Discriminant Value of 18FDG-PET in Amyotrophic Lateral Sclerosis (P3.011)

Objective. To assess the discriminant value of 18FDG-PET at rest in the largest series of ALS patients investigated so far. Background. ALS is characterized by a the loss of spinal and bulbar motor neurons and corticospinal tracts degeneration. A disease marker of upper motor neuron (UMN) degeneration is still lacking. Studies on 1H-MR spectroscopy and DTI hold promise for detecting and quantifying subclinical UMN damage. Few data have been reported on cerebral 18FDG-PET. Methods. 195 ALS patients (77 women and 118 men, mean age 63.2 [SD 12.1]) and 40 control subjects (11 women and 29 men, mean age 62.6 [SD 14.1]) underwent brain 18FDG-PET. Twenty-six cortical and sub-cortical brain regions were segmented by the Pick Atlas tool in SPM2 and relative metabolic uptakes individually normalized by whole brain values. Factorial (FA) and Discriminant (DA) analysis were performed using the 52 bilateral regions as well as age and sex. The strict statistical constraint of post-hoc cross-validation was applied. Results. FA identified 8 factors in controls and 10 factors in ALS explaining 90% and 87% of total variance,. As compared to controls, ALS showed factors with stronger laterality, networking separately left and right temporal lobes and left somatosensory and superior parietal cortex. In ALS FA gathered also selectively insulas, and caudate and thalamic nuclei. DA performed on all 235 subjects showed following cross-validation an accuracy of 88% (sensitivity 89% and specificity 83%) when all 52 regions were taken into account and of 80% (equal sensitivity and specificity) when analyzing the discriminant value of the 10 factors. Conclusions. In ALS different cortical and subcortical networks were found as compared to normal controls. When 52 functional regions were submitted to discriminant analysis the overall accuracy was 88% with a sensitivity of 89%. This findings show for the first time the usefulness of 18FDG-PET in differentiating ALS patients from normal controls and paves the way to a larger implementation of such methodology in ALS diagnosis. Disclosure: Dr. Pagani has nothing to disclose. Dr. Calvo has nothing to disclose. Dr. Moglia has nothing to disclose. Dr. Nobili has nothing to disclose. Dr. Fania has nothing to disclose. Dr. Morbelli has nothing to disclose. Dr. Valentini has nothing to disclose. Dr. Canosa has nothing to disclose. Dr. Cistaro has nothing to disclose.