Meeting Banner
Abstract #2460

A Novel Algorithm for Improved Pixel-By-Pixel T2* Mapping

Yanqiu Feng1, Peter David Gatehouse2, 3, David N. Firmin2, 3, Dudley J. Pennell2, 3, Wufan Chen1, Taigang He2, 3

1School of Biomedical Engineering, Southern Medical University, Guangzhou, China; 2Cardiovascular Biomedical Research Unit, Royal Brompton Hospital, London, United Kingdom; 3National Heart and Lung Institute, Imperial College, London, United Kingdom

MRI relaxometry mapping (pixel-by-pixel) is sensitive to the noise especially when high-resolution or rapid scans are required. Considering decay signals as the objects to be averaged, we aimed at improving the precision of T2* mapping through selectively Averaging the Decay signals with Similar Underlying Relaxation rates (ADSUR) before curve fitting. This novel algorithm was tested on T2* mapping of simulation data and an ex vivo heart with iron overload. The results demonstrate that the T2* relaxometry mapping can be improved by filtering the serial images with the proposed ADSUR algorithm, independent of curve-fitting models used.

Abstract PDF

Keywords

adjacent aimed applied audience averaging biomedical blurring calculation cardiovascular cause certain changing china clinically clinicians close college considering containing contrast controls cost curve decay decaying denotes details directly distance distortion domain edges engineering engineers ensure equal errors especially euclidean evaluate excitation filter filtered filtering fitted fitting formulated fourth function gatehouse heart hospital impact imperial improve improved improving in vivo independent indicating influence institute intensities interfaces introduced iron kingdom known leads liver local lung mapping maps medical model models national noise normalizing novel objects offset original output overload patterns peter pixel pixels precision precisions preserve preserving produces proposed quantitatively rapid reduce reduces repeated representing required resolution respectively root royal scanner school selectively sensitive serial similarity simulated simulation smooth smoother southern spacing spatial spatially square suppress synthesized table target third traditional true truncation underlying unit utilized variances varying vector wise