Skip to main content

Advanced Development of TIES - A Clinical Text Search Engine

By Rebecca Crowley

University of Pittsburgh

You must login to access the full resource.

Published on


Archived human tissues are an essential resource for translational research. Formalin-fixed, paraffin-embedded (FFPE) tissues from cancer patients are used in a wide range of assays, including RT-PCR, SNP profiling, multiplex biomarkers, imaging biomarkers, targeted exome, whole exome, and whole genome sequencing. Remainder FFPE tissues generated during patient care are 'retrospective'; use of these tissues under specific conditions does not require consent. For personalized medicine researchers, these retrospective specimens are vital resources enabling biomarker validation, detailed molecular analysis, and systems modeling before application is made to individual patients. But access to these human tissues is often a rate limiting factor in cancer research. The TIES system uses natural language processing to de-identify, structure and annotate textual clinical documents associated with human tissues. Over the past seven years, the system has been used extensively at University of Pittsburgh Cancer Institute and also at other institutions and cancer centers. In this new NCI U24 grant, we propose advanced development of the TIES software to (1) increase institutional capacity for using FFPE to support molecular characterization of human tumors, (2) increase access to tissues within cancer centers, and (3) improve the ability to share tissues and associated phenotype data among cancer centers. With this U24, we will also launch the TIES Cancer Research Network across four cancer centers: University of Pittsburgh, University of Pennsylvania, Roswell Park Cancer Institute and Georgia Regents University. The resulting software and network will provide a unique test-bed for enhancing access to cancer research resources.


Girish Chavan

University of Pittsburgh