- Email: [email protected]
Development of Three-Dimensional Printed Craniocerebral Models for Simulated Neurosurgery
Accepted Manuscript Development of 3D Printed Craniocerebral Models for Simulated Neurosurgery Qing Lan, Ailin Chen, Tan Zhang, Guowei Li, Qing Zhu, Xiaomin Fan, Cheng Ma, Tao Xu PII:
S1878-8750(16)30196-6
DOI:
10.1016/j.wneu.2016.04.069
Reference:
WNEU 3998
To appear in:
World Neurosurgery
Received Date: 18 March 2016 Revised Date:
17 April 2016
Accepted Date: 20 April 2016
Please cite this article as: Lan Q, Chen A, Zhang T, Li G, Zhu Q, Fan X, Ma C, Xu T, Development of 3D Printed Craniocerebral Models for Simulated Neurosurgery, World Neurosurgery (2016), doi: 10.1016/ j.wneu.2016.04.069. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Title page Development of 3D Printed Craniocerebral
RI PT
Models for Simulated Neurosurgery
Qing LAN, Ailin CHEN, Tan ZHANG, Guowei LI, Qing ZHU, Xiaomin FAN, Cheng MA, Tao XU
Qing Lan, Ailin Chen, Tan Zhang,
M AN U
University, Suzhou 215004, China
SC
Department of Neurosurgery, Second Affiliated Hospital of Soochow
Guowei LI, Qing Zhu ; Bio-Manufacturing Center, Department of Mechanical Engineering, Tsinghua University, Beijing 100084, China Xiaomin Fan, Cheng Ma, Tao Xu
TE D
Corresponding author: Qing LAN, Email: [email protected] Tao XU, Email: [email protected]
Abstract
EP
Objective: To use 3D printed craniocerebral models to guide
AC C
neurosurgery and design the best operative route preoperatively. Method: CT, MRI, CTA and fMRI images of the patients were collected as needed, reconstructed to form multicolor 3D craniocerebral images, and printed to form solid 3D models. The hollow aneurysm model was printed with rubber-like material; craniocerebral models were printed with resin or gypsum. Results: The 3D printed hollow aneurysm model was highly representative of what was observed during the surgery. The model had
ACCEPTED MANUSCRIPT
realistic texture and elasticity and was used for preoperative simulation of aneurysm clipping for clip selection, which turned out to be the same as what was used during the surgery. The craniocerebral aneurysm model
RI PT
clearly demonstrated the spatial relation between the aneurysm and surrounding tissues, which can be used to select the best surgical approach in the preoperative simulation, to evaluate the necessity of drilling the
SC
anterior clinoid process, and to determine the feasibility of using
M AN U
contralateral approach. The craniocerebral tumor and anatomical model displayed the spatial relation between tumor and intracranial vasculatures, tractus pyramidalis and functional areas, which was helpful when selecting the optimal surgical approach to avoid damages on brain function and
TE D
helpful for learning the functional anatomy of the craniocerebral structure, and preoperative selection of surgical spaces in the sellar region. Conclusion: 3D printing provides neurosurgeons with solid craniocerebral
EP
models that can be observed and operated on directly and effectively, which
AC C
further improves the accuracy of neurosurgeries. Keywords 3D printing; Precision surgery; Solid model; Neurosurgery Surgical simulation
Three dimensional (3D) printing is an emerging technology based on digital 3D imaging and multi-layer continuous printing. Using digital
ACCEPTED MANUSCRIPT
craniocerebral images, sinterable powdery materials can be printed to form solid models through the layer-by-layer approach with 3D printers. These solid craniocerebral disease models can be used for 3D observations and be
RI PT
operated on directly, which is helpful for selecting surgical approach preoperatively, aversion of surgical risks, implementation of simulated operations, and can potentially replace cadaveric specimens in anatomical
SC
studies.
M AN U
To date, 3D printed model with tactical sensation of real bone have been utilized as an endoscopic endonasal transsphenoidal approach training system. Drilling can be used on the models to reveal anatomical structures such as planum sphenoidale, internal carotid artery, optic nerve, sella 17
; 3D model-based training system of paracentesis for
TE D
turcica, and clivus
external ventricular drainage has also been established for resident doctors 16
. Benet et al. reported to print 3D models of basilar apex and middle
EP
cerebral artery aneurysms from real patient data, and implanted them into
AC C
the anatomic regions of cadaveric heads to create realistic, individualized aneurysm models 2 . Toshihiro et al. conducted trainings such as brain tissue traction
9
and microvascular decompression
7
by using 3D printed
craniocerebral models. Thus, 3D printed craniocerebral models have great potential in neurosurgery. In 2009, it took 3~7 days and about $300~400[7] to make a 3D printed model of an intracerebral aneurysm. In recent years, it reported that only 6
ACCEPTED MANUSCRIPT
hours and 600 yen (approximately $6) material fee was needed to make a simple aneurysm model and 15000 yen (approximately $150) [15] to prepare a whole skull model. With the improvement of 3D printing technology and
RI PT
development of materials, solid model-based surgical simulation is gaining popularity.
In this study, we have created 3D printed craniocerebral models of
SC
intracranial vasculatures, nerves, aneurysms, and tumors. At a high
M AN U
resolution of 0.016 mm, it took 2 hours to print the hollow aneurysm model, 20 hours to print the whole craniocerebral model at a material cost of $20-200.
These models were used in preoperative simulation of 65
complicated clinical cases, and achieved promising results.
TE D
Materials and Methods
Image data collection: Data from clinical craniocerebral imaging examinations, including CT, MRI and CTA was obtained. Other MR data,
EP
such as DTI and BOLD images were added when necessary. As much as
AC C
possible digital information of the craniocerebral structure was collected through thin-layer chromatographic (TLC) scanning, saved in DICOM format, and exported from the computer. Images of bone tissue and vasculatures were collected with CT and CTA, respectively. The layer thickness of CT was 0.5mm. Images of soft tissues such as craniocerebral tumor and brain tissue were collected by MRI with a layer thickness of 1.0 mm.
ACCEPTED MANUSCRIPT
3D reconstruction of DICOM images: The image data was processed by the Mimics 3D image reconstruction software, which sorted tissues according to their grayscale, and converted tissues within a certain range of
RI PT
grayscale into 3D images through the algorithm of the software. 3D image processing: Data with the same grayscale but belonged to different tissues, such as the skull, vasculatures, tumor, and the optic nerve
SC
was segmented, visualized with different colors, and saved as separate files,
M AN U
which can be exported as 3D PDF files for viewing (Figure 1). Pre-printing preparation: Using the software provided by the 3D printer, digital files of different tissues or organs were processed for material selection and color identification. Objet350 Connex3 printer from
TE D
Stratasys (Rehovot, Israel) was used in this study, as it can utilize materials with different hardness, color, or transparency for 3D printing. Similar to 2D inkjet printer, rigid opaque photopolymers—VeroCyan, VeroMagenta
EP
and VeroYellow with the three primary colors can produce hundreds of
AC C
vivid colors after blending. These materials can be mixed with the Polyjet photosensitive polymers, including resins, hard materials, rubber-like materials, transparent and high-temperature resisting materials from Stratasys to simulate standard and high-temperature resisting engineering plastics. The color, softness and other features of the materials were selected; digital images were imported into the 3D printer to print models through layer-by-layer accumulation of the polymeric materials.
ACCEPTED MANUSCRIPT
Printing and post-printing processing: During printing, the liquid polymeric “ink” changed to solid state through the stereo lithography apparatus prototyping technology. To maintain its shape, the 3D model was
RI PT
supported by a type of easy-forming material during the printing process, which was removed by washing after the completion of printing.
Hollow aneurysm model: A digital 3D aneurysm reconstruction was
SC
created using CTA images. To create a hollow model, the thickness of the
M AN U
aneurysm wall was extracted manually and set to 0.5 mm. This data was used for 3D printing, after completion and removal of supporting materials, a hollow aneurysm model with wall thickness of 0.5 mm was created. Craniocerebral models: The CT and CTA digital images were used for
TE D
3D reconstruction of the skull and vasculatures. Similarity, MRI images were used for 3D reconstruction of brain tumors, nerves and conduction bundles. These images were combined to form an integrated craniocerebral
Results
AC C
EP
digital image, which was used to create the 3D printed model.～ ～
Hollow aneurysm model: hollow 1:1 replica of patient’s aneurysm was
created with 3D printing. The model was highly consistent with operative findings, and accurately replicated the vasculature system down to vessels as small as 1 mm in diameter. The model was of good texture and elasticity and can be used for simulating aneurysm clipping (Figure 2), aneurysm clips selected with the model were the same as the ones used during the
ACCEPTED MANUSCRIPT
surgery (Figure 3). In order to evaluate the elasticity of the model, a 3D printing hollow aneurysm was clipped repeatedly 20 times without broken. And the rigidity of the model was tested by a tensile strength test instrument
RI PT
(HY3080, HengYi testing instruments co., LTD, Shanghai, China) with clamping length 10mm, effective length 40mm and stretching velocity 200mm/min. Results showed that the maximum load was (8.57±1.00)N,
M AN U
25.31)%, elasticity modulus (0.45±0.02).
SC
and strength of extension(Stress) (0.73±0.08)MPa, elongation (163.68±
Craniocerebral aneurysm model: the craniocerebral aneurysm model (Figure 4) clearly displayed surgical space in the sellar region, which was consistent with what was showed during the surgery (Figures 5); this model
TE D
directly reflected the relation between aneurysm and the surrounding tissues, elucidated the relation between the anterior clinoid process and aneurysm, which helped the surgeon to decide if the anterior clinoid process needed to
EP
be abraded (Figure 5). This model was also helpful for selecting the surgical
AC C
approach, and can be used to preoperatively explore the feasibility of keyhole approach or contralateral approach (Figure 3C). Craniocerebral tumor and anatomical model: Digital images of the
brain tumor, vasculatures, and nerve conduction bundle were fused and used to create 3D printed model, which clearly demonstrated the spatial relation between the tumor and the conduction bundle (Figure 6). When necessary, the eloquent area images can be added and the resulting 3D
ACCEPTED MANUSCRIPT
printed model can be used for operative route selection in minimal invasive surgery. Furthermore, bone tissue can be printed with transparent material to show damage to the bone caused by tumor, and to display the relation
Discussion
RI PT
between the tumor and vasculature inside the bone (Figure 7).
SC
Image-guided preoperative simulation is a technique that has been
M AN U
long wanted by surgeons, as it can be used for designing an effective plan to avoid the risk of surgery, improving surgical skills and increasing the confidence of the surgeon, and shortening the operation time. Modern information technology makes it possible to reconstruct,
TE D
process, manipulate, and analyze 3D images of craniocerebral structures with computers. Furthermore, simulated surgery can be performed on complex lesions with the augmented-reality model created on computer.
EP
However, this virtual model does not provide realistic tactical sensation,
AC C
lacks the ability of 3D analysis, and cannot be used for actual surgical operation. Thus, the result of surgical simulation with augmented-reality model is far from satisfactory and has not been widely used in clinical practices.
In 1990, 3D printing technology was first applied in the field of neurosurgery, as Mankovich et al.
11
successfully printed a 3D skull model;
in 2004, Wurm et al.(20) printed a 3D cerebrovascular aneurysm model for
ACCEPTED MANUSCRIPT
surgery design. After more than 20 years of development, 3D printed models have been used more and more frequently in neurosurgeries (1,3,13). 3D printed craniocerebral models for approach selection: The size of
RI PT
sellar interspaces is closely related to the design of operative route. Fully understanding of the sellar region space and the selection of an appropriate operative route is the prerequisite of a successful surgery. For example,
, clipping of the internal carotid aneurysm via the contralateral
M AN U
10,12
SC
clipping of basilar artery aneurysm with the supraorbital keyhole approach
approach, and excision of sellar tumors(21) cannot be achieved without effective selection of the sellar region space. Some researchers used cisternal CT images to estimate the size of prechiasmatic interspace and
TE D
whether the aneurysm was in the cavernous sinus(18); others chose multiple image markers, such as the length of optic nerve, optic chiasma position, optic nerve shifting and the surrounding structure of the anterior clinoid
EP
process, tuberculum sellae, cavernous sinus dural rings to analyze and
AC C
estimate the size of sellar interspaces. (8) However, these assessments are based on 2D images and the interference between various parameters is hard to be eliminated. Integration of 3D images solves this issue because the interspaces can be observed directly and intuitively, which is very helpful for the precise preoperative design of the surgery. 3D printed craniocerebral models for aneurysm clipping: In aneurysm clipping, clip selection is key to the success of the operation. Through
ACCEPTED MANUSCRIPT
preoperative surgical simulation with the 3D model, the required length and angles of the clips, combined clipping scheme of multiple aneurysm clips, and clipping angle can be effectively determined to prevent unfavorable
RI PT
events, such as repeated intraoperative clip adjustments, incomplete aneurysm clipping, and parent artery distortion from happening.
Accurately determine the spatial relation between the anterior clinoid
SC
process and the parent artery or the aneurysm can elucidate if the anterior
M AN U
clinoid process needs to be abraded, or if internal carotid artery should be controlled with balloon catheter, or if neck incision with tie control is needed. Although the cerebral angiography can display this relation at a certain plane, it is not accurate because of the view angle. 3D printed solid
TE D
model can be directly observed from the angle of the operative approach, which is helpful for avoiding the interference from the anterior clinoid process.
EP
In one case study, Khan et al.
5
successfully created a 3D solid model
AC C
of an unruptured right paraclinoidal aneurysm from its images. In another study, Kosuke et al.
6
used CTA images to print the 3D model of an
unruptured aneurysm including the base of the skull, Willis artery circle, and aneurysm for anatomic demonstration in microsurgery. The length of the arteries was similar to that of the actual vessels, however, the vessel wall thickness and size of the aneurysm was not accurately captured. In our study, to create the hollow aneurysm model, CTA data was used to
ACCEPTED MANUSCRIPT
reconstruct the 3D image, and the aneurysm wall thickness was manually extracted and controlled to be 0.5mm. Although the thickness of aneurysm wall was not accurately replicated due to the resolution of the images, the
RI PT
size of the vasculatures was accurately captured in this model. Multiple intracranial aneurysms often locate at the two sides of the cranial cavity. When unilateral craniotomy is performed for bilateral
SC
aneurysm clipping, it is almost non-invasive to the contralateral aneurysm.
M AN U
However, if the indication is not suitable enough, contralateral approach will lead to less optimal exposure of the surgical region, poor control of the parent artery, incomplete aneurysm clipping, or even catastrophic consequences such as uncontrollable aneurysm hemorrhage.
TE D
Preoperative surgical simulation with the printed 3D model helps with feasibility assessment and selection of the optimal surgical plan (Figure 8). In the 3D printed model, an appropriate surgical approach can be selected,
EP
and the inner plate of the skull can be abraded to reflect what have been
AC C
performed in the real surgery; under microscope, the intracranial structure can be easily observed from multiple angles through the keyhole approach to carry out the corresponding operation procedures. These are difficult to accomplish with augmented-reality computerized models. 3D printed craniocerebral models for brain tumor resection: The ideal outcome of surgical treatment of intracranial tumors is total removal of the tumor while maintaining the original functions of the brain; or even repair
ACCEPTED MANUSCRIPT
of the impacted function caused by tumor extrusion or brain edema. The prerequisite of maintaining the function is to effectively avoid damages to the cranial nerve, vasculatures, and the functional areas.
RI PT
In the printed craniocerebral tumor model, not only tumor peripheral nerve and vascular structure can be viewed in 3D, but also the skull can be printed in transparent resin, thus the spatial relation between the tumor that
SC
destroys petrous bone and the petrous segment of internal carotid artery can
M AN U
be observed through the petrous bone (Figure 7). Structures such as the pyramidal tract, optic tract, optic radiation, and functional areas that cannot be observed during the surgery can be printed in the 3D model (Figure 6), thus be excluded from the surgical route during preoperative planning; 3D
TE D
printed models could also help the design of joint keyhole approach for treating giant intracranial tumors, avoiding the functional areas while performing total resection of brain tumor. These assistances provided by 3D
printed
craniocerebral
models
for
anatomy
study:
The
AC C
3D
EP
printed models could greatly improve the safety of the surgery.
craniocerebral anatomy is complex, and neurosurgeons should have a strong knowledge of it, which is mainly gained from textbooks and dissection on cadavers, and gradually strengthened in surgical operations. Because of the limitations of clinical surgeries, and ethical issues, there is a decreasing trend of new surgeon training through clinical operations. Furthermore, there is a lack of surgical experience of certain complicated
ACCEPTED MANUSCRIPT
cases due to their rarity. 2D images can increase the subjective understanding of individual lesions and their anatomy, but cannot be practiced on for the surgeons to justify and progress with what has been
RI PT
observed. With 3D printed craniocerebral models, neurosurgeons can gradually train themselves from the most basic operations, such as design of incision,～ craniotomy, simple tumor resection and vascular～ dissection . This preoperative practice is particularly important for the professional
SC
19
M AN U
growth of young neurosurgeons, because only getting plenty of exercises in the model can they become sufficiently skillful and confident in the actual surgery.
Modern precision neurosurgery requires the surgeon not only has a
TE D
good understanding of the craniocerebral microanatomy, but also a sufficient understanding of the functional anatomy not reflected in cadaveric craniocerebral specimens to avoid function impairment caused by
EP
surgery. For example, the conduction bundle and functional areas cannot be
AC C
seen during the surgery. Even with intraoperative electrophysiology monitoring, irreparable damages may have already occurred before warning is issued. In 3D printed model, the spatial relation between tumor, conduction bundle and functional areas becomes clear, which helps the surgeon to select the best operative route to avoid important areas. Based on two cases of lesions at the peripheral of the motor center, Spottiswoode et al.
14
created a 3D printed model that included the functional areas and
ACCEPTED MANUSCRIPT
clearly indicated the location and extent of a tumor relative to brain surface features and important adjacent structures with an average error of less than 0.5 mm. The solid craniocerebral models demonstrated in this study had
RI PT
well-constructed skull, vasculatures, nerves, conduction bundle, and brain tissue. They not only reflected normal visible organizational structures, but also showed the location of the conduction bundle and functional areas that
SC
cannot be displayed by cadaveric specimens. Thus, 3D printed models are
M AN U
of great help for learning the craniocerebral microanatomy.
Deficiencies of 3D printed craniocerebral models: Certainly, 3D printing needs further improvement as there are still several limitations when creating craniocerebral lesion models: it cannot completely replicate
TE D
the normal craniocerebral structure under physiological conditions; it is difficult to establish the arachnoid system and to reflect hemodyamics; and it is difficult to simulate aneurysm wall strength and adhesion with
Conclusions:
AC C
EP
surrounding tissues (4).
In summary, patient-specific 3D printed model allows the surgeon to repeatedly design and practice before operation, which greatly improves the quality of preoperative surgical planning. Simulation with 3D printed solid models also ensures the feasibility of surgery and improves skills of the surgeons. For surgeons in training or junior surgeons, this model is also helpful for the study of craniocerebral and functional anatomy.
ACCEPTED MANUSCRIPT
References 1. AlAli AB, Griffin MF, ButlerPE. Three-Dimensional Printing Surgical Applications. Eplasty, 2015; 15(37): 352-67.
RI PT
2. Benet A, Plata-Bello J, Abla AA, Acevedo-Bolton G, Saloner D, Lawton MT: Implantation of 3D-Printed Patient-Specific Aneurysm Models into Cadaveric Specimens: A New Training Paradigm to Allow for Improvements in Cerebrovascular Surgery and Research. BioMed research international 2015, 2015:939387.
SC
3. Giesel FL, Hart AR, Hahn HK, et al. 3D reconstructions of the cerebral ventricles and volume quantification in children with brain malformations. AcadRadiol, 2009; 16(5):
M AN U
610-7.
4. GabrieleWurm, Michael Lehner, Berndt Tomancok, et al. Cerebrovascular Biomodeling for Aneurysm Surgery: Simulation-Based Training by Means of Rapid Prototyping Technologies. Surgical Innovation, 2011,18(3):294-306. 5. Khan IS, Kelly PD, Singer RJ. Prototyping of cerebral vasculature physical models.
TE D
SurgNeurolInt, 2014; 5(11).
6. Kosuke K, Masaaki N, Hiroyuki M, et al. Anatomical reproducibility of a head model molded by a three-dimensional printer. Neurol Med Chir, 2015; 55: 592-98.
EP
7. Kimura T, Morita A, Nishimura K, Aiyama H, Itoh H, Fukaya S, et al: Simulation of and training for cerebral aneurysm clipping with 3-dimensional models. Neurosurgery
AC C
2009, 65(4):719-725; discussion 725-716. 8. Kakizawa Y, Tanaka Y, Orz Y, et al. Parameters for contralateral approach to ophthalmic segment aneurysms of the internal carotid artery. Neurosurgery, 2000,47(5):1130-1136; discussion 1136-1137. 9. Kaneko N, Watanabe E: Training in Brain Retraction Using a Self-Made Three-Dimensional Model. World neurosurgery 2015, 84(2):585-590. 10. Lan Q, Zhu Q, Li G. Microsurgical Treatment of Posterior Cerebral Circulation Aneurysms Via Keyhole Approaches. World Neurosurg, 2015, 84(6):1758-1764.
ACCEPTED MANUSCRIPT 11. Mankovich NJ, Cheeseman AM, Stoker NG. The display of three-dimensional anatomy with stereolithographic models. J Digit Imaging, 1990; 3(3): 200-3. 12. Ma Y, Lan Q. An anatomic study of the occipital transtentorial keyhole approach. World Neurosurg, 2013, 80(1-2):183-189.
RI PT
13. Rengier F, Mehndiratta A, von Tengg-Kobligk H, et al. 3D printing based on imaging data: review of medical applications. Int J Comput AssistRadiolSurg, 2010; 5(4): 335-41.
14. Spottiswoode BS. van den Heever DJ. Chang Y, et al. Preoperative
SC
three-dimensional model creation of magnetic resonance brain images as a tool to assist neurosurgical planning. StereotactFunct Neurosurg,2013, 91(3):162-169. et al. Development of
M AN U
15. Toshihiro Mashiko, Keisuke Otani, Ryutaro Kawano
Three-Dimensional Hollow Elastic Model for Cerebral Aneurysm Clipping Simulation Enabling Rapid and Low Cost Prototyping. World Neurosurg, 2015,83 (3): 351-361. 16. Tai BL, Rooney D, Stephenson F, Liao PS, Sagher O, Shih AJ, et al: Development of a 3D-printed external ventricular drain placement simulator: technical note. Journal of
TE D
neurosurgery 2015, 123(4):1070-1076.
17. Tai BL, Wang AC, Joseph JR, Wang PI, Sullivan SE, McKean EL, et al: A physical simulator for endoscopic endonasal drilling techniques: technical note. Journal of
EP
neurosurgery 2015:1-6.
18. Yamada K, Hayakawa T, Oku Y, et al. Contralateral pterional approach for carotidophthalmic aneurysm: usefulness of high resolution metrizamide or blood computed
AC C
tomographic cisternography. Neurosurgery, 1984,15(1):5-8. 19. Waran V, Narayanan V, Karuppiah R, et al. Utility of multimaterial 3D printers in creating models with pathological entities to enhance the training experience of neurosurgeons. J Neurosurg, 2014; 120(2):489-92. 20. Wurm G, Tomancok B, Pogady P, Holl K, Trenkler J
Cerebrovascular
stereolithographic biomodeling for aneurysm surgery. Technical note. J Neurosurg 2004
100:139–145.
21. Zhu Q, Yuan LQ, Xu L, et al. Craniopharyngioma removal via supraorbital keyhole approach. Austin J Neurosurg, 2014, 1(2): 9-17.
ACCEPTED MANUSCRIPT
Disclosure: The authors have no conflict of interest.
Acknowledgments:
RI PT
This study was financially supported by the following programs: 1. Program of medical innovation team and leading talent of Jiangsu province,China (No LJ201150).
SC
2. Science and technology plan projects of JiangSu province, China (No
M AN U
BL2012048).
AC C
EP
TE D
3. National Natural Science Foundation of China (No. 81170551).
ACCEPTED MANUSCRIPT
EP
TE D
M AN U
SC
RI PT
Figure Legends:
AC C
Figure 1. CT, CTA, and MRI image fused 3D reconstructed model. Bone was imaged with CT (grey), vasculatures were imaged with CTA (red), optic nerve, optic chiasma, and optic tract were imaged with MRI (yellow). This image can be freely rotated in the software for observational purposes.
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
EP
Figure 2. 3D printed hollow aneurysm and parent artery model for clip selection. A. Incomplete clipping of aneurysm by a straight clip, with the
AC C
residual of the aneurysm neck under the clip; B, C and D. Complete clipping of the aneurysm by the gun-type clip.
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
Figure 3. Left middle cerebral artery aneurysm clipping via the right supraorbital keyhole approach. A. The left middle cerebral artery aneurysm was seen during surgery; B. A gun-type clip was used to clip the aneurysm during surgery; C. On the 3D printed model, clipping angle was simulated via the right supraorbital approach; D. the size, shape and direction of the printed aneurysm was consistent with what was observed during the surgery, and a gun-type chip was selected in the simulation.
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Figure 4. 3D printed model that contains the skull, vasculatures, and
TE D
the optic nerve can be used for the selection of operative route, especially
AC C
EP
the assessment and selection of surgical space in the sellar region.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Figure 5. Right posterior communicating artery aneurysm clipping via
the right supraorbital keyhole approach. A. The communicating artery aneurysm was adjacent to the anterior clinoid process, as observed during the surgery; B. The 3D printed craniocerebral aneurysm model displayed the spatial relation between the aneurysm, anterior clinoid process, and internal carotid artery as observed during the surgery. Because it was not
ACCEPTED MANUSCRIPT
covered by the dura, the exposure condition of optic nerve after falciform ligament opening can be observed with this model; C. Aneurysm clipping was limited due to the blocking of anterior clinoid process; D. Abrading of
RI PT
the anterior clinoid process was practiced on the 3D printed model; E. Abrading of the anterior clinoid process was performed in the actual
AC C
EP
TE D
M AN U
SC
surgery.
SC
RI PT
ACCEPTED MANUSCRIPT
M AN U
Figure 6. 3D printed model that demonstrates the spatial relation between brain tumor (blue), pyramidal tract (green) and cerebral vasculatures (pink) is helpful for reducing functional impairment when
AC C
EP
TE D
choosing the surgical approach.
SC
RI PT
ACCEPTED MANUSCRIPT
M AN U
Figure 7. 3D printed model of bone-damaging neurinoma. A: A model printed with gypsum that demonstrated the spatial relation between tumor (blue), skull (grey), and vasculatures(red); B. A model printed with transparent resin that demonstrated the damaging condition of tumor (light
TE D
yellow) to the petrous bone (white to transparent) and the spatial relation
AC C
EP
between the neurinoma and the internal carotid artery.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Figure 8: Right posterior communicating artery aneurysm clipping via the right supraorbital keyhole approach. A: A long, curved clip was used for aneurysm clipping after temporary clipping of the internal carotid artery;
ACCEPTED MANUSCRIPT
B. Preoperative clipping simulation with the 3D printed craniocerebral aneurysm model; C. Incomplete clipping was observed in the preoperative simulation with the 3D printed hollow aneurysm model; D. Incomplete
RI PT
aneurysm clipping was observed with intraoperative endoscopy; E F. Aneurysm clipping was simulated on the hollow and craniocerebral
aneurysm models, respectively. Due to incomplete clipping of the first clip,
SC
a second fenestrate straight clip was used to clip the residual aneurysm neck;
M AN U
G. A second clip was used for clipping the residual aneurysm neck during
AC C
EP
TE D
the actual surgery.
ACCEPTED MANUSCRIPT
1.With the improvement of 3D printing technology and the development of materials, the solid model of brain surgical simulation is becoming more and more feasible.
RI PT
2.In our study, under the high precision of 0.016 mm, it takes approximately 2 hours to print the aneurysm cavity model, and approximately 20 hours to prepare the brain entity model, and 20～200
SC
dollars of materials cost.
M AN U
3.Individual disease model allows the surgeon to repeatedly design and practice before operation, which will greatly improve the preoperative surgical design.
4.The model is also conducive to the learning of brain anatomy and
TE D
functional anatomical knowledge.
5.Though there are some articles about 3D printed models for neurosurgery, few of that refer to nerve bundle, anterior clinoidectomy,
EP
selection of the sellar region space, feasibility demonstration of
AC C
contralateral approach and tumor observation through transparent skull resin.
ACCEPTED MANUSCRIPT
Three dimensional:
3D
Computed Tomography:
CT MRI
Diffusion Tensor Imaging:
DTI
AC C
EP
TE D
M AN U
SC
RI PT
Magnetic Resonance Imaging:
S1878-8750(16)30196-6
DOI:
10.1016/j.wneu.2016.04.069
Reference:
WNEU 3998
To appear in:
World Neurosurgery
Received Date: 18 March 2016 Revised Date:
17 April 2016
Accepted Date: 20 April 2016
Please cite this article as: Lan Q, Chen A, Zhang T, Li G, Zhu Q, Fan X, Ma C, Xu T, Development of 3D Printed Craniocerebral Models for Simulated Neurosurgery, World Neurosurgery (2016), doi: 10.1016/ j.wneu.2016.04.069. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Title page Development of 3D Printed Craniocerebral
RI PT
Models for Simulated Neurosurgery
Qing LAN, Ailin CHEN, Tan ZHANG, Guowei LI, Qing ZHU, Xiaomin FAN, Cheng MA, Tao XU
Qing Lan, Ailin Chen, Tan Zhang,
M AN U
University, Suzhou 215004, China
SC
Department of Neurosurgery, Second Affiliated Hospital of Soochow
Guowei LI, Qing Zhu ; Bio-Manufacturing Center, Department of Mechanical Engineering, Tsinghua University, Beijing 100084, China Xiaomin Fan, Cheng Ma, Tao Xu
TE D
Corresponding author: Qing LAN, Email: [email protected] Tao XU, Email: [email protected]
Abstract
EP
Objective: To use 3D printed craniocerebral models to guide
AC C
neurosurgery and design the best operative route preoperatively. Method: CT, MRI, CTA and fMRI images of the patients were collected as needed, reconstructed to form multicolor 3D craniocerebral images, and printed to form solid 3D models. The hollow aneurysm model was printed with rubber-like material; craniocerebral models were printed with resin or gypsum. Results: The 3D printed hollow aneurysm model was highly representative of what was observed during the surgery. The model had
ACCEPTED MANUSCRIPT
realistic texture and elasticity and was used for preoperative simulation of aneurysm clipping for clip selection, which turned out to be the same as what was used during the surgery. The craniocerebral aneurysm model
RI PT
clearly demonstrated the spatial relation between the aneurysm and surrounding tissues, which can be used to select the best surgical approach in the preoperative simulation, to evaluate the necessity of drilling the
SC
anterior clinoid process, and to determine the feasibility of using
M AN U
contralateral approach. The craniocerebral tumor and anatomical model displayed the spatial relation between tumor and intracranial vasculatures, tractus pyramidalis and functional areas, which was helpful when selecting the optimal surgical approach to avoid damages on brain function and
TE D
helpful for learning the functional anatomy of the craniocerebral structure, and preoperative selection of surgical spaces in the sellar region. Conclusion: 3D printing provides neurosurgeons with solid craniocerebral
EP
models that can be observed and operated on directly and effectively, which
AC C
further improves the accuracy of neurosurgeries. Keywords 3D printing; Precision surgery; Solid model; Neurosurgery Surgical simulation
Three dimensional (3D) printing is an emerging technology based on digital 3D imaging and multi-layer continuous printing. Using digital
ACCEPTED MANUSCRIPT
craniocerebral images, sinterable powdery materials can be printed to form solid models through the layer-by-layer approach with 3D printers. These solid craniocerebral disease models can be used for 3D observations and be
RI PT
operated on directly, which is helpful for selecting surgical approach preoperatively, aversion of surgical risks, implementation of simulated operations, and can potentially replace cadaveric specimens in anatomical
SC
studies.
M AN U
To date, 3D printed model with tactical sensation of real bone have been utilized as an endoscopic endonasal transsphenoidal approach training system. Drilling can be used on the models to reveal anatomical structures such as planum sphenoidale, internal carotid artery, optic nerve, sella 17
; 3D model-based training system of paracentesis for
TE D
turcica, and clivus
external ventricular drainage has also been established for resident doctors 16
. Benet et al. reported to print 3D models of basilar apex and middle
EP
cerebral artery aneurysms from real patient data, and implanted them into
AC C
the anatomic regions of cadaveric heads to create realistic, individualized aneurysm models 2 . Toshihiro et al. conducted trainings such as brain tissue traction
9
and microvascular decompression
7
by using 3D printed
craniocerebral models. Thus, 3D printed craniocerebral models have great potential in neurosurgery. In 2009, it took 3~7 days and about $300~400[7] to make a 3D printed model of an intracerebral aneurysm. In recent years, it reported that only 6
ACCEPTED MANUSCRIPT
hours and 600 yen (approximately $6) material fee was needed to make a simple aneurysm model and 15000 yen (approximately $150) [15] to prepare a whole skull model. With the improvement of 3D printing technology and
RI PT
development of materials, solid model-based surgical simulation is gaining popularity.
In this study, we have created 3D printed craniocerebral models of
SC
intracranial vasculatures, nerves, aneurysms, and tumors. At a high
M AN U
resolution of 0.016 mm, it took 2 hours to print the hollow aneurysm model, 20 hours to print the whole craniocerebral model at a material cost of $20-200.
These models were used in preoperative simulation of 65
complicated clinical cases, and achieved promising results.
TE D
Materials and Methods
Image data collection: Data from clinical craniocerebral imaging examinations, including CT, MRI and CTA was obtained. Other MR data,
EP
such as DTI and BOLD images were added when necessary. As much as
AC C
possible digital information of the craniocerebral structure was collected through thin-layer chromatographic (TLC) scanning, saved in DICOM format, and exported from the computer. Images of bone tissue and vasculatures were collected with CT and CTA, respectively. The layer thickness of CT was 0.5mm. Images of soft tissues such as craniocerebral tumor and brain tissue were collected by MRI with a layer thickness of 1.0 mm.
ACCEPTED MANUSCRIPT
3D reconstruction of DICOM images: The image data was processed by the Mimics 3D image reconstruction software, which sorted tissues according to their grayscale, and converted tissues within a certain range of
RI PT
grayscale into 3D images through the algorithm of the software. 3D image processing: Data with the same grayscale but belonged to different tissues, such as the skull, vasculatures, tumor, and the optic nerve
SC
was segmented, visualized with different colors, and saved as separate files,
M AN U
which can be exported as 3D PDF files for viewing (Figure 1). Pre-printing preparation: Using the software provided by the 3D printer, digital files of different tissues or organs were processed for material selection and color identification. Objet350 Connex3 printer from
TE D
Stratasys (Rehovot, Israel) was used in this study, as it can utilize materials with different hardness, color, or transparency for 3D printing. Similar to 2D inkjet printer, rigid opaque photopolymers—VeroCyan, VeroMagenta
EP
and VeroYellow with the three primary colors can produce hundreds of
AC C
vivid colors after blending. These materials can be mixed with the Polyjet photosensitive polymers, including resins, hard materials, rubber-like materials, transparent and high-temperature resisting materials from Stratasys to simulate standard and high-temperature resisting engineering plastics. The color, softness and other features of the materials were selected; digital images were imported into the 3D printer to print models through layer-by-layer accumulation of the polymeric materials.
ACCEPTED MANUSCRIPT
Printing and post-printing processing: During printing, the liquid polymeric “ink” changed to solid state through the stereo lithography apparatus prototyping technology. To maintain its shape, the 3D model was
RI PT
supported by a type of easy-forming material during the printing process, which was removed by washing after the completion of printing.
Hollow aneurysm model: A digital 3D aneurysm reconstruction was
SC
created using CTA images. To create a hollow model, the thickness of the
M AN U
aneurysm wall was extracted manually and set to 0.5 mm. This data was used for 3D printing, after completion and removal of supporting materials, a hollow aneurysm model with wall thickness of 0.5 mm was created. Craniocerebral models: The CT and CTA digital images were used for
TE D
3D reconstruction of the skull and vasculatures. Similarity, MRI images were used for 3D reconstruction of brain tumors, nerves and conduction bundles. These images were combined to form an integrated craniocerebral
Results
AC C
EP
digital image, which was used to create the 3D printed model.～ ～
Hollow aneurysm model: hollow 1:1 replica of patient’s aneurysm was
created with 3D printing. The model was highly consistent with operative findings, and accurately replicated the vasculature system down to vessels as small as 1 mm in diameter. The model was of good texture and elasticity and can be used for simulating aneurysm clipping (Figure 2), aneurysm clips selected with the model were the same as the ones used during the
ACCEPTED MANUSCRIPT
surgery (Figure 3). In order to evaluate the elasticity of the model, a 3D printing hollow aneurysm was clipped repeatedly 20 times without broken. And the rigidity of the model was tested by a tensile strength test instrument
RI PT
(HY3080, HengYi testing instruments co., LTD, Shanghai, China) with clamping length 10mm, effective length 40mm and stretching velocity 200mm/min. Results showed that the maximum load was (8.57±1.00)N,
M AN U
25.31)%, elasticity modulus (0.45±0.02).
SC
and strength of extension(Stress) (0.73±0.08)MPa, elongation (163.68±
Craniocerebral aneurysm model: the craniocerebral aneurysm model (Figure 4) clearly displayed surgical space in the sellar region, which was consistent with what was showed during the surgery (Figures 5); this model
TE D
directly reflected the relation between aneurysm and the surrounding tissues, elucidated the relation between the anterior clinoid process and aneurysm, which helped the surgeon to decide if the anterior clinoid process needed to
EP
be abraded (Figure 5). This model was also helpful for selecting the surgical
AC C
approach, and can be used to preoperatively explore the feasibility of keyhole approach or contralateral approach (Figure 3C). Craniocerebral tumor and anatomical model: Digital images of the
brain tumor, vasculatures, and nerve conduction bundle were fused and used to create 3D printed model, which clearly demonstrated the spatial relation between the tumor and the conduction bundle (Figure 6). When necessary, the eloquent area images can be added and the resulting 3D
ACCEPTED MANUSCRIPT
printed model can be used for operative route selection in minimal invasive surgery. Furthermore, bone tissue can be printed with transparent material to show damage to the bone caused by tumor, and to display the relation
Discussion
RI PT
between the tumor and vasculature inside the bone (Figure 7).
SC
Image-guided preoperative simulation is a technique that has been
M AN U
long wanted by surgeons, as it can be used for designing an effective plan to avoid the risk of surgery, improving surgical skills and increasing the confidence of the surgeon, and shortening the operation time. Modern information technology makes it possible to reconstruct,
TE D
process, manipulate, and analyze 3D images of craniocerebral structures with computers. Furthermore, simulated surgery can be performed on complex lesions with the augmented-reality model created on computer.
EP
However, this virtual model does not provide realistic tactical sensation,
AC C
lacks the ability of 3D analysis, and cannot be used for actual surgical operation. Thus, the result of surgical simulation with augmented-reality model is far from satisfactory and has not been widely used in clinical practices.
In 1990, 3D printing technology was first applied in the field of neurosurgery, as Mankovich et al.
11
successfully printed a 3D skull model;
in 2004, Wurm et al.(20) printed a 3D cerebrovascular aneurysm model for
ACCEPTED MANUSCRIPT
surgery design. After more than 20 years of development, 3D printed models have been used more and more frequently in neurosurgeries (1,3,13). 3D printed craniocerebral models for approach selection: The size of
RI PT
sellar interspaces is closely related to the design of operative route. Fully understanding of the sellar region space and the selection of an appropriate operative route is the prerequisite of a successful surgery. For example,
, clipping of the internal carotid aneurysm via the contralateral
M AN U
10,12
SC
clipping of basilar artery aneurysm with the supraorbital keyhole approach
approach, and excision of sellar tumors(21) cannot be achieved without effective selection of the sellar region space. Some researchers used cisternal CT images to estimate the size of prechiasmatic interspace and
TE D
whether the aneurysm was in the cavernous sinus(18); others chose multiple image markers, such as the length of optic nerve, optic chiasma position, optic nerve shifting and the surrounding structure of the anterior clinoid
EP
process, tuberculum sellae, cavernous sinus dural rings to analyze and
AC C
estimate the size of sellar interspaces. (8) However, these assessments are based on 2D images and the interference between various parameters is hard to be eliminated. Integration of 3D images solves this issue because the interspaces can be observed directly and intuitively, which is very helpful for the precise preoperative design of the surgery. 3D printed craniocerebral models for aneurysm clipping: In aneurysm clipping, clip selection is key to the success of the operation. Through
ACCEPTED MANUSCRIPT
preoperative surgical simulation with the 3D model, the required length and angles of the clips, combined clipping scheme of multiple aneurysm clips, and clipping angle can be effectively determined to prevent unfavorable
RI PT
events, such as repeated intraoperative clip adjustments, incomplete aneurysm clipping, and parent artery distortion from happening.
Accurately determine the spatial relation between the anterior clinoid
SC
process and the parent artery or the aneurysm can elucidate if the anterior
M AN U
clinoid process needs to be abraded, or if internal carotid artery should be controlled with balloon catheter, or if neck incision with tie control is needed. Although the cerebral angiography can display this relation at a certain plane, it is not accurate because of the view angle. 3D printed solid
TE D
model can be directly observed from the angle of the operative approach, which is helpful for avoiding the interference from the anterior clinoid process.
EP
In one case study, Khan et al.
5
successfully created a 3D solid model
AC C
of an unruptured right paraclinoidal aneurysm from its images. In another study, Kosuke et al.
6
used CTA images to print the 3D model of an
unruptured aneurysm including the base of the skull, Willis artery circle, and aneurysm for anatomic demonstration in microsurgery. The length of the arteries was similar to that of the actual vessels, however, the vessel wall thickness and size of the aneurysm was not accurately captured. In our study, to create the hollow aneurysm model, CTA data was used to
ACCEPTED MANUSCRIPT
reconstruct the 3D image, and the aneurysm wall thickness was manually extracted and controlled to be 0.5mm. Although the thickness of aneurysm wall was not accurately replicated due to the resolution of the images, the
RI PT
size of the vasculatures was accurately captured in this model. Multiple intracranial aneurysms often locate at the two sides of the cranial cavity. When unilateral craniotomy is performed for bilateral
SC
aneurysm clipping, it is almost non-invasive to the contralateral aneurysm.
M AN U
However, if the indication is not suitable enough, contralateral approach will lead to less optimal exposure of the surgical region, poor control of the parent artery, incomplete aneurysm clipping, or even catastrophic consequences such as uncontrollable aneurysm hemorrhage.
TE D
Preoperative surgical simulation with the printed 3D model helps with feasibility assessment and selection of the optimal surgical plan (Figure 8). In the 3D printed model, an appropriate surgical approach can be selected,
EP
and the inner plate of the skull can be abraded to reflect what have been
AC C
performed in the real surgery; under microscope, the intracranial structure can be easily observed from multiple angles through the keyhole approach to carry out the corresponding operation procedures. These are difficult to accomplish with augmented-reality computerized models. 3D printed craniocerebral models for brain tumor resection: The ideal outcome of surgical treatment of intracranial tumors is total removal of the tumor while maintaining the original functions of the brain; or even repair
ACCEPTED MANUSCRIPT
of the impacted function caused by tumor extrusion or brain edema. The prerequisite of maintaining the function is to effectively avoid damages to the cranial nerve, vasculatures, and the functional areas.
RI PT
In the printed craniocerebral tumor model, not only tumor peripheral nerve and vascular structure can be viewed in 3D, but also the skull can be printed in transparent resin, thus the spatial relation between the tumor that
SC
destroys petrous bone and the petrous segment of internal carotid artery can
M AN U
be observed through the petrous bone (Figure 7). Structures such as the pyramidal tract, optic tract, optic radiation, and functional areas that cannot be observed during the surgery can be printed in the 3D model (Figure 6), thus be excluded from the surgical route during preoperative planning; 3D
TE D
printed models could also help the design of joint keyhole approach for treating giant intracranial tumors, avoiding the functional areas while performing total resection of brain tumor. These assistances provided by 3D
printed
craniocerebral
models
for
anatomy
study:
The
AC C
3D
EP
printed models could greatly improve the safety of the surgery.
craniocerebral anatomy is complex, and neurosurgeons should have a strong knowledge of it, which is mainly gained from textbooks and dissection on cadavers, and gradually strengthened in surgical operations. Because of the limitations of clinical surgeries, and ethical issues, there is a decreasing trend of new surgeon training through clinical operations. Furthermore, there is a lack of surgical experience of certain complicated
ACCEPTED MANUSCRIPT
cases due to their rarity. 2D images can increase the subjective understanding of individual lesions and their anatomy, but cannot be practiced on for the surgeons to justify and progress with what has been
RI PT
observed. With 3D printed craniocerebral models, neurosurgeons can gradually train themselves from the most basic operations, such as design of incision,～ craniotomy, simple tumor resection and vascular～ dissection . This preoperative practice is particularly important for the professional
SC
19
M AN U
growth of young neurosurgeons, because only getting plenty of exercises in the model can they become sufficiently skillful and confident in the actual surgery.
Modern precision neurosurgery requires the surgeon not only has a
TE D
good understanding of the craniocerebral microanatomy, but also a sufficient understanding of the functional anatomy not reflected in cadaveric craniocerebral specimens to avoid function impairment caused by
EP
surgery. For example, the conduction bundle and functional areas cannot be
AC C
seen during the surgery. Even with intraoperative electrophysiology monitoring, irreparable damages may have already occurred before warning is issued. In 3D printed model, the spatial relation between tumor, conduction bundle and functional areas becomes clear, which helps the surgeon to select the best operative route to avoid important areas. Based on two cases of lesions at the peripheral of the motor center, Spottiswoode et al.
14
created a 3D printed model that included the functional areas and
ACCEPTED MANUSCRIPT
clearly indicated the location and extent of a tumor relative to brain surface features and important adjacent structures with an average error of less than 0.5 mm. The solid craniocerebral models demonstrated in this study had
RI PT
well-constructed skull, vasculatures, nerves, conduction bundle, and brain tissue. They not only reflected normal visible organizational structures, but also showed the location of the conduction bundle and functional areas that
SC
cannot be displayed by cadaveric specimens. Thus, 3D printed models are
M AN U
of great help for learning the craniocerebral microanatomy.
Deficiencies of 3D printed craniocerebral models: Certainly, 3D printing needs further improvement as there are still several limitations when creating craniocerebral lesion models: it cannot completely replicate
TE D
the normal craniocerebral structure under physiological conditions; it is difficult to establish the arachnoid system and to reflect hemodyamics; and it is difficult to simulate aneurysm wall strength and adhesion with
Conclusions:
AC C
EP
surrounding tissues (4).
In summary, patient-specific 3D printed model allows the surgeon to repeatedly design and practice before operation, which greatly improves the quality of preoperative surgical planning. Simulation with 3D printed solid models also ensures the feasibility of surgery and improves skills of the surgeons. For surgeons in training or junior surgeons, this model is also helpful for the study of craniocerebral and functional anatomy.
ACCEPTED MANUSCRIPT
References 1. AlAli AB, Griffin MF, ButlerPE. Three-Dimensional Printing Surgical Applications. Eplasty, 2015; 15(37): 352-67.
RI PT
2. Benet A, Plata-Bello J, Abla AA, Acevedo-Bolton G, Saloner D, Lawton MT: Implantation of 3D-Printed Patient-Specific Aneurysm Models into Cadaveric Specimens: A New Training Paradigm to Allow for Improvements in Cerebrovascular Surgery and Research. BioMed research international 2015, 2015:939387.
SC
3. Giesel FL, Hart AR, Hahn HK, et al. 3D reconstructions of the cerebral ventricles and volume quantification in children with brain malformations. AcadRadiol, 2009; 16(5):
M AN U
610-7.
4. GabrieleWurm, Michael Lehner, Berndt Tomancok, et al. Cerebrovascular Biomodeling for Aneurysm Surgery: Simulation-Based Training by Means of Rapid Prototyping Technologies. Surgical Innovation, 2011,18(3):294-306. 5. Khan IS, Kelly PD, Singer RJ. Prototyping of cerebral vasculature physical models.
TE D
SurgNeurolInt, 2014; 5(11).
6. Kosuke K, Masaaki N, Hiroyuki M, et al. Anatomical reproducibility of a head model molded by a three-dimensional printer. Neurol Med Chir, 2015; 55: 592-98.
EP
7. Kimura T, Morita A, Nishimura K, Aiyama H, Itoh H, Fukaya S, et al: Simulation of and training for cerebral aneurysm clipping with 3-dimensional models. Neurosurgery
AC C
2009, 65(4):719-725; discussion 725-716. 8. Kakizawa Y, Tanaka Y, Orz Y, et al. Parameters for contralateral approach to ophthalmic segment aneurysms of the internal carotid artery. Neurosurgery, 2000,47(5):1130-1136; discussion 1136-1137. 9. Kaneko N, Watanabe E: Training in Brain Retraction Using a Self-Made Three-Dimensional Model. World neurosurgery 2015, 84(2):585-590. 10. Lan Q, Zhu Q, Li G. Microsurgical Treatment of Posterior Cerebral Circulation Aneurysms Via Keyhole Approaches. World Neurosurg, 2015, 84(6):1758-1764.
ACCEPTED MANUSCRIPT 11. Mankovich NJ, Cheeseman AM, Stoker NG. The display of three-dimensional anatomy with stereolithographic models. J Digit Imaging, 1990; 3(3): 200-3. 12. Ma Y, Lan Q. An anatomic study of the occipital transtentorial keyhole approach. World Neurosurg, 2013, 80(1-2):183-189.
RI PT
13. Rengier F, Mehndiratta A, von Tengg-Kobligk H, et al. 3D printing based on imaging data: review of medical applications. Int J Comput AssistRadiolSurg, 2010; 5(4): 335-41.
14. Spottiswoode BS. van den Heever DJ. Chang Y, et al. Preoperative
SC
three-dimensional model creation of magnetic resonance brain images as a tool to assist neurosurgical planning. StereotactFunct Neurosurg,2013, 91(3):162-169. et al. Development of
M AN U
15. Toshihiro Mashiko, Keisuke Otani, Ryutaro Kawano
Three-Dimensional Hollow Elastic Model for Cerebral Aneurysm Clipping Simulation Enabling Rapid and Low Cost Prototyping. World Neurosurg, 2015,83 (3): 351-361. 16. Tai BL, Rooney D, Stephenson F, Liao PS, Sagher O, Shih AJ, et al: Development of a 3D-printed external ventricular drain placement simulator: technical note. Journal of
TE D
neurosurgery 2015, 123(4):1070-1076.
17. Tai BL, Wang AC, Joseph JR, Wang PI, Sullivan SE, McKean EL, et al: A physical simulator for endoscopic endonasal drilling techniques: technical note. Journal of
EP
neurosurgery 2015:1-6.
18. Yamada K, Hayakawa T, Oku Y, et al. Contralateral pterional approach for carotidophthalmic aneurysm: usefulness of high resolution metrizamide or blood computed
AC C
tomographic cisternography. Neurosurgery, 1984,15(1):5-8. 19. Waran V, Narayanan V, Karuppiah R, et al. Utility of multimaterial 3D printers in creating models with pathological entities to enhance the training experience of neurosurgeons. J Neurosurg, 2014; 120(2):489-92. 20. Wurm G, Tomancok B, Pogady P, Holl K, Trenkler J
Cerebrovascular
stereolithographic biomodeling for aneurysm surgery. Technical note. J Neurosurg 2004
100:139–145.
21. Zhu Q, Yuan LQ, Xu L, et al. Craniopharyngioma removal via supraorbital keyhole approach. Austin J Neurosurg, 2014, 1(2): 9-17.
ACCEPTED MANUSCRIPT
Disclosure: The authors have no conflict of interest.
Acknowledgments:
RI PT
This study was financially supported by the following programs: 1. Program of medical innovation team and leading talent of Jiangsu province,China (No LJ201150).
SC
2. Science and technology plan projects of JiangSu province, China (No
M AN U
BL2012048).
AC C
EP
TE D
3. National Natural Science Foundation of China (No. 81170551).
ACCEPTED MANUSCRIPT
EP
TE D
M AN U
SC
RI PT
Figure Legends:
AC C
Figure 1. CT, CTA, and MRI image fused 3D reconstructed model. Bone was imaged with CT (grey), vasculatures were imaged with CTA (red), optic nerve, optic chiasma, and optic tract were imaged with MRI (yellow). This image can be freely rotated in the software for observational purposes.
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
EP
Figure 2. 3D printed hollow aneurysm and parent artery model for clip selection. A. Incomplete clipping of aneurysm by a straight clip, with the
AC C
residual of the aneurysm neck under the clip; B, C and D. Complete clipping of the aneurysm by the gun-type clip.
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
Figure 3. Left middle cerebral artery aneurysm clipping via the right supraorbital keyhole approach. A. The left middle cerebral artery aneurysm was seen during surgery; B. A gun-type clip was used to clip the aneurysm during surgery; C. On the 3D printed model, clipping angle was simulated via the right supraorbital approach; D. the size, shape and direction of the printed aneurysm was consistent with what was observed during the surgery, and a gun-type chip was selected in the simulation.
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Figure 4. 3D printed model that contains the skull, vasculatures, and
TE D
the optic nerve can be used for the selection of operative route, especially
AC C
EP
the assessment and selection of surgical space in the sellar region.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Figure 5. Right posterior communicating artery aneurysm clipping via
the right supraorbital keyhole approach. A. The communicating artery aneurysm was adjacent to the anterior clinoid process, as observed during the surgery; B. The 3D printed craniocerebral aneurysm model displayed the spatial relation between the aneurysm, anterior clinoid process, and internal carotid artery as observed during the surgery. Because it was not
ACCEPTED MANUSCRIPT
covered by the dura, the exposure condition of optic nerve after falciform ligament opening can be observed with this model; C. Aneurysm clipping was limited due to the blocking of anterior clinoid process; D. Abrading of
RI PT
the anterior clinoid process was practiced on the 3D printed model; E. Abrading of the anterior clinoid process was performed in the actual
AC C
EP
TE D
M AN U
SC
surgery.
SC
RI PT
ACCEPTED MANUSCRIPT
M AN U
Figure 6. 3D printed model that demonstrates the spatial relation between brain tumor (blue), pyramidal tract (green) and cerebral vasculatures (pink) is helpful for reducing functional impairment when
AC C
EP
TE D
choosing the surgical approach.
SC
RI PT
ACCEPTED MANUSCRIPT
M AN U
Figure 7. 3D printed model of bone-damaging neurinoma. A: A model printed with gypsum that demonstrated the spatial relation between tumor (blue), skull (grey), and vasculatures(red); B. A model printed with transparent resin that demonstrated the damaging condition of tumor (light
TE D
yellow) to the petrous bone (white to transparent) and the spatial relation
AC C
EP
between the neurinoma and the internal carotid artery.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Figure 8: Right posterior communicating artery aneurysm clipping via the right supraorbital keyhole approach. A: A long, curved clip was used for aneurysm clipping after temporary clipping of the internal carotid artery;
ACCEPTED MANUSCRIPT
B. Preoperative clipping simulation with the 3D printed craniocerebral aneurysm model; C. Incomplete clipping was observed in the preoperative simulation with the 3D printed hollow aneurysm model; D. Incomplete
RI PT
aneurysm clipping was observed with intraoperative endoscopy; E F. Aneurysm clipping was simulated on the hollow and craniocerebral
aneurysm models, respectively. Due to incomplete clipping of the first clip,
SC
a second fenestrate straight clip was used to clip the residual aneurysm neck;
M AN U
G. A second clip was used for clipping the residual aneurysm neck during
AC C
EP
TE D
the actual surgery.
ACCEPTED MANUSCRIPT
1.With the improvement of 3D printing technology and the development of materials, the solid model of brain surgical simulation is becoming more and more feasible.
RI PT
2.In our study, under the high precision of 0.016 mm, it takes approximately 2 hours to print the aneurysm cavity model, and approximately 20 hours to prepare the brain entity model, and 20～200
SC
dollars of materials cost.
M AN U
3.Individual disease model allows the surgeon to repeatedly design and practice before operation, which will greatly improve the preoperative surgical design.
4.The model is also conducive to the learning of brain anatomy and
TE D
functional anatomical knowledge.
5.Though there are some articles about 3D printed models for neurosurgery, few of that refer to nerve bundle, anterior clinoidectomy,
EP
selection of the sellar region space, feasibility demonstration of
AC C
contralateral approach and tumor observation through transparent skull resin.
ACCEPTED MANUSCRIPT
Three dimensional:
3D
Computed Tomography:
CT MRI
Diffusion Tensor Imaging:
DTI
AC C
EP
TE D
M AN U
SC
RI PT
Magnetic Resonance Imaging: