Currently detection of breast tumors is most often accomplished with x-ray mammography. This is true despite the fact that mammography misses 5-15% of palpable breast cancers and has a positive predictive value of only 15-30%. Consequently, mammography requires a much higher rate of biopsy than occurrences of malignancy. Alternatively, contrast-enhanced magnetic resonance imaging (MRI) has been shown to be highly sensitive for breast cancer. However, these benefits are offset by the fact that it too has frequent false positive findings, and has therefore not been routinely adopted. More recently, T1-weighted dynamic contrast- enhanced magnetic resonance imaging (DCE-MRI) combined with two-compartment pharmacokinetic modeling has demonstrated potential for better differentiating benign from malignant tumors. Still, these DCE methods have not had the impact on specificity as originally hoped. This has been attributed to the fact that the indices derived from the DCE approach, such as Ktrans, the plasma-tissue contrast agent transfer constant, are influenced by several factors together including blood volume, flow and vascular permeability, often making interpretation difficult and thus sub-optimal for distinguishing benign from malignant lesions. Consequently, we propose an alternative """"""""dynamic-contrast"""""""" method, which directly measures tumor blood volume and other vascular parameters individually, as a way to provide both specific and sensitive markers for breast cancer imaging. The approach is based on T2*-weighted dynamic susceptibility contrast (DSC) MRI methods. The challenge with the DSC approach, and the reason why it has only rarely been used for perfusion measurements outside of the brain, is that it can be confounded by T1 and T2/T2* changes, which occur when contrast agent leaks out of the vessels. Thus for DSC imaging to be feasible and meaningful for breast MR imaging, where Gd contrast agents readily leak out of the vessels, these confounding factors must be minimized and/or corrected. In this regard, based on 10+ years of performing DSC studies in leaky brain tumors, and a recently study performed by us to compare several different DSC methods, we have developed what we hypothesize to be the complete solution to breast tumor perfusion imaging. The primary approach is to collect dual-echo spiral based data, which directly negates T1 effects, and then apply a post-processing algorithm to diminish T2 leakage effects. The additional advantage of this approach is that T1-weighted DCE data can be collected, from the first echo, at the same time. This novel combination of technologies is the basis for an invention disclosure, for which Imaging Biometrics LLC will have an exclusive license. It will form the foundation of Imaging Biometric's initial MRI breast perfusion analysis product. Given the statistics of breast cancer and recent recommendations from the American Cancer Society (ACS), the commercial market potential and societal gain for this technology is outstanding. In the U.S. the incidence of breast cancer is approximately 240,000 in 2006 with a compounded annual growth rate of 4.1%. Each year 40,000 women die of breast cancer and one out of eight women will get breast cancer at some point in her lifetime. An estimated 50 million MRI procedures will be performed in 2007 and 45% of those will involve a contrast agent. And just recently the ACS American Cancer Society (March 28, 2007) released new guidelines advising annual MRI screening for women with high lifetime risk of breast cancer, defined as 20% or more. Of the 70 million woman in the U.S. between 30 and 70 years of age, 1-1.5 million fall into one of the high risk groups recommended for routine screening with breast MRI, along with mammography. In response to the need for better and more specific MRI methods, and recent ACS recommendations, Imaging Biometrics LLC proposes to develop and market the novel DSC perfusion technology by addressing two aims.
Aim 1 will test the feasibility and clinical relevance of Imaging Biometric's perfusion technology with preliminary studies performed in patients with breast masses detected on x-ray mammography. The patients studied will be those who are already scheduled for MRI and follow-up biopsy. In parallel, we will work to productize the analysis algorithms (Aim 2) for ease of use and implementation into the platforms of interested vendors such as Confirma Inc (Bellevue, WA). Confirma is a market leader in breast MRI CAD (computer aided detection). It is therefore of great significance that they have already expressed interest and willingness to work with Imaging Biometrics LLC to develop this perfusion technology. Their letter of support is included with this application.

Public Health Relevance

In the U.S. the incidence of breast cancer was approximately 240,000 in 2006 with a compounded annual growth rate of 4.1%. Each year 40,000 women die of breast cancer and one out of eight women will get breast cancer at some point in her lifetime. Given these staggering statistics, it is of paramount importance that we have available to us the best technology available to both detect breast tumors and determine whether they are benign or malignant. Currently, X-ray mammography is the technology most often used to detect breast tumors. This is true despite the fact that mammography misses 5-15% of palpable breast cancers and is not a reliable way to determine if a tumor is benign or malignant. Given that the rate of malignancies is much greater than the number of biopsies, this results in a large number of unnecessary biopsies and anxiety. Consequently, the ultimate goal for breast cancer screening is a technology that can both sensitively detect the presence of breast tumors and determine if the tumor is benign or malignant. Magnetic resonance imaging (MRI) has demonstrated the potential to fulfill this goal. Multiple research studies have confirmed improved cancer detection with contrast-enhanced MRI. Specifically, when a MRI contrast-agent (or dye) is administered intravenously, it often increases the brightness of the breast tumor making it easier to see. However, this type of contrast-enhanced MRI, though more sensitive, is often not much better than x-ray mammography in distinguishing benign from malignant breast tumors. But, if one acquires the MRI images very fast (ie dynamically) while the contrast agent is being administered, additional information about the tumor can be derived. Specifically information about the blood flow and the number of blood vessels can be derived. This information may help to distinguish benign from malignant tumors since malignant tumors typically have a greater number of blood vessels compared to benign tumors. In this regard, our laboratory has developed a dynamic MRI method that has been shown to successfully measure blood volume in brain tumors. We believe that this method should also work for the measurement of blood volume in breast tumors. Our approach is unique, compared to other dynamic MRI approaches that have been tried in the breast, in that it gives a direct measure of blood volume. The other dynamic methods provide indices that are influenced by many factors, making them sometimes difficult to interpret and less useful, than originally hoped, for distinguishing benign from malignant tumors. It is therefore the hope of Imaging Biometrics LLC that this approach will prove feasible and clinically relevant and provide an improved diagnostic technology for breast cancer. To further test and develop the feasibility of this approach for breast imaging we propose to perform pilot studies in patients who are scheduled for MRI and will be undergoing biopsy. In these studies we will compare the information derived from our measurement of tumor blood volume to the indices derived with standard contrast-enhanced MRI and see how each compare to the biopsy results. At the same time, Imaging Biometrics LLC will develop the image analysis software as a low-cost perfusion analysis product. The formation of Imaging Biometrics LLC, and thus the development and marketing of this software, is from our perspective the best way to get this advanced imaging technology into clinical practice in the most time and cost-efficient manner. ? ? ?

Agency
National Institute of Health (NIH)
Institute
National Cancer Institute (NCI)
Type
Small Business Innovation Research Grants (SBIR) - Phase I (R43)
Project #
1R43CA138568-01A1
Application #
7538091
Study Section
Special Emphasis Panel (ZRG1-SBMI-T (10))
Program Officer
Beylin, David M
Project Start
2008-09-01
Project End
2009-02-28
Budget Start
2008-09-01
Budget End
2009-02-28
Support Year
1
Fiscal Year
2008
Total Cost
$107,000
Indirect Cost
Name
Imaging Biometrics, LLC
Department
Type
DUNS #
792265121
City
Elm Grove
State
WI
Country
United States
Zip Code
53122