Introduction

Many hepatic cellular cancer (HCC) patients have a background of diffuse liver disease of which liver fat, iron, and combined overload are common manifestations and may cause fibrosis, cirrhosis. A variety of quantitative MR imaging techniques can be used to non-invasively evaluate the above pathological changes, providing valuable information for the clinical diagnosis and treatment of HCC patients. Multi-echo magnetic resonance spectroscopy (MRS) sequences have proven to be valuable tools for the evaluation of steatosis and iron overload in the liver^1,2. T2 relaxation time has also been reported for detection and staging of liver fbrosis^3,4. But there is no literature to compare them together. The purpose of this study was to prospectively evaluate a new radial turbo spin echo (TSE) sequence for the assessment of hepatic fibrosis and steatosis in comparison with multi-echo MRS in a group of HCC cases.

Methods

This study included 40 patients (12 female, age 58 +/- 15 years) with hepatic cellular cancer (HCC) who underwent magnetic resonance imaging (MRI) including multi-echo MRS (HISTO)⁵ and a prototype radial TSE sequence with echo-sharing reconstruction and T2 mapping⁶ before surgery on a 3T system (MAGNETOM Prisma, Siemens Healthcare, Erlangen, Germany). The HISTO sequence was executed with TE =12/24/36/48/72ms, TR=3000ms, bandwidth=1200Hz, averages=1, voxel size=2×2×2cm³, breath-hold and acquisition time=15sec. The spectroscopy voxel was positioned using three-plane localizers to avoid the main vessels. The Radial_TSE sequence was executed twice, with and without fat saturation (FS), all other parameters remained the same: TR/TE1 = 3000/9ms, FOV =380×380mm², scan matrix=320×320, voxel size=1.2×1.2×4.0mm³, slices=38, trajectory=radial, radial views=360, free breathing and acquisition time = 2min18sec. T2 parametric maps and synthetic T2W images were generated inline after data acquisition. Fat fraction (HISTO_FF) and HISTO_R2_water values were calculated inline from HISTO, while HISTO_T2_water was manually calculated as HISTO_T2_water = 1/HISTO_R2_water. RadTSE_T2 (without fat saturation) and RadTSE_FS_T2 (with fat saturation) values were obtained from two different T2 parametric maps with the ROI drawn in the same position as the voxel from HISTO（Figure 1）, and their difference RadTSE_ΔT2 = RadTSE_T2 - RadTSE_FS_T2 was calculated. All patients underwent surgical resection of partial liver tissue within 5 days after examination and were pathologically confirmed to have different degrees of liver fibrosis and steatosis. Statistical analyses were performed using MedCalc Software (MedCalc Software Ltd, Belgium). Paired t-test, Pearson coefficient, and Bland-Altman analysis were used to assess the difference, correlation, and agreement between RadTSE and HISTO sequences. p <0.05 was considered statically significant.

Results

As shown in Figure 1, the RadTSE sequence had good image quality using free breathing. Figure 2 shows the moderate and high correlations of T2 values between the RadTSE (r=0.54, p<0.05) and RadTSE_FS (r=0.69, p<0.05) and HISTO sequences, and good agreement (difference value T2=2.5ms) of T2 values between the RadTSE_FS and HISTO sequences. Table 1 shows T2 values were significantly higher for RadTSE_FS (30.59±4.89 ms) compared with HISTO (28.26±3.99 ms) (p<0.05). Figure 3 shows RadTSE ΔT2 had excellent correlation with HISTO_FF.

Discussion

This study showed that RadTSE_FS_T2 had a better correlation with HISTO_T2_water than RadTSE_T2. The reasons may be that both RadTSE_FS_T2 and HISTO_T2_water exclude the influence of fat signals in their calculations. Although RadTSE_FS_T2 was larger than HISTO_T2_water, the difference of 2.5ms was small. This suggests that RadTSE_FS_T2 may be interchangeable with HISTO_T2_water in quantitative assessment of liver iron deposition or fibrosis. In addition, RadTSE_FS sequence can be excuted with free-breathing and can quantify the whole liver, whereas HISTO requires breath-holding and can only measure parts of the liver. Takeshi et al⁷. proposed that the T2W image with FS can be considered water only, whereas the T2W image without FS is the water-plus-fat. The fat fraction can be calculated from T2W images with and without FS using a specific equation. In this study, we found that the RadTSE_FS_T2 was lower than that of RadTSE_T2 and their difference had excellent correlation with HISTO_FF. This finding indicates RadTSE ΔT2 may be a indicator for hepatic steatosis quantification. However, the underlying reason behind this is unclear and require further investigation.

Conclusions

Radial TSE and multi-echo MRS sequences may have similar value for the simultaneous evaluation of fibrosis and steatosis in liver with free breathing.

Acknowledgements

Thank Dr. Stephan Kannengiesser(Application Predevelopment, Siemens Healthcare GmbH, Erlangen, Germany) for his precious comments during the revision of this abstract.