Thank you very much. And I thank the happy SIG for inviting me to give this talk. So I'm not going to talk too much on other barriers and strategy for XCC prevention besides risk scores. So these are my disclosures. So before we talk about risk score, we have to understand what risk scores are. Actually, there are too many factors that may influence the risk of XCC, be it patient factor, virus factor, disease factor, and other comorbid illnesses. And a lot of doctors find extremely confusing to assess the risk of XCC because there are too many factors, and the weight of these factors may be different in different patient groups. So to make the lives of doctors easier, some researchers try to make use of the very big longitudinal patient cohorts and pull all these factors into a computer, perform multivariate analysis, and identify independent risk factors for XCC. And then according to the weight of these independent risk factors, they generate scores. And these are very simple scores that can be used by clinicians to assess the risk of liver cancer. So initially, this is actually an academic exercise, and most investigators have not thought how to use these scores in clinical practice. But at least with these scores, we should be able to differentiate or classify patient risk. So there will be patients with low risk, there are some intermediate risk patients, and there are some high risk patients. But this is not good enough for XCC prevention. For a score that can be used for XCC prevention, it has to pass two tests. One test is risk reduction. Can it identify patients for antiviral therapy? Because with antiviral therapy, we can reduce the risk of cancer. And second is for prognosis improvement. Theoretically speaking, high risk patients will be treated, then whether these risk scores can tell us which patients should undergo XCC screening and which patients do not need XCC screening. So let's see whether these scores can pass these two tests on risk reduction. So my question is, when to start antiviral therapy? Can XCC risk score help? So the first XCC scores being derived in untreated patients predominantly come from Asia. There are two scores coming from Hong Kong. One is from our cohort from CHK, and other is from our sister university in Hong Kong. And the other is the biggest cohort from Taiwan, the review HPV cohorts. The Hong Kong cohorts have mixed patients with cirrhosis and non-cirrhosis. The review cohort is predominantly in non-cirrhosis patients. And when the review HPV cohort score was generated, cirrhosis patients were deliberately excluded because we believe that these patients should be treated and risk score is useless. So if you look into the parameters, actually HPV DNA is in the score. Cirrhosis is in the score of the Hong Kong cohorts, and also other parameters like age, and sometimes gender, sometimes ALT. So different scores using different risk factors. But nonetheless, besides these scores are proven by the big cohorts in the original study, these scores are also externally validated. So this is only one of the examples. These three scores were validated in Toronto in more than 2,000 patients, 79% Asian, a quarter cirrhosis, and a third treated by interferon on NA. And the scores can be validated in both Asian patients and non-Asian patients. And all the scores can differentiate patient with a low risk group, intermediate risk group, and a high risk group with different XCC risks. So far, so good. So if you look through deeper into what it exactly means, so I try to use the estimated five-year XCC risk predicted by all these scores. And you must, number one, look into the estimated XCC risk in the original cohort. Because if the original cohort have a high cancer risk, then of course, the patient with high risk will be more, and there's a higher chance that the patient will have an overall higher risk in the entire predicted scores. So looking at the CXCC score, GACXCC score, and RISC-B score, if you think about starting treatment, then probably we should start treatment in those with intermediate risk and high risk, but not the low risk. Because if you also start with an old risk, then these scores are useless. So apparently, the CXCC score and GACXCC score, we start treatment beyond at least 2% in five years. In RISC-B score, it's most difficult, because overall, the cancer risk is lower. There's no cirrhosis patient in development cohort. And intermediate risk actually bridges from 0.5% to 3.3% in two years. Of course, we should treat all patients in high risk scores. But now, we have guidelines. We just heard from Professor Torod that all three regional guidelines actually make use of disease severity, including fibrosis, cirrhosis, inflammation, that is ALT level, HPV DNA, and also classify patient according to the angiostatus. And all these are based on a lot of previous studies suggesting that patients with active liver disease, high viremia, and cirrhosis, they have a high risk of progression into liver complications and cancer. And therefore, we should treat these patients. So in short, for patients with cirrhosis, all require treatment. For those with significant histology, they require treatment. And if they have high ALT level because or related to high HPV DNA, then they should be treated. So let's dissect to see whether the scores are better or worse than the guidelines. So the first question is treatment of non-cirrhotic patients. If you look at all these scores, actually, all the risk scores mainly based on HPV DNA and age to classify the age of patient, whereas Yelingen and ALT are largely unimportant in the scores. And risk B score only apply to patients older than 30 year old. So we look into the scores. If HPV DNA is more than 200,000 IU per mL, which is pretty high level compared to guidelines. Guidelines, we talk about 2,000, 200,000, 20,000. But now I use 200,000 IU per mL. If you see HCC score, we only treat patients beyond age of 50. If you get HCC score, we almost do not treat anyone. In risk B score, we almost treat everyone. Men or age over 30, or ALT abnormal, older than 45, or E-positive patient. So if you use the scores as compared to guidelines, if you use HCC score, you probably undertreat. You just risk B score, you probably overtreat. You just get HCC score, you never treat. The second challenge is in cirrhotic patients. Now in all regional guidelines, all cirrhotic patients with active viremia should be treated. Of course, in risk B score, again, cirrhosis was excluded. Now we go to the free risk scores. If you see HCC score for cirrhosis, immediately go to high risk group. So again, all should be treated. Get HCC score, you only treat men beyond age of 53, female beyond age of 67. In other words, you will not treat young cirrhotics. In risk B score, you should treat all cirrhotics because this is the assumption of the score behind. And therefore, again, CCC score, well, it concurs with guideline. Risk B, not applicable because you do not need the score. In get HCC score again, you are not treating the right patients. But then you see that diagnosis of cirrhosis is very important in the scores. And in the guidelines, it's recommended we should use non-invasive assessment, or liver biopsy whenever we see patient with active viremia but low AOT levels. And all patients need cirrhosis assessment in risk score estimation because CCC score, get CCC score, this is important criteria. Risk B score is not applicable in cirrhotic patient. Again, cirrhosis patients should be treated. So you can see that cirrhosis is a very big weighting in all the scores, and you need to make the diagnosis right. But sometimes this is a challenge. And in the risk score estimation, there's no guideline when you should assess fibrosis. It seems that for every single patient, you need to know the cirrhosis stage. Therefore, there is some modification of some scores. One example is a CCC score, and we modify the cirrhosis with liver stiffness measurement. So the LS-MHCC score is an upgraded version of the CCC score with slightly improved performance. There's no criteria for cirrhosis, but we use liver stiffness instead. So liver stiffness below eight, eight to 12, and beyond 12. And all other parameters remain the same as a CCC score. So if you use the LS-MHCC score to decide when to treat, then you treat all patients with liver stiffness measurement beyond 12, but most of these patients should have advanced fibrosis or cirrhosis. For those between eight to 12, normally we call them in the gray zone. Then we treat patient beyond age of 50 or DNA more than 200,000 IU per mL. For those with LS-MHCC less than eight, usually not advanced fibrosis. Then we need to treat beyond age of 50 and DNA beyond 200,000. So we use that to guide treatment as compared to guidelines. It's more or less comparable, except patients with lower DNA than 200,000 and non-cirrhotic. So again, it slightly undertreat patient even if you use LS-MHCC score. So overall, if you compare the risk scores with guidelines, it seems that they do not match exactly, and then you need to pick whichever one you want to follow. Now the second test is prognosis improvement. It's when or when not to perform cirrhosis or X-ray surveillance. With this one clinical challenge is that patient with significant risk of cancer should be treated with antiviral therapy. So in other words, all antitreated patients should undergo X-ray surveillance. But theoretically speaking, the cancer risk can be reduced by antiviral therapy. So my question is, is there any antitreated patients who do not need X-ray surveillance? So it can save a lot of money, a lot of resource without increasing the risk of missing important cancer. So we try to use the scores derived from untreated patients. In antitreated patients, it doesn't work. So this is a very good study looking into the use of the free scores derived from untreated patient in Asia. In ED, we treat the patient and the performance is very poor. So you cannot use those scores. So there are a few reasons for that. Number one is those scores are derived from untreated patients. Therefore, HPV DNA is irrelevant in antitreated patients. So they should not be included. And cirrhosis also may regress on a treatment and the risk of efficacy may be reduced and there may be some differences between Caucasian and Asian patients which is actually speculated in that particular paper. So among treated patients, the first score developed is the PHB score which is a very simple score, only make use of platelet count, age, and gender. So there's no HPV DNA there. Age and gender are general demographic risk factors of cancer and platelet partly reflect portal hypertension and partly reflect the severity of cirrhosis. And this is a very carefully done study with 1,300 patient in D-reference cohort and validated by 500 patients. In the low-risk group, that is PHB score 710, no patient in the original study developed cancer in five years, so pretty good. So after the PHB score, there are a few Asian scores derived, again, taking away HPV DNA from their score and trying to predict XCC risk in antitreated patients. So this included one case called M-RICH-B score which, again, used liver stiffness measurement to replace HPV DNA in RICH-B score. So M-PAGE score which adds serum albumin to the PHB score and also another CAMD score derived from Taiwanese and validated in Hong Kong. So you see no HPV DNA there. So there are a couple of studies looking into the XCC risk of these four scores in antitreated patient. The initial four studies are mainly from Korea. The Taiwan study is a study derived the CAMD score. And the last study is actually one we just completed. Our analysis in Hong Kong paper is still in preparation. It's the biggest study so far to validate all these scores. I want to highlight that the M-PAGE-B score and PHB scores are these scores that with the best data and it shows less than 1% XCC risk in five years. So if I try to use this particular score that is a low risk group, that I provocatively speculate that maybe they may not need XC surveillance. It's because if the five-year risk is below 1%, the estimated annual risk is below 0.2% which is below the cost-effective threshold ASA recommends to undergo XC surveillance. What does it mean by having a low PHB and low M-PAGE-B score? So it means that if it's a man, then he has to be young at the age of 40 and the platelet has to be over 200,000. If it's a woman, then if the platelet is very normal, she has to be under the age of 60. If the platelet is between 150 to 200,000, then she has to be under the age of 50. But with a normal M-PAGE-B, the serum albumin has to be over 35. In other words, there should not be any decarbonization. So it seems that it looks pretty good, but I have a word of caution. It's remember all these scores are developed in a cohort of patient with a mixture of non-serotic and serotic. If you just look at a serotic, I'm sure the M-PAGE-B will go down and estimated cancer risk will go up. So probably these scores may not be applicable to purely a cohort of serotic patients. But there are also other caveat is these scores have not taken into consideration of functional cure, as Jordan just mentioned. But this is a recent paper we published among a large cohort of patient on NA. There is a small group of patient having undergone ST recurrence. You can see the XTC risk is much, much lower than those on NA, but with complete viral suppression. The eight-year cancer risk is 0.6%, with an estimated annual risk of 0.075%. So probably XTC surveillance is also not necessary. So take away the message from this talk that is where XTC score can help to prevent XTC, for XTC prevention. Number one, these scores are simple. They are similar weighted score based on demographic and clinical variables. But if you use this risk reduction, that is to decide when to set antiviral, it seems that it's not matched well with the guidelines. For non-serotic, the XTC score will undertreat, which will overtreat, get XTC score, never treat. Cirrhosis, only the XTC score matches the guideline. If you use liver stiffness measurement, it can avoid the diagnosis of cirrhosis, but still it may undertreat a little bit. For cirrhosis, for prognosis improvement, whether we can take away some NA-treated patient from XTC surveillance, well, we can apply the PHB score and NPHE scores, and for young patients, female patients, and those with normal platelet, probably they may not need XTC surveillance. With that, thank you very much. Thank you.

risk scores

liver cancer prevention

patient factors

antiviral therapy

cirrhosis