Quantitative Analysis of Hydrocephalus Impact on FDG-PET Statistical Parametric Mapping in Dementia Evaluation
Andres Ricaurte-Fajardo1, Ana M. Franceschi3, Brielle Intorcia1, Carlyn Wisherop1, Valentina Velasco4, Arsalan Haghdel1, Anna Nordvig2, Joseph R. Osborne1, Gloria Chia-Yi Chiang1, Jana Ivanidze1
1Department of Radiology, 2Department of Neurology, Weill Medical College of Cornell University, 3Department of Radiology, Lenox Hill Hospital, Donald and Barbara Zucker School of Medicine at Hofstra/Northwell, 4School of Medicine, Pontificia Universidad Javeriana
Objective:

Our objective was to assess the impact of hydrocephalus on SPM analysis, particularly focusing on the cingulate gyrus.

Background:

The evaluation of cortical glucose metabolism in dementia diagnosis is commonly conducted using [18F]fluorodeoxyglucose (FDG) PET scans. Statistical parametric mapping (SPM) is a method employed to compare standardized uptake values (SUVs) in specific brain regions across individuals. Region-specific Z-scores, generated by SPM, quantitatively indicate the deviation of SUVs from an age-matched reference database of cognitively normal subjects allowing for a more objective and sensitive assessment. Accurate measurement of cingulate metabolism is vital for neurodegenerative disorder assessments. In the context of cognitive impairment, communicating hydrocephalus is a crucial consideration.  We hypothesize that ventricular enlargement and associated lack of FDG avidity can affect the accuracy of SPM analysis of circumventricular structures.  

Design/Methods:

We reviewed brain FDG-PET scans conducted between 2014-2022. Inclusion criteria encompassed FDG-PET scans for dementia evaluation and indications of hydrocephalus on MRI. Our analysis targeted the anterior, middle, and posterior cingulate gyrus (ACG, MCG, PCG) and used the cerebellum as a reference region due to its minimal susceptibility to hydrocephalus. Non-parametric t-tests were employed to compare each cingulate region with the reference area, utilizing GraphPad Prism software.

Results:

49 patients were evaluated (22/49, 45% female; mean age 78), the cerebellum, ACG, MCG, and PCG displayed Z-score means of -1.2 (-6 to -5), -4.3 (-13.8 to 7.7), -6.9 (-21.6 to 1.8), -3.2 (-7.9 to 1.8), respectively. Z-scores in the ACG, MCG, and PCG were markedly lower than cerebellar Z-scores (p <0.0001 in each analysis). Interestingly, this disparity was not evident in qualitative assessments.

Conclusions:

In the evaluation of dementia using FDG-PET SPM analysis, hydrocephalus can act as a confounding factor, especially when examining the cingulate gyrus. It is imperative to conduct meticulous qualitative reviews to prevent misinterpretation of SPM data.

10.1212/WNL.0000000000205877