Premium
P1‐175: CSF and Plasma Biomarkers for Diagnosing Dementia in Outpatient Clinic
Author(s) -
Orikasa Masayuki,
Kawarabayashi Takeshi,
Wakasaya Yasuhito,
Nakamura Takumi,
Nakahata Naoko,
Shoji Mikio
Publication year - 2016
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2016.06.923
Subject(s) - dementia , medicine , neurology , cerebrospinal fluid , alzheimer's disease , degenerative disease , disease , gastroenterology , pathology , psychiatry
Background:Not necessary all diagnostic test data used in neurodegenerative diseases are needed to reach high diagnostic accuracy for all patients. A simple battery of tests may provide enough information for accurate diagnosis in some cases while a complete test battery may be needed in other more challenging cases. The objective of this work is to study whether a systematic stratified diagnostic workflow could enable more cost-efficient diagnostics. Methods: The data from five diagnostics groups were used (Table 1). The data consisted of clinical and neuropsychological test data, CSF-biomarkers and MRI-data (T1 & FLAIR) quantified using both visual ratings and automatically computed imaging biomarkers. For the stratified approach, the data were grouped into five batteries and a cost was given to each battery: NP-S (MMSE, memory tests, fluency tests and trail makings tests, 100 V), NP-F (full neuropsychological battery, NP-S being a subset, 300 V), MRI-V (MRI data acquisition including visual scorings, 400 V), MRI-A (additional imaging biomarkers from six quantification tools, 80 V) and CSF-biomarkers (300 V). Figure 1 shows a schematic presentation of the stratified workflow. The disease-state index (DSI) method, extended to differential diagnostics, was used to assess patients’ fit to previously diagnosed cases in the database [1]. A patient was labeled as ‘clear’ if DSI was higher than a given threshold for a certain combination of batteries. The threshold producing the accuracy>85% for the clear cases was chosen. No extra cost was given to wrong diagnosis or to undiagnosed cases needing follow-up. Cross-validation was used in validation. Results: Table 2 and Figure 2 show the results. The best sequence was when stratification was used with automatically computed imaging biomarkers: the cost per patient was 741 V, 73.6 % of all cases were labeled as clear and the accuracy for the clear cases was 85.7 %. Conclusions: The results show that the stratified approach has potential to reduce costs in diagnostics. In addition, imaging biomarkers from automated tools can be beneficial. [1] J. Mattila et al., Disease State Fingerprint for Evaluating the State of Alzheimer’s Disease in Patients. Journal of Alzheimer’s Disease 27: 163-176, 2011.