Computational and Mathematical Methods in Medicine
Volume 2013 (2013), Article ID 647548, 11 pages
Research Article

Automatic Vertebral Column Extraction by Whole-Body Bone SPECT Scan

1Department of Medical Informatics, Tzu Chi University, Hualien 97004, Taiwan
2Institute of Medical Sciences, Tzu Chi University, Hualien 97004, Taiwan
3School of Medicine, Chung Shan Medical University, Taichung 40201, Taiwan
4Department of Nuclear Medicine, Chung Shan Medical University Hospital, Taichung 40201, Taiwan
5Department of Computer Science and Information Engineering, Asia University, Taichung 41354, Taiwan
6Department of Nuclear Medicine, Buddhist Tzu Chi General Hospital, Taipei Branch, New Taipei City 23142, Taiwan
7Department of Nuclear Medical, Buddhist Tzu Chi Hospital, Hualien 97004, Taiwan
8Department of Radiological Technology, Tzu Chi College of Technology, Hualien 97005, Taiwan

Received 28 November 2012; Revised 8 March 2013; Accepted 10 March 2013

Academic Editor: Chung-Ming Chen

Copyright © 2013 Sheng-Fang Huang et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.


Bone extraction and division can enhance the accuracy of diagnoses based on whole-body bone SPECT data. This study developed a method for using conventional SPECT for automatic recognition of the vertebral column. A novel feature of the proposed approach is a novel “bone graph" image description method that represents the connectivity between these image regions to facilitate manipulation of morphological relationships in the skeleton before surgery. By tracking the paths shown on the bone graph, skeletal structures can be identified by performing morphological operations. The performance of the method was evaluated quantitatively and qualitatively by two experienced nuclear medicine physicians. Datasets for whole-body bone SPECT scans in 46 lung cancer patients with bone metastasis were obtained with Tc-99m MDP. The algorithm successfully segmented vertebrae in the thoracolumbar spine. The quantitative assessment shows that the segmentation method achieved an average TP, FP, and FN rates of 95.1%, 9.1%, and 4.9%. The qualitative evaluation shows an average acceptance rate of 83%, where the data for the acceptable and unacceptable groups had a Cronbach’s alpha value of 0.718, which indicated reasonable internal consistency and reliability.