Differential diagnosis of dementia combining web‐based cognitive testing and MRI: Neuropsychology/computerized neuropsychological assessment. (7th December 2020)

• Record Type: Journal Article
• Authors: Rhodius‐ Meester, Hanneke F.M.; Koikkalainen, Juha; Paajanen, Teemu; Mahdiani, Shadi; Barkhof, Frederik; Herukka, Sanna‐Kaisa; Hänninen, Tuomo; Ngandu, Tiia; Kivipelto, Miia; van Gils, Mark; Hasselbalch, Steen Gregers; Mecocci, Patrizia; Remes, Anne; Soininen, Hilkka; Scheltens, Philip; van Der Flier, Wiesje; Lötjönen, Jyrki
• Is Part Of: Alzheimer's & dementia. Volume 16(2020)Supplement 6
• Page Start: n/a
• Page End: n/a
• Publication Date: 2020-12-07
• Subjects: Alzheimer's disease -- Periodicals; Alzheimer Disease -- Periodicals; Dementia -- Periodicals; Démence; Maladie d'Alzheimer; Périodique électronique (Descripteur de forme); Ressource Internet (Descripteur de forme); 616.83
• DOI: 10.1002/alz.042626
• Languages: English
• ISSNs: 1552-5260

Abstract

Abstract: Background: Differential diagnostics in dementia is challenging. To date, the basic assessment still includes imaging of the brain and cognitive testing with pen and paper. Web‐based cognitive tests however hold potential for standardized and low‐cost screening in clinical workup. How they perform when combined with imaging of the brain is unknown. We therefore evaluated the accuracy of a new web‐based cognitive battery (Muistikko [1]) detecting different types of dementia, when combined with brain MRI, and compared this to traditional cognitive testing and MRI. Method: We included 229 subjects from two memory clinic cohorts (PredictND and VPH‐DARE), consisting of 188 controls, 29 patients with Alzheimer's dementia (AD), 7 with frontotemporal dementia (FTD) and 5 with vascular dementia (VaD) (Table 1). All patients performed a traditional cognitive test battery (consisting of MMSE, RAVLT, TMT‐A and B, Animal Fluency), web‐based cognitive testing and had MRI of the brain. Although Muistikko is composed of seven subtasks, only global cognitive score (GCS) was used as defined in [1]. From MRI, multiple imaging biomarkers were defined [2]. Disease‐state index classifier was developed from the predictors [2]. Cross‐validation was used to calculate balanced accuracy (BACC; average of sensitivities for each diagnostic group). Given the class imbalance, we also calculated prevalence corrected accuracy (PACC). Result: BACC was 66 % and PACC 64% when using the traditional cognitive test battery + MRI. Both BACC and PAC were 69 % when using the web‐based cognitive testing + MRI (Table 2). Of note, since we compare four diagnostic groups, BACC by guessing would be 25%. Conclusion: This study shows that combining web‐based cognitive tests with MRI data results in high accuracy when separating different types of dementia. The results were comparable with the standard traditional work‐up. Web‐based cognitive testing is therefore a promising tool to support the clinician in the daily challenge of differential diagnostics, especially when combined with MRI data. References: [1] Paajanen, S. et al. Detecting cognitive disorders using Muistikko web‐based cognitive test battery. Alzheimer's & Dementia 13(7):Supplement, P234‐P235, 2017 [2] Bruun, M. et al. Evaluating combinations of diagnostic tests to discriminate different dementia types. Alzheimers Dement 2018 Aug17;10:509‐51.