Meeting Banner
Abstract #0379

A Vectorized Formalism for Efficient SAR Computation in Parallel Transmission

Mihir Pendse 1 and Brian Rutt 1

1 Radiology, Stanford University, Stanford, CA, United States

We formulate an expression for a normalized SAR cost function that encapsulates all local/global SAR and average power terms into a single expression that is evaluated to determine whether a given RF pulse sequence abides by all time-averaged safety/hardware limits. A vectorized oracle for evaluating this function and its derivatives is described and implemented. With the use of a GPU, the computation time of the oracle is shown to be approximately constant even for large number of local SAR voxels and subpulses. The vectorized oracle allows for direct SAR-aware pulse optimization without the need for SAR matrix formation and compression.

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

Join Here