To the best of our knowledge, this is the first large cohort study on children with CHDs to demonstrate the potential of liver native T1 mapping. We also investigated 75 patients with Fontan circulation among 113 patients who underwent single ventricle repair.
FALD is increasingly recognized as a common complication in patients after Fontan operations. Increased CVP and chronic low cardiac output after the Fontan operation cause hepatic congestion and tissue hypoxia, which may instigate fibro genic pathways in the liver [13]. FALD can range from benign nodular hyperplasia to liver cirrhosis and failure [6], furthermore, it is often accompanied by hepatocellular carcinoma [1–5, 13–16]. At present, multiple structural and functional liver abnormalities have been identified in patients after Fontan operations, but it is often unclear whether they are incidental or pathological [2].
Although no healthy control was included in this study, the mean LT1 of healthy controls measured by 1.5-T MRI at our institution was 552 ± 42 ms (n = 23, mean age: 7.2 ± 4.9 years) [17]. Leah et al. reported that the median LT1 in controls measured by 1.5-T MRI, was 569 ± 39 ms (range, 529–646; n = 12; mean age, 12.2 ± 3.1 years), similar to our control data. In this report, there was no significant association between age and LT1 [18], but this was not observed in our study, similar to our previous report [17]. Compared with these control data, the LT1 of the BVR and bidirectional Glenn circulation groups were almost the same as the control data, while the LT1 of the Fontan circulation group was higher in our study. This is consistent with the study by Shiina et al., which reported that, in adults, the LT1 of patients with Fontan circulation was significantly higher than that of the controls and patients with tetralogy of Fallot [5]. Beigh et al. reported that the LT1 was increased in single ventricular patients both before and after Fontan operations (683 ± 82 ms and 727 ± 49 ms, respectively) in comparison to TGA patients (587 ± 58 ms, P < 0.001) [19].
The LT1 of all patients with CHD had a significant positive correlation with CVP in addition to ALT and the Fib-4 index (Table 2). According to these results, LT1 is influenced by many factors, including liver fibrosis markers and hemodynamic congestion data represented by CVP. Additionally, the T1 relaxometry time does not unequivocally distinguish fibrosis, congestion, or inflammation [5]. In previous reports, high CVP and severe atrioventricular valve regurgitation in the early postoperative period were independent predictors of liver cirrhosis and hepatocellular carcinoma [16]. Therefore, LT1 may be an alternative indicator to CVP and a predictor of liver cirrhosis and hepatocellular carcinoma owing to the strong correlation between CVP and LT1.
The LT1 of Fontan patients was correlated with CVP and showed a significant positive correlation with the Fib-4 index. Additionally, the LT1 was weakly correlated with the IVC return flow of Fontan patients. LT1 might represent congestion because both CVP and IVC return flow are positively correlated with LT1 (Table 3). The difference between Fontan patients and all other patients is that the correlation between IVC return flow and LT1 was observed only in Fontan patients. This may be explained by the hemodynamic characteristics of Fontan circulation, where IVC return drains to the pulmonary artery without passing through the atrium. In a previous study, an increased LT1—both before and after the Fontan stages—suggested that intrinsic liver abnormalities occur due to a combination of edema from passive venous congestion early during single ventricle palliation, and that LT1 may be an early marker of liver disease [19].
In this study, we showed that laboratory data such as ALT and the Fib-4 index might have some clinical implications (Table 2, 3). Strong correlations between LT1 and ALT and the Fib-4 index were observed in the overall population. Additionally, a significant correlation between LT1 and Fib-4 index was observed in Fontan patients. In a previous study, Kogiso et al. reported that the hyaluronic acid level and Fib-4 index were significantly higher in FALD-HCC patients, and that the t-bil level and MELD XI score tended to be higher in FALD-HCC patients [16]. Shiina et al. reported that MELD XI and ALBI scores were significantly correlated with LT1 [5]; however, liver biopsy was not performed in the present study. According to the results of our study and previous studies, in daily practice, it may be more appropriate to evaluate FALD using LT1 MRI and blood markers of liver fibrosis in addition to ultrasonography.
In a bile duct ligation mouse model, the T1 value was not correlated, but the T2 value, or the difference between pre-enhanced T1 and post-enhanced T1 values, was significantly correlated with the hepatic fibrosis rate [20]. Another study reported that the T1 mapping-based hepatocyte fraction method was more accurate than the reduction rate for determining the stage of liver fibrosis, suggesting that the calculation of hepatocyte fractions is a superior method for non-invasive liver fibrosis staging [21]. The methods reported in the two aforementioned papers require contrast medium, which was not used in the present study. While T1 mapping of the liver can be acquired relatively easily and generated inline on the same scanner and at the same time as cardiac MRI, the accuracy of T1 mapping needs to be further validated with liver biopsy and other imaging studies.