|Breast MRI: Optimizing & Predicting Outcomes|
NMR Shutter-Speed Discrimination of Malignant and
Benign Breast Tumors Using ROI Data
Xin Li1, Wei Huang2, Elizabeth A. Morris2, Luminita A. Tudorica3, William D. Rooney1, Ya Wang2, Jingang Xu1, Charles S. Springer, Jr. 1
1Oregon Health & Science University, Portland, USA; 2Memorial Sloan Kettering Cancer Center, USA; 3State University of New York at Stony Brook, USA
Dynamic-Contrast-Enhanced (DCE) MRI data were acquired from the lesions of 22 breast cancer patients, who had been screened positive by institutional protocols, but before biopsy and pathology analyses were performed. The data were subjected to pharmacokinetic analyses by both the Standard Model (SM) and the Shutter Speed Model (SSM). For ROIs circumscribing every one of the 15 lesions subsequently proven benign, the SM returned essentially the same value for Ktrans (a measure of CR extravasation rate) as the SSM. However, for every one of the 7 malignant tumors, the SM underestimated Ktrans. This allows complete separation of malignant from benign tumors.
Effect of the Spatial Resolution of A-Priori Data on
the Quality of Single-Point Dixon Fat-Water Images
Elizabeth Ramsay1, Donald Plewes1
1Sunnybrook Health Sciences Centre, Toronto, Canada
In this work, the effect of reducing the spatial resolution of the a priori phase information on Single-Point Dixon (SPD) images is examined, considering both noise performance and the quality of the fat-water separation. Decreasing the spatial resolution of the a priori images improves the noise properties of fat-water images, as well as reducing acquisition time. Human and phantom studies show that is possible to reduce the size of the acquisition matrix for the phase mapping scans to about 64x64 (voxel size 3mm) without compromising the quality of fat-water separation.
BOLD Contrast in the Breast at 3T
Rebecca Rakow-Penner1, Laura Pisani1, Bruce Daniel1, Stephanie Oberfoell1, Christine Law1, Gary H. Glover1
1Stanford University, Stanford, California , USA
BOLD contrast imaging applied to breast tumors may provide useful clinical information on tumor oxygenation. Before studying tumors, we are developing a robust method for measuring BOLD contrast in healthy breast tissue at 3T. The method includes heart saturation to diminish cardiac motion artifact, along with corrections for respiratory motion and cardiac pulsation. A rapid multi-interleave spiral sequence is used to acquire the images. Results indicate that in healthy breast tissue that there is an inverse correlation between BOLD contrast and oxygenation.
Multi-Frequency Off-Resonance Correction and
Water/Fat Separation for Spiral Breast Imaging
Kristin L. Granlund1, 2, Brian A. Hargreaves1
1Stanford University, Stanford, California , USA
Off-resonance is a problem in spiral imaging because it causes image blurring and water/fat separation to fail. The problem may be severe when imaging over large volumes, such as in bilateral breast MRI. This project simultaneously corrects for off-resonance and performs water/fat separation using a multi-frequency reconstruction for 3D stack-of-spiral breast images with multiple echo times.
3-D Tracking of the Mammary Ductal Tree Using
Diffusion Tensor MR Imaging
Erez Eyal1, Edna Furman-Haran1, Hadassa Degani1
1Weizmann Institute of Science, Rehovot, Israel
Mammary malignancies typically develop from the ductal epithelial cells, and spread within the ducts. Consequently, the ductal structures are an important area of investigation of both normal breast development and malignant breast transformation. Our aim is to develop an MRI method for tracking the spatial anisotropic diffusion of the water in the breast ducts and thereby map the ductal tree. The preliminary results show that diffusion in the mammary tissue has an anisotropic nature and that the orientation of the primary diffusion direction indicates that the tube-like micro structure of the ductal system can be tracked by DT-MRI
Can Contrast-Enhanced MRI Be Used to Identify Those
Breast Tumors at High-Risk for Disease Recurrence with High Specificity
Even Prior to Preoperative Chemotherapy?
Ka-Loh Li1, Savannah C. Partridge2, Bonnie N. Joe1, Jessica Gibbs1, Ying Lu1, Evelyn Proctor1, Laura Esserman1, Nola M. Hylton1
1University of California, San Francisco, San Francisco, USA; 2University of Washington, Seattle, Washington, USA
This study aimed to evaluate the predictive value of high spatial resolution signal enhancement ratio (SER) MR imaging performed prior to neoadjuvant chemotherapy in patients with breast cancers. The major finding was that breast tumor volume calculated from the number of voxels with high SER value and the volume (in voxels) of cancerous breast tissue infiltrating into the parenchyma (Such infiltrating voxels were of low SER) were important predictors of recurrence. 6/10 of patients with early recurrence (within 1.6 years) and 6/9 of deceased patients were identified. This study demonstrated the potential of pre-neoadjuvant chemotherapy contrast-enhanced MRI in predicting disease recurrence.
Prediction of Long Term Breast Cancer
Survival Using MR Metabolomics
Tone Frost Bathen1, Beathe Sitter1, Hans Fjøsne2, David E. Axelson3, Steinar Lundgren1, 4, Ingrid S. Gribbestad1
1NTNU, Trondheim, Norway; 2St Olav Univeristy Hospital, Trondheim, Norway; 3MRi_Consulting, Kingston, Canada; 4St Olav University Hospital, Trondheim, Norway
MR metabolomics is the systematic study of MR visible small-molecules. Metabolite profiles of breast cancer tissue have been found to correlate to lymph node and hormone receptor status. In this study, breast cancer tissue from patients with invasive ductal carcinomas were analysed by HR MAS MRS. The spectra were further analysed in a supervised manner with Support Vector Machine, relating selected variables from the spectra to the actual survival status 4 years after surgery. This preliminary investigation suggests that MR metabolomics can predict long term breast cancer survival. Early identification of patients with poor prognosis may be important for the administration of more aggressive treatment and closer follow-up of this patient group.
Statistical Metrics to Determine When Water Exchange
Should Be Incorporated Into DCE-MRI Analysis: Simulations and
Experimental Breast Cancer Results
Thomas Yankeelov1, E. Brian Welch2, Anuradha Chakravarthy3, Robert Lee3, Darla Freehardt3, Ingrid Mayer3, Ingrid Meszoely3, Mark Kelley3, Julie Means-Powell3, John C. Gore3
1Vanderbilt University, Nashville, USA; 2Philips Medical Systems, USA; 3Vanderbilt University, USA
We apply three statistical measures to determine when water exchange effects should be incorporated in DCE-MRI data analysis. Simulations predict that after ti becomes above ~0.1 s, the FXR model will be preferred and this is what is seen in the breast cancer data. We conclude that, at least in the case of human breast cancer, water exchange effects should be explicitly incorporated in the analysis of DCE-MRI data.
Combined 2D MR Spectroscopy and Dynamic Contrast
Enhanced MRI for Breast Cancer Detection
Scott Logan Lipnick1, Xiaoyu Liu1, Nanette DeBruhl1, M. Albert Thomas1
1University of California, Los Angeles, Los Angeles, California , USA
The reported specificity of DCE-MRI has been variable, ranging from 37% to 97%, resulting in many unnecessary biopsies of benign lesion. Studies of breast cancer cell lines and breast tumors have consistently shown that choline (Cho) is elevated in malignant lesions. Magnetic Resonance Spectroscopy (MRS) allows noninvasive measurements of the concentrations of metabolites such as Cho in human breast. The purpose of this pilot study was to combine the DCE breast MRI analysis with localized two-dimensional MR correlated spectroscopy (2D L- COSY) in order to increase the specificity of breast cancer detection.
The Interaction of Gadolinium Based MR Contrast
Agents with Choline
Robert E. Lenkinski1, Mostafa Elian1, Xiaoen Wang1, S Nahum Goldberg1
1Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA
There is increasing interest in applying proton MRS to increase the specificity of characterizing breast lesions. The most common approach is to employ DCEMRI to identify suspicious lesions followed by localized proton MRS to determine either the presence of choline containing compounds or their absolute concentrations. The implicit assumption is that that the administration of a Gadolinium based contrast agent has little or no effect on the choline resonance. We report both in vitro and in vivo investigations of the effects of six commercially available contrast agents on the choline resonance. All of the negatively charged chelates showed significant effects.