Intensity-invariant texture analysis for classification of bi-rads category 3 breast masses

Chung Ming Lo, Woo Kyung Moon, Chiun Sheng Huang, Jeon Hor Chen, Min Chun Yang, Ruey Feng Chang

Research output: Contribution to journalArticle

19 Citations (Scopus)

Abstract

Radiologists likely incorrectly classify benign masses as Breast Imaging Reporting and Data System (BIRADS) category 3. A computer-aided diagnosis (CAD) system was developed in this study as a second viewer to avoid misclassification of carcinomas. Sixty-nine biopsy-proven BI-RADS category 3 masses, including 21 malignant and 48 benign masses, were used to evaluate the CAD system. To improve the texture features, gray-scale variations between images were reduced by transforming pixels into intensity-invariant ranklet coefficients. The textures of the tumor and speckle pixels were extracted from the transformed ranklet images to provide more robust features than in conventionalCADsystems. As a result, tumor texture and speckle texture with ranklet transformation achieved significantly better areas under the receiver operating characteristic curve (Az) compared with those without ranklet transformation (Az = 0.83 vs. 0.58 and Az = 0.80 vs. 0.56, p value <0.05). The improved CAD system can be a second reader to confirm the classification of BI-RADS category 3 masses.

Original languageEnglish
Pages (from-to)2039-2048
Number of pages10
JournalUltrasound in Medicine and Biology
Volume41
Issue number7
DOIs
Publication statusPublished - 2015
Externally publishedYes

Fingerprint

breast
Breast
textures
tumors
pixels
Information Systems
ROC Curve
data systems
gray scale
Neoplasms
readers
Carcinoma
Biopsy
receivers
cancer
curves
coefficients
Radiologists

Keywords

  • Breast cancer
  • Breast imaging and reporting data system
  • Computer-aided diagnosis
  • Ranklet
  • Ultrasound

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging
  • Radiological and Ultrasound Technology
  • Biophysics

Cite this

Intensity-invariant texture analysis for classification of bi-rads category 3 breast masses. / Lo, Chung Ming; Moon, Woo Kyung; Huang, Chiun Sheng; Chen, Jeon Hor; Yang, Min Chun; Chang, Ruey Feng.

In: Ultrasound in Medicine and Biology, Vol. 41, No. 7, 2015, p. 2039-2048.

Research output: Contribution to journalArticle

Lo, Chung Ming ; Moon, Woo Kyung ; Huang, Chiun Sheng ; Chen, Jeon Hor ; Yang, Min Chun ; Chang, Ruey Feng. / Intensity-invariant texture analysis for classification of bi-rads category 3 breast masses. In: Ultrasound in Medicine and Biology. 2015 ; Vol. 41, No. 7. pp. 2039-2048.
@article{34cf457936584ec88ed3f64bf4850c85,
title = "Intensity-invariant texture analysis for classification of bi-rads category 3 breast masses",
abstract = "Radiologists likely incorrectly classify benign masses as Breast Imaging Reporting and Data System (BIRADS) category 3. A computer-aided diagnosis (CAD) system was developed in this study as a second viewer to avoid misclassification of carcinomas. Sixty-nine biopsy-proven BI-RADS category 3 masses, including 21 malignant and 48 benign masses, were used to evaluate the CAD system. To improve the texture features, gray-scale variations between images were reduced by transforming pixels into intensity-invariant ranklet coefficients. The textures of the tumor and speckle pixels were extracted from the transformed ranklet images to provide more robust features than in conventionalCADsystems. As a result, tumor texture and speckle texture with ranklet transformation achieved significantly better areas under the receiver operating characteristic curve (Az) compared with those without ranklet transformation (Az = 0.83 vs. 0.58 and Az = 0.80 vs. 0.56, p value <0.05). The improved CAD system can be a second reader to confirm the classification of BI-RADS category 3 masses.",
keywords = "Breast cancer, Breast imaging and reporting data system, Computer-aided diagnosis, Ranklet, Ultrasound",
author = "Lo, {Chung Ming} and Moon, {Woo Kyung} and Huang, {Chiun Sheng} and Chen, {Jeon Hor} and Yang, {Min Chun} and Chang, {Ruey Feng}",
year = "2015",
doi = "10.1016/j.ultrasmedbio.2015.03.003",
language = "English",
volume = "41",
pages = "2039--2048",
journal = "Ultrasound in Medicine and Biology",
issn = "0301-5629",
publisher = "Elsevier USA",
number = "7",

}

TY - JOUR

T1 - Intensity-invariant texture analysis for classification of bi-rads category 3 breast masses

AU - Lo, Chung Ming

AU - Moon, Woo Kyung

AU - Huang, Chiun Sheng

AU - Chen, Jeon Hor

AU - Yang, Min Chun

AU - Chang, Ruey Feng

PY - 2015

Y1 - 2015

N2 - Radiologists likely incorrectly classify benign masses as Breast Imaging Reporting and Data System (BIRADS) category 3. A computer-aided diagnosis (CAD) system was developed in this study as a second viewer to avoid misclassification of carcinomas. Sixty-nine biopsy-proven BI-RADS category 3 masses, including 21 malignant and 48 benign masses, were used to evaluate the CAD system. To improve the texture features, gray-scale variations between images were reduced by transforming pixels into intensity-invariant ranklet coefficients. The textures of the tumor and speckle pixels were extracted from the transformed ranklet images to provide more robust features than in conventionalCADsystems. As a result, tumor texture and speckle texture with ranklet transformation achieved significantly better areas under the receiver operating characteristic curve (Az) compared with those without ranklet transformation (Az = 0.83 vs. 0.58 and Az = 0.80 vs. 0.56, p value <0.05). The improved CAD system can be a second reader to confirm the classification of BI-RADS category 3 masses.

AB - Radiologists likely incorrectly classify benign masses as Breast Imaging Reporting and Data System (BIRADS) category 3. A computer-aided diagnosis (CAD) system was developed in this study as a second viewer to avoid misclassification of carcinomas. Sixty-nine biopsy-proven BI-RADS category 3 masses, including 21 malignant and 48 benign masses, were used to evaluate the CAD system. To improve the texture features, gray-scale variations between images were reduced by transforming pixels into intensity-invariant ranklet coefficients. The textures of the tumor and speckle pixels were extracted from the transformed ranklet images to provide more robust features than in conventionalCADsystems. As a result, tumor texture and speckle texture with ranklet transformation achieved significantly better areas under the receiver operating characteristic curve (Az) compared with those without ranklet transformation (Az = 0.83 vs. 0.58 and Az = 0.80 vs. 0.56, p value <0.05). The improved CAD system can be a second reader to confirm the classification of BI-RADS category 3 masses.

KW - Breast cancer

KW - Breast imaging and reporting data system

KW - Computer-aided diagnosis

KW - Ranklet

KW - Ultrasound

UR - http://www.scopus.com/inward/record.url?scp=84942990272&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84942990272&partnerID=8YFLogxK

U2 - 10.1016/j.ultrasmedbio.2015.03.003

DO - 10.1016/j.ultrasmedbio.2015.03.003

M3 - Article

C2 - 25843514

AN - SCOPUS:84942990272

VL - 41

SP - 2039

EP - 2048

JO - Ultrasound in Medicine and Biology

JF - Ultrasound in Medicine and Biology

SN - 0301-5629

IS - 7

ER -