Technical note Three-dimensional anatomical accuracy of cranial models created by rapid prototyping techniques validated using a neuronavigation station Vicknes Waran a,b,⇑ , P. Devaraj a , T. Hari Chandran a , K.A. Muthusamy a , Alwin Kumar Rathinam b , Yuwaraj Kumar Balakrishnan b , Tan Su Tung b , R. Raman c , Z.A.A. Rahman b,d a Division of Neurosurgery, Department of Surgery, Faculty of Medicine, University Malaya, Lembah Pantai, Kuala Lumpur 50603, Malaysia b Centre for Biomedical and Technology Integration, University Malaya, Kuala Lumpur, Malaysia c Department of Otorhinolaryngology, Faculty of Medicine, University Malaya, Kuala Lumpur, Malaysia d Division of Maxillofacial Surgery, Faculty of Dentistry, University Malaya, Kuala Lumpur, Malaysia article info Article history: Received 24 May 2011 Accepted 25 July 2011 Keywords: Neuronavigation Rapid prototyping technique Surgical heads abstract In neurosurgery and ear, nose and throat surgery the application of computerised navigation systems for guiding operations has been expanding rapidly. However, suitable models to train surgeons in using nav- igation systems are not yet available. We have developed a technique using an industrial, rapid prototyp- ing process from which accurate spatial models of the cranium, its contents and pathology can be reproduced for teaching. We were able to register, validate and navigate using these models with com- mon available navigation systems such as the Medtronic StealthStation S7 Ò . Ó 2011 Elsevier Ltd. All rights reserved. 1. Introduction Teaching models today that allow procedures to be simulated for both trainees and practicing surgeons, and these models are especially useful if cadavers are difficult to obtain or if the equip- ment and tools for cadaveric dissections are too expensive. 1–9 In neurosurgery and ear, nose and throat surgery the applica- tion of computerised navigation systems for guiding surgery has been expanding rapidly since 2000. 10 These systems are used to plan and define tumours and vital structures and to track surgical instruments. They are used in most operations, including biopsies, and for tracking devices such as endoscopes during surgery. 11–15,6,9 Unfortunately suitable models to train surgeons in the use of these systems have been unavailable thus far. Rapid prototyping (RP) or three-dimensional (3D) printing ma- chines have been available since 2000 to convert 3D computer- aided drawings into 3D objects made from powder or other mate- rials. These objects are printed in layers from the base upwards un- til the entire object has been created. These machines have been used mainly in engineering and the automobile manufacturing industry to prototype models of intended parts. We have developed a technique using industrial RP processes to reproduce accurate spatial models of the cranium, its contents and pathology. 11–26,2–7,9 We have been able to register and navigate on these spatial models using a commonly available navigation sys- tem. Therefore, this study aimed to assess the accuracy of these models compared to patient CT scans for neuronavigation. 2. Methods 2.1. Model creation Each cranial model was created by initially obtaining a fine slice (1-mm) CT scan in the Digital Imaging and COmmunications in Medicine (DICOM) format. This information was then processed using a computing algorithm to generate a three-dimensional (3D) biomodel in silico (performed on a computer or via computer simulation). The structural data for the biomodel were then read by an industrial RP machine (model ZPrinter Ò 450, manufactured by Z Corporation, 32 Second Avenue, Burlington, MA 01803 USA), which reproduced a physical model of the patient’s skull (Fig. 1). 26 Using the same technique we were able to create the soft tissue representing the face. This soft tissue section of the model was de- signed to overlay accurately over the model skull (Fig. 2). This overlay allowed surface registration techniques to be used when registering the model to the navigation system. 2.2. Model verification Six skulls were created from CT data of patients scanned for var- ious reasons using the above technique. The datasets were loaded into a StealthStation S7 Ò navigation station (Medtronic;Minneapo- 0967-5868/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jocn.2011.07.031 ⇑ Corresponding author. Tel.: +60 3 7949 2070; fax: +60 3 7958 6360. E-mail address: cmvwaran@gmail.com (V. Waran). Journal of Clinical Neuroscience 19 (2012) 574–577 Contents lists available at SciVerse ScienceDirect Journal of Clinical Neuroscience journal homepage: www.elsevier.com/locate/jocn