UK Biobank imaging adapted for a real‐world memory clinic sample

Abstract

AbstractBackgroundThe Oxford Brain Health Centre (BHC) provides real‐world research data, aligned with the UK Biobank imaging study (UKB), routinely captured in a clinical service from patients under‐represented in dementia research. In this work we present how we 1) adapted the UKB MRI acquisition protocol to be suitable for memory clinic patients, 2) modified the imaging analysis pipeline to extract measures that are in line with radiology reports and 3) compare measures from BHC patients to the biggest brain MRI study in the world (ultimately 100,000 participants).MethodAdaptations of the UKB acquisition protocol for BHC patients include dividing the scan into core and optional sequences to improve patients’ tolerance. We adapted the UKB structural MRI analysis pipeline to extract quantitative measures from the scans, taking into account the characteristics of a memory clinic population (e.g. high level of white matter hyperintensities, WMH, causing misclassifications in tissue‐type segmentation). We compared the measurements extracted from the scans to visual ratings from radiology reports and to reference distributions derived from UKB data.ResultOf the 108 BHC attendees since August 2020 (average age 78.5 years, range 65‐101 years, 50.9% female), 100 (92.6%) completed the clinical scans, 100 (92.6%) consented to use of data for research, and 69 (63.9%) consented to additional research sequences, demonstrating that the protocol is well tolerated. Modified tissue‐type segmentation improved grey matter volume estimation, which showed stronger correlation with visual ratings of global atrophy after correction (r=‐0.34; p=0.001) than with the original analysis pipeline (r=‐0.31; p=0.002). Correlations between hippocampal volumes and MTA scale (r(L)=‐0.59, r(R)=‐0.71) and between WMH volume and Fazekas scale (r=0.78) were also significant (p<0.001). Because UKB participants are younger than memory clinic patients, additional scans on elderly healthy controls are planned to improve reference distributions to detect pathological deviations. Measures from BHC patients within UKB age range are comparable to those from UKB participants, with several patients falling in the lower percentiles.ConclusionWe adapted UKB research imaging for clinical practice. Current and future work aims to integrate automated quantitative measures in the radiology reports and evaluate their clinical utility.