SEARCH

SEARCH BY CITATION

Keywords:

  • cirrhosis;
  • computerized tomography;
  • HCC;
  • magnetic resonance;
  • screening abdominal ultrasounds

Abstract

  1. Top of page
  2. Abstract
  3. The population to be tested
  4. The surveillance strategies
  5. The recall policy
  6. Is liver-related mortality reduced by screening?
  7. Is surveillance in patients with cirrhosis cost-effective?
  8. Conclusions
  9. Conflicts of interest
  10. References

Early diagnosis of hepatocellular carcinoma (HCC) is feasible because HCC develops in the background of well-known, readily identifiable and potentially avoidable environmental risk factors. According to the American Association for the Study of the Liver Diseases and the European Association for the Study of the Liver, patients with cirrhosis and carriers of chronic viral hepatitis are the target of surveillance to be investigated with abdominal ultrasounds (US) every 6 or 12 months. The diagnostic confirmation of a ≥2 cm nodule in patients with cirrhosis detected during surveillance is possible with any imaging technique among second-generation contrast US, contrast computed tomography and gadolinium-contrast magnetic resonance imaging. HCC shows an early hyperenhanced arterial vascularization, followed by enhanced hypoattenuation (wash-out) in the late phase of imaging. In patients with a ≤2 cm nodule, two imaging techniques are required for the final diagnosis, which, however, have a relatively low diagnostic sensitivity (33%). Nodules with negative imaging findings need to be investigated further with an echo-guided liver biopsy or enhanced follow-up with imaging (every 3 months) to reach a final diagnosis. The cost-effectiveness ratio of surveillance depends on multiple factors, like HCC incidence, the cost and accuracy of diagnostic tests and the costs and outcome of the therapeutic interventions.

Hepatocellular carcinoma (HCC) is unique in that it occurs in the context of well-recognized and potentially preventable environmental risk factors. Surveillance improves the detection and treatment of small tumours and, following standardization of the diagnostic tests and recall procedures, it has become the standard of care for patients with an HCC (1, 2). The guidelines generated by a panel of the European Association for the Study of the Liver (EASL) experts in 2001 have been largely superceded by the American Association for the Study of the Liver Diseases (AASLD) guidelines of 2005.

Surveillance for HCC is feasible because the target population is readily identifiable, and the test adopted [abdominal ultrasound (US)] is user friendly and acceptable to the population, while it has a high diagnostic accuracy. The diagnostic tests and recall procedures have been standardized, while early cancer definition also incorporates suitability into potentially curative treatments. The definition of early cancer has been evolving during the past decades to incorporate the term ‘very early tumour’, which indicates a <2 cm nodule with an indistinct nodular pattern and a hypoattenuated vascular pattern on contrast imaging (3–6). This tumour has a better prognosis than equally small HCCs, which have a distinct nodular pattern, but are hypervascular on imaging and infiltrating at histology.

The population to be tested

  1. Top of page
  2. Abstract
  3. The population to be tested
  4. The surveillance strategies
  5. The recall policy
  6. Is liver-related mortality reduced by screening?
  7. Is surveillance in patients with cirrhosis cost-effective?
  8. Conclusions
  9. Conflicts of interest
  10. References

Patients with cirrhosis and carriers of hepatitis B virus (HBV) are the target population for surveillance (Table 1) (1, 2, 7). Cirrhosis is the leading cause of HCC worldwide, the risk of HCC increasing in parallel with disease severity. The annual rate of conversion of virus-related cirrhosis to liver cancer is approximately 3%, however, with important regional variations (from 1.5–2% in Europe to 3–8% in Asia) (1). HCC risk can further increase in the presence of cofactors known to accelerate progression of HCC like aflatoxin B1 in HBV carriers, and alcohol consumption, iron and overweight in hepatitis C virus (HCV) carriers (8). In many developed countries, diabetes and non-alcoholic steatohepatitis have emerged as important risk factors for cryptogenic cirrhosis and HCC in patients lacking HBV or HCV. Although the transition from cirrhosis to HCC is accelerated in livers expressing histopathological markers of increased cell proliferation or cell dysplasia, screening should not be restricted to patients with histological markers of increased risk until the cost-effectiveness of this policy is demonstrated. Age cut-offs have also been indicated for starting surveillance of a selected population of patients like Africans and Asians, because the risk of HCC in these patients often starts earlier than in other populations as a result of early intrafamiliar infection with HBV (9). Theoretically, surveillance programmes could be improved by patient stratification by clinical and histological scores, but the cost-effectiveness of this policy is unknown. By combining sex, age and serum α-foetoprotein (AFP) levels, patients with compensated cirrhosis can be stratified into categories at increasing risk for the tumour. The lowest risk group (1.5% rate per year) included women younger than 53 years of age with normal AFP levels compared with men older than 53 years of age with high AFP levels, who are at the highest risk (10%) for developing HCC (10). In alcoholic patients (11), sex, age, disease severity and liver cell dysplasia identified a subset of patients with an exceedingly high (72%) risk of developing HCC during a follow-up of 3 years. In Spain, similar criteria of disease aetiology and severity (excluding cell dysplasia) led to the identification of patients at a high risk (30.1%) and those at a low risk (2.3%) of developing HCC during 4 years of surveillance (12). Patients with severe comorbidities and those with advanced liver disease (Child–Pugh C class patients), who do not fit the criteria for curative therapies, should not enter into surveillance. Moreover, surveillance programmes should be restricted to individuals below 30 years and those above 75 years of age, because of the low risk of cancer in younger people, while the older ones would not have significant benefit if diagnosed with an HCC. However, these criteria for selection are not evidence based and may not be suitable for all geographical areas. Importantly, patient stratification by risk level does not change the entry of the patients into surveillance, because the growth rate of tumours is the only variable that influences the outcome of screening.

Table 1.   Groups of patients for whom surveillance is recommended by the European (European Association for the Study of the Liver), American (American Association for the Study of the Liver Diseases) and Japanese (The Japan Society of Hepatology) associations for the study of the liver
GuidelinesTarget population
Chronic hepatitis B or CCirrhosis
  1. AASLD, American Association for the Study of the Liver Diseases; A1AT, α1-antitrypsin; AIH, autoimmune hepatitis; ALT, alanine aminotransferase; EASL, European Association for the Study of the Liver; HBV, hepatitis B virus; HCV, hepatitis C virus; JSH, The Japan Society of Hepatology; LT, liver transplantation; NASH, non-alcoholic steatohepatitis.

EASL 2001 (2)HBV: not specified HCV: histological transition to cirrhosisChild–Pugh A and B Child–Pugh C if LT available
AASLD 2005 (1)HBV: all cirrhotics  non-cirrhotics: ALT+DNA+specific age cut-offs for ethnic groupsAll aetiologies: A1AT, AIH, NASH
JSH 2008 (7)High risk Increasing risk: sex, age, alcoholVery high risk: HBV/HCV

The surveillance strategies

  1. Top of page
  2. Abstract
  3. The population to be tested
  4. The surveillance strategies
  5. The recall policy
  6. Is liver-related mortality reduced by screening?
  7. Is surveillance in patients with cirrhosis cost-effective?
  8. Conclusions
  9. Conflicts of interest
  10. References

Ultrasound is the method of choice for screening, because it has adequate sensitivity, specificity, positive and negative predictive values (2). Based on studies of tumour volume doubling time, a duration of 6 months has been selected by most experts in the West as the ideal interval of screening with US, whereas in the Far East, a 3-month interval for screening is largely adopted. In Italy, a study in haemophiliacs chronically infected with HCV demonstrated similar efficacies for 6- and 12-month intervals of screening in the identification of potentially curable HCC (13). In France, a multicentre study in patients with alcoholic and hepatitis C-related cirrhosis showed that increasing the frequency of US examinations every 3 months did not increase the accuracy of screening, whereas it resulted in increased costs to treat patients with false-positive results (14).

The serum assay AFP is no longer considered for screening by AASLD and EASL because of the high rates of false-positive and false-negative results in patients with chronic liver disease. Whether other serum markers like des-γ-carboxyprothrombin and fucosylated AFP should be incorporated into surveillance for HCC is being debated (1).

The recall policy

  1. Top of page
  2. Abstract
  3. The population to be tested
  4. The surveillance strategies
  5. The recall policy
  6. Is liver-related mortality reduced by screening?
  7. Is surveillance in patients with cirrhosis cost-effective?
  8. Conclusions
  9. Conflicts of interest
  10. References

In cirrhotic patients with a liver nodule detected during surveillance, HCC can be diagnosed by imaging techniques without the need for a histological examination (Table 2) (15). Computed tomography scan, magnetic resonance imaging and contrast-enhanced US serve the purpose whenever arterial enhancement of the nodule is demonstrated and followed by wash-out of contrast in the delayed venous phase (16). The accuracy of radiological diagnosis, however, is tumour size dependent: the diagnosis is attained in all nodules >2 cm in diameter with one imaging technique only, whereas nodules between 5 and 20 mm in size escape diagnosis in two-third of the cases despite application of two imaging techniques (15). Liver nodules not accurately diagnosed by imaging will require histological examination with a liver biopsy or close monitoring with US carried out at 3-month intervals (enhanced follow-up). Radiological diagnosis of HCC is challenged by false-positive results in the presence of artero-venous shunts and macroregenerative nodules with dysplastic liver cells. Currently, dysplastic macronodules can be discriminated from early tumours by morphological criteria based on exclusion of microscopic stromal invasion (17). Immunostaining for glypican-3, and structural and functional analysis of the genetic profile of the nodules may also distinguish between macronodules and tumours (18). Falsely negative nodules at contrast imaging are also a problem when searching for very early tumours, accounting for approximately 20% of all HCCs 1–2 cm in size (19).

Table 2.   Radiological diagnosis of hepatocellular carcinoma in patients with cirrhosis according to the guidelines of the European (European Association for the Study of the Liver) and American (American Association for the Study of the Liver Diseases) associations for the study of the liver (1, 2, 7)
  1. CT, computed tomography; MRI, magnetic resonance imaging; US, ultrasound.

Imaging techniquesContrast-enhanced US, contrast-enhanced spiral CT and gadolinium-enhanced MRI
 1–2 cm noduleTwo imaging techniques showing a hyperenhanced nodule in the arterial phase and a hypoenhanced nodule in the portal phase (wash-out)
>2 cm noduleOne imaging technique
Prospective validation (15)89 patients with a 5–20 mm nodule
 Sensitivity33%
 Specificity100%
 Positive prediction value100%
 Negative prediction value42%

Is liver-related mortality reduced by screening?

  1. Top of page
  2. Abstract
  3. The population to be tested
  4. The surveillance strategies
  5. The recall policy
  6. Is liver-related mortality reduced by screening?
  7. Is surveillance in patients with cirrhosis cost-effective?
  8. Conclusions
  9. Conflicts of interest
  10. References

In the last 5-year period, two-thirds of all patients presented with a TNM I/II tumour compared with the 1980s, when only a minority of the patients had a small HCC (20). In a national survey of 554 members of the AASLD, the vast majority of responders indicated that they routinely screened patients with cirrhosis using US (21).

A surveillance programme of semi-annual determinations of AFP in HBV carriers in Alaska led to the identification of curable HCC in 40% of the affected patients, a fact that was perceived as beneficial because before the AFP screening programme the case-fatality rate for HCC in Alaskan natives was 100%, with an average survival of 3 months (22). A randomized-controlled study conducted in Shanghai area using abdominal US and serum AFP every 6 months among individuals with chronic hepatitis and other risks for HCC showed a reduction in the mortality rates in a screened vs an unscreened population (83.2 vs 131.5 per 100 000 inhabitants) (23). The study, however, has limitations in the design and the way it was conducted, because the proportion of patients with cirrhosis is unknown, transplantation was not included among the radical therapies and the compliance of the population to the programme was suboptimal (58%) (Table 3).

Table 3.   A population-based screening for hepatocellular carcinoma: the importance of early diagnosis for improving liver-related mortality (23)
FindingsScreened group (pp × yr=38 444)Control group (pp × yr=41 077)
  1. CI, confidence interval; HCC, hepatocellular carcinoma.

HCC occurrence
 Cases8667
 Early cancer390
 Total incidence (per 100 000)223.7163.1
 Rate ratio (95% CI)1.37 (0.99, 1.89)Reference
Deaths from HCC
 Deaths3254
 Total mortality (per 100 000)83.2131.5
 Rate ratio (95% CI)  

The liver-specific mortality rates were reduced in cirrhotic patients with an HCC detected during surveillance, likely as a consequence of both improvements in detection and treatment. In a re-analysis of 112 patients with an HCC detected during surveillance, higher survival rates were demonstrated for patients who were treated for a liver cancer that developed during the last 5 years of surveillance than previously (90 vs 55%, P=0.0009) (Table 4) (10). Increased survivals could confidently be attributed to a significant reduction in the mortality rates of treated patients (from 34 to 5%, P=0.003), owing to a wider application of curative treatments and improved selection of patients undergoing surgical or ablative treatments. In Taiwan, between 1989 and 1998, there has been a significant increase in survival among 3345 patients with an HCC in the last 5-year period (from 29 to 35%) that was only in part (34%) because of advancement in medical care, but mostly (66%) attributable to early detection (24).

Table 4.   A clinic-based surveillance for hepatocellular carcinoma in 447 cirrhotics: the importance of treatment refinement for improving liver-related mortality (10)
Outcomes1987–19911992–19961997–2001P
  1. HCC, hepatocellular carcinoma.

HCC, no.523723 
HCC size (cm)3.7 (1.5–8)3.0 (1.5–6.0)2.2 (1.4–3.1)0.02
Radical treatments (%)2838430.02
Mortality in treated (%)342850.024
Mortality in untreated (%)6910092NS
Overall mortality (%)4537100.0009

Is surveillance in patients with cirrhosis cost-effective?

  1. Top of page
  2. Abstract
  3. The population to be tested
  4. The surveillance strategies
  5. The recall policy
  6. Is liver-related mortality reduced by screening?
  7. Is surveillance in patients with cirrhosis cost-effective?
  8. Conclusions
  9. Conflicts of interest
  10. References

Moderators of treatment outcome, like compliance, heterogeneity and aetiology of liver disease and treatment effectiveness, may have an impact on the cost-effectiveness of surveillance. Patient compliance is an important variable of surveillance, as 1–15% of all patients failed to comply with clinic-based surveillance programmes in Europe and Japan [reviewed by Thompson Coon et al. (25)], compared with the 42% of the population enrolled in a study in Shanghai (23). The cost-effectiveness of screening is largely also influenced by the heterogeneity of the natural history of tumour disease, the annual risk of HCC ranging from 1% in HBV patients to 5–8% in patients with compensated cirrhosis caused by HCV (8). Because the cost for the tumour detected is inversely proportional to the tumour incidence, screening is considered valuable in populations with ≥1.5% incidence of HCC. However, in population-based screening programmes with an HCC incidence lower than 1.5%, the low cost-effectiveness of surveillance is counter-balanced by the high numbers of targeted individuals with a preserved liver function, who have more chances of receiving curative treatments compared with patients enrolled in clinic-based surveillance programmes. Finally, because the cost-effectiveness of a surveillance largely depends on the outcome of treatments, studies like the Shanghai programme lacking access to liver transplantation do not help in predicting the cost-effectiveness of surveillance in the West, where liver transplantation is the first option for selected patients with cirrhosis and a small tumour. In the absence of randomized trials, decision analysis (Markov model) has provided the best means to estimate the cost-effectiveness of HCC surveillance in this patient population. A cost-effectiveness ratio represents the cost for life-year saved compared with a cost–utility ratio, which describes the cost per quality-adjusted life-year (QALY) saved. Surveillance programmes in which cost–utility ratios are measured at <US$50 000 × QALY are thought to be cost-effective (Table 5) (26).

Table 5.   Markov decision models to simulate the cost–utility ratio of surveillance according to the American Association for the Study of the Liver Diseases/European Association for the Study of the Liver (25, 26)
StudyAetiology of cirrhosisIncremental cost–utility ratio (US$/ QALY)Prerequisite
  • Surveillance is thought to be cost-effective in the presence of a willingness to pay of ≤US$50 000/QALY.

  • *

    Referenced by Thompson Coon et al. (25).

  • HCC, hepatocellular carcinoma; HCV, hepatitis C virus; OLT, orthotopic liver transplantation; QALY, quality-adjusted life-year; US, ultrasound.

Sarasin et al. (1996)*Mixed48 29360% survival 3 years after resection
Everson et al. (2000)*Mixed35 0002.5% HCC × year
Saab et al. (2003)*Wait list74 000 
Arguedas et al. (2003)*HCV26 68950 years old eligible for OLT
Lin et al. (2004)*HCV73 789 
Patel et al. (2005)*HCV26 100Hepatic resection
 46 700Cadaveric liver transplant
 50 400Living donor liver transplant
Thompson Coon et al. (25)Mixed£31 900Most alcohol-related
Anderson et al. (26)Mixed30 700US alone

Conclusions

  1. Top of page
  2. Abstract
  3. The population to be tested
  4. The surveillance strategies
  5. The recall policy
  6. Is liver-related mortality reduced by screening?
  7. Is surveillance in patients with cirrhosis cost-effective?
  8. Conclusions
  9. Conflicts of interest
  10. References

In Europe and Japan, screening has become the standard of care for patients at risk of HCC because it improves the treatment of HCC. Indeed, the disparity in outcomes between patients diagnosed with an early HCC compared with those with a more advanced tumour strongly supports screening for HCC.

To widen the application of screening, which is restricted to hospital and outpatient facilities, serum biomarkers with high diagnostic accuracy for HCC need to be developed to aid in the implementation of community-based screening programmes. Molecular imaging techniques might improve the radiological diagnosis of very early tumours, thus further improving the treatment outcome of patients under surveillance.

References

  1. Top of page
  2. Abstract
  3. The population to be tested
  4. The surveillance strategies
  5. The recall policy
  6. Is liver-related mortality reduced by screening?
  7. Is surveillance in patients with cirrhosis cost-effective?
  8. Conclusions
  9. Conflicts of interest
  10. References