3D‐printed, patient‐specific intracranial aneurysm models: From clinical data to flow experiments with endovascular devices. Issue 4 (17th February 2021)

• Record Type: Journal Article
• Authors: Pravdivtseva, Mariya S.; Peschke, Eva; Lindner, Thomas; Wodarg, Fritz; Hensler, Johannes; Gabbert, Dominik; Voges, Inga; Berg, Philipp; Barker, Alex J.; Jansen, Olav; Hövener, Jan‐Bernd
• Is Part Of: Medical physics. Volume 48:Issue 4(2021)
• Page Start: 1469
• Page End: 1484
• Publication Date: 2021-02-17
• Subjects: 3D printing -- 4D flow MRI -- flow diverter stent -- intracranial aneurysm -- vascular models flow modulation device; Medical physics -- Periodicals; Medical physics; Geneeskunde; Natuurkunde; Toepassingen; Biophysics; Periodicals; Periodicals; Electronic journals; 610.153
• DOI: 10.1002/mp.14714
• Languages: English
• ISSNs: 0094-2405

Abstract

Abstract : Purpose: Flow models of intracranial aneurysms (IAs) can be used to test new and existing endovascular treatments with flow modulation devices (FMDs). Additionally, 4D flow magnetic resonance imaging (MRI) offers the ability to measure hemodynamics. This way, the effect of FMDs can be determined noninvasively and compared to patient data. Here, we describe a cost‐effective method for producing flow models to test the efficiency of FMDs with 4D flow MRI. Methods: The models were based on human radiological data (internal carotid and basilar arteries) and printed in 3D with stereolithography. The models were printed with three different printing layers (25, 50, and 100 µm thickness). To evaluate the models in vitro, 3D rotational angiography, time‐of‐flight MRI, and 4D flow MRI were employed. The flow and geometry of one model were compared with in vivo data. Two FMDs (FMD1 and FMD2) were deployed into two different IA models, and the effect on the flow was estimated by 4D flow MRI. Results: Models printed with different layer thicknesses exhibited similar flow and little geometric variation. The mean spatial difference between the vessel geometry measured in vivo and in vitro was 0.7 ± 1.1 mm. The main flow features, such as vortices in the IAs, were reproduced. The velocities in the aneurysms were similar in vivo and in vitro (mean velocity magnitude: 5.4 ± 7.6 and 7.7 ± 8.6 cm/s, maximum velocity magnitude: 72.5 and 55.1 cm/s). By deploying FMDs, the mean velocity was reduced in the IAs (from 8.3 ± 10 to 4.3 ± 9.32 cm/s for FMD1 and 9.9 ± 12.1 to 2.1 ± 5.6 cm/s for FMD2). Conclusions: The presented method allows to produce neurovascular models in approx. 15 to 30 h. The resulting models were found to be geometrically accurate, reproducing the main flow patterns, and suitable for implanting FMDs as well as 4D flow MRI.