Optimized multi-level elongated quinary patterns for the assessment of thyroid nodules in ultrasound images

Raghavendra U 1 , Gudigar A 2 , Maithri M 3 , Gertych A 4 , Meiburger KM 5 , Yeong CH 6 Show all authors , Madla C 7 , Kongmebhol P 7 , Molinari F 5 , Ng KH 6 , Acharya UR 8

Affiliations 

  • 1 Department of Instrumentation and Control Engineering, Manipal Institute of Technology, Manipal Academy of Higher Education, Manipal, 576104, India. Electronic address: raghavendra.u@manipal.edu
  • 2 Department of Instrumentation and Control Engineering, Manipal Institute of Technology, Manipal Academy of Higher Education, Manipal, 576104, India
  • 3 Department of Mechatronics Engineering, Manipal Institute of Technology, Manipal Academy of Higher Education, Manipal, 576104, India
  • 4 Department of Surgery, Department of Pathology and Laboratory Medicine, Cedars-Sinai Medical Center, Los Angeles, CA, USA
  • 5 Department of Electronics and Telecommunications, Politecnico di Torino, Italy
  • 6 Department of Biomedical Imaging, University of Malaya, Kuala Lumpur, Malaysia
  • 7 Department of Radiology, Faculty of Medicine, Chiang Mai University, Chiang Mai, 50200, Thailand
  • 8 Department of Electronics and Computer Engineering, Ngee Ann Polytechnic, Clementi, 599489, Singapore; Department of Biomedical Engineering, School of Science and Technology, SIM University, Clementi, 599491, Singapore; Department of Biomedical Engineering, Faculty of Engineering, University of Malaya, Kuala Lumpur, 50603, Malaysia
Comput Biol Med, 2018 04 01;95:55-62.
PMID: 29455080 DOI: 10.1016/j.compbiomed.2018.02.002

Abstract

Ultrasound imaging is one of the most common visualizing tools used by radiologists to identify the location of thyroid nodules. However, visual assessment of nodules is difficult and often affected by inter- and intra-observer variabilities. Thus, a computer-aided diagnosis (CAD) system can be helpful to cross-verify the severity of nodules. This paper proposes a new CAD system to characterize thyroid nodules using optimized multi-level elongated quinary patterns. In this study, higher order spectral (HOS) entropy features extracted from these patterns appropriately distinguished benign and malignant nodules under particle swarm optimization (PSO) and support vector machine (SVM) frameworks. Our CAD algorithm achieved a maximum accuracy of 97.71% and 97.01% in private and public datasets respectively. The evaluation of this CAD system on both private and public datasets confirmed its effectiveness as a secondary tool in assisting radiological findings.

* Title and MeSH Headings from MEDLINE®/PubMed®, a database of the U.S. National Library of Medicine.

MeSH terms
Similar publications