HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Download to citation manager Google Scholar Articles by Deeb, E. Articles by Ziolko, S. PubMed Articles by Deeb, E. Articles by Ziolko, S.

Visual diagnosis vs statistical FDG PET analysis to distinguish between normal, mild cognitive impairment (MCI) and Alzheimer’s (AD)

¹ Radiology, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania

2023

Objectives: Brain PET for AD has traditionally relied upon findings of hypometabolism in the parietal/temporal lobes among other areas. Softwares for image analysis have been developed which compare the PET scan to a database of normal controls and generate areas of statistically significant hypometabolism. We compared the accuracy of visual assessment of PET to the outputs of 2 softwares: 3D-SSP and PMOD.

Methods: Patients were screened through the University of Pittsburgh Alzheimer’s Research Center using NINCDS-ADRDA criteria. 19 normal controls (NC), 17 MCI, and 16 AD underwent FDG PET. The scans were processed by the 3DSSP and PMOD algorithms. Scans were randomly distributed for visual interpretation and at a different time all were again randomized and rated using the 3D SSP and PMOD algorithms. Visual analysis was performed by two expert brain PET scan interpreters. Raters scored each scan from 1 – 4 where 1 was definitely normal, 2 was questionably normal, 3 was questionably abnormal, and 4 was definitely abnormal. An accurate call for NC would be a rating of 1 or 2. MCI subjects were considered correct if called 2 or 3. The same rating scale was used for assessment of AD where 1 was definitely not AD and 4 was definitely AD.

Results: Rater 1 was correct in 13/19 NC, 8/17 MCI, and 11/ 16 AD. Rater 2 correctly identified 11/19 NC, 10 /17 MCI, and 9 /16 AD. When 3DSSP was combined with visual reads, rater 1 and 2 agreed on 9/10 NC, all MCI, and 5 out of 6 AD. When using PMOD rater 1 and 2 agreed on the classification of all NC, 8/10 MCI, and all but 1 AD.

Conclusions: Visual assessment provides fairly accurate interpretation of NC from AD, but when combined with statistical methods the accuracy significantly increases for discriminating NC and AD.

This Article Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Download to citation manager Google Scholar Articles by Deeb, E. Articles by Ziolko, S. PubMed Articles by Deeb, E. Articles by Ziolko, S.