TIES over the years

Development of the TIES system began in 2002 as part of the Shared Pathology Informatics Network. University of Pittsburgh worked with Clem McDonald at Regenstrief University to develop the first federated system for querying pathology reports across institutions. As part of that grant, the TIES development team created the Pittsburgh information extraction system using GATE which was a newly released open source Software Architecture for Language Engineering (SALE). That system was developed and a few queries were issued to it across the network of four institutions, but the vision of SPIN was never fully realized.

In 2004, the same development team began work on developing caTIES as part of a caBIG contract with the National Cancer Institute Center for Bioinformatics (NCICB) through Booz Allen Hamilton. caTIES become one of the first four development projects to use the nascent caGRID. With funding from caBIG, we significantly matured our product. caBIG provided our first external adopters. Working with Mike Feldman and Tara McSherry at University of Pennsylvania, we performed the first evaluation of TIES. Other adopters followed. During this time a very early version of caTIES was integrated into caTISSUE.

As caBIG was ending, we turned our focus back towards our own institution. During this time, we implemented caTIES at the University of Pittsburgh and began working with our own Tissue Bank to transfer administration of the system over to them. With funding from our Clinical and Translational Sciences Institute, the period from 2006-2009 helped us make TIES easier to install, easier to fit into a complex academic medical center, easier to process millions of documents, and easier for researchers to use. We became a UPMC approved, production level system during this time, and we concentrated on recruiting users at Pitt and modifying our system based on their feedback.

In 2009, we were very fortunate to be funded by an NCI Software Development and Maintenance Grant (R01 CA132672). The work resulting from this grant enabled us to extend TIES to other database management systems beyond MySQL and other document types. The need for better and faster installations eventually drove us to a VM deployment which we released at the end of the funding period for this grant.

In 2010, the TIES team began working with the Department of Radiology at University of Pittsburgh to extend our approach to the processing of more than 20 million Radiology documents across all UPMC hospitals. Increasing our document set by a factor of 10 required an entirely new approach, including the replacement of the MMTx concept recognition system with our own algorithm called NOBLE coder. Many other significant enhancements were made at that time. Working with Allesandro Furlan and the Suzanne Burdin, we successfully implemented TIES for Radiology notes and tied the system in with the Radiology Honest Brokers and their existing processes.

Meanwhile, hoping to finally stand up a real federated data and tissue network, we applied for a CTSA supplement to develop a federated network using TIES at University of Pittsburgh and University of Pennsylvania. That funding ran from 2011 – 2012 and set the stage for the current TCRN. During this time we worked out the first network agreement between Pitt and Penn which formed the basis for our current agreements. We finally had a functioning network but the sand in the hourglass ran out. Without a funding source, our agreement lapsed and we could not exchange data anymore.

Lucky for us, the NCI began a new and exciting chapter in their commitment to Biomedical Informatics. The NCI Informatics Technology for Cancer Research(ITCR) program started and began funding technology development projects using the U24 and U01 mechanisms. As part of the first group of funded projects, our ITCR U24 (U24CA180921) is supporting the deployment of the system across two additional sites (Roswell Park Cancer Institute and Augusta University). That grant will also fund a number of major enhancements to the system making it possible to include virtual slides, and to use the system for tissue microarray creation. A major goal of the U24 grant is to get 16 separate scientific projects using the TIES system across institutions over the next four years.

The most exciting period of TIES development is just around the corner. We are confident now that we will have an operationalized federated network supporting real cancer research by the end of 2014. Our goal is to develop TCRN into a national data and tissue sharing network and a major national clinical and translational research asset.