Purpose: To compare the homogeneity and degree of fat saturation when using water excitation (WE) at different binominal scheme (11, 121, 1331, 14641) in iShim (sequence integrated shimming) diffusion-weighted MRI (DWI) of liver. And to identify if different binominal scheme have effects on apparent diffusion coefficient (ADC) values reproducibility.

Materials and Methods: This prospective study was approved by the appropriate ethics committee and written informed consent was obtained from all healthy volunteers. Between August 2015 and September 2015, 30 healthy volunteers (16men, 14 women; mean age, 23 years) underwent iShim diffusion-weighted imaging using water excitation technique with binominal scheme 11 (A), 121 (B), 1331 (C), 14641 (D) and T2WI images with TSE sequence at 3.0 T. The signal intensity of liver, spleen and subcutaneous fat area in abdominal wall, background noise (standard deviation of air signal intensity) of the iShim sequence and the signal intensity of subcutaneous fat area in abdominal wall were measured. The ADC values of liver, spleen, muscle were measured on the ADC maps generated by using prototype software (Siemens Healthcare, Erlangen, Germany). The SNR, CNR, the standard deviation of signal intensity in subcutaneous fat area in abdominal wall (SIvar), the difference of the mean signal intensity in subcutaneous fat area between T2WI TSE and iShim DWI sequence (dertaFAT) were calculated. The above quantitative parameters were compared with Repeated Measures ANOVA and a bonferroni adjustment was applied for multiple comparisons. The image quality about ghost artifacts, image distortion, fat saturation efficiency were scored by two radiologists with more than 10 years experience independently on the basis of subjective four-point image quality scales and were compared with Friedman signed rank test. A P <0.05 indicate statistically significant difference.

Results: The subjective quality scores(SQS) of Group C (3.45±0.51) and D (3.60±0.37) were higher than Group A (1.10±0.21) and B (2.10±0.83), whose difference was statistically significant (P<0.001), but the difference between Group C and D did not demonstrate statistically significant (P<0.001). The SIvar of Group D(5.91±3.95) and Group C(6.87±2.28) were lower than the other groups(12.12±7.02, 9.59±3.20) with the difference statistically significant( P<0.001),but there were no significant difference between Group C and Group D for the SIvar(P>0.05). No statistically significant difference were found for SNR values (28.29±7.22, 41.49±4.09, 19.56±0.30, 20.81±7.98), CNR values (239.31±93.00, 89.38±37.97, 86.88±38.60, 140.87±6.43),dertaFAT(7.15±2.58,7.16±0.69,8.36±0.51,9.95±2.85),liver’s ADCs[(1.154±0.206)×10-3 mm 2/s, (1.165±0.120)×10-3 mm 2/s, (1.099±0.135)×10-3 mm 2/s, (1.297±0.044)×10-3 mm 2/s], spleen’s ADCs[ (0.702±0.121)×10-4 mm 2/s, (0.775±0.031)×10-4 mm 2/s, (0.788±0.020)×10-4 mm 2/s, (0.742±0.090)×10-4 mm 2/s].

Conclusion: The binominal scheme 1331 and 14641 in WE can improve the homogeneity of fat saturation in abdominal iShim DWI without effecting ADC values.