Detecting mitotic figures in breast cancer histopathology images

M. Veta, P.J. Diest, van, J.P.W. Pluim

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

25 Citations (Scopus)

Abstract

The scoring of mitotic figures is an integrated part of the Bloom and Richardson system for grading of invasive breast cancer. It is routinely done by pathologists by visual examination of hematoxylin and eosin (H&E) stained histology slides on a standard light microscope. As such, it is a tedious process prone to inter- and intra-observer variability. In the last decade, whole-slide imaging (WSI) has emerged as the "digital age" alternative to the classical microscope. The increasing acceptance of WSI in pathology labs has brought an interest in the application of automatic image analysis methods, with the goal of reducing or completely eliminating manual input to the analysis. In this paper, we present a method for automatic detection of mitotic figures in breast cancer histopathology images. The proposed method consists of two main components: candidate extraction and candidate classification. Candidate objects are extracted by image segmentation with the Chan-Vese level set method. The candidate classification component aims at classifying all extracted candidates as being a mitotic figure or a false object. A statistical classifier is trained with a number of features that describe the size, shape, color and texture of the candidate objects. The proposed detection procedure was developed using a set of 18 whole-slide images, with over 900 manually annotated mitotic figures, split into independent training and testing sets. The overall true positive rate on the testing set was 59.5% while achieving 4.2 false positives per one high power field (HPF). © 2013 SPIE.
LanguageEnglish
Title of host publicationSPIE Medical Imaging Symposium 2013: Digital Pathology, 10 - 11 February 2013, Lake Buena Vista, Florida, USA
EditorsM.N. Gurcan, xx Madabhushi
Place of PublicationWashington
PublisherSPIE
Pages867607-1/7
ISBN (Print)9780819494504
DOIs
StatePublished - 2013

Publication series

NameProceedings of SPIE
Volume8676
ISSN (Print)0277-786X

Fingerprint

Microscopes
Imaging techniques
Histology
Testing
Pathology
Image segmentation
Image analysis
Classifiers
Textures
Color

Cite this

Veta, M., Diest, van, P. J., & Pluim, J. P. W. (2013). Detecting mitotic figures in breast cancer histopathology images. In M. N. Gurcan, & X. Madabhushi (Eds.), SPIE Medical Imaging Symposium 2013: Digital Pathology, 10 - 11 February 2013, Lake Buena Vista, Florida, USA (pp. 867607-1/7). (Proceedings of SPIE; Vol. 8676). Washington: SPIE. DOI: 10.1117/12.2006626
Veta, M. ; Diest, van, P.J. ; Pluim, J.P.W./ Detecting mitotic figures in breast cancer histopathology images. SPIE Medical Imaging Symposium 2013: Digital Pathology, 10 - 11 February 2013, Lake Buena Vista, Florida, USA. editor / M.N. Gurcan ; xx Madabhushi. Washington : SPIE, 2013. pp. 867607-1/7 (Proceedings of SPIE).
@inproceedings{f15c780dcfb0482a99f8e2968e0bf9ff,
title = "Detecting mitotic figures in breast cancer histopathology images",
abstract = "The scoring of mitotic figures is an integrated part of the Bloom and Richardson system for grading of invasive breast cancer. It is routinely done by pathologists by visual examination of hematoxylin and eosin (H&E) stained histology slides on a standard light microscope. As such, it is a tedious process prone to inter- and intra-observer variability. In the last decade, whole-slide imaging (WSI) has emerged as the {"}digital age{"} alternative to the classical microscope. The increasing acceptance of WSI in pathology labs has brought an interest in the application of automatic image analysis methods, with the goal of reducing or completely eliminating manual input to the analysis. In this paper, we present a method for automatic detection of mitotic figures in breast cancer histopathology images. The proposed method consists of two main components: candidate extraction and candidate classification. Candidate objects are extracted by image segmentation with the Chan-Vese level set method. The candidate classification component aims at classifying all extracted candidates as being a mitotic figure or a false object. A statistical classifier is trained with a number of features that describe the size, shape, color and texture of the candidate objects. The proposed detection procedure was developed using a set of 18 whole-slide images, with over 900 manually annotated mitotic figures, split into independent training and testing sets. The overall true positive rate on the testing set was 59.5{\%} while achieving 4.2 false positives per one high power field (HPF). {\circledC} 2013 SPIE.",
author = "M. Veta and {Diest, van}, P.J. and J.P.W. Pluim",
year = "2013",
doi = "10.1117/12.2006626",
language = "English",
isbn = "9780819494504",
series = "Proceedings of SPIE",
publisher = "SPIE",
pages = "867607--1/7",
editor = "M.N. Gurcan and xx Madabhushi",
booktitle = "SPIE Medical Imaging Symposium 2013: Digital Pathology, 10 - 11 February 2013, Lake Buena Vista, Florida, USA",
address = "United States",

}

Veta, M, Diest, van, PJ & Pluim, JPW 2013, Detecting mitotic figures in breast cancer histopathology images. in MN Gurcan & X Madabhushi (eds), SPIE Medical Imaging Symposium 2013: Digital Pathology, 10 - 11 February 2013, Lake Buena Vista, Florida, USA. Proceedings of SPIE, vol. 8676, SPIE, Washington, pp. 867607-1/7. DOI: 10.1117/12.2006626

Detecting mitotic figures in breast cancer histopathology images. / Veta, M.; Diest, van, P.J.; Pluim, J.P.W.

SPIE Medical Imaging Symposium 2013: Digital Pathology, 10 - 11 February 2013, Lake Buena Vista, Florida, USA. ed. / M.N. Gurcan; xx Madabhushi. Washington : SPIE, 2013. p. 867607-1/7 (Proceedings of SPIE; Vol. 8676).

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

TY - GEN

T1 - Detecting mitotic figures in breast cancer histopathology images

AU - Veta,M.

AU - Diest, van,P.J.

AU - Pluim,J.P.W.

PY - 2013

Y1 - 2013

N2 - The scoring of mitotic figures is an integrated part of the Bloom and Richardson system for grading of invasive breast cancer. It is routinely done by pathologists by visual examination of hematoxylin and eosin (H&E) stained histology slides on a standard light microscope. As such, it is a tedious process prone to inter- and intra-observer variability. In the last decade, whole-slide imaging (WSI) has emerged as the "digital age" alternative to the classical microscope. The increasing acceptance of WSI in pathology labs has brought an interest in the application of automatic image analysis methods, with the goal of reducing or completely eliminating manual input to the analysis. In this paper, we present a method for automatic detection of mitotic figures in breast cancer histopathology images. The proposed method consists of two main components: candidate extraction and candidate classification. Candidate objects are extracted by image segmentation with the Chan-Vese level set method. The candidate classification component aims at classifying all extracted candidates as being a mitotic figure or a false object. A statistical classifier is trained with a number of features that describe the size, shape, color and texture of the candidate objects. The proposed detection procedure was developed using a set of 18 whole-slide images, with over 900 manually annotated mitotic figures, split into independent training and testing sets. The overall true positive rate on the testing set was 59.5% while achieving 4.2 false positives per one high power field (HPF). © 2013 SPIE.

AB - The scoring of mitotic figures is an integrated part of the Bloom and Richardson system for grading of invasive breast cancer. It is routinely done by pathologists by visual examination of hematoxylin and eosin (H&E) stained histology slides on a standard light microscope. As such, it is a tedious process prone to inter- and intra-observer variability. In the last decade, whole-slide imaging (WSI) has emerged as the "digital age" alternative to the classical microscope. The increasing acceptance of WSI in pathology labs has brought an interest in the application of automatic image analysis methods, with the goal of reducing or completely eliminating manual input to the analysis. In this paper, we present a method for automatic detection of mitotic figures in breast cancer histopathology images. The proposed method consists of two main components: candidate extraction and candidate classification. Candidate objects are extracted by image segmentation with the Chan-Vese level set method. The candidate classification component aims at classifying all extracted candidates as being a mitotic figure or a false object. A statistical classifier is trained with a number of features that describe the size, shape, color and texture of the candidate objects. The proposed detection procedure was developed using a set of 18 whole-slide images, with over 900 manually annotated mitotic figures, split into independent training and testing sets. The overall true positive rate on the testing set was 59.5% while achieving 4.2 false positives per one high power field (HPF). © 2013 SPIE.

U2 - 10.1117/12.2006626

DO - 10.1117/12.2006626

M3 - Conference contribution

SN - 9780819494504

T3 - Proceedings of SPIE

SP - 867607-1/7

BT - SPIE Medical Imaging Symposium 2013: Digital Pathology, 10 - 11 February 2013, Lake Buena Vista, Florida, USA

PB - SPIE

CY - Washington

ER -

Veta M, Diest, van PJ, Pluim JPW. Detecting mitotic figures in breast cancer histopathology images. In Gurcan MN, Madabhushi X, editors, SPIE Medical Imaging Symposium 2013: Digital Pathology, 10 - 11 February 2013, Lake Buena Vista, Florida, USA. Washington: SPIE. 2013. p. 867607-1/7. (Proceedings of SPIE). Available from, DOI: 10.1117/12.2006626