Using computer-extracted image phenotypes from tumors on breast magnetic resonance imaging to predict breast cancer pathologic stage

Elizabeth S. Burnside, Karen Drukker, Hui Li, Ermelinda Bonaccio, Margarita Zuley, Marie Ganott, Jose M Net, Elizabeth J. Sutton, Kathleen R. Brandt, Gary J. Whitman, Suzanne D. Conzen, Li Lan, Yuan Ji, Yitan Zhu, Carl C. Jaffe, Erich P. Huang, John B. Freymann, Justin S. Kirby, Elizabeth A. Morris, Maryellen L. Giger

Research output: Contribution to journalArticle

23 Citations (Scopus)

Abstract

BACKGROUND The objective of this study was to demonstrate that computer-extracted image phenotypes (CEIPs) of biopsy-proven breast cancer on magnetic resonance imaging (MRI) can accurately predict pathologic stage. METHODS The authors used a data set of deidentified breast MRIs organized by the National Cancer Institute in The Cancer Imaging Archive. In total, 91 biopsy-proven breast cancers were analyzed from patients who had information available on pathologic stage (stage I, n = 22; stage II, n = 58; stage III, n = 11) and surgically verified lymph node status (negative lymph nodes, n = 46; ≥ 1 positive lymph node, n = 44; no lymph nodes examined, n = 1). Tumors were characterized according to 1) radiologist-measured size and 2) CEIP. Then, models were built that combined 2 CEIPs to predict tumor pathologic stage and lymph node involvement, and the models were evaluated in a leave-1-out, cross-validation analysis with the area under the receiver operating characteristic curve (AUC) as the value of interest. RESULTS Tumor size was the most powerful predictor of pathologic stage, but CEIPs that captured biologic behavior also emerged as predictive (eg, stage I and II vs stage III demonstrated an AUC of 0.83). No size measure was successful in the prediction of positive lymph nodes, but adding a CEIP that described tumor "homogeneity" significantly improved discrimination (AUC = 0.62; P =.003) compared with chance. CONCLUSIONS The current results indicate that MRI phenotypes have promise for predicting breast cancer pathologic stage and lymph node status.

Original languageEnglish (US)
Pages (from-to)748-757
Number of pages10
JournalCancer
Volume122
Issue number5
DOIs
StatePublished - Mar 1 2016

Fingerprint

Lymph Nodes
Magnetic Resonance Imaging
Breast Neoplasms
Phenotype
Area Under Curve
Neoplasms
Biopsy
National Cancer Institute (U.S.)
ROC Curve
Breast

Keywords

  • breast cancer stage
  • magnetic resonance imaging (MRI)
  • prognosis
  • quantitative image analysis

ASJC Scopus subject areas

  • Cancer Research
  • Oncology

Cite this

Burnside, E. S., Drukker, K., Li, H., Bonaccio, E., Zuley, M., Ganott, M., ... Giger, M. L. (2016). Using computer-extracted image phenotypes from tumors on breast magnetic resonance imaging to predict breast cancer pathologic stage. Cancer, 122(5), 748-757. https://doi.org/10.1002/cncr.29791

Using computer-extracted image phenotypes from tumors on breast magnetic resonance imaging to predict breast cancer pathologic stage. / Burnside, Elizabeth S.; Drukker, Karen; Li, Hui; Bonaccio, Ermelinda; Zuley, Margarita; Ganott, Marie; Net, Jose M; Sutton, Elizabeth J.; Brandt, Kathleen R.; Whitman, Gary J.; Conzen, Suzanne D.; Lan, Li; Ji, Yuan; Zhu, Yitan; Jaffe, Carl C.; Huang, Erich P.; Freymann, John B.; Kirby, Justin S.; Morris, Elizabeth A.; Giger, Maryellen L.

In: Cancer, Vol. 122, No. 5, 01.03.2016, p. 748-757.

Research output: Contribution to journalArticle

Burnside, ES, Drukker, K, Li, H, Bonaccio, E, Zuley, M, Ganott, M, Net, JM, Sutton, EJ, Brandt, KR, Whitman, GJ, Conzen, SD, Lan, L, Ji, Y, Zhu, Y, Jaffe, CC, Huang, EP, Freymann, JB, Kirby, JS, Morris, EA & Giger, ML 2016, 'Using computer-extracted image phenotypes from tumors on breast magnetic resonance imaging to predict breast cancer pathologic stage', Cancer, vol. 122, no. 5, pp. 748-757. https://doi.org/10.1002/cncr.29791
Burnside, Elizabeth S. ; Drukker, Karen ; Li, Hui ; Bonaccio, Ermelinda ; Zuley, Margarita ; Ganott, Marie ; Net, Jose M ; Sutton, Elizabeth J. ; Brandt, Kathleen R. ; Whitman, Gary J. ; Conzen, Suzanne D. ; Lan, Li ; Ji, Yuan ; Zhu, Yitan ; Jaffe, Carl C. ; Huang, Erich P. ; Freymann, John B. ; Kirby, Justin S. ; Morris, Elizabeth A. ; Giger, Maryellen L. / Using computer-extracted image phenotypes from tumors on breast magnetic resonance imaging to predict breast cancer pathologic stage. In: Cancer. 2016 ; Vol. 122, No. 5. pp. 748-757.
@article{7c37240c43264c79923b7b96823585cc,
title = "Using computer-extracted image phenotypes from tumors on breast magnetic resonance imaging to predict breast cancer pathologic stage",
abstract = "BACKGROUND The objective of this study was to demonstrate that computer-extracted image phenotypes (CEIPs) of biopsy-proven breast cancer on magnetic resonance imaging (MRI) can accurately predict pathologic stage. METHODS The authors used a data set of deidentified breast MRIs organized by the National Cancer Institute in The Cancer Imaging Archive. In total, 91 biopsy-proven breast cancers were analyzed from patients who had information available on pathologic stage (stage I, n = 22; stage II, n = 58; stage III, n = 11) and surgically verified lymph node status (negative lymph nodes, n = 46; ≥ 1 positive lymph node, n = 44; no lymph nodes examined, n = 1). Tumors were characterized according to 1) radiologist-measured size and 2) CEIP. Then, models were built that combined 2 CEIPs to predict tumor pathologic stage and lymph node involvement, and the models were evaluated in a leave-1-out, cross-validation analysis with the area under the receiver operating characteristic curve (AUC) as the value of interest. RESULTS Tumor size was the most powerful predictor of pathologic stage, but CEIPs that captured biologic behavior also emerged as predictive (eg, stage I and II vs stage III demonstrated an AUC of 0.83). No size measure was successful in the prediction of positive lymph nodes, but adding a CEIP that described tumor {"}homogeneity{"} significantly improved discrimination (AUC = 0.62; P =.003) compared with chance. CONCLUSIONS The current results indicate that MRI phenotypes have promise for predicting breast cancer pathologic stage and lymph node status.",
keywords = "breast cancer stage, magnetic resonance imaging (MRI), prognosis, quantitative image analysis",
author = "Burnside, {Elizabeth S.} and Karen Drukker and Hui Li and Ermelinda Bonaccio and Margarita Zuley and Marie Ganott and Net, {Jose M} and Sutton, {Elizabeth J.} and Brandt, {Kathleen R.} and Whitman, {Gary J.} and Conzen, {Suzanne D.} and Li Lan and Yuan Ji and Yitan Zhu and Jaffe, {Carl C.} and Huang, {Erich P.} and Freymann, {John B.} and Kirby, {Justin S.} and Morris, {Elizabeth A.} and Giger, {Maryellen L.}",
year = "2016",
month = "3",
day = "1",
doi = "10.1002/cncr.29791",
language = "English (US)",
volume = "122",
pages = "748--757",
journal = "Cancer",
issn = "0008-543X",
publisher = "John Wiley and Sons Inc.",
number = "5",

}

TY - JOUR

T1 - Using computer-extracted image phenotypes from tumors on breast magnetic resonance imaging to predict breast cancer pathologic stage

AU - Burnside, Elizabeth S.

AU - Drukker, Karen

AU - Li, Hui

AU - Bonaccio, Ermelinda

AU - Zuley, Margarita

AU - Ganott, Marie

AU - Net, Jose M

AU - Sutton, Elizabeth J.

AU - Brandt, Kathleen R.

AU - Whitman, Gary J.

AU - Conzen, Suzanne D.

AU - Lan, Li

AU - Ji, Yuan

AU - Zhu, Yitan

AU - Jaffe, Carl C.

AU - Huang, Erich P.

AU - Freymann, John B.

AU - Kirby, Justin S.

AU - Morris, Elizabeth A.

AU - Giger, Maryellen L.

PY - 2016/3/1

Y1 - 2016/3/1

N2 - BACKGROUND The objective of this study was to demonstrate that computer-extracted image phenotypes (CEIPs) of biopsy-proven breast cancer on magnetic resonance imaging (MRI) can accurately predict pathologic stage. METHODS The authors used a data set of deidentified breast MRIs organized by the National Cancer Institute in The Cancer Imaging Archive. In total, 91 biopsy-proven breast cancers were analyzed from patients who had information available on pathologic stage (stage I, n = 22; stage II, n = 58; stage III, n = 11) and surgically verified lymph node status (negative lymph nodes, n = 46; ≥ 1 positive lymph node, n = 44; no lymph nodes examined, n = 1). Tumors were characterized according to 1) radiologist-measured size and 2) CEIP. Then, models were built that combined 2 CEIPs to predict tumor pathologic stage and lymph node involvement, and the models were evaluated in a leave-1-out, cross-validation analysis with the area under the receiver operating characteristic curve (AUC) as the value of interest. RESULTS Tumor size was the most powerful predictor of pathologic stage, but CEIPs that captured biologic behavior also emerged as predictive (eg, stage I and II vs stage III demonstrated an AUC of 0.83). No size measure was successful in the prediction of positive lymph nodes, but adding a CEIP that described tumor "homogeneity" significantly improved discrimination (AUC = 0.62; P =.003) compared with chance. CONCLUSIONS The current results indicate that MRI phenotypes have promise for predicting breast cancer pathologic stage and lymph node status.

AB - BACKGROUND The objective of this study was to demonstrate that computer-extracted image phenotypes (CEIPs) of biopsy-proven breast cancer on magnetic resonance imaging (MRI) can accurately predict pathologic stage. METHODS The authors used a data set of deidentified breast MRIs organized by the National Cancer Institute in The Cancer Imaging Archive. In total, 91 biopsy-proven breast cancers were analyzed from patients who had information available on pathologic stage (stage I, n = 22; stage II, n = 58; stage III, n = 11) and surgically verified lymph node status (negative lymph nodes, n = 46; ≥ 1 positive lymph node, n = 44; no lymph nodes examined, n = 1). Tumors were characterized according to 1) radiologist-measured size and 2) CEIP. Then, models were built that combined 2 CEIPs to predict tumor pathologic stage and lymph node involvement, and the models were evaluated in a leave-1-out, cross-validation analysis with the area under the receiver operating characteristic curve (AUC) as the value of interest. RESULTS Tumor size was the most powerful predictor of pathologic stage, but CEIPs that captured biologic behavior also emerged as predictive (eg, stage I and II vs stage III demonstrated an AUC of 0.83). No size measure was successful in the prediction of positive lymph nodes, but adding a CEIP that described tumor "homogeneity" significantly improved discrimination (AUC = 0.62; P =.003) compared with chance. CONCLUSIONS The current results indicate that MRI phenotypes have promise for predicting breast cancer pathologic stage and lymph node status.

KW - breast cancer stage

KW - magnetic resonance imaging (MRI)

KW - prognosis

KW - quantitative image analysis

UR - http://www.scopus.com/inward/record.url?scp=84958879262&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84958879262&partnerID=8YFLogxK

U2 - 10.1002/cncr.29791

DO - 10.1002/cncr.29791

M3 - Article

VL - 122

SP - 748

EP - 757

JO - Cancer

JF - Cancer

SN - 0008-543X

IS - 5

ER -