Evaluation of Computed Tomography Head scan reporting on Brain Frailty Markers in an Acute Medical Cohort

4424

A Yu1; J Gan2; S T Pendlebury2,3

1. Oxford (Thames Valley) Foundation School, Oxford University Hospitals NHS Foundation Trust, Oxford, UK; 2. Wolfson Centre for Prevention of Stroke and Dementia, Nuffield Department of Clinical Neuroscience, University of Oxford, Oxford, UK; 3. Departme

Scientific Research

SP - Neurology & Neuroscience

Abstract

Introduction

Computed tomography head (CTH) scans are commonly used in the acute medical setting to detect acute intracranial pathology. CTH scan information on cerebrovascular burden and atrophy, key markers of brain frailty, could also be used to identify patients at risk of cognitive decline. In an observational study, we assessed reporting brain frailty markers in routinely acquired scans.

Method

Anonymised reports of CTH scans performed during or <1-year of index admission were retrieved on consecutive acute medicine patients aged >65 years admitted over six 8-week periods (2010-2018). Information on small vessel disease (SVD), cerebral atrophy, and acute/old stroke were manually extracted from scan reports, including severity of SVD and global/regional atrophy categorised as none/mild/moderate/severe or using visual rating scales (Fazekas for SVD, Global Cerebral Atrophy-GCA Scale, medial temporal atrophy-MTA).

Results

Among 1,146 patients, CTH reports were retrieved from 557 (49%) patients (female=291, mean/SD age=81.7/8.1 years). SVD was mentioned in 378/557 (67.9%) reports, of which 245 (64.8%) quantified disease as absent=2.4%, mild=37.6%, moderate=35.9%, severe=25.0%. Only one used the Fazekas scale. Atrophy was mentioned in 391/557 (70.2%) reports and described as “age-appropriate” in 221 (56.8%) and suggestive of Alzheimer’s dementia in 10 (3%). Severity was graded in 117/391 (30.0%) as none=1.7%, mild=43.6%, moderate=27.4%, severe=25.6%. None used the GCA, three used the MTA. Old stroke(s) were reported in 154/557 (28%) and new strokes in 35 (6%). In 14 scans, reporting was limited owing to structural abnormalities (e.g., mass, stroke) or movement artefact.

Conclusion

The majority (70%) of CTH scans reported SVD/atrophy, but only two-thirds of SVD and one-third of atrophy were quantified using broad severity categories. Visual rating scales were almost never used. Exploitation of currently under-used brain frailty information contained in CTH scans will require automated brain image analysis tools providing standardised quantification without the need for time-consuming human input.