Editor's note: SearchHealthIT.com received this feedback in response to the article New Photoshop image deblurring feature could change radiology. It has been edited for style but not for content.
Your article fails to mention that this raises a classic example of why the FDA reviews products that will impact the quality of images that are used to diagnose patients.
Think about it. A child has neurological symptoms and an MRI is ordered. Should the site be expected to have a "Propeller" type feature on its MRI system, one that corrects the raw data upon MRI acquisition in the scanner -- a sequence that has been reviewed by the FDA and does not diminish diagnostic effectiveness -- or is it ok to use a Photoshop tool that creates new lines and edges within the image data, possibly creating "new image data"? Do you think correcting an image of a sunset taken through a car window on a bumpy road is comparable to that pediatric MRI in terms of image integrity issues? Does Photoshop really want the FDA looking over their shoulder, and the liability?
I think that without any other data in your article other than it's "de-blurring" the image, the idea is terrible from a quality and safety point of view. Radiology is not about making pretty images; it's about making accurate, diagnostic images to better manage the patient. We've had image-space filters for decades; that's nothing new-- high pass and low pass filters. Contextvision filters built into MRI and CT scanners, independent workstations for low field MRI and Siemens IRIS. They are all image space filters.
The article raises major questions of image quality/integrity and FDA 510K or PMA requirements -- that were neither answered nor addressed.
Let us know what you think about the story; email Jean DerGurahian, Executive Editor.
This was first published in December 2011