Background and aims: The gut has a relevant role in the pathophysiology of type 2 diabetes. Different intestinal tracts harbour different microbiota communities, and fecal characterisation cannot map the microbiota of the upper intestine. Sorbitol is utilised by a variety of bacteria. The aim of this study was to use positron emission tomography (PET) imaging of 2-deoxy-2-[18F]fluoro-d-sorbitol (18FDS) to visualize gut microbiota in vivo in relation to glycemia. Material and methods: The study was conducted in 3 groups of C57BL/6 mice, undergoing: a) no treatment (n=12), b) treatment with oral antibiotics (abx) to deplete the gut microbiota (n=12), or c) with antibiotics followed by probiotics (n=8). 18FDS was administered orally for microbiota binding, and its distribution was monitored over 5 hours by PET imaging under fasting conditions or glucose loading. Radioactivity was measured in the stomach and caecum, and compared with ex vivo 16S-RNA sequencing data, including total bacterial amplicon sequence variants (ASV). Results: In untreated mice, both in vivo (18FDS: 12.6±1.9 vs 4.1±1.1 MBq/cc p=0.005) and ex-vivo (ASV: 11594±1104 vs 6373±916 p=0.004) bacteria signals were higher in caecum than stomach. These differences were not seen in abx treated and probiotic mice (n.s.). In the caecum, 18FDS accumulation was lower in abx (p=0.049) and probiotic (p=0.016) compared to untreated mice, and it was correlated with total ASV counts (R=0.38, p=0.037), whereas no such findings were observed in the stomach. Caecal 18FDS activity was associated with gram-negative bacteria counts, including Bacteroidetes (R=0.32, p trend=0.078), Muribaculaceae (R=0.35, p=0.050), Prevotella (R=0.45 p=0.013) and Oscillibacter (R=0.41 p=0.025). Conversely, Bifidobacterium and Streptococcus counts were increased in the probiotic compared to abx and untreated groups (p<0.001), but did not reflect the 18FDS signal. Fasting glycemia was lower in probiotic treated than untreated mice (8.1±0.6 vs 9.5±0.4 mmol/L, p=0.05); glucose-loading glycemia was correlated with caecal 18FDS signal at 3 (R=0.56, p=0.01) and 5 hours (R=0.45, p=0.055). Conclusion: Imaging of orally administered 18FDS by PET seems a promising approach to quantify specific microbial features, mostly belonging to gram-negative taxa, across different tracts of the intestine in vivo, associating with glucose levels.