An exploratory pilot study on diffusion MRI measurement of cortical microstructural neurodegeneration for the assessment of dementia in a Japanese hospital setting

Abstract

AbstractBackgroundThe use of diffusion MRI (dMRI) to assess cortical microstructure has demonstrated promise for measuring neurodegeneration in Alzheimer’s disease (PMID:33174658, PMID:36281682) and frontotemporal dementia (PMID:32641807, PMID:34686217) using retrospective research datasets. The ultimate translation of research findings to clinical practice requires prospective studies in hospital settings. A pilot study at the National Hospital Organization, Niigata, was designed and successfully delivered.MethodThe study was registered on Oct 5th 2021 (jRCT1032210367). Scanning took place between Oct 13th 2021 and June 30th 2022, on a 1.5T Philips Ingenia. MRI data was transferred using a secure cloud gateway (Cimar). A total of 10 healthy controls (HC), 10 Alzheimer’s disease (AD) patients, and 5 frontotemporal dementia (FTD) patients were recruited. Structural (3D T1‐weighted) and dMRI (2mm isotropic, 32 directions at b = 1000 s/mm2) scans were used to calculate cortical mean diffusivity and three neuropathologically‐inspired cortical diffusivity measures: the angle between the radial minicolumnar direction and the principal diffusion direction (AngleR); the principal diffusion component parallel with the minicolumns (ParlPD), and the diffusion components perpendicular to the minicolumns (PerpPD+). Metrics were also summarized over AD signature regions (Ridgway et al., CTAD 2022).ResultOf the n = 25 cases, 18 had clear clinical diagnoses after neurological and radiological assessment, while 7 were more complex (including focal bleeding, focal ischemia, meningioma, supra‐sellar tumour). Characteristics are given in Table 1. Figure 1 plots two whole‐brain measures: PerpPD+ and AngleR. A threshold of PerpPD+ > 1.325 has 100% sensitivity to the 8 clear AD cases, with 86% specificity for 7 controls vs AD. Results were less clear for FTD using whole‐brain measures alone. The Cohen’s d effect size for clear AD versus clear HC was 2.82 using whole‐brain PerpPD+. Using signature regions (Figure 2b) improved separation to d = 3.08 for PerpPD+, compared to d = 1.87 for the cortical thickness signature (Figure 2a).ConclusionThe pilot study demonstrated practical feasibility and good performance for distinguishing AD from HC using cortical diffusivity measures acquired at the 1.5T field‐strength common in hospitals in Japan. Future work will investigate FTD signature regions. A larger multicentre study assessing diagnostic accuracy using additional AD biomarkers is warranted.