Non-Uniformly Sampled MR Correlated Spectroscopic Imaging in Breast Cancer and Nonlinear Reconstruction
Abstract
Differentiation of aggressive from indolent cancer will be a major task of this project. Dynamic contrast enhanced magnetic resonance imaging (DCE-MRI) offers higher sensitivity enabling improved diagnosis for breast cancer. However, varying specificity is a major concern. The MRI provides information regarding structural properties of the tissue of interest, whereas magnetic resonance spectroscopy (MRS) is a noninvasive technique that has great potential to provide tumor metabolism, which may be used in tumor diagnosis and evaluating the therapeutic response of the tumor. The combination of structural and metabolic information will provide medical professionals and clinical researchers with more information on which to base diagnosis and prognosis. The purpose of the proposed project is to speed up MRI acquisition so that the breast cancer patient suffers less during the MRI scan and improve the diagnostic accuracy using multi-voxel based 2D MRS in combination with diffusion weighted imaging (DWI), another powerful functional imaging technique. Three major goals are proposed here: (1) To further optimize the echo-planar correlated spectroscopic imaging (EP-COSI) and DWI, and to evaluate this multi-volume of interest based technique in patients with malignant and benign breast cancer and healthy breasts. (2) To develop and further optimize the EP-COSI reconstruction algorithms. (3) To correlate the changes in metabolite and lipid levels with apparent diffusion coefficient changes in breast cancer patients and healthy women. The study aims to include 50 malignant tumor patients, 20 patients with benign tumors, and 20 healthy volunteers. Using these goals, the proposed research will test three major hypotheses. First, the novel EP-COSI technique, with acceleration of more than 8 times, will enable two-dimensional (2D) spectra combined with MRS imaging (MRSI) to be acquired much faster than the currently used techniques based on selecting a single region of interest. The second hypothesis is that a novel group sparsity based reconstruction will facilitate higher acceleration of data acquisition. Third, breast tissue water movement measured by DWI can be correlated negatively with choline and positively with lipid levels quantified by the EP-COSI technique. False positive rate of mammography, the most popular primary screening and diagnostic tool for breast cancer detection for over three decades, is in the range of 60%-80%. Despite years of development, mammography, ultrasound, MRI, and even DCE-MRI are unable to reliably distinguish between malignant and benign tissues. Accelerated EP-COSI will reduce the total acquisition time that will be within clinically viable scan times and can acquire information not present in any other non-invasive modality, which will improve the cancer diagnosis. The metabolic information describes the metabolic structure of the chemicals including the metabolites and lipids present in the breast tissue, as well as their respective concentrations. This information can be used to assist in diagnosis, prognosis, and measurement of therapeutic efficacy. The proposed research will be the foundation for future metabolic imaging studies of the breast and will serve as a complementary method for tissue characterization of the breast compared with what is presently available. The additional information from 2D MRS compared to 1D MRS will enable better assignment of metabolites and lipids that will be important. The proposed research is capable of reducing the time required for MRS acquisition and improving diagnostic accuracy of breast cancer and thus, the outcome will facilitate shorter waiting time before treatment as well as reducing the number of false positives.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jan 31, 2017
- Source ID
- W81XWH1610524
Entities
People
- Michael Thomas
Organizations
- United States Army
- University of California, Los Angeles