[18F]flutemetamol uptake in the colon of a memory clinic population and its association with brain amyloidosis and the gut microbiota profile: an exploratory study

Abstract

Purpose: Some Alzheimer's disease (AD) patients report gastro-intestinal symptoms and present alterations in the gut microbiota (GM) composition. Elevated colonic amyloid immunoreactivity has been shown in patients and animal models. We evaluated the colonic uptake of the amyloid positron emission tomography (PET) imaging agent [18F]flutemetamol (FMM) in a memory clinic population and investigated its association with brain amyloidosis and GM composition.

Methods: Forty-five participants underwent (i) abdominal and cerebral FMM PET, acquired at 40 (early phase) and 120 min (late phase) after tracer injection, (ii) abdominal computed tomography, and (iii) cerebral T1-weighted MRI. Colonic standardized uptake value ratio (SUVr) was determined through manual tracing and automatic segmentation (TotalSegmentator), using the aortic blood signal as a reference region. Fecal GM composition was assessed using 16 S rRNA sequencing. Amyloid positive (A+) and negative (A-) participants, based on cortical FMM quantification (PetSurfer), were compared in terms of SUVr and GM features.

Results: Increased colonic early SUVr was reported in A+ than A- (manual, p =.008; automated, p =.035). Altered GM composition was found in A + as shown by lower Pielou's evenness (p =.023), lower abundance of Eubacterium hallii group, and higher abundance of several genera. High UC5-1-2E3 abundance positively correlated with high colonic early SUVr (whole group: manual, p =.012, automated, p =.082; A+: manual, p =.074; automated, p =.016).

Conclusion: This exploratory study showed that subjects with cerebral amyloidosis have greater colonic FMM uptake than subjects with normal cerebral amyloid load, correlating with altered GM composition. Further analysis is needed to determine if these changes denote amyloid-related changes or other phenomena.

Keywords: Alzheimer’s disease; Amyloid PET; Colon; Gut microbiota.

Fig. 1

Manual and automated segmentation procedures of volumes of interest. Manual procedure (panel A): volumes of interest for the sigmoid wall and aorta bloodstream were manually drawn on CT. Automated procedure (panel B): (i) aorta (green) and colon (cyan) binary masks were automatically computed (panel B.1); (ii) CT images were resliced to PET space (panel B.2); (iii) the inverse transformation was applied to aorta and colon binary masks to match PET resolution, and the uptake values were computed (panel B.3)

Fig. 2

Amyloid pathology status in the brain (negative [A-] / positive [A+]) on [18 F]Flutemetamol (FMM) mean and maximum SUVr of sigmoid wall (manual tracing; panel A) and colon (automated segmentation; panel B). Early and late PET phases (e-FMM and l-FMM, respectively) were tested. Each dot represents a subject, whiskers denote the values’ range, the horizontal line within boxes is the group median. Significant results are reported in bold.

Fig. 3

Amyloid pathology status in the brain (negative [A-] / positive [A+]) and alpha diversity measured by Pielou’s index (panel A) and on fecal microbiota composition at the genus level (panel B, linear discriminant analysis (LDA) effect size (LEfSe) algorithm, LDA > 3). Association matrix (panel C) between [18F]flutemetamol mean SUVr in the gut based on manual tracing (sigmoid) and automated segmentation (colon) with MMSE, cerebral mean SUVr and fecal microbiota features.