Tools for monitoring patients with hepatocellular carcinoma on the waiting list and after liver transplantation

Abstract

For patients who are initially being listed for liver transplantation and for patients who are seeking to retain their active status on the wait list, the first step in the diagnostic consideration of hepatocellular carcinoma (HCC) is the accurate identification and staging of the tumors. Advances in imaging and biomarkers have improved our ability to diagnose HCC and in parallel have reduced the incidence of incidental HCC after liver transplantation. For the initial diagnosis, high sensitivity and high specificity are paramount considerations. The requirements change, however, when we are considering listing for transplantation: then, the prediction of the risk for posttransplant tumor recurrence and patient survival assume primacy. We must understand with confidence not only that the patient has a tumor but also that the tumor will affect the 5-year patient survival rate through the risk of tumor recurrence. Our first task in making the decision to list a patient with HCC for liver transplantation is to rule out factors representing absolute contraindications: extrahepatic disease and macrovascular tumor invasion. Beyond these variables, an understanding of the risk of posttransplant HCC recurrence is dominated by 2 major variables: tumor mass and tumor biology. Three major types of information can inform the decision to list a patient with HCC for liver transplantation. The first and usually dominant factors are the imaging characteristics, which include the size and number of the malignant lesions. The second type consists of biomarkers, which are most commonly abnormal proteins in the circulation that signify the presence of more dedifferentiated malignant cells with a higher propensity to invade and metastasize. The third type consists of factors from histological evaluations of tumor tissue obtained by biopsy or excision; these include factors such as the tumor grade, the presence of microvascular invasion, and protein, messenger RNA (mRNA), or DNA markers that have been correlated with more aggressive tumor behavior. Our team was tasked with generating recommendations for the best protocol for monitoring HCC patients on the wait list and for screening them for tumor recurrence after transplantation. The role of pretransplant histological factors in decision making is controversial even for listing. Does histological information add significantly to the prognostic power of the Milan criteria, the University of California San Francisco (UCSF) criteria, the rule of 7, or the total tumor volume (TTV)/alpha-fetoprotein (AFP) criteria as they are used at various centers? Do the risk of biopsy and a patient's aversion to biopsy trump the information gained? When we focus on the specific question being addressed here, however, we find no published evidence or guidelines from experts or center protocols that advocate the use of routine biopsy for evaluating possible alterations in the tumor characteristics of patients on the wait list. Therefore, we leave this question for discussions of the initial listing criteria. We focus our attention on the use of imaging modalities and tumor biomarkers for monitoring patients on the wait list and after transplantation. Wait-list screening addresses the question of what impact a change in imaging or biomarkers has on the risk of HCC recurrence and on patient survival. We must recognize that the impact of such changes may vary in the presence or absence of interventions such as ablation or angiographic chemotherapy or radiotherapy aimed at the control or elimination of tumors in the liver. Modifications in posttransplant immunosuppression or the use of adjuvant chemotherapy may also have an influence. Our task is also more than establishing the simple presence of an impact on the risk of posttransplant recurrence and survival; we must determine whether the risk is great enough to justify withholding or withdrawing liver transplantation as a suitable therapeutic modality from a patient with specific characteristics. In practice, these recommendations may require a flexibility that recognizes the dynamic balance between the limitations in access to liver transplantation in various countries and the transplant benefits accrued by both patients with HCC and patients without tumors. Changing the stringency of the listing or maintenance characteristics that are required for tumor patients will alter the posttransplant outcomes, change the number of patients who qualify for transplantation, and thus affect others on the waiting list. 3D, 3-dimensional; AFP, alpha-fetoprotein; CI, confidence interval; CT, computed tomography; DC-MRI, dual-contrast magnetic resonance imaging; DCP, des-gamma carboxyprothrombin; FAT SAT, fat saturation; [18F]-FDG, fludeoxyglucose F 18; HCC, hepatocellular carcinoma; HR, hazard ratio; MDCT, multidetector computed tomography; MRI, magnetic resonance imaging; mRNA, messenger RNA; PET, positron emission tomography; RFA, radio frequency ablation; S-MRI, superparamagnetic iron oxide–enhanced magnetic resonance imaging; TACE, transarterial chemoembolization; TTV, total tumor volume; UCSF, University of California San Francisco. The keywords used in the searches varied with each question. For electronic searches, we concentrated on MEDLINE, but we also reviewed the Cochrane database. Reference articles were ranked according to the University of Oxford Centre for Evidence-Based Medicine classification. For imaging strategies, we performed a search of the MEDLINE database for January 2002 to March 2010 with the medical subject headings hepatocellular carcinoma, primary liver cancer, liver cell carcinoma, and liver cancer as free text words in combination with liver transplantation, waiting list, and imaging. All titles and abstracts were reviewed, and appropriate studies were further assessed. We further reviewed the reference sections of all these publications to identify studies that may have been missed during the primary search. The main analysis concentrated on randomized controlled trials, controlled studies, and meta-analysis reviews. A second analysis concentrated on case-control analytical studies. Different concordant clinical studies and studies reporting the opinions of respected experts were also considered for inclusion. For biomarkers, the same medical subject headings were combined for the same period with liver transplantation, waiting list, waitlist progression, biomarkers, alpha-fetoprotein, and prothrombin produced by vitamin K absence or antagonism II. The same progressive search strategy was then employed. What imaging strategy is optimal for monitoring patients with HCC on the liver transplant waiting list? What biomarker monitoring is best for patients with HCC on the liver transplant waiting list? What is the most effective and efficient method for monitoring patients for recurrence after transplantation? The recommendations generated by our group may require modifications to mesh appropriately with those of the group on imaging criteria for transplant candidacy because wait-list monitoring is in practical terms a direct extension of the initial evaluation of the patient for transplant eligibility. We found that evidence-based studies of the best imaging approaches for patients before liver transplantation, on the waiting list, and after liver transplantation were extremely limited. We propose in general to accept the conclusions established by the experts on imaging at the National Conference on Liver Allocation in the United States.1 We agree with these experts that the current United Network for Organ Sharing criteria are inadequate, and we support the consensus position on the need for minimal hardware requirements and image acquisition protocols, which are outlined here in Tables 1 and 2. These experts also proposed a new classification system for local liver lesions. These precautions are justified by the unacceptable inaccuracy of staging reported by some centers.2 We support contrast-enhanced magnetic resonance imaging (MRI) and computed tomography (CT) as the gold-standard imaging examinations, as reported previously,3-7 with appropriate specifications of their minimal requirements. Three additional considerations merit consideration. First, an increase in quality can be obtained with dual-contrast magnetic resonance imaging (DC-MRI), which has been shown to be significantly better at detecting small HCCs than CT or superparamagnetic iron oxide–enhanced magnetic resonance imaging (S-MRI).8 DC-MRI consists of baseline magnetic resonance images, superparamagnetic iron oxide–enhanced images, and gadobenate dimeglumine–enhanced dynamic images. DC-MRI provides not only the T1 and T2 signal character of hepatic nodules and superparamagnetic iron oxide–enhanced T2 or vT2* signal information but also the dynamic enhancement character of the nodules. In Lee et al.'s study,8 the mean sensitivities for detecting small HCCs (<2 cm) were 37.3%, 39.1%, and 55.5% for CT, S-MRI, and DC-MRI, respectively (P < 0.001). Second, outcomes equivalent to those obtained with contrast-enhanced CT and MRI have been reported by several centers with contrast-enhanced ultrasound, which we believe fulfils some of the requirements for a third diagnostic option when the appropriate hardware and expertise exist. The lack of protocolized/regimented image capture with contrast-enhanced ultrasound remains its main drawback and prevents a recommendation for equivalence with CT and MRI. Routine sonography (ie, not contrast-enhanced), however, is not recommended as the sole imaging modality for lesion detection before transplantation. Despite this, many institutions worldwide still use sonography as the study of choice for surveillance because of its relatively low cost, its lack of ionizing radiation, and its availability.9 Although their research was not performed in a transplant setting, Thompson Coon et al.10 reported that a combination of AFP testing and an ultrasound examination every 6 months was an effective surveillance strategy. The model estimated that in comparison with no surveillance, this strategy may triple the number of people with operable tumors at diagnosis and almost halve the number of people who die from HCC. Third, consideration is also required for a potentially important role for positron emission tomography (PET).11-14 The preoperative AFP level and vascular invasion in poorly differentiated HCCs were found to be significantly associated with positive fludeoxyglucose F 18 ([18F]-FDG) PET findings,11, 13, 14 and its uptake during PET was specifically predictive of microvascular invasion and tumor recurrence after liver transplantation for HCC.13 Furthermore, [11C]-choline PET requires further study as a potential tracer that could complement [18F]-FDG because it is better at detecting moderately differentiated HCC.12 Despite the advances in PET/CT and its clinical applications, [18F]-FDG PET/CT was recently reported to be unsuitable as a screening tool for recurrence after liver transplantation because of its limitations for small lesions, intrahepatic lesions, and brain lesions. Nevertheless, [18F]-FDG PET/CT could provide additional information for the diagnosis of suspected recurrence beyond that provided by conventional modalities, and it could contribute to the clinical management of HCC recurrence after liver transplantation, especially in patients with extrahepatic recurrence.15 We found no publications addressing the specific points of timing or optimal imaging modalities for screening liver transplant candidates on the wait list. Our recommendations are based on the evidence for the accuracy of imaging modalities in the diagnosis and staging of HCC as well as our understanding of the biology of HCC (eg, tumor doubling times) and program practices in the area of tumor surveillance for liver transplant patients on the wait list. Wait-list monitoring should be performed every 3 months with contrast-enhanced CT or MRI. The quality of the imaging studies during the evaluation for listing and the radiation exposure for each patient should be given due consideration during the selection of the follow-up imaging modality (grade C). DC-MRI and PET/CT require further study for the evaluation and monitoring of patients with HCC on the liver transplant wait list. Patients who have progressed beyond the criteria acceptable for listing for liver transplantation should be placed on hold, and consideration should be given to therapeutic interventions. If no intervention [eg, transarterial chemoembolization (TACE), radio frequency ablation (RFA), or transarterial radioembolization] is felt to be appropriate or if the tumors cannot be adequately controlled by the intervention, the patients should be removed from the wait list (grade C). Algorithm for monitoring HCC patients on the liver transplant waiting list. *Imaging hardware should meet interventional guidelines. **One should consider increasing the frequency of monitoring to every 4 to 6 weeks in accordance with the center's down-staging protocol. ***The disease is stable for 3 or 6 months in accordance with the center's down-staging protocol. The liver transplant listing criteria for patients with HCC have been the subject of intense debate since the landmark publication of Mazzaferro et al. in 1996.16 Research and especially clinical practice have focused largely on the impact of variations in the morphological criteria, perhaps because the initial transformative report by Mazzaferro et al. did not find the AFP level to be associated with outcomes. Recent advocacy for the expansion of candidacy from the Milan criteria to guidelines such as the UCSF criteria17 or the rule of 7 (as suggested by application of the Metroticket predictive model18) has continued to focus on morphological factors. The importance of microvascular invasion, the tumor grade, and the presence of microsatellite tumors to HCC recurrence and survival is apparent from several analyses, and this finding adds the consideration of aggressive tumor behavior or biology to information about the size and number. Unfortunately, this information is obtained from the study of explant pathology findings and cannot be directly used for listing decisions. As a source for such information, biopsy suffers from a triple challenge: sampling error, the risk of needle tract tumor seeding, and invasiveness with patient aversion. Serum biomarkers that are available by relatively noninvasive means correlate well with microvascular invasion, the tumor grade, and microsatellitosis, and they should provide a method for considering the biological nature of HCC as well as morphological information and thus for improving the predictive power. In fact, Toso et al.19 found AFP and TTV to be equally powerful and independent predictors of posttransplant patient survival; this is a very strong argument for both morphological and biological factors to be included when the patient criteria for listing for liver transplantation are being considered. AFP is a protein produced in normal fetal hepatocytes that is lost during later development and maturation, but it can reappear in dedifferentiation situations such as oncogenesis. It is produced by approximately 60% of HCCs, and its usefulness in the diagnosis of liver tumors is limited by this lack of sensitivity. AFP has major advantages: it is simple, relatively inexpensive, noninvasive, and widely available. AFP has been studied in a series of mainly single-center retrospective reviews for its ability to predict posttransplant outcomes; predictive power has been reported at cutoff values ranging from >10 to > 1000 ng/mL.17, 19-30 Despite an abundance of data, AFP has not yet been widely adopted as a selection criterion. Toso et al.19 reviewed 6478 patients in the Scientific Registry of Transplant Recipients database who underwent transplantation for HCC between 2002 and 2008, and they reported that only the TTV [hazard ratio (HR) = 1.4, 95% confidence interval (CI) = 1.03-2.0, P < 0.05] and the AFP level (HR = 1.1, 95% CI = 1.1-1.19, P < 0.001) predicted patient survival. The combination of a TTV < 115 cm3 and an AFP level < 400 ng/mL accurately predicted posttransplant survival (HR = 2.95, 95% CI = 1.7-2.4, P < 0.001). Patients not meeting these criteria had a survival rate lower than 50% at 3 years. The report concluded that a composite patient selection score combining the TTV and the AFP level was the most effective of all the criteria tested for predicting patient survival. Other criteria that they studied included the largest tumor diameter, the number of tumors, the Milan criteria, and the UCSF criteria. The composite TTV/AFP criteria have been adopted by the University of Alberta and the University of Geneva for selecting patients for liver transplantation and for monitoring patients on the wait list. Limited published information on the use of AFP for following patients on the wait list is available. Two articles are noteworthy. Publications from Montreal31 and Paris32 have reported the importance of AFP changes as a dominant predictor of HCC recurrence after transplantation, and they have proposed 50 and 15 ng/mL/month, respectively, as optimal cutoff values for the velocities of these changes. A univariate analysis of 48 patients who underwent transplantation for HCC in Montreal revealed a positive correlation between the preoperative AFP slope and vascular invasion (P = 0.045), the total tumor diameter in the explant (P = 0.040), the Cancer of the Liver Italian Program score (P = 0.017), and recurrence-free survival (P = 0.028). In a multivariate analysis, a preoperative AFP slope > 50 ng/mL/month (risk ratio = 11.64, 95% CI = 1.59-85.39, P = 0.016) and a total tumor diameter > 7 cm (according to a pathological examination; risk ratio = 37.06, 95% CI = 3.29-417.69, P = 0.003) were identified as independent predictors of tumor recurrence. With a receiver operating characteristic value > 50 ng/mL/month, the sensitivity was 36%, the specificity was 97%, the positive predictive value was 80%, and the negative predictive value was 84% for the ability to predict recurrence. The 1-year recurrence-free survival rate was 40% for patients with a preoperative AFP slope > 50 μg/mL/month and 90% for those with a slope < 50 ng/mL/month (P < 0.001). The difference in the last AFP value before transplantation (63 ng/mL with recurrence versus 18 ng/mL without recurrence) was not significant (P = 0.54), possibly because of a lack of power in the study. All patients within the Milan criteria received percutaneous alcohol injections in this series, whereas all patients beyond the Milan criteria received a course of 3 TACE treatments (any patient who did not undergo transplantation within 2 months received a fourth treatment). Vibert et al.32 retrospectively reviewed prospectively gathered database records for 153 subjects with elevated AFP levels (>7 ng/mL) at listing. They employed an aggressive tumor control regimen that included routine initial TACE for all Child A and B patients at listing and follow-up RFA or TACE for those patients progressing morphologically or in their AFP levels. Patients were removed from the wait list only because of the development of extrahepatic metastases or macrovascular invasion (as evidenced by tumoral portal vein thrombosis). This may explain the very low cutoff value in their study; a failure to prevent even a modest rise in the AFP level appears important in the face of aggressive intervention. Importantly, reducing the AFP elevation to <15 ng/mL/month after TACE or RFA resulted in outcomes equal to those for patients whose initial values were in the low range. No static value of AFP was a statistically significant predictor of outcomes in this study, although the 5-year recurrence-free survival rates were 23% lower with AFP levels greater than 400 ng/mL and 28% lower with AFP levels greater than 1000 ng/mL. The lack of significant predictive power despite differences in survival of 23% to 28% at 5 years may again be associated with the limited number of patients in this single-center study. The study was criticized in an editorial for its poor positive and negative predictive values (0.41 and 0.38, respectively), and in a receiver operating characteristic analysis, it had an area under the curve of only 0.56. The importance of an increase of 15 ng/mL/month may also mean a very different prognosis if the baseline value is 20 or 300 ng/mL, and the use of the slope may prove more useful. Although the differences were statistically significant, even the group with this low rate of progression achieved a 5-year survival rate of 54%. With the widespread experience of longer times on the waiting list and the increasing interest in tumor down-staging to qualify patients for liver transplantation who are outside the current guidelines at presentation, measures such as changes in AFP levels and the progression of morphological characteristics in patients on the waiting list may increase in importance. We recently published an analysis of 6817 patients in the Scientific Registry of Transplant Recipients in which we examined the feasibility of down-staging for the transplantation of patients with HCC with AFP levels > 400 ng/mL.33 We investigated the relevance and relative impact of absolute values of AFP levels and changes in patients on the waiting list with respect to dropout and survival rates, and we studied whether AFP down-staging could be a meaningful pretransplant criterion. A local HCC treatment (TACE, RFA, or percutaneous ethanol injection) before transplantation was used in 41% of the patients on the waiting list. Patients with AFP levels > 400 ng/mL at the time of listing whose levels were down-staged to ≤400 ng/mL before transplantation had a better intent-to-treat survival rate than patients whose AFP levels were not reduced to ≤400 (81% versus 48% at 3 years, P ≤ 0.001), and their survival rate was similar to the rate for patients with stable AFP levels ≤ 400 ng/mL (74%, P = 0.14). Patients whose AFP levels were down-staged to ≤400 ng/mL and patients with levels persistently ≤ 400 ng/mL also had similar rates of dropout from the list (10% in both groups) and similar posttransplant survival rates (89% versus 78% at 3 years, P = 0.11). This AFP down-staging was associated with good survival whatever the original AFP level was (even if it was originally >1000 ng/mL). Only the pretransplant AFP level independently predicted survival in a multivariate Cox regression analysis (unlike the AFP level at listing or AFP changes). The intent-to-treat HR for the pretransplant AFP level was 1.59 (95% CI = 1.43-1.76, P ≤ 0.001); the posttransplant survival HR was 1.49 (95% CI = 1.29-1.72, P < 0.001). Down-staging HCC patients with high AFP levels (>400 ng/mL) is feasible and leads to intent-to-treat and posttransplant survival rates similar to those of patients with AFP levels persistently ≤ 400 ng/mL. Wait-list progression beyond the accepted initial listing criteria for either morphological or biological (AFP) values indicates that a patient is no longer acceptable as a candidate for liver transplantation. It may be most appropriate to suspend the patient's candidacy while down-staging interventions are attempted to regain satisfactory tumor control or, if this fails, to permanently remove the individual from consideration for transplantation. Therefore, a high initial AFP level or a rising AFP level while the patient is on the wait list should not be taken as an indication to remove the candidate from consideration for transplantation but rather should be considered an indication for a need for an intervention (or a more effective intervention) to control the tumors. Excellent survival outcomes according to an intent-to-treat analysis can be achieved as long as AFP levels less than 400 ng/mL can be achieved and maintained at the time of transplantation. In addition to these outcomes with changing AFP levels in patients on the waiting list, the impact of the stability of the tumor size and number over time on the posttransplant recurrence risk has also been reported. Yao et al.34 and Roberts et al.35 reported impressive survival data with a down-staging protocol that required tumor stability or regression over a 3-month period after the intervention. They proposed the consideration of delaying transplantation for most patients and for all patients with advanced tumors after interventional treatments for HCC; they hypothesized that the potential for tumor stability after an intervention reflects the biology of the tumor and, therefore, the risk of recurrence after transplantation. Total AFP can be divided into 3 different glycoforms (AFP-L1, AFP-L2, and AFP-L3) according to the capacity for binding to the lectin Lens culinaris agglutinin. AFP-L3 is the major glycoform of AFP in the serum of HCC patients and can be detected in 35% of patients with small HCCs (<3 cm).36, 37 A high percentage of AFP-L3 is closely related to poor differentiation and biologically malignant characteristics of HCC, and HCC patients who are positive for AFP-L3 have worse liver function, poorer tumor histology, and larger tumor masses.36-38 This marker is also significantly elevated in the presence of metastatic HCC.39-41 AFP-L3 is significantly related to portal vein invasion and patient outcomes and appears to be a useful prognostic marker for HCC.42 DCP (also known as prothrombin produced by vitamin K absence or antagonism II) is an abnormal, inactive prothrombin resulting from an acquired posttranslational defect of the prothrombin precursor in HCC cell lines. DCP is present in 50% to 60% of all HCC patients but in only 15% to 30% of early HCC cases. Abnormal prothrombin levels do not correlate well with serum AFP levels.43, 44 Because AFP (and AFP-L3) and DCP levels do not correlate in patients with HCC, it seems reasonable to determine the levels of both markers to improve the accuracy of an HCC diagnosis. A limited study of DCP has been reported in the context of liver transplantation for HCC. Yamashiki et al.36 found that in 6 of 9 patients with posttransplant recurrence, the DCP level was raised beyond 40 mAU/mL. At this cutoff, the sensitivity was 54.5%, and the specificity was 96.6%. Carr et al.42 found a combination of Lens culinaris–reactive AFP (AFPL3), AFP, and DCP levels to be superior to any marker of HCC alone in determining the outcomes of patients with HCC. No such study has been performed for monitoring patients for HCC recurrence after liver transplantation. Serum ferritin (>200 ng/mL).45 Blood AFP mRNA.46-48 Telomerase reverse-transcriptase mRNA.38 Other molecular factors have been associated with HCC invasion and have potential prognostic value. Molecular markers for the malignant cell phenotype include changes in ploidy, cell proliferation markers (proliferating cell nuclear antigen and Ki-67), nuclear morphology, the p53 gene and other cell cycle regulators (cyclins and cyclin-dependent kinases), oncogenes and their receptors (c-myc, hepatocyte growth factor, and c-met), and factors related to apoptosis and telomerase activity. Molecular markers involved in invasion and metastasis include adhesion molecules (epithelial cell adhesion molecule, E-cadherin, and catenins), and proteinases involved in extracellular matrix degradation have also been considered as biomarkers for the malignant phenotype of HCC and have been related to prognosis and treatment results.49, 50 For the first group of candidates, data are limited in a transplant context and are nonexistent for wait-list monitoring. For the second group, data exist only in a diagnostic context or for prognostication outside a transplant context. All share the limitations of a limited capacity for analysis by the great majority of liver transplant programs at present. AFP levels should be monitored every 3 months for all patients on the liver transplant wait list (grade B). Any increase in the AFP level for patients on the wait list merits the consideration of (1) an intervention to achieve tumor control before transplantation in patients with known HCC or (2) a further diagnostic imaging evaluation for patients without a previous diagnosis of HCC (grade D). Patients whose AFP level progresses beyond 400 ng/mL should be placed on hold until the initiation of therapies for down-staging the AFP level to this range (grade C). If the AFP level increases at a rate > 50 ng/mL/month for a patient on the wait list, placing the patient on hold until the outcomes of interventions are known should be seriously considered. This indicator requires further evaluation in well-structured prospective trials (grade D). Patients whose AFP levels cannot be adequately controlled despite interventions or for whom interventions are not considered appropriate should be removed from the wait list (grade C). DCP and AFP L3 merit further study in the context of HCC outcome predictions after transplantation and in the context of monitoring patients on the wait list. The literature is barren with respect to studies addressing the question of rational screening protocols for monitoring patients for HCC recurrence after liver transplantation. The sole publication specifically addressing this issue is a summary of proceedings from 2005.51 The authors noted the dearth of published information on this topic and provided an expert opinion supported by references for the issue. They concluded that only approximately 1% of patients undergoing liver transplantation for HCC would ultimately prove to have a curable recurrence, and at least 80% of such recurrences would occur within the first 2 years after transplantation. Regular imaging studies and monitoring for tumor biomarkers are performed by most programs for at least 2 to 3 years after liver transplantation for HCC. In contrast to Roberts' published opinion,51 a recent poster from Toronto52 at the 2010 American Transplant Congress reported on the ability to provide potentially curative therapies to more than one-third of patients with recurrent HCC after transplantation within a program of regular imaging and AFP monitoring. In a recently published study using an upper limit of normal of 20 ng/mL, the sensitivity and specificity of elevated AFP levels as a sign of recurrence were 67% and 100%, respectively; with 10 ng/mL, the sensitivity and specificity were 78% and 98%, respectively.53 Data are limited on the diagnostic ability of the broad range of alternative candidate biomarkers for HCC recurrence. The high cost of imaging studies appropriate for the diagnosis of HCC recurrence within the liver (contrast-enhanced CT and MRI) or elsewhere (PET/CT) should limit their application to 6-month intervals for the first 3 to 5 years after transplantation in patients who have indicators of increased risk for recurrence (tumors beyond the Milan criteria, last AFP level before transplantation > 400 mg/dL, or high-risk markers such as microvascular invasion, a high tumor grade, and microsatellitosis according to explant pathology findings) or for whom suspicions are raised by abnormal biomarker studies or for clinical reasons. Regular ultrasound is less accurate but is also much less expensive and may add value when it is applied at 3-month intervals, especially in high-risk cases. AFP monitoring is much less costly and may be appropriate at more frequent intervals. There are insufficient data to recommend the use of other biomarkers. Considerations of cost efficiency raise the question of whether groups of patients can be defined for whom the risk of HCC recurrence is so low that routine monitoring after liver transplantation is not advisable. The risk of recurrence in patients within the Milan criteria has been regularly reported to be 6% to 10%.54 The inclusion of additional defining criteria such as pretransplant AFP levels or a risk factor analysis based on explant pathology findings should certainly be capable of reducing this risk. In the University of Alberta experience, although 2 of 4 of patients (50%) who were within the Milan criteria, had AFP levels > 400 ng/mL, and survived more than 6 months after liver transplantation experienced HCC recurrence, only 2 of 66 patients (3%) who were within the Milan criteria and had AFP levels < 400 ng/mL experienced recurrence. The absence of risk factors such as microvascular invasion according to explant pathology findings would almost certainly further reduce the risk. With the risk of recurrence being less than 3% and with curative therapy being possible for no more than one-third of these cases (1% overall), the costs and benefits of imaging studies for monitoring these patients certainly make their value questionable. Because of the markedly lower costs and the high likelihood of background hepatitis C virus/hepatitis B virus and other risk factors in HCC patients (and thus the additional risk of de novo tumor development), posttransplant monitoring for AFP appears advisable. Patients who have an increased risk of HCC recurrence (as indicated by staging beyond the Milan criteria at listing or at transplant, by AFP levels > 400 mg/dL at transplant, or by high-risk criteria such as microvascular invasion, a high tumor grade, or microsatellitosis identified by explant pathology findings) should undergo contrast-enhanced CT or MRI every 6 months for 3 to 5 years. Patients with no indicators of increased risk (staging within the Milan criteria, AFP levels < 400 mg/dL before transplantation, and no high-risk criteria according to explant pathology findings) do not merit regular imaging for follow-up (grade D). AFP levels should be monitored every 3 months for 2 years and every 6 months thereafter in all patients with HCC that was diagnosed before transplantation or during the explant examination (grade C). Any rise in AFP levels beyond 20 ng/mL merits liver-focused imaging (contrast-enhanced CT or MRI) and systemic imaging (PET/CT) for the diagnosis of potentially treatable recurrences (grade C). Aggressive interventions (liver or pulmonary resection or ablation or chemotherapy) should be considered for posttransplant patients with HCC recurrence if the liver function and the overall condition are favorable. Posttransplant monitoring of HCC patients.