Computer-aided focal liver lesion detection
- PDF / 1,321,612 Bytes
- 15 Pages / 595.276 x 790.866 pts Page_size
- 0 Downloads / 182 Views
ORIGINAL ARTICLE
Computer-aided focal liver lesion detection Yanling Chi · Jiayin Zhou · Sudhakar K. Venkatesh · Su Huang · Qi Tian · Tiffany Hennedige · Jimin Liu
Received: 30 September 2012 / Accepted: 11 March 2013 / Published online: 31 March 2013 © CARS 2013
Abstract Purpose Our aim is to develop an automatic method which can detect diverse focal liver lesions (FLLs) in 3D CT volumes. Method A hybrid generative-discriminative framework is proposed. It first uses a generative model to describe nonlesion components and then identifies all candidate FLLs within a 3D liver volume by eliminating non-lesion components. It subsequently uses a discriminative approach to suppress false positives with the advantage of tumoroid,
Y. Chi (B) · S. Huang · Q. Tian · J. Liu Singapore Bioimaging Consortium, Agency for Science, Technology and Research, 30 Biopolis Street, #07-01, Matrix, 138671 Singapore, Singapore e-mail: [email protected]; [email protected] S. Huang e-mail: [email protected] Q. Tian e-mail: [email protected] J. Liu e-mail: [email protected] J. Zhou Institute for Infocomm Research, Agency for Science, Technology and Research, Singapore, Singapore e-mail: [email protected] S. K. Venkatesh · T. Hennedige Department of Diagnostic Imaging, National University Hospital, Singapore, Singapore T. Hennedige e-mail: [email protected] S. K. Venkatesh Mayo Clinic, Rochester, MN, USA e-mail: [email protected]
a novel measurement combining three shape features spherical symmetry, compactness and size. Results This method was tested on 71 abdominal CT datasets (5,854 slices from 61 patients, with 261 FLLs covering six pathological types) and evaluated using the freeresponse receiver operating characteristic (FROC) curves. Overall, it achieved a true positive rate of 90 % with one false positive per liver. It degenerated gently with the decrease in lesion sizes to 30 ml. It achieved a true-positive rate of 36 % when tested on the lesions less than 4 ml. The average computing time of the lesion detection is 4 min and 28 s per CT volume on a PC with 2.67 GHz CPU and 4.0 GB RAM. Conclusions The proposed method is comparable to the radiologists’ visual investigation in terms of efficiency. The tool has great potential to reduce radiologists’ burden in going through thousands of images routinely. Keywords Computer-aided detection · 3D focal liver lesion detection · Background subtraction · Non-lesion modeling Introduction Imaging techniques such as ultrasound (U/S), computer tomography (CT) and magnetic resonance imaging (MRI) are the common methods for the noninvasive diagnosis and management of liver diseases. Among them, multi-detector computer tomography (MDCT) has been increasingly used for the liver since 2000 due to its various advantages over conventional spiral CT, especially for the detection of FLLs [1–4]. Several recent studies reported the investigations on MDCT-based hepatocellular carcinoma (HCC) detection, with the resultant sensitivities ranging from 64 to 89
Data Loading...
