Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2016 Mar 1;122(5):748-57.
doi: 10.1002/cncr.29791. Epub 2015 Nov 30.

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

Elizabeth S Burnside  1 , Karen Drukker  2 , Hui Li  2 , Ermelinda Bonaccio  3 , Margarita Zuley  4 , Marie Ganott  4 , Jose M Net  5 , Elizabeth J Sutton  6 , Kathleen R Brandt  7 , Gary J Whitman  8 , Suzanne D Conzen  2 , Li Lan  2 , Yuan Ji  9   10 , Yitan Zhu  10 , Carl C Jaffe  11 , Erich P Huang  11 , John B Freymann  11 , Justin S Kirby  11 , Elizabeth A Morris  6 , Maryellen L Giger  2
Affiliations

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

Elizabeth S Burnside et al. Cancer. .

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. Cancer 2016;122:748-757. © 2015 American Cancer Society.

Keywords: breast cancer stage; magnetic resonance imaging (MRI); prognosis; quantitative image analysis.

PubMed Disclaimer

Figures

Figure 1
Figure 1
Representative breast MRI case depicting a sagittal fat saturated T1-weighted image with an annotation identifying the tumor and measurement (“radiologist-size”).
Figure 2
Figure 2. Schematic of the MRI-based computer-extracted image phenotypes
Figure 3
Figure 3
The “radiologist size” distribution (the mean value for the maximal tumor diameter assigned by the 3 expert radiologists) for a) pathologic stage, and b) pathologic lymph node status.
Figure 3
Figure 3
The “radiologist size” distribution (the mean value for the maximal tumor diameter assigned by the 3 expert radiologists) for a) pathologic stage, and b) pathologic lymph node status.
Figure 4
Figure 4
Distribution of CEIP values for pathologic stage: a) surface area, b) enhancement texture homogeneity (the inverse difference moment of the gray-level co-occurrence matrix calculated within the tumor at the first post-contrast image), and distribution of CEIP for “radiologist size” stratified by pathologic lymph node (LN) status c) irregularity, and d) enhancement texture homogeneity
Figure 4
Figure 4
Distribution of CEIP values for pathologic stage: a) surface area, b) enhancement texture homogeneity (the inverse difference moment of the gray-level co-occurrence matrix calculated within the tumor at the first post-contrast image), and distribution of CEIP for “radiologist size” stratified by pathologic lymph node (LN) status c) irregularity, and d) enhancement texture homogeneity
Figure 4
Figure 4
Distribution of CEIP values for pathologic stage: a) surface area, b) enhancement texture homogeneity (the inverse difference moment of the gray-level co-occurrence matrix calculated within the tumor at the first post-contrast image), and distribution of CEIP for “radiologist size” stratified by pathologic lymph node (LN) status c) irregularity, and d) enhancement texture homogeneity
Figure 4
Figure 4
Distribution of CEIP values for pathologic stage: a) surface area, b) enhancement texture homogeneity (the inverse difference moment of the gray-level co-occurrence matrix calculated within the tumor at the first post-contrast image), and distribution of CEIP for “radiologist size” stratified by pathologic lymph node (LN) status c) irregularity, and d) enhancement texture homogeneity
Figure 5
Figure 5
Example dynamic contrast-enhanced images of primary tumor (both pathologic stage II) at the first post-contrast time point: a) demonstrates a tumor proven to be lymph node negative with “low” enhancement texture homogeneity (imaged at 1 minute, 29 seconds post-contrast), and b) demonstrates a tumor proven to be lymph node positive with “high” enhancement texture homogeneity (imaged at 1 minute, 16 seconds post-contrast).
Figure 5
Figure 5
Example dynamic contrast-enhanced images of primary tumor (both pathologic stage II) at the first post-contrast time point: a) demonstrates a tumor proven to be lymph node negative with “low” enhancement texture homogeneity (imaged at 1 minute, 29 seconds post-contrast), and b) demonstrates a tumor proven to be lymph node positive with “high” enhancement texture homogeneity (imaged at 1 minute, 16 seconds post-contrast).
See this image and copyright information in PMC

References

    1. Lee SC, Jain PA, Jethwa SC, Tripathy D, Yamashita MW. Radiologist's role in breast cancer staging: providing key information for clinicians. Radiographics : a review publication of the Radiological Society of North America, Inc. 2014 Mar-Apr;34(2):330–342. - PubMed
    1. Bagaria SP, Ray PS, Sim MS, et al. Personalizing breast cancer staging by the inclusion of ER, PR, and HER2. JAMA surgery. 2014 Feb;149(2):125–129. - PubMed
    1. Yi M, Mittendorf EA, Cormier JN, et al. Novel staging system for predicting disease-specific survival in patients with breast cancer treated with surgery as the first intervention: time to modify the current American Joint Committee on Cancer staging system. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2011 Dec 10;29(35):4654–4661. - PMC - PubMed
    1. Trop I, LeBlanc SM, David J, et al. Molecular classification of infiltrating breast cancer: toward personalized therapy. Radiographics : a review publication of the Radiological Society of North America, Inc. 2014 Sep-Oct;34(5):1178–1195. - PubMed
    1. [Accessed September 13, 2014];Breast Cancer Version 3.2014. 2014 http://www.nccn.org/professionals/physician_gls/pdf/breast.pdf.

Publication types