Getting a jump on cancer with a genomic risk classifier

Institution: Stanford University
Investigator(s): Robert West, M.D., Ph.D. -
Award Cycle: 2013 (Cycle 19) Grant #: 19IB-0124 Award: $251,119
Award Type: IDEA
Research Priorities
Etiology and Prevention>Etiology: the role of environment and lifestyle



Initial Award Abstract (2013)

This project was supported in part by a generous donation from the Katie Ann Buzbee Trust.

Non-technical overview of the research topic and relevance to breast cancer:
Prevention is the ideal strategy for breast cancer treatment as it reduces the patient suffering both from the cancer and from the conventional treatments. Recent large studies that measure alterations in cancer DNA have identified and confirmed numerous recurrent changes to genes that can be found in the vast majority of breast cancers. In contrast to the cancers, little is understood about these genetic changes associated with the early stages of progression to breast cancer: changes from normal tissue to early neoplasias to carcinoma in situ to cancer. The evaluation of these “pre-invasive neoplasias” found in screening breast biopsies relies solely on how these lesions look under light microscope, based on concepts developed decades ago. These pre-invasive neoplasias (such as atypical ductal hyperplasia and columnar cell lesions) represent a dilemma in clinical management as, for example, only about 20% of atypical ductal hyperplasia lesions progress to cancer (though 80% do not). These lesions are becoming increasingly detected with more and more sensitive radiographic screening.

The question(s) or central hypotheses of the research:
Because breast cancer is composed of recurrent, common alterations in the DNA, we hypothesize that the identification of these changes in pre-invasive neoplasia will help us identify pre-invasive neoplasms that have a high likelihood of progressing to cancer.

The general methodology:
We propose a case-control study design that will enable us to measure the frequency of common DNA alterations in pre-invasive neoplasia that is adjacent to cancer (cases) compared to the frequency of these mutations in pre-invasive neoplasia with no adjacent cancer (controls). Our data will be derived from cancers from both women treated at an academic hospital as well as women in the multi-ethnic “Equality in Breast Cancer Care” (EBCC) study. We will use targeted sequencing and fluorescent in situ hybridization (FISH) to observe these DNA differences. We will then build a classifier using this data that is most effective in stratifying pre-invasive lesions for their likelihood of being associated with cancer. Our methods can be readily translated to the clinical laboratory for validation in a larger dataset and subsequently implemented through policy guidelines to the community level.

Innovative elements of the project:
The innovative elements of the project are the use of recurrent, common alterations in the DNA of breast cancers to predict likelihood of progressing to cancer. The concept of applying genomic data to breast cancer prevention is also highly innovative. Improved knowledge of which proliferative lesions will develop into cancer will help reduce cancer incidence and mortality by indicating when and where aggressive clinical treatment, like complete surgical excision, chemoprophylaxis, and/or close follow-up with the possible addition of intensive surveillance techniques such as breast magnetic resonance imaging (MRI), will benefit the patient and prevent cancer development.




Final Report (2016)
Our project uses the untapped potential of the recent breast cancer genomic data to better refine our current risk classification system of pre-invasive breast neoplasia. Ultimately, our goal is to provide more relevant and predictive clinical results to women who are diagnosed with pre-invasive breast neoplasia on screening biopsies, that will more appropriately guide therapy and future screening with the ultimate goal of reducing incidence of and mortality from invasive breast carcinoma, while at the same time reducing unnecessary surgical interventions. Thus, our primary research question is whether we can develop a test to stratify pre-invasive breast neoplasia for risk of development to invasive breast carcinoma, using established and emerging clinical platforms to measure the genomic changes that we and others have identified.

In the months spanning this grant, we have pursued two specific goals. We have generated DNA copy number data and mutation data for a cohort of cases of ductal carcinoma in situ from an academic center. The FISH data generated can predict the presence of invasive ductal carcinoma better than conventional biomarkers in this cohort. We are finishing a tissue microarray of pre-invasive neoplasia samples from underserved patients in the Bay Area for which we will also generate DNA copy number data and mutation data.

We were able to overcome a barrier of high through put for FISH measurements on a large series of cases using automated image recognition software.

We have generated a multi DNA amplicon classifier for distinguishing ductal carcinoma in situ cases that are associated with invasive ductal carcinoma or not. Through targeted sequencing, we have identified the mutations in genes commonly mutated in invasive ductal carcinoma, and are comparing these data to the DNA copy number data. We have generated a tissue microarray of pre-invasive neoplasia samples from underserved patients in the Bay Area and are generating DNA copy number data.

We have built a tissue microarray of hyperplasias and will generate DNA copy number data and mutation data for these samples in part guided by our experience with the ductal carcinoma in situ cases.

Publications:
Chromosomal copy number alterations for associations of ductal carcinoma in situ with invasive breast cancer
Navigating choices when applying multiple imputation in the presence of multi-level categorical interaction effects

Publications:
Synergistic drug combinations from electronic health records and gene expression



Navigating choices when applying multiple imputation in the presence of multi-level categorical interaction effects
Periodical:Statistical Methodology
Index Medicus: Stat Methodol
Authors: Mitani, AA, Kurian AW, Das AK, Desai M,
Yr: 2015 Vol: 27 Nbr: Abs: Pg:82-99

Chromosomal copy number alterations for associations of ductal carcinoma in situ with invasive breast cancer DOI: 10.1186/s13058-015-0623-y
Periodical:Breast Cancer Research
Index Medicus: Breast Cancer Res
Authors: Afghahi A, Forgo E, Mitani A, Desai M, Varma S, Seto T, Rigdon J, et al.
Yr: 2015 Vol: 17 Nbr: 108 Abs: Pg: