Breast cancer is one of the most prevalent tumors for women in Taiwan. Higher incident rate of breast cancer in young age of 35 years old and the population of molecular subtypes in Taiwan differ from western countries. In 2013, the 13th St. Gallen International Breast Cancer Conference expert panel updated the consensus of diagnosis and treatment strategies for early breast cancer. The consensus described the definition of genetic surrogates for categorizing molecular subtypes of breast cancer. The prognosis and adjuvant therapies were further recommended for each breast cancer subtype. However, the high costs of the multi-gene assay prohibit its clinical applications. Developing a non-invasive imaging-based approach can reduce screening costs and benefit the patients with breast cancer in the aspects of diagnosis, treatment decision, and longitudinal monitoring. An emerging cancer research field, termed as Radiogenomics, focuses on discovering the relationships between imaging phenotypes and the underlying gene expression in tumors. Several studies in brain tumor (glioblastoma) and lung cancer have revealed that the regulations of gene can be characterized by non-invasive MRI or CT imaging techniques. We hypothesize that the multi-modality and multi-feature MRI (Radiomics) can be utilized to estimate gene expressions associated with tumor proliferation, invasion, HER2, and estrogen in breast cancer, and accordingly achieve the ultimate goal of “imaging gene” in clinical. In this two-year project, the study aims encompass, Aim 1: Perform multi-feature radiomic analysis using the TCGA/TCIA database to classify the molecular subtypes of breast cancer. Aim 2: Collect patients’ multi-modality MR data and tissue samples for RNA microarray to unravel the relationships between imaging phenotypes and genotypes (Radiogenomics) in Taiwanese population. Aim 3: Establish image-based recurrence score and compare it with 21-gene recurrence score in predicting the risk of tumor recurrence; Aim 4: Explore the tumor heterogeneity from both perspectives of imaging features and gene expressions. To fulfill our study aims, we will collect imaging and gene data of breast cancer from The Cancer Genome Atlas (TCGA)/The Cancer Imaging Archive (TCIA) database and patient population in Taipei Medical University Hospital (TMUH). During the first three months of the 1st year, the datasets downloaded from TCGA/TCIA will be used to construct the platform in calculating the parametric maps and 55 quantitative radiomic features (intensity-based, geometry-based, and textural analyses). Multi-feature analysis will then be applied to investigate the discrimination power of molecular subtypes based on imaging features. To investigate the Taiwanese characteristics of breast cancer, around 200 patients with diagnostic breast cancer will be recruited in TMUH during this two-year project. Data of MRI, pathology, and gene expression (RNA microarray for 60 of 200 patients due to limited budget) will be collected for each patient. Anticipated outcomes from this project include the establishment of an imaging-based platform in categorized molecular subtypes of breast cancer, reveal of relationships between imaging phenotypes and gene expression, publication of 1 to 2 SCI journal articles, and training of relevant personnel and students to foster the clinical applications of MR imaging and quantitative analyses in diagnosis and treatment of breast cancer.
|Effective start/end date||8/1/16 → 10/31/17|
- breast cancer
- molecular subtypes
- recurrence score