Studies estimating liver fibrosis in animal models and liver cirrhosis in humans using magnetic resonance imaging (MRI) T1rho have been reported^1-3. In all these studies, consistent spin-lock times (TSL) of 1, 10, 20, 30, 40 and 50 milliseconds (ms) were used for T1rho measurement. The purpose of this study was to evaluate the potential of T1rho for liver fibrosis evaluation and to explore the best TSL points with optimized diagnostic efficiency.

Thirteen healthy control subjects (mean age=38.4±13.2 years; F/M=5/8) and eight patients with clinically diagnosed liver fibrosis (mean age=50.5±9.1 years; F/M=3/5) were scanned on a 3.0T scanner (Ingenia, Philips, Healthcare, Best, the Netherlands) using a multi transmit RF system and a 32 channel phased-array receiver coil. For T1rho measurement, a rotary echo spin-lock pulse was implemented in a 3D balanced turbo field echo (b-TFE) sequence (TR/TE=3.8/1.82 ms, FOV=300×360 mm, resolution=2.5×2.81×6.00 mm, slice thickness=6 mm, slice number=8, NSA=2, B1max=11.5μT, TFE factor=64 ). The spin lock frequency was set to 500HZ and the TSL were 1, 10, 20, 30, 40, 50 and 60 ms. The T1rho map was generated on a pixel-by-pixel basis on Philips Research Integrated Development Environment (PRIDE) software written in Interactive Data Language using a mono-exponential decay model: M(TSL)=M0*exp(-TSL/T1rho). T1rho maps were constructed by using all seven TSLs points, six TSLs points of 1, 10, 20, 30, 40 and 50 ms, or three TSLs of 10, 30 and 50 ms, respectively. Mean T1rho values of healthy control subjects and patients with liver fibrosis were calculated and compared using Student t-test. The differences of T1rho values derived from different TSLs were compared using one way ANOVA by using IBM SPSS Statistics 20.0 (Armonk, New York, USA). In addition, a receiver operating characteristic (ROC) curve analysis was performed to find the best TSL points for the prediction of liver fibrosis. P<0.05 indicated a significant difference.

Typical T1rho maps are shown in Figure1. Mean T1rho values of all TSLs combination were higher in patients with liver fibrosis than those of healthy subjects, but significant differences between the two groups were only found in 7-TSLs and 6-TSLs (Table 1). Both 7-TSLs and 6-TSLs had moderate diagnostic efficacy, and AUC of 7-TSLs was slightly higher than 6-TSLs (0.769>0.740). Mean T1rho values of different TSLs combination had no significant differences in both control subjects and patients with fibrosis (P>0.05).At 3.0T, severe artifacts caused by spin locking due to B0 and B1 inhomogeneity are observed. We used dual-source parallel RF excitation and deliberately placed pencil-beam shimming to reduce dielectric shading and improve the B1 and B0 homogeneity. The results of this study demonstrated that 7-TSLs and 6-TSLs had highest and second highest significant differences between control subjects and patients with fibrosis. This difference indicated the potentiality of MR T1rho for the diagnosis of liver fibrosis, which is consistent to the results of animal models^1,2. The reduction of TSL number can decrease the total scantime and reduce the SAR and RF energy deposition⁴. In our study, T1rho values of 7-TSLs, 6-TSLs and 3-TSLs all had no significant differences in the two groups, while 3-TSLs did not provide diagnostic capacity for liver fibrosis.

7-TSLs and 6-TSLs are recommended for the diagnosis of liver fibrosis, and 7-TSLs method has slightly higher diagnostic performance.