Basics of MRI Image Interpretation

Basic Aspects of Interpreting MR Images

In the next half hour, I'm going to talk about some of the basic aspects for interpreting MR Images.

We're gonna go through a little bit of the lexicon as we go along.

Mostly I'm gonna show you a lot of different images.

When we approach Mr Interpretation, it's important to remember that we really need to know the reason that we're doing the examination.

And we also need to know most importantly, the history and risk of the patient, and also the mammography and ultrasound findings.

When we sit down to interpret best, ma, we really have all of that information to us, including the prior imaging 'cause it's important to correlate the standard mammography and ultrasound imaging with the MR imaging as well.

And then we look at the morphology.

We are very familiar with morphology from those of you who do mammography and ultrasound, and I really do think that in order to interpret breast MRI, the radiologist needs to have a thorough grounding in ultrasound, MRI, all aspects of brace breast imaging and breast pathology, et cetera.

And then we also, in addition to looking at the morphology, which we have very good idea about from our other breast imaging studies, we need to look at the kinetic characteristics of the enhancing lesions that we identify on MRI.

And it's certainly true that for interpretation of masses on MRI, it's fairly straightforward.

We're very familiar with the morphologic aspects.

The real challenges in diagnosis are characterizing foci and non mass like enhancement now called non mass enhancement.

Introduction to Images and Protocol

Just by way of introduction to the images, I'm gonna show you the protocol.

Most of the images that I'm gonna show are taken from a 16 channel system at three Tesla.

Some of them are 1.5 also with a 16 channel system.

The T two weighted images are not fat suppressed.

Reason being, we want high morphology in this sequence.

So we have one pre contrast data set consisting of a T two weighted TSE image with the same slice sickness as the T one.

So they're matched slice, slice by slice for the subsequent enhanced images that you're gonna see on the T 1D CE part of the study dynamic contrast enhancement.

Then we do a pre injection T one, inject contrast.

And then we do a number of post six post contrast images.

So this is our protocol.

So the dynamic sequence takes seven minutes.

The T two pre contrast non-fat suppress takes about four to five minutes.

So the total exam time is less than 15 minutes.

We match spatial resolution on T two and T one in plain resolution is 0.8 isotropic voxels.

And here we can see just down below the type of images you're gonna be seeing.

This is a T two 80 non-fat suppressed image showing a lesion here, moderate high signal, pre contrast T one.

These are all source nons subtracted images showing an enhancing mass correlating with the T two finding here.

Here, you can sort of see it on the pre contrast image as well because it's a fatty breast, subtracted image.

And then here we have the angio map, which the CAD produces giving us an idea of the color coding of the delayed face.

And we can create a time course curve.

So these are the basic images that we're gonna be looking at and the basic images that you need to in interpret a breast MRI study, some just by way of explanation, some many sites will not use this kind of pre contrast sequence that we use and will use a fat suppressed T two and then add a non-fat suppressed T one image.

So you may be doing that in your practice.

We collapse 'em both into one sequence.

So it's important to look at the images, you look at the mammography and ultrasound images if they're available, if they show a finding before you look at the MRI study.

And then when you acquire MRI data, here's the non-fat suppressed T two we are acquiring in the axial plane.

Most sites are acquiring in the axial plane in the US today.

It's important to look at it in the sagittal and coronal planes as I'm showing you here.

So we can look at each set of images in three planes.

And as you see here down below, this is a pre contrast axial T one showing a marker clip within the lesion.

Here we have got contrast injection showing heterogeneous enhancement.

And again, we can look at this enhancing mass in the axial plane, which you see here in the sagittal plane and the coronal plane.

And we can look at a thin MIP image showing the enhancing mass with some non mass like enhancement, extending both anteriorly and posteriorly from the mass.

And then we can look at the kinetic characteristics of the lesion, which in this case include heterogeneity of uptake with some wash out of contrast, some persistent and some plateau.

So that's basically the tools that we have that enable us to look at abnormal findings that we've seen on MRI.

MRI Lexicon

Now the MRI lexicon second addition has been coming out shortly for quite a while.

Should be out here pretty soon, I think.

And it's important to remember that this is a very good way of stratifying the lesions that we see on MRI, both normal findings and abnormal.

And when we have abnormal findings, bi rads assessment facilitate a an approach to reporting and management of the MR findings, which are useful.

And Can the descriptors that we are gonna go through now convey a predictable degree of malignant risk and mandate a specific management direction.

And the new lexicon attempts to bring terminology in line with the mammography and ultrasound lexicon, whatever possible.

As you'll see.

So the first thing we should report when we look at MRI is the fibro glandular parenchymal assessment.

This is a new term, a new classification, and it's asking you to look at the parenchymal volume on MRI that is how much breast tissue is actually there.

Fibro glandular tissue is visible on the image.

It's sort of similar to a density assessment of mammography.

But volume is the appropriate term for MRI and classified as in the past mammography into one through four quartiles of amount of tissue that you see.

So this is not enhancement, this is just the amount of tissue an anatomically that exists.

So here I'm showing you a mammogram, a lateral view, and then a sagittal view on mr.

And we are being asked to assess estimate in quartiles how much breast tissue there is on mr.

Here you can see just again, this is the T two non-fat suppressed data, looking at the sagittal AAL image and comparing it with a mammogram, sagittal and axial image, assessing the parenchymal volume.

And here you can see quite nicely with our three dimensional technique, being able to look at the sagittal and coronal and axial images.

We have a very nice way of looking and assessing how much fiber glandular tissue a patient may have.

Background Parenchymal Enhancement (BPE)

And then we're asked to assess the background parenchymal enhancement, also known as BPE, and describe on every report a qualitative assessment of the actual enhancement of that parenchymal volume that you've just estimated.

In other words, how much of that fiber glandular tissue that you assessed in terms of volume is actually enhancing following injection of contrast.

And again, these are the non minimal we can put together.

These are the approximate estimates that you can see non minimal, mild, moderate marked.

So you're gonna look at the volume of enhancement and then the amount, intensity of enhancement.

Now, background parenchymal enhancement, as you know, is actually non mass enhancement, right?

And normal breast tissue enhances post contrast.

So most of the non mass enhancement that you see at MR MRI is actually normal breast tissue.

And the trick is to differentiate normal parenchymal enhancement from actual pathological non mass enhancement.

That's what we're gonna go through.

We know that estrogen causes parenchymal hyperemia.

We expect greater background enhancement in pre-menopausal women.

That's been shown, and we've also know that there's a degree in intensity of enhancement varies during the menstrual cycle and is most intense in the immediate premenstrual period.

So all of these things, menstrual cycle pregnancy, lactation patients on hormonal replacement therapy may have profound effects upon the enhancement of the normal breast tissue.

Now, the breast parenchymal enhancement doesn't always occur symmetrically across the parenchyma in both breasts.

In fact, usually it does not.

Usually it enhances the periphery of the fibro glandular tissue.

As you see in these examples here where we have more intense enhancement at the cortical aspect of the parenchyma, the periphery of the parenchyma following the outline of the fibro glandular fat junction, which we're very familiar with from mammography.

So here you can see in the sagittal T one post contrast nons subtracted view and the similar technique in the axial plane.

We can see enhancement at the periphery of the fibro glandular fat junction.

Here's another example of the same thing, more enhancement at the periphery than the central mount of the parenchyma.

Here you see the same thing in a sagittal view, it's often symmetrical, but not always.

Here in this patient, we have enhancement all the way around the periphery of the parenchyma.

And again, you see it here.

So this is normal physiologic parenchymal enhancement.

You need to look at the non subtracted images.

If you just look at the subtracted images, you're not gonna be able to see the relationship of the enhancing parenchymal with the non enhancing parenchymal.

You'll just see the enhancement.

So in order to determine the exact distribution of the enhancement, you need to look at the non-contrast, the nons subtracted contrast enhanced images that I've just shown you.

This is the problem with nodule stipple pattern of enhancement with a lot of little enhancing foci is a problem that we encounter not infrequently at high, at high risk screening.

And this can be a real problem for us in interpreting mr.

We have such intense parenchymal enhancement and you can see on the angio map that we have even wash out within it.

But fortunately this is relatively uncommon.

We don't see this too much, but we do on occasion.

So parenchymal enhancement is our enemy and menstrual cycle scheduling, avoiding week four, can help diminish the amount of enhancement that you see.

The ideal time to do MRIs has been to actually is in week two, but you need to avoid week four at all possible.

So tips for differential diagnosis and diagnosing BPE.

Always look at the source images, not just the subtraction.

Always look in three planes.

You can make thin MIPS and thicken the slice of any area that you're worried about that can help you decide whether it's assuming, for example, a more linear form for that might suggest DCIS versus just diffuse parenchyma peripheral enhancement.

Look for differential kinetics.

And I will tell you that a higher field strength magnet does help, and we'll go into that later.

Non-Enhancing Findings

Then we have to look in the lexicon.

There's a new section on non enhancing findings.

On the pre contrast acquisition.

A lot of us forget to look at the pre contrast data, the T two and the T one pre contrast.

There's quite a lot of information actually there that can help you in differential diagnosis and interpretation.

We often see high contrast in the ducts on the pre contrast imaging.

And usually that's due to benign changes, fluid within the duct.

Simple cysts, complicated cysts, edema, non enhancing masses can be visible as in this patient with a large mass.

You can see the mass lesion on the T two pre contrast, duct distortion and clips we can sometimes see in this particular case is a finding that we often see on patients with large aggressive cancers.

Here's the mass, but you'll notice we have fluid coming back towards the petrol muscle representing distended lymphatics.

Again, a secondary sign of malignancy.

Here's some more examples of that.

This is due to lymph angiogenesis.

Here again, we see linear fluid extending posteriorly to the petrol muscle.

And this patient with a mass that you can see visible low signal on the T two image.

We can also see seromas, which you can see here in three planes.

Axial coronal and sagittal plane post-surgical seromas are visible.

Here's the fluid in the ducts that you can see on the P contrast T one.

We see this not infrequently, and as long as it subtracts out on the post contrast examination and there's no ductal enhancement, we don't worry about it.

This is a cyst, or this is actually is the same patient with a seroma.

Here we can see coronal imaging showing these ducts filled with fluid and debris converging towards the nipple.

And here's the sagittal plane showing the same thing.

Here's a mag view of the sagittal plane.

So we often see high signal and ducts on pre contrast acquisition, and most of the time, if there's no enhancement on the of the ducts on the T one post contrast, we don't worry about it.

We can also look at implants and lymph nodes.

We get a very nice visualization of the lymph nodes here.

We can see nice fat hilum and three planes.

And here you can see a mag view.

We can look at the cortex quite nicely.

So all this we can see on the T two.

Here we can see cysts.

Here we have cysts.

They're not fluid filled.

They're probably complicated cysts, maybe hemorrhagic cysts.

We see at least a couple here.

Here we see little bit of here we see the cysts with maybe a little enhancing of the wall, maybe not, maybe it's just the normal wall.

Here's post contrast, but it subtracts out on the subtracted image.

And this is just a cyst, maybe a hemorrhagic cyst, fat containing lesions, again, lymph nodes, fat necrosis, hamartoma, and postoperative seroma with fat.

We can see by looking at the non-fat suppressed images and identifying fat within the lymph node, as you see here.

And on the post contrast image high resolution, we can actually see the lymph nodes and the hilum and the vessels coming from it.

Lesion Interpretation

So moving on to actual lesion interpretation.

Now we've dealt with the pre contrast findings and the parenchymal findings that are all part of the new lexicon.

We look at pathology now and image interpretation.

We need to look at the morphologic features of each lesion we see, and we need to look at the dynamic features.

And if as far as the kinetic data is concerned, we need to look at the uptake of contrast signal intensity over time.

And we plot curves as to how the enhancement of the lesion is measured over time.

And the delayed phase is color coded as a wash out of contrast from the lesion being more commonly seen in malignant lesions and persistent enhancement be more commonly seen in benign lesions.

Plateau type enhancement can be seen in about 60% of malignant lesions, so could be either one.

Morphology will help us further analyze the lesions.

So we look at the morphology first, and then we look at the enhancement of the lesion.

And in the lexicon, we have three buckets into which we have to place an any enhancing finding that we see.

So if we see an enhancing finding on MR and we think it might be abnormal, it's not the parenchymal enhancement we talked about.

We have to decide is it a focus, is it a mass?

And I actually think those two usually can be put together, but right now, separate focus less than five millimeters a mass, in which case we look at the shape margins and internal enhancement characteristics and non mass like enhancement.

We look at the distribution and the internal enhancement characteristics.

So it's fairly straightforward focus, small, less than five millimeters, difficult to characterize mass as a mass shape and or is it non mass?

And I'm gonna go quickly through those.

But remember that the kinetic behavior of the lesions that you identify is a reflection of underlying tumor biology, and therefore can provide some information about the underlying pathology of the lesion.

Kinetics

Just quickly, when you look at your signal intensity time curves, you're looking at relative single signal increase over time.

And here's the curves, typical curves you get from cancer, fibroadenoma, hyperplasia, and parenchyma.

Cancers usually have rapid uptake and many of them will wash out.

Normal parenchymal enhancement is usually under 50%.

And we can set thresholds of enhancement and tell the computer only show me on the angio map lesions that are enhancing more than 50% or 80%, and then it'll just show you that.

So you can set that.

And the usual pattern of malignant enhancement is rapid up slope peak at one to two minutes post injection with 150 to 200% signal intensity, usually compared to the pre injection image.

Okay, report the worst looking curve within a lesion.

Many malignant lesions have very heterogeneous uptake within in fact, that's a hallmark of malignancy.

It's actual actually the degree of heterogeneity within the lesion.

So when you're evaluating and producing a kinetic curve or looking at the kinetic data report, the worst looking kinetic curve within the lesion and select your thresholds, you can select two sets of thresholds, many so 50 to capture the medium uptake and a hundred percent to capture rapid uptake as you see here.

Or you can just select an 80% threshold if you like.

And then the angioma will color code the delayed phase as you see on this sagittal image, the red representing in this case washout and a hundred percent uptake, more than a hundred percent uptake.

Also, you the commercial CAD systems can give you a breakdown of the amount of washout persistent or plateau kinetics within a lesion.

This is a measure of heterogeneity of the lesion.

And we have reported, and I hope we'll soon publish this data in our practice showing that the heterogeneity of uptake is actually as important as the single time intensity course curve for malignancy diagnosis.

Foci

Focus foci quickly.

Small lesions less than five millimeters in size cannot be otherwise characterized.

Single or multiple previously reported in the literature is incidental enhancing lesions, et cetera.

Actually at three Tesla, we can interrogate these foci better than we can at 1.5.

That's certainly been true in our practice.

This is a little cancer.

You can see here that it's got an irregular margin to it.

And you can almost suggest heterogeneity in this four millimeter lesion.

Most of the time they're multiple and a physiologic, and they represent enhancing tlu and fibrocystic change within the breast.

They may be associated with cystic change, as you see here, here we've got cysts and enhancing nodular parenchyma.

They're common in high risk screening exams.

And younger women, usually they're physiologic.

If you see them within the same quadrant as an index cancer, more than 50% of them are likely to be malignant.

We usually biopsy those when we stage cancer.

If they're in a different quadrant, they're more often benign.

So we do look at where they're located and what the clinical scenario is.

Solitary focus and a high risk women we're worried about it.

Maybe the kinetic's a little different.

Sometimes we'll bring the patient back in a very very short time, like one to two months after the MRI's been done.

We don't have any prior study.

We'll just repeat the dynamic sequence only.

And if the focus still is troubling, then we'll biopsy it at that time.

We'll set up for a biopsy.

Here's one patient with two foci.

Very different here we can see in the lateral breast, we can see on the T two a high signal, smoothly marginated, oval little mass with homogeneous persistent kinetics.

And in the anterior breast we see a low signal on T two, enhancing little mass, a little bit smaller lesion.

This is actually a three millimeter lesion, but it hasn't had very smooth.

It's not a smooth oval.

The kinetics are wash out.

This is an irregular focus.

We're gonna biopsy this.

This is a little fiber adenoma high signal on T two helps a very smooth margins.

And persistent kinetics are very helpful.

This had been stable on our screening examination for quite a while.

Here's an example of a fiber adenoma.

Hallmark of fiber adenoma.

Again, easier to see at three T or unen enhancing internal septations.

I'm gonna show you a magnified view.

Here's a sagittal reformatting clearly shows an un enhancing internal septation, also visible on Theron and as well, what type of benign lesion presents with washout kinetics?

I told you washout kinetics were bad sign.

Typically papillo intra lymph node of course 'cause direct feeding vessels and a young fiber adenomas.

They may present with washout kinetics, when to biopsy foci if it, if the margin focus, if it's a margin is irregular.

Re enhancement is present.

Kinetics are wash up plateau.

It looks different than the other foci that may be present.

If you've got multiple ones, any focus that shows similar findings that I've already listed.

And CAD can be helpful in selected cases if the enhancement of one particular focus is very different than all the others.

And you're going to be more likely to biopsy them if they're if you're staging a patient and she's got foci within the same quadrant as the index lesion.

Masses

Masses are straightforward.

I'm gonna go quickly through them in the last five minutes.

We look at shape margins and internal enhancement characteristics.

Here we have an example of round and oval masses and irregular and lobulated have been put together in the new lexicon as one finding.

Here very nice examples of speculation and irregular masses.

The term in the new lexicon is circumscribed and not circumscribed.

To bring it in line with the mammography and our ultrasound lexicon.

And in the not circumscribed masses, we the subdivisions are irregular, indistinct and speculated.

Look at the pre and the post.

It's a speculated mass.

We do see a little blooming.

So edge fudginess in the later post contrast series.

Here's a subtraction image and here's the heterogeneous uptake within this legion.

On the angio map.

Look at the T two findings whenever you have a malignant lesion.

Here we have edema on the T two.

We have some skin thickening as well.

And also look, we have additional little foci, which are very evident in the sagittal reformat.

So she's got a spiculated mass and multiple foci associated with it.

And here you can see the coronal plane.

Very helpful.

So example of homogeneous and heterogeneous enhancement, rim enhancement, which you can see here all mal high high grade malignancies I'm showing you.

Now.

Here's an example of rim enhancement.

When you look at the T two, there's fat within it.

This is an example of fat necrosis, T two being very helpful and enhancing internal septations.

Again, seen on the pre, even on the T two.

You can see it on the pre contrast data.

And then again on the post contrast images and homogeneous enhancement, typical for fibroadenoma.

Here's another one, fibroadenoma high signal on T two.

Here we can see unen enhancing septations on the T one post contrast.

Here you can see them again and again here.

Very classic for fibroadenoma.

Here is an example of a patient with a palpable mass mammography and ultrasound malignant lesion.

And this has got high signal on T two masses that have high signal on T two can be due to very high tumor grade.

This is a necrotic tumor actually.

A triple negative tumor can also see a mucinous carcinoma as well, where you have mucinous fluid within the lesion that gives high signal.

Here's an example of a different type of cancer that gives different kinetics.

Here we can, those of you who are mammographers can see distortion in the ML and CC view.

And here we have a mag view of the same.

And here we can see a distortion.

This is a lesion ultrasound co of a little mass, irregular mass.

And on mr, we can see in enhance them, but it's not very robust.

It has persistent kinetics, but the margins are very irregular.

There's an ultrasound correlative distortion, but even if they were not, this is pretty typical for invasive lobular carcinoma.

Doesn't have the same kind of robust kinetics as invasive duct carcinoma usually.

And here we can see the subtracted image showing its irregular mass changing its shape.

So it's really a non mass like enhancement with wish washy kinetics.

So the variable presentation of invasive cancer, this is an invasive lesion presenting as non mass enhancement happens sometimes either in the ductal or with a lobular histologic type.

Here you can see it very beautifully in the sagittal view.

This whole area of invasive cancer wrapping around the lateral side of the left breast heterogeneous kinetics, but heterogeneous kinetics, but mostly of the persistent type.

And this was another invasive lobular.

So for depiction of invasive cancer in general, for ductal carcinoma, usually presents as masses, usually irregular shape and margins.

Heterogeneous uptake, rapid uptake.

And kinetic shive curve is related to the underlying lesion physiology.

So grade three lesions are more likely to have a more uptake and wash out the lower grade lesions.

And the differential diagnosis is fiber adenoma, papillo and fat necrosis.

Now I'm going to this is just a classic appearance of an invasive cancer.

Here we can see it on the T two little bit of lymph node enlargement.

Here.

Here we can see the sagittal.

We notice a little bit of edema around it.

Here we can see the axial image, rim enhancement, sagittal rim enhance.

You'll notice that there's a little blood vessels around the lesion itself.

You can see there is a little bit of non mass like enhancement here, which could represent some associated in cyto lesion.

Here is the enhancement of the angio map showing very rapid enhancement.

And here is a case here is a just to run through the axial plane for you.

We can see Nice fat suppression uniform.

We're coming through.

This is a non subtracted data and here We can see our lesion very nicely.

RIM enhancement.

This is a first post contrast dataset.

Non-Mass Enhancement

What is non mass enhancement?

That's pathologic.

We know most of it is normal.

Why is this diagnosis often challenging?

It accounts for approximately 80% of MR guided biopsies.

There's no space occupying lesion.

There are no pre contrast findings, and there's no clear interface or margins with normal parenchyma.

And the MR depiction of non-invasive cancer is very different than the cancers that are invasive.

I've just been showing you most of them have been invasive presents often as a little area of non mass enhancement, a little focal area, segmental linear area.

So most of the time cancer in situ presents as non mass lesions segmental distribution in a clumped pattern.

The kinetics tend to be medial in media, initial uptake and plateau curve shape.

So these are often the kinds of cancers that we find at screening.

And so in my next talk when I talk about MR screening, I will show some examples of in of incy cancer and their typical appearance on MRI.

And it's very important that we be able to find these small lesions and differentiate them for normal breast tissue.

Thank you very much.