Radiology-pathology correlation in staging of liver fibrosis using superb microvascular imaging.

PURPOSE

Progression of liver fibrosis to end-stage disease can potentially be prevented with antiviral treatment. Thus, diagnosis of fibrosis is important in determining treatment protocols. This study aims first, to determine the sensitivity of a novel Doppler method, superb microvascular imaging (SMI), in detecting small vascular structures of the liver compared with other Doppler methods; and second, to choose the best method among these Doppler applications to determine the morphologic changes that occur due to chronic fibrosis. By doing so, the study would be able to provide an ultrasound grading that might differentiate and predict mild and severe liver fibrosis, thus giving rise to a possible alternative to biopsy.

METHODS

A total of 43 patients diagnosed with chronic hepatitis and scheduled for liver biopsy were included. Color Doppler, power Doppler, advanced dynamic flow (ADF) Doppler, color SMI (cSMI) and monochrome SMI (mSMI) Doppler were performed in subcapsular areas of right anterior lobe. Depth from the capsule of the most peripherally located detectable vessel was measured for each Doppler subgroup. Appearance of the vascular tree was categorized into four groups and correlated with pathology results. ROC curve analysis was used to determine if this Doppler classification was statistically significant in differentiating mild and severe forms of fibrosis. Finally, multiple regression analysis was used to determine which Doppler parameter can significantly predict severity.

RESULTS

mSMI and cSMI were found to be superior to other Doppler techniques in detecting the most superficially located vessels of the liver, 4.4 mm and 3.3 mm deep from the capsule, respectively (P < 0.001). Among the changes identified in the vascular tree, small vessel blunting was the most prevalent finding in predicting the presence of severe fibrosis (multiple regression test, t=5.969, P < 0.0001). ROC analysis identified that the presence of at least two pathologic findings in the vascular tree was highly predictive of severe fibrosis (AUC=0.881, sensitivity 86.67%, specificity 89.29%, positive and negative predictive values 8.09 and 0.15, respectively).

CONCLUSION

Our study proves that SMI is superior to other Doppler techniques in detecting the smallest vessels visible to ultrasound. Using this method, it is possible to determine the vascular changes in terms of blunting and tortuosity and thus predict the severity of fibrosis. This method might be a practical alternative to biopsy.