Liver transplantation for hepatocellular carcinoma - factors influencing outcome and disease-free survival

Abstract

Hepatocellular carcinoma is one of the leading causes of cancer-related death worldwide. Liver transplantation can be a curative treatment in selected patients. However, there are several factors that influence disease-free survival after transplantation. This review addresses the pre-, intra- and postoperative factors that influence the risk of tumor recurrence after liver transplantation.

Key Words: Hepatocellular carcinoma; Survival; Risk factor; Diagnostics; Recurrence; Liver transplantation

Core tip: Hepatocellular carcinoma is one of the leading causes of cancer-related death worldwide. Liver transplantation can be a curative treatment in selected patients. This review addresses the pre-, intra-, and postoperative factors that influence disease-free survival and the risk of tumor recurrence after liver transplantation. Furthermore, novel diagnostic methods are presented and discussed.

INTRODUCTION

Hepatocellular carcinoma (HCC) is one of the leading causes of cancer-related mortality worldwide[1,2]. Most HCC co-occurs with liver cirrhosis, and the etiology of liver cirrhosis leading to HCC differs among global regions, e.g., hepatitis C virus infection in North America, Europe, and Japan and hepatitis B virus infection in China, South Korea, and Taiwan[1].

Depending on hepatic function, liver transplantation (LT) is a curative option for selected HCC patients. However, the recurrence rate after LT is reported to be between 5% and 23%[3-6].

This review addresses the issue of HCC recurrence after LT and is subdivided into three sections: preoperative, intraoperative, and postoperative factors.

PREOPERATIVE FACTORS

Size and number of HCC nodules

In 1996, Mazzaferro et al[7] introduced the Milan criteria (MC), which showed a survival benefit in HCC patients with one tumor nodule < 5 cm diameter or three tumor nodules with a maximum diameter < 3 cm for each nodule. The MC remain the gold standard for decision-making in LT settings for HCC patients.

The MC have been challenged by different authors and work-groups aiming at an expansion of the criteria (see Table 1). However, there is only a consensus that “modest expansion of the criteria should consider the dynamics of the waiting list and whether a worse prognosis could be tolerated, if there is no prejudice for patients without HCC”[4].

Table 1 Overview of the selection criteria for liver transplantation.

Name	Criteria	Ref.
Milan	1 tumor < 5 cm in diameter or	Mazzaferro et al[7]
	≤ 3 tumor nodules, each ≤ 3 cm in diameter
	No extrahepatic manifestation
	No vascular invasion
Up-to-seven criteria (“new Milan”)	Seven as the sum of the size of the largest tumor (in cm) and the number of tumors	Mazzaferro et al[166]
Kyoto	≤ 10 tumors, all ≤ 5 cm in diameter	Takada et al[167]
	PIVKA-II > 400 mAU/mL
UCSF	Solitary tumor ≤ 6.5 cm or	Yao et al[9]
	≤ 3 nodules with largest lesion ≤ 4.5 cm and total tumor diameter ≤ 8 cm
	No gross vascular invasion
Shanghai Fudan	Solitary tumor ≤ 9 cm in diameter or	Fan et al[168]
	≤ 3 lesions with the largest ≤ 5 cm and total tumor diameter ≤ 9 cm
Hangzhou	Total tumor diameter ≤ 8 cm or	Zheng et al[169]
	Total tumor diameter more than 8 cm with histopathological grade I or II and preoperative α-fetoprotein ≤ 400 ng/mL	Lee et al[170]
Asan	Largest tumor diameter ≤ 5 cm
	Hepatocellular carcinoma number ≤ 6
	No gross vascular invasion

What we do know is the so-called “Metroticket paradigm”: the larger the tumor burden, the lower the post-transplant expected survival. However, reliance on tumor size and number of nodules is not the most ideal approach because biological parameters, such as the response to local-ablative therapy or tumor grading, are not included.

In practice, there are two important pipelines in liver allocation: center-based allocation, meaning that a graft is offered to a center that chooses the ideal recipient for a donor organ, and MELD-based allocation, which involves a waiting list where patients with a high MELD score receive an organ sooner than patients with a low MELD score. This is important in HCC patients because most have good hepatic function and subsequently a low MELD score. These patients have the chance of requesting an “exceptional” MELD score when fulfilling certain criteria. This phenomenon brings us back to the MC: in most countries around the world, the MC is the deciding tool for an exceptional MELD request (see Table 2).

Table 2 Overview of prioritization systems in different transplant regions worldwide.

Region	Country	Basic listing	Standard exception	Patient benefit
Eurotransplant	Germany		1 tumor > 2 and < 5 cm up to 3 tumors > 1 and < 3 cm	Initial listing with MELD 22; upgrading every 3 mo by 10% mortality risk
	The Netherlands		1 tumor > 2 and < 5 cm up to 3 tumors > 1 and < 3 cm	Initial listing with MELD 20; upgrading every 3 mo by 10% mortality risk
				However, “test of time”: patient must have been on the waiting list for 6 mo prior
	Austria	Possible (if Milan criteria are met); however, irrelevant with center-based allocation	No
Europe	United Kingdom	Single lesion < 5 cm		No prioritization on the waiting list
		Up to 5 lesions < 3 cm
		Single lesion between 5 and 7 cm without progression over 6 mo
		No extrahepatic tumor
		No macrovascular invasion AFP < 1000 U/L
	France	Complex French Liver Allocation Score under consideration of
		Lab-MELD-Scores
		Tumor stage (T2 ranked higher than T1)
		Elapsed waiting time
		Distance between donor and Recipient hospital
	Switzerland		1 tumor > 2 and < 5 cm	Lab-MELD + 1.5 points per month
			up to 3 tumors > 1 and < 3 cm
North America	United States		1 tumor > 2 and < 5 cm	Initial listing with MELD 22; upgrading every 3 mo by 10% mortality risk
			up to 3 tumors > 1 and < 3 cm
South America	Brazil		1 tumor < 5 cm	Initial listing with 20 points, increase to 24 points after 3 mo and 29 points after 6 mo
			up to 3 tumors of less than 3 cm each

However, there is increasing evidence that a decision on organ allocation based only on radiological findings does not reflect the reality of tumor biology. A large single-center study of over 800 patients undergoing LT because of HCC revealed that in addition to tumor size, other clinicopathological parameters are helpful and necessary to identify patients with a lower risk of tumor recurrence[8]. Varona et al[5] showed that the French prognostic model, including pre-transplant α-fetoprotein (AFP), tumor size, and number of nodules, was better at detecting patients with HCC recurrence after LT than the MC or up-to-7 criteria. Yao et al[9] showed that expanded tumor size had no negative effect on patient survival or tumor recurrence. In conclusion, the development of the MC was an important step towards improved outcome after LT in HCC, but these criteria are likely too narrow, and further adaptations are necessary.

AFP

The biomarker AFP, which is encoded in humans by the AFP gene located on chromosome 4[10-12], is a frequently used serum parameter for the detection of HCC. Unfortunately, the sensitivity of AFP is limited because non-tumor liver diseases are also associated with high serum AFP levels[13,14], and AFP levels are not always increased in HCC[15]. Generally speaking, the higher the tumor differentiation, the lower the AFP level. However, elevated AFP levels predict HCC recurrence in a multi-predictive model, together with elevated liver enzymes, lactate dehydrogenase, small resection margins, and advanced tumor stage[16]. In an analysis of approximately 1500 patients with liver cirrhosis, AFP levels showed a sensitivity of 99% and a specificity of approximately 72% to detect HCC in combination with ultrasound[17]. Recently, several studies showed that during hepatitis treatment, AFP levels were helpful to detect HCC development[18-20].

The recurrence of HCC after LT is a problem; therefore, patients with a high risk of recurrence must be detected. A retrospective cohort study showed that a combination of biomarkers, such as AFP and the MC, was more sensitive for predicting tumor recurrence than the MC alone[21]. In a multivariate analysis, AFP was the only pre-transplant predictor of HCC recurrence and mortality in a cohort of 1074 patients transplanted for HCC[22]. Grąt et al[23] analyzed 121 patients undergoing LT for HCC and demonstrated that the combination of up-to-7 criteria and University of California, San Francisco (UCSF) criteria with AFP levels less than 100 ng/mL was associated with a minimized risk for tumor recurrence. Previously, in a small cohort of 20 patients, a cut-off level of 100 ng/mL for AFP was shown to be a predictor for HCC recurrence after LT[24].

AFP remains the gold standard biomarker for HCC detection and a prognostic marker for post-transplant tumor recurrence, particularly in combination with other morphological tumor criteria or diagnostic tools.

Response to bridging

To avoid tumor progression and waitlist drop-off, patients waiting for LT undergo local bridging therapies[25] such as transarterial chemoembolization (TACE), which has been shown to be effective and was first proposed as a treatment option in 1977[26]. TACE produces a combination of localized chemotherapy and ischemia by occluding feeding vessels with concomitant tumor necrosis[27]. Other bridging treatments are selective internal irradiation (SIRT) or radiofrequency ablation (RFA)[28,29]. The question of who is likely to benefit from locoregional therapy remains controversial, but, currently, the proposed optimal candidates for TACE have almost normal liver function, a tumor localized to the liver, an estimated survival of 16 mo[30], tumors > 3 cm in size, and signs of hypervascularization[31]. Some authors noted that poor liver function, vascular invasion, extrahepatic tumor load, bilobar tumors, arterioportal fistula, portal vein thrombosis, or renal dysfunction are contraindications for TACE[30,32]. A systematic review analyzed the different techniques and substances used for TACE[33]. Doxorubicin, epirubicin, or cisplatin alone or in combination are used as local chemotherapy. In addition to these drugs, lipiodol, gelatin sponge particles, or beads are injected after the chemotherapeutic substance to embolize the tumor-supplying artery. This should be performed in a selective or superselective manner to avoid damage to the nontumorous liver[31,33]. Acute liver failure, acute renal failure, gastrointestinal bleeding, and abscess formation are side effects of TACE, and a treatment related mortality rate of 2.4% was reported[33], mainly because of liver failure. The timing of TACE before LT remains unclear[34]. Decaens et al[35] showed that TACE only had a beneficial effect in patients with a waiting time longer than 4 mo. Others showed a positive effect of TACE on survival and tumor recurrence in patients with at least 6 mo on the waiting list[34,36]. The complete response is reported to be up to 30%[37], but rates of progressive disease following treatment have been reported to be 20%[38]. Overall, several studies have demonstrated that TACE prior to LT was associated with a good response rate for advanced tumor stages with acceptable survival rates after LT of 40% to 90% after at least 4 years, even in patients with HCC outside the MC[39-43]. A recently published study investigated 204 patients with HCC undergoing LT and showed reduced survival in patients with TACE prior to transplantation. The authors concluded that this might be because of a higher amount of pulmonary and distant metastasis[44]. Although most results appear promising, there are no controlled prospective trials to investigate the benefits of TACE prior to LT; thus, this issue remains controversial.

In contrast to TACE, SIRT is more effective in cases of large and multifocal HCC[28]. Negative side effects are rare[45-47]. Portal vein thrombosis is not an absolute contraindication for selective internal irradiation and, therefore, offers an alternative to TACE[48]. Selective internal irradiation leads to downstaging or tumor size reduction in 32% to 56% of patients with HCC[49,50]. Selective internal irradiation therapy is useful in bridging patients with HCC until LT[49,51].

During local tumor therapy with RFA, the tumor is destroyed because of local hyperthermia or chemical injury. RFA is performed percutaneously or during a surgical procedure[29]. Thus far, RFA has been shown to be a safe and effective procedure as a bridging therapy prior to transplantation[29,52-55]. A combination of different bridging therapies, such as RFA and TACE, are safe and effective regarding tumor necrosis[56]. However, a Cochrane analysis demonstrated that there is a lack of randomized clinical trials and no evidence demonstrating the superiority of RFA for bridging therapy over no intervention, chemotherapy, or transplantation[57].

Positron emission tomography

Positron emission tomography (PET) measures the metabolic activity of tumor tissue. In most oncologic cases, the tracer ¹⁸F-fluorodeoxyglucose (¹⁸F-FDG), an analog of glucose, is used[58].

The role of ¹⁸F-FDG-PET as a diagnostic tool in HCC patients is controversial. One of the first studies dealing with PET in HCC was published by Teefey et al[59]. In this work, ultrasound, computed tomography, magnet resonance imaging, and PET were compared. PET showed the worst sensitivity but displayed good specificity. A Chinese group noted the value of PET scans for the diagnosis of HCC lymph node metastases in a pre-transplant setting and for the detection of tumor recurrence after LT[60]. A study by Lee et al[61] showed an excellent predictive value for the ratio of the maximum standard uptake volume of the tumor to the maximum standard uptake volume of non-tumoral liver tissue, which predicted tumor recurrence after LT. Similar results were confirmed by several workgroups[62-66].

In non-transplant settings, ¹⁸F-FDG-PET is useful for estimating tumor differentiation[67], the probability of microvascular invasion[68], the diagnosis of bone metastases[69,70], and the estimation of tumor-free or overall survival[67,71-77].

One major issue is the dependency of PET results on tumor grade. Well-differentiated tumors have nearly the same metabolic activity as the surrounding liver tissue and, therefore, have similar tracer uptake.

However, as shown above, ¹⁸F-FDG-PET correlated well with microvascular invasion and post-transplant outcome. ¹⁸F-FDG-PET is a good marker for the biological behavior of HCC[58]. Therefore, ¹⁸F-FDG-PET appears to be a promising approach to evaluate the aforementioned findings using a prospective approach.

Metabolomics, proteomics, and transcriptomics

AFP is the standard serum biomarker for the detection of HCC. As previously mentioned, AFP is not specific for HCC, and, therefore, diagnosis on the basis of this biomarker is limited. However, until now, no other marker has replaced AFP as the new standard in routine clinical use. To overcome this dilemma, new and more sensitive markers have been investigated using metabolomic, transcriptomic, and proteomic techniques.

Metabolomics is a global unbiased analysis of biomarkers that identifies small molecular metabolites reflecting normal biological or pathological processes in biological fluids, tissues, organs, and organisms[78-81]. Metabolomics can measure the metabolite complements in living and diseased systems[81]. In contrast, proteomics is able to analyze all proteins within a biological system[82].

Thus far, several metabolites have been investigated and proposed as potential biomarkers for HCC, such as the aspartate metabolism pathway[83], 1-methyladenosine[81], and aberrant lipid metabolism[84]. Urinary liquid chromatography-hybrid triple quadrupole linear ion trap mass spectrometry (LC-QTRAP MS) revealed that butyrylcarnitine and hydantoin-5-propionic acid were markers to distinguish patients with HCC from patients with liver cirrhosis without HCC[85]. Huang et al[86] showed that betaine and propionylcarnitine in tissue and serum could distinguish HCC from hepatitis or cirrhosis better than AFP alone. A combination of metabolomics and transcriptomics reported reduced cellular glucose levels and reduced metabolites of cellular energy production in HCC[87]. A proteomic analysis detected 87 differently expressed proteins in patients with early HCC recurrence involved in catalytic pathways, signal transduction, and cell organization[88] and quantified novel phosphorylation sites that might be important for tumor progression in HCC[89]. The fields of metabolomics, transcriptomics, and proteomics are still emergent fields in cancer research. For most candidate metabolites or proteins, a definite role in HCC development and tumor progression is unclear, and further investigations are required. To date, there is no clinically available biomarker for HCC detection other than AFP.

MicroRNAs

MicroRNAs (miRNAs) are small noncoding RNA molecules that are not transcribed into proteins but are important for regulating the stability and translation of protein-coding messenger RNAs[90,91]. miRNAs were first described in 1993[92,93]; since then, hundreds of miRNAs have been identified. miRNAs play an important role in tumorigenesis[94-96]. In HCC, a number of miRNAs have been identified with partially prognostic significance[90,97,98]. These miRNAs can be downregulated, e.g., miR-122[99,100] and miR-199[101,102], or up-regulated, e.g., miR-21[103], miR-221[104,105], and miR-222[104,106,107]. This is only a partial list of the previously reported miRNAs. Therefore, miRNAs could be important diagnostic and therapeutic targets in the future of cancer and HCC therapy.

Circulating tumor cells

Despite the radical resection of localized HCC with hepatectomy or LT, postoperative tumor recurrence and metastasis are frequently observed[108,109], with the transplanted liver the most frequent site of early recurrence[110]. After access of the primary tumor cells to the blood stream, circulating tumor cells (CTCs) are postulated to be responsible for tumor recurrence and tumor metastasis after complete surgical resection[110-113]. Several methods have been investigated in the past to detect CTCs, mainly based on the identification of tumor-specific antigens or epithelial cell surface antigens that are present on the primary tumor[114,115]. One of the most frequently used markers for the detection of circulating tumor cells is the epithelial cell adhesion molecule, which is only expressed on a small proportion of HCC tumors[116]. In addition, in one-third of patients, only low numbers of CTCs are detectable[117]. Therefore, this technique is not suitable for the routine detection of HCC CTCs[118]. Novel approaches to detect CTCs showed promising results[119,120], but until CTC detection is able to guide the therapy of HCC patients, further basic and clinical research is required.

INTRAOPERATIVE FACTORS

Ischemia time

Intraoperative factors are less likely to be associated with tumor recurrence in the long-term. However, there is some evidence that ischemia times play a significant role in the recurrence of HCC.

A recent published study by Nagai et al[121] showed a significant effect of both cold (CIT) and warm ischemia time (WIT) on post-transplant HCC recurrence. In the multivariate analysis, a CIT of more than 10 h and a WIT of more than 50 min were risk factors for the development of a recurrent HCC[121]. These results were confirmed by a Munich workgroup[122].

This observation is explained by ischemia-reperfusion injury, which leads to hepatic microcirculatory barrier dysfunction and activates cell signals related to invasion and migration[121]. Furthermore, a cellular cascade leading to angiogenesis, cellular proliferation, and growth is activated by ischemia.

This theory was confirmed by an analysis by Croome et al[123], who showed an inferior outcome in HCC patients who received an organ from a donor who underwent cardiac death. These grafts are exposed to additional WIT.

Transfusion

Bleeding during LT remains a major problem, and sometimes large amounts of blood products are required[124-127]. Over the last decades, there has been a significant reduction in the need for transfusions[128]. Several studies have shown that blood loss and transfusion during LT were associated with decreased overall survival and increased complications[129,130]. In addition, perioperative transfusion is associated with earlier tumor recurrence and cancer-related mortality in colorectal cancer resection[131-135] and liver resection for colorectal liver metastases[136,137]. Shiba et al[138] showed that a reduction of blood supply during liver resection for HCC was associated with increased survival. In a meta-analysis of 5635 patients undergoing surgery for HCC, survival, tumor recurrence, and complications were negatively correlated with blood transfusion[139]. However, the use of intraoperative autotransfusion during liver surgery because of malignancy showed no negative effects in terms of survival or tumor dissemination[140,141]. Several studies have investigated the safety of blood salvage autotransfusion regarding tumor recurrence during LT in HCC patients. The authors concluded that in cases where nonruptured HCC tumor cells were filtered, or particularly when a leukocyte depletion filter was used, there was no increase in risk of tumor recurrence[142-144].

POSTOPERATIVE FACTORS

Immunosuppression

There is a relationship between the inflammatory state and carcinogenesis[7,145]. Pro-inflammatory cells and cytokines play a pivotal role in tumor growth, tumor invasion, and tumor spread[146,147]. Therefore, immunosuppression after transplantation can modify the inflammatory state and influence tumor recurrence. Most immunosuppression has a negative effect on the outcome of patients with HCC undergoing LT. Steroids[148], basiliximab[149], and calcineurin inhibitors (CNIs)[150] are postulated to be associated with an increased risk of HCC recurrence. In contrast, several studies reported that mammalian target of rapamycin inhibitors (mTORi) have positive effects on tumor recurrence and are favored drugs in HCC patients after LT[151-153]. A meta-analysis of five studies demonstrated a decreased recurrence rate and increased recurrence-free and overall survival in patients with sirolimus-based immunosuppression compared to patients with CNIs[154]. One reason for the positive effect of mTOR could be the inhibition of the phosphatidylinositol 3 phosphate kinase (PI3K)/Akt/mTOR pathway, which is a key regulator of the cell cycle and is responsible for cell proliferation and cancer[155]. However, thus far, randomized controlled prospective studies with long-term follow-up investigating the influence of this immunosuppression treatment are lacking[156].

Adjuvant treatment

Even after the use of a potentially curative treatment for gastrointestinal tumors, adjuvant therapy is used in most cases depending on the tumor stage. For HCC, this concept was not accepted for a long time[157]. A review by Duvoux et al[158] identified the problem with adjuvant therapy protocols after LT: most studies were small and retrospective with a low level of evidence. The authors concluded that the homogeneous and ethical selection of patients with a high risk of recurrence, stratification by confounding factors such as pre-transplant therapies and post-transplant immunosuppression, relevant endpoints focusing on recurrence, and appropriate follow-up are key points for appropriate studies on this issue. Similar conclusions were drawn 2 years later by Fujiki et al[159].

Even recently published trials lack the inclusion of a sufficient number of patients. A study by Teng et al[160] showed a beneficial effect of sorafenib in a case-control study. However, sorafenib was used in only 11 patients with HCC beyond the MC in either a curative intervention (n = 5) or a palliative regime after LT. Another prospective trial showed a benefit of sorafenib application after LT in seven patients with HCC beyond the MC compared to a historic control group[161].

The increased toxicity of sorafenib in patients after LT should be taken into account. This increased toxicity, which is not mechanistically understood, should lead to a dose reduction in affected patients[162].

In addition to sorafenib, Licartin has been proposed as a potential adjuvant treatment for HCC[158,159]. One Chinese study evaluated Licartin, an ¹³¹I-radiolabeled murine monoclonal antibody, in a transplantation setting with excellent results regarding the tumor recurrence rate and the overall survival in the treatment group[163]. Additional data or even multicenter data for Licartin after LT are unfortunately still lacking.

Some authors have used conventional chemotherapy protocols for the treatment of HCC after LT. Zhang et al[164] showed good short-term (1 year) results for a treatment protocol using the FOLFOX regime. The overall survival was even better in the midterm (3 years), but the recurrence rate did not differ significantly from the control after this time period. Wu et al[165] performed a randomized three-arm study and showed that the gemcitabine regimen and conventional chemotherapy significantly improved the survival rate and disease free survival rate of HCC patients who had major vascular invasion and/or microvascular invasion after LT compared to a best supportive care group.

In summary, the best approach for adjuvant treatment protocols has not been identified. Large, prospective, randomized studies should be performed in the future.

CONCLUSION

LT is the treatment of choice for selected patients with HCC. There are several factors that should be taken into account, particularly in preoperative settings. The previously used selection criterion of pure morphometric variables should be modified to include biological parameters, which offer a better risk stratification for tumor recurrence.

In addition to intraoperative parameters that influence the post-transplant prognosis (ischemia times, transfusion), immunosuppression is an important tool to prevent or at least reduce the recurrence of HCC. Adjuvant protocols have not yet been established for HCC.