Neuradicon: operational representation learning of neuroimaging reports
This enables operational optimization in radiology by allowing quantitative monitoring of report content, addressing a domain-specific bottleneck in healthcare.
The authors tackled the problem of analyzing unstructured neuroradiological reports by developing Neuradicon, an NLP framework that converts them into quantitative representations, achieving excellent generalizability across 336,569 reports from two healthcare institutions.
Radiological reports typically summarize the content and interpretation of imaging studies in unstructured form that precludes quantitative analysis. This limits the monitoring of radiological services to throughput undifferentiated by content, impeding specific, targeted operational optimization. Here we present Neuradicon, a natural language processing (NLP) framework for quantitative analysis of neuroradiological reports. Our framework is a hybrid of rule-based and artificial intelligence models to represent neurological reports in succinct, quantitative form optimally suited to operational guidance. We demonstrate the application of Neuradicon to operational phenotyping of a corpus of 336,569 reports, and report excellent generalizability across time and two independent healthcare institutions.