INTRODUCTION
Hepatocellular carcinoma (HCC) is the most common malignancy cancer with fast rising incidence in both men and women and the second major contributor to cancer-related mortality worldwide^{1, 2}. Diffusion-weighted imaging (DWI), an unenhanced, robust and reliable technique which has been widely used in the evaluation of liver tumors and showed its usefulness and potential in HCC diagnosis^3-6. However, acquired high b-value results in low signal-to-noise ratios (SNRs) and artifacts such as eddy current-induced distortions, with prolonged acquisition time as requirement of longer echo times (TEs)⁷. By contrast, synthetic DWI is a mathematical computation technique using derived apparent diffusion coefficient (ADC) and acquired b-values to synthesize desired b-values images with at least two different b-values⁷, which may tackle the limitations of conventional directly acquired DWI by achieving background suppression of a very high b-value DWI without additional acquisition scan time. Though previous studies have demonstrated the usefulness of synthetic DWI in detection of prostate cancers, pancreatic cancers, hepatic metastases, and breast cancers^8-15, almost no study has focused on synthetic DWI in HCC evaluation. Therefore, the purpose of this study is to investigate the clinical feasibility and diagnosis performance of synthetic DWI in patients with HCC compared with conventional acquired DWI qualitatively and quantitatively.
METHODS
This prospective study was approved by the Institutional Review Board with written informed consent obtained. Twenty-seven patients with HCC were prospectively involved undergoing conventional acquired DWI at b-values of 50, 800, 1000, 1200, and 1400 s/mm² and synthetic DWI, of which synthetic images were calculated at b=1000, 1200, and 1400 s/mm² from acquired images using b-values of 200 and 600 s/mm². Two readers reviewed image sets of both synthetic and conventional acquired images of 1000, 1200, and 1400 s/mm² independently and rated image quality and lesion conspicuity. Overall image quality was assessed using 4-point scale: 1=poor, 2=moderate, 3=good, 4=excellent, and lesion conspicuity based on 5-point scale: 1=barely conspicuous, 2=fairly conspicuous, 3=moderately conspicuous, 4=well seen, 5=very well and sharply defined. Signal intensity (SI) was measured by placing round regions of interest (ROIs) on lesions and liver parenchyma avoiding any focal lesions and major vessels. Quantitative analysis was performed as follows: signal-to-noise ratio (SNR) = SI_lesion / SD_{background noise}, contrast-to-noise ratio (CNR) = (SI_lesion - SI_{liver parenchyma}) / SD_{background noise}, SIR= SI_lesion / SI_{liver parenchyma}, cancer-to-parenchyma contrast ratio =(SI_{liver parenchyma} - SI_lesion) / (SI_{liver parenchyma} + SI_lesion). The grading scores at different b values were compared using the Friedman test. SNR, CNR, and SIR were compared using analysis of variance. In case of significant results, post-hoc analysis was performed. The interobserver agreement were also evaluated by intraclass correlation coefficient (ICC) for quantitative assessment. Two-sided p value <0.05 was considered as statistically significant.
RESULTS
There were no significant differences in overall image quality at synthetic DWI₁₀₀₀, conventional DWI₁₀₀₀, synthetic DWI₁₂₀₀, and conventional DWI₁₂₀₀ (p>0.05), whereas synthetic DWI₁₄₀₀ showed lower overall image quality than other DWI sets (p<0.001). Comparison of lesion conspicuity at both synthetic and conventional acquired three DWI sets revealed no significant differences (p>0.05). SNR, CNR and SIR showed a continuous decrease with increasing b-values, with significant differences between each of b-values sets (P<0.001), while comparison between synthetic DWI₁₀₀₀ and conventional DWI₁₀₀₀, synthetic DWI₁₂₀₀ and conventional DWI₁₂₀₀, synthetic DWI₁₄₀₀ and conventional DWI₁₄₀₀ presented no significant difference (p>0.05). Cancer-to-parenchyma contrast ratio almost unchanging showed no significant difference (p>0.05). Both readers showed similar assessment in SNR, CNR, SIR and cancer-to-parenchyma contrast ratio at all DWI sets, with moderate to good interobserver agreement (ICC=0.63-0.81, P<0.05).
CONCLUSION
Synthetic DWI is of great comparable image quality and lesion conspicuity in HCC diagnosis versus conventional acquired DWI, while reducing scan time and avoiding the disadvantages of performing DWI at high b-values.

Acknowledgements

No acknowledgement found.