A comparative study at 3 T of sequence dependence of T2 quantitation in the knee

Abstract

T2 mapping has been used widely in detecting cartilage degeneration in osteoarthritis. Several scanning sequences have been developed in the determination of T2 relaxation times of tissues. However, the derivation of these times may vary from sequence to sequence. This study seeks to evaluate the sequence-dependent differences in T2 quantitation of cartilage, muscle, fat and bone marrow in the knee joint at 3 T. Three commercial phantoms and 10 healthy volunteers were studied using 3 T MR. T2 relaxation times of the phantoms, cartilage, muscle, subcutaneous fat and marrow were derived using spin echo (SE), multiecho SE (MESE), fast SE (FSE) with varying echo train length (ETL), spiral and spoiler gradient (SPGR) sequences. The differences between these times were then evaluated using Student's t test. In addition, the signal-to-noise ratio (SNR) efficiency and coefficient of variation of T2 from each sequence were calculated. The average T2 relaxation time was 36.38+/-5.76 ms in cartilage and 34.08+/-6.55 ms in muscle, ranging from 27 to 45 ms in both tissues. The times for subcutaneous fat and marrow were longer and more varying, ranging from 41 to 143 ms and from 42 to 160 ms, respectively. In FSE acquisition, relaxation time significantly increases as ETL increases (P<.05). In cartilage, the SE acquisition yields the lowest T2 values (27.52+/-3.10 ms), which is significantly lower than those obtained from other sequences (P<.002). T2 values obtained from spiral acquisition (38.27+/-6.45 ms) were higher than those obtained from MESE (34.35+/-5.62 ms) and SPGR acquisition (31.64+/-4.53 ms). These differences, however, were not significant (P>.05). T2 quantification can be a valuable tool for the diagnosis of degenerative disease. Several different sequences exist to quantify the relaxation times of tissues. Sequences range in scan time, SNR efficiency, reproducibility and two- or three-dimensional mapping. However, when choosing a sequence for quantitation, it is important to realize that several factors affect the measured T2 relaxation time.