Hi, this is Elliot Fishman, and welcome to our newest lecture on CT evaluation of hepatic masses: a focused approach to signs and enhancement patterns. Now, I will tell you as I put together this talk, I collected a lot of stuff on the liver, and this may be one of our longest talks.

So we'll see how it goes. I have lots of information, lots of interesting cases, and lots of things to share with you. So let's get started. When we think about the liver in general, in over, well, not quite 50 years, the goals of CT of the liver have not really changed when you look at the basics. Detection of the presence of disease, defining the extent of disease, determining what the etiology is (is it tumor, is it infectious, is it acquired?), and to help in patient management decisions (do we need further studies for more information, do we need a biopsy, should the patient go to surgery, does the patient need medical management?). All of these things can be done with a single, well-performed CT scan.

Now, one of the things we know over time, and this still continues, with better detectors, better technology, we pick up ever smaller lesions. Now, in the old days, a small lesion was 2 to 3 centimeters, then it became a centimeter, and now it's typically probably 5 to 8 millimeters. We also know that even though we can pick up more lesions now than ever, it doesn't necessarily mean we're more accurate. And in fact, in many ways, it could be problematic because many of the small lesions we pick up are simply cysts or hemangiomas, and then in a patient with malignancy, be it a colon cancer, be it pancreatic cancer, what do you do? Does the patient have liver mets? Well, you can get another study, but MR, pet scan usually is not going to be helpful on tiny lesions either. Do you do surgery? Do you assume it's benign? Do you assume it's malignant? Perhaps going forward, liquid biopsies will make a difference. But simply detecting the lesion is not enough. And so, in this talk, what I'm going to focus on a lot of is looking at what lesions look like, what they could be, and more, what they actually are.

Now, if I was teaching a computer to look at a CT of the liver, what would I teach the computer? First of all, look at the demographics of the patient. What you're going to see in an 8-year-old or a 78-year-old is different. A big mass might be hepatoblastoma in an 8-year-old, and in a 78-year-old, hepatoma or metastatic colon cancer. What's the clinical presentation? Is it fever? Is it abdominal pain? Is it extrahepatic findings? Is it a lung nodule? Is it an adrenal mass? What else is going down?

And what's the sex of the patient? Certain lesions, hepatic adenomas, for example, are typically much more common in females. So again, that can be very helpful.

Now we look at the scan. Most people, including ourselves, do not routinely get non-contrast CT. In the old days, non-contrast CT was obtained. It wasn't really clear what it was obtained for. We still would do a non-contrast CT if we were looking at a focal hepatic lesion and trying to look at its sequence from non-contrast to arterial, to venous, and delayed. But in routine imaging, we do not do non-contrast scans.

We'll typically do a dual-phase acquisition: arterial phase at about 30 seconds, and venous more like 70 seconds. Occasionally we will get delayed scans, typically at 4 to 5 minutes later. But most of the time, we're doing two phases.

What are we going to look at? Is the lesion vascular on arterial phase or not? Does the lesion enhance on venous phase? Does it change between arterial and venous phases? Is there a central scar? Is there calcification? Is there fat in the tumor? Is the liver normal parenchyma, or is it cirrhosis? If you have a mass in a cirrhotic liver, you're always going to think about ruling out a hepatoma first. Is the lesion single or multiple? Most FNHs are going to be single, but they can be multiple. Hemangiomas are usually single but can be multiple. Metastasis can be single or can be multiple. Hepatoma is usually a single lesion, but not always. So sometimes it can be somewhat helpful, but at times it may not be helpful. And of course, extrahepatic findings: are there lung lesions present like lung metastasis, are there lesions in the spleen, and is there nodes present? Maybe we're dealing with lymphoma. Is there a renal lesion present? Is there something in the bowel or anywhere else in the abdomen? And maybe the liver lesion is going to be metastatic disease. So again, these are the things we think about.

Now, prep, oral contrast: we use water, about 1000 cc's is ideal, this distends the stomach and bowel.

IV contrast, iohexol or iodixanol, usually about 100 to 120 ml, we inject at 5 cc's a second.

On arterial phase, we typically go at about 30 seconds. Some people wrote articles about early phase at 15 seconds, we don't do that. And venous phase is typically at about 70 seconds. If we do get delayed scans, which is infrequent, we'll typically come back at about 4 minutes. One place where a delayed scan can be helpful is if we are thinking about a cholangiocarcinoma, those lesions may become brighter on the 4-minute scan. So those are the things we think about.

So again, talking about all of the things, detection, classification, identification, and management. And we need to put all of that together.

So let's just look at a few basic things we think about when we look at images. This is arterial phase. There's a lesion in the right lobe, there's a blush around the lesion. When I start seeing perfusion changes around the lesion, I often do worry about malignancy. This patient has no cirrhosis. It's a very vascular lesion, which quickly on venous phase fills in. You see, you go from here to here. The filling in, that pattern of filling in is most classic for hepatic hemangioma. Hemangiomas typically fill in from peripheral to central. They may totally become filled in as this case is, but often, particularly with larger lesions, will not fill in in their entirety. Here's the same lesion on the MIP imaging, we can see the hepatic artery feeding the lesion. You can see it here as well on the cinematic rendering where the perfusion changes around the hemangioma are well seen. So we typically like to think that if you see perfusion changes around the lesion, you've got to be thinking about malignancy. But the truth is, most vascular lesions, whether it's hemangioma or whether it's a hepatoma, can have perfusion changes. Other benign lesions and things like ductal dilatation and things like infection can also have perfusion changes.

Now, the washout is important, and you need to be very strict. Here's a patient with a liver is cirrhotic, and there are lesions in the right lobe by the dome. One's very vascular, and one's hypodense. What are we dealing with? Are these two different pathologies? What exactly is this? Is this hemangioma? But remember in the last case, the hemangioma had peripheral enhancement and central filling in. You can get very tiny hemangiomas that fill in with flash filling, but those are uncommon. And when I see a cirrhotic liver like this, and I see that second lesion, I'm surely going to worry, even with just the one lesion, that I'm dealing with a hepatoma. And so in cirrhotic livers, I feel very uncomfortable calling things hemangiomas. Remember hemangiomas are really cystic lesions of blood, and in a cirrhotic liver with fibrosis, these will collapse. Take a look at your next hundred cirrhotic livers, you'll never see a hemangioma. And here's that lesion again, you can see the prominent vessels leading to the lesion, not very helpful. Here it is also on the cinematic display, again, not awesomely helpful in that regard, I'm still worrying this is not hemangioma. And then when I get the washout, you can see the lesion washes out centrally, but there's peripheral puddling. It's kind of the reverse of what I spoke about when I spoke about hemangioma. That other lesion in the right lobe of the liver by the dome is still hypodense. And we biopsied this liver lesion, this was a hepatoma. So, in a cirrhotic liver, you got to be worried about hepatoma. And this lesion did not behave like hemangioma, it wasn't peripheral to central filling in, it was a vascular lesion which became hypodense but the rim became denser, kind of the reverse pattern you'd expect to see with a hemangioma of the liver.

Now, other vascular lesions: here's a vascular lesion in a non-cirrhotic liver. Now, again, you can have a hepatoma in a non-cirrhotic liver, but vascular lesions, particularly like this that are irregular, I got to think about hepatic adenoma. I got to think about metastasis from a neuroendocrine tumor or renal cell carcinoma. I have to think perhaps about an atypical hemangioma. Here, there's no mass effect, but you can see the mass. You can see the hepatic artery going to the mass. And as you look at the venous phase, the lesion is still very well defined. It doesn't have the look of hemangioma, focal nodular hyperplasia typically will fill in in its entirety. We still could be thinking about hepatoma, we could be thinking about the possibility of hepatic adenoma. And here you look at this on venous phase, image on your right, the volume-rendered image showing you nicely the portal vein, which is widely patent. Here's the cinematic rendering showing you the mass, again, that rim-like pattern, but I am concerned. And that lesion washes out, not like FNH, not like hemangioma. It could be hepatoma, but it was hepatic adenoma.

Hepatic adenoma is a very unusual and very difficult lesion. Hepatic adenomas and hepatomas are on the spectrum, that is, you have to worry that hepatic adenomas can become malignant. There are a number of reasons for hepatic adenomas, the classic hereditary one is with glycogen storage disease. Most hepatic adenomas can be related to birth control pills or in males due to androgen use. Hepatic adenomas like the last case can be very well seen and very vascular, or they can be in this case where there's a mass in the left lobe of the liver. It's lobulated, you can kind of see it, but it's easy to walk by. On the coronal views, particularly the volume views, the pattern of enhancement, the increased vascularity, the mass effect, is more obvious. You can see on the MIP imaging stretching of the left hepatic artery. And you can see that when the lesion goes to later phase, it washes out. Yes, I can see the mass effect there, but the lesion is not quite, but nearly isodense. Hepatic adenomas can become nearly isodense.

Now, one thing I've learned, when you use cinematic rendering, you get a lot of good texture mapping information, and we've spoken about this before, but you can see that hepatic lesion, which looked nearly isodense, when you do the texture mapping is very obvious. And so, sometimes if you're not careful, the windowing will make a lesion look like it's gone, the cinematic will show it better that it's really still there.

Now, another case, again, we're looking at our protocol arterial phase. Here's a very vascular lesion, very well defined, the liver has mild textural changes. You know, one of the things about hepatomas, you always think about big irregular lesions, lots of neovascularity, maybe necrosis, maybe calcification, but hepatomas can be very well defined and have a pseudocapsule. And if you look at the same case with MIP imaging, look at that neovascularity within that lesion, very nicely shown. Here it is on the cinematic rendering, again, showing you this dominant mass, a very nice appearance to that mass, very well defined. Here you can see it very well defined with a pseudocapsule on the later phase imaging where the lesion washes out, and that was also a hepatoma.

Now, when we think about the liver, again, I think about benign lesions and malignant lesions. In the malignant category, we typically think about hepatoma, maybe cholangiocarcinoma, metastasis, we don't often think about lymphoma. This is a patient who had fever, some weight loss. When you look at the liver, there's some mild textural change on the arterial phase, not all that substantial. And then if you look hard, it looks like the vessels off the hepatic artery are somewhat splayed. When you go from the image on your left, the arterial phase, to the venous phase, look how much more obvious the lesions are. This was lymphoma. Now, it's interesting, lymphoma is one of the lesions that shows best on venous phase. Again, looking back on the arterial phase, you can see the lesions, but nowhere as nicely as you do on the venous phase imaging. Again, could this be metastasis? The answer is yes. Multifocal hepatoma? The answer is probably no. I don't see any liver lesions, I don't see any splenic lesions beyond the widespread liver lesions, but I don't see anything that would push me to hepatoma or push me to lymphoma. So again, very tricky.

Another case: arterial phase, there are infiltrating lesions in the liver, kind of a funny pattern. Almost looks like some of these might be infarcts, but it's involving a substantial portion of the patient's liver. I don't see adenopathy, I think there are textural changes within the liver proper. And then when you go to venous phase, look at all of these vessels. There's prominent vascularity everywhere. People thought this might be atypical hemangiomas that just didn't enhance well early, and it kind of looks like hemangiomas in some ways, but there's too many of them, and it doesn't have that peripheral pattern of enhancement. I got to be thinking something else. Could this be vascular metastasis? It's a little tricky because why did the arterial phase look so good? Essentially vascular mets almost always, always is a big word, but almost always are better seen arterial rather than venous. What could we be dealing with here? We just weren't sure. This patient ended up with a biopsy. Look at those multiple areas of vascularity in the liver, the textural changes. It looks like the lesion has also bled. This ended up being a very unusual tumor, which was an angiosarcoma of the liver. We talk about angiosarcomas of the spleen, we talk about angiosarcomas of the liver, both are going to be very, very unusual tumors.

Now, in this talk, I want to go through very specific lesions, show you the classic appearance, and then also show you some of the non-classic findings as well. We're not going to have time for parenchymal disease, we'll mention it only if it relates to the primary masses in the liver, and we are going to talk about infectious lesions, only in part because they can be great mimics. Patients with FUO can have metastasis, can have lymphoma, can have hepatoma, or can have a liver abscess. So you at least need to be thinking about that when you're making a differential diagnosis.

Now, I mentioned before incidental lesions in the liver are one of the challenges to us, particularly when they're small, they're usually hamartomas or hemangiomas or cysts, and they can be challenging because how do you manage them? Do you simply follow them, or do you simply ignore them? This white paper looked carefully at the incidental liver lesions, dividing them up into categories: under 1 cm, 1 to 1.5 cm, or over 1.5 cm. And again, typically, under 1 cm, it would ignore, over 1.5, it would be more aggressive. And it became important looking at this, to try to figure out how good we were in determining what tumors were really present and what was their etiology. So let's do this: let's stop here, and let's pick up with benign hepatic tumors, and say, let's take a 10-minute break. And I'll see you in 10 minutes. Bye, guys.