Challenges Remain in Hepatocellular Carcinoma Surveillance of Patients With NASH

Amit Singal, MD, MS

Amit Singal, MD, MS

Research has shown that the incidence of hepatocellular carcinoma (HCC) in patients with nonalcoholic steatohepatitis (NASH)-related cirrhosis is high enough for surveillance to be cost-effective in this patient population. However, few of these patients receive regular surveillance for HCC.

“Overall, when you look at all patients with cirrhosis, less than 20% receive regular surveillance every 6 months, and it’s even lower for that subgroup with NASH cirrhosis,” said Amit Singal, MD, MS, at the 2019 International Liver Cancer Association (ILCA) Annual Meeting.

Retrospective studies have shown that providers have a hard time recognizing patients with NASH-related cirrhosis, which makes surveillance in this population more challenging. One issue with surveillance in the NASH population is providing access to the test for patients.

Another issue with surveillance in this patient population is providing the patients with effective screening tests for HCC. Currently, ultrasound and blood test for the alpha-fetoprotein (AFP) biomarker are the recommended tests for HCC. However, ultrasound has a sensitivity rate of around 47% for detecting HCC at an early stage. AFP tests when combined with ultrasound can bring the sensitivity rate to around 60% to 65%, but more work is needed to improve these tests further.

There are curative treatment options for early-stage HCC, including liver resection, local ablation, and transplantation. However, the disease needs to be caught earlier in patients within the NASH population in order for them to benefit from these therapies. Systemic therapies are also available for patients with intermediate- to advanced-stage HCC.

In an interviewTargeted Oncology, Singal, medical director of the Liver Cancer Program at UT Southwestern Medical Center, discussed issues with surveillance in patients with NASH, with or without cirrhosis. He also highlighted how combatting these issues can help improve the survival of patients with NASH-related HCC.

TARGETED ONCOLOGY: What does the treatment landscape look like now for patients with NASH-related HCC?

Singal:At this point, we don’t have different treatment options that are really based on etiology of liver cancer. The treatments that we have for patients with NASH-related cirrhosis in HCC are similar to those that we have for other etiologies. If a patient is found at an early stage, we have curative therapies such as surgical resection, local ablation, and liver transplantation, which can afford a 5-year survival rate of over 60%. If you’re found at a more intermediate- or advanced-stage, then we deliver palliative therapies, and that can be either locoregional, such as chemoembolization, radioembolization, radiation-based therapies; or systemic-based therapies with either tyrosine kinase inhibitors or the checkpoint inhibitors.

TARGETED ONCOLOGY: What are the issues with surveillance of the NASH population?

Singal:Given the strong association between early detection and curative therapy and improved survival, there’s been a strong push to do surveillance in this population to find these tumors at an early stage. Surveillance is typically recommended to be done with ultrasound and a blood test for AFP. When you talk about the effectiveness when surveillance is applied in a population of patients, there are 2 main drivers. The first is that you need to have access to the test, and second is the test needs to succeed and be effective in that patient population.

When you talk about the population of NASH patients having access to surveillance, you really have to break this down into patients with NASH with cirrhosis and NASH without cirrhosis. For patients with NASH with cirrhosis, we have good data showing that the incidence of HCC in that patient population is sufficiently high for surveillance to be cost effective. However, when we take a look at retrospective cohort studies, we find that providers actually have a hard time recognizing patients with NASH with cirrhosis. That is actually 1 of the most common reasons for patients not to have surveillance done. Overall, when you look at all patients with cirrhosis, less than 20% receive regular surveillance every 6 months, and it’s even lower for that subgroup with NASH cirrhosis.

In that second group that I mentioned of patients with NASH without cirrhosis, this is a cohort of interest because prior studies have suggested that up to 25% to 30% of patients with NASH in HCC actually develop HCC in the absence of cirrhosis. This caused a lot of alarm because it changes the paradigm of delivering surveillance to that patient population. Recently, there have been data that came out of the national Veterans Affairs which showed even though a large proportion of HCC occurs in the absence of cirrhosis, when you take a look at the actually incidence rate in a non-cirrhotic NASH population, it is actually extremely low. The incidence of HCC in a non-cirrhotic NASH population is 0.008 per 100 person years. This study nicely shows that HCC surveillance is actually not cost effective in this patient population. Recent grade-based recommendations say that you should not do HCC surveillance in this population, largely based on these data. One of the things I think we have to push for in the future is to actually develop predictive models where we can identify a subgroup of patients with non-cirrhotic NASH that are at sufficiently high risk to get better HCC surveillance, but we currently don’t have anything that is ready for prime time.

The second area that I talked about, besides having access to the test, was the tests actually have to work. Going back to the current recommended tests, it’s ultrasound plus or minus this biomarker, AFP. When you take a look at how well ultrasound works in a general cirrhotic patient population, you find that ultrasound has a sensitivity rate around 47% for finding HCC at an early stage. However, when you take a look at the correlates for ultrasound quality, you find that patients with NASH-related cirrhosis are much more prone to having poor quality ultrasounds, and so these data lead us to believe that the sensitivity of ultrasound in NASH cirrhosis may be even lower than that 47%. This highlights that we need better tests. Now those better tests can include incorporation of more biomarkers or better imaging techniques in the future. Right now, the only biomarker that has had sufficient validation is AFP, but while AFP does add value to ultrasound, it just doesn’t have the sensitivity for early tumor detection where we need it to be. Even when you use ultrasound and AFP in combination, you still talk about a sensitivity that is around 60% to 65% for finding cancer at an early stage. The hope is that by incorporating new novel biomarkers, we can get that sensitivity for early HCC detection much higher and hopefully closer to 75% to 85% in the future.

TARGETED ONCOLOGY: How can these issues in surveillance be addressed in HCC?

Singal:The next steps are multi-fold. The first is we talked about the under use of surveillance in this patient population. I think that [some of] the interventions to address this can include provider education, so that simple step has actually shown to actually increase HCC surveillance. You can talk about doing things like best practice alerts in the electronic medical record. We’ve evaluated population health programs like mailed-outreach strategy approaches in patients with cirrhosis, but I think we have to think about interventions to increase that HCC surveillance utilization in the future.

Now for the second part, that’s where a lot of research is going on right now in terms of trying to find better surveillance modalities. We’ve briefly talked about biomarker development that is going on. There are large cohorts that are now maturing, which will allow validation of some of these biomarkers to see if they can actually improve early HCC detection. Likewise, there’s now imaging techniques that are being evaluated. One example is an abbreviated MRI protocol where you take a full diagnostic MRI and you shorten it to the 15 minutes of required sequences that would allow for a screening exam, and that would allow us to make the test more cost effective, hopefully.

These things are all in early development, and we need further validation for these tests to be used in clinical practice. Overall, if you ask what I’m most hopeful for, it’s if we can come up with a highly effective biomarker that can actually solve both test effectiveness as well as the utilization. It’s easy to apply a blood test much more than a radiographic surveillance test.

TARGETED ONCOLOGY: Is there anything else important to note about these challenges?

Singal:We focused on the application of surveillance in this high-risk population, but surveillance is only the first step in a longer process. Surveillance can help detect these tumors, but then it’s important that if these tests work and we find tumors at an earlier stage, it’s important that once again these patients are referred for curative therapy. There are also large amounts of data that unfortunately show under-use of curative therapies for those patients found at an early stage. I’m hoping that if we improve that first step, that the process works, and we can truly maximize our curative therapies for these patients.

TARGETED ONCOLOGY: What do you hope oncologists take away from this?

Singal: