Benefits

Based on fair evidence, screening of persons at elevated risk does not result in a decrease in mortality from hepatocellular cancer.

Magnitude of Effect: No reduction in mortality.

• Study Design: Randomized controlled trials.
• Internal Validity: Fair.
• Consistency: Multiple studies, large number of participants.
• External Validity: Fair.

Harms

Based on fair evidence, screening would result in rare but serious side effects associated with needle aspiration cytology such as needle-track seeding, particularly of lesions more than 2 cm in diameter, and hemorrhage, bile peritonitis, and pneumothorax. Transjugular liver biopsy is rarely associated with major complications such as perforation of the hepatic capsule or cholangitis.

Magnitude of Effect: Good evidence for uncommon but serious harms.

• Study Design: Randomized controlled trials and observational studies.
• Internal Validity: Fair.
• Consistency: Multiple studies, large number of participants.
• External Validity: Good.

Incidence, Mortality, and Risk Factors

In 2020, liver cancer was the sixth most common cancer and third leading cause of cancer death in the world. In the United States, it is estimated that there will be 41,630 new cases diagnosed in 2024 and 29,840 deaths due to this disease. There is a distinct male preponderance among all ethnic groups in the United States.

Chronic hepatitis B and C are recognized as the major factors worldwide increasing the risk of HCC, with risk being greater in the presence of coinfection with hepatitis B virus and hepatitis C virus. The incidence of HCC in individuals with chronic hepatitis is as high as 0.46% per year. In the United States, chronic hepatitis B and C account for about 30% to 40% of HCC. Chronic hepatitis G infection is not associated with HCC in either hepatitis B surface antigen–positive carriers or noncarriers.

Cirrhosis is also a risk factor for HCC, irrespective of the etiology of the cirrhosis. The annual risk of developing HCC among persons with cirrhosis is between 1% and 6%. Other risk factors include alcoholic cirrhosis, hemochromatosis, alpha-l-antitrypsin deficiency, glycogen storage disease, porphyria cutanea tarda, tyrosinemia, and Wilson disease, but rarely biliary cirrhosis. A retrospective case-control study found that features suggestive of nonalcoholic steatohepatitis, including obesity, type 2 diabetes, dyslipidemia, and insulin resistance, were more frequently observed in patients with HCC associated with cryptogenic cirrhosis than in those with HCC of viral or alcohol etiology. Aflatoxins, which are mycotoxins formed by certain Aspergillus species, are a frequent contaminant of improperly stored grains and nuts. In parts of Africa, the high incidence of HCC in humans may be related to ingestion of foods contaminated with aflatoxins. This association, however, is blurred by the frequent coexistence of hepatitis B infection in those population groups. The likely etiology of HCC is summarized in the following table.

Likely Etiology of HCC

Rationale for Screening

The rationale for screening for hepatocellular carcinoma (HCC) is based on the concept that populations at high risk for HCC, such as those with cirrhosis, can be identified. However, 20% to 50% of patients presenting with HCC have previously undiagnosed cirrhosis. These patients would not be recruited into a surveillance program if the presence of cirrhosis is used to define a target population. The modalities potentially available for screening include serum alpha-fetoprotein (AFP) and ultrasonography. Abnormal screening results may lead to liver biopsy for diagnosis. Complications of liver biopsy are reported in 0.06% to 0.32% of patients, and typically occur within the first few hours after the biopsy.

Tumor Markers for the Detection of Hepatocellular Carcinoma

There are four categories of tumor markers that are currently being used or studied for the detection of hepatocellular carcinoma. These include oncofetal antigens and glycoprotein antigens; enzymes and isoenzymes; genes; and cytokines.

Alpha-fetoprotein

Serum AFP, a fetal-specific glycoprotein antigen, is the most widely used tumor marker for detecting patients with HCC. The reported sensitivity of AFP for detecting HCC varies widely in both hepatitis B virus (HBV)-positive and HBV-negative populations, which is attributable to overlap between screening and diagnosis study designs. When AFP is used for screening of high-risk populations, a sensitivity of 39% to 97%, specificity of 76% to 95%, and a positive predictive value (PPV) of 9% to 32% have been reported. AFP is not specific for HCC. Titers also rise in acute or chronic hepatitis, in pregnancy, and in the presence of germ cell tumors.

A prospective, 16-year, population-based, observational study of screening for HCC included 1,487 Alaska Native individuals chronically infected with HBV. The study compared survival among screen-detected patients with HCC with a historical comparison group of clinically diagnosed patients with HCC. The screening program’s target was AFP determination every 6 months. It achieved 97% sensitivity and 95% specificity (excluding pregnant women) for HCC. Such high sensitivity and specificity have not been found for other high-risk groups, such as individuals with cirrhosis. Whether screening actually improved survival is not clear.

A case-control study conducted within the U.S. Veterans Affairs (VA) health care system assessed whether screening with AFP and/or ultrasound reduced HCC mortality. The cases were 238 patients with cirrhosis who died of HCC from 2013 to 2015 and who had been in VA care with a diagnosis of cirrhosis for 4 years or more before the diagnosis of HCC. The controls, who did not die of HCC and had also been in VA care for 4 years or more, were matched for date of entry (or focal time) and for age, sex, race, model for end-stage liver disease (MELD) score, and etiology of cirrhosis (mainly hepatitis C virus). The study examiners, blinded to outcome status, used chart extraction to assess exposure to ultrasound and AFP screening. The reason for testing (screening vs. other indication) was assessed, also blinded to outcome. The study found that there was no difference between cases and controls regarding the proportion of patients who underwent screening ultrasound (52.9% vs. 54.2%), AFP screening (74.8% vs. 73.5%), or both. The lack of difference persisted for tests within 1, 2, or 3 years of the outcome. Given the paucity of randomized controlled trials and their lack of strength, as noted elsewhere in this section, this case-control study—done with great care to avoid bias—comprised perhaps the strongest evidence about the efficacy of AFP or ultrasound screening; however, it showed no benefit in HCC mortality.

Hepatic Ultrasonography

Limitations in the sensitivity and specificity of AFP in surveillance of high-risk populations led to the use of ultrasonography as an additional method for detection of HCC. Studies in both healthy hepatitis B surface antigen carriers and in patients with cirrhosis have defined the performance characteristics of ultrasound as a screening test for HCC. Sensitivity in the former was 71% and in the latter 78%, with 93% specificity. The PPVs were 14% and 73%, respectively. In a study of patients who were on a waiting list for liver transplant, ultrasonography was found to have a sensitivity of 58%, a specificity of 94%, a negative predictive value of 91%, and a PPV of 68%.

A case-control study conducted in the VA population assessed whether screening with AFP and/or ultrasonography reduced HCC mortality. For more information, see the Alpha-fetoprotein section.

Computed Tomography

Limitations in the sensitivity and specificity of AFP and ultrasonography in surveillance of high-risk populations, such as individuals with cirrhosis, led to the assessment of computed tomography (CT) as an additional method for detection of HCC. Studies in patients with cirrhosis suggest that CT may be a more sensitive test for HCC than ultrasonography or AFP more than 20 μg/L.

Efficacy of Screening and Surveillance Programs

A controlled trial of 18,816 individuals aged 35 to 59 years with hepatitis B in Shanghai randomly assigned patients to a screening group using AFP and ultrasonography every 6 months versus a usual-care group. HCC mortality was lower in the screened group (83.2 vs. 131.5 per 100,000; mortality rate ratio of 0.63 ). While these results are promising, there were problems, including the following:

• The results varied in different publications.
• The comparison group was not actively followed.
• The CI was near 1.0.
• Intention-to-treat analysis was not used.
• Assessment of outcome was not blinded.
• Generalizability to other populations is uncertain.

A randomized controlled trial studied 5,581 men aged 30 to 69 years who were chronic carriers of HBV between 1989 and 1995 in Qidong County, China. Of these men, 3,712 were randomly assigned to a screening group and 1,869 to a control group. Screening entailed 6-monthly AFP assays, with follow-up of patients having an abnormal (≥20 μg/L) test result. All patients were followed up for liver cancer and/or death. The overall sensitivity and specificity of the program were 55.3% and 86.5%, respectively. In patients who complied with all scheduled screening tests, sensitivity was 80% and specificity was 80.9%. The mortality rate in the screening group (1,138 per 100,000 person-years) was not significantly different from that in the control group (1,114 per 100,000 person-years), although AFP screening resulted in an earlier diagnosis of liver cancer (i.e., percentage of cases in stage I was significantly higher in the screened group than in the control group ). A review concluded that the method of measuring AFP was not sensitive enough to detect HCC, affecting interpretation of the negative result of this trial.

Two kinds of harms or complications may result from screening. Direct harms may result from complications of liver biopsy done as part of the diagnostic workup. Such complications are reported in 0.06% to 0.32% of patients, and typically occur within the first few hours after the biopsy. Complications include hemorrhage, bile peritonitis, penetration of viscera, and pneumothorax. Rarely, death occurs as a direct result of liver biopsy (0.009%–0.12%). About one third of patients experience pain at the site of entry, in the right upper quadrant, or in the right shoulder. Needle aspiration cytology and liver biopsy are rarely associated with needle-track implantation of malignant cells. Lead-time bias (earlier diagnosis in the natural history of hepatocellular carcinoma rather than improved survival from earlier diagnosis and treatment), length bias (earlier detection of slower-growing and less aggressive tumors through screening), and/or overdiagnosis of HCC (detection of tumors that will not affect morbidity or mortality) may wholly or partially account for the improved 5-year and 10-year survival rates reported.