Background: This study investigated the hypothesis that the combination of regional CT brain atrophy measurements and semiquantitative SPECT regional blood flow ratios could produce a diagnostic test for Alzheimer's disease (AD) with an accuracy comparable to that achieved with the present clinical gold standard of the NINCDS-ADRDA criteria.
Methods: Single proton emission computed tomography (SPECT) and CT head scans were performed on 122 subjects referred an UBC Alzheimer clinic and diagnosed as either 'not demented' (ND-37) or 'possible/probable AD' (AD-85) by the NINCDS-ADRDA criteria. Stepwise discriminant analysis (SDA) was performed on the bilateral SPECT regions of interest and compared to bilateral CT qualitative/quantitative assessment in the frontal, parietal and temporal lobes to determine which were most accurate at ND/AD distinction. Receiver operating curves (ROC) were then constructed for these variables individually and for their combined discriminant function.
Results: The left temporal qualitative cortical atrophy score (CT) and left temporal perfusion ratio (SPECT) were selected in the SDA. The combined discriminant function was more specific at AD/ND distinction than either of CT or SPECT alone. The accuracy of AD/ND distinction with the combined discriminant function was below that achieved by clinical diagnosis according to the NINCDS-ADRDA criteria and was not significantly different from that achieved with SPECT or CT alone as defined by ROC curve analysis.
Conclusion: The measurements of left temporal cortical atrophy and regional cerebral blood flow were most indicative of AD; however they lacked the sensitivity and specificity to recommend their use as a diagnostic test for AD.