Abstract
The differential diagnosis across different variants of degenerative diseases is sometimes controversial. This study aimed to validate a qualitative scoring method for the pentagons copy test (QSPT) of Mini-Mental State Examination (MMSE) based on the assessment of different parameters of the pentagons drawing, such as number of angles, distance/intersection, closure/opening, rotation, closing-in, and to verify its efficacy to differentiate dementia with Lewy Body (DLB) from Alzheimer's disease (AD). We established the reliability of the qualitative scoring method through the inter-raters and intra-subjects analysis. QSPT was then applied to forty-six AD and forty-six DLB patients, using two phases statistical approach, standard and artificial neural network respectively. DLB patients had significant lower total score in the copy of pentagons and number of angles, distance/intersection, closure/opening, rotation compared to AD. However the logistic regression did not allow to establish any suitable modeling, whereas using Auto-Contractive Map (Auto-CM) the DLB was more strongly associated with low scores in some qualitative parameters of pentagon copying, i.e. number of angles and opening/closure and, for the remaining subitems of the MMSE, in naming, repetition and written comprehension, and for demographic variables of gender (male) and education (6-13 years). Twist system modeling showed that the QSPT had a good sensitivity (70.29%) and specificity (78.67%) (ROC-AUC 0.74). The proposed qualitative method of assessment of pentagons copying used in combination with non-linear analysis, showed to be consistent and effective in the differential diagnosis between Lewy Body and Alzheimer's dementia.