4D Flow MRI data of a reference aneurysm model with an inserted commercial flow-diverter and combined 3D-printed aneurysm-flow-diverter models

This repository contains eight in vitro 4D Flow MRI datasets acquired from four aneurysm models. Each model was imaged twice using different velocity encoding (venc) settings, resulting in two datasets per model. One aneurysm model represents the untreated case, while the remaining three correspond to the same aneurysm after treatment with flow diverter stents.

The datasets were used in the publication "Development of 3D-printed flow diverting implants for studying the effect of aneurysm treatment in vitro," published in Annals of 3D Printed Medicine, DOI: 10.1016/j.stlm.2025.100196 (Bautz, 2025)

In the paper, we utilized 4D Flow MRI to assess flow conditions in 3D-printed patient-specific aneurysm models induced by flow diverter stents. Three different flow diverters were studied, including a commercially available (DERIVO Embolisation Device, ACANDIS GmbH, Pforzheim, Germany) and two 3D-printed stents with a number of wires of 6 and 16. The 3D-printed stents were designed and fabricated in-house using the developed protocol described in the paper (Bautz, 2025).

Please cite the associated paper when using the data in your work (Bautz, 2025).

__________________________________________________________________________

Investigated aneurysm models

A digital model of a paraophthalmic aneurysm of the left internal carotid arteries was designed out of clinical 3D rotational angiography data (x-ray angiography system, Allura Xper FD20/10 biplane, Philips Healthcare, Best, Netherlands). Segmentation and simplification of the vascular system and the necessary extensions to create flow models were carried out according to a workflow previously published by our group (Pravdivtseva, 2021). The aneurysm models were produced using stereolithography 3D printing (Form 3B, Formlabs, Somerville, Massachusetts, USA).

Two models were printed empty, while two more contained mesh-like structures covering the aneurysm orifice and thereby mimicking flow diverters. The design of the flow diverter replicas comprises these main steps: 1) reconstruction of the healthy vessel, 2) design of the braided stent structure, and 3) integration into the digital patient-specific vasculature. More specific design descriptions can be found in the associated paper (Bautz, 2025) and manual (DOI: 10.5281/zenodo.14987034). All models were equipped with connectors and MRI-visible makers to ensure reproducible placement of the MR imaging volume and 2D planes. Before MRI, the models were placed in containers filled with agarose gel

In total, four models were investigated with 4D Flow MRI: M1-empty without any flow diverter and was used to assess flow before treatment, M2-commercial had an inserted commercial flow diverter. The models with a 3D-printed flow diverter were distinguished by the number of wires: M4-printed-16 contained 16 wires, and M5-printed-6 contained six wires. The digital aneurysm models finalized for 3D printing (M1-empty, M4-printed-16 and M5-printed-6) can be downloaded here (DOI: 10.5281/zenodo.14987034).

__________________________________________________________________________

Magnetic Resonance Imaging 

MR Imaging was performed on a clinical 3T MR system equipped with a 32-channel head coil (Ingenia CX, R5 V6.1, Philips Healthcare, Best, Netherlands). The aneurysm models were integrated into a flow pump setup and supplied with a flow rate of 4.2ml/s at a constant period of 0.8s (PD-1100, BDC Laboratories, Wheat Ridge, Colorado, USA). The imaging protocol used consisted of three angiographic pulse sequences: Time-of-Flight (ToF), 2D Phase-Contrast (2D PC), and 4D Flow MRI. 

4D Flow MRI was used to measure velocities within the 3D imaging volume covering the aneurysm sac and parental vessel. Due to different flow conditions in the aneurysm and parental vessel, two 4D Flow sequences with different venc values were performed per model. To accurately acquire velocities in the parental vessel, venc values of 80cm/s (M1-empty), 100cm/s (M2-commercial), and 50cm/s (M4-printed-16/M5-printed-6) were used. To minimize the noise in the aneurysm sac, venc values of 55cm/s (M1-empty), 20cm/s (M2-commercial), 5cm/s (M4-printed), and 25cm/s (M5-printed-6) were chosen. The optimal venc values were determined by acquiring a series of 2D phase-contrast datasets with varying venc values.

4D Flow MRI is based on a 3D spoiled gradient-echo sequence and was accelerated with the compressed-sensing technique (Philips). The 4D Flow MRI sequence comprised the following parameters: echo time (TE) /repetition time (TR) – 5.0/8.3ms; field of view – (110 x 100 40.125) mm3; voxel size (0.75 x 0.75 x 0.75) mm3; acceleration factor – 4.5. A number of 20 cardiac phases were acquired in a total scan time of about 26min.

References

Bautz, L. et. al. Development of 3D-printed flow-diverting stents for studying the effect of aneurysm treatment in vitro. Ann. 3D Print. Med. (2025) DOI: 10.1016/j.stlm.2025.100196 

Pravdivtseva, M. S. et al. 3D‐printed, patient‐specific intracranial aneurysm models: From clinical data to flow experiments with endovascular devices. Med. Phys. 48, 1469–1484 (2021). DOI: 10.1002/mp.14714