Background Alzheimer's disease and its complications are the leading cause of death in adults with Down syndrome. Studies have assessed Alzheimer's disease in individuals with Down syndrome, but the natural history of biomarker changes in Down syndrome has not been established. We characterised the order and timing of changes in biomarkers of Alzheimer's disease in a population of adults with Down syndrome. Methods We did a dual-centre cross-sectional study of adults with Down syndrome recruited through a populationbased health plan in Barcelona (Spain) and through services for people with intellectual disabilities in Cambridge (UK). Cognitive impairment in participants with Down syndrome was classified with the Cambridge Cognitive Examination for Older Adults with Down Syndrome (CAMCOG-DS). Only participants with mild or moderate disability were included who had at least one of the following Alzheimer's disease measures: apolipoprotein E allele carrier status; plasma concentrations of amyloid β peptides 1-42 and 1-40 and their ratio (Aβ 1-42/1-40 ), total tau protein, and neurofilament light chain (NFL); tau phosphorylated at threonine 181 (p-tau), and NFL in cerebrospinal fluid (CSF); and one or more of PET with ¹⁸F-fluorodeoxyglucose, PET with amyloid tracers, and MRI. Cognitively healthy euploid controls aged up to 75 years who had no biomarker abnormalities were recruited from the Sant Pau Initiative on Neurodegeneration. We used a first-order locally estimated scatterplot smoothing curve to determine the order and age at onset of the biomarker changes, and the lowest ages at the divergence with 95% CIs are also reported where appropriate.

These results support a biphasic model of changes in AD, which could affect the selection of patients for clinical trials and the use of magnetic resonance imaging as a surrogate marker of disease modification.

No abstract

Cortical mean diffusivity has been proposed as a novel biomarker for the study of the cortical microstructure in Alzheimer's disease. In this multicentre study, we aimed to assess the cortical microstructural changes in the behavioural variant of frontotemporal dementia (bvFTD); and to correlate cortical mean diffusivity with clinical measures of disease severity and CSF biomarkers (neurofilament light and the soluble fraction beta of the amyloid precursor protein). We included 148 participants with a 3 T MRI and appropriate structural and diffusion weighted imaging sequences: 70 patients with bvFTD and 78 age-matched cognitively healthy controls. The modified frontotemporal lobar degeneration clinical dementia rating was obtained as a measure of disease severity. A subset of patients also underwent a lumbar puncture for CSF biomarker analysis. Two independent raters blind to the clinical data determined the presence of significant frontotemporal atrophy to dichotomize the participants into possible or probable bvFTD. Cortical thickness and cortical mean diffusivity were computed using a surface-based approach. We compared cortical thickness and cortical mean diffusivity between bvFTD (both using the whole sample and probable and possible bvFTD subgroups) and controls. Then we computed the Cohen's d effect size for both cortical thickness and cortical mean diffusivity. We also performed correlation analyses with the modified frontotemporal lobar degeneration clinical dementia rating score and CSF neuronal biomarkers. The cortical mean diffusivity maps, in the whole cohort and in the probable bvFTD subgroup, showed widespread areas with increased cortical mean diffusivity that partially overlapped with cortical thickness, but further expanded to other bvFTD-related regions. In the possible bvFTD subgroup, we found increased cortical mean diffusivity in frontotemporal regions, but only minimal loss of cortical thickness. The effect sizes of cortical mean diffusivity were notably higher than the effect sizes of cortical thickness in the areas that are typically involved in bvFTD. In the whole bvFTD group, both cortical mean diffusivity and cortical thickness correlated with measures of disease severity and CSF biomarkers. However, the areas of correlation with cortical mean diffusivity were more extensive. In the possible bvFTD subgroup, only cortical mean diffusivity correlated with the modified frontotemporal lobar degeneration clinical dementia rating. Our data suggest that cortical mean diffusivity could be a sensitive biomarker for the study of the neurodegeneration-related microstructural changes in bvFTD. Further longitudinal studies should determine the diagnostic and prognostic utility of this novel neuroimaging biomarker.

One outstanding challenge for machine learning in diagnostic biomedical imaging is algorithm interpretability. A key application is the identification of subtle epileptogenic focal cortical dysplasias (FCDs) from structural MRI. FCDs are difficult to visualize on structural MRI but are often amenable to surgical resection. We aimed to develop an open-source, interpretable, surface-based machine-learning algorithm to automatically identify FCDs on heterogeneous structural MRI data from epilepsy surgery centres worldwide. The Multi-centre Epilepsy Lesion Detection (MELD) Project collated and harmonized a retrospective MRI cohort of 1015 participants, 618 patients with focal FCD-related epilepsy and 397 controls, from 22 epilepsy centres worldwide. We created a neural network for FCD detection based on 33 surface-based features. The network was trained and cross-validated on 50% of the total cohort and tested on the remaining 50% as well as on 2 independent test sites. Multidimensional feature analysis and integrated gradient saliencies were used to interrogate network performance. Our pipeline outputs individual patient reports, which identify the location of predicted lesions, alongside their imaging features and relative saliency to the classifier. On a restricted ‘gold-standard’ subcohort of seizure-free patients with FCD type IIB who had T1 and fluid-attenuated inversion recovery MRI data, the MELD FCD surface-based algorithm had a sensitivity of 85%. Across the entire withheld test cohort the sensitivity was 59% and specificity was 54%. After including a border zone around lesions, to account for uncertainty around the borders of manually delineated lesion masks, the sensitivity was 67%. This multicentre, multinational study with open access protocols and code has developed a robust and interpretable machine-learning algorithm for automated detection of focal cortical dysplasias, giving physicians greater confidence in the identification of subtle MRI lesions in individuals with epilepsy.