Biomarkers
Bhalerao GV., Markiewicz PJ., Thomas DL., de Vita E., Parkes L., Thompson G., MacKewn J., Krokos G., Wimberley C., Hallett W., Su L., Smith S., Malhotra P., Hoggard N., Taylor JP., Ritchie C., Wardlaw JM., Matthews PM., Aigbirho FI., O'Brien JT., Hammers A., Fox NC., Herholz K., Barkhof F., Miller K., Matthews J., Griffanti L.
BACKGROUND: The DPUK PET-MR Harmonisation Study aims to quantify within-site and across-site variability of brain scans across scanning equipment. Brain images were acquired from 8 PET-MR scanners to optimise clinical trial design and evaluate PET-MR methods. Scans were collected from healthy elderly participants in three groups: Repeated, Intra-manufacturer, and Inter-manufacturer (Table 1). This study examines variability in imaging-derived phenotypes (IDPs) from T1-weighted MRI scans and evaluates harmonisation methods. METHOD: All T1w scans were processed using the UK Biobank (UKB) pipeline to extract volumetric IDPs (total brain, tissue-specific, hippocampus). Within-subject (scanner) variability was quantified using the Coefficient of Variation (CoV) for each subject pair and IDP. In the inter-manufacturer group, CoV and cross-subject variability (biological variability measured as across subjects' standard deviation of within-subject mean IDP) were assessed following three harmonisation approaches: IQM-based: Identify image quality metrics (IQMs) (e.g., noise, contrast) that significantly differ across scanners and exhibit low correlation (-0.5