Compressive Slice Encoding for Metal Artifact Correction in MRI

Severe metal-induced artifacts often render MR images non-diagnostic. Recently proposed Slice Encoding for Metal Artifact Correction (SEMAC) fully corrects metal-induced artifacts in MR images. However, the clinical utility of SEMAC is hampered with prolonged scan times and relatively low signal-to-noise ratio (SNR). This work presents an improved SEMAC reconstruction technique referred to as Compressive SEMAC that exploits the compressibility of SEMAC-encoded slices. The SEMAC-encoded slices are vastly down-sampled along phase encoding directions. The down-sampled SEMAC-encoded slices are reconstructed by iterating between joint noise suppression of SEMAC-encoded slices and imposing k-space consistency. In-vivo experimental results demonstrate that the proposed compressive SEMAC greatly reduces scan times incurred to fully correct metal artifacts without compromising the SNR of SEMAC-corrected results.