Automatic detection of bone metastasis in vertebrae by using CT images

Sheng Fang Huang, Kuo Hsien Chiang

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

Commonly diagnosed in advanced breast cancer patients, especially in vertebrae, bone metastases can appear lytic, sclerotic, or anywhere in between these extremes. Given its ability to alter the therapeutic strategy, bone metastases is a critical issue in staging and follow-up of breast cancer. This work presents a novel computer-aided diagnosis (CAD) system to detect metastasis in vertebrae by using whole body computed tomography (CT). An automated method is developed to extract ROIs of trabecular centrum from vertebrae. Eleven texture features and their inter-slice differences are then calculated for each ROI. Next, total 33 features are fed into an artificial neural network (ANN) to determine whether any abnormality occurs in the trabecular centrum. The datasets include 35 breast cancer patients who underwent a whole-body PET/CT scan between 2007 and 2011. The average sensitivity, specificity, and accuracy are 85.4%, 91.8%, and 89.7%, respectively. Capable of identifying possible bone lesions by using CT images, the proposed CAD system can incorporate with features of nuclear medicine images to increase diagnostic accuracy in an automated CAD system for estimating bone metastases quantitatively.

Original languageEnglish
Title of host publicationLecture Notes in Engineering and Computer Science
EditorsA. M. Korsunsky, S. I. Ao, David WL Hukins, Len Gelman, Andrew Hunter
PublisherNewswood Limited
Pages1166-1171
Number of pages6
ISBN (Print)9789881925213
Publication statusPublished - 2012
Event2012 World Congress on Engineering, WCE 2012 - London, United Kingdom
Duration: 2012 Jul 42012 Jul 6

Publication series

NameLecture Notes in Engineering and Computer Science
Volume2198
ISSN (Print)2078-0958

Conference

Conference2012 World Congress on Engineering, WCE 2012
CountryUnited Kingdom
CityLondon
Period12-07-0412-07-06

All Science Journal Classification (ASJC) codes

  • Computer Science (miscellaneous)

Fingerprint Dive into the research topics of 'Automatic detection of bone metastasis in vertebrae by using CT images'. Together they form a unique fingerprint.

Cite this