UK Imaging Informatics Group
Solving the riddle of related procedures PreviousNext
UK Imaging Informatics Group > Questions & Answers > PACS Integration & Standards >
 Link to this message Joel Lidstrom  posted on Tuesday, March 18, 2008 - 03:16 pm Edit Post Delete Post Print Post
My partner and I have created, an automated website for determining relationships among radiologic procedures. This has grown out of a desire to automatically create related-procedures files for Fuji Synapse.

We have had tremendous success, and would like to migrate our ideas to other PACS platforms. In researching how we might do this I came across this forum, and your National Interim Descriptions for Diagnostic Imaging Procedures. While you are engaged in this long-needed effort to standardize procedure names and descriptions, we have spent our time finding similarity among procedures no matter how variously they are named.

This is important for two reasons. Many radiology departments struggle with what your National Descriptions seek to remedy, i.e., poor naming conventions, free-texted data entry at the modality, and synonyms that computers don't recognize as being related. But perhaps equally important is the continuous need to re-define correspondence among radiology's growing complexity of studies. No matter how accurately we devise language to describe individual studies, those studies bear a relationship to a host of other functionally similar procedures.

This has been RadMapps' focus--to always assure when viewing a current study that any related study, no matter how it is named, automatically hangs for comparison.

What do we mean by "related"? Relationships are defined, not only by body part but also by functionality. For example, when an MR of the knee is read, the femoral artery runs through it but isn't looked at to any significant extent. Conversely, if an angio CT of the knee area is read, the bones of the knee are of no interest. These studies, both of the knee, bear no functional relationship.

Contrast this with studies whose foci are various: A Renal Calc CT, a Pancreas Oncology scan, a Gallbladder Oncology scan, a Hepatic Hemangioma study, a CT Urography study. Each has a different intent, yet all would reasonably hang as priors to a CT_Abdomen.

The future of PACS automation will rely heavily on the standardization of radiologic terminology. But it will also rely on software that can unfailingly find not just identical procedures, but similar and functionally related ones.

Please take a look at to see if there is anything there to complement your fine work. You can assess it one of two ways. Either upload a spreadsheet of procedures for RadMapps to map (go to, or look at a mishmash of mapped procedures at by entering the address "", and the password "DemoUser".

(Fuji users will soon be able to download XML Related Procedures files from our site. Please inquire!)


Joel Lidstrom
Add Your Message Here
Username: Posting Information:
This is a private posting area. Only registered users may post messages here.
Options: Automatically activate URLs in message