Alright. Thank you very much for the invitation and as I mentioned or as was mentioned, I'll be talking about the non-invasive approach to diagnosis and just a quick disclosure slide and just some learning objectives that vary a little bit from what you have in your syllabus. So here's the case that we are presented. A 61 year old man with alcoholic cirrhosis. Very important to recognize that it's an at-risk patient for HCC. Routine MRI and found to have what's described as an unusual mass, two centimeters within the liver. This was described as having a thick enhancing viable rim and diffusion restricting rim. So one thing I want to point out is that we do have a very precise approach to the diagnosis on imaging and this is going to be provided by Lyrad. So if you want to learn more about that, I'm not citing it throughout, but basically Google Lyrad's ACR. There's a manual that has over 300 pages with 16 chapters and it provides four individual imaging contexts for HCC diagnosis and we start with screening surveillance, which is predominantly referring to ultrasound. Two diagnostic manuals here, both CEUS and then CT and MRI, as well as a treatment response assessment manual. And the one thing I do want to point out is that I kind of used quotations when I'm describing the case that was provided because the terminology used in the case, unusual, rind, viable, these things are not necessarily the approved lexicon that we want to be using for our imaging studies. So that's one other thing that you can find are kind of the approved terms. And as we move forward, we're trying to unify kind of worldwide the terms that we're using to describe HCC and the imaging findings. So with that, we're going to focus here on the CTMR algorithm because we're provided with an MRI case. So this algorithm is meant to be applied in a stepwise process and really if we follow those steps along, the idea here is that we can eliminate a lot of other entities and what we should be left with are solid hepatocellular nodules. And of these nodules, we can then define whether they represent a definite HCC according to stringent criteria. Those are LR5s or if they represent a probable HCC or one a nodule of intermediate probability. So let's go ahead and walk through this stepwise algorithm just so you get the flavor of it. The first step is that we want to assess, do we have adequate imaging to evaluate the nodule at hand? And really what we're looking for with the imaging on CTNMR is a multi-phase contrast enhanced study. So we need to have multiple phases so that we can assess the enhancement pattern and assign probability. If we don't have that, you might come up with a report that says LRNC. That's not categorizable. So maybe we see an observation. We don't have enough information to really narrow that differential down meaningfully into something that's benign or malignant. The next step and the reason this happens so early in the algorithm is that this algorithm initially was designed to be used in the transplantation setting. Though now we have unification with the ASLD for these diagnostic criteria, initially we were sort of mirroring along with the OPTN criteria and so given the transplantation context, we really want to exclude tumor in vein and advance disease fairly early on. We do that by identifying unequivocal enhancing soft tissue within the vein. We keep moving down the list. Now we're going to exclude benign entities. Probably most importantly, we'll be excluding things like hemangiomas and excluding things like arterial portal shunts so that we don't mistake them as having arterial phase hyper-enhancement and representing intermediate or probable nodules. And then here's where we're going to get stuck today based on the case that we have. And that is with this category, the LRM. So LRM is a probable or definite malignancy, but it doesn't have features specific for HCC. And really this is the next question before we get to that sort of hepatocellular nodule category is that we want to exclude things that are not hepatocellular in origin. And that is the primary reason for the LRM category is basically so by the time we get to the LR5, definite HCC, we have a very highly specific diagnosis for HCC. We have a very highly specific diagnosis for HCC and a very high positive predictive value for HCC on the imaging. So what is LRM? LRM, as I mentioned, is going to be malignant, but not specific for HCC. And we actually do provide fairly specific features for LRM. These are the targetoid features and that applies to three different aspects of the imaging study. The dynamic contrast enhanced, so that's that multi-phase post-contrast imaging. The diffusion-weighted imaging. And if we're using E-A-Vist in the United States or PrimaVist in Europe, then we're going to also apply this to the hepatobiliary phase. So why targetoid? It's kind of a funny sounding name. We chose that because basically it represents sort of the lesion having different constituents throughout. And we recognize that targetoid tends to refer, most commonly in the literature, to things like cholangiocarcinomas, intrapathic mass-forming cholangiocarcinomas. And we can look at the pathology of these lesions and understand that peripherally we have high cellularity. Centrally, rather, we have necrosis and fibrosis. And so that helps to inform some of these imaging features that we have. So if we talk about the targetoid dynamic enhancement pattern, the RIM-AFI peripheral washout, corresponding to the outside of the lesion or the periphery of the lesion, happens to correspond to that cellular component, typically. The delayed central enhancement can be explained intuitively by thinking about the contrasts getting trapped within that extracellular matrix of fibrous tissue. So here are just a couple of examples. So this is peripheral RIM-AFI. Just another example of RIM-AFI. And here's just a big lesion with RIM-AFI. Now, this doesn't mean that it can't have arterial phase hyperenhancement in other components, that it's not in the periphery, but when you see RIM-AFI, that is sufficient for saying that this could be an LRM lesion. So RIM-AFI is probably the most important feature that we use for discriminating between lesions. It's most commonly seen in cholangiocarcinomas. But just to make note, we do actually rarely see it in a minority of HCCs. And it's been described in combined tumors. And why RIM? So RIM-AFI is in contradistinction to non-RIM-AFI. Non-RIM-AFI is the central or confluent enhancement that we tend to see and ascribe to HCC. So let's shift gears and talk about the targetoid diffusion-weighted imaging. So again, we can think back to the cholangiocarcinoma, the inter-hepatic mass-forming cholangio that has peripheral cellularity. We know diffusion-weighted imaging is a surrogate marker of cellularity. So we're going to expect the highest degree of diffusion restriction within the periphery of that lesion. And that's what we mean when we say targetoid restriction. On this ADC map, you see that there is a dark RIM around the periphery of this lesion. This is a targetoid pattern. Here are other more obvious targetoid patterns. Hepatobiliary contrast. So Eovist is taken up actively by the hepatocytes, by transporters, but that's not what we think is responsible for this targetoid appearance that we see in cholangiocarcinomas. We don't expect cholangiocarcinomas to take up this contrast, but we do see retention of the contrast within the lesions. And we believe it's, again, probably due to some trapping of that contrast agent within necrotic areas, potentially within fibrotic areas. It just cannot leave and it's not being taken up by the cellular membrane transporters and it's not being excreted into the bile. So we see sort of trapping in this targetoid appearance. This can be a transient phenomenon. So we oftentimes will acquire a couple of different hepatobiliary phases. You may only see it on the earlier of them. And this is a well-documented feature in the literature and supportive of a diagnosis of a non-HCC malignancy. In addition to targetoid features, we also talk about non-targetoid features for LRM. And that is if you don't have an LR5 lesion and you don't have TIV, you have something that looks malignant and has these features, it would also be an LRM. So what are these features? These are marked diffusion restriction. So this is not a targetoid mass. This is not a classic HCC, but it certainly looks malignant in the right population. And we've got marked diffusion restriction. So this would be an LRM. Infiltrative appearance. So we think about this with regards to HCC. And this can be in the case of a ceratomimetic type of growth pattern. And here we can see very distinctly this area on a hepatobiliary phase that does not take up the contrast. And then we compare that to the background liver where you do have retention of contrast. Diffusion can be very helpful for identifying this region of infiltration as well. But when we have an infiltrative mass that's not an LR5, it's clearly malignant, but it's not showing us those specific features for HCC. So that, again, is going to be classified as an LRM. Necrosis or ischemia. We already mentioned cholangiocarcinomas notoriously can do this. So we have regions of non-enhancement, non-liquefactive regions of non-enhancement that represent ischemia on imaging. Other features. We leave up to the discretion of radiologists. Things like bile duct dilation, capsular retraction. These also go along with features that we typically see with cholangiocarcinomas. So now that you have kind of an understanding of what the LRM lesions are, let's think about what these actually represent when we apply them to imaging studies. And primarily what we see in the literature is that LRM observations are cholangiocarcinomas, combined tumors, specifically the combined hepatocellular cholangiocarcinomas, metastatic disease and important to recognize this category here of atypical HCC. There was a systematic review in 2018 looking at this and I think the takeaway point here is that if something's called LRM, highly likely to be malignant and it doesn't exclude the diagnosis of HCC. About a third of them were actually atypical HCCs. And here's just an example of that. If you're very astute and I'm not sure how well it's showing here, but this is a pre-contrast image and we have some bright signal in there. That's hemorrhage. We have RIM AFI. This is actually an atypical HCC that has RIM AFI, probably because the center of the lesion has bled. Here's just another example, RIM AFI and then progressive enhancement on biopsy. This was called LRM but was biopsy. This was an HCC. Combined tumors, important to recognize that these can have features of both HCC and cholangiocarcinoma. Although in the literature, they predominantly look like cholangiocarcinomas. And here's a nice example of targetoid appearance on hepatobiliary phase. Abscess. We always have to, in the back of our mind, think about this because it's never something that's our first thought. But my pearl on this is if you have a very peculiar looking lesion in the liver and it doesn't look like anything you've seen before, it could be an abscess. And you must ask the question as a radiologist, are there any symptoms, fevers, why blood cell count, anything, you know, a GI infection that could explain that. And always in the back of your mind, at least consider that possibility. So we talked a lot about LRM as a radiologic category. But I want to think about it too as a radiologic phenotype. And as a phenotype, it's important to ask the question, is it actually prognostic? And I think there's emerging literature to suggest that the LRM appearance is, in fact, prognostic. So what is it? Why is it prognostic? It's not entirely clear. But if we look at the subtypes of HCC, we talk about atypical HCCs. Could it be that we're predicting the macrotrabecular mass of HCCs? Some emerging research from France suggests that these have massive central necrosis in them. Could it be that we're predicting the scurus HCCs? There's not much research out there on the imaging appearance of neutrophil-rich. But some of it suggests that they may have sarcomatoid features. Sarcomatoid HCCs in the literature are typically LRM. So let's just look very briefly at this study by Dr. Choi's group. And here we have a Lyrids classification and prognosis of primary liver cancers using Eovist. So they had a group of HCCs, combined tumors, and cholangiocarcinomas. And if we look at it according to their survival based on the pathology, the HCCs do the best. And the cholangios and the combined tumors do worse. Let's look at it according to LR category. Okay, so the LR4s and 5s do better than the LRMs. So I think we're kind of getting a picture here. And then if we actually break it down by both and look at specifically the combined tumors, the combined tumors that have LR4-5 phenotype do better than the combined tumors that have LRM phenotype. So it's possible that that imaging may be telling us something more even than the pathology itself is. And here's another study of primary malignancies after resection. And here we can see that there is in fact a worse prognosis for LRM HCCs after resection than for LR5 HCCs. So why is it prognostic? It's possible. Even if you get the pathology back and it's HCC, perhaps we're predicting CK19 positive HCCs. Perhaps we're predicting or recognizing combined tumors. So I think the main question here is can we use this imaging to guide therapy? If we have an LRM, we biopsy and it comes back as an HCC, should we treat it differently than we would treat just a typical or conventional HCC? Unfortunately, we don't have any prospective data yet to guide us in this decision. So let's take a look again here at our case. As I mentioned, this is a patient who's at risk, so we're going to apply LIRADS. It's an MRI, so we're using the CTMR algorithm. And we have this lesion here. If we look at our features that we've discussed, it has RIM AFI. It has peripheral washout corresponding to the region that had AFI. And it has central delayed enhancement. So this is an LRM lesion with a very nice example of a targetoid enhancement pattern. LIRADS does provide a management algorithm, but we don't go very far in telling you exactly what to do. Basically, we recommend multidisciplinary discussion for the tailored workup and probably biopsy if it's an LRM to help guide therapy. So the key take-home points here is that I want to address the fact that when you have a lesion in an at-risk patient, LIRADS does provide a lexicon for us to use so that we're all using the same terminology and speaking the same language. And it provides a diagnostic algorithm that helps to hone that probability of HCC and hopefully to achieve a specific diagnosis by identifying lesions that are not HCC and those falling into the LRM category. And finally, LRM lesions are almost all malignant. So they're not something that we want to ignore when the radiologist says that there's an LRM lesion, although it doesn't necessarily mean that if you have a transplant candidate that they would entirely be excluded from that consideration because we do have to consider that there are still some HCCs that can appear as LRMs. And finally, to establish that diagnosis and further direct management, oftentimes biopsy comes into play with these lesions.

liver cirrhosis

MRI

LIRADS

LRMs

cholangiocarcinomas