Analyzepro 1

Teeth samples were analyzed using a μCT 50 scanner (Scanco Medical) at 70 kVp, 76 μA, 0.5 Al filter, 900‐ms integration, and 10‐μm voxel size. AnalyzePro 1.0 (AnalyzeDirect) was used to reconstruct and analyze images. Images were anatomically oriented and calibrated to a linear standard curve created using hydroxyapatite phantoms with known densities. Enamel was segmented ≥1650 mg HA/cm³, and dentin/cementum at 600–1649 mg HA/cm³. HFTC dentin and cementum were indistinguishable by density, so dentin and cementum were combined for analysis. Pulp was segmented using a combination of manual and semiautomatic functions. Heat maps were generated for density visualization.

The 3D clinical rotational angiography data was acquired using an angiography system (Allura Xper FD 20/20, Philips Healthcare, Best, Netherlands) with the following acquisition parameters: 5 s acquisition time, 220° rotation, 150 individual images at a frame rate of 30/s, $(15\times 48)$  cm² detector FOV with a $512\times 512$ pixel acquisition matrix. The images were reconstructed with a soft tissue kernel of isotropic voxel size (edge length 0.238 mm).
Segmentation was performed with Analyze Pro 1.0 (AnalyzeDirect, Overland Park, Kansas, USA), while a semi-automated approach using a region growth algorithm was used. For this purpose, seed areas were defined at different points in the vessels. Small perforator arteries were removed if the diameter was smaller than 0.1 mm. The segmented anatomy was converted into stereolithography (.stl) format. A rectangular box with a size of $(36.5\times 13.1\times 22.2)$ mm³ was designed to enclose the target vessels of the anatomy and an adapter for the sample holder of the MPI scanner was attached to the model. Printing was performed on a stereolithography printer (Form2, Formlabs Inc.) using a 25 µm layer thickness and clear photopolymer resin (Formlabs Inc.). After the printing process and post-processing steps, the application of clear coat was performed⁵⁰ .