HTML5 DICOM Viewer Best Practices

[Eren PARLAKGÜMÜŞ]

Hi Everybody,

We have a PACS server solution which is stable and deployed to many sites. But the Web Viewer side is Silverlight. Now we are planning to move our viewer to HTML5. We have found "cornerstone" to be the best for this kind of implementation. Now we are planning to start. What I want to ask you is, what is the best practice for this kind of purpose ? Has anyone implemented this kind of solution tho the existing PACS system ? Being more specific, are the following steps appropriate ?

- First fetch the DICOM images to client side (By the help of WADO loader probably)

- Use cornerstone libraries to parse the DICOM file(s)

- Use again cornerstone libraries to perform operations (like annotations, measurements etc)

Is there any other option ?

And also do we have to get all the DICOM images at initial connection of the viewer ? Is it better to get only a portion of the study then get the others if the user needs them ? If you can give me some advice for the start point I will be so glad.

Thanx in advance. 

Eren

[Chris Hafey]

Hi Eren,

Cornerstone has been integrated with a wide variety of back end systems including PACS.  The cornerstoneWADOImageLoader is probably the most widely used mechanism for loading images and works very well.  It does have limited support for transfer syntaxes so make sure your data is in a supported transfer syntax or your PACS can transcode.  You do not need to load all the images initially - cornerstone is designed to load them "on demand".  You may want to study at the source code for the cornerstoneDemo which is a good starting point for building your own viewer:

See it live here:

http://chafey.github.io/cornerstoneDemo/

Source code for the viewer:

https://github.com/chafey/cornerstoneDemo/tree/gh-pages

You will notice that the cornerstoneDemo first loads a json file describing each study.  Most people have to create a REST endpoint to generate a JSON like this from their database.  The new DICOMweb standards will allow us to do this in a standard way via QIDO-RS and WADO-RS Retrieve Metadata, check my blog for more info on these:

http://chafey.blogspot.com/

Chris