A.V. Protopopov – Post-graduate Student, Moscow Institute of Physics and Technology (State University)
(Dolgoprudniy, Moscow region); Federal Scientific Clinical Center of Children's Hematology, Oncology and Immunology n.a. Dmitry Rogachev (Moscow) E-mail: proto.alex@hotmail.com
M.V. Gulyaev – Ph.D. (Phys.-Math.), Senior Research Scientist, Faculty of Fundamental Medicine, M.V. Lomonosov Moscow State University E-mail: mihon-epsilon@ya.ru
Yu.A. Pirogov – Dr.Sc. (Phys.-Math.), Professor, Faculty of Physics, M.V. Lomonosov Moscow State University; National Research Nuclear University MEPhI, Engineering Physics Biomedical Institute E-mail: yupi937@gmail.com
In magnetic resonance imaging (MRI), the gradient recalled echo (GRE) sequence became a powerful tool for classification of biological tissues due its sensitivity to structural heterogeneities of mesoscopic and macroscopic scale. The atomic scale heterogeneities, determined by spin-spin interaction of protons, can be imaged in contrast of transverse relaxation time T2, obtained in spin-echo (SE) sequence. For many years, the SE sequence was the only way to measure T2 accurately because, in GRE sequence, this parameter is masked by more rapid decay of the signal due to magnetic gradients. First attempts to evaluate - even not exactly measure - T2 in GRE sequences revealed serious algorithmic problems. Recently introduced new solution to this problem - multipoint algorithm with clamping (MPC) - opens the possibility for accurate and fast measurements of T2 in GRE sequences. The present short communication presents quantitative statistical confirmation of its performance on phantoms. It is shown that MPC algorithm produces accurate measurements of T2 in GRE sequences with average discrepancies of about 2% on small magnetic heterogeneities and 8% on the large ones with Pearson coefficients in the range 0.9-0.95. Commonly accepted in practice evaluation of T2 on GRE sequences by means of monoexponential fitting (the so-called values) gives inadequate average values with errors of about 30% and Pearson coefficient of only 0.7.