Management of hepatocellular carcinoma

Abstract

Purpose

Hepatocellular carcinoma (HCC), a common cause for cancer-related death, is increasing worldwide. Over the past decade, survival and quality of life of HCC patients have significantly improved due to better prevention strategies, early diagnosis, and improved treatment options. We performed this narrative review to synthesize current status on the HCC management.

Methods

Literature search for publications especially over the last decade, which has changed the paradigm on the management of HCC.

Results

Hepatitis B vaccination and treatment of chronic hepatitis B and C are important measures for HCC prevention. Screening and surveillance for HCC using ultrasonogram and alpha-fetoprotein estimation are directed toward cirrhotics and hepatitis B patients at high risk of HCC. If detected at an early stage, curative treatments for HCC can be used such as tumor resection, ablation and liver transplantation. HCC patients without curative options are managed by loco-regional therapies and systemic chemotherapy. Loco-regional treatments include trans-arterial chemoembolization, radioembolization and combinations of loco-regional plus systemic therapies. Currently, sorafenib is the only FDA-approved systemic therapy and newer better chemotherapeutic agents are being investigated. Palliative care for terminally ill patients with metastatic disease and/or poor functional status focusses on comfort care and symptom control.

Conclusions

In spite of significant advancement in HCC management, its incidence continues to rise. There remains an urgent need to continue refining understanding of HCC and develop strategies to increase utilization of the available preventive measures and curative treatment modalities for HCC.

Introduction

Hepatocellular carcinoma (HCC) is the third leading cause of cancer-related death worldwide and the ninth leading cause in the USA. The incidence of HCC is rising worldwide with a reported annual incidence of 2.1/100,000 in 2001 and 3.2/100,000 in 2006 (El-Serag 2001, 2002, 2004; Jepsen et al. 2007; Parkin et al. 2005). According to the International Agency for Research on Cancer, HCC is the fifth most common cancer in men (523,000 cases per year with 7.9 % of all cancer cases) and the seventh most common cancer in women (226,000 cases per year with 6.5 % of all cancer cases) (Kudo 2011). An effective treatment regimen involves a multidisciplinary approach, involving hepatologists, surgeons, interventional radiologists and LT teams. Physicians taking care of patients with cirrhosis and chronic liver disease should be familiar with the strategies designed to prevent HCC, regular surveillance for early diagnosis and strategies to manage the various stages of HCC (Bruix et al. 2011).

Prevention of hepatocellular carcinoma

Hepatocellular carcinoma is one of the best examples to justify the phrase “Prevention is better than cure.”

Primary prevention

Approximately 80 % of HCC is reported from East Asia and sub-Sahara Africa where the main risk factor is hepatitis B virus (HBV) infection. HBV vaccination is a safe and effective method of preventing HBV infection. A population-based study from Taiwan showed that the incidence of HCC among children aged 6 years after the introduction of HBV vaccination program in 1984 decreased from 0.7/100,000 (1981–1986) to 0.57/100,000 (1986–1990) and to 0.36/100,000 (1990–1994) (Chen et al. 1987). HBV vaccine is recommended for all newborns and children. Children born to HBV-positive mothers should also receive HB immunoglobulin. Adults at high risk of acquiring HBV infection should also be vaccinated (Bruix et al. 2011).

Secondary prevention

Antiviral treatment for chronic HBV and hepatitis C virus (HCV) infections reduces the incidence of HCC. A large placebo-controlled, randomized study in patients with HBV-related, compensated cirrhosis showed the beneficial effect of lamivudine 100 mg per day in reducing the incidence of HCC from 7.4 % in placebo-treated patients to 3.9 % in patients treated with lamivudine (P < 0.05) (Liaw et al. 2004). A meta-analysis of 6 controlled trials comparing treated vs. untreated patients showed a reduction in HCC among treated patients (9.7 vs. 3.3 %, P = 0.01), with an overall 56 % reduction in the occurrence of HCC with anti-viral treatment (Singal and Fontana 2011). It is important to realize that, in spite of complete viral suppression of the virus, patients do remain at risk of HCC, and these patients should continue to receive regular surveillance. In a recently reported meta-analysis of 45 studies involving 9,330 patients who received various oral anti-viral drugs for HBV infection, the pooled rate of HCC development was 1.4 % per 100 person years at follow-up, with a higher incidence in individuals aged 50 years or older (2 vs. 1 %), those with underlying cirrhosis (3 vs. 0.3 %) and patients with detectable HBVDNA at the time of HCC occurrence (2 vs. 1 %) (Singal and Fontana 2011).

The main risk factor for developing HCC in North America, Europe and Japan is hepatitis C virus (HCV) infection (El-Serag 2011). Currently, we do not have a vaccine to prevent HCV infection. Pooled data from 14 controlled studies showed that HCV treatment with a sustained virologic response results in about 57 % reduction in HCC occurrence (Singal et al. 2010a, b). Development of newer and more potent drugs for HCV is encouraging and is likely to further reduce the HCC incidence in future.

Tertiary Prevention

In a pooled analysis of five randomized studies evaluating the use of interferon after resection or ablation of HCV related HCC, there was a reduction in the HCC recurrence by 74 % and mortality by 69 % (Singal et al. 2010a, b). Similarly, in a retrospective analysis of 4,569 chronic HBV patients (518 treated), the use of oral nucleoside analogs after resection of HCC reduced the HCC recurrence rate (HR 0.67, CI 0.55–0.81) (Wu et al. 2012).

Diagnosis of hepatocellular carcinoma

Screening and surveillance

Screening refers to the application of diagnostic tests to asymptomatic subjects who are at risk of acquiring HCC. Surveillance is the repeated application of screening tests to these subjects. The prognosis of HCC detected after the onset of symptoms is grim, with 0–10 % survival at 5 years.(El-Serag 2011). In contrast, 5-year survival of small HCC if detected at an early stage with screening and surveillance is over 50 % (Bruix et al. 2011; El-Serag 2011; El-Serag and Davila 2011). A large population-based randomized controlled study from China examined the impact of HCC surveillance in 18,816 patients (9,373 screened) with chronic HBV. The study showed a 37 % reduction in HCC-related mortality over 5 years among screened subjects as compared to the unscreened group (Zhang et al. 2004). In a meta-analysis of 47 studies on over 15,000 patients, there was about twofold increase in early-stage detection of HCC, receipt of curative treatments and overall survival among 41 % of patients, who had HCC detected by surveillance strategy (Singal et al. 2014).

At risk population

Surveillance is considered beneficial if it improves survival by 100 days (Naimark et al. 1994). Surveillance is recommended for patients with cirrhosis or with chronic HBV infection (Table 1) (Sarasin et al. 1996).

Table 1.

Surveillance recommendations for patients at high risk of hepatocellular carcinoma (Evan et al. 2014)

A. Patients with high risk of HCC for whom surveillance is indicated

(a) Patients with liver cirrhosis

Hepatitis C with cirrhosis

Hepatitis B with cirrhosis

Stage 4 primary biliary cirrhosis

Hemochromatosis with cirrhosis

Alpha 1-antitrypsin deficiency with cirrhosis

Cirrhosis secondary to other etiologies

(b) High-risk hepatitis B patients

Asian male HBV carrier aged 40 or more

Asian female HBV carrier aged 50 or more

HBV carrier with family history of HCC

African American HBV carrier aged 20 or more

B. Patients with increased risk of HCC for whom surveillance indication is uncertaina

Male hepatitis B carriers younger than 40

Female hepatitis B carriers younger than 50

Hepatitis C with stage 3 fibrosis

NAFLD without cirrhosis

^aBased on the benefit of surveillance if annual HCC incidence is 1.5 % or more for patients with cirrhosis and more than 0.2 % for chronic HBV patients

Recent data from Denmark questioned the use of HCC screening in alcoholic cirrhosis. However, these findings need to be confirmed before the guidelines on screening patients with alcoholic cirrhosis are changed (Jepsen et al. 2012).

Surveillance tests

Alpha-fetoprotein (AFP) is a tumor marker, and a value of 20 ng/mL is 60 % sensitive and 90 % specific for the diagnosis of HCC (Marrero et al. 2003; Saab et al. 2003; Trevisani et al. 2001). Lowering this cutoff value improves the sensitivity but decreases the specificity with false positive results. Descarboxyprothrombin (DCP) or prothrombin induced by vitamin K absence II (PIVKA-II) is another tumor marker reported to be of some use for HCC screening. Data from post hoc analysis on 39 HCC patients from the HALT-C trial using AFP and DCP levels at the time of diagnosis and 12 months prior to diagnosis of HCC concluded that neither test was an optimal method for surveillance (Marrero et al. 2003).

Ultrasonogram

In an Italian cohort of 313 patients followed prospectively, 61 patients developed HCC detected by ultrasonogram (USG) with a sensitivity of 65–80 % and specificity of over 90 % (Bolondi et al. 2001). HCC lesions are commonly described as hypoechoic on ultrasound; however, they could also present as isoechoic with a halo or have a mixed echogenicity (Bolondi et al. 2001). USG is operator dependent and is of limited value in patients with obesity or ascites. In a randomized study on 163 patients with cirrhosis, biannual screening with USG compared to yearly CT scan had better sensitivity (71 vs. 67 %) and similar specificity (97 vs. 94 %) for HCC detection (Pocha et al. 2013). In a retrospective study on cirrhotics, initial USG was 77 % sensitive and 99 % specific for identifying a lesion confirmed to be HCC on the CT scan. The sensitivity of USG combined with AFP >20 ng/mL increased to 87 % with a positive predictive value of 88 % and negative predictive value of 94 % (Welch et al. 2010). Similar data were reported in another retrospective study, but without additional benefit of AFP values on the overall accuracy of the USG (Evan et al. 2014). In a pooled data of 13 studies, USG was 94 % sensitive for all HCC and 63 % sensitive for early-stage HCC. AFP did not provide additional benefit to the accuracy of USG (Singal et al. 2009).

Compliance of physicians with the HCC screening guidelines was assessed in a study on 13,002 veterans from 128 VA medical centers. The study found an underutilization of surveillance with only 42 % of patients with HCV-related cirrhosis receiving 1 or 2 surveillance tests during the first year after diagnosis of cirrhosis. Moreover, routine surveillance was performed in only 12 % of these patients during a follow-up period of 2–4 years (Davila et al. 2011). Therefore, strategies are needed to educate the medical community on the importance of surveillance of high-risk patients.

Surveillance interval

It is recommended that HCC surveillance be performed at an interval of 6 months for patients at risk of HCC (Santagostino et al. 2003; Trevisani et al. 2002). The surveillance interval does not need to be changed for elderly individuals, patients with coinfections, diabetics, decompensated cirrhotics, patients with high HBV DNA levels or with a family history of HCC (Bruix et al. 2011).

Diagnosis of hepatocellular carcinoma

Alpha-fetoprotein

Although a recommended marker for screening, AFP is not very accurate in the diagnosis of HCC (Bruix et al. 2011). Apart from its moderate sensitivity for diagnosis of HCC, elevation of AFP can occur in other conditions such as intrahepatic cholangiocarcinoma, colon cancer, primary gastric carcinoma, teratomas and choriocarcinomas (Adachi et al. 2003; Sato et al. 1994). Nonetheless, when a patient with cirrhosis or advanced liver disease has a liver lesion, an AFP elevated >200 ng/mL is highly suspicious for HCC (Bruix et al. 2011).

Radiological diagnosis

The cornerstone for diagnosis of HCC is a dynamic 4-phase imaging study comprising of an unenhanced phase (obtained at baseline without contrast), an arterial phase (obtained with bolus injection of contrast), a portal venous phase (obtained 35–55 s after initiation of the arterial phase) and a delayed phase (at >120 s after contrast injection).

The normal liver parenchyma receives dual blood supply from the hepatic artery and portal vein. With the development of HCC and arterializations, the lesion loses its portal vein supply and is dependent on the arterial blood supply (Fig. 1). This forms the basis for arterial enhancement (Forner et al. 2008; Levy et al. 2001). During the portal venous phase, the liver receives the venous blood containing the contrast, while the HCC with the arterial supply no longer contains any contrast. This results in the lesion becoming less enhanced during the portal venous phase, also described as the “washout” phase (Forner et al. 2008; Levy et al. 2001). This characteristic appearance of a lesion with arterial enhancement and “washout” on portal venous phase is a classic finding and is diagnostic of HCC (Fig. 2a–c).

Fig. 1.

Algorithm for screening, surveillance and diagnosis of hepatocellular carcinoma

Fig. 2.

Evolutional change in the proportion of blood supply (portal venous supply shown in blue vs. hepatic arterial supply shown in red) with changing liver histology from cirrhosis to regenerative nodule to dysplastic nodule to HCC. Patients with a poorly differentiated HCC are supplied almost entirely by hepatic artery

CT or MR scanning may be used to obtain 4-phase liver imaging. Pooled data from ten studies evaluating CT and nine studies evaluating MR showed better sensitivity with the CT (81 vs. 68 %) and better specificity with the MR (93 vs. 85 %) (Colli et al. 2006). The sensitivity of MR can be increased using diffusion-weighted imaging (DWI) or a hepatobiliary contrast agent, such as Eovist (Lee et al. 2012). Comparison of CT and MR scans reveals that both modalities have about 95 % accuracy for HCC >2 cm. However, for lesions <2 cm, MR was slightly more accurate compared with CT (45–90 vs. 40–55 % accuracy) (Mueller et al. 2003).

Compared with MR, the advantages of CT are as follows: lower cost, less time needed to complete the imaging, quicker assessment and reading of films, less complexity and artifacts, better patient acceptance and no interference with metals. In contrast, the advantages of MR include ability to undergo the imaging with the use of thinner cannula, need for less contrast (10–20 cc for MR scan vs. 1.5 cc/kg body weight for CT scan), safety of use in patients with borderline renal function (estimated GFR 30–60 mL/min) and no radiation exposure to the patient. The choice of test in routine practice may vary based on the center policy, personal preference of the radiologist and the clinical situation. Findings on imaging are described based on the Organ Procurement Transplant Network (OPTN) classification (Table 2).

Table 2.

Organ Procurement Transplant Network (OPTN) classification of liver lesions on CT or MR scan and their management

OPTN	Classification	Action
0	Inadequate study	Repeat the study
1	Negative study	Follow surveillance protocol
2	Benign study	Follow surveillance protocol
3A	Lesion <2 cm not HCC	Repeat study 12 months
3B	Lesion ≥2 cm not HCC	Repeat study 6 months
4A	Lesion <2 cm indeterminate for HCC	Repeat study 6 months
4B	Lesion ≥2 cm indeterminate for HCC	Repeat study 3 months
5A	Lesion <2 cm classic HCC	Early HCC
5B	Lesion ≥2 cm classic HCC	Evaluate for treatment by BCLC staging

Lesions >1 cm with features that are not diagnostic of HCC on one modality can be confirmed using the other modality. If both imaging techniques do not show the classic findings of HCC, tissue diagnosis using liver biopsy should be obtained (Bruix et al. 2011). For lesions <1 cm, the recommendation is repeat liver imaging at 3- to 6-month intervals to look for any interval increase in lesion size (Fig. 3) (Bruix et al. 2011).

Fig. 3.

Classical radiological imaging findings of HCC on dynamic 4-phase CT scan with enhancement of the lesion during the arterial phase, washout of contrast in the portal venous and delayed phases

Liver biopsy

The sensitivity of liver biopsy in detecting HCC is about 90 %. The accuracy is affected by the location of the nodule (Sala et al. 2004). Patients with a negative biopsy require enhanced surveillance with repeat biopsy and/or imaging. The incidence of needle track seeding of the tumor is 2.7 % overall and 0.9 % per year as reported in a meta-analysis of pooled data from seven studies (Silva et al. 2008). In addition to the microscopic criteria, certain histological stains can help differentiating HCC from high-grade dysplastic nodules. The characteristic histological appearance of HCC is positive staining for vascular endothelium (CD34) and negative staining for biliary epithelium (CK7 and CK19) (Park et al. 2007).

Staging of hepatocellular carcinoma

Staging of HCC is complex but essential in determining the treatment modality as well as the prognosis. The Barcelona Clinic Liver Cancer (BCLC) classification is used worldwide and involves the assessment of tumor stage, liver function, performance status and cancer-related symptoms (Bruix and Llovet 2002). This extensively validated system connects the staging and treatment options, providing well laid-out algorithms for managing HCC into four stages (Fig. 4) (Bruix et al. 2011). The Model for End-Stage Liver Disease (MELD) score is a good predictor of LT free mortality in patients with cirrhosis, but not for HCC (Huo et al. 2007). Tumor–node–metastasis (TNM) classification includes evidence of vascular invasion, which is often not obtainable until the availability of the surgical specimen (Vauthey et al. 2002). The Okuda classification does not encompass patients with early or indeterminate HCC (Okuda et al. 1985). Moreover, both TNM and Okuda staging systems do not take into consideration the performance status, an important variable in selecting treatment for HCC.

Fig. 4.

Algorithm for management of HCC based on the Barcelona Classification of Liver Cancer

Treatment of hepatocellular carcinoma

Treatment of HCC is best carried out using a multidisciplinary approach. Tumor resection, ablation and LT are curative treatment options with 5-year survival rates of over 50 % among candidates appropriate for these modalities (El-Serag 2011). Loco-regional therapies (trans-arterial chemoembolization or TACE and radioembolization) and systemic chemotherapy are options for treating patients not amenable to curative options. Patients with metastatic disease, very high tumor burden and poor performance status can only be managed with a palliative approach (Fig. 4) (Bruix et al. 2011). The overall survival and not the disease-free survival, is considered a better endpoint to assess efficacy when various treatment modalities are compared, as the latter is confounded by the progression of underlying cirrhosis, a common underlying condition in HCC patients (Llovet et al. 2008a, b). Although HCC is increasingly diagnosed at an early stage, the use of curative treatments has remained underutilized. In a recent regression analysis of 18 registries from the SEER database (representing 28 % of the US population), only 23 % of cases received potentially curative treatment. In 2000–2005, the rate of treatment with curative intent increased by 17.6 % per year but declined by −2.9 % per year during 2005–2010 (P < 0.001) (Ulahannan et al. 2014).

Surgical resection

About 5 % of HCC patients in the western hemisphere and 40 % in Asia are amenable to surgical resection (El-Serag 2011). Non-cirrhotic patients with a single lesion without any extrahepatic spread are best served by surgical resection. Among cirrhotics, the presence of portal hypertension and elevated serum bilirubin level predict the survival following resection of HCC. In a retrospective analysis of 77 patients undergoing HCC resection, the 5-year survival was 25 % in 27 patients with elevated serum bilirubin (≥1 mg/dL) and portal hypertension; 51 % among 25 patients with either of these two abnormalities; and 74 % among 35 patients without portal hypertension and with normal bilirubin (Fig. 5) (Llovet et al. 1999). Thrombocytopenia (<150,000/cm) is a predictor of portal hypertension, and patients selected for resection should have a normal platelet count (Maithel et al. 2011). Patients without clinical (ascites, variceal hemorrhage and/or hepatic encephalopathy), hematological (platelet count), endoscopic (varices and/or portal hypertensive gastropathy) and imaging evidence of portal hypertension (splenomegaly, dilated portal vein, collaterals and/or varices) should undergo measurement of hepatic venous pressure gradient (HVPG) prior to resection in order to assure that this value is <10 mmHg (Bruix et al. 1996, 2011). A prospective study of 100 patients showed that liver stiffness measured by elastography is as effective as HVPG measurement in predicting decompensation from portal hypertension in cirrhotics (Robic et al. 2011). Thus, liver stiffness could be useful in predicting postoperative decompensation and hepatic failure in patients with HCC undergoing resection.

Fig. 5.

The outcome of HCC treated by resection, with the worst survival in patients with portal hypertension (platelet count <150,000) and high bilirubin (>1.1 mg/dL), the best survival in those without these factors and intermediate survival in patients with one of these factors

About 80 % of liver resections are performed by the laparoscopic route and can achieve negative resection margins in 96 % of patients, with an average hospital stay of 4 days and a 30-day mortality of 0.8 % (Winslow and Hawkins 2013). A systemic review of 152 studies (including two randomized studies and 27 retrospective case studies) showed a median 5 year post-resection overall survival of 67 %, disease-free survival of 37 % and operative mortality rate of 0.7 % (Lim et al. 2012). Pre-resection TACE did not result in any additional postoperative survival benefit (Yamasaki et al. 1996).

MELD score is a powerful predictor of outcomes following liver resection in cirrhotics. For example, PHLF (post-hepatectomy liver failure) occurred in none of the patients with a MELD score of <9, 3.6 % with MELD of 9–10 and 37.5 % with MELD score of >10. In another study, the perioperative mortality for minor (three or less segments) or major (four or more segments) hepatectomy was 29 % in patients with MELD score of >8 and 0 with MELD score of 8 or less (Teh et al. 2005). Similarly, mortality after liver resection was higher for patients with MELD score >8 compared with lower MELD scores (4 vs. 0.6 %, P = 0.004) on analysis of over 1,100 HCC patients undergoing resection between 1991 and 2005 at a single center in Taiwan (Hsu et al. 2009). In yet another study, the perioperative mortality was 19 % with MELD >8 vs. 0 for lower MELD scores (Cucchetti et al. 2012; Delis et al. 2009). The same investigators in Italy reported another study on the concept of conditional survival (sum of the survival from the time of diagnosis and the postsurgical survival) and concluded that a MELD score of <9 achieved the best conditional survival (Cucchetti et al. 2012). In patients with MELD >9, further risk stratification can be made based on the extent of surgery and serum sodium levels (Cescon et al. 2009).

The National Comprehensive Cancer Network recommends that after resection, a CT or MR scan should be performed with AFP estimation every 4 months for 2 years and then every 6 months for another 3 years (Yamasaki et al. 1996). Recurrence after resection can be as high 75 % at 5 years (Adachi et al. 1995; Shirabe et al. 1991). Risk factors for recurrence include vascular invasion and additional tumors found abutting the primary lesion (Adachi et al. 1995; Imamura et al. 2003). Most recurrences are due to vascular invasion and dissemination occurring in the first 2–3 years after resection (Morimoto et al. 2003). A second resection may be an option for these patients if they meet selection criteria, as mentioned above. Transplantation is usually not an option for these patients due to the aggressive nature of these tumors and poor liver function (Minagawa et al. 2003). However, if these high-risk features for recurrence are identified in the resected specimens, these patients could be listed for salvage LT immediately after resection (Hama and Kusano 2005). Salvage LT is also an effective option for managing de novo recurrence of HCC in an underlying cirrhotic liver. Preoperative or postoperative adjuvant treatment modalities of any kind have not resulted in improved survival or reduced recurrence of HCC (Schwartz et al. 2002).

In summary, resection offers the best chance for curative treatment for select HCC patients. The outcome depends on liver function, portal hypertension, tumor characteristics and post-resection hepatic reserve.

Ablative therapies

Ablation is the treatment of choice for patients with early cancer defined as a lesion <2 cm, no extrahepatic spread, no portal vein invasion and performance status of 0 without any cancer-related symptoms (Forner et al. 2012; Livraghi et al. 2008). However, it is usually difficult to diagnose such lesions, since they are generally hypovascular with ill-defined margins and are usually discovered during examination of explants (Kojiro 2004). However, if diagnosed prior to LT, these lesions are best managed by ablation with 5-year DFS of almost 100 % (Forner et al. 2012; Okada 1999). Ablation with necrosis of tumor can be achieved chemically (ethanol, acetic acid or boiling saline) or thermally (cryotherapy, radiofrequency, laser and microwave). Both techniques are usually performed under ultrasound guidance.

Percutaneous ethanol injection (PEI) and radiofrequency ablation (RFA) are commonly used ablative therapies. Four randomized controlled studies involving 652 patients compared radiofrequency ablation to PEI. The results showed a better 3-year survival with the RFA compared with the PEI (63–81 vs. 48–67 %). The benefit of RFA may be substantial in tumors larger than 2 cm (Cho et al. 2009). Complete response rates were similar (90–96 % with RFA and 80–88 % with PEI), but with a fewer number of sessions with RFA as compared to PEI (1.1–2.1 vs. 4.8–6.4) (Cho et al. 2009). Based on these data, RFA is considered the ablative modality of choice over PEI (Bruix et al. 2011; Cho et al. 2009).

A follow-up dynamic CT scan is recommended after 1 month to assess treatment efficacy, as determined by the presence of tumor necrosis or absence of tumor enhancement (Bruix et al. 2001). Treatment efficacy is reported as about 80 % complete response for tumors <3 cm and about 50 % for lesions 3–5 cm in size, with 40–70 % survival at 5 years (Llovet and Bruix 2008; Orlando et al. 2009). RFA is associated with about 6 % incidence of major complications such as pain, liver abscess, pleural effusion, bowel perforation and hemoperitoneum (Giorgio et al. 2005). Tumor seeding of the needle tract can occur with RFA, although rare, and is related to factors such as subcapsular location of the tumor, poor histological differentiation and elevated baseline AFP (Llovet et al. 2001). There is no defined follow-up interval; however, it is suggested that dynamic imaging should be done every 3–4 months for the first 2 years. The interval can be increased if there is no evidence of recurrence.

Whether resection or ablation is considered as the first-line treatment has been addressed in two randomized studies. In the first study, 140 patients with HCC <3 cm, with Child–Pugh stages A and B were randomized to resection (n = 70) or ablation (n = 70). There was no survival difference with 50–55 % survival with both modalities (Chen et al. 2006). Another study randomized 160 HCC patients to resection or ablation (n = 80 in each arm) and found similar 3-year survival of 70–80 % (Feng et al. 2012). In a meta-analysis of fifteen studies, including the above-mentioned two randomized studies, comparing resection (n = 1,233) and ablation (n = 1,302), the mortality was lower by about 40–50 % at 1, 3 and 5 years of follow-up among resected patients. However, the complication rate was about sixfold higher with resection compared to ablation (Xu et al. 2012). Similar data were reported in a national survey on more than 54,000 patients with HCC treated in Japan (Sato et al. 2012). Based on these data, ablation may be considered as the first-line option for small HCC (Bruix et al. 2011).

Liver transplantation

The presence of HCC was considered a contraindication for LT until 1990s. However, data from observational studies showed that the post-LT outcome of patients with incidentally detected HCC in explants was similar to patients without detected HCC. Later, in a retrospective study on 48 HCC patients (single lesion <5 cm or up to 3 lesions each <3 cm) undergoing LT, the actuarial survival was 75 % at 4 years with only 8 % recurrence of HCC. Based on this landmark study, LT became an established treatment option for patients with HCC for solitary lesions <5 cm in size or up to 3 lesions each <3 cm each (Milan criteria) (Mazzaferro et al. 1996). Patients should have abdominal and chest CT scans as recommended by UNOS to exclude extrahepatic spread of the cancer (Mazzaferro et al. 1996).

Patients with HCC with a low MELD score have a risk of tumor progression while waiting for LT leading to death or disease progression which may exclude them from receiving an organ. On the other hand, HCC patients with higher MELD have been shown to have worse post-transplant survival compared to comparable MELD in non-HCC patients (Ioannou et al. 2008). Therefore, in February 2003, UNOS accepted a policy of awarding MELD exception points for HCC patients within the Milan criteria. This resulted in a considerable increase in the proportion of transplants performed for HCC from 4.6 % during 1997–2002 to 26 % during 2002–2007 (Ioannou et al. 2008). Therefore, in March 2005, this policy was modified to award 22 MELD exception points instead of 24 points to patients with HCC to more accurately reflect their risk of dying (Wiesner et al. 2004). Even with this policy, the odds of a patient with HCC receiving the organ remain substantially higher compared with a non-HCC patient (Massie et al. 2011; Washburn et al. 2010).

On the other hand, about 29 % of patients with HCC are dropped from the waiting list due to tumor progression suggesting consideration of other factors such as number of tumors, alpha-fetoprotein levels, tumor biology and biological or calculated MELD score for transplanting HCC patients (Ioannou et al. 2008; Park et al. 2011). However, many of these factors are also associated with tumor recurrence after transplantation and a higher probability of dropout from the waiting list is directly related to a higher risk of recurrence of HCC after LT (Cucchetti et al. 2011). Hence, a balanced approach is needed to optimize the use of livers for transplanting HCC patients aiming at maintaining post-transplant survival as well as optimizing the chance of receiving a transplant. In this regard, the concept of dropout equivalent MELD (deMELD) points has been introduced taking into consideration the risk of dropout from the waiting list based on MELD score and other HCC characteristics. For example, two patients with similar MELD scores may be given different exception points based on their dropout risk from the waiting list after considering factors such as the number of tumors, maximum tumor size, age of the patient, etiology of the liver disease and AFP levels (Toso et al. 2012).

To avoid delisting from the transplant list, it is recommended that patients with HCC should be transplanted within 6 months of listing. It has been observed that the dropout risk exceeds 25 % if the wait time is more than 12 months, lowering the intent-to-treat survival from 75 % to below 60 % at 5 years (Yao et al. 2002). Preoperative treatment of HCC using RFA or TACE is recommended if the waiting period on the transplant list is longer than 6 months (Majno et al. 1997). Other options for transplanting HCC patients especially when the wait time is high are the use of marginal donors such as donor after cardiac death, domino transplants and the use of live donors (LDLT). These donors also allow for transplanting patients outside the Milan criteria, although there are no firm recommendations in this regard in routine clinical practice (Pomfret et al. 2010). The 5-year survival rate for living donor liver transplant in HCC was as high as 76 % when the selection criteria are broadened beyond the Milan criteria. Further, LDLT was cost effective among patients with wait period over 7 months (Sarasin et al. 2001). However, one should take into account the morbidity of 20–30 % for the donors related to biliary leak, bleeding and infection. Further, there is a potential for donor mortality of about 0.3–0.5 % (Trotter et al. 2002). In a balanced approach, one should avoid including aggressive tumors on the transplant list as they have a high risk of recurrence after LT. In this regard, it is suggested that one should wait for at least 3 months before listing to avoid aggressive tumors from being listed (Ioannou et al. 2008; Pomfret et al. 2010).

The current management of immunosuppression in the post-LT period does not differ from non-HCC patients. The mTOR inhibitors (mTORis), such as sirolimus and everolimus, have anti-proliferative properties and may have beneficial effects in reducing HCC recurrence and improving post-LT survival (Toso et al. 2010). In a meta-analysis of five studies involving 2,950 patients after LT, the use of sirolimus-based regimen compared with non-sirolimus-based regimens was associated with better overall survival and reduced HCC recurrence at 1 and 5 years without any differences in post-transplant complications (Liang et al. 2012). In a recent analysis of 42 studies, mTOR inhibitor-based regimen was associated with lower HCC recurrence compared with CNI-based regimens (8 vs. 13.8 %, P < 0.001). The everolimus-based regimen had the lowest recurrence rate compared with sirolimus- and CNI-based regimens (4.1 vs. 10.5 vs. 13.8 %, respectively, P < 0.05) (Cholongitas et al. 2014). The results of an open-labeled, prospective, randomized study comparing sirolimus-based regimens to non-sirolimus-based regimens in patients undergoing OLT for HCC are eagerly awaited (http://www.clinicaltrials.gov/NCT00355862.133) (Schnitzbauer 2010). HCC recurs after LT recurs in about 8–22 % of patients (Zimmerman et al. 2008). Although strong data on guidelines for surveillance after LT are lacking, expert opinion and National Comprehensive Cancer Network suggest CT or MR scan every 3–6 months with AFP for 2 years and then annually until 5 years of follow-up (Clavien et al. 2012; “National Comprehensive Cancer Network Guidelines”).

Trans-arterial chemoembolization

Trans-arterial embolization aims at obliterating the blood supply to the HCC. The procedure may be performed with chemotherapeutic administration followed by obliteration of the arterial supply i.e., TACE. Two commonly used chemotherapeutic agents are adriamycin and cisplatin (Bruix et al. 2004). Previously, the conventional technique involved mixing of the drug with lipiodol. Being oil based, this was associated with separation of the drug leading to systemic side effects and less delivery of the drug into the lesion (Malagari et al. 2008; Varela et al. 2007).

The lack of portal vein blood flow is a relative contraindication for this treatment. TACE should also be avoided in patients with advanced liver disease due to the increased risk of liver failure and death. Patients should be well hydrated prior to the procedure to prevent contrast-induced nephropathy, especially in those at high risk. The procedure can be done on an outpatient basis in most patients except those with thrombocytopenia (<50,000/cm) and/or coagulation abnormalities (INR >1.5), who may require admission for administration of relevant blood products prior to the procedure. The routine use of antibiotic prophylaxis is not recommended. Post-embolization syndrome secondary to tumor necrosis with the development of abdominal pain, fever and ileus may occur in about 30–50 % patients after the procedure (Castells et al. 1995). Rarely, the development of a liver abscess has been reported following TACE (Hama and Kusano 2005; Huang et al. 2003).

Treatment response should be assessed by reimaging at 1–2 months after the procedure. Tumor response is measured by the European Association for Study of Liver Diseases criteria by assessment of the extent of tumor necrosis (the endpoint of loco-regional therapy) or by the modified RECIST criteria by assessing the area of remaining viable tumor (Fig. 6) (Georgiades et al. 2012). A reasonable response is defined as necrosis of >50 % of the lesion and is achieved in about 50 % of patients. Complete response is very unusual and is seen in only about 2 % patients (Bruix et al. 2004).

Fig. 6.

CT scan imaging showing pre- and post-TACE treatment of the HCC. TACE achieves necrosis of the tumor as shown by reduced area of arterial enhancement (mRECIST criteria)

Pooled data obtained from seven randomized controlled trials involving 545 patients showed improved 2-year survival with TACE (55 vs. 35 %, P < 0.05) with objective response rate of about 35 % (Llovet and Bruix 2003). Based on these data, it can be concluded that TACE is an effective palliative treatment option for patients who are not candidates for curative treatment options and do not have portal vein involvement, extrahepatic spread and have performance status of 0 (Bruix et al. 2004). Pooled data on the combination of TACE and RFA have shown improved survival rates, but more randomized studies are needed to address this question before combined therapy can be recommended for routine clinical practice (Yan et al. 2012).

Drug-eluting bead-trans-arterial chemoembolization (DEB-TACE) consists of embolization with beads loaded with sustained-release chemotherapeutic agents, most commonly doxorubicin. This variation of TACE was developed in an effort to limit the systemic and local side effects of TACE and to increase the concentration of drug delivered to the tumor site (Malagari et al. 2008; Varela et al. 2007). A number of studies have shown DEB-TACE to be as effective as TACE. In a multicenter, prospective randomized controlled trial comparing DEB-TACE (n = 93) and TACE (n = 108), DEB-TACE tended to achieve greater objective response rates (52 vs. 44 %, P = 0.11) and disease control at 6 months (63 vs. 52 %, P = 0.11). Among patients with Child–Pugh B stage, the objective response was statistically significant with DEB-TACE compared with TACE (52.4 vs. 34.7 %, P = 0.038). Complication rates were similar to the two techniques 20 % DEB-TACE patients having 28 events versus 19 % TACE patients reporting 24 events (P = 0.86). DEB-TACE was associated with lower liver toxicity: The mean maximum rise in AST and ALT was 41 % with DEB-TACE versus 50 % less with TACE (P < 0.001) (Lammer et al. 2010). A meta-analysis of seven studies involving 700 patients demonstrated about twofold higher tumor necrosis with 36 % lower mortality with DEB-TACE compared with TACE (Huang et al. 2014). DEB-TACE may also have a role as a therapeutic bridge to LT (Nicolini et al. 2013).

Radioembolization

Radioembolization is a type of loco-regional therapy using radiation-emitting beads to embolize the vessels supplying the tumor. Yttrium-90 (Y90)-labeled microspheres are used as the radiation-emitting beads in the USA. Studies suggest that radioembolization may be an effective treatment for unresectable HCC. In a recent prospective longitudinal study on 291 patients with unresectable HCC treated with Y90-radioembolization, the overall response rates regardless of the disease stage were 42 % by modified RECIST criteria and 57 % by EASL criteria. Of note, patients with Child–Pugh B stage with portal vein thrombosis (PVT) and/or with metastatic disease did poorly post Y90-radioembolization. The most common adverse effect was fatigue, seen in >50 % of patients (Salem et al. 2010). In a retrospective cohort study of 463 HCC patients treated with either Y90-radioembolization or TACE, the time to progression (TTP) was longer after Y90-radioembolization (13.3 vs. 8.4 months, P = 0.046) with a similar median survival time (20.5 vs. 17.4 months, P = 0.23). Y90-radioembolization was associated with less frequent abdominal pain and elevations in transaminase levels, compared with TACE (P < 0.05) (Salem et al. 2011). Another study showed that the quality of life scores tended to be higher among patients treated with Y90-radioembolization compared with TACE (Salem et al. 2013). NCCN suggests imaging every 3–6 months for 2 years and then annually. It also recommends serum AFP (if initially elevated) every 3 months for 2 years and then every 6 months (“National Comprehensive Cancer Network Guidelines”).

Due to a lack of randomized trials on the efficacy and safety of radioembolization, the AASLD does not recommend radioembolization outside of clinical trials. However, radioembolization may be useful among patients with unresectable HCC with PVT (Mazzaferro et al. 2013).

Systemic chemotherapy

Currently, sorafenib (a multikinase inhibitor acting on the vascular endothelial growth factor receptor) is the only systemic chemotherapeutic agent approved by the FDA for treating HCC not amenable to curative treatments and TACE (Fig. 4). In the pivotal double-blind placebo-controlled randomized trial on 602 treatment-naïve patients with advanced HCC, sorafenib improved survival compared to placebo (median survival time: 10.7 vs. 7.9 months). The incidence of adverse effects with sorafenib was 80 %, leading to treatment interruption in 38 % and permanent drug discontinuation in 11 % of patients. The most frequent adverse events were gastrointestinal, constitutional and dermatological, occurring during the first 2–6 weeks of treatment (Llovet et al. 2008a, b).

In a recent meta-analysis of five randomized controlled trials on 1,462 patients with unresectable HCC, sorafenib use compared to placebo improved the disease control rate (RR 1.85, 95 % CI 1.55–2.20, P < 0.001), decreased tumor progression (HR 0.61, 95 % CI 0.51–0.73, P < 0.001) and reduced mortality (HR 0.71, 95 % CI 0.56–0.89, P < 0.001). Subgroup analysis showed results to be unaffected by HCC etiology, performance status and the BCLC stage (Shen et al. 2013).

With the limited survival benefit and significant adverse effects with sorafenib, other drugs are being investigated. A randomized double-blind phase 3 trial on 1,155 patients comparing sorafenib with brivanib (a tyrosine kinase inhibitor) as first-line therapy showed that brivanib did not improve median overall survival compared with sorafenib (9.5 vs. 9.9 months) (Johnson et al. 2013). Another randomized double-blind phase 3 trial on 395 patients with intolerance or refractory to sorafenib treatment, brivanib compared to placebo improved tumor response as assessed by modified RECIST criteria (10 vs. 2 %). However, the median survival did not show any difference (9.4 vs. 8.2 months, HR 0.89, 95 % CI 0.69–1.15, P = 0.33) (Llovet et al. 2013).

In an open-labeled, phase 2 randomized study on 31 patients, who failed sorafenib therapy, regorafenib (multikinase inhibitor) provided promising results with median overall survival of 13.8 months (Bruix et al. 2013). A phase 3 placebo-controlled randomized study is currently ongoing examining the benefit of regorafenib in patients with advanced HCC who have failed sorafenib (http://www.clinicaltrials.gov/NCT01774344). The study is estimated to recruit 530 patients and expected to be completed in 2016.

Several other drugs including tamoxifen, interferon, octreotide, anti-androgen compounds, systemic chemotherapy and intra-arterial injection of chemotherapeutic agents have been tried but have not been found to be effective and are not recommended (Chow et al. 2002; Llovet et al. 2000; Manesis et al. 1995; Nerenstone et al. 1988; Nowak et al. 2005; Raoul et al. 1994).

Combination of TACE and sorafenib

If a patient with unresectable HCC has poor performance status, PV invasion or extrahepatic tumor spread, the treatment option suddenly shifts from a more optimistic choice of TACE to the final treatment option of sorafenib (Fig. 4). There is a growing interest in the combination of local and systemic therapies especially for managing these patients. The rationale of combining TACE and sorafenib is based on the findings of an upregulation of hypoxia-induced factor, tumor metabolism and angiogenesis with the use of TACE. The vascular endothelium growth factor (VEGF) levels have been shown to increase and predict survival after TACE (Sergio et al. 2008).

A recent meta-analysis of six heterogeneous studies on 1,254 patients receiving combination therapy or TACE alone showed TACE combined with sorafenib to be better in lowering the mortality by 35 % and tumor response rate by 6 % without any effect on the progression-free survival. Grade III/IV adverse reaction was more common with the combination treatment as compared to TACE alone (Zhang et al. 2014). Data are also emerging on combining TACE with other systemic chemotherapeutic agents such as brivanib (Kudo et al. 2014).

Currently, outside of clinical trials, the combination approach is not recommended in routine practice. The results of an ongoing phase III STORM (Sorafenib as Adjuvant Treatment in the Prevention of Recurrence of HCC) trial are eagerly awaited and may provide an answer to this question (http://clinicaltrials.gov/show/NCT00692770).

Palliative care

Patients with advanced disease (Child–Pugh stage C), with multiple comorbidities and with extremely poor performance status are not candidates for any of the specific HCC treatments (Fig. 4). These patients are provided palliative care with focus on symptom control and comfort care measures. There has been interest in developing quality of life scores that might predict who will benefit from therapy and who will benefit from palliative care at the time of diagnosis. Scores with better discriminatory ability are still needed (Diouf et al. 2013).

Future management modalities

Many newer agents targeting specific oncogenes are being investigated. These include (a) imatinib (bcr-abl oncogene), (b) erlotinib, gefitinib or lapatinib (target epidermal growth factor), (c) bevacizumab (vascular endothelial growth factor monoclonal antibody), (d) sunitinib and brivanib (target multiple pathways including fibroblast growth factor), (e) everolimus and sirolimus (mTOR inhibitors), (f) perifosine (phosphatidyl inositol-3 kinase AkT) and (g) drugs targeting retinoic acid receptor (Porta and Paglino 2010).

In summary, the prognosis of HCC has changed substantially from an inevitable death sentence to a potentially curable disease, mainly due to its early detection. In spite of the huge progress in the diagnostic approach to HCC, its incidence continues to rise. Hence, strategies are urgently needed to improve the utilization of the currently available screening modalities and of curative treatment options in the management in order to improve survival of patients with HCC.