1. May 21-27, 2016 American Society of Neuroradiology (ASNR) Washington, DC Ramin Javan, MD, Shahram Majidi, MD, Ardalan Tangestanipoor, MD George Washington University Medical Center Abstract No: O eEdE-173

2. Neither I nor my immediate family members have a financial relationship with a commercial organization that may have a direct or indirect interest in the content. Financial Disclosure

3. Purpose Several algorithms have been developed for perfecting automatic and semi-automatic segmentation methods, especially in the setting of three-dimensional (3D) reconstruction, using commercial or open source software.

4. Edge Enhancement Region Growing

5. Purpose Automatic segmentation techniques are inherently limited in their ability to create error-free results due to lack of adequate contrast between tissue interfaces, high degree of anatomical variability in the body or at times suboptimal quality of acquired images. Therefore, manual refining of segmentation becomes an essential step in creating ideal results.

6. Purpose The purpose of this exhibit is to demonstrate the use of pen displays in manual segmentation in the setting of 3D printing.

7. What is 3D Printing? Rapid prototyping or 3D printing consists of various techniques to construct a physical 3D scale model utilizing computer aided design. In radiology, RP has allowed for the creation of 3D anatomic models for teaching, research, and interventional planning. Siamese Twins - 3D Print © Royal College of Art London

8. Background Rapid prototyping has become an innovative method of fast and cost-effective production of 3D models for manufacturing. It comprises a variety of automated manufacturing techniques such as selective laser sintering, stereolithography, and three- dimensional printing, which use virtual 3D data sets to fabricate solid forms in a layer-by-layer technique. Segmentation is the most important initial step in processing DICOM images, allowing for formation of 3D surface reconstructions that can eventually be 3D printed.

9. Methodology The Mouse has limitations: Inability to take full advantage of the high level of detail afforded by the dexterity of the fingers in performing remarkably fine tasks. Lines and borders that are being traced are displayed on a monitor, which is separate from the user's surface of interaction, i.e., the mouse pad.

10. Pen Display Tracing on screen Expensive Special order Full manual dexterity Pen Tablet Tracing off screen Cheap Somewhat available Full manual dexterity Input Devices Mouse Tracing off screen Very Cheap Widely available Slightly limited manual dexterity

11. Pen Displays were 15% faster than traditional mouse. No significant difference between Pen Tablet and traditional mouse in time saving. No difference between the devices in precision of measurement.

14. Discussion The important factors to consider and compare: Accuracy Precision Rapidity Cost User Friendliness Practicality

15. Discussion Based on the results of the 3 studies that were found regarding this topic, Pen Displays should serve as the natural progression in manual segmentation. Manual segmentation has become more vital nowadays due to the popularity of 3D printing. Upfront cost and the initial learning curve in becoming familiar and comfortable with Pen Displays are likely the biggest challenges. Initial cost likely negated by the time-saving aspect

16. Summary Display-based user interfaces, specifically pen displays, may improve accuracy and efficiency of manual segmentation. These devices do incur a higher initial cost and the early learning curve is steep for most users, but the long-term benefits likely outweigh these factors especially in high- volume settings. Comparison studies and user surveys are necessary to support this theory.

17. References Chen JY, Seagull FJ, Nagy P, Lakhani P, Melhem ER, Siegel EL, Safdar NM.Computer input devices: neutral party or source of significant error in manual lesion segmentation? J Digit Imaging. 2011 Feb;24(1):135-41. Perandini S, Faccioli N, Inama M, Pozzi Mucelli R. Freehand liver volumetry by using an electromagnetic pen tablet: accuracy, precision, and rapidity. J Digit Imaging. 2011 Apr;24(2):360-5. McWalter EJ, Wirth W, Siebert M, von Eisenhart-Rothe RM, Hudelmaier M, Wilson DR, Eckstein F. Use of novel interactive input devices for segmentation of articular cartilage from magnetic resonance images. Osteoarthritis Cartilage. 2005 Jan;13(1):48-53. Esses SJ, Berman P, Bloom AI, Sosna J. "Clinical Applications of Physical 3D Models Derived From MDCT Data and Created by Rapid Prototyping." AJR 2011 Jun;196(6):W683-8. Barnes E. "3D CT printing simplifies neurosurgical planning.” Auntminnie.com,, Accessed April 9, 2012. “Selective Laser Sintering.” Wikipedia,, Accessed April 9, 2012. i.Materialise.com., Accessed April 9, 2012.

18. Ramin Javan, MD George Washington University Medical Center Washington, DC ramin.javan@mfa.gwu.edu Siamese Twins - 3D Print © Royal College of Art London