Discriminative use of SPECT in frontal lobe-type dementia versus (senile) dementia of the Alzheimer's type

Abstract

Dementia of the Alzheimer's type [(S)DAT] and dementia with frontal features (FLD) are nosological entities with different prognoses and presumed pathophysiology. There is a need for noninvasive differential diagnostic tools. To evaluate whether SPECT perfusion imaging could discriminate between these neurodegenerative disorders, we performed a comparative study.

Methods: SPECT scans using 99mTc-hexamethylpropylene amine oxime (99mTc-HMPAO) of 21 patients with FLD were compared with those obtained in a group of 19 age- and severity-matched patients suffering from (S)DAT. Brain SPECT perfusion deficits were scored by visual qualitative analysis with respect to location, lateralization and severity. A total severity score of cerebral hypoperfusion (maximal value = 18) was calculated by adding all severity scores (scored between 0 and 3; 0 = no perfusion deficit; 1 = 13%-30% hypoperfusion; 2 = 30%-50%, hypoperfusion and 3 = > 50% hypoperfusion including breaching of the cortex) for right and left frontal, parietal and temporal lobes. Moreover, bifrontal hypoperfusion (Fa) was scored, yielding a value between 0 and 6 by adding the two frontal severity scores.

Results: No significant correlation was found between MMSE scores and total severity scores on SPECT. A statistically significant correlation was found between the Middelheim frontality score and frontal severity score. Statistically more significant bilateral hypoperfusion of the parietal lobes was found in the (S)DAT group. Conversely, bifrontal hypoperfusion was found more in the FLD group. Stepwise logistic regression analysis identified the severity of bifrontal hypoperfusion as the most significant contributing parameter to correctly classifying (S)DAT versus FLD on SPECT. The probability of predicting (S)DAT based on the SPECT scan is calculated with the following formula: [equation: see text] Using this equation, a value above 0.5 was predictive for (S)DAT and a calculated value under 0.5 was predictive for FLD. Using this model, 81% of the FLD group and 74% of the (S)DAT were correctly classified.

Conclusion: Technetium-99m-HMPAO SPECT may help in discriminating FLD from (S)DAT. Bifrontal hypoperfusion was found to be the most powerful predictor of clinical classification. Further validation of the presented logistic regression model is warranted.