Abstract #1455

In-Vivo Sub-Minute rNOE Mapping Using AutoCEST: a Machine-Learning Approach for CEST/MT Protocol Invention and Quantitative Reconstruction

Or Perlman¹, Bo Zhu^1,2, Moritz Zaiss^3,4, Naoyuki Shono⁵, Hiroshi Nakashima⁵, E. Antonio Chiocca⁵, Matthew S. Rosen^1,2, and Christian T. Farrar¹

¹Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital and Harvard Medical School, Charlestown, MA, United States, ²Department of Physics, Harvard University, Cambridge, MA, United States, ³Magnetic Resonance Center, Max Planck Institute for Biological Cybernetics, Tübingen, Germany, ⁴Department of Neuroradiology, University Clinic Erlangen, Erlangen, Germany, ⁵Brigham and Women’s Hospital and Harvard Medical School, Boston, MA, United States

The long acquisition-time and the semi-quantitative nature of the typical CEST-MRI experiment constitute a major obstacle for its clinical adoption. Recently, a machine-learning approach termed AutoCEST was developed, for the automatic design of the optimal acquisition schedule and the reconstruction of quantitative 2-pool CEST maps. Here, we expand this approach for in-vivo scenarios, by incorporating the semisolid-pool into the underlying computational-graph and allowing 3 pools. AutoCEST was evaluated for quantitative rNOE mapping using a GBM mouse model, resulting in a total acquisition and reconstruction times of 49.15s. The tumor rNOE volume-fraction was significantly decreased, in agreement with previous human studies.

This abstract and the presentation materials are available to members only; a login is required.

Join Here