Texture features from mammographic images and risk of breast cancer

Armando Manduca, Michael J. Carston, John J. Heine, Christopher G. Scott, V. Shane Pankratz, Kathleen R Brandt, Thomas A. Sellers, Celine M Vachon, James R Cerhan

Research output: Contribution to journalArticle

91 Citations (Scopus)

Abstract

Mammographic percent density (PD) is a strong risk factor for breast cancer, but there has been relatively little systematic evaluation of other features in mammographic images that might additionally predict breast cancer risk. We evaluated the association of a large number of image texture features with risk of breast cancer using a clinic-based case-control study of digitized film mammograms, all with screening mammograms before breast cancer diagnosis. The sample was split into training (123 cases and 258 controls) and validation (123 cases and 264 controls) data sets. Age-adjusted and body mass index (BMI)-adjusted odds ratios (OR) per SD change in the feature, 95% confidence intervals, and the area under the receiver operator characteristic curve (AUC) were obtained using logistic regression. A bootstrap approach was used to identify the strongest features in the training data set, and results for features that validated in the second half of the sample were reported using the full data set. The mean age at mammography was 64.0 ± 10.2 years, and the mean time from mammography to breast cancer was 3.7 ± 1.0 (range, 2.0-5.9 years). PD was associated with breast cancer risk (OR, 1.49; 95% confidence interval, 1.25-1.78). The strongest features that validated from each of several classes (Markovian, run length, Laws, wavelet, and Fourier) showed similar ORs as PD and predicted breast cancer at a similar magnitude (AUC = 0.58-0.60) as PD (AUC = 0.58). All of these features were automatically calculated (unlike PD) and measure texture at a coarse scale. These features were moderately correlated with PD (r = 0.39-0.76), and after adjustment for PD, each of the features attenuated only slightly and retained statistical significance. However, simultaneous inclusion of these features in a model with PD did not significantly improve the ability to predict breast cancer.

Original languageEnglish (US)
Pages (from-to)837-845
Number of pages9
JournalCancer Epidemiology Biomarkers and Prevention
Volume18
Issue number3
DOIs
StatePublished - Mar 2009

Fingerprint

Breast Neoplasms
Area Under Curve
Odds Ratio
Mammography
Confidence Intervals
Case-Control Studies
Body Mass Index
Logistic Models
Datasets

ASJC Scopus subject areas

  • Epidemiology
  • Oncology

Cite this

Texture features from mammographic images and risk of breast cancer. / Manduca, Armando; Carston, Michael J.; Heine, John J.; Scott, Christopher G.; Pankratz, V. Shane; Brandt, Kathleen R; Sellers, Thomas A.; Vachon, Celine M; Cerhan, James R.

In: Cancer Epidemiology Biomarkers and Prevention, Vol. 18, No. 3, 03.2009, p. 837-845.

Research output: Contribution to journalArticle

Manduca, Armando ; Carston, Michael J. ; Heine, John J. ; Scott, Christopher G. ; Pankratz, V. Shane ; Brandt, Kathleen R ; Sellers, Thomas A. ; Vachon, Celine M ; Cerhan, James R. / Texture features from mammographic images and risk of breast cancer. In: Cancer Epidemiology Biomarkers and Prevention. 2009 ; Vol. 18, No. 3. pp. 837-845.
@article{2a565bd06d164bd1a08451a00f01c999,
title = "Texture features from mammographic images and risk of breast cancer",
abstract = "Mammographic percent density (PD) is a strong risk factor for breast cancer, but there has been relatively little systematic evaluation of other features in mammographic images that might additionally predict breast cancer risk. We evaluated the association of a large number of image texture features with risk of breast cancer using a clinic-based case-control study of digitized film mammograms, all with screening mammograms before breast cancer diagnosis. The sample was split into training (123 cases and 258 controls) and validation (123 cases and 264 controls) data sets. Age-adjusted and body mass index (BMI)-adjusted odds ratios (OR) per SD change in the feature, 95{\%} confidence intervals, and the area under the receiver operator characteristic curve (AUC) were obtained using logistic regression. A bootstrap approach was used to identify the strongest features in the training data set, and results for features that validated in the second half of the sample were reported using the full data set. The mean age at mammography was 64.0 ± 10.2 years, and the mean time from mammography to breast cancer was 3.7 ± 1.0 (range, 2.0-5.9 years). PD was associated with breast cancer risk (OR, 1.49; 95{\%} confidence interval, 1.25-1.78). The strongest features that validated from each of several classes (Markovian, run length, Laws, wavelet, and Fourier) showed similar ORs as PD and predicted breast cancer at a similar magnitude (AUC = 0.58-0.60) as PD (AUC = 0.58). All of these features were automatically calculated (unlike PD) and measure texture at a coarse scale. These features were moderately correlated with PD (r = 0.39-0.76), and after adjustment for PD, each of the features attenuated only slightly and retained statistical significance. However, simultaneous inclusion of these features in a model with PD did not significantly improve the ability to predict breast cancer.",
author = "Armando Manduca and Carston, {Michael J.} and Heine, {John J.} and Scott, {Christopher G.} and Pankratz, {V. Shane} and Brandt, {Kathleen R} and Sellers, {Thomas A.} and Vachon, {Celine M} and Cerhan, {James R}",
year = "2009",
month = "3",
doi = "10.1158/1055-9965.EPI-08-0631",
language = "English (US)",
volume = "18",
pages = "837--845",
journal = "Cancer Epidemiology Biomarkers and Prevention",
issn = "1055-9965",
publisher = "American Association for Cancer Research Inc.",
number = "3",

}

TY - JOUR

T1 - Texture features from mammographic images and risk of breast cancer

AU - Manduca, Armando

AU - Carston, Michael J.

AU - Heine, John J.

AU - Scott, Christopher G.

AU - Pankratz, V. Shane

AU - Brandt, Kathleen R

AU - Sellers, Thomas A.

AU - Vachon, Celine M

AU - Cerhan, James R

PY - 2009/3

Y1 - 2009/3

N2 - Mammographic percent density (PD) is a strong risk factor for breast cancer, but there has been relatively little systematic evaluation of other features in mammographic images that might additionally predict breast cancer risk. We evaluated the association of a large number of image texture features with risk of breast cancer using a clinic-based case-control study of digitized film mammograms, all with screening mammograms before breast cancer diagnosis. The sample was split into training (123 cases and 258 controls) and validation (123 cases and 264 controls) data sets. Age-adjusted and body mass index (BMI)-adjusted odds ratios (OR) per SD change in the feature, 95% confidence intervals, and the area under the receiver operator characteristic curve (AUC) were obtained using logistic regression. A bootstrap approach was used to identify the strongest features in the training data set, and results for features that validated in the second half of the sample were reported using the full data set. The mean age at mammography was 64.0 ± 10.2 years, and the mean time from mammography to breast cancer was 3.7 ± 1.0 (range, 2.0-5.9 years). PD was associated with breast cancer risk (OR, 1.49; 95% confidence interval, 1.25-1.78). The strongest features that validated from each of several classes (Markovian, run length, Laws, wavelet, and Fourier) showed similar ORs as PD and predicted breast cancer at a similar magnitude (AUC = 0.58-0.60) as PD (AUC = 0.58). All of these features were automatically calculated (unlike PD) and measure texture at a coarse scale. These features were moderately correlated with PD (r = 0.39-0.76), and after adjustment for PD, each of the features attenuated only slightly and retained statistical significance. However, simultaneous inclusion of these features in a model with PD did not significantly improve the ability to predict breast cancer.

AB - Mammographic percent density (PD) is a strong risk factor for breast cancer, but there has been relatively little systematic evaluation of other features in mammographic images that might additionally predict breast cancer risk. We evaluated the association of a large number of image texture features with risk of breast cancer using a clinic-based case-control study of digitized film mammograms, all with screening mammograms before breast cancer diagnosis. The sample was split into training (123 cases and 258 controls) and validation (123 cases and 264 controls) data sets. Age-adjusted and body mass index (BMI)-adjusted odds ratios (OR) per SD change in the feature, 95% confidence intervals, and the area under the receiver operator characteristic curve (AUC) were obtained using logistic regression. A bootstrap approach was used to identify the strongest features in the training data set, and results for features that validated in the second half of the sample were reported using the full data set. The mean age at mammography was 64.0 ± 10.2 years, and the mean time from mammography to breast cancer was 3.7 ± 1.0 (range, 2.0-5.9 years). PD was associated with breast cancer risk (OR, 1.49; 95% confidence interval, 1.25-1.78). The strongest features that validated from each of several classes (Markovian, run length, Laws, wavelet, and Fourier) showed similar ORs as PD and predicted breast cancer at a similar magnitude (AUC = 0.58-0.60) as PD (AUC = 0.58). All of these features were automatically calculated (unlike PD) and measure texture at a coarse scale. These features were moderately correlated with PD (r = 0.39-0.76), and after adjustment for PD, each of the features attenuated only slightly and retained statistical significance. However, simultaneous inclusion of these features in a model with PD did not significantly improve the ability to predict breast cancer.

UR - http://www.scopus.com/inward/record.url?scp=64549104287&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=64549104287&partnerID=8YFLogxK

U2 - 10.1158/1055-9965.EPI-08-0631

DO - 10.1158/1055-9965.EPI-08-0631

M3 - Article

VL - 18

SP - 837

EP - 845

JO - Cancer Epidemiology Biomarkers and Prevention

JF - Cancer Epidemiology Biomarkers and Prevention

SN - 1055-9965

IS - 3

ER -