Handling of Multiple CAD SRs for one image
Heaney, Dave
The current text in the IHE Technical Framework regarding how an Image Display should handle multiple CAD SRs applicable to an image is rather ambiguous. I have created the attached CP to clarify how this situation should be handled.
There is still a question regarding what the default behaviour should be when multiple CAD SRs exist. Should the latest CAD SR (by Content Date/Time) be displayed by default if multiple exist or should this be based on some other characteristic of CAD SR objects?
Manager, Standards Group
McKesson Provider Technologies, Medical Imaging Group
130 - 10711 Cambie Road
Richmond, BC
Canada
V6X 3G5
Tel: (604) 279-5422 x6306
Fax: (604) 279-5468
E-mail: David....@McKesson.com
Confidentiality Notice: This e-mail message, including any attachments, is for the sole use of the intended recipient(s) and may contain confidential and privileged information. Any unauthorized review, use, disclosure or distribution is prohibited. If you are not the intended recipient, please contact the sender by reply e-mail and destroy all copies of the original message.
Hoecker, Renate
Von: ihe-mam...@googlegroups.com [mailto:ihe-mam...@googlegroups.com] Im Auftrag von Heaney, Dave
Gesendet: Freitag, 14. Januar 2011 21:53
An: ihe-mam...@googlegroups.com
Betreff: [ihe-mammography227] Handling of Multiple CAD SRs for one image
You received this message because you are subscribed to the Google Groups "IHE Mammography" group.
To post to this group, send email to ihe-mam...@googlegroups.com.
To unsubscribe from this group, send email to ihe-mammograp...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/ihe-mammography?hl=en.
Topher Gedeon
Hi David,
I think more clarity can be given to the situation of having multiple Mammography CAD Structured Reports (SR) for a patient case and more consideration needs to be put into the resolution based on how the multiple SRs occurred and how the results should be displayed.
Here are a few scenarios on how multiple CAD SRs could occur for a single patient:
- Four standard view (RCC, LCC, RMLO, LMLO) images are sent to CAD for processing. A single CAD SR is produced and sent to the review station. Two follow up images are ordered (perhaps they are diagnostic follow ups). There will now be another CAD SR generated for this patient case and sent to the review station. If the review station only displays the most recent CAD SR, the radiologist will only see CAD marks on the last two images and not have results on the first four standard view images, which is most likely what they want to see, and can only see the results of the first four standard views if the review station can have each SR user selectable for display. Additionally, a radiologist may or may not want to see CAD marks on diagnostic images. The filtering out of diagnostic images could potentially be controlled by the CAD server or it could be a feature of the review station to display only the findings of standard views.
- Four standard view (RCC, LCC, RMLO, LMLO) images are sent to CAD for processing. A single CAD SR is produced and sent to the review station. Two follow up images are ordered (perhaps there were issues with the original images, e.g. image blur or positioning issues). There will now be another CAD SR generated for this patient case and sent to the review station. If the review station only displays the most recent CAD SR, the radiologist will only see CAD marks on the last two images and not have results on the first four standard view images unless the review station can have each SR user selectable for display. Additionally the radiologist might not want to see CAD findings on images that had quality issues. The filtering out of reacquired images could potentially be controlled by the CAD server or it could be a feature of the review station to display only the findings of the most recent standard views.
- With most sites transitioning to digital, there will most likely be digital priors along with a prior CAD SR. CAD systems can include metrics on findings in an SR such as area and diameter, which is useful for temporal comparisons. The radiologist might want to not only see the most recent SR, but also the SR of the prior year with the metrics displayed to see the change over time.
The onus could be placed on CAD systems with the issue of follow up images, where the CAD system could cache or group images over a period of time so that all images are represented in a single CAD SR. So in the case of the four standard images, initially there would be a CAD SR for the four images, but if two follow up images were acquired and sent to CAD, the CAD server could cache or group the follow up images with the previous four images and therefore generate a single CAD SR that contained the results of the six images.
One issue that could occur with having the CAD server group images is if the CAD server is using case based processing, that is, using the sum of all images in a patient case to make a decision. For example, if four standard views are sent to the CAD server, CAD results will be generated for those four images. If additional images are sent to the CAD server and are grouped together, the results for the first four images may now be different based on the additional images acquired. In this case, displaying the most recent CAD SR would be desired, with the exception of if the user wanted to see CAD on priors by default.
In its simplest workflow, having a statement indicating that “all CAD SRs be made available (user selectable) for display on the review workstation with the most recent CAD SR (Content Date/Time) being displayed by default” would probably satisfy most users.
Most CAD SRs reference a single case that contains multiple images. If multiple CAD SRs exist for a patient case, another solution would be to not display the most recent CAD SR for the “case”, but rather display the most recent CAD results for each “image”. This would be useful if there were multiple results for a single image due to follow images in a case being acquired.
These are just some thoughts that come to mind when considering this IHE Change Proposal.
Regards,
Topher Gedeon
iCAD, Inc.
IMPORTANT NOTICE:This message and any attachments are confidential and intended only for the individual or entity named above and may contain information that is privileged and/or proprietary. If you are not the intended recipient, you are notified that disclosing, copying, distributing, disseminating or using the contents of this message and/or any attachments for any purpose is strictly prohibited and may be unlawful. If you received this message in error, or you otherwise are not an authorized recipient, please immediately delete the message and any attachments and notify the sender.
P Please consider the environment before printing this email.