Intracranial vasculature 3D printing: review of techniques and manufacturing processes to inform clinical practice

Abstract

Background: In recent years, three-dimensional (3D) printing has been increasingly applied to the intracranial vasculature for patient-specific surgical planning, training, education, and research. Unfortunately, though, much of the prior literature regarding 3D printing has focused on the end-product and not the process. In addition, for 3D printing/manufacturing to occur on a large scale, challenges and bottlenecks specific to each modeled anatomy must be overcome.

Main body: In this review article, limitations and considerations of each 3D printing processing step, as they relate to printing individual intracranial vasculature models and providing an active clinical service for a quaternary care center, are discussed. Relevant advantages and disadvantages of the available acquisition techniques (computed tomography, magnetic resonance, and digital subtraction angiography) are reviewed. Specific steps in segmentation, processing, and creation of a printable file may impede the workflow or degrade the fidelity of the printed model and are, therefore, given added attention. The various available printing techniques are compared with respect to printing the intracranial vasculature. Finally, applications are discussed, and a variety of example models are shown.

Conclusion: In this review we provide insight into the manufacturing of 3D models of the intracranial vasculature that may facilitate incorporation into or improve utility of 3D vascular models in clinical practice.

Keywords: 3D printing; Cerebral angiography; Intracranial vasculature; Patient specific models.

Fig. 1

Overview of the 3D printing process. (1) Acquire angiographic data using an imaging protocol designed with the intent on 3D printing. (2) Import the imaging to 3D printing software such as Materialise: Mimics/3-Matic, 3D systems: Geomagic, OSIRIX, Toshiba: Vital Images, Autodesk, Freeware: 3D slicer, Sketchup, Blender, and many more. (3) Segment or separate the parts of the anatomy you want to print (intracranial vasculature) from the rest of the anatomy. (4) Export the STL file or segmented 3D model to the CAD program. In this step parts needed for stability must be designed, such as cylinders to hold the vertebral arteries to the skull. Surgical osteotomies maybe planned as shown in pink. (5) Print your model using your choice of print technology and material based on the need for colors, model flexibility, and support material. (6.) Post-process the model, including removal support material as shown in this example

Fig. 2

Segmentation of 3DRA, CTA, and MRA. a Initial 3D volume from thresholding a right carotid injection 3DRA includes only the vasculature due to the density of contrast and spatial resolution of angiography. b Region growing removes any external branches in one mouse click. 3DRA results in the most accurate 3D model of the intracranial vasculature in the fastest time. c Initial 3D volume from thresholding an MRA includes non-arterial structures to be removed. d Region growing removes any soft tissue and venous contamination in one mouse click, producing the second fastest and second most accurate vascular model with less robust distal vasculature than 3DRA. e CTA initial threshold shows that the arterial tree cannot be separated from bone, venous contamination, and some soft tissue by Hounsfield units alone. f After region growing, non-arterial structures, such as the dural venous sinuses (superior sagittal sinus, shaded light blue), remain that must be removed by manual trimming tools

Fig. 3

Examples of processing steps in CAD software. a Original segmented data set from a 3 T MRA. b Original data set can be wrapped to reinforce small vessels, (arrow) but this may result in merging or overlapping of vessels and loss of detail (arrowheads). c Example of over-smoothing. Smoothing creates a more realistic model removing triangulation from the mesh, though over-smoothing can result in attenuation or loss of small distal vessels and distortion of the anatomy. d Triangular mesh. 3D file is made of hundreds of thousands of triangles. Before printing, the mesh often has to be fixed to remove overlapping triangles, bad edges, and inverted normals. e Hollowed models used for patient-specific simulation or intracranial device research need to be modified in CAD software to create common outflow channels and lofted parts that can be assembled with physiologic pump systems. f Cylinders (blue) may need to be placed to support fragile anatomy so that parts that do not physically touch may be 3D printed to maintain anatomical relationships

Fig. 4

Examples of different 3D printing techniques and the required support structures. a-b Material extrusion, FDM (Stratasys Fortus, Eden Prairie, MN) model of the intracranial vasculature. a Intracranial vasculature with a water-soluble support structure. b The final model can be printed in only 1–2 materials and 1–2 colors. Resolution and fragility of printable models are dependent on the printer manufacturer and allowable materials. c-d Material jetting (Stratasys Objet 500 Polyjet, Eden Prairie, MN) model of the intracranial vasculature. c A model with surrounding soluble support material that can be removed by placement in a lye or water bath with ultrasonic agitation. d Final model printed in color. Material jetting allows full color multilateral printing. e-f Vat polymerization: SLA (Formlabs Form 2, Boston, MA) model of a posterior communicating aneurysm. e Vascular model with scaffolding support material is standard in this technology and can limit the ability to print complex internal architecture of hollow parts. f Final model with support removed. While flexible materials are available, they do not withstand physiologic pressure for device testing or simulation. g-h Binder jetting (3D Systems Projet 660, Eden Prairie, MN) model of an intracranial aneurysm printed in relation to the skull. g Support material is powder surrounding the print which is easily vacuumed and brushed away, saving significant post-processing time. h Final multicolor model is impregnated with cyanoacrylate to improve durability

Fig. 5

Clinical applications for 3D printing of the intracranial vasculature. a Hollow model in a patient with multiple intracranial aneurysms used for patient-specific simulation of aneurysm coiling (Biomodex, Paris, France). b Parasagittal arteriovenous malformation (FDM) printed in relation to the skull and used for patient education in addition to surgical planning. c Skull base tumor (Material jetting, Objet 500). The surgical approach was changed after seeing the model with the detachable posterior component. d Insular glioma (Material jetting, Objet 500) in which 3D printing was used to demonstrate the relationship of the intracranial vasculature to critical structures. e Intracranial aneurysm printed in relation to the skull in a 7-year-old (Material jetting, Objet 500), and (f) vein of Galen malformation printed in relation to the skull (Material jetting, Objet 500) in a 6 month old. Both e and f are patient-specific models of rare pathology used for pre-surgical planning and ongoing Accreditation Council for Graduate Medical Education radiology, neurosurgery, and neuropathology education. The model of the vein of Galen malformation (f) changed the treatment to a two stage approach of interventional therapy followed by surgical intervention from a single interventional treatment plan