The goal of this research is to develop quantitative image-based surrogate markers of breast cancer tumors for use in predicting response to therapy and ultimately aiding in patient management. There is a large variation in the clinical presentation of breast cancer in women, and it has been shown that in many instances, biological characteristics, i.e., features, of the primary tumor correlate with outcome. Methods to assess such biological features for the prediction of outcome, however, may be invasive, expensive or not widely available. Our hypothesis is that MRI-based features obtained through quantitative image analysis will prove useful as non-invasive biomarkers for the assessment of, and prediction of, the response of breast cancer to neoadjuvant therapy. We propose to validate such image-based biomarkers using magnetic resonance (MR) images of breast tumors from the ACRIN 6657 clinical trial, which includes pathological response data. Specifically, (1) We will investigate the relationship of breast cancer therapy outcome and MR image-based tumor characteristics (features), and changes in these features over time, using a University of Chicago database and the ACRIN 6657 I-SPY clinical trial dataset of breast cancer tumors from patients who have undergone neoadjuvant treatment, (2) We will develop and evaluate the MRI-derived `signatures' of breast cancer tumors for the prediction of, and assessment of, response to therapy using the ACRIN 6657 dataset, and (3) We will conduct preliminary, initial stratification and association of the MRI features with cancer subtype and other clinical/histopathological data from the ACRIN dataset. We will build on our 25-year history of taking innovation to the clinical setting by extending our prior development, validation, and translation of quantitative image analysis methods for computer- aided diagnosis to the post-diagnosis, predictive component in order to assess response to neoadjuvant therapy. Our research addresses the development and validation of algorithms using the existing ACRIN 6657 dataset with the goal of improving the ability to measure the response of targeted tumors to therapy quantitatively. Our proposed research is aligned with the QIN U01 PAR-11-150 goals of including robustness investigations and multi-site trial data (UChicago and ACRIN). Through this QIN grant, our participation in the QIN community will yield deliverables including an open-platform system that will provide tools for linking segmentation/feature extraction/classification, for comparing performance metrics across acquisition and/or analysis systems, and for discovery through dimension reduction techniques. Our research will yield a set of validated lesion signatures that will serve as quantitative tools for use in clinical studies/trials to predict and/or assess tumor response. Given that other studies/trials may use different treatments, we will make available to the QIN community our tools for training, testing, and presenting the the quantitative signatures so that predictive signatures for a range of treatments can be determined.

Public Health Relevance

Accurate characterization of the breast cancer tumor is critical to assessing response to therapy and thus, successful patient management. This proposal represents a major effort to validate our image-based tumor features from MRI as image-based predictive biomarkers and in assessing response to therapy within the datasets of the ACRIN I-SPY-1 trial (ACRIN 6657).

Agency
National Institute of Health (NIH)
Institute
National Cancer Institute (NCI)
Type
Research Project--Cooperative Agreements (U01)
Project #
5U01CA195564-03
Application #
9249507
Study Section
Special Emphasis Panel (ZCA1)
Program Officer
Tata, Darayash B
Project Start
2015-04-15
Project End
2020-03-31
Budget Start
2017-04-01
Budget End
2018-03-31
Support Year
3
Fiscal Year
2017
Total Cost
Indirect Cost
Name
University of Chicago
Department
Radiation-Diagnostic/Oncology
Type
Schools of Medicine
DUNS #
005421136
City
Chicago
State
IL
Country
United States
Zip Code
60637
Mendel, Kayla; Li, Hui; Sheth, Deepa et al. (2018) Transfer Learning From Convolutional Neural Networks for Computer-Aided Diagnosis: A Comparison of Digital Breast Tomosynthesis and Full-Field Digital Mammography. Acad Radiol :
Mendel, Kayla R; Li, Hui; Lan, Li et al. (2018) Quantitative texture analysis: robustness of radiomics across two digital mammography manufacturers' systems. J Med Imaging (Bellingham) 5:011002
Li, Hui; Giger, Maryellen L; Huynh, Benjamin Q et al. (2017) Deep learning in breast cancer risk assessment: evaluation of convolutional neural networks on a clinical dataset of full-field digital mammograms. J Med Imaging (Bellingham) 4:041304
Antropova, Natalia; Huynh, Benjamin Q; Giger, Maryellen L (2017) A deep feature fusion methodology for breast cancer diagnosis demonstrated on three imaging modality datasets. Med Phys 44:5162-5171
Huynh, Benjamin Q; Li, Hui; Giger, Maryellen L (2016) Digital mammographic tumor classification using transfer learning from deep convolutional neural networks. J Med Imaging (Bellingham) 3:034501
Li, Hui; Zhu, Yitan; Burnside, Elizabeth S et al. (2016) MR Imaging Radiomics Signatures for Predicting the Risk of Breast Cancer Recurrence as Given by Research Versions of MammaPrint, Oncotype DX, and PAM50 Gene Assays. Radiology 281:382-391
Li, Hui; Weiss, William A; Medved, Milica et al. (2016) Breast density estimation from high spectral and spatial resolution MRI. J Med Imaging (Bellingham) 3:044507
Li, Hui; Zhu, Yitan; Burnside, Elizabeth S et al. (2016) Quantitative MRI radiomics in the prediction of molecular classifications of breast cancer subtypes in the TCGA/TCIA data set. NPJ Breast Cancer 2:
Burnside, Elizabeth S; Drukker, Karen; Li, Hui et al. (2016) Using computer-extracted image phenotypes from tumors on breast magnetic resonance imaging to predict breast cancer pathologic stage. Cancer 122:748-57
Zhu, Yitan; Li, Hui; Guo, Wentian et al. (2015) Deciphering Genomic Underpinnings of Quantitative MRI-based Radiomic Phenotypes of Invasive Breast Carcinoma. Sci Rep 5:17787

Showing the most recent 10 out of 12 publications