gammaSTAR : A framework for the development of dynamic, real‐time capable MR sequences

PURPOSE: To present the real-time capability and advanced MR sequence library of the MR sequence development framework gammaSTAR. METHODS: The presented platform consists of four different components: (1) a frontend for sequence development combined with a Python backend for sequence generation; (2) a Lua backend for the creation of hardware instructions; (3) a vendor-specific driver for translation of these instructions into scanner-specific objects; and (4) an interface for real-time feedback capability. In vivo measurements of the same volunteer were performed for comparison of imaging and spectroscopy sequences implemented in this framework with those of one main vendor (Siemens Healthineers) at magnetic field strengths of 3 T and 1.5 T. Prospective motion correction was integrated into a spin echo EPI sequence to demonstrate the real-time feedback capability. RESULTS: /s in white matter. Results of pseudo-continuous arterial spin labeling gradient and spin-echo (pCASL GRASE) and 3D radial UTE imaging demonstrate the ability to run complex sequences without long sequence preparation times. Prospective motion correction is possible by means of real-time feedback and shows much fewer movement artifacts with mean voxel displacement of 1.63 mm (uncorrected) versus 0.37 mm (corrected). All images were reconstructed using the vendor's reconstruction pipeline. CONCLUSION: The platform gammaSTAR allows for MR sequence development with real-time feedback capability demonstrated by a large number of MR sequences and applications.