Digital Pathology and Visuvi Visual Search

Reading time: 2 – 4 minutes

At the 2009 JavaOne Conference in June, Visuvi Inc., a Redwood City, California-based company developing visual search solutions, announced a Java-based visual search technology that integrates a JavaFX front-end with PathXchange, a Web 2.0 pathology portal. The Visuvi Java application enables medical professionals to take a digital biopsy image of 50,000 x 60,000 pixels (that’s 3000 mega pixels) and search medical images in leading databases for similarities.

The power of this technology is that the medical images in the databases have corresponding case information. Thus, images that are found to be similar also provide information on patient diagnosis, treatments, outcomes, etc. This is particulary useful in oncology since, in more cases than you’d expect, pathologists disagree on whether a particular image shows cancer or not [1-3]. By comparing pathology images, the technology greatly accelerates the review process and allows a diagnosis to be more objective than what is currently done.

Check out their demo at the conference in the video below. James Gosling (the inventor of Java) was clearly impressed with the significance of the technology. Chris Boone, CEO and president of Visuvi, and Florian Brody, VP Marketing, highlighted and captured a section of an actual prostate cancer biopsy image and searched 90,000 images in 0.3 seconds to find related patient cases.

PathXchange is a not-for-profit professional networking portal for the global pathology community. PathXchange brings the field of pathology into the digital age with Web 2.0 features designed to promote exchange of pathological cases, knowledge and information, combining the elements of a case gallery, community content contribution and professional networking.

References

  1. Farmer et al. Discordance in the histopathologic diagnosis of melanoma and melanocytic nevi between expert pathologists. Hum Pathol. 1996 Jun;27(6):528-31.
    View abstract
  2. Lettieri et al. Discordance between general and pulmonary pathologists in the diagnosis of interstitial lung disease. Respir Med. 2005 Nov;99(11):1425-30. Epub 2005 Apr 21.
    View abstract
  3. Lodha et al. Discordance in the histopathologic diagnosis of difficult melanocytic neoplasms in the clinical setting. J Cutan Pathol. 2008 Apr;35(4):349-52.
    View abstract
About the Author

Walter Jessen, Ph.D. is a Data Scientist, Digital Biologist, and Knowledge Engineer. His primary focus is to build and support expert systems, including AI (artificial intelligence) and user-generated platforms, and to identify and develop methods to capture, organize, integrate, and make accessible company knowledge. His research interests include disease biology modeling and biomarker identification. He is also a Principal at Highlight Health Media, which publishes Highlight HEALTH, and lead writer at Highlight HEALTH.