CEUS detection of biliary ischaemia during the first 4 weeks after liver transplantation predicts non-anastomotic biliary stricture

BACKGROUND: Biliary ischaemia is an important factor in the pathogenesis of non-anastomotic biliary stricture (NAS) after liver transplantation (LT). Contrast-enhanced ultrasound (CEUS) can be used to detect biliary ischaemia, but no study has examined the utility of CEUS in predicting NAS. OBJECTIVE: To evaluate whether repeated CEUS as a non-invasive method of biliary ischaemia can identify NAS. METHODS: Consecutive LT patients who underwent CEUS examinations at 1–4 weeks after LT from September 2012 to December 2015 at our institution were included. The CEUS images and clinical data were analysed. RESULTS: Among 116 eligible LT patients, 39 (33.6%) were diagnosed with NAS within 1 year after LT. The patients with NAS had a significantly higher CEUS score at weeks 2–4 (all P < 0.05) and a higher slope of CEUS score progression (0.480 vs –0.044, P < 0.001). The accuracy of CEUS in identifying NAS improved over time after LT, reaching its maximum at week 4, with a sensitivity of 66.7%, a specificity of 87.9%, a positive predictive value (PPV) of 75.9%, a negative predictive value (NPV) of 82.3%, and an accuracy of 80.2%in the full cohort when a CEUS score≥3 was used as the cut-off. Multivariate analysis identified gamma-glutamyl transpeptidase (GGT), alanine transaminase (ALT) and the CEUS score at week 4 as independent predictors of NAS. In the task of identifying NAS, an NAS score combining the above 3 variables at week 4 showed areas under the receiver operating characteristic curve of 0.88 (95%CI, 0.78–0.99) in the estimation group (n = 60) and 0.82 (95%CI, 0.69–0.96) in the validation group (n = 56). An NAS score cut-off of 0.396 identified 87.2%of NAS cases in the estimation group, with a PPV of 93.3%; and 75.0%of NAS cases in the validation group, with a PPV of 58.8%. CONCLUSIONS: CEUS examination during the first 4 weeks is useful in assessing the risk of NAS within 1 year after LT. In particular, an NAS score combining the CEUS score, GGT level, and ALT level at week 4 can be used to accurately predict the risk of NAS in LT patients.

Keywords: Ultrasonography; contrast agent BR1; liver transplantation; microcirculation; ischemia; biliary tract diseases

Non-anastomotic biliary stricture (NAS) is recognized as the most troublesome biliary complication after liver transplantation (LT) [[1]], with a reported incidence of 5–30%[[3], [5]]. It may occur at focal or multiple locations in the biliary tree, and it can produce a wide spectrum of changes ranging from mild non-progressive stenoses to frank biliary necrosis [[4], [6]].

Early manifestations of NAS through symptoms or liver function tests are often unspecific [[7]], and the diagnosis is usually confirmed by endoscopic retrograde cholangiopancreatography (ERCP), percutaneous transhepatic-cholangial drainage (PTCD), or magnetic resonance cholangiopancreatography (MRCP) [[2], [9]]. Nevertheless, these methods are usually based on the evaluation of abnormalities in the biliary tree, when the graft has already shown irreversible morphological changes at a median time of 3–4 months after LT [[4], [10], [12]].

Biliary ischaemia is an important factor in the pathogenesis of NAS [[1]], especially in cases that present within 1 year after LT [[4]]. It is believed that morphological changes, e.g., fibrous stenosis, bile duct necrosis and bilomas, develop predominantly secondary to biliary ischaemia [[13], [15]]. Therefore, early identification of biliary ischaemia theoretically has the potential to help predict NAS and even guide treatment.

Contrast-enhanced ultrasonography (CEUS), a non-invasive clinical methodology for the display of tissue perfusion, has been demonstrated to display perfusion at very fine structural scales, including microvascular invasion of hepatocellular carcinoma (HCC) [[17]] and biliary perfusion [[18]]. In addition, it has reasonably good diagnostic performance, not only for biliary ischaemia in animal experiments [[18]] but also for NAS in patients who have undergone LT [[19]]. However, it remains unknown whether CEUS in the early post-LT period could predict the development of NAS. Therefore, we performed a prospective cohort study to evaluate whether repeated CEUS examination during the first 4 weeks after LT can identify patients at risk of developing NAS.

This study was approved by our institutional review board ([2010] 2–27), and informed consent was obtained from each LT recipient who participated. Consecutive patients who underwent LT at our institution from September 2012 to December 2015 were prospectively considered for the study. We excluded patients who 1) underwent LT due to sclerosing cholangitis or primary sclerosing cholangitis (PSC); 2) had contraindications to CEUS (age < 18 years or severe heart/lung problems); 3) died within 1 week after LT or before confirmation of NAS; 4) were diagnosed with hepatic artery thrombosis before confirmation of NAS; 5) did not undergo ERCP, PTCD, or MRCP to confirm NAS; or 6) received fewer than two CEUS examinations during the 4-week postoperative period due to severe conditions.

The surgical technique has been described previously [[21]]. In brief, organ procurement was performed according to standard techniques using University of Wisconsin preservation fluid or histidine-tryptophan-ketoglutarate solution. All liver transplants were performed with the piggyback technique. There were two options for arterial anastomosis: if both the recipient and donor livers had good arterial condition, end-to-end anastomosis was performed; if the arterial condition of either the recipient or the donor liver was poor, the bypass technique was used, with most donor vessels sourced from the lower limb of the recipient. A duct-to-duct bile duct anastomosis was preferred, and T-tubes were not used during the surgery [[23]].

Cyclosporine A or tacrolimus and prednisone were used for immunosuppression induction after LT. Tacrolimus was started at 0.1–0.15 mg/kg twice daily to achieve a plasma concentration of 8–12 ng/mL. After 3 months, the target plasma concentration was 6–10 ng/mL. Cyclosporine A was added for patients whose target plasma concentration could not be maintained. Acute rejection episodes were confirmed by liver histology and treated with steroids if they were moderate or severe.

Patients underwent repeated biliary CEUS examinations at 1–4 weeks after LT. All CEUS examinations were performed by three radiographers (each having more than 4 years of experience in liver CEUS) using an Acuson Sequoia 512 (Siemens, USA) or an Aplio 500 (Toshiba, Japan) with a low-mechanical-index, real-time, contrast-specific, contrast-enhanced sonographic mode [[19]]. In brief, the long-axis view of the hilar bile duct with the thickest wall on baseline ultrasonography (US) was magnified and selected for CEUS evaluation. The hilar biliary wall and surrounding liver parenchyma were scanned continuously for 6 min after intravenous injection of 1.5 mL of the contrast agent SonoVue (BR1; Bracco, Milan, Italy).

Two radiologists (each having more than 4 years of experience in liver CEUS imaging) independently analysed the images and video recordings while blinded to all clinical information. Discrepant results were reassessed to achieve consensus. With the results of our preliminary studies [[18], [20]] as a reference, we assigned the enhancement pattern of each hilar biliary wall a CEUS score based mainly on the enhancement in the arterial phase (Table 1, Fig. 1), where higher scores meant larger decreases in enhancement.

Table 1 CEUS score according to biliary duct wall enhancement pattern

Note. The enhancement level of the biliary wall is assessed with respect to that of the adjacent liver parenchyma at the same time.

Graph: Fig. 1 Cases of CEUS enhancement pattern of CEUS Scores 1, 4 and 5. White arrows: biliary wall; *: liver parenchyma. Score 1: Biliary wall appeared clearly on B-mode US, and hyper-enhancement on CEUS in the arterial phase comparing with that of the adjacent liver parenchyma and iso-enhancement in the portal venous phase. Score 4: Biliary wall appeared thickened in B-mode US, and hypo-enhancement on CEUS in the arterial phase and the portal venous phase comparing with that of the adjacent liver parenchyma. Score 5: Biliary wall appeared clearly in B-mode US, and non-enhancement on CEUS in the arterial phase and the portal venous phase comparing with that of the adjacent liver parenchyma.

The outcome in our study was whether the transplant recipients developed NAS within 1 year after LT. NAS was defined as any stricture, dilatation, or irregularity of the intra- or extrahepatic bile ducts of the liver graft, either with or without biliary sludge formation, after exclusion of hepatic artery thrombosis by either Doppler US or conventional angiography [[4]]. MRCP (using a Signa Excite II system, GE Healthcare, Waukesha, WI, USA), ERCP (using a TJF-260 system, Olympus, Tokyo, Japan), or PTCD (using an LCV + system, GE Healthcare, Waukesha, WI, USA) was performed to confirm the diagnosis of NAS when patients showed increased gamma-glutamyl transpeptidase (GGT) levels, cholangitis, or jaundice. MRCP was interpreted by two radiologists, and ERCP and PTCD were performed by one radiologist; each of the three radiologists had more than 10 years of experience in biliary imaging and was blinded to the CEUS results.

Potential risk factors for NAS were collected, including donor-related, recipient-related, surgical, and postoperative variables.

Regular follow-up was usually conducted once a week during the first month after discharge, twice a month during the second and third months and then monthly until the end of the first year. Liver function was tested at each follow-up, and US examination was performed every year. Additional follow-up and examinations were added if the patients showed abnormal liver function or had cholangitis or jaundice. The last follow-up was in November 2019.

Statistical analysis was performed with STATA 11 (StataCorp, College Station, TX, USA) and the SPSS 17.0 package (IBM Corp., New York, USA). All reported P values were two-sided, and P values < 0.05 were considered statistically significant. Quantitative variables are expressed as medians (range). Differences between qualitative variables were assessed with a chi-square test or Fisher's exact test. Differences between quantitative variables were analysed with a Mann-Whitney test, Kruskal-Wallis test or t-test.

We estimated the linear slope of the CEUS score for NAS versus non-NAS patients using a longitudinal mixed model for repeated measurements. To analyse the risk factors for NAS and construct a predictive model, patients were randomly divided into two data sets: data from 60 (52%) patients were used to estimate the model parameters, and data from the remaining 56 (48%) were used for validation. Risk factors for NAS were analysed by univariate analysis, and variables with a P value < 0.05 were included in multivariate logistic regression; a backward step-down selection process with the Akaike information criterion was used to determine the independent predictors of NAS. Bootstrapping was used to perform an additional internal validation by generating 10,000 resampling sets with replacement. The diagnostic accuracy of CEUS to identify patients at risk of developing NAS was assessed using the area under the receiver operating characteristic curve (AUROC), and the optimal cut-off was selected based on the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for the identification of NAS.

From September 2012 to December 2015, a total of 204 consecutive patients who underwent LT at our institution were prospectively considered for the study. We excluded patients who 1) underwent LT due to PSC (n = 2); 2) had contraindications to CEUS (n = 10); 3) died within 1 week after LT (n = 14) or before confirmation of NAS (n = 1); 4) were diagnosed with hepatic artery thrombosis before confirmation of NAS (n = 3); 5) did not undergo ERCP, PTCD, or MRCP to confirm NAS (n = 32); and 6) received fewer than two CEUS examinations during the 4-week postoperative period due to the patient's severe conditions (n = 26). Therefore, 116 LT patients were included in our study. The causes of LT were HCC (n = 46), fulminant hepatitis (n = 36), viral hepatitis-related cirrhosis (n = 23), NAS (n = 2), primary biliary cirrhosis (n = 1), alcoholic cirrhosis (n = 4), liver metastases (n = 1), congenital bile duct dilatation (n = 1), Wilson's disease (n = 1), and polycystic liver (n = 1). The baseline characteristics of all patients were summarized in Table 2.

Table 2 Baseline characteristics of the LT patients (validation group, n = 60 and estimation group, n = 56)

Qualitative variables are shown in n (%) and quantitative variables are expressed as medians (range).

To confirm the presence or absence of NAS, 17 patients (14.6%) underwent PTCD, 35 (30.2%) underwent MRCP, and the remaining 64 (55.2%) had constant normal liver function during follow-up, without any sign of biliary complications detected by computed tomography (CT) or US. There were 39 patients (33.6%) confirmed as having NAS within 1 year after LT; 23 among them (59.0%) were included in the estimation group. The median interval between LT and the diagnosis of NAS was 2 months (range, 0.5–13 months). Thirteen patients (33.3%) were confirmed to have NAS by PTCD, and 26 (66.7%) were confirmed to have NAS by MRCP.

The follow-up continued to November 2019. Among 39 patients confirmed as having NAS, 12 (30.8%) died, with a survival time from 0.5 months to 3 years. The causes of death included infection (n = 5), recurrence of HCC (n = 3), NAS (n = 1), acute myocardial infarction (n = 1), multiple organ failure (n = 1), and haemorrhagic shock (n = 1). Four patients (10.3%) underwent secondary LT 2–5 years after the primary LT, with the causes of NAS (n = 3) and infection (n = 1). Five patients (12.8%) were lost to follow-up after the confirmation of NAS, with a follow-up time of 2–6 months. The remaining 18 patients (46.2%) were followed up for 1–7 years. At the last follow-up, 9 of the remaining 18 patients (50.0%) showed normal liver function, while the others showed abnormalities, and no recurrence of HCC was found.

For a variety of reasons associated with the conditions of the patients, 77 patients (66.4%) underwent CEUS examinations at week 1 after LT, 101 (87.1%) at week 2, 95 (81.9%) at week 3, and 91 (78.4%) at week 4. Twenty-five patients (21.6%) underwent two CEUS examinations, 50 (43.1%) underwent 3 examinations, and the remaining 41 (35.3%) underwent 4 examinations. A total of 364 valid CEUS scores were available during the first 4 weeks after LT.

The progression of CEUS scores during the first 4 weeks after LT was different between the NAS and non-NAS groups (Fig. 2). In the non-NAS group (n = 77), the CEUS score did not significantly increase during the first 4 weeks after LT, with a median CEUS score of 1 at weeks 1–4 and a slope of the CEUS score progression of –0.044 (P = 0.484). In contrast, the NAS patients (n = 39) showed a progressive increase over time (P < 0.001), with median CEUS scores of 1, 2, 3 and 3 at weeks 1–4, respectively, and a significantly increased slope of 0.480 (P < 0.001).

Graph: Fig. 2 Progression of CEUS score after LT in non-NAS patients (n = 77) and NAS patients (n = 39) using a mathematical mixed model for repeated measurements. The slope of linear function in NAS patients (0.480) was significantly higher than the slope of non-NAS patients (–0.044, P = 0.001).

The accuracy of the CEUS score to identify NAS improved over time after LT, with AUROCs of 0.60 (95%CI: 0.47–0.72) at week 1, 0.62 (95%CI: 0.51–0.73) at week 2, 0.69 (95%CI: 0.57–0.80) at week 3, and 0.81 (95%CI: 0.71–0.90) at week 4 in all LT patients. When a CEUS score≥3 was used as the cut-off to identify NAS, CEUS had a sensitivity of 66.7%, a specificity of 87.9%, a PPV of 75.9%, an NPV of 82.3%, and an accuracy of 80.2%in the full cohort at week 4. The ability of CEUS to predict NAS in the estimation and validation datasets was shown in Fig. 3 and Table 3.

Graph: Fig. 3 Diagnostic accuracy of CEUS score to predict NAS in the estimation (n = 60) and validation (n = 56) groups at 1, 2, 3 and 4 weeks.

Table 3 Predictive values of CEUS score (at 4 weeks after LT) for predicting early NAS (≤1year) in the estimation and validation groups of LT patients

S: sensitivity, Sp: specificity; PPV: positive predictive value; NPV: negative predictive value; LR+: likelihood ratio positive; LR-: likelihood ratio negative.

Univariate analyses were performed in the estimation group (n = 60) to identify the variables associated with NAS (Table 4). Five variables were identified at 4 weeks after LT: aspartate aminotransferase, alanine transaminase (ALT), GGT, alkaline phosphatase, and CEUS scores. The multivariate analysis showed that only GGT, ALT and CEUS scores were independent predictors of NAS at 4 weeks after LT. An NAS score was established to identify NAS: NAS score = –4.783 + 0.469×GGT + 0.449×ALT + 0.813×CEUS. Both GGT and ALT were defined as the fold change compared to the upper limit of the normal range at week 4, and CEUS was defined as the CEUS score at week 4. The NAS score achieved an AUROC of 0.88 (95%CI, 0.78–0.99) in the estimation group and 0.82 (95%CI, 0.69–0.96) in the validation group (Fig. 4). The results of the internal bootstrap validation showed that the NAS score had a good predictive value for NAS, with an AUROC of 0.87 (95%CI, 0.77–0.97). An NAS score cut-off of 0.396 identified 87.2%of NAS cases in the estimation group, with a PPV of 93.3%; and 75.0%of NAS cases in the validation group, with a PPV of 58.8%(Fig. 4 and Table 5).

Table 4 Univariate analysis in the estimation group (n = 60) to identify the variables associated with NAS within 1 year

Qualitative variables are shown in n (%) and quantitative variables are expressed as medians (range). ULNL: upper limit of normal level.

Graph: Fig. 4 Diagnostic accuracy of NAS score at 4 weeks to identify patients to develop NAS in the estimation(n = 60) and validation (n = 56) groups.

Table 5 Predictive values of the NAS score obtained from the estimation and validation groups to identify NAS within 1 year

S: sensitivity, Sp: specificity, PPV: positive predictive value, NPV: negative predictive value; LR+: likelihood ratio positive, LR-: likelihood ratio negative. NAS score = –4.783 + 0.469×GGT + 0.449×ALT + 0.813×CEUS. Both GGT and ALT were defined as the folds of the upper limits of the normal lever at week 4, and CEUS was defined as the CEUS score at week 4.

We conducted this longitudinal study to evaluate whether repeated biliary CEUS examinations during the first 4 weeks after LT could identify patients at risk for the development of NAS. The results showed that CEUS is useful for predicting whether patients will develop NAS within 1 year after LT.

NAS can occur in different periods after LT and can be categorized into early (< 1 year) and late (>  1 year) types mainly based on the radiological presentation [[4]]. This study is focused on early presentation for the following reasons: 1) NAS within 1 year has been considered a predictor of radiological progression of biliary abnormality and severe adverse outcomes, including death and re-transplantation due to NAS [[4], [25], [27], [29]]; therefore, the detection of this early type is important in clinical practice. 2) Early NAS is more frequently associated with ischaemic factors [[4]] and usually develops secondary to biliary ischaemia [[1]]. CEUS, an imaging method that has been demonstrated to represent biliary perfusion and has reasonably good diagnostic ability for biliary ischaemia [[18]], can be applied as an early warning system for early NAS.

Considering that 85%of NAS cases showed biliary lesions around the common bile duct [[4]], the hilar biliary wall was selected as the target of CEUS. In our current study, CEUS could predict this type of NAS as early as the first 4 weeks after LT. Patients with NAS showed a progression of increasing biliary ischaemia of the hilar biliary wall during the first 4 weeks, while non-NAS patients showed constant and steady normal biliary perfusion. The ability of CEUS to predict NAS increased over time, showing moderate accuracy at weeks 1–3 (AUROC, 0.60–0.69) and good accuracy at week 4 (AUROC, 0.81).

Four weeks appeared to be an important time point to predict NAS, not only because the accuracy of CEUS was sufficient (showing a sensitivity of 66.7%, a specificity of 87.9%, a PPV of 75.9%, an NPV of 82.3%, and an accuracy of 80.2%in the full cohort when using a CEUS score≥3 as the cut-off) but also because GGT and ALT levels were independent risk factors. Que et al. showed that a GGT level more than 5 times the upper limit of normal was diagnostic for biliary stricture [[30]]. When an NAS score was established by combining all 3 variables, it achieved an AUROC of 0.82–0.88 and identified 75.0%–87.2%of NAS cases with a PPV of 58.8%–93.3%.

We acknowledge several limitations for this study. First, NAS can present with biliary lesions in different regions of the biliary tree, but we chose only the hilar biliary duct for evaluation through CEUS. In addition to the fact that most cases show abnormalities in the common bile duct, as mentioned above [[4]], the visibility of the biliary wall on CEUS is a precondition. The hilar biliary duct can be displayed more easily than intrahepatic ducts, especially in the condition that intrahepatic ducts do not show dilation. Second, CEUS examinations beyond the first 4 weeks after LT were not included in our current study, so whether a longer period could achieve better prediction of NAS remains unknown. In our institute, most LT patients are discharged at 4 weeks unless their recovery is unsatisfactory or their condition deteriorates. Although these patients who remain hospitalized after 4 weeks still undergo weekly CEUS, it's difficult to maintain a uniform routine and ensure that other patients perform unnecessary follow-up. Third, although the diagnosis of NAS in our study was confirmed by MRCP and PTCD, most non-NAS patients were diagnosed by follow-up showing constant normal liver function and no biliary complications detected by CT or US. It is almost inevitable that the number of NAS cases was underestimated. Fourth, as a study of predictive model, this study lacked external testing. Considering that biliary CEUS examination is highly dependent on the expertise of the operator, we believe that special training to learn the standard scanning approach and imaging analysis methods is essential to avoid incorrect evaluation. Future studies with larger sample sizes, stricter diagnostic criteria, longer follow-up periods and external validation are needed.

In conclusion, CEUS examination during the first 4 weeks allows the prediction of NAS within 1 year after LT. An NAS score combining the CEUS score, GGT level, and ALT level at week 4 could be used to accurately predict the risk of NAS in LT patients.