Level set breast mass segmentation in contrast-enhanced and non-contrast-enhanced breast CT

Hsien Chi Kuo, Maryellen L. Giger, Ingrid Reiser, John M Boone, Karen K Lindfors, Kai Yang, Alexandra Edwards

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

3 Citations (Scopus)

Abstract

Dedicated breast CT (bCT) is an emerging technology that produces 3D reconstructed images of the breast, thus allowing radiologists to detect and evaluate breast lesions in 3D. In previous work, we have developed an algorithm that combines radial gradient index (RGI) segmentation and a modified level set model for segmentation of lesions in contrast-enhanced bCT images; yielding an average overlap ratio (OR) of 0.69, which is higher than 0.4, the overlap ratio that is generally deemed "acceptable". In this study, this segmentation algorithm, with the same parameter settings, was applied to the corresponding non-contrastenhanced bCT images. The results show that the OR obtained on non-contrast images was 0.62, with the segmented lesion volumes tending to be slightly smaller as compared with those obtained on the corresponding contrast-enhanced images. These results imply that while use of contrast improves segmentation performance, the increase may not be significant, and thus, the role of noncontrast-enhanced breast CT should be further investigated.

Original languageEnglish (US)
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages697-704
Number of pages8
Volume7361 LNCS
DOIs
StatePublished - 2012
Event11th International Workshop on Breast Imaging, IWDM 2012 - Philadelphia, PA, United States
Duration: Jul 8 2012Jul 11 2012

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume7361 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other11th International Workshop on Breast Imaging, IWDM 2012
CountryUnited States
CityPhiladelphia, PA
Period7/8/127/11/12

Fingerprint

Level Set
Segmentation
Overlap
CT Image
3D Image
Gradient
Imply
Evaluate
Model

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Kuo, H. C., Giger, M. L., Reiser, I., Boone, J. M., Lindfors, K. K., Yang, K., & Edwards, A. (2012). Level set breast mass segmentation in contrast-enhanced and non-contrast-enhanced breast CT. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 7361 LNCS, pp. 697-704). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7361 LNCS). https://doi.org/10.1007/978-3-642-31271-7_90

Level set breast mass segmentation in contrast-enhanced and non-contrast-enhanced breast CT. / Kuo, Hsien Chi; Giger, Maryellen L.; Reiser, Ingrid; Boone, John M; Lindfors, Karen K; Yang, Kai; Edwards, Alexandra.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 7361 LNCS 2012. p. 697-704 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7361 LNCS).

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

Kuo, HC, Giger, ML, Reiser, I, Boone, JM, Lindfors, KK, Yang, K & Edwards, A 2012, Level set breast mass segmentation in contrast-enhanced and non-contrast-enhanced breast CT. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 7361 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 7361 LNCS, pp. 697-704, 11th International Workshop on Breast Imaging, IWDM 2012, Philadelphia, PA, United States, 7/8/12. https://doi.org/10.1007/978-3-642-31271-7_90
Kuo HC, Giger ML, Reiser I, Boone JM, Lindfors KK, Yang K et al. Level set breast mass segmentation in contrast-enhanced and non-contrast-enhanced breast CT. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 7361 LNCS. 2012. p. 697-704. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-642-31271-7_90
Kuo, Hsien Chi ; Giger, Maryellen L. ; Reiser, Ingrid ; Boone, John M ; Lindfors, Karen K ; Yang, Kai ; Edwards, Alexandra. / Level set breast mass segmentation in contrast-enhanced and non-contrast-enhanced breast CT. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 7361 LNCS 2012. pp. 697-704 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{bc264ebfc3584d43a38bb0ef453b6f50,
title = "Level set breast mass segmentation in contrast-enhanced and non-contrast-enhanced breast CT",
abstract = "Dedicated breast CT (bCT) is an emerging technology that produces 3D reconstructed images of the breast, thus allowing radiologists to detect and evaluate breast lesions in 3D. In previous work, we have developed an algorithm that combines radial gradient index (RGI) segmentation and a modified level set model for segmentation of lesions in contrast-enhanced bCT images; yielding an average overlap ratio (OR) of 0.69, which is higher than 0.4, the overlap ratio that is generally deemed {"}acceptable{"}. In this study, this segmentation algorithm, with the same parameter settings, was applied to the corresponding non-contrastenhanced bCT images. The results show that the OR obtained on non-contrast images was 0.62, with the segmented lesion volumes tending to be slightly smaller as compared with those obtained on the corresponding contrast-enhanced images. These results imply that while use of contrast improves segmentation performance, the increase may not be significant, and thus, the role of noncontrast-enhanced breast CT should be further investigated.",
author = "Kuo, {Hsien Chi} and Giger, {Maryellen L.} and Ingrid Reiser and Boone, {John M} and Lindfors, {Karen K} and Kai Yang and Alexandra Edwards",
year = "2012",
doi = "10.1007/978-3-642-31271-7_90",
language = "English (US)",
isbn = "9783642312700",
volume = "7361 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "697--704",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Level set breast mass segmentation in contrast-enhanced and non-contrast-enhanced breast CT

AU - Kuo, Hsien Chi

AU - Giger, Maryellen L.

AU - Reiser, Ingrid

AU - Boone, John M

AU - Lindfors, Karen K

AU - Yang, Kai

AU - Edwards, Alexandra

PY - 2012

Y1 - 2012

N2 - Dedicated breast CT (bCT) is an emerging technology that produces 3D reconstructed images of the breast, thus allowing radiologists to detect and evaluate breast lesions in 3D. In previous work, we have developed an algorithm that combines radial gradient index (RGI) segmentation and a modified level set model for segmentation of lesions in contrast-enhanced bCT images; yielding an average overlap ratio (OR) of 0.69, which is higher than 0.4, the overlap ratio that is generally deemed "acceptable". In this study, this segmentation algorithm, with the same parameter settings, was applied to the corresponding non-contrastenhanced bCT images. The results show that the OR obtained on non-contrast images was 0.62, with the segmented lesion volumes tending to be slightly smaller as compared with those obtained on the corresponding contrast-enhanced images. These results imply that while use of contrast improves segmentation performance, the increase may not be significant, and thus, the role of noncontrast-enhanced breast CT should be further investigated.

AB - Dedicated breast CT (bCT) is an emerging technology that produces 3D reconstructed images of the breast, thus allowing radiologists to detect and evaluate breast lesions in 3D. In previous work, we have developed an algorithm that combines radial gradient index (RGI) segmentation and a modified level set model for segmentation of lesions in contrast-enhanced bCT images; yielding an average overlap ratio (OR) of 0.69, which is higher than 0.4, the overlap ratio that is generally deemed "acceptable". In this study, this segmentation algorithm, with the same parameter settings, was applied to the corresponding non-contrastenhanced bCT images. The results show that the OR obtained on non-contrast images was 0.62, with the segmented lesion volumes tending to be slightly smaller as compared with those obtained on the corresponding contrast-enhanced images. These results imply that while use of contrast improves segmentation performance, the increase may not be significant, and thus, the role of noncontrast-enhanced breast CT should be further investigated.

UR - http://www.scopus.com/inward/record.url?scp=84864857095&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84864857095&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-31271-7_90

DO - 10.1007/978-3-642-31271-7_90

M3 - Conference contribution

AN - SCOPUS:84864857095

SN - 9783642312700

VL - 7361 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 697

EP - 704

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

ER -