Description
In 2017, the Precision Medicine track (TREC-PM) of the Text REtrieval Conference (TREC) aimed to facilitate the design of automatic systems capable of providing useful precision-medicine related information to clinicians treating patients with cancer. Given a query describing a patient's tumor, genetic variants, and demographics, systems were evaluated on their ability to retrieve and rank (1) scientific articles from MEDLINE describing targeted treatments, or (2) pertinent clinical trials for which the patient may be eligible. In this poster, we explore whether automatically inferring targeted medications and treatments improved performance in the 2017 TREC-PM evaluation.
Learning Objective 1: Evaluate the impact of inferring treatments on information retrieval for precision medicine
Learning Objective 2: Infer target treatments for genetic variants of cancer from multiple knowledge sources
Authors:
Travis Goodwin (Presenter)
University of Texas at Dallas
Sanda Harabagiu, University of Texas at Dallas
Presentation Materials: