Author: biho/ktg

Source: RSNA 2017

Last Updated: December 5, 2017

Primovist® in Korean Guidelines

A Korean study confirms the advantages of having included specific features of HCC appearance with Primovist®-MRI in the 2016 Korean Society of Abdominal Radiology (KSAR) consensus statement.

Shinhye Hwang, Yonsei University, Seoul, Korea, performed a retrospective study to validate three of the new Korean Society of Abdominal Radiology (KSAR) consensus statements for liver MR imaging using the liver-specific contrast agent Primovist®. Evaluation following the LI-RADS lexicon (v. 2017) served as standard of truth.

Method

Hwang included 177 patients with 241 liver nodules. All patients had undergone a 3T Primovist®-enhanced MRI.

Two experienced radiologists evaluated the images independently of each other. If necessary, final diagnosis was reached in consensus.

KSAR recommendations followed in lesion assessment

  1. Washout appearance should be determined on either the portal-venous phase (PVP) or the in the transitional phase of Primovist®-MRI.

    1. Hwang assessed hypointensity in the transitional phase
  2. It needs to be considered that capsule appearance looks differently with Primovist<sup>®</sup>-MRI than with extracellular contrast agents

    1. Hwang interpreted a non-enhancing rim as capsule
  3. For HCCs larger than one centimeter, additional diagnostic criteria other than typical vascular changes are not required for diagnosis.

    1. Hwang set a cut-off at one centimeter for lesions showing arterial phase hyperenhancement
    2. Hwang performed univariate and multivariate analysis and calculated sensitivity, specificity and accuracy. LI-RADS lexicon (v. 2017) was used as standard of truth.

Results

149 HCCs and six other malignancies had been histopathological proven. 86 lesions were benign, 16 of which had been histopathologically confirmed and 70 clinically diagnosed.

Uni- and multivariate analysis showed a statistically significant increase in sensitivity compared to LI-RADS:
Sensitivity: LI-RADS: 57.7%, uni- and multivariate analysis: 67.8-81.9% (p<0.001 for all).

Specificity was similar: LI-RADS: 97.8%, uni- and multivariate analysis: 93.5-97.8%. Specificity was slightly lower for multivariate analysis including all three consensus statements (93.5%, p=0.041) and for univariate analysis with size cutoff at 1cm (94.6%, p=0.078).

Accuracy was statistically significantly different: LI-RADS: 73%, uni- and multivariate analysis: 79.3-86.3% (p<0.001 for all).

Conclusion

Hwang concluded that the new KSAR guidelines improve radiologic practice. Hypointensity in the transitional phase and non-enhancing rim as capsule appearance increased sensitivity of HCC diagnosis without impairing specificity. A cut-off of one centimeter for hyperintense appearance on the arterial phase also increased sensitivity but at the same time slightly decreased specificity.

References

Korean Society of Abdominal Radiology. Diagnosis of Hepatocellular Carcinoma with Gadoxetic Acid-Enhanced MRI: 2016 Consensus Recommendations of the Korean Society of Abdominal Radiology. Korean J Radiol2017;18(3):427-43.