ACR US LI-RADS for Screening and Surveillance in HCC - HD

Introduction and Disclosures

I'm gonna be talking about the ACR ultrasound LI-RADS for screening and surveillance in HCC.

These are my disclosures. I just want to disclose that I am the chair of the ultrasound LI-RADS working group. I'm a member of the version 2017 LI-RADS Writing Group and a member of the LI-RADS Steering Committee. So I have a personal vested interest in all of you adopting LI-RADS and especially ultrasound LI-RADS and CEUS. Yes. A plug for CEUS too. It's very important.

A lot of you probably use LI-RADS already, but perhaps you have not heard of ultrasound LI-RADS. It's actually brand new. We just released it in August of 2017 on the ACR website.

What is Ultrasound LI-RADS?

Ultrasound LI-RADS stands for Ultrasound Liver Imaging Reporting and Data System. And we provide a standardized system for screening and surveillance ultrasounds in patients who are at risk. That's really important. These are at risk patients only, for developing HCC.

It's kind of interesting. All major societies actually do recommend using ultrasound for screening and surveillance. But none of them go into detail about how to perform the examination, how to interpret the examination, or how to report the ultrasounds. So there's a huge void in the existing literature. And so ultrasound LI-RADS helps to fill in this void and hopefully by having this system we can then help with data collection, which can then perhaps influence future directions in HCC screening and surveillance.

Components of LI-RADS

LI-RADS as a whole has four major components. There's the ultrasound LI-RADS, which is for screening and surveillance in HCC. It's different, separate from contrast enhanced ultrasound, which is Stephanie's baby. There's CT/MRI LI-RADS, which most of you're probably familiar with. The most recent iteration came out in July of 2017. And they also included a treatment response algorithm within that, iteration.

I'm really fortunate to be working with a large group of very talented individuals who are all experts in ultrasound. Most people also do CT and MR in their practice, so it really helps to put ultrasound in perspective, in the greater scheme of HCC imaging.

Screening and Surveillance in General

I'm gonna first talk about screening and surveillance, in general, and then I'll talk about ultrasound, the ultrasound LI-RADS algorithm.

As you know, HCC is a cancer of the hepatocyte is the sixth most common cause of cancer, but the second most common cause of cancer death in the world. So it's quite lethal, but it only occurs in the setting of chronic liver disease, most commonly cirrhosis in 80 to 90% of patients. So when you have a disease that is so prevalent and lethal, it's really important to do screening and surveillance to improve survival.

And the goal of screening and surveillance in HCC is to be able to detect preclinical HCC when patients are not symptomatic, to identify HCCs when they're potentially curable either with local therapy or with transplantation. And the ideal interval has been shown to be every six months semi-annual surveillance, and I'll talk about that in a little bit more detail. And this may also be a sounds question.

I talk about screening and surveillance, but there's actually a difference. So screening is where you apply a diagnostic test to a population that is at risk for developing a disease. And so the first time you apply that test, you are detecting the prevalence of the disease. So that's screening every subsequent time that you apply that same diagnostic test to the same at-risk population at a defined interval for that particular disease that is now surveillance. And now you're gonna be detecting the incidents of the disease.

So with HCC, there are actually two major tests out there for screening and surveillance. There's ultrasound and there's AFP. So ultrasound is actually recommended by all international societies with the, surveillance interval of six months being the most common. It has the advantage of being widespread in its availability. It's non-invasive as you all know, it's accepted, well accepted by patients and physicians. There's moderate cost associated with ultrasound. It's a lot cheaper than for instance, CT or MRI. So it is cost effective as a screening and surveillance test sensitivity on the hand other hand is somewhat variable depending on what you read in the literature, somewhere between 60 to 89%. And this variability and sensitivity is something that we actually do address with our ultrasound LI-RADS.

Specificity on the other hand with ultrasound is actually very high. 90% in the literature when you use it as a screening surveillance test, AFP on the other hand, is only advocated by Asian societies and they use a threshold of 200 nanograms per milliliter, which actually only has a sensitivity of 22%. If you lower your threshold to 20 nanograms per milliliter, then the sensitivity increases to 60%. But then you have a lot of false negative or false positives. And so because of this variable performance, the AASLD does not currently advocate AFP, although it is being revisited.

Evidence for Using Ultrasound

So what is the evidence for using ultrasound? Well, actually there isn't that much evidence out there. There's actually been only one large randomized controlled study that was performed in China where they had almost 19,000 patients with chronic hep B with and without cirrhosis. And in this study they actually had two groups, those patients who underwent screening and surveillance with ultrasound and AFP every six months. And those who didn't. And then the patients who had screening and surveillance, even though they had a pretty poor compliance rate of about 60%, they were still able to show a 37% reduction in HCC related mortality.

There was another randomized controlled study where they only used AFP and that did not change mortality. So we think that the first study, we can attribute that decrease in mortality to ultrasound. There was another mass screening study published out of Taiwan where they used ultrasound one time and they were able to show a mortality decrease of 31%. We don't think that there's gonna be another large randomized controlled study to be replicated like these because of the ethical consequences of not doing screening and surveillance.

Surveillance Population

Okay, so when you do screening and surveillance, it's important to identify your surveillance population. The AASLD is currently defines the surveillance population in the United States to include cirrhotic patients from any etiology. For those patients, the incidence of HCC exceeds 1.5% per year. So that is considered to be cost effective. And they also include non-cirrhotic chronic hep B patients in whom the incidence of HCC exceeds 0.2% per year.

So it's not all chronic hep B patients, it's only certain subsets that exceed this threshold and that includes Asian males over 40 Asian women, over 50 African and North American blacks of any age, and patients who have a family history of HCC and have chronic Hep B.

So there are other patients who are have an increased risk of HCC, but their risk does not exceed these defined thresholds. And so in those patients, the benefit of surveillance is actually uncertain. And so the AASLD does not officially advocate surveillance in those patients. But it depends on your local regional preferences. So your institution may choose to do screening and surveillance in these patients below non-cirrhotic, chronic Hep C patients with stage three fibrosis, non ciro hep B patients that don't meet these criteria non-cirrhotic NAFLD patients or transplant candidates who don't meet those criteria, but they're not currently advocated for by the AASLD.

That said, it actually depends on where you live. So LI-RADS is North America, so we align ourselves with the AASLD for our screening and surveillance population, but it's actually different if you live in Japan. The Asian Pacific Association for the Study of Liver Disease has a slightly different population. The Korean society and the European Society, they all have slightly different populations. I just wanna point out that Child-Pugh C Cirrhotics are not currently advocated for screening and surveillance. And the reason for that is that their life expectancy is not long enough to warrant screening and surveillance unless they are waiting a liver transplant.

Surveillance Interval

Okay, so I talked about the surveillance interval and the surveillance interval actually depends on the growth rate of the tumor, not the risk of HCC development. And so all major societies do recommend every six month, imaging. And the evidence out there is that there's two studies that have shown that six months is better or improved survival compared to 12 months. There's one study that's a little bit older that looked at six versus 12 months and showed that they were equivalent. That was from 2002. And then there are a couple of other studies that used shorter interval surveillances less than six months and didn't show any benefit to survival. So six months seems to be the sweet spot for HCC.

Ultrasound LI-RADS Algorithm

Okay, so that's screening and surveillance. Let's get to the meat of this, the ultrasound LI-RADS algorithm.

Okay, so in our algorithm we do talk about how to do the ultrasound examination. It's really important that you use proper technique. This is probably all obvious to you guys. But you do want to optimize by making sure that the patient hasn't eaten at least four to six hours prior to the ultrasound. To minimize bowel gas, you wanna optimize the image to adequately penetrate the whole liver looking, so that you can see the entire liver and diaphragm get transverse and longitudinal views through the liver. You're really looking specifically for any focal abnormality or diffuse parenchymal abnormalities, which we'll talk about.

We do recommend that when you look at the main portal vein, that you look at it with both gray scale as well as color doppler, not just color doppler because the blooming artifact from color Doppler could potentially obscure a subtle non occlusive thrombus, which you might be able to see with gray scale better. We leave it optional whether you get color Doppler views of the branches of the portal vein right and left branches right middle and left hepatic veins. Some institutions do that like Stanford, but others do not. So that's optional. Unless you see a vein thrombus, then we do recommend that you look at that more carefully with color or spectral doppler.

Okay, so this is kind of a busy slide, but these are just all the different recommended ultrasound views and longitudinal and transverse. This is available online. Again, optional views include the color and spectral tracings of the veins. And then cine loops we leave as optional. Although we have found anecdotally that cine loops helps to improve your sensitivity.

So the way this all fits together is that a patient should undergo screen, surveillance every six months with ultrasound. If the ultrasound is positive ultrasound three, then they should go on to further characterization with a contrast enhanced multiphasic CT MR or contrast enhanced ultrasound. So the goal of the ultrasound, the surveillance ultrasound is to detect a lesion or detect an abnormality. The goal of the diagnostic study is to characterize the lesion.

Ultrasound Categories

Okay, so ultrasound LI-RADS in a nutshell, you should provide two scores. Okay? The first score is the ultrasound category and that determines the follow up. And there are three possible ca three, three parts to this or three possible categories. Ultrasound one, negative ultrasound two sub-threshold, and ultrasound three positive. We'll talk about each of these in more detail. And then we recommend that you provide a visualization score. So the visualization score addresses that sensitivity. We talked about the visualization score. You can have visualization A, no or minimal limitations visualization. B, moderate limitations or visualization. C, severe limitations. And this communicates the expected level of sensitivity of the ultrasound in the particular individual, but does not determine follow up.

Okay, looking at the ultrasound category first, we've color coded it just like the rest of LI-RADS. So ultrasound one negative, it's green, that's good. You can continue on with your surveillance every six months. Ultrasound two sub-threshold is yellow, it means caution. And so we would recommend short interval surveillance three to six months. And then ultrasound three positive stop. It's red. Go on to further characterization with with CT MR or CEUS and notice that these are numbers as opposed to visualization scores, which are letters A, B, or C. Ultrasound category 1, 2, 3.

Okay, so these are the different management. Follow up. So ultrasound one, negative six month routine surveillance ultrasound two sub threshold, short interval ultrasound in three to six months. And then ultrasound three positive further characterization with a multiphasic study.

So in a little bit more detail, ultrasound one negative. This means that there is no evidence of HCC. So either you see nothing or some or you see something, but it is definitely benign. And by that we mean a simple cyst, focal fat spearing, most commonly around the gallbladder. A previously confirmed benign lesions such as a hemangioma or FNH for instance or calcifications, those would all be considered ultrasound one negative studies. And here's an example of a patient who has a normal liver except for a cyst.

Ultrasound two subthreshold is where you have an observation that may warrant short term ultrasound surveillance. And by that we actually mean that we see something observation that measures less than 10 millimeters in longest dimension. That's not definitely benign and it can be of any echogenicity. So here's an example. Here we see a hyper echoic subs centimeter lesion. We would say that this is ultrasound two subthreshold. We see these very frequently in patients who have chronic liver disease. And so we don't really want to be chasing all of these. Every single one of these, many of them do not have CT or MR. Correlates. And so these would warrant a short interval follow up, another patient with a sub centimeter hyper echoic lesion. Again, short interval follow up.

The reason for that is that although many patients have these with chronic liver disease, we don't wanna miss a subtle early HCC. This patient actually got lost to follow up and one year later, he came back and this lesion had grown substantially into this large artily enhancing lesion with washout. So this is what you do not want to happen. You wanna detect that lesion early. Another example of a hypoechoic sub centimeter sub-threshold, lesion ultrasound three positive is where you see an observation that warrants for their characterization.

I'm trying to use the word observation instead of lesion because observation does not convey an opinion whether positive or negative or suspicious or non suspicious observation is more neutral. So we do try to use observation in LI-RADS. In any case, an ultrasound three positive would be either an a focal observation that measures greater than or equal to 10 millimeters in longest dimension or a new thrombus in a vein. And I just wanna also mention that we include in this parenchymal distortion, as part of an ultrasound three positive, and I'll show you some examples of that.

Okay, so ultrasound three positive. Here's an example. Iso echoic observation measures over 10 millimeters. That should be characterized further with a multiphasic study. Another example of an ultrasound. Three positive, observation quite a bit larger. That should also be characterized. Another patient with a slightly hyper echoic observation increased through transmission. There's some detectable internal vascularity here. How many of you would say this is suspicious for an HCC and a patient who is at risk?

Okay, well this is a trick question. You're not characterizing this on ultrasound. Ultrasound in this case is used to detect a lesion or detect an observation. This should be characterized definitively on a multiphasic study. In this case, we see that it's hyper enhancing on a trail phase washes out on portal venous and delayed phase, it measures over two centimeters in size. And so by that, by LI-RADS, this meets LI-RADS five criteria definite HCC. But that determination needs to be made with a contrast study. Another example, hypoechoic observation over one centimeter. Again, we're not trying to characterizes where detecting it. In this case, this has rim enhancement or OID enhancement that persists on arterial portal venous and delayed phase. It's retaining contrast. And so this is actually quite characteristic of a cholangiocarcinoma.

Parenchymal distortion in my opinion, is a lot more challenging to identify, but if you're looking for it, you will see this. So with parenchymal distortion, we define that as an area. It'll define area of heterogeneity, refractive edge shadows, or an area where you've lost your normal hepatic architecture. And here's an example. It's quite subtle, but we've lost our normal hepatic architecture here and there's a little bit of mass effect on the port of hetus. These areas of parenchymal distortion very frequently correspond to infiltrative HCCs, so you don't wanna miss those. Another example in this patient, they have parenchymal distortion, manifest by, refractive edge shadows throughout the liver in an area where we've lost our normal portal triads. And that corresponds to this large infiltrative HCC.

If you see a thrombus in a vein, you wanna look at that more carefully with color doppler. In this case with color doppler, we see there's a lot of internal vascularity and detectable arterialized flow. So this is quite characteristic on ultrasound as a tumor in vein. I just wanna let you know that LI-RADS no longer uses tumor thrombus. It's tumor in vein. That's the correct term tumor in vein. So sometimes you can actually see tumor in vein, quite characteristically on ultrasound, but not always. And so that's why we don't actually require that you try to characterize it on gray scale ultrasound.

This is a different patient with really marked parenchymal distortion in the liver. That alone would, classify this as an ultrasound three positive study. But in this case, they also have a thrombus in the portal vein. It turned out to be a tumor in vein, but really hard to tell in this color Doppler ultrasound.

Okay, so those are the ultrasound categories which were red, yellow, and green.

Ultrasound Visualization Score

Now we're gonna talk about the ultrasound visualization score. So the visualization score communicates the expected level of sensitivity in an individual patient. And so you'll see now we are in gray scale, we don't have colors anymore, and we're using letters A, B or C, a no or minimal limitations. B, moderate limitations and c, severe limitations. And so the way, we think about this is it's very analogous to mammography, where in mammography you are describing breast density and that breast density, if it's a very dense breast, that just conveys the expected level of sensitivity of that mammogram in that patient. But it doesn't necessarily change management.

So very similarly, the ultrasound visualization score conveys the expected level of sensitivity of that ultrasound, but it does not change management.

Okay, so visualization a no or minimal limitations. This is where you think if you have any limitations, they're really unlikely to meaningfully affect sensitivity. So here's a normal looking liver, different patient with a visualization, a no or minimal limitations, some mild cirrhosis here, nodularity.

Once you get to visualization B, we call this moderate limitations, where you think that there are some limitations that may potentially obscure small masses. By that we mean less than a centimeter in size, whether the liver's moderately heterogeneous as we see here, or there may be moderate beam attenuation that may obscure some parts of the liver or diaphragm. In this case, there is a moderate limitation here. But we also see, a focal liver lesion. So this is an ultrasound three, positive visualization. B case visualization.

C is where we say there are severe limitations, limitations that significantly lower the sensitivity for detection of focal liver lesions. This is where you think, okay, this ultrasound is worthless. I cannot exclude an HCC in this patient, whether it's because of severe heterogeneity, because of cirrhosis. Maybe there's a tumor, I don't know, maybe that's in background liver, or maybe there's such severe beam attenuation or shadowing that you can't see the diaphragm. We all have these cases or maybe there's so much, obscuration of the liver, whether it's bowel gas or or ribs or lung, where you feel like you're not seeing a majority of the liver. These would all be considered visualization. C severe limitations.

So in summary, assign an ultrasound category A, uh, one, two, or three. These have colors associated with them. This determines management and an ultrasound visualization score. A, B, or C gray scale does not affect management, but informs the expected level of sensitivity of the examination.

Experience at Stanford and Other Institutions

So I just wanna briefly tell you about our experience at Stanford. So we started implementing ultrasound LI-RADS at our institution. And we found, when we looked back at our cases, we found that 90% were coded as ultrasound. One 5% were ultrasound, two 4% were ultrasound three, 85% were visualization. A 11% visualization B and 4% were visualization C.

So you might be thinking, okay, well I'm at Stanford and everyone in California is skinny and not that much cirrhosis. And so it doesn't apply to my institution. But actually Shichi Rogers at Einstein in Penn, Philadelphia looked at her numbers. Her numbers are eerily similar to Stanford's, even though she has a very different patient population. David Fetzer at UT Southwestern also looked at his numbers. He does have fewer percentagewise ultrasound ones, more twos and more threes, fewer as, percentagewise, more bs, but I'm actually surprised that he has fewer Percentagewise CS than Einstein or Stanford. So we'll see how this goes.

Handling Visualization C

I also just wanted to discuss just visualization. See, this is probably the most frequently asked question that I have received. And I actually did a little blurb in the SRU newsletter letter about ask the expert on this particular topic, what should you recommend for those cases? Patients with visualization See livers where you think the ultrasound is just worthless for screening and surveillance. What should you recommend?

And the answer is that you cannot blanketly recommend CT or MR at this time. And the reason for that is that first of all, it's not, it has not been shown to be cost effective, and there is no evidence to show that that is more effective than ultrasound currently. So that said, a lot of patients do get screening and surveillance with CT and MR, we acknowledge that, but that decision really needs to be made on a case by case basis between the gastroenterologist, the patient, everyone else taking care of the patient, not a blanket statement at this time.

However, as we are able to look at our data and we get performance and outcomes data, then perhaps we might, change our recommendations based on our visualization scores. But that'll really depend on what kind of quality and strength of evidence we are able to obtain.

Summary

Okay, so in summary, ultrasound LI-RADS screening and surveillance, it was just released in August. It's fresh off the press. I hope that you implement it at your institution. There are two scores. The ultrasound category score, one, two, or three that determines follow up. And then the visualization score, A, B, or C that conveys the expected level of sensitivity but does not determine follow up.