Shared Pathology Informatics Network: Information retrieval for the genomic era

It is now widely recognized that of all the components of the mission to translate the "new biology" into the "new medicine" it is not genomic measurements which are the major challenge. Through the efforts of researchers world-wide, genomic measurements have achieved commodity-level status. Contrariwise, phenotypic annotations to the genomic measurements have remained irreproducibly labor-intensive, dependent on experts (typically clinicians) and therefore are the true rate limiting step in translating the new biology. The Shared Pathology Informatics Network (SPIN) is a constructive and highly successful response to this problem funded in response to an RFA from the National Cancer Institute. We describe here part of the SPIN effort, that led by a consortium of Harvard-affiliated and UCLA-affiliated hospitals entitled the Consented High-performance Indexing and Retrieval of Pathology Specimens (CHIRPS).

   

CHIRPS' has had the following goals:

  • Establish a scaleable representation of specimens and derivatives, starting with surgical pathology, and growing to include tissue banks and tumor registries.
  • Support a taxonomy of consent for use of tissue within the specimen annotations.
  • Develop a Peer-to-Peer (P2P) distributed architecture for locating specimens that leverages the Internet and promotes local control of data.
  • Minimize the effort (time, resources, politics) required to participate in the SPIN network.

.

   
CHIRPS has been highly successful in several dimensions. Even though it is in the 4th year of a proof-of-principle research plan, it has already generated the following achievements.
  • Developed a P2P network that provides a searcharble index to millions of specimens across pathology departments coast to coast.
  • Defined and implemented the privacy protection mechanisms that such a network entails.
  • Even before going live with CHIRPS on a national scale, it has been implemented within the Dana-Farber Harvard Cancer Center (DFHCC) to provide access to over one million specimens across the Harvard-affiliated hospitals. This included obtaining DFHCC-wide approval by several Institutional Review Boards to ensure that CHIPRS meets the highest levels of ethical research
    

References

1. A. H. Namini, D. A. Berkowicz, I. S. Kohane, H. Chueh, A submission model for use in the indexing, searching, and retrieval of distributed pathology case and tissue specimens, Medinfo 2004, 1264 (2004).
2. A. M. Holzbach, H. Chueh, A. J. Porter, I. S. Kohane, D. Berkowicz, A query engine for distributed medical databases, Medinfo 2004, 1519 (2004).

Related Sites:

NCI's Cancer Diagnosis Program

SPIN News.

Presentation by Jon Braun in 2005

A text scrubber suite developed within SPIN (it all also "chunks" the pathology report, parses it, and autocodes it.

SPIN Implementation in Boston

 

 

Frequently asked questions

Q: Can I join SPIN or CHIRPS?

A: SPIN is still in a research phase. However, there is room for highly motivated collaborators who are prepared to participate in our alpha testing and refinement of our software and procedures.

Q:

On what hardware or software platform does CHIRPS run?

A: CHIRPS is implemented in Java and has been successfully run on Windows, Linux and Macintosh OS X platforms.