Skip Navigation
National Cancer Institute U.S. National Institutes of Health www.cancer.gov
NCI Wiki New Account Help Tips
Skip to end of metadata
Go to start of metadata

cTAKES 1.3

Contents of this Page

Introduction

Clinical Text Analysis and Knowledge Extraction System (cTAKES) is a system through which one creates one or more pipelines to processes clinical notes and to identify clinical named entities. For example, a pipeline can identify mentions of drugs, diseases/disorders, signs/symptoms, anatomical sites and procedures. Each named entity that is found is given attributes for the text span, the ontology mapping code, the context (probable/possible, family history of, or history of), and negated/not negated.

Overview

cTAKES is built on the UIMA framework. cTAKES 1.3 does not provide a GUI of its own for installation or processing. The documentation shows how to use the GUIs provided by the UIMA framework, and how to run cTAKES from a command line.

Before using cTAKES you need to know that cTAKES does not provide any mechanisms of its own to handle patient data securely. It is assumed that cTAKES is installed on a system that can process patient data, or that any data being processed by cTAKES has already been through a deidentification step in order to comply with any applicable laws.

What's New

The cTAKES community brings you the following functions in 1.3:
*Inclusion of a set of UMLS dictionaries (SNOMED-CT and RxNorm). The Dictionary Lookup component now includes these dictionaries for doing SNOMED CT and RxNorm Named Entity Recognition out-of-the-box. You no longer need to take an extra step to get production level dictionaries except to supply a UMLS username and password.
*Constituency Parser - A new annotator component
*Coreference resolver - A new annotator component

Download and Install

There are two kinds of cTAKES users. One is the user that wishes to use the tool, without any compile steps or use of development environments. The goal is primarily to get the tool up and running and to apply it to clinical documents. These users will install the software like any other product that you might buy or download. Note that there still may be file configuration required by using an editor.

The second is a developer that will be taking the cTAKES software to extend or modify it here and there to suit project needs. More parameters and such can be customized in the developers' environment for fine tuned control, but use does require that you have an integrated development environment in which to modify, build and deploy the code.

There is a third install path for developers, so it does not count as a third user type. It is possible to build a back-end server out of the software. That is, there will be no graphical user interface available. The developer will have only application programming interfaces (APIs) available with which to interact. Instructions for this are embedded in the Developer install instructions. At one point during the install, pay attention to the Important icon which marks steps only required if you want a back-end server.

Just as with cTAKES 1.2, this release provides you the option to use icTAKES, an integrated version of Mayo Clinic's cTAKES for end users and developers, or cTAKES with the original project and directory structures. To distinguish them, cTAKES (downloadable [from SourceForge|http://sourceforge.net/projects/ohnlp/files/cTAKES) is distributed as cTAKES 1.3.1 and icTAKES (also downloadable from SourceForge is distributed as cTAKES 1.3.2. The table below links to the proper download and install instructions depending on the user. Consideration is also given to the install environment you wish to use. The icTAKES release is recommended, which has a simple install mechanism.

cTAKES Download and Install Selection

Documentation

Support

We encourage use of our forums as the first line for inquiries. The forums are split into these places for discussion which match up to the user types:

If you have an issue that can not be placed into public forums then we also have an email address for you to use: clinicalnlp@mayo.edu