Robust extrapolated semi-solid magnetization transfer reference fitting for amide proton transfer (APT) imaging quantification

Motivation: The commonly used APTw metric is significantly confounded by the NOE effect. New quantification method is needed for more accurate APT imaging. Goal(s): To propose a robust EMR fitting method for APT imaging quantification on data with various noises and imperfections. Approach: A bias term was added to the original EMR model, and a two-step coarse-to-fine fitting was implemented for more robust and accurate fitting. Results: The proposed EMR fitting showed good performances in terms of fitting errors in phantom and contrast and image quality in tumor patients. Preliminary results showed that it can improve the differentiation of treatment effect and tumor recurrence. Impact: The proposed method enabled more robust APT imaging quantification, which provides stronger APT signal intensity values than traditional MTR asymmetry analysis. Preliminary results proved its effectiveness in distinguishing treatment effect and tumor recurrence and more applications are expected.