By Erik L. Ridley, AuntMinnie staff writer
November 11, 2013

Thursday, December 5 | 10:50 a.m.-11:00 a.m. | SSQ16-03 | Room S505AB
In this session, researchers will discuss how semiquantitative indices can tackle the bane of FDG-PET brain imaging: interreader variability.

To improve uniformity, radiology is moving toward standardized reporting of imaging studies. Furthermore, with a life-altering diagnosis such as Alzheimer's disease, inconsistency in scan interpretation can have devastating consequences, said Dr. Jacob Dubroff, PhD, from the Perelman School of Medicine at the University of Pennsylvania.

"The exact reason the [U.S. Food and Drug Administration] initially withheld approval of F-18 florbetapir was out of fear of interpretation inconsistency," he said. "This is not limited to these specific scans, as one of the weaknesses of FDG-PET brain imaging for memory loss is imperfect agreement on study interpretation."

The researchers hypothesized that integrating quantification into the scan interpretation process would help reduce interreader variability. In a study involving five readers and 29 patients, the use of standardized uptake value ratios (SUVr) generated by commercially available brain analysis software (syngo Scenium, Siemens Healthcare) significantly improved interreader variability for interpreting F-18 florbetapir PET brain exams.

While no imaging test is perfect, reducing interreader variability can help standardize patient results, Dubroff said.

"Our study showed that by using this automatically generated quantitative value [SUVr], interreader variability in interpretation of these scans was dramatically improved," he said. "This suggests that by using this technique, patients will be more likely to receive a consistent interpretation regardless of reader."

Last Updated hh 11/7/2013 4:46:04 PM