Focal Liver Lesions: Non-Cirrhotic

Working Up the Incidental Liver Mass

Lots of benign lesions out there, present in 20% of the population at least.

Because of this, we really need a practical and economical and ethical approach to the differential diagnosis.

I am going to postulate that we can diagnose most of these effectively and definitively without biopsy.

But to do so, you really need to consider a number of factors.

I often hear when I speak at meetings like this, you guys and academics get all this information about your patients, and we don't have that in private practice.

Obviously this stuff isn't provided on the tracking form, but we really do solicit this information either from the doctors when appropriate or from the medical records.

All of our workstations have a workstation that can link to the medical records.

We look up this information in a lot of our cases, I would say almost routinely.

We consider the age and gender of the patient presence or absence of underlying chronic liver disease.

Any history of extra hepatic malignancy.

Drug exposure is important for a young woman, for instance, where you're considering hepatic adenoma.

You want to know if she's on birth control pills, other disease processes such as ulcerative colitis, heart disease, polycystic disease, and so forth.

If you do only a poorly performed contrast enhanced scan at some indeterminate time after administering the contrast, basically every liver lesion is going to look like a hypodense lesion in the liver.

If you're limited to saying hypodense liver lesion, every abnormality in the book can be included on that.

We really need to define the characteristics.

More specifically, we wanna talk about the density and the enhancement profile relative to liver, relative to blood pool, which is very important, as you'll see when we talk about cavernous angios.

We need to talk about it depending on what phase of imaging.

If you were to say to me, is a liver metastasis from a pancreatic neuroendocrine tumor, a hypodense or hyperdense lesion, the answer is, it depends on what phase of imaging.

On non-contrast, it's either hypo or iso dense.

On arterial phase, it's usually heterogeneously, hyper dense.

On delayed phase, it's hypodense.

We need to talk about the size and margins of the lesions.

We wanna look for things like a peritumoral halo.

As the name implies, that's almost always a malignant process.

We wanna look for morphologic features like central scars or capsules.

Some lesions characteristically have overlying capsular retraction.

Some lesions, particularly hepatic adenomas and HCC often have spontaneous hemorrhage.

We wanna look for signs of invasiveness.

I'm always dismayed when I read an outside report and here's this mass that is obstructing bile ducts and blood vessels and so forth.

Then I read in the differential diagnosis, this could be a cavernous angio, it could be this, could be that.

I'm thinking, cavernous angios don't bite off blood vessels and so forth.

We want to consider all of those features.

We're not gonna go through this chart in detail.

This is taken from the white paper on incidental Lomas, but I will refer to every element of this chart in prose and in examples during these presentations today.

Hepatic Cysts

Hepatic cyst. These are extremely common.

They're derived from biliary endothelium, but they don't contain bile, nor do they communicate with the biliary tree.

Polycystic Disease

Polycystic disease. While you can consider this in patients who have more than 10 cysts, the number of cysts is pretty irrelevant.

I certainly have seen many patients with simple hepatic cysts that are numerous, and I'm gonna talk about biliary hamartomas, an entity that may be unfamiliar to you, but I guarantee you see it in your practice.

The way we diagnose polycystic disease is by demonstrating cysts in other organs or a family history of polycystic disease.

There are genetic markers as well.

Let's look at some criteria.

Water density, usually, most importantly, no enhancement, much like we talked about with renal cysts yesterday.

No visible wall.

It's okay to have a couple of thin septa that doesn't make it worrisome or neoplastic.

Differential diagnosis, we can consider cystic tumors, abscesses and hamartomas.

The latter occur almost exclusively after trauma or in the setting of liver transplantation.

A hepatic artery thrombosis or stenosis where you get biliary necrosis and a collection of bile within the liver.

We're not gonna talk about that today.

The problem, as we all know, is that small cysts don't measure water attenuation due to partial volume averaging.

Problem solving sonography is very good.

Anytime you're considering a cyst in any organ, sonography is going to show the morphology well, and particularly if you can get the lesion visible on a high frequency transducer.

In a thin patient or obviously in the operating room or endoscopic ultrasound, very good.

MR is a very good problem solver, but too expensive for us to use it in the great majority of cases, just to document hepatic cyst.

Here's some pearls I use every day in my practice.

If I see a small hypodense lesion less than two centimeters in a non-cancer patient, this is almost always benign, and that's just not my own imagining.

This has been documented in huge series of thousands of patients.

What I put in my dictation is indeterminate, but statistically very likely to be benign.

That's a very different statement than cannot rule out malignancy cannot rule out metastases.

That latter is a very bad thing to do unless you've got a real reason to strongly consider that.

Here's another little pearl.

A lesion less than blood density on an unenhanced CT scan is almost always benign.

Either cyst biliary hamartoma or a thrombosed or hyalinized hemangioma.

The latter probably also a term that's unfamiliar to you, but there's a lot of those out there.

Hemangiomas sometimes stay the same for years.

Sometimes they get a little bigger and sometimes they thrombose and turn into a little fibrous nubbin.

If you see an extremely well-defined, sharply defined lesion on a contrast enhanced scan in the liver, that is almost never a metastasis from the usual sources and almost never a metastasis or a lymphoma deposit.

Here is a patient with a number of cysts.

I also use internal standards.

Even if this one, which looks a little fuzzy due to partial volume averaging, even if the ROI did not measure exactly water density, you can see it's essentially the same density as the opacified bile within the gallbladder.

Here's a patient that has multiple lesions in the liver.

The largest of these are clearly identified as cyst, sharp wall, no mural nodularity, no enhancing contents.

An ROI here will in fact show water density.

This patient actually was a cancer patient.

In our outpatient center, I just wheeled this patient across the hall and we did ultrasound.

Every one of these lesions, including the little ones, was a cyst on ultrasound.

Instead of saying on the CT scan, there are multiple lesions too small to characterize in a cancer patient, we have to strongly consider metastases.

What I really said was, these are probably all cysts and this is supported by the ultrasound.

We'll continue to follow some of these as this patient is receiving the chemotherapy and so forth.

But at this point, there is likely no metastatic disease.

Here's a hepatic cyst on CT and ultrasound.

There probably is a little septum in there.

That's okay, don't worry about it.

You note the through transmission, lack of internal echos and so forth.

Here's the simple cyst on coronal and axial MR very bright, uniformly bright on T two weighted images, very dark on T one weighted images.

Here we have some lesions that are obviously cyst and then a lesion here.

If we had only the contrast enhanced scan, this might give you pause because this is obviously higher than water density.

You might say, mural nodularity, this is a cystic neoplasm.

It's very helpful to have the non enhanced scan in this case where we can clearly see that this is clotted blood.

This is gonna measure about 65 Hounsfield units characteristic of clotted blood, putting an ROI on it before contrast and after contrast shows no enhancement whatsoever.

This is spontaneous bleeding into a large cyst.

In this case, we see lots of lesions, dozens, and a lot of people would probably dictate this as being either multiple cysts or polycystic disease.

I would say no.

This is a very characteristic appearance of biliary hamartomas.

If you look at these lesions, there's a little bit of mural nodularity in some of these.

The dozens of lesions that are present are all less than 15 millimeters in size.

Basically, this never happens in polycystic disease.

If you have an adult patient as this is with autosomal dominant polycystic liver disease, there are going to be many, many cysts of varying size, some of them very large routinely in polycystic disease.

There is spontaneous hemorrhage within some of the cysts.

Even in multiple simple, you will almost never see dozens of lesions, all less than 15 millimeters in size.

Biliary Hamartomas

Let's talk about this entity of biliary hamartomas.

I was just writing this chapter the other day for my new book.

In reviewing the information on this and a lot of textbooks continue to say that these are rare lesions.

These are not rare lesions.

I see these on a regular basis in my practice.

They go by a number of other names, including bile duct hamartomas, bile duct micro hamartomas, or micro adenomas, von Meyenburg Complex and so forth.

This is a benign proliferation of bile ductules and stroma.

That's important. Actually, there is stroma in these things, characteristically small, less than 15 millimeters, few to innumerable.

It's actually, I think a common cause for the multiple small, too small to characterize lesions that we encounter on an everyday basis.

Ideologically, this is part of the fibropolycystic disease spectrum.

We've written an article, there's an article in radiographics from some years ago by Brancato and myself about this.

All these entities, and I'll just name 'em here and then we'll go back to this.

Biliary hamartomas, autosomal dominant polycystic disease, Caroli's disease congenital hepatic fibrosis.

They're all part of the same spectrum.

They all occur during the embryologic development of the so-called portal triads and ductal plates, depending on exactly when in fetal development this occurs, you wind up with one or more of these entities.

This is a physician who has this very scary looking liver at an outside hospital.

He was told he had a diffuse metastatic disease, which came as a great surprise to him since he felt fine.

On non-contrast scan. You can see many more low density lesions than on the enhanced scan, which seems paradoxical, right?

Because ordinarily we see more lesions on the contrast enhanced scan.

The reason we see fewer on the contrast enhanced is that these little tiny hamartomas have fibrous tissue in the wall that enhances and makes them shrink visibly.

A lot of these are disappearing on the enhanced scan.

Ordinarily I just look at these and say, this is what it is.

This fellow wanted proof of this and so I did biopsy these and gave a heads up to the pathologist that I was strongly considering biliary hamartomas because they have trouble sometimes recognizing this with that knowledge.

They did make a confident diagnosis of this and he went on to live happily ever after.

Many of these are really a little small to characterize by themselves at one snapshot in time by CT ultrasound or MR.

Sort of paradoxically. Ultrasound actually shows these as being quite hypoechoic.

On CT or MR. They again, innumerable cysts, but they're too uniform in size and small to be polycystic disease in general.

These are hard to biopsy because they're so small.

It wasn't hard in that guy because I could throw a dart from across the room and hit his biliary hamartomas.

Here's some more cases.

These are all very recent, so some of these may not even be in your syllabus.

Here's a patient I see in outpatient imaging and he's got these innumerable low density lesions in his liver.

Now I wanna show you the ultrasound from the same day and you say, I don't see that, I don't see anywhere near this number of cysts on the ultrasound.

I've already told you ultrasound is generally very sensitive and specific for showing cysts.

What I see instead are some lesions, some of which have echogenic stuff in the wall.

Some of the lesions are diffusely echogenic.

If you look in the background liver, this is not a fatty liver imparting this echogenic texture.

His most of his biliary hamartomas are these tiny echogenic things.

What you have is fibrous tissue and fluid within many of these small biliary hamartomas and the sound wave is reverberating there and giving you echogenic foci.

This is the typical appearance of biliary hamartomas on CT and ultrasound.

Here's a patient on non-contrast CT scan.

You can see the cystic lesions interposed among the unopacified blood vessels.

Again, some of them disappear into background liver.

Others are now more clearly defined.

Here's an MR in this patient T one and T two.

This is the very typical appearance of biliary hamartomas and that's the way I dictate.

Dictated this out. The patient had no cysts in other organs, no family of polycystic disease.

This is not a patient who is immunosuppressed where I'm worried about micro abscesses.

This is a healthy patient with an incidental finding.

Autosomal Dominant Polycystic Liver Disease

Here are two patients with autosomal dominant polycystic liver disease.

This patient has a few cysts in the kidneys.

This one has gross distortion of the kidneys.

Notice on this T two weighted image that some of the cysts in the liver come in different colors.

That's because this lesion has had spontaneous hemorrhage at some point in the past.

As we alluded to yesterday in some of our talks, hemorrhage in cysts, whether in the kidney or adrenal or liver, will often result eventually in a ring rim of calcification.

On CT in somebody, an older patient with autosomal dominant polycystic kidney disease, we will often see cysts that are a little higher than water density with a rim complete or otherwise of calcification.

This is a slam dunk.

You don't need a differential diagnosis and so forth.

Biliary Cystadenoma

Now let me mention something that's actually quite an uncommon lesion, but one that you can diagnose and that is a biliary cystadenoma.

I want to emphasize several features about this because this is really gonna help limit your differential diagnosis.

I'll challenge you with a case in the subsequent session on this.

These are almost invariably large solitary, complex multi septate cystic masses.

As my drawing indicates here, there are blood vessels within the septa, it, this is viable tissue separating these cystic spaces.

You should consider these all malignant or pre-malignant.

I had one case in my practice in Pittsburgh that they had followed for seven years without change.

Then lo and behold, at year eight or nine or whatever, they did a repeat scan and this thing had turned into an ugly biliary cystadenocarcinoma.

Biopsy resection, proven biliary cyst adenocarcinoma.

These need to come out and by the way, they need to come out completely.

I've had the fortune of working in two places with great hepatic surgeons and they obviously know this and what'll happen a lot of times in a big lesion as I'm drawing in here, part of the lesion will abut the portal vein or IVC or something and they'll say, I'll say, Hey, did you get it all out?

They go, boy, I got almost all of it out except that part that was adherent to the portal vein.

I couldn't get that quite out.

Then we both kind of cringe 'cause we know this lesion's probably gonna come back.

Here's another important pearl.

These occur almost exclusively in women.

Here's the drawing and here's a case that came along sometime later.

Pretty good match okay, so here is non-contrast CT scan.

As I recall, this lesion had somewhat higher than water density.

Here's a MR in this and you can see that some of the spaces are not absolutely uniform, and that's to be expected.

This is not simple serous fluid in all cases it's got protein in it, it's got varying consistencies.

You will see the septa probably a little better on MR, certainly better on ultrasound than on CT.

But this is a characteristic appearance of a biliary cystadenoma slash cystadenocarcinoma.

Here's ultrasound. Here's CT.

I think this was a very recent case, so it may not be in your syllabus.

55-year-old woman.

Very characteristic middle aged woman is the mental image you want to have for the presentation.

Here again on MR, this is a coronal.

Not all of the cystic spaces have uniformly bright signal.

Might even be some settling in this one here.

We're not gonna call this polycystic disease because it's one big encapsulated lobulated mass.

Here's the ultrasound and what else do we see here?

We see on both MR and on ultrasound mural nodularity.

This has to be dictated as biliary cystadenoma with the mural nodularity strongly suggesting malignant degeneration.

That was proven at resection.

Cystic Metastases

Let's talk about cystic metastases.

They can occur from any primary tumor, but there's two that I want you to think about in particular.

Either the primary tumor is cystic, like a ovarian cystadenocarcinoma or the family of sarcomas.

In that I want you to consider GI stromal tumors, which are a form of, sort of a form of sarcoma.

Metastases from those entities account for the great majority of cystic metastases.

Almost all of these will show complexity on CT ultrasound. And MR. And ultrasound again is a good problem solver.

Here we have T one and T two weighted images.

Yeah, there's fluid in this larger cystic space, but I hope nobody would consider this to be in any way. A simple cyst there is enhancing mural, nodularity a very thick wall.

There's settling of junk within the largest lesion.

This is proven metastasis from a primary sarcoma.

Here's another case of metastasis from a primary sarcoma.

Sarcoma can be anywhere, could be in the lower extremity, it could be in the bowel, could be stomach, could be whatever, enhancing mural nodularity that's tumor every time if it's enhancing right?

Here's a this one actually caught me by surprise because I knew these were complex cysts and I was worried about metastatic cancer.

Here's the mural nodularity can't call these simple cyst.

What surprised me was the origin a thyroid cancer.

That's an unusual source for cystic metastases.

Now let me mention something and I've got a longer version of this.

But when you have treated tumors, particularly treated metastases from GI stromal tumors, we have all of these anti vascular agents and so forth.

You take somebody who's got a GI stromal tumor metastatic to the liver, put 'em on Gleevec, and that tumor subsequently can turn into such a simple looking cyst that you might easily blow that off if you didn't pay attention to what was there on prior studies as a simple cyst.

It's really a treated GI stromal tumor metastasis.

Do your homework.

Look at the prior scans as we of course always need to do in oncologic patients.

Liver Abscesses

Liver abscesses. Again, you would say, that's not an incidental finding and most often it's not.

But I certainly have seen a lot of patients in whom the referring physician was not thinking abscess at the time they ordered a CT scan.

It may simulate a cystic tumor.

Pyogenic abscesses have this sort of appearance, a cluster of grapes or multi septated non often non encapsulated mass.

These often occur in the right lobe of the liver, often in a subdiaphragmatic or per diaphragmatic location.

As a result of the latter, there will often be a pleural effusion, tenting of the diaphragm and some basilar atelectasis.

I do use that to support an inflammatory rather than neoplastic and infectious rather than neoplastic etiology check symptoms.

We will sometimes aspirate these for diagnosis.

This patient was actually sent, this is from my Pittsburgh days as I recall.

It was actually sent to UPMC for please consider resection of this hepatic tumor.

Could that be a biliary cystadenoma?

Well, I suppose, but to me this looked like a septated lesion with a thick wall slightly higher than water density contents.

We called the clinicians and said, would pyogenic abscess even be a consideration?

They went, I hadn't really thought of that, but yeah, maybe.

We stuck a needle in this.

Got pus, we actually stuck a catheter in.

You don't have to suck out all the pus from these.

You just need, what I usually try to do is get most of the pus out of the biggest lesion, get 'em on the right antibiotics, and these will almost invariably go away on their own.

Here's another one that came in with, oh wow, that's a surprise.

We've got this complex lesion right under the right hemidiaphragm note, the right lower lobe atelectasis in a slight pleural effusion.

This was the patient on admission.

I called him up and I said, I don't think this is a neoplasm.

I think this is actually a pyogenic abscess.

This is us sticking a needle in a subsequent catheter into that.

We actually got very little pus out as you might expect because this is not completely liquified yet, but very shortly.

I mean really within hours of starting the, as of doing the aspiration, even though we didn't get much pus out, getting the patient on the right antibiotic, he started to feel better right away.

Let me give you a little anecdote here.

Here they're repeating this scan eight days later and I'm talking to the clinicians on this and they're going, oh, now we're getting nervous.

You know, it's been eight days and this hasn't gone away.

It ain't gonna go away in eight days, right?

This, this lesion can't go away until you get all the pus out until this destroyed part of the liver regenerates.

You have to do some handholding here.

A lot of times they get really anxious and they wanna keep repeating scans.

You go, we got the diagnosis, relax, it's gonna go away.

But just not gonna go away fast.

One of the most common causes of these pyogenic abscesses outside of the more obvious postoperative post-trauma, whatever, I think actually the most common cause is diverticulitis.

It's not the patient who has obvious diverticulitis.

It's the patient who has smoldering low level diverticulitis that's never been diagnosed or treated.

This patient continues to seed bacteria into the mesenteric venous system and they present with right upper quadrant symptoms.

I've got a whole lecture on that as well.

Septic thrombophlebitis usually from smoldering diverticulitis, sometimes from undiagnosed appendicitis.

Hemangiomas

Human hemangiomas a dozen, right?

We see these literally every day in our practice.

Women more than men, we don't, we rarely care about these other than screwing up and mistaking it for something else.

It's basically a bag of blood. With fibro septa and vascular spaces, they can range in size from a few millimeters to 20 centimeters.

They can be solitary or dozens of lesions.

Sometimes it still says this in some textbooks you can diagnose hemangiomas correctly on CT in only 50% of cases.

That is absolutely wrong.

If you use the right criteria and the right technique, you should be able to diagnose hemangiomas confidently in over 80% of cases.

When I examine on boards and I ask the candidates, how do you diagnose a cavernous hemangioma?

They say, you get my B contrast and you get some peripheral enhancement, then you do a delayed scan and it fills in.

Those are not reliable criteria.

I could show you dozens of metastatic lesions and cholangiocarcinomas that fulfill those criteria.

Forget those criteria.

In small capillary hemangiomas, they don't have nodular peripheral enhancement.

They have flash filling.

People say, then I guess you can't diagnose capillary hemangiomas confidently wrong.

You can very large cavernous hemangiomas always have central scars with or without calcification.

Those aren't going to fill in with contrast no matter how many delayed or how long delayed scans you take.

Some people say, then if it doesn't fill in completely, you can't say cavernous hemangioma nonsense.

Here's the criteria that work.

The contents are isodense against the blood vessels on a non enhanced CT scan.

'cause that's what it is, it's blood peripheral nodular discontinuous or cloud-like enhancement isodense to blood vessels.

Yes, most do have progressive fill in.

It's the least useful criterion.

It's okay to have a central scar and small hemangiomas, which are sometimes called capillary hemangiomas enhance homogeneously.

The key is they remain isodense to blood vessels on all phases of imaging.

Here's a typical CT non enhanced isodense to blood vessels.

Portal venous phase nodular peripheral enhancement, isodense to blood vessels.

That's a cavernous hemangioma.

This is happens to be an ultrasound on the same patient done for a different reason.

But uniform bright echogenicity is the typical though not the only appearance of hemangioma on sonography.

Here's a different patient that had sonography for evaluation of right upper quadrant pain.

In this case there's a peripheral ring of very bright echogenicity and another lesion had a peripheral ring and some more echogenicity inside.

We thought those were probably hemangiomas, but we weren't a hundred percent positive.

Patient happened to have a CT.

If I had a follow-up exam, you could do CT or MR.

But here is CT in this patient isodense against to blood on non-contrast uniform, bright opacification, nearly isodense against to hepatic blood vessels on arterial phase, late portal venous phase still isodense against with blood.

This is not HCC, this is not metastatic malignancy.

This is diagnostic of hemangioma.

This patient has two different hemangiomas, each of which is very characteristic of its size.

This small one isodense with blood, a non-contrast isodense with blood on arterial phase isodense with blood on delayed phase.

The larger lesion, most of it is isodense with blood. A non-contrast cloud-like peripheral enhancement isodense with blood.

The enhanced portions of the lesion are actually now it's almost becoming hyperdense, the blood because it's circulating very slow and the blood vessels start, are starting to clear out.

We don't care that these scars never filled in.

You don't have to keep taking images till those fill in.

They're not gonna fill in. These are characteristic diagnostic of hemangiomas.

Here's a lesion that's very large, has a large central scar, nodular or cloud-like peripheral enhancement, isodense with blood.

Here's an MR very bright lesion and that central scar is even brighter.

Now you might correctly say, if you're so damn confident that that's an hemangioma why did you get an MR.

The reason is this is the chairman of the neurosurgery department and as far as he's concerned, MR is truth and CT is a poor, poor substitute.

He has large and multiple hemangiomas.

Here's the one that fools people all the time.

You see a lesion and you say, Ooh, it's hyperdense to underlying liver.

Wow, that's got, there's a narrow differential for that.

I don't even know what even put on my differential.

Then here's the venous phase, enhanced lesion, nodular, peripheral enhancement isodense against to blood non-contrast.

Blood, blood.

This is just a typical cavernous hemangioma in a fatty liver.

Don't compare, I mean you can look at the surrounding liver and compare it to that, but that's not a reliable criterion when you're dealing with a cavernous hemangioma or any other lesion.

You have to account for the fact that this is a fatty liver and that's going to change the relationship, visual relationship of whether a lesion is hyper or hyperdense, hyper or hypodense to the underlying liver.

Problem solving small lesion.

If it's not hyalinized on MR, it may look like an hemangioma cyst.

If it is thrombosed hyalinized, it may look like a hypovascular tumor.

But the ones that are not fibrosed will have the characteristic enhancement pattern that I've shown you.

Many examples of ultrasound, usually the hemangiomas are quite echogenic.

Again, in a fatty liver, a hemangioma may actually be iso or hypoechoic compared to the underlying echogenic liver.

MR. Very bright on T two, same enhancement pattern.

A little note I just added, it may not even be in your syllabus, but do not, if you have a lesion and you suspect this is a cavernous hemangioma, don't do your MR scan with Eovist or Vist.

The reason is it the tate is a not very good vascular enhancing agent.

You'll miss the optimal demonstration of the nodular peripheral enhancement and on the delayed hepatobiliary phase of imaging an hemangioma looks like every other bad lesion including metastasis.

Don't use Eovist to evaluate possible hemangioma.

Here's a we probably did this scan in an acute abdomen setting with, you know, a single phase of contrast enhancement.

I said, this looks like nodular or cloud-like peripheral enhancement isodense against the blood.

I'm about 90% sure that this is just a cavernous hemangioma.

They said, can you get us up to a higher degree of confirmation?

Here's the MR, it's gonna show the same nodular peripheral enhancement and here's the T two.

Okay, now we've got a hundred percent confirmation.

Do we need to do that in every case? No, absolutely not.

But MR, if I had one study to do for definitive diagnosis of a cavernous hemangioma, it would absolutely be MR with one of the usual nonspecific gadolinium based contrast media.

Focal Nodular Hyperplasia (FNH)

Let's talk about FNH focal nodular hyperplasia. Very common.

Sometimes you'll read in, particularly older articles or textbooks. This is a rare lesion. Don't worry about it. It's not rare.

It's present probably 5% of the population.

If you're doing good bolus enhanced scans, you're gonna come across these on a regular basis.

Usually solitary though you can see multiple, usually about two to 10 centimeters in size.

Once again, we don't really care about these at all other than mistaking them for something bad.

Unlike adenomas, these don't bleed, they don't have any potential for malignant degeneration and we just need to make a confident diagnosis.

The etiology of these is believed to be a focal ischemic injury of the liver.

It's scars and then it recruits blood vessels and there is an over exuberant hyperplastic response of the liver.

Knowing that it helps you understand the imaging, there's always a central scar, though we don't always see it on CT.

There's always hyper vascularity and the tissue is essentially normal liver in a disorganized fashion.

It's got hepatocytes, it's got portal venous radicles, it's got bile ducts, it's got Kupffer cells, it's got the whole package.

But in a disorganized fashion in our drawing, we've colored in some bile staining because these do produce bile, but the disorganized bile ductules cannot draw the bile away quickly.

That's gonna be pertinent as we talk about some of the imaging features.

Every FNH that's resected and we shouldn't see many resected FNHs, they always have a central scar that's scar is generally isodense, or I'm sorry, the FNH should be nearly isodense with liver. A non-contrast portal venous phase and delay because it is essentially normal liver, but it's going to be uniformly hypervascular on hepatic arterial phase CT or MR, it's not an encapsulated lesion, though it may appear to have a capsule with a bunch of draining veins that have to suck away all that blood, a small scar while present histologically in all patients, we see the scar on imaging in about two thirds of large FNHs arbitrarily.

I'm defining that as an FNH greater than three centimeters in diameter, but in only a third of small FNHs.

Here are smorgasbord of resected. FNHs all of which have fibrous scar.

Here's a slam dunk, a lesion, a young woman, nearly isodense with liver, a little hypodense on the non-con uniformly hypervascular central scar, large draining veins, which by the way always drain to the hepatic veins, not to the portal vein on late portal venous phase, there's that basket of draining veins.

There's the central scar.

The lesion itself has reverted to being as almost isodense to underlying liver.

This is FNH doesn't need biopsy, doesn't need follow up, doesn't need MR.

This just is, this patient has multiple FNHs.

Each of them is essentially typical of its size.

Here's the larger one, which is uniformly hypervascular with a central scar.

The smaller lesion can't even be seen on the non-contrast scan uniformly hypervascular.

At this point you might say, gee, couldn't that be a cavernous hemangioma?

No, 'cause a cavernous hemangioma would be isodense with blood vessels and this was not.

Then on portal venous phase, the small one essentially disappears.

Most of the larger lesion is isodense with liver.

This sort of looks like a capsule, but these are actually all draining veins.

If you happen to do a real delayed scan, delayed phase scan like 10 minutes, that central scar will typically fill in with contrast.

A good takeaway point is fibrous tissue of any etiology will typically show delayed persistent enhancement on CT or contrast. Enhanced MR doesn't speak to benign or malignant, it just speaks to fibrous stroma.

We're gonna revisit that point a couple of times.

Here's a young man that had a symptomatic FNH.

These are often in my experience, symptomatic when they're hanging off the left lobe because that's in the epigastric region.

It gets compressed by seat belts and steering wheels and so forth.

This person comes in with a symptomatic lesion, MR on arterial phase and enhances homogeneously central scar.

The central scar on T two weighted images is bright, unlike fibrolamellar carcinoma, which we'll talk about in a few minutes.

This is an FNH.

Here's a little more heterogeneous lesion than we usually see, but it's got the other criteria.

Hypervascular on arterial phase, central scar, mostly isodense to liver on the portal. Venous phase, kind of a basket of draining veins into the hepatic veins.

You may see I, I see this sometimes when people send me cases they'll say, you know, I had this patient as a liver and some of the lesions look like cyst and some of them look like hemangiomas or some look like FNH and some look like adenomas or capillary. Hemangiomas, and I'm really confused.

Well, cavernous hemangiomas and cysts are very common and they may be seen in the same liver and that's a random relationship.

But in some work that we did in published, cavernous hemangiomas and FNH occur pretty commonly together and it's not a random relationship.

Here's a little pearl for you.

FNH occurs more commonly in any liver that has any kind of vascular disturbance or other vascular lesions.

Adenoma, ischemic liver, Budd Chiari, et cetera, et cetera.

Here's a patient who has a hemangioma well seen on a T one weighted image well seen on a T two weighted image.

Well, what do we have emerging here?

A very vaguely defined lesion.

That's an FNH bright central scar characteristic on T two weighted image.

Here's the hepatic arterial phase image uniform. Very bright enhancement central scar, incidental FNH and hemangioma.

Here's a delayed scan.

Probably unnecessary, although we pretty routinely do that anyway in our practice to evaluate liver masses.

That scar will show delayed persistent enhancement with the usual non-specific gadolinium based contrast.

Median, not Vist, we're not talking about actually we've seen that patient before with typical findings.

MR criteria for FNH nearly iso intense to liver on unenhanced T one and T two weighted imaging.

Central scar is bright on T two.

Bright uniform enhancement on a hepatic arterial phase.

Prolonged uniform enhancement greater than liver on delayed phase imaging with Eovist.

This is the most specific imaging tests, but usually not necessary.

There are two agents widely available in North America that have both a hepatobiliary and renal excretion.

Gadoxetate dimeglumine or MultiHance is great majority renal only.

A little bit of hepatobiliary.

You have to do the delayed phase imaging at one to three hours and therefore I don't like this one so much if I'm really trying to evaluate the functionality of a mass like an FNH or adenoma and so forth, the gadoxetate or Eovist or Vist or if you work in other countries, Primovist, 50% hepatobiliary 50% renal.

We do the delayed phase imaging at 10 to 20 minutes and I'd like to use this more often, but it's pretty darn expensive.

Here from the literature, small lesion central scar, this is a delayed phase imaging the liver itself is quite bright but the FNH is even brighter because it has taken up, made bile and then the bile can't get out of the lesion quickly because the bile ductules are all screwed up.

If you have a lesion that is uniformly hyperintense to underlying liver on the 10 or 20 minute Eovist enhanced scan, it's a slam dunk that is an FNH.

Some foci of hyperintensity may be seen in some adenomas and some well-differentiated HCCs, but it effectively is never this uniform.

Here's a 23-year-old man we saw not so long ago, again a palpable lesion that bothered him in the epigastric region.

It is pretty homogeneously hypervascular on arterial phase imaging has a central scar.

It's also exophytic, which is part of the reason it was bothering him.

There is. We were about, oh I don't know, maybe 90% sure this is an FNH, the little bit of heterogeneity.

The fact that it was symptomatic in fact that this was a man rather than woman led us to consider, why don't we nail this down with an MR.

Here's the T one and T two weighted images.

The lesion is nearly iso intense with liver.

The central scar is bright.

Now I'm up to about 98% sure that this is FNH, but we had 'em on the table.

We gave him some of the Vist.

It enhances not as brightly as it would if we gave him one of the non-specific or usual contrast agents on arterial phase imaging.

But the slam dunk is you do the delayed 20 minute and this is brighter than the underlying liver in a fairly uniform homogeneous pattern.

This is diagnostic of FNH.

In fact, I would say this is more diagnostic than a biopsy.

Sometimes when you have a lesion where you're saying, eh, it's either FNH or adenoma an HCC in my experience, you give the pathologist a biopsy and they go, it's either FNH or adenoma or a well differentiated HCC.

I don't think we want to abdicate our responsibility here.

We can make this diagnosis by imaging.

Fibrolamellar HCC

Alright, little challenge case here to wake you up in the morning and keep you engaged.

21-year-old woman's got two different liver lesions.

Lesion A and lesion B.

Lesion A is very hypointense to liver on non enhanced T one and quite bright on T two.

Lesion B is invisible to the liver on a portal venous phase image uniformly hypervascular on arterial phase and then promptly disappears.

Again. This is done with Magnevist by the way.

This lesion a bit nodular and this lesion is starting to disappear on delayed imaging and the nodules look nearly iso intense to blood vessels.

Lesion A is a, I think I heard murmuring it's what lesion A is a hemangioma lesion B is FNH.

Here's Eovist as expected.

The FNH can't be seen at all or it's either gonna be completely iso intense or it's going to be hyperintense to liver on the 20 minute delayed.

The hemangioma helpless, hopeless, no help at all on the delayed Eovist it looks like any other bad lesion of the liver.

Eovist is not a good agent to be using here.

Everybody says not everybody but I've heard this about a hundred times FNH and fibrolamellar HCC both occur in young people non-cirrhotic livers.

They're both hypervascular.

They both have a central scar paragraph. Impression.

We can't distinguish the two.

No you can fibrolamellar HCCs almost invariably come in when they are large tumors.

They are heterogeneous. I should have that underlined.

They are always heterogeneous.

The central scar in over two thirds of cases has calcification and two thirds of patients have obvious signs of malignancy, lymphadenopathy, metastases, vascular invasion, biliary obstruction, et cetera.

You really shouldn't confuse FNH with fibrolamellar HCC with rare exceptions.

This is an absolutely typical fibrolamellar.

It's a 21-year-old man as I recall.

Non-contrast, large calcified scar metastatic nodes to the cardiophrenic region.

Heterogeneous enhancement remains heterogeneous.

The nodes enhance and de enhance. Just like the primary tumor, there's biliary obstruction lower down.

There is no way in the world this is FNH.

You will never ever see an FNH with a large calcified scar. Like that just does not happen.

This is fibrolamellar HCC, every time.

Hepatic Adenoma

Hepatic adenoma. Great majority occur in young women.

The rest in glycogen storage disease and people on anabolic steroids.

These are rare relative to FNH mets and HCC.

These often bleed when they're larger.

They rarely have malignant degeneration.

Larger ones should be resected.

50% are solitary, which leaves others that may be multiple.

There is an entity called adenomatosis sort of arbitrarily defined as more than 10 adenomas.

I've seen patients with hundreds of adenomas, rarely pathology.

These are benign lesions.

By and large they've got hepatocytes, fewer Kupffer cells, no bile ducts.

They always have excessive glycogen, often have excessive lipid, often have intra and peritumoral hemorrhage and they often have areas of necrosis which makes them much more complex on imaging by and large than FNH.

I had the artist draw this image and keep this one in mind.

I think I've got a nice case to show you about this.

I said give me a mass that's encapsulated hypervascular and then give me a subcapsular hemorrhage to illustrate the spontaneous hemorrhage.

Let me jump ahead here for a minute and I swear to God I did the drawing and six months later this woman comes in.

I mean is that like a dead ringer for this or what?

Heterogeneous hypervascular lesion.

Subcapsular hematoma, sentinel clot. I love that case.

Alright, so they usually have these other foci that are evident on imaging.

Particular MR no question MR shows these elements of fat necrosis, hemorrhage better than CT.

They're usually hypervascular, they're often encapsulated.

This is something I wrote in an article, but I think I actually see a partial capsule more frequently nowadays.

This woman came into ED, spontaneous bleeding, hypervascular mass, non-cirrhotic liver in a woman slam dunk.

It's an adenoma every time. Here's another one.

CT scan non enhanced.

We have slightly hyperdense material that's spontaneous hemorrhage within this lesion.

Here's hypervascular enhancement of the viable parts of the tumor.

She had other hypervascular tumors in her liver.

Now I mentioned that she's an obese young woman because there's a specific reason for that.

Hepatic steatosis, which is extremely common in obesity, is a fertile field for development of adenomas and we did an article on that a few years ago.

It stimulates the growth of adenomas because it's a hyper estrogen environment and estrogen feeds adenomas.

We're going to see more adenomas with the obesity epidemic that we are in.

Young woman on birth control pills and encapsulated hypervascular mass.

Little pearl for you, that's a pearl in case you were wondering.

Encapsulated mass in a non-cirrhotic liver. Think adenoma. An encapsulated mass in a cirrhotic liver is almost always HCC

MR findings, MR shows features better than CT such as the presence of lipid within the lesion.

At least 70% of cases on MR, only 10% on CT intralesional hemorrhage.

Bright on T one and T two steatosis in the surrounding liver.

MR with hepatic specific contrast. Agents Vist most adenomas are hypointense on delayed imaging because they don't have enough functioning hepatocytes and they don't have bile ducts.

So it looks like a hypointense lesion.

A small number of them will have heterogeneous uptake in retention of Eovist.

In some reports they said this, we confused this with FNH.

The published pictures I've seen, I don't think I would've confused with FNH because it's way too heterogeneous.

Here we have a typical case, arterial phase, CT portal venous phase CT heterogeneous hypervascular enhancement.

Here's the MR as expected hypervascular enhancement on arterial phase T two weighted MR the lesion is fairly bright in phase and out of phase imaging we see dropout of signal from the lesion, therefore it has lipid.

Let's think about this.

We have a partially encapsulated lesion that contains lipid.

It's enhancing like a neoplastic lesion.

It's a non-cirrhotic liver, it's an adenoma.

We have a young woman with multiple hepatic lesions that I can barely see unless I stand on my tiptoes.

They are here we have an outof phase image.

In phase image we have a selective signal dropout, on the outof phase image telling me that these are lipid containing masses.

Multiple lipid containing encapsulated masses in a young woman is adenoma.

Lipid or fat containing hepatic mass in a non-cirrhotic liver is almost always adenoma.

If it's not just a focal deposition of steatosis focal fat in a mass, in a cirrhotic liver is HCC.

Another woman with steatosis on CT we see encapsulated mass with even greater fat deposition within a lesion here.

Contrast enhanced scan encapsulated multifocal mass in a steatotic liver in a young woman think adenomas.

This is the article that we published on the stimulation of adenoma growth and number in the steatotic liver.

This patient we followed for five years with a biopsy proven adenoma and then it got bigger and uglier and had foci of calcification and foci of fat.

We said, you know, you really had to be thinking about resecting this.

They resected it and somewhere in this mess they found foci of hepatocellular carcinoma.

You can also see foci of gross fat and hemorrhage and so forth.

Now I'm not gonna get into it today.

Actually I just learned a lot about this as I was writing the chapter on adenoma a couple weeks ago for the new book.

They have now divided adenomas into three distinct categories based on their genetics and so forth.

It helps to explain the variable appearance of these and the behavior on MR CT and so forth.

There's some nice review articles on that.

I haven't had a chance to put it in your literature 'cause I just came across it myself.

But keep an eye on that.

Just I have to leave it at that.

Not all adenomas look exactly the same.

Some of them, but not all of them have malignant potential and that's all being worked out now and it's in the literature.

I'll refer that to you for further homework.

Summary

Let's summarize. FNH and adenoma. FNH is common.

Adenoma is uncommon. FNH is uniformly hypervascular.

Adenoma is heterogeneous.

FNH you're gonna see a scar in the larger lesions.

Adenomas don't have a central scar.

FNHs don't have a real capsule, but they have draining veins that may mimic a capsule.

You're gonna see a capsule and at least 25% of adenomas, I think actually higher calcification is extraordinarily rare in FNH.

Much more common in adenomas. Focal fat is very rare in FNH.

Focal fat and hemorrhage are characteristic of adenoma on MR FNH Virtually never bleeds.

Adenomas have a bad habit of bleeding.

FNHs are usually solitary.

Adenomas are often multiple, especially in a steatotic liver, FNH on Vist, delayed persistent homogeneous enhancement, slam dunk diagnosis.

Adenomas rarely show persistent enhancement with Vist and it's usually heterogeneous.

In summary, evaluation of patients with liver masses is challenging.

Biopsy may be necessary occasionally for primary and metastatic malignancies.

Having said that, I do a tiny fraction of biopsies of liver masses compared to what we did 10 or 20 years ago.

That's because we make a more confident diagnosis by imaging.

CT plus clinical information is often sufficient to make the diagnosis.

Specifically if you're gonna recommend additional imaging, think about what specific imaging test you're going to recommend.

Hopefully you wanna nail it with one additional test.

If you think it's a cyst, maybe recommend an ultrasound.

If you think it's a hemangioma, recommend an MR and think about whether you want to do the MR with Vist versus just a usual non-specific contrast medium.

Look at the lesion relative to the liver.

Pay attention to whether the rest of the liver is steatotic or iron overloaded and so forth.

'cause that's gonna change the appearance of the lesion And think about the lesion relative to blood vessels.

'cause that's gonna be critical when you're dealing with hemangiomas.