AbstractCerebral small vessel disease is a heterogeneous disease in which various underlying etiologies can lead to different types of white matter hyperintensities (WMH). WMH shape features might aid in distinguishing these different types. In this proof of principle study in patients with type 2 diabetes mellitus (T2DM), we present a novel approach to assess WMH using shape features. Our algorithm determines WMH volume and different WMH shape and location features on 3T MRI scans. These features were compared between patients with T2DM (n = 60) and a matched control group (n = 54). Although a more traditional marker (WMH volume) was not significantly different between groups (natural log transformed Beta (95% CI): 0.07 (−0.11↔0.24)), patients with T2DM showed a larger number of non-punctuate WMH (median (10^th–90^th percentile), patients: 40 lesions per person (16–86); controls: 26 (5–58)) and a different shape (eccentricity) of punctuate deep WMH (Beta (95% CI): 0.40 (0.23↔0.58)) compared to controls. In conclusion, our algorithm identified WMH features that are not part of traditional WMH assessment, but showed to be distinguishing features between patients with T2DM and controls. Future studies could address these features to further unravel the etiology and functional impact of WMH.