P1-309: agreement, sensitivity, and specificity between the neuropathological and clinical diagnoses of alzheimer's disease (ad) dementia: a multicenter longitudinal study

The definite diagnosis of AD dementia relies on neuropathology. The literature reports an agreement of 83-93% between clinical and neuropathological diagnosis. In vivo markers could improve the diagnosis accuracy.Objectives: To analyze the agreement between the clinical and the neuropathological diagnosis & to establish the sensitivity and specificity of diagnostic markers. A cohort of 604 patients is being followed (“cohort FRA”) annually with clinical follow-up including neuropsychological standardized evaluation, MRI and molecular imaging. So far 76 patients deceased, of which 39 accepted the postmortem brain neuropathological exam. We reviewed their pre-mortem and the neuropathological data. The agreement for the clinical and neuropathological diagnosis was subsequently established and the agreement between each exam (neuropsychological evaluation, CSF biomarkers, MRI, SPECT/PET) and the clinical and neuropathological diagnosis. The sensitivity and specificity of each marker was also calculated. The mean age was 60,0 ±10,2 years at dementia onset. They died at a mean age of 69,5±10,3 years. The prevalence of AD pathology was (31/39) 79,5%. AD patients had at least another pathological finding among which the two more frequent were amyloid angiopathy (65 %) and Lewy body accumulation (45 %, half diffuse). The agreement between clinical and neuropathological diagnosis was 93,3% with a sensitivity and specificity of 100% for the diagnosis of AD dementia and the agreement between neuropsychological evaluation and both clinical and neuropathological diagnosis was 96% notwithstanding associated pathology (Lewy pathology was clinically underestimated). For CSF biomarkers assay, the agreement with both clinical and neuropathological diagnosis was 100%. For the patients who underwent functional imaging (SPECT/PET), the exam had an agreement of 90,0% for the clinical diagnosis and of 82,6% for the neuropathological diagnosis. MRI in this sample was unhelpful in 40-50% of the cases. The agreement, sensitivity and specificity between clinical and neuropathological diagnoses in this sample were high even though the cohort was multicentric and the mixed pathology prevalence elevated. The exams that have proven more useful were the extensive and early neuropsychological evaluation and the biomarkers assay in CSF.