Contrast-Enhanced Ultrasonography-Based Hepatic Perfusion for Early Prediction of Prognosis in Acute Liver Failure

Abstract

Background and aims: Acute liver failure (ALF) is a rare but dramatic clinical syndrome characterized by massive hepatic necrosis leading to multiorgan failure. It is difficult to predict the outcomes in patients with ALF using existing prognostic models. We aimed to analyze hepatic perfusion using contrast-enhanced ultrasound and Doppler ultrasound in patients with ALF and investigate its utility as a prognostic biomarker.

Approach and results: In this prospective observational study, 208 patients with acute liver injury/ALF were enrolled from 2015 to 2019. We evaluated 50 consecutive patients with ALF with Doppler ultrasound and contrast-enhanced ultrasound performed on admission. The cases were divided into the following two groups: survivors (recovered without surgical intervention) and nonsurvivors (died of ALF or underwent liver transplantation). The time to peak and peak intensity of hepatic artery, portal vein, hepatic vein, and liver parenchyma were calculated using the time-intensity curve analysis. The hepatic artery (HA) resistive index was calculated using the fast Fourier transform analysis of Doppler ultrasound. The time interval (TI) between the time to peak of HA and liver parenchyma (LP) was significantly shorter in the nonsurvivors than in the survivors (P < 0.0001). The area under the receiver operating curve values for TI (HA, LP), Japanese scoring system, HE prediction model, Model for End-Stage Liver Disease score, and King's College Hospital criteria for the prediction of poor prognosis were 0.953, 0.914, 0.861, 0.816, and 0.731, respectively. The most appropriate cutoff value of TI (HA, LP) was 6.897 seconds; the sensitivity, specificity, positive and negative predictive values were 94.4%, 90.6%, 85.0%, and 96.7%, respectively.

Conclusions: TI (HA, LP) accurately predicts the outcome in patients with ALF and may be useful in clinical decision making.

FIG. 1

Flow chart of eligible patients with ALF.

FIG. 2

Intensities of the HA, PV, HV, and LP were measured by setting circular ROIs using ImageJ software. (A) The ROIs were set at a depth of 6‐8 cm (±3 cm from the focus point) from the surface. (B) Schematic TIC describing events after the bolus injection. The TTP indicates the duration from the first appearance of the contrast agent in the HA until maximum enhancement was reached. The PI indicates the maximum enhancement after subtracting the baseline intensity. The TI indicates the duration between two curves of TTP. The transit time (TT) describes the period of arrival time between two curves. TIC of the liver in a survivor (C) and nonsurvivor (D). The orange line is the signal intensity of the HA; the green line is the signal intensity of the PV; the blue line is the signal intensity of the HV; and the gray line is the signal intensity of the LP. In the nonsurvivor, the slope of the HA and LP are steeper, and the TTP is short compared with that observed in the survivor. CEUS perfusion imaging of the liver in a survivor (E) and a nonsurvivor (F). In the liver of the nonsurvivor, the HA and LP were enhanced steeply, whereas the PV was enhanced slowly.

FIG. 3

TI (HA, LP) for all of the patients in each of the three categories: patients recovered with intensive therapies, including artificial liver support (recovered, n = 32 [8.31 (7.27‐10.12)]; survived with LT, n = 6 [5.77 (5.38‐6.29)]) and died without LT (died, n = 12 [4.73 (3.35‐5.85)]). There were significant differences between patients who “recovered” versus “LT” (P < 0.0001), and “recovered” versus “those who died without LT” (P = 0.002), but not between “LT” and “those who died” (P = 0.389).

FIG. 4

Serial changes in the TI (HA, LP) of survivors (A), nonsurvivors (B), and ΔTI (HA, LP) (C). In survivors, the median TI (HA, LP) extended from 9.35 (7.17, 12.27) seconds to 10.01 (8.07, 12.62) seconds on day 7 (P < 0.0001). In contrast, in nonsurvivors, it was reduced from 5.19 (3.63, 5.75) seconds to 4.80 (3.23‐5.48) seconds. The ΔTI (HA, LP) was +0.51 (0.12, 1.75) in survivors and −0.12 (−0.59, −0.06) in nonsurvivors.