Home

Cancer Data Element Benchmark Documentation

Welcome to Cancer Data Element Extraction Benchmark documentation.

Goals

We aim to create a multi-site cancer data element extraction benchmark dataset to support the evaluation of fine-tuning strategies.

• We curate gold standard annotations from clinical texts for various COVID-19 related NLP tasks:
  • COVID-19 signs and symptoms
  • Adverse events of vaccinations
  • Long COVID
  • Social determinants of health
  • Context Classification
• We design and develop infrastructure for concept extraction efforts from clinical texts.
  • N3C NLP Infrastructure Overview
  • OHNLP Backbone
  • Online Rule Editor
• We work with participating sites to implement the OHNLP federated evaluation process.

Please join us if you would like to know more or contribute!