Imaging, Genomics, and Glycoproteomics for Cancer Detection

Institution: Stanford University
Investigator(s): Sharon Pitteri, Ph.D. -
Award Cycle: 2013 (Cycle 19) Grant #: 19IB-0140 Award: $235,348
Award Type: IDEA
Research Priorities
Detection, Prognosis and Treatment>Imaging, Biomarkers, and Molecular Pathology: improving detection and diagnosis

Initial Award Abstract (2013)

Non-technical overview of the research topic and relevance to breast cancer:
Mammography is currently the best available screening tool for breast cancer early detection. However, limitations of mammography include relatively high false positive (abnormal mammogram with no cancer) and false negative (cancer missed by mammogram) rates. For example, mammography’s ability to detect breast cancer is limited in women with dense breast tissue. This study seeks to improve early detection of breast cancer by combining information traditional breast imaging information (mammography, ultrasound, and MRI), breast density measurements, genomic information about the tumor, and protein alterations in the blood.

The question(s) or central hypotheses of the research:
We hypothesize that there is a correlation among primary breast cancer imaging characteristics, (tumor) tissue gene expression, and corresponding serum (blood) protein levels, and that some proteins from compartments of the cells, molecules, and blood vessels surrounding the tumor are present in the blood at the earliest stages of breast cancer development. If successfully answered, the identified molecules have the potential to contribute to decreased mortality by creating a blood test for breast cancer early detection to accompany other screening modalities.

The general methodology:
For these studies, we will establish a Stanford tissue bank of matched needle core biopsy samples, information from imaging studies (mammography, MRI, and ultrasound), breast density measurements, and blood samples. We will generate profiles of genes expressed in tumor tissue and compare these profiles to genes expressed in normal tissue. We will measure levels of and characterize proteins in the blood from women with and without breast cancer. The imaging data, tumor gene expression data, breast density measurements, and blood protein data will be integrated to identify correlations between the imaging, tumor, and blood data.

Innovative elements of the project:
This dataset will represent the most comprehensive analysis of breast cancer imaging, breast density measurements, tumor gene expression, and matched protein plasma analysis to date, and will lead to a blood test that will complement existing radiographic screening tests. The molecular information from genomic and proteomic analysis will be combined with imaging information to ultimately lead to improved breast cancer detection and prognosis through the use of a blood test in conjunction with imaging.

Final Report (2015)

Screening mammography is the most widely available tool for breast cancer detection. However, mammography suffers from both false negative and false positive results. False negatives can occur when breast tissue is dense, causing the tumor to be missed. False positives can occur when suspicious benign lesions are identified. The primary objective of this study is to understand the relationship between clinical imaging characteristics and molecular features of breast cancer (both at the genomic and proteomic level in the tumor and blood respectively), to ultimately improve breast cancer screening. This project tests the hypothesis that there is a correlation among breast imaging characteristics, tissue gene expression, and corresponding plasma glycoprotein levels, and that some proteins from the tumor microenvironment are present in the blood before proteins from cancer cells during early breast cancer development.

To this end, we have 1) recruited patients and obtained imaging data, breast density measurements, breast biopsy tissues, and blood samples; 2) generated gene expressing profiles of breast cancer and normal adjacent tissue samples by RNA-sequencing; and 3) performed quantitative glycoproteomic profiling of matched patient plasma samples. We are in the process of integrating this data to understand the correlation between the imaging, genomics, proteomics, and clinical patient information. We have successfully completed our aims of obtaining patient samples and generating molecular profiles. The data analysis is ongoing. We did not encounter any substantial barriers in the project. We plan to publish the results of this work following further verification and validation studies within one year. We plan to apply for federal funding to support continuation of the project, with the specific direction to be determined by our findings once the data is integrated and the validation studies are performed.