Abstract #3066
DCE-MRI Analysis using Model-Based Classification Shapes with Non-Negative Least-Squares
Zaki Ahmed 1 and Ives R Levesque 1,2
1
Medical Physics Unit, McGill University,
Montreal, Quebec, Canada,
2
Research
Institute of the McGill University Health Center,
Montreal, Quebec, Canada
We describe a new analysis method for DCE-MRI which is
based on shape analysis and uses the Tofts model to
define the classification shapes. Non-negative
least-squares (NNLS) is used to identify the weight of
these shapes in measured data. We apply this method to a
dataset of breast cancer patients undergoing neoadjuvant
chemotherapy, and show that our method can predict
pathologic complete response (pCR) in images from
pre-treatment or after one cycle of therapy. Our results
also suggest that the method is robust to inaccuracies
in the T1 and arterial input function (AIF).
This abstract and the presentation materials are available to members only;
a login is required.
Join Here
