RSNA 2009 

Abstract Archives of the RSNA, 2009


SSE19-01

Visual Assessment versus Quantitative FDG PET Neuroimaging for Detection of Mild Cognitive Impairment and Alzheimer's Disease

Scientific Papers

Presented on November 30, 2009
Presented as part of SSE19: Nuclear Medicine (Imaging the Brain: Structure and Function)

 Trainee Research Prize - Resident

Participants

Vance Lehman MD, Presenter: Nothing to Disclose
Patrick James Peller MD, Abstract Co-Author: Nothing to Disclose
Daniel O. Claassen MD, Abstract Co-Author: Nothing to Disclose
Robert Charles Murphy MD, PhD, Abstract Co-Author: Nothing to Disclose
Ronald C. Petersen MD, PhD, Abstract Co-Author: Nothing to Disclose
Val John Lowe MD, Abstract Co-Author: Research grant, General Electric Company Research grant, Siemens AG Research grant, Avid Radiopharmaceuticals, Inc

PURPOSE

The aim of this study was to examine the potential clinical impact of commercially available quantitative software using three-dimensional surface projection (3D-SSP) on the diagnostic accuracy of functional brain imaging with FDG PET in mild cognitive impairment (MCI) and Alzheimer disease (AD).

METHOD AND MATERIALS

 Twenty-three normal controls (age 72.7±9.0; DRS 12.5±2.1) and 31 cognitive impairment patients (18 MCI and 13AD; age 74.0±7.1; DRS 6.4±3.0) were enrolled in the Mayo Alzheimer's Disease Research Center and underwent FDG PET imaging using the standard ADNI protocol. The transaxial FDG PET images were rated by 4 readers separately, 2 beginners (<1 year experience) and 2 experts (>10 years experience) in 2 sessions. In the first interpretation only black and white and color transaxial images were available and in the second 3-D-SSP quantitation images were added. The readers rated normality, degree of abnormality and diagnostic confidence. All readers were blinded to clinical information. At a mean followup of 3.4 years, all 23 normal individuals remained cognitively normal, 13 of 13 AD patients progressed and 12 of 18 MCI patients developed AD.

RESULTS

The accuracy of experience readers (70%) was moderately better than the beginner interpreters (61%). The addition of 3D-SSP quantitation improved accuracy in both groups, with beginners (76%) benefiting more than experts (82%). The majority this improvement came from significant gains in specificity. The beginner’s specificity rose from 26% to 63% and experts from 57% to 87%. Both beginners and experts adequately assigned degree of abnormality in patients with moderate and severe AD. Expert confidence ratings consistently were 22% higher than beginners but both groups rose 20% with the addition of 3-D SSP quantitation.

CONCLUSION

The use of 3D-SSP software quantitation improved the diagnostic accuracy of FDG PET interpretation in patients with MCI or AD. Both beginners and expert readers were aided by 3D-SSP images, especially in the detection of normality and improved diagnostic confidence.

CLINICAL RELEVANCE/APPLICATION

Accurate diagnosis of AD with quantitative FDG PET would help direct appropriate treatment early in the disease course and facilitate future testing of potential disease modifying drugs.

Cite This Abstract

Lehman, V, Peller, P, Claassen, D, Murphy, R, Petersen, R, Lowe, V, Visual Assessment versus Quantitative FDG PET Neuroimaging for Detection of Mild Cognitive Impairment and Alzheimer's Disease.  Radiological Society of North America 2009 Scientific Assembly and Annual Meeting, November 29 - December 4, 2009 ,Chicago IL. http://archive.rsna.org/2009/8007416.html