Reports Using a Syntacto-semantic Parsing Approach
Venue Workshop on Biomedical Natural Language Processing
Session
Year 2013
Authors

In order to integrate heterogeneous clinical information sources, semantically correlating information entities have to be linked. Our discussions with radiologists revealed that anatomical entities with pathological findings are of particular interest when linking radi- ology text and images. Previous research to identify pathological findings focused on sim- plistic approaches that recognize diseases or negated findings, but failed to establish a holis- tic approach. In this paper, we introduce our syntacto-semantic parsing approach to clas- sify sentences in radiology reports as either pathological or non-pathological based on the findings they describe. Although we operate with an incomplete, RadLex-based linguistic resource, the obtained results show the effec- tiveness of our approac...