Why do we screen for hepatocellular carcinoma in cirrhosis?
Preventative care in patients with liver disease is essential and complicated; when the liver goes down, it takes other organs down with it. In addition to frailty assessment, polypharmacy evaluation, and variceal screening, we also look out for hepatocellular carcinoma (HCC). Have you ever wondered why we screen for HCC in patients with cirrhosis? Let’s find out!
How it all began:
The story of HCC screening begins with finding our little friend, alpha feto-protein (AFP). While first isolated in the 1944 in fetal serum of different animals, it was later noted be present in the serum and ascitic fluid of pediatric patient with a hepatoma. Later studies evaluated its potential as a biomarker of disease activity for HCC.
In 1948, Walter Sheldon and his colleagues reported one of the earliest cases series of primary liver carcinoma arising from a background of cirrhosis, giving evidence to a link between chronic liver disease and liver cancer. A workshop in Alaska in 1990 held by McMahon and colleagues gathered experts in the field to review data on screening high risk groups for HCC. The workshop concluded from available data that HCC could be detected at an early stage with serum and radiologic tests and postulated that early detection and resection of these cancers would lead to prolonged survival. The very first randomized clinical trial for HCC screening was not conducted until 2004. In this study, 19,000 patients with hepatitis B infection or other chronic hepatitis were followed for 5 years. Combined AFP and ultrasonography examination biannually demonstrated a 37% reduction in HCC mortality with a 58% adherence to screening. Keep reading to learn more about modalities for HCC screening below!
What is the epidemiology of HCC?
The incidence of HCC has been rapidly rising over the past 20 years and is projected to continue to rise in the next decade. The greatest burden of HCC globally is seen in Asia largely due to the high prevalence of HBV and specifically, is most concentrated in China and Southeast Asia.
Overall, HCC demonstrates a male predominance, showing a 3-4x greater prevalence in males than females. Risk factors include HBV- recent studies have estimated that majority of HCC cases worldwide are due to this virus even in the absence of cirrhosis. Chronic Hepatitis C infection (HCV) in setting of the opioid epidemic, and rising rates of diabetes and obesity leading to nonalcoholic fatty liver disease are responsible for the growing incidence of HCC in Western nations.
This figure demonstrates the HCC incidence per 100,000 individuals by geographic region as of 2012.
Encouragingly, some countries in Asia, including Taiwan, have observed a declining incidence of HCC; this is considered an early indicator of the success of national HBV vaccination strategies in select areas and preventative measures against aflatoxin.
What is happening at the cellular level in HCC?
HCC generally arises from a cirrhotic liver, with recurrent inflammation and subsequent fibrosis leading to dysplasia and eventual neoplasia. Alterations in several molecular pathways have been demonstrated to be involved in HCC. Across the genome, chromosomal deletions are identified in nearly all tumors.; less frequently, somatic mutations including TP53, a tumor suppressor gene is mutated, leading to dysregulation of DNA repair, apoptosis, and key stages of the cell cycle. Increased telomerase activity is noted in nearly 90% of HCC tumors leading to abnormal methylation and dysfunctional mismatch repair; the latter is particularly common in HBV-associated HCC, as the HBV integration into host hepatocyte DNA directly affects telomerase reverse transcriptase and creates significant genomic instability. A host of epigenetic changes seen in HCC result in hypermethylation that silences tumor suppressor genes, including E-cadherin, RASSF1, BRCA1, B-Catenin, and P16INK4a.
Beyond genetic irregularities, several signaling pathways have been implicated in HCC development. The MAPK (RASRAF/MEK/ERK) pathway has shown to be upregulated in hepatocarcinoma cells, but not dysplastic cells, and is one of many second messenger signaling cascades currently being investigated for future therapies.
More recently, several proteins in the epidermal growth factor family have been shown to have a pivotal role in HCC proliferation. Overexpression of epidermal growth factor receptor serves as an independent negative prognostic indicator for extrahepatic metastases and early tumor recurrence. Sorafenib, a tyrosine kinase inhibitor, targets this receptor to inhibit both cell proliferation and angiogenesis.
This figure beautifully depicts the cellular signaling pathways thought to be involved in HCC pathogenesis.
Screening in specific disease states
Because underlying cirrhosis is found in most patients with newly diagnosed HCC, it is safe to say that anyone with cirrhosis should be considered at risk for HCC. But what about in specific diseases?
Hepatitis B Virus (HBV)
HBV deserves special mention. Unlike most other etiologies of HCC, HBV has a direct oncogenic effect irrespective of stage of fibrosis.
This excellent figure from a recent HBV-associated HCC review highlights the different mechanisms of HCC development in HBV, including DNA damage from viral genome integration, chromosomal instability, telomerase reactivation, and neo-angiogenesis. Other factors in patients with HBV that increase HCC risk include HBV genotype (particularly genotype 3), high HBV DNA levels, and genomic mutations.
Vaccination strategies and HBV therapies have decreased HCC incidence, confirming HBV’s importance in HCC pathogenesis.
So, what does this mean? You don’t need advanced fibrosis or cirrhosis in HBV infected patients to develop HCC, but HCC risk increases with development of cirrhosis.
Hepatitis C Virus (HCV)
While global efforts to eliminate HCV are gaining attraction among public health experts, HCC screening in this population has been established by most international medical organizations
A large Taiwanese study identified a 20x greater risk of HCC in patients infected with HCV, implying a compelling contribution to hepatocarcinogenesis. The cumulative incidence of HCC in anti-HCV positive patients was 2.98 percent, compared to 0.13 percent seronegative individuals.
The advent of direct acting antivirals (DAA) has changed the therapeutic landscape for HCV patients, and by extension, their risk for HCC. demonstrated a 71% reduction in HCC incidence in patients with DAA induced sustained viral response (SVR); subgroup analysis showed that absolute risk reduction was much greater in cirrhotics than non-cirrhotics.
European society guidelines currently suggest screening in patients with HCV and advanced fibrosis due the annual HCC incidence > 1.5%; the American Association for Study Liver Diseases (AASLD) and Asian Pacific Association for Study of the Liver (APASL) do not make specific recommendations for non-cirrhotic HCV patients.
Non-alcoholic fatty liver disease (NAFLD)
Based on current estimates, NAFLD is the most rapidly growing cause of liver transplant candidates in the United States. While NAFLD may not have as high of a HCC risk as HCV infected patients, a long term study from Japan highlighted a 25-fold increase in HCC development in NAFLD with fibrosis, compared to those without. Other literature has suggested mitigation of metabolic risk factors can also reduce HCC risk. Hepatic lipid accumulation may promote carcinogenesis through oxidative DNA damage, providing potential data for future guidelines to carefully consider.
Similar to HCV, guidelines differ in recommendations for screening. European society guidelines recommend HCC screening in NAFLD patients with cirrhosis AND advanced fibrosis as they appear to have ~1-1.5% annual incidence; AASLD guidelines recommend screening in patients with cirrhosis.
Alcohol-related liver disease
Alcohol use greater than 80g/day for more than a decade elevates the risk of HCC almost 5-fold. Notably, the risk does not decrease with alcohol cessation. A seminal Italian study of 1,829 patients with cirrhosis observed HCC in ~10% of patients with cirrhosis purely due to alcohol – nearly the same as HCC prevalence in patients with HCV. The hepatocarcinogenic effects of alcohol are thought to be due not only to development of cirrhosis but also due to direct genotoxic effects including production of reactive oxygen species and DNA fragmentation.
Currently, guidelines recommend screening for HCC for patients with alcohol associated liver disease if they have evidence of cirrhosis.
What should we be using to screen for HCC?
- There are no physical exam findings early in the disease course specific to HCC
- The liver has substantial functional capacity so lab abnormalities may only be seen late in the tumor life cycle
- Metastases develop late in the disease
So what modalities have been used historically and what do we use now?
As discussed above, early studies investigated serum biomarkers for early detection of HCC – most notably, serum a-fetoprotein (AFP). Difficulties with AFP soon became apparent. As studied by Lok and colleagues in a Chinese cohort, AFP levels may be elevated in the setting of cirrhosis without HCC and could remain normal in those with underlying .
This schematic serves as a window to HCC screening algorithms in the past, including serum markers, ultrasonography, computed tomography (CT)/magnetic resonance imaging (MRI), or liver biopsy. Now that we have better screening modalities, this is not an algorithm that is followed any longer.
Interestingly, a large Veterans Affairs study recently validated a HCC early screening algorithm by incorporating AFP (specifically rate of change from prior AFP levels), age, platelets, and ALT to predict 6-month HCC risk in patients with cirrhosis – offering an increased sensitivity of 53%, compared to 48% by using AFP alone.
As AFP became popularized in the early 1980s for screening and diagnosis for HCC, something became painfully obvious. While AFP was a good test, up to 15% of patients with advanced HCC could have normal levels. Further, more than half of tumors detected with elevated AFP were already large, making screening less effective. An early study by Sheu and colleagues evaluated early detection of HCC by real time ultrasonography in 528 patients. In HCC patients with liver masses less than 5cm, almost 50% had normal AFP and would have been missed otherwise.
A specific limitation of ultrasound is evident in patients with obesity. Current AASLD guidelines do not have different recommendations for HCC screening modalities based on body mass index (BMI); an evaluation of 116 liver transplant patients with HCC observed on explanted livers showed a significantly lower sensitivity of HCC detection in patients with obesity compared to patients with BMI < 30 kg/m2. CT scan detected HCC in 98% of obese patients that was not detected on ultrasound. These data highlight the need for more sensitive imaging modalities in our patients with obesity, especially giving the rising prevalence of obesity in the United States.
While the shortcoming of ultrasonography includes operator dependence and low specificity for echographic features, real time scanning and high sensitivity allowed easier detection of small tumors, paving the way for its inclusion in current practice.
CT and MRI are most costly than ultrasound and lacked early data in effectiveness for HCC screening. Recent cohort studies, however, challenge this notion. For example, in evaluation of patients with cirrhosis, MRI had a lower false positive rate than ultrasound in HCC detection (3 vs. 5.6%, p < 0.05). Abbreviated MRI protocols have been developed to minimize contrast exposure and optimize hepatic anatomy visualization; a recently published meta-analysis found non-contrast abbreviated MRI to be a low cost and a comparable modality to abdominal ultrasonography for HCC detection.
Fast forward to present day – here is a summary of most recent guidelines for HCC screening by three major liver societies by Sangiovanni et al. Generally speaking, US with or without AFP is the screening modality of choice.
We know that HCC screening is important, but is it cost-effective?
A seminal study by Sarasin and colleagues in 1998 evaluated the cost effectiveness of HCC screening in a hypothetical cohort of patients with Child Pugh class A cirrhosis; surveillance achieved a 90-day increase in life expectancy and was deemed cost effective if annual incidence of cirrhosis was at least 1.5%. The annual incidence of HCC in patients with cirrhosis is ~2-4%, which make surveillance a cost-effective endeavor. More specifically, ultrasound with AFP optimizes early detection of HCC with lowest costs compares to no surveillance or ultrasound alone - accounting for false positives and harms of screening.
Surveillance of patients without cirrhosis is generally guided by a balance of risk and effectiveness of testing modality. Individuals with a family history of HCC, HBV infected Africans or African Americans, HBV infected Asian males above 40 years old, and HBV infected Asian females above 50 years old should be offered HCC screening per AASLD guidelines.
While screening may detect early-stage HCC, we must balance benefits of early diagnosis with potential screening related harms. A hot-off-the-press prospective study evaluating benefits and harms of HCC surveillance in patients with cirrhosis. It showed early detection of HCC in almost two-thirds of patients, minimal surveillance harms, and almost half of the cohort was receiving curative therapy. Further data on patients without cirrhosis are needed.
All in all, screening for HCC is a preventative measure to detect liver cancer early. There remains significant work to be done in serum biomarkers, clarifying the role of screening in patients without advanced fibrosis, and in evaluating each patient as an individual with their risk factors and preferences in mind.
So how are we doing in HCC screening?
Unfortunately, only ~20% of patients with cirrhosis undergo screening. Barriers to screening include limited clinic time, unclear designation of screening provider, and lack of widespread awareness.
A recent clinical study sought to assess if mailed outreach invitation would be effective; 1,800 patients randomized to three arms received usual care, mailed outreach, or mailed outreach with patient navigation. Screening rates were significantly higher in the outreach arms than in those with usual care, including in historically underscreened populations, such as non-Caucasians.
Clearly we have work left to do.