The value of periportal hyperintensity sign from gadobenate dimeglumine-enhanced hepatobiliary phase MRI for predicting clinical outcomes in patients with decompensated cirrhosis

Objectives To determine the value of periportal hyperintensity sign from gadobenate dimeglumine (Gd-BOPTA)-enhanced hepatobiliary phase (HBP) magnetic resonance imaging (MRI) for predicting clinical outcomes in patients with decompensated cirrhosis. Methods A total of 199 cirrhotic patients who underwent Gd-BOPTA-enhanced MRI were divided into control group (n  =  56) and decompensated cirrhosis group (n  =  143). The presence of periportal hyperintensity sign on HBP MRI was recorded. The Cox regression model was used to investigate the association between periportal hyperintensity sign and clinical outcomes. Results There was a significant difference in the frequency of periportal hyperintensity sign on HBP between compensated and decompensated cirrhotic patients (p  <  0.05). After a median follow-up of 29.0 months (range, 1.0–90.0 months), nine out of 143 patients (6.2%) with decompensated cirrhosis died. Periportal hyperintensity sign on HBP MRI was a significant risk factor for death (hazard ratio (HR)  =  23.677; 95% confidence interval (CI)  =  4.759–117.788; p  =  0.0001), with an area under the curve (AUC) of 0.844 (95% CI  =  0.774–0.899). Thirty patients (20.9%) developed further decompensation. Periportal hyperintensity sign on HBP MRI was also a significant risk factor for further decompensation (HR  =  2.594; 95% CI  =  1.140–5.903; p  =  0.023). Conclusions Periportal hyperintensity sign from Gd-BOPTA-enhanced HBP MRI is valuable for predicting clinical outcomes in patients with decompensated cirrhosis. Critical relevance statement Periportal hyperintensity sign from gadobenate dimeglumine-enhanced hepatobiliary phase magnetic resonance imaging is a new noninvasive method to predict clinical outcomes in patients with decompensated cirrhosis. Key points • There was a significant difference in the frequency of periportal hyperintensity sign on HBP between compensated and decompensated cirrhotic patients. • Periportal hyperintensity sign on the hepatobiliary phase was a significant risk factor for death in patients with decompensated cirrhosis. • Periportal hyperintensity sign on the hepatobiliary phase was a significant risk factor for further decompensation in patients with decompensated cirrhosis. Graphical Abstract


Introduction
Decompensated cirrhosis is determined by the presence of ascites, hepatic encephalopathy, variceal bleeding, or other events [1,2].Following the first appearance of any of these symptoms, the disease usually progresses rapidly, and the patients are at a high risk of mortality [3].The 5-year survival rate in decompensated cirrhotic patients is about 14 to 35%, which is shorter than that in compensated cirrhotic patients [2,4].Besides a significant mortality rate, decompensated cirrhosis is associated with considerable cost of treatment and patient suffering [5].
The Child-Pugh and model for end-stage liver disease (MELD) scores are well-recognized prognostic models of decompensated cirrhosis [6][7][8].The advantage of these two scores is that they can be readily determined based on clinical and laboratory information [9].Despite advances in the clinical outcomes of these methods for non-invasive liver assessment, their drawbacks have limited their extensive use.For example, there is no evidence that the cut-off levels for the Child-Pugh score are optimal [10], and the levels of creatinine and bilirubin included in the MELD score are easily altered by therapeutic interventions, hemolysis, or sepsis [11].Therefore, new tools are needed to evaluate the prognosis in patients with decompensated cirrhosis.
Gadobenate dimeglumine (Gd-BOPTA) is a commonly used hepatocyte-specific contrast agent, which is utilized to characterize focal liver lesions [12].Our previous studies have determined that Gd-BOPTAenhanced hepatobiliary phase (HBP) magnetic resonance imaging (MRI) is a good approach for assessing liver function and clinical outcomes in cirrhotic patients [13,14].Gd-BOPTA-enhanced portal vein imaging on HBP can predict the prognosis in chronic liver disease patients [15].Periportal hyperintensity sign on HBP is defined as enhancement manifested as a periportal ring that surrounds the portal veins and has been confirmed to be a useful indicator for predicting advanced liver fibrosis [16,17].However, no study has reported on the value of periportal hyperintensity sign on HBP for predicting clinical outcomes in patients with decompensated cirrhosis.
Therefore, the present study estimated the value of periportal hyperintensity sign from Gd-BOPTA-enhanced HBP MRI for predicting clinical outcomes in patients with decompensated cirrhosis.

Patients
Patients with decompensated cirrhosis who underwent a Gd-BOPTA-enhanced liver MRI examination between November 2012 and December 2021 were retrospectively enrolled in this study.Cirrhosis was defined based on pathologic or clinical evidence, including nodularity/ splenomegaly on liver imaging and/or thrombocytopenia [1].Hepatic decompensation was identified by the presence of ascites, variceal bleeding, or hepatic encephalopathy (HE) [18,19].Exclusion criteria were as follows: the presence of malignancy, acute hepatitis, surgery involving the biliary tract, insufficient imaging quality, missing biochemical parameters, loss to follow-up, or renal impairment.Patients who underwent liver transplant surgery were withdrawn from the study.Overall, 56 compensated cirrhotic patients were included in the control group and 143 patients (91 men and 52 women) were included in the decompensated cirrhosis group.Decompensated cirrhotic patients were categorized as Child-Pugh A (n = 23), Child-Pugh B (n = 67), and Child-Pugh C (n = 53) in accordance with their clinical manifestations.
Serum markers tested within 2 weeks of MRI were obtained using electronic medical records [13].Outcomes, including death and further decompensation, were followed up in all patients.Follow-up time was the period from the first MRI to the time of the event or the end of the follow-up.Further decompensation was defined as recurrent acute variceal bleeding, refractory ascites, recurrent HE, spontaneous bacterial peritonitis, hepatorenal syndrome, acute-on-chronic liver failure (ACLF), and liver-related death [20,21].

MRI
All patients were examined on a 3-T MR scanner (MAG-NETOM Verio or Prisma, Siemens).The liver protocols included pre-and post-contrast T1-weighted sequences.Dynamic sequences were obtained in 20 s (arterial phase), 50 s (late arterial phase), 80 s (portal venous phase), and 90 min (HBP) after contrast media injection at a concentration of 0.05 mmol/kg (0.1 mL/kg) of body weight followed by a 20-mL saline flush.Image parameters were as follows: repetition time, 3.31 ms; echo time, 1.3 ms; slice thickness, 3 mm; number of partitions, 72; matrix, 182 × 320; flip angle, 9°; and acquisition time, 17 s.

Imaging analysis
Two radiologists (observers 1 and 2, with 10 and 13 years of experience, respectively), blinded to patients' clinical history, independently reviewed all images [13].Disagreements were resolved by discussion and achieving consensus.The presence of periportal hyperintensity sign on HBP was recorded.Periportal hyperintensity sign was defined as a periportal ring or tramline enhancement around the intrahepatic portal veins present within several (more than one) hepatic segments [16,22].

Statistical analysis
Student's t test or Mann-Whitney U test was used for the comparison of the two groups.Interobserver agreement of categorical variables was estimated using Cohen's kappa (κ) statistics.The Cox regression model was generated to identify factors related to death and further decompensation in patients with decompensated cirrhosis.Performance was assessed using the area under the curve (AUC).Then, the cumulative incidences of death and further decompensation were calculated using the Kaplan-Meier method.p < 0.05 was considered significant.SPSS Statistics (version 25.0, IBM) and MedCalc (version 15.6.1,MedCalc Software) were used for all statistical analysis.
Periportal hyperintensity sign was observed in 21 patients (10.5%) on HBP.The interobserver agreement of the presence of periportal hyperintensity sign on HBP was excellent (κ = 0.884 [95% CI = 0.772, 0.995]).There was a significant difference in the frequency of periportal hyperintensity sign on HBP between compensated and decompensated cirrhotic patients (p < 0.05) (Table 1, Fig. 1).The MELD score and serum liver function parameters, including AST, ALT, and total bilirubin, were significantly higher in patients with periportal hyperintensity sign on HBP than in those without the sign (p < 0.05) (Table 2).

Discussion
In the present study, there was a significant difference in the frequency of periportal hyperintensity sign on HBP between compensated and decompensated cirrhotic patients.Moreover, periportal hyperintensity sign from Gd-BOPTA-enhanced HBP MRI was a significant risk factor for death and further decompensation in patients with decompensated cirrhosis.
The timing of liver uptake and excretion of Gd-BOPTA begins at 40 min after injection and can last until 120 min at the hepatobiliary phase.In clinical practice, it is widely accepted that 90 min was a feasible liver uptake and excretion time point [13,23].Our study found that there was a significant difference in the frequency of periportal hyperintensity sign on HBP MRI between the control group and decompensated cirrhosis group, suggesting that periportal hyperintensity sign on HBP MRI may distinguish decompensated cirrhotic patients from compensated cirrhotic patients.The reason might be that patients with decompensated cirrhosis have poor liver function, resulting in a poor enhancement of liver parenchyma and a relative increase of signal intensity around the portal vein [17,24].Ascites is the most common complication in patients with decompensated cirrhosis [25].As stated by Ciolina et al. [26], pleural and peritoneal fluids appear hyper/isointense on HBP MRI in most patients after Gd-BOPTA injection, while fluids remain hypointense on HBP MRI after gadoxetate disodium (Gd-EOB-DTPA) injection.
As the risk of death in decompensated cirrhotic patients is about four to five times higher than that in compensated cirrhotic patients [27], early identification and treatment are crucial for patients who are at higher risk of death.In the present study, periportal hyperintensity sign on HBP MRI was a significant risk factor for death in patients with decompensated cirrhosis.
Several studies about periportal hyperintensity on HBP MRI have been done.Lampichler et al. [22] found that periportal hyperintensity sign on HBP MRI may correspond to active inflammation or periportal fibrosis.Zheng et al. [28] demonstrated that periportal hyperintensity sign on HBP MRI can be useful for predicting advanced liver fibrosis.As liver fibrosis accumulates and inflammation activity increases, liver function deteriorates rapidly [29].The fact that liver function determines patient survival could have accounted for our finding [6,30].The identification of patients with periportal hyperintensity sign would be invaluable in deciding when and in which patients to intensify medical care.It could also aid in selecting candidates for liver transplantation, thereby facilitating focused resource allocation by identifying those at high risk of death [31].In addition, the MELD score was also a significant risk factor for death in patients with decompensated cirrhosis in this study.The MELD score may play an important role in evaluating clinical outcomes of patients with decompensated cirrhosis, which is similar to previous studies [6,25].
The present study also showed that periportal hyperintensity sign on HBP MRI was a significant risk factor for further decompensation in patients with decompensated cirrhosis.These results may be explained by the persistent inflammatory state in these patients, which may easily trigger further decompensation [25,32].Therefore, periportal hyperintensity sign on HBP MRI may be useful for clinicians to pursue relevant interventions in order to prevent further decompensation and to stabilize the disease progression.Additionally, the finding that the MELD score was associated with further decompensation is consistent with that described by a previous study by Zanetto et al. [21].
There were several limitations in the study.First, this was a retrospective, single-center study, which may introduce inherent selection bias.Second, clinical records were the principal source of information on cirrhosis diagnosis and not all cases were histopathologically confirmed [17].Third, the sample size of patients with periportal hyperintensity sign was small [24].Therefore, further clinical trials in larger populations are necessary in order to validate the present study findings.
In conclusion, periportal hyperintensity sign from Gd-BOPTA-enhanced HBP MRI is valuable for predicting clinical outcomes in patients with decompensated cirrhosis.This information might be helpful for clinicians to guide treatment and eventually improve clinical outcomes.

Fig. 1 2
Fig.1The pre-contrast T1-weighted image (a) and HBP image (b) were selected from a 57-year-old compensated cirrhotic patient.Periportal hyperintensity on HBP is not observed in this patient.The pre-contrast T1-weighted image (c) and HBP image (d) were selected from a 48-year-old decompensated cirrhotic patient.Periportal hyperintensity on HBP is observed (black arrow) in this patient.HBP, hepatobiliary phase

Fig. 2
Fig. 2 Kaplan-Meier curves for patients with decompensated cirrhosis.a Cumulative incidence of death in patients with periportal hyperintensity sign on HBP compared to those without the sign.b Cumulative incidence of further decompensation in patients with periportal hyperintensity sign on HBP compared to those without the sign.HBP, hepatobiliary phase

Table 1
Comparison of baseline characteristics in control and decompensated cirrhosis groupsData are presented as median (interquartile range) or data (percentage) PT prothrombin time, INR international normalized ratio, MELD model for end-stage liver disease a Data are means ± standard deviation with ranges in parentheses

Table 3
Cox survival analysis for predicting death in patients with decompensated cirrhosisHR hazard ratio, CI confidence interval, PT prothrombin time, INR international normalized ratio, MELD model for end-stage liver disease

Table 4
Cox survival analysis for predicting further decompensation in patients with decompensated cirrhosis HR hazard ratio, CI confidence interval, PT prothrombin time, INR international normalized ratio, MELD model for end-stage liver disease