Accessible and Informative Sectioned Images and Surface Models of a ...

ANATOMICAL STUDY

Accessible and Informative Sectioned Images and Surface Models of a Cadaver Head Dong Sun Shin, PhD,* Hae Gwon Jang, MS,Þ Jin Seo Park, PhD,þ Hyung Seon Park, PhD,§ Sangho Lee, PhD,§ and Min Suk Chung, MD, PhD* Abstract: The sectioned images and surface models of a cadaver head in the Visible Korean Project would be more beneficial if they were accessible and informative. To this aim, 3 policies were established: (1) the data are promptly obtainable and observable; (2) the graphic data are accompanied by explaining the anatomic terms; and (3) two-dimensional images and three-dimensional models are shown together. According to the policies, the following were attempted. Two hundred thirty-five couples of sectioned images and outlined images (intervals, 1 mm) of the head were prepared. Browsing software was developed where the 2 corresponding images were displayed simultaneously. In addition, the structures in the images were recognized with the aid of automatic labeling. From the outlined images, surface models of 178 head structures were constructed. The two-dimensional surface models including the sectioned images were embedded into the three-dimensional surface models. All the models were categorized into systems and arranged to be inputted to a PDF (portable document format) file. The finalized PDF file containing comprehensive head data could be explored on Adobe Reader. If the user clicked on the surface models, their anatomic names were highlighted. The sectioned images, outlined images, and surface models in the browsing software and PDF file can be downloaded from the homepage (anatomy.co.kr) free of charge. The state-of-the-art graphic information will hopefully assist medical students in learning head anatomy. In addition, the raw data are expected to contribute to the various clinical practice simulations. Key Words: Anatomic models, computer-assisted image processing, head, neuroanatomy, three-dimensional imaging, user-computer interface, visible human projects (J Craniofac Surg 2012;23: 1176Y1180) From the *Department of Anatomy, School of Medicine, and †Graduate School of Information and Communication, Ajou University, Suwon; ‡Department of Anatomy, Dongguk University College of Medicine, Gyeongju; and §Korea Institute of Science and Technology Information, Daejeon, Republic of Korea. Received February 7, 2012. Accepted for publication March 17, 2012. Address correspondence and reprint requests to Dr. Min Suk Chung, Department of Anatomy, Ajou University School of Medicine, 164 Worldcup-ro, Suwon, Republic of Korea, 443-749; E-mail: [email protected] This research was supported by Basic Science Research Program through the National Research Foundation of Korea funded by the Ministry of Education, Science and Technology (grant no. 2010-0009950). The authors report no conflicts of interest. Copyright * 2012 by Mutaz B. Habal, MD ISSN: 1049-2275 DOI: 10.1097/SCS.0b013e31825657d8

1176

I

n the Visible Korean Project, sectioned images and surface models of the head have been created as follows. The head of a male cadaver was serially sectioned and photographed to provide the sectioned images. Outlines of different structures were drawn on the sectioned images at 1-mm intervals.1,2 The contours of each structure were stacked in sequence, and a surface reconstruction was then executed to build a surface model.3Y5 The sectioned images, outlined images, and surface models that correspond with one another would be more valuable if they were more user-friendly and instructive. To that end, 3 policies have been set up: (1) the image data in reduced resolution can be downloaded from a homepage to be readily observed; (2) the sectioned images and surface models are automatically labeled with the structures’ names; and (3) two-dimensional images and threedimensional models are mixed for display. According to the principles, we already released the software that can be used to browse the sectioned images and accompanying outlined images of the entire male body. A small problem in the program is that setup files and update files have to be downloaded and installed separately.6 This problem could be alleviated by unifying 2 kinds of files in terms of the head data. Surface models of various structures from different cadavers have been built and distributed. The data can be opened using only a three-dimensional viewer such as Maya version 2012 (Autodesk, Inc, San Rafael, CA). The use of this viewer is undesirable because it is expensive to acquire and complicated for the typical users to install and handle.4,5,7Y11 To overcome this problem, we decided to use Adobe Reader Windows version 9 (Adobe Systems, Inc, San Jose, CA), a free software package where our surface models, collected in a portable document format (PDF) file, could be easily manipulated. It was also decided to make the surface models capable of being overlapped by their raw data and sectioned images and to make all models linked with their anatomic terms. The purpose of this research was to make the sectioned images and surface models of cadaver head accessible and informative. Our intent was to create a system that would allow the user to explore the two- and three-dimensional data on the browsing software and Adobe Reader, respectively, to study the topographic anatomy of the head including the brain. It is hoped that interested researchers will evaluate the complete data, with the aim of improving the educational software and/or the virtual reality simulator.

MATERIALS AND METHODS Sectioned Images and Outlined Images Our research team had already produced the sectioned images and outlined images of a head from a male cadaver (67 years old at the time of death from myasthenia gravis; height, 1.62 m; weight, 45 kg). The cadaver head was serially sectioned by means of a cryomacrotome to generate sectioned images (intervals, 0.1 mm; pixel size, 0.1 mm; color depth, 48-bit color; file format, TIFF

The Journal of Craniofacial Surgery

& Volume 23, Number 4, July 2012

Copyright © 2012 Mutaz B. Habal, MD. Unauthorized reproduction of this article is prohibited.

The Journal of Craniofacial Surgery

& Volume 23, Number 4, July 2012

Surface Models of a Cadaver Head

FIGURE 2. Surface models of the right head with serial layers removed. Surface models are artificially painted and the skin color is made semitransparent.

The browsing software including all data was transformed into an installation file on the NSIS (Nullsoft Scriptable Install System; NSIS Media), a script language. NSIS was easily handled with HM NIS Edit, a kind of user interface, by us, even with no knowledge of the computer language. Both software packages are obtainable at no cost. FIGURE 1. Full data (A, C) and reduced data (B, D) of the sectioned image (A, B) and surface model with mesh revealed (C, D).

[tagged image file format]). The sectioned images comprising the brain and adjacent cranium were compiled to generate coronal and sagittal images to satisfying neuroanatomy interest.1 In every 10 sectioned images (intervals, 1 mm), 231 structures were delineated semiautomatically or manually on Photoshop CS3 version 10 (Adobe Systems) to produce outlined images. All outlines of each structure were filled with a specific color (color depth, 8-bit color; file format, BMP [bitmap]).1,3,4

Browsing Software of Two-Dimensional Images Two hundred thirty-five pairs of sectioned images and outlined images with 1-mm intervals were prepared. The resolution of all images was decreased to be 1000  997 to lower their file size. The images before and after reducing resolution represented full and reduced data, respectively (Figs. 1A, B). The software was developed to browse the sectioned and outlined images in the C# language of Microsoft Visual Studio .NET 2003 (Microsoft Corp, Seattle, WA).6

Surface Models Among the 231 outlined structures, 178 had been constructed at the time of submission (Table 1). Mimics version 10.01 (Materialise, Leuven, Belgium) recognized and clustered the serial outlines of each structure considering the brightness of the designated colors. After stacking the outlines, surface reconstruction of all structures was simultaneously achieved.12Y16 The surface models were painted individually or by systems to distinguish the collective structures (Table 1). The color of skin was made semitransparent to show inside (Fig. 2). In each surface model, the outlines were removed, and triangular surfaces were appropriately reduced in number. Inadequate surface models, resulting from incorrect demarcation, were revised by anatomists.4,7,17 To further reduce the file size, the number of triangular surfaces was decreased as far as possible without altering the original shape of the models. For very complicated structures (eg, temporal bone), the decreased surface triangulation was very slight. The surface models before and after this simplification were named full data and reduced data, in that order (Figs. 1C, D). File format of the composites of the reduced data was converted from Mimics to stereolithography to be readable on the Maya software.

TABLE 1. 178 Structures That Were Surface Reconstructed Among 231 Outlined Ones Systems

Structures

Integumentary (2) Muscular (44)

Skin* Masseter,* lateral pterygoid muscle,* medial pterygoid muscle,* longus colli,* longus capitis,* sternocleidomastoid muscle,* rectus capitis posterior major,* rectus capitis posterior minor,* rectus capitis anterior,* rectus capitis lateralis,* obliquus capitis superior,* obliquus capitis inferior,* digastric muscle,* stylohyoid muscle,* mylohyoid muscle,* sternohyoid muscle,* longissimus,* trapezius ,* levator scapulae,* thyrohyoid muscle,* splenius capitis,* semispinalis cervicis* Cranium without mandible,* mandible, hyoid bone, first cervical vertebra, second cervical vertebra, third cervical vertebra, fourth cervical vertebra, fifth cervical vertebra, sixth cervical vertebra Intervertebral disc (C IIYC III), intervertebral disc (C IIIYC IV), intervertebral disc (C IVYC V), intervertebral disc (C VYC VI) Common carotid artery,* external carotid artery,* superficial temporal artery,* maxillary artery,* internal carotid artery,* vertebral artery,* anterior cerebral artery, pericallosal artery, middle cerebral artery, posterior communicating artery, posterior cerebral artery, basilar artery Internal jugular vein,* transverse sinus,* sigmoid sinus,* inferior petrosal sinus,* superior petrosal sinus,* emissary vein,* confluence of sinuses, occipital sinus, superior sagittal sinus, inferior sagittal sinus, straight sinus, cavernous sinus, diploic vein, superior cerebral vein, inferior cerebral vein, great cerebral vein Cornea and sclera,* lens,* levator palpebrae superioris,* superior rectus,* inferior rectus,* medial rectus,* lateral rectus,* superior oblique muscle,* inferior oblique muscle* External acoustic meatus,* stapes,* incus,* malleus,* auditory tube,* cochlea,* anterior semicircular duct,* posterior semicircular duct,* lateral semicircular duct* Substantia nigra,* red nucleus,* thalamus,* cerebral hemisphere,* lateral ventricle,* caudate nucleus,* putamen,* globus pallidus,* spinal cord, brainstem, fourth ventricle, cerebellum, corpus callosum, fornix Optic nerve,* oculomotor nerve,* ophthalmic nerve,* frontal nerve,* chorda tympani,* abducens nerve,* facial nerve,* vestibular nerve,* cochlear nerve,* optic chiasm Pituitary gland

Skeletal (10) Articular (4) Vascular (artery) (18)

Vascular (vein) (22)

Sensory (eye) (18) Sensory (ear) (18) Central nervous (22) Peripheral nervous (19) Endocrine (1)

Numbers in parentheses are numbers of structures. *Both bilateral structures were surface reconstructed.

* 2012 Mutaz B. Habal, MD

Copyright © 2012 Mutaz B. Habal, MD. Unauthorized reproduction of this article is prohibited.

1177

The Journal of Craniofacial Surgery

Shin et al

FIGURE 3. Horizontal, coronal, and sagittal sectioned images, superimposed on the surface models.

& Volume 23, Number 4, July 2012

FIGURE 5. Paper introducing the surface models, involved in other pages of the PDF file.

Maya was used to open the three-dimensional surface models (reduced data), into which serial two-dimensional surface models including the sectioned images were embedded. For this, horizontal, coronal, and sagittal sectioned images were chosen at 10-mm intervals in advance. We then elaborated two-dimensional surface models that were the planes with no thickness and wrapped the prepared sectioned images on the models (Fig. 3).18 All files of matching three- and two-dimensional surface models were saved as stereolithography.

PDF File of Two- and Three-Dimensional Surface Models On the Deep Exploration Standard (Right Hemisphere, San Ramon, CA), the surface models were categorized into 11 systems. In each system, some of the models were divided into right and left sides, and the models were arranged in official anatomic terms (Table 1).19 The two-dimensional surface models comprising sectioned images were regarded as other systems during this task. After finishing the coordination, all models were gathered in 1 VRML (virtual reality modeling language) file. By use of the 3D Reviewer software accompanying Acrobat 9.0 Pro Extended (Adobe Systems), the VRML file was inputted to a PDF file.20 When the PDF file was opened on Adobe Reader, the systematized anatomic terms of the surface models were displayed in the left top (model tree window) (Fig. 4A), and surface models were exhibited to the right (Fig. 4C). It was advantageous to anatomy learning that the models were assembled according to dissection levels. So, in Acrobat, we created bookmarks of the PDF file. The bookmarks were as follows: all structures (oblique view), half (left) removed, skin removed, muscles removed, bones removed, blood vessels removed, eyes and ears removed, and cerebral hemisphere removed (Figs. 2 and 4B). Using Acrobat, we attached the published articles explaining the sectioned images and surface models of male head. In the PDF files, page 1 was already occupied by all surface models, so the articles were put into the remaining pages (Fig. 5).

RESULTS The installation file (Browsing software [male head]; size, 93 MB) and PDF file (PDF file [male head]; size, 83 MB) that were constructed are freely available at the homepage of the Department

FIGURE 4. PDF file, composed of the model tree window (A), bookmark window (B), and surface models (C).

1178

FIGURE 6. Browsing software displaying structure name when the mouse pointer flies over a specific structure on the outlined image (A) and the sectioned image (B).

of Anatomy at Ajou University School of Medicine (anatomy.co.kr). The files can be downloaded without charge or registration.

Browsing Software of Two-Dimensional Images Implementation of the installation file resulted in automatic production of the main folder including executing file as well as the 2 subfolders containing sectioned images and outlined images. The sectioned and outlined images (reduced data) could be used for diverse achievements by others. If no longer needed, entire folders and files could be erased by the uninstallation procedure. The browsing software consisted of a tool bar for selecting images in the left side and a couple of images in the right side. Among the 2 images, one was larger; the sizes of the images could be interchanged by dragging the mouse (Fig. 6). Using the software, images could be easily browsed in real time. The user was able to select the images either by using the scroll bar or by typing an image number into the program. The neighboring images were continuously displayed by clicking software tools (Fig. 6). The software images showed names of the outlined structures. If a user located the mouse pointer on a structure in the outlined images, its name appeared as the tool tip text beside the mouse pointer. The structures’ names could be seen even in the sectioned images (Fig. 6).

PDF File of Two- and Three-Dimensional Surface Models Surface models in the PDF file were explored on Adobe Reader as follows. Users were capable of mixed display of three-dimensional surface models using the check boxes of the systems and individual structures (Fig. 7; Table 1). Two-dimensional surface models of

FIGURE 7. Highlighting of the surface models ‘‘cerebellum’’ (A), ‘‘central nervous system’’ (B), ‘‘all structures’’ (C) by repeated mouse clicking, which coincides with highlighting of their texts in the model tree window.

* 2012 Mutaz B. Habal, MD

Copyright © 2012 Mutaz B. Habal, MD. Unauthorized reproduction of this article is prohibited.

The Journal of Craniofacial Surgery

& Volume 23, Number 4, July 2012

FIGURE 8. Inferior view (A) and superior view (B) of the surface models and sectioned image showing locational relationship of caudate nucleus, lentiform nucleus, thalamus, and lateral ventricle.

sectioned images in horizontal, coronal, and sagittal planes could be exhibited too as if they were items of the other 3 systems. As a result, it was possible to superimpose the sectioned images on the threedimensional surface models (Fig. 3). The established bookmarks enabled users to see the surface models with sequential structures removed (Fig. 2). The surface models could be suitably zoomed in and zoomed out. They were freely rotated and shifted. The colors of surface models, already decided by developers, might become semitransparent to demonstrate the behind models. Besides, toolbar of the Adobe Reader was available for diverse expositions of the surface models. If the user clicked a model, its anatomic name was highlighted in the model tree window. The second, third, and fourth clicks of the same model resulted in spotlight of larger and larger groups involving the structure, until all the systems of the head were selected. Reverse highlighting was possible as well; clicking interesting texts in the model tree window prompted the appearance of the matching surface models individually or by system (Fig. 7). The supplementary article reprints in the presented PDF file could be demonstrated by picking the different pages (Fig. 5). Using Acrobat, the surface models (reduced data) could be extracted from the PDF file to another three-dimensional viewer software. This was possible because the authors did not use the protection option during the procedure to put surface models in PDF file on 3D Reviewer. Locational relationship of the adjacent structures was conserved after taking out the individual surface models of the PDF file.

DISCUSSION The current research aimed to make the sectioned images and surface models of the head generated as part of the Visible Korean Project accessible and informative. For that, 3 policies were established. The first policy is that the data have to be readily obtainable from a homepage and observable without difficulty. This policy was fulfilled by downloading the sectioned images and surface models for viewing on the special software and Abode Reader, respectively. Resolution of the illustrations was reduced for easy distribution and fast monitoring. The reduced data are still interpretable by medical students (Fig. 1)6,7 and can be extracted from the browsing software and PDF file to be evaluated by other researchers. They will be supplied with the full data of sectioned images, outlined images, and surface models after agreement process with authors. Our persisting goal is the wide distribution of Visible Korean data. It is hoped that the complete version could be a rich information resource for use in virtual reality for medical students and clinicians. The second policy is that the graphic data are explained by annotation. Without labels, the two-dimensional images and threedimensional models are merely fancy pictures and not a rich source of learning. Automatic annotation is accomplished using the outlined images. A structure’s name becomes visible according to the location of the mouse pointer either on the outlined images or on the corresponding sectioned images. This labeling is realized by

Surface Models of a Cadaver Head

programming based on the fact that the computer identifies the specific colors, with which the individual structures are filled (Fig. 6).6 Computerized annotation is feasible in the PDF file format. Highlighting of surface models and texts coincides with each other (Fig. 7). This function facilitates the advantageous acquisition of anatomy knowledge indeed. The surface models are built from the outlined images, so delineation is regarded as the crucial means for the computerized annotation in both two-dimensional images and three-dimensional models. The third policy is that the two-dimensional images and threedimensional models from the identical subject are displayed together. The combined demonstration is required because the sectioned images provide users with comprehensive information of the surface models. Conversely, the stereoscopic surface models help users grasp the complicated sectioned images. In the PDF file, not only the horizontal sectioned images but also the coronal and sagittal ones are overlapped in the three-dimensional models (Fig. 3). A potential merit is that the sectioned images may be replaced with the equivalent outlined images.1 Among the innumerable combinations of the two-dimensional images and three-dimensional models, suitable selections are required as an anatomic learning tool. For example, a student can produce the best mixture of twoand three-dimensional resources to account for magnetic resonance images showing caudate nucleus, lentiform nucleus, and thalamus (Fig. 8). Even without the two-dimensional images, the PDF file can be used by students to make the best scene for explaining a given subject (eg, morphology of the dural venous sinuses in the skull) (Fig. 9). In terms of the combination, the browsing software compensates two-dimensional images in the PDF file. In the software, the sectioned images in high resolution can be seen at 1-mm intervals rather than at 10-mm intervals. The software is equipped with the outlined images of 231 structures rather than 178 reconstructed structures.6 Despite this exciting beginning, improvements of the browsing software and PDF file are possible. First, the more detailed structures are delineated, the better will be the rendering of the outlined images and surface models. The high-quality sectioned images have already been published in an atlas, where thousands of brain constituents are annotated. On the basis of the extraordinarily detailed annotation in the book, someone might attempt additional delineation and surface reconstruction for one’s own intention.21 Second, full versions of two- and three-dimensional data will enable better visualization. We used a reduced version of sectioned image with lower resolution and surface models with fewer polygons (Fig. 1). In addition, original intervals (0.1 mm) of the sectioned images were increased to 1 mm in the browsing software and even to 10 mm in the PDF file. The restriction would naturally disappear by further improvement of the computer and communication. Third, the browsing software and PDF file can be combined into 1 program containing all two-dimensional images, threedimensional models, and satisfactory tools for user convenience. To achieve this goal, well-organized programming after collection of the client demands is required. The sectioned images, outlined images, and surface models of male head in the browsing software and PDF file will hopefully

FIGURE 9. Surface models of right dural venous sinuses and cranium.

* 2012 Mutaz B. Habal, MD

Copyright © 2012 Mutaz B. Habal, MD. Unauthorized reproduction of this article is prohibited.

1179

The Journal of Craniofacial Surgery

Shin et al

assist students in the study of the topographic anatomy and sectional anatomy of the head, as well as neuroanatomy. In addition, it is expected that the raw data will stimulate researchers to develop three-dimensional medical simulators for improvement of clinical practice in neurosurgery and otorhinolaryngology.

10.

REFERENCES

12.

1. Park JS, Chung MS, Shin DS, et al. Sectioned images of the cadaver head including the brain and correspondences with ultrahigh field 7.0 T MRIs. Proc IEEE 2009;97:1988Y1996 2. Park JS, Chung MS, Park HS, et al. A proposal of new reference system for the standard axial, sagittal, coronal planes of brain based on the serially-sectioned images. J Korean Med Sci 2010;25:135Y141 3. Park JS, Chung MS, Chi J-G, et al. Segmentation of cerebral gyri in the sectioned images by referring to volume model. J Korean Med Sci 2010;25:1710Y1715 4. Jang HG, Chung MS, Shin DS, et al. Segmentation and surface reconstruction of the detailed ear structures, identified in sectioned images. Anat Rec 2011;294:559Y564 5. Shin DS, Chung MS, Park JS. Systematized methods of surface reconstruction from the serial sectioned images of a cadaver head. J Craniofac Surg 2012;23:190Y194 6. Shin DS, Chung MS, Park HS, et al. Browsing software of the Visible Korean data used for teaching sectional anatomy. Anat Sci Educ 2011;4:327Y332 7. Park JS, Shin DS, Chung MS, et al. Technique of semiautomatic surface reconstruction of the Visible Korean Human data using commercial software. Clin Anat 2007;20:871Y879 8. Shin DS, Chung MS, Lee JW, et al. Advanced surface reconstruction technique to build detailed surface models of liver and neighboring structures from the Visible Korean Human. J Korean Med Sci 2009;24:375Y383 9. Shin DS, Park JS, Lee S-B, et al. Surface model of the

1180

11.

13.

14.

15.

16.

17.

18. 19.

20. 21.

& Volume 23, Number 4, July 2012

gastrointestinal tract constructed from the Visible Korean. Clin Anat 2009;22:601Y609 Shin DS, Chung MS, Park JS, et al. Three-dimensional surface models of detailed lumbosacral structures reconstructed from the Visible Korean. Anal Anat 2011;193:64Y70 Shin DS, Park JS, Shin B-S, et al. Surface models of the male urogenital organs built from the Visible Korean using popular software. Anat Cell Biol 2011;44:151Y159 Lee YS, Seon JK, Shin VI, et al. Anatomical evaluation of CT-MRI combined femoral model. Biomed Eng Online 2008;7:6 Benkhadra M, Savoldelli G, Fournier R, et al. A new anatomical technique to investigate nerves by imagery. Surg Radiol Anat 2009;31:221Y224 Yavuz MS, Buyukkurt MC, Tozoglu S, et al. Evaluation of volumetry and density of mandibular symphysis bone grafts by three-dimensional computed tomography. Dent Traumatol 2009;25:475Y479 Storck C, Juergens P, Fischer C, et al. Three-dimensional imaging of the larynx for pre-operative planning of laryngeal framework surgery. Eur Arch Otorhinolaryngol 2010;267:557Y563 Puchwein P, Schildhauer TA, Scho¨ffmann S, et al. Three-dimensional morphometry of the proximal ulna: a comparison to currently used anatomically preshaped ulna plates. J Shoulder Elbow Surg 2011 [Epub ahead of print Oct 19] Park JS, Jung Y-W, Lee JW, et al. Generating useful images for medical applications from the Visible Korean Human. Comput Methods Programs Biomed 2008;92:257Y266 Lanier L. Advanced Maya Texturing and Lighting. Indianapolis, IN: Wiley, 2008 FCAT (Federative Committee on Anatomical Terminology). Terminologia Anatomica: International Anatomical Terminology. New York: Thieme, 1998 Sahlin D. How to Do Everything Adobe Acrobat 9. New York: McGraw-Hill, 2008 Cho ZH. 7.0 Tesla MRI Brain Atlas. In Vivo Atlas With Cryomacrotome Correlation. New York: Springer, 2009

* 2012 Mutaz B. Habal, MD

Copyright © 2012 Mutaz B. Habal, MD. Unauthorized reproduction of this article is prohibited.