MY CONTENT

This proof-of-concept investigation used open-source generative adversarial networks to create synthetic lumbar spine MRI STIR from T1 and T2 sequences. Evaluating radiologists found the synthetic volumes were of equal or better quality in 77% of test patients and demonstrated equivalent or decreased motion artifacts in 78% of test patients. The synthetic volumes had high positive predictive value (75%-100%) but lower sensitivity (0%-67%) for common pathologies. The results from this study are a promising first step toward expediting imaging protocols.