Adapting a natural language processing tool to facilitate clinical trial curation for personalized cancer therapy.
Lists
Only you see this
Abstract
The design of personalized cancer therapy based upon patients' molecular profile requires an enormous amount of effort to review, analyze and integrate molecular, pharmacological, clinical and patient-specific information. The vast size, rapid expansion and non-standardized formats of the relevant information sources make it difficult for oncologists to gather pertinent information that can support routine personalized treatment. In this paper, we introduce informatics tools that assist the retrieval and curation of cancer-related clinical trials involving targeted therapies. Particularly, we adapted and extended an existing natural language processing tool, and explored its applicability in facilitating our annotation efforts. The system was evaluated using a gold standard of 539 curated clinical trials, demonstrating promising performance and good generalizability (81% accuracy in predicting genotype-selected trials and an average recall of 0.85 in predicting specific selection criteria).
BibTeX
@inproceedings{zeng2014adapting,
author = {Jia Zeng and Yonghui Wu and Ann M. Bailey and Amber M. Johnson and Vijaykumar Holla and Elmer V. Bernstam and Hua Xu and Funda Meric‐Bernstam},
title = {{Adapting a natural language processing tool to facilitate clinical trial curation for personalized cancer therapy.}},
booktitle = {PubMed},
year = {2014},
}Loading related content...