Chemical shift imaging at 3 Tesla: effect of echo time on assessing bone marrow abnormalities

Our purpose is to test the effect of varied in-phase (IP) and opposed-phase (OP) sequence order on characterizing marrow signal changes at 3T. The study was HIPAA compliant and IRB approved. Informed consent was waived. At 3T, IP and OP sequences were acquired in three patients with biopsy-proven osteosarcomas, using two methods: approach 1 (OP acquisition before IP acquisition) and approach 2 (OP after IP). Signal intensity (SI) measurements in 12 locations of biopsy-proven osteosarcoma and in six locations with normal bone marrow were performed independently by two experienced musculoskeletal radiologists. The signal intensity ratio (SIR) was measured within the marrow where there was T1 signal lower than skeletal muscle. A SIR < 20 % on the OP compared with IP imaging was considered positive for marrow replacement, while SIR > = 20 % was considered negative. Interobserver agreement was measured by the Lin concordance correlation coefficient (CCC). In 75 % (18/24) of locations within the biopsy-proven tumors, the SIR was > 20 % (SI drop more than 20 % in OP compared to IP) using approach 2 and in 100 % (24/24) of the locations the SIR was < 20 % (SI drop less than 20 % in OP compared to IP) using approach 1, indicating a high percentage of false-negative results by approach 2, and no false-negative results with approach 1. There was good agreement between observer measurement (CCC = 0.96). At 3T, the OP sequence should be acquired prior to the IP sequence, because susceptibility artifacts on a later-acquired OP sequence may lead to an erroneous interpretation of marrow signal abnormalities.