A comprehensive workflow for CCTA and OCT data fusion with 3D printing validation: advancing patient-specific testing environments for percutaneous coronary intervention devices

Fig. 1. The red segment denotes data derived from OCT, while the blue segment corresponds to data obtained exclusively from CCTA. The reconstructed vessel wall is shown in transparent red. a Right coronary artery of patient 12 with a detailed view of the stenosis. b Left coronary artery of patient 7 with calcified plaque highlighted in yellow

Fig. 2. Superimposed calcium contours from CCTA and OCT illustrating selected frames for a patient 9, b patient 4, and c patient 7

Fig. 3. Linear regression analysis comparing a the OCT-derived area stenosis (%AS) and b minimum lumen area (MLA) determined for each patient and corresponding phantom

Fig. 4. a Superimposed original model (OM) and derivative model (DM) of patient 12 including heatmap to indicate vertex distance (VD) between the two models. b z-score normalized lumen area comparison

Fig. 5. Workflow for generating high-resolution 3D models of coronary arteries (left) and processing of OCT images to obtain vessel contours (right): a semi-automatic lumen delineation, b binary mask generation, c convex hull application, d B-spline fitting, and e conversion from pixel to physical space

Fig. 6. Custom-built software for the fusion of OCT and CCTA images, featuring a GUI. a Main dialog box for selecting processing parameters and providing user instructions. b Toolset for centerline adjustment. c Visualization and manipulation of the fused OCT-CCTA data. d Optional transitional outputs to verify intermediate results and data alignment

Fig. 7. The 2D OCT frame displaying clearly defined calcification borders was registered to the 3D calcification model reconstructed from the CCTA. The intersection of the two datasets was used to generate a 2D cross-section of the CCTA-derived calcification

Fig. 8. Comparison of CCTA (left) and OCT (right) imaging of the lumen in the right coronary artery of patient 12. OCT provides higher resolution and clearer delineation of the lumen boundary compared to CCTA

Fig. 9. Merged point clouds of the CCTA (blue) and OCT (red) before (left) and after (right) removal of overlapping points

Fig. 10. Schematic illustration of the validation process: generation of a DM followed by superimposition of the OM and DM to evaluate vertex distance and lumen area

Fig. 11. Conceptual registration between OCT and CCTA using the carina point of a bifurcation, identifiable in both a the OCT image and b the CCTA. c In our software, the projections of the registration point onto the main and side branches were manually selected and adjusted within the graphical user interface. These projections were then used to compute the actual carina point and achieve precise alignment between OCT and CCTA

Fig. 12. a Known ground truth of a simplified coronary artery model. b A reduced point cloud with spatial resolution similar to photon-counting CCTA. c 3D printed negative of the lumen