Jason Daehn Balkman MD, Presenter: Nothing to Disclose

Yves Martelli MS, Abstract Co-Author: Nothing to Disclose

Sharing medical images across institutions for educational or clinical purposes frequently requires access to remote data and viewing software, as well as compliance with various file formats. The movement towards cloud technology and zero footprint web applications should enable a nonproprietary solution to this challenge.

A zero footprint web-based DICOM viewer written entirely in JavaScript and HTML5 for mobile phones, tablets, and computers was uploaded to an open source repository, GitHub (github.com/ivmartel/dwv). Image windowing was made programmatically available by caching pixel data. Mechanisms for viewing local, online and PACS images were developed as part of the software. The DICOM viewer was deployed to a universally accessible server domain, venosviewer.appspot.com, and integrated with an existing cloud database of images on a separate server domain, developed using Google Cloud SQL and Blobstore.

The DICOM viewer and cloud database were freely accessible via the VeNOS website running on a modern web browser with no additional software requirement. Local and online DICOM or JPEG images could be loaded into the viewer using the “File” dialog box. Local DICOM images loaded at a rate of approximately 2-3 slices per second for 529 kilobyte size slices. Images from the cloud database could be selected through the “Cross Domain Cloud-based Case File” dialog box. Cloud-based sequences also loaded into the viewer at approximately 2-3 slices per second for a typical wireless connection, though latencies were disproportionately greater for sequences containing over 50 slices. Methods for image annotation, scrolling, zooming, and windowing were functional features of the web application.

A vendor neutral open source DICOM viewer to display local PACS or online cloud-based medical images was developed and integrated with an existing cloud database on a separate server domain. This architecture demonstrates the feasibility of viewing remote cloud-based images without the need for downloadable or proprietary software. Increased latencies experienced when loading large cloud-based image stacks were likely the result of numerous cross-domain HTTP requests.

This work combines a zero footprint DICOM viewer web application and a cloud-based case file, demonstrating the potential for open source medical image sharing between institutions and across server domains.

Balkman, J, Martelli, Y, VeNOS: A Vendor Neutral Open Source DICOM Solution for Cloud-based Medical Image Sharing.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14006020.html