A review of the use of transient elastography in the assessment of fibrosis and steatosis in the post–liver transplant patient

Liver biopsy is considered the gold standard method for diagnosing and staging liver disease, particularly in the post‐liver transplant setting. Given the invasive nature of biopsy, alternate means for accurately assessing liver fibrosis and steatosis are preferred especially as the number of patients with fatty liver disease is increasing. Transient elastography has been validated as a useful tool for evaluation of liver fibrosis, as has controlled attenuation parameter index as a tool for assessing steatosis. It is a non‐invasive, rapid, and highly reproducible approach to demonstrate the presence of fibrosis among non‐transplant patients with chronic liver disease of various etiologies. However, it has not yet found wide acceptance in liver transplant recipients. There are few published studies evaluating the merits and applicability of transient elastography to assess allografts after liver transplantation. We review the published data on the use of transient elastography with concurrent controlled attenuation parameter in liver transplant recipients and recommend its greater use to follow allograft function over time.


| INTRODUC TI ON
Early identification of fibrosis recurrence is essential in the management of liver transplant (LT) recipients as it permits appropriate interventions and management modifications. Liver biopsy is the established method for assessing hepatic fibrosis, particularly in the post-transplant setting. LT recipients require longitudinal follow-up to assess allograft fibrosis. Thus, it is impractical to use liver biopsy as a tool for serial monitoring in the post-transplant setting due to its invasiveness, cost, and potential for sampling error. 1

| ME THODS
We performed a Medline search using the key terms transient elastography: liver transplantation, Fibroscan ® , and controlled attenuation parameter (CAP) going back for the last 10 years. We sought publications reviewing the specific application of transient elastography in assessing liver disease post-LT.

| Transient elastography
The gold standard to assess liver allograft function is liver biopsy.
In the early days of LT, protocol liver biopsy was standard practice mostly in patients with hepatitis C (HCV). Much of the decisionmaking revolving around HCV treatment was predicated on the presence of moderate to advanced fibrosis or rapid progression of fibrosis. Currently, however, most LT centers no longer perform protocol liver biopsies since the advent of direct-acting antiviral (DAA) drug therapy for HCV infection. Protocol liver biopsies helped show that mild acute cellular rejection (ACR) could be present in the setting of normal liver chemistry tests and not be clinically relevant, and that often long-term survivors of LT had abnormal liver histology. 7 Protocol liver biopsies are much less frequently performed in non-HCV patients.
TE was first introduced as a non-invasive tool to assess the degree of LS in patients with chronic liver disease in 2003. 8 Since then, it has found wide-ranging acceptance for staging disease prior to LT.
Several studies have demonstrated good reproducibility of TE results in a wide array of liver diseases, with an intra-observer agreement of 96%-96% and an inter-observer agreement of 89%-98%. [9][10][11][12] The system is equipped with an ultrasound transducer that transmits vibrations of mild amplitude and low frequency to tissues. It involves assessing the velocity rate of propagation of an elastic shear wave generated by mechanical impulse from an applied transducer. This rate is determined by cohesiveness of the liver. The degree of cohesiveness increases in direct proportion to the degree of fibrosis.
A measurement depth of 25-45 mm is considered reliable and 10 successful measurements are required. The velocity measurements are computer-processed and an average of the 10 measurements is reported in kilopascal (kPa). The addition of CAP, introduced in 2010, 13 has allowed the grading of steatosis by assessing the degree of ultrasound attenuation due to hepatic fat using the TE probe simultaneously with LS measurements. CAP has been an important addition to the applicability of TE, but important caveats apply.
Whereas an elevated LS is uniformly associated with worse clinical outcomes in patients with chronic liver disease, such as a higher risk of decompensating events, 14 HCC development, 15,16 and decreased survival, 17 steatosis as measured by CAP has not been associated with the development of liver cancer or decompensating events. 18,19 Thus, steatosis is an unreliable marker for disease progression and does not portend a poor clinical course, unlike elevations in LS.
Overall, TE is a novel, adjunctive modality that may replace the gold standard, liver biopsy, when clinically warranted. By comparison, ultrasonography, frequently used as a non-invasive method to assess for liver cirrhosis, has been demonstrated to have a low sensitivity, particularly in diagnosing non-nodular forms of cirrhosis. 20,21 The recent European Association for the Study of the Liver guidelines on non-invasive diagnosis of liver fibrosis suggest that non-invasive methods could replace liver biopsy when used in combination with the pre-transplant setting. 22 In 2013, the Food and Drug Administration approved the use of TE in both adults and children with liver disease. In the post-LT population, however, data regarding the use of TE are sparse.

| Graft selection
The optimal selection of liver graft is essential to the success of the transplant. 23 Determination of graft steatosis is a key segment of the selection process. The precise estimation of graft steatosis is challenging, and although a liver biopsy is most commonly performed to evaluate this, it remains prone to sampling error. [24][25][26] CAP measured using TE is a potentially rapid, non-invasive, reliable tool to pre-operatively assess and quantify graft steatosis. 13

| Live donors
The application of CAP with LS has subsequently been evaluated in the living liver donor population. Kim et al evaluated a group of 79 living donors by TE using LS and CAP, computed tomography (CT), and magnetic resonance imaging and compared these non-invasive modalities to intra-operative frozen sections. The authors found that CT liver-spleen attenuation differences and the measurements obtained via TE correlated with hepatic steatosis present on biopsy, and a weighted combination of these two modalities yielded a higher predictive capability of hepatic steatosis, with an AUROC of 86.6% (P = .001). 35 In another study by Hong and colleagues, 55 potential live donors were evaluated by CAP and the results were compared with biopsy. CAP had an AUROC of 78% for detecting at least S1 steatosis and 88% for at least S2. 36 Most recently, Yen et al 37 compared TE with intra-operative liver biopsy in 54 living donors. Using a cutoff of 257 dB/m, the authors were able to achieve a 100% sensitivity for hepatic steatosis, though only 7 of the 12 patients with CAP values >257 had biopsy-proven steatosis. Although these results are promising, this study was performed in an East Asian population with an average BMI of 24 kg/m 2 potentially limiting its applicability to a Western population. Most potential donors undergo MRI for assessment of vascular and biliary anatomy. As part of this evaluation, elastography and assessment for steatosis can be performed.
Both have similar sensitivities as TE in quantifying fibrosis and steatosis, respectively. Exclusion of other important liver histology in the setting of abnormal liver tests is not possible with TE or CAP, so liver biopsy appears warranted in this setting, especially if baseline MRI does not suggest steatosis or fibrosis.

| Acute cellular rejection
The inflammatory cascade driving acute cellular rejection (ACR), as with any inflammatory process, might be expected to increase LS. 38 Overall, data are limited regarding this population; a recent systematic review found only 3 studies comprising 33 total patients with ACR who were evaluated with TE. 40 Further, although an increased LS may in fact suggest underlying ACR, it does not add much to a clinician's decision-making process in the acute setting, that is, whether a liver biopsy is necessary or whether an empiric increase in immunosuppression is warranted. Liver histology would be invaluable in assessing the degree of ACR and whether there was evidence of chronic ductopenic rejection, plasma cell hepatitis rejection, or antibody-mediated rejection that might influence the timing and duration of therapy. 41 Thus, we feel that TE plays little role in the diagnosis and management of ACR.

| Recurrent fibrosis post-transplant
Progression of graft fibrosis is significantly accelerated in immu- Finally, in patients transplanted for HCV who were treated post-LT and achieved SVR, LS has been shown to accurately assess the presence of advanced fibrosis and clinically significant portal hypertension. 53 LS measurements to rule out and rule in advanced fibrosis were 10.6 and 14 kPa, respectively, with an AUROC of 0.902. LS cutoffs to rule out and rule in clinically significant portal hypertension were 11.3 kPa and 23 kPA, respectively, with an AUROC of 0.888.
Given that patients with lower hepatic venous pressure gradient have been shown to have a lower incidence of clinical decompensation, 54 LS may be an important prognostic tool in the post-transplant population who have achieved an SVR.

| Non-alcoholic fatty liver disease
NAFLD affects about 24% of the world population and is becoming the most common etiology of end-stage liver disease in the industrial world. 55,56 Patients on immunosuppression are at risk of developing or exacerbating risk factors that lead to progression of NAFLD, such as weight gain, diabetes, hyperlipidemia, and hypertension. This places these patients at risk for the development of hepatic graft steatosis and recurrence of non-alcoholic steatohepatitis (NASH) post-LT. [57][58][59] Additionally, de novo NAFLD in LT patients not transplanted due to NASH is being increasingly observed. One systematic review reported a 26% pooled weighted prevalence of biopsy-proven de novo NAFLD in a LT patient population. 60 In a single-center study of patients undergoing LT for NASH, Malik et al identified recurrent NAFLD, NASH, and stage 2 fibrosis in 70%, 25%, and 18% of patients, respectively, with a mean follow-up time of 18 months. 61 In another study, recurrent steatohepatitis was noted in one third of patients transplanted for NASH. 62 Accuracy of TE in diagnosing fatty liver disease in non-transplant patients has been established and validated. 63 The utility of TE in detecting NAFLD post-LT has not yet been corroborated. One concern regarding the use of TE is that accuracy is affected by obesity, resulting in 15% of readings being unreliable. 64 Given obesity is a common comorbidity in the NAFLD population, the utility of TE could be called into question. The XL probe was designed to improve the usefulness of TE in the obese patient, allowing for measurements of depths ≥35 mm below the skin, compared with 25 mm for the M probe. 65 In one study comparing use of TE using the XL probe and the standard probe (M) to liver biopsy, the XL probe gave reliable results 23% more often than the M probe. 64 With the addition of CAP, a program designed to measure hepatic steatosis, the accuracy of the XL probe increases even more. 66 Karlas et al 67 evaluated post-LT steatosis by TE using CAP measurements in a population of 204 patients who underwent LT, 50% of whom were transplanted for alcoholic cirrhosis and only 2% transplanted for NASH. At the time of the study, CAP was not available using the XL probe and thus only the M probe was used. Perhaps as a result, only 157 of the cases were able to achieve valid results; however, TE detected steatosis in 44% of these patients, with 24% having advanced steatosis, indicating a high prevalence of post-LT de novo steatosis. Although it has not been compared with biopsy in the evaluation of grafts post-LT, the promise of non-invasive evaluation of graft steatosis with CAP and TE using the XL probe is compelling and is worthy of further study. Recurrent or de novo NASH may be present in the setting of normal or near normal aminotransferases.
Demonstration of NASH via TE and CAP should alert the clinicians and patient to this additional comorbidity and the potential greater likelihood for complications related to insulin resistance. In some recipients, despite achieving SVR fibrosis may continue to progress.
The presence of appreciable steatosis in the setting of advanced fibrosis will identify patients for both closer follow-up and for HCC surveillance. 68  69 The cutoffs for "significant fibrosis" also varied from 7.9 kPa to 10.1 kPa.

| Limitations of transient elastography
In a large, more recent prospective study of 259 patients evaluated by TE and liver biopsy, the cutoffs proposed for all patients to achieve a specificity of ≥90% were 8.1 kPa for F ≥ 1, 12.3 kPa for F ≥ 2, 15.1 for F ≥ 3, and 16.7 for F = 4. When only patients with HCV were analyzed, the cutoffs were adjusted to 8.1 kPa for F ≥ 1, 12.3 kPa for F ≥ 2, 16.5 for F ≥ 3, and 17.6 for F = 4 to achieve a specificity of ≥90%. 70 Although a large study such as this shows promise, many of the other published studies dedicated to post-LT TE have been small, single-center series and have lacked a validation cohort.
Other difficulties complicate the regular use of TE in the post-LT population. These patients may have differences relative to the anatomic position of the liver due to instances of hemi-diaphragmatic elevation or geometric mismatch between graft and abdominal cavity.
Liver stiffness, traditionally associated with degree of fibrosis, has also been shown to correlate with factors that affect sinusoidal pressure (SP). Factors that affect SP, such as inflammation and hepatic congestion (see below), alter LS even prior to the development of fibrosis. 71 Similarly, factors that increase mean arterial pressure, such as vasopressors or aerobic exercise, increase LS measurements in both animal and human models. 72  The presence of active inflammation increases LS measurements regardless of fibrosis stage. For example, one study found that among patients with the same level of fibrosis by histology, those with concurrent increased histological inflammation had higher LS than those with less inflammation. 75 LS is elevated in a number of comorbid conditions in the absence of liver fibrosis. 71,76 For instance, factors that increase central venous pressure, such as right heart failure, have been shown to increase LS. 77,78 Extra-hepatic cholestasis has also been shown to affect LS measurements. In a study of 15 patients undergoing endoscopic retrograde cholangiopancreatography for biliary obstruction in the absence of liver disease, successful biliary drainage reduced LS by 8.1 kPa. 79 In sum, understanding the clinical context is critical when interpreting TE results ( Table 1).
The presence of obesity has historically limited the use of TE in patients with suspected liver disease. 80 Hepatic steatosis in patients transplanted for both NASH and those with de novo NASH is an important, emerging issue in the LT population. Utilizing TE to assess the degree of steatosis in these patients may provide useful information for clinicians treating patients in the post-transplant setting. In particular, it may motivate physicians to aggressively treat exacerbating co-morbidities such as obesity and diabetes mellitus. Further, the armamentarium of agents to treat NASH as well as anti-fibrotic agents is expected to grow as compounds currently being evaluated in clinical trials become available. 86,87 The utility of these agents in the post-LT setting will ultimately need to be examined. One can envision TE as a tool to better understand and study patients who may be enrolled in trials evaluating these agents in the post-LT setting.

| Summary
The long-term results of liver transplantation are excellent. Liver fibrosis induced by various causes remains one of the principle reasons for graft dysfunction. Early identification of appreciable fibrosis or allograft injury in LT recipients is of paramount importance to permit risk stratification, ascertain prognosis, and thereby provide targeted interventions. It is important to establish the utility of new non-invasive methods, including TE, in this population. LS measurement by TE appears to be a promising tool for non-invasive monitoring of liver fibrosis progression after LT. Further studies with validation in large cohorts are needed to establish the precise association of TE values and cutoff values, with corresponding histological lesions and collagen content of liver biopsies using optimally sized biopsy specimens as a reference standard. Longitudinal assessment of fibrosis by means of these non-invasive tests may reduce the need for serial liver biopsies.
Given the non-invasive nature and ability to perform serial measurements, non-invasive testing for liver fibrosis could be used in the post-transplant clinical setting as an adjunctive tool for suspected recurrent or de novo liver disease. The high accuracy noted in the few published post-LT studies, especially for TE, suggests that these tests have similar diagnostic value as in the pre-transplant setting. More data are necessary in assessing the sensitivity of CAP for hepatic steatosis in the post-LT setting.

TA B L E 1 Factors affecting liver stiffness measurements
Of course, liver biopsy remains a cornerstone for the clinical management of LT recipients, as non-invasive tests cannot differentiate between pathologies that may coexist in this setting, such as acute or chronic rejection. Nonetheless, when the etiology of recurrent or de novo disease has been established, these non-invasive tests may be helpful in following fibrosis progression longitudinally and implementing preventive therapies in a timely manner.

CO N FLI C T O F I NTE R E S T
None.