Pitfalls: Breast Imaging, Part 2

Review of Pitfalls in Breast Imaging

I still have no disclosures.

This is just a quick review of how perceptual, interpretive, and technical errors can all contribute to pitfalls in any area of radiology.

I'm focusing on breast imaging.

In the last talk, I focused on screening, the screening setting.

In this talk, this will be a case-based review of some of these common pitfalls in the diagnostic setting.

Along the way, I'll again refer to some useful references for evidence-based algorithmic practice, most of which are based on ACR appropriateness criteria.

Image Sequencing in Diagnostic Evaluation

The two main areas that I'd like to focus on today are pitfalls in the context of diagnostic evaluation, specifically image sequencing and incidental findings.

To start with image sequencing, I'm specifically referring to women who come in with a clinical finding.

We're in the diagnostic setting.

Generally a woman is presenting with some sort of focal breast sign or symptom.

Image sequencing comes up because we have two modalities at our disposal in the diagnostic clinic, mammography and ultrasound.

These questions come up. When is mammography enough?

When should a workup begin with a mammogram and be followed by an ultrasound?

Conversely, when should a workup begin with an ultrasound and be followed by a mammogram?

Finally, when is ultrasound enough?

The answers to these questions often depend on the patient, her clinical presentation, and whatever findings are identified on the modality that we begin with.

But luckily there are some algorithms that we can use to help us along the way.

Generally in the context of a focal breast sign or symptom, if the woman is under screening age or less than 40, it would be reasonable to start with an ultrasound first, in part because it has a very high negative predictive value and because there's no associated ionizing radiation.

You could then follow with a mammogram.

MX is the abbreviation that I'm using for diagnostic mammogram.

If whatever is revealed at ultrasound can be categorized as greater than or equal to a BI-RADS three or probably benign finding, or if the patient is a high risk patient.

Now, if the woman is of screening age or older, so 40 years or older, typical guidelines suggest that you begin the workup with a mammogram.

With a BB, a skin BB placed at the side of the patient's palpable area of concern, you would then go on to follow with an ultrasound if there's no specific benign mammographic finding that explains the patient's focal breast sign or symptom, or if the mammogram is negative.

Along the way, it's very important to be able to recognize specific benign findings, whether you're starting with ultrasound or mammography, because if you can do this, then in many ways you can expedite the workup and obviate the need for biopsy.

In general, it's very important to recommend clinical follow-up if the imaging workup turns out to be negative.

Specific Benign Findings

Before I go through a series of cases illustrating the use of this algorithm, I just wanted to review some specific benign findings that we should all be able to recognize on mammography and ultrasound.

Normal anatomy can explain focal breast signs or symptoms.

Here's an ultrasound image of it, just a ridge of dense tissue here, and that can often explain a woman's palpable abnormality.

Here's a prominent fat lobule.

That can be something that can present as a focal breast sign or symptom and ectatic duct without any other abnormality identified in or around the duct.

Here's a rib. Many women come in saying, I have this mass.

But it turns out it's a rib.

Here's a classic oval circumscribed hypoechoic avascular simple cyst, a benign finding.

Here's a round circumscribed mammographic mass that has these popcorn calcifications that are classic for a benign calcified fibroadenoma.

Here's a beautiful little retroareolar lymph node that has this thin hypoechoic cortex and echogenic fatty hilum.

Then here's a fat containing mass.

Fat containing masses are easily identified on mammography because of their radiolucency.

Whether it's an oil cyst, a lipoma, hematoma, or even fat necrosis, this should be a finding that you can take to the bank and stop with mammography in terms of the patient's workup.

If you don't see one of these specific benign findings, it would be important to consider employing additional imaging, whether it's mammography or ultrasound.

Case-Based Review

Case 1: 38-Year-Old Woman with Palpable Mass

Let's start with this case of a 38-year-old woman.

She presented with a palpable mass in the right breast for about five weeks.

She's under screening age.

The first test of choice was an ultrasound.

This ultrasound reveals this hypoechoic irregular mass with some peripheral vascularity and some shadowing, and very irregular margins.

What would your BI-RADS assessment be?

Actually, a hundred percent are correct in that it could be considered suspicious or highly suggestive of malignancy.

Great job. In this case, we started with an ultrasound, but based on the ultrasound features, and given the fact that we're recommending biopsy, it would be appropriate to then recommend a diagnostic mammogram just to confirm there aren't additional abnormalities or see any other associated mammographic findings that may be sonographically occult.

Then recommend ultrasound guided core biopsy.

Here is the woman's mammogram.

There's a BB placed over her palpable abnormality, which also corresponds to the ultrasound finding.

We see on these two views a focal asymmetry, very irregular, hard to characterize it specifically as a mass, but certainly an asymmetry.

The rest of her breasts look negative.

She has scattered glandular density.

We felt like this was all that was in store for her because she had this suspicious finding.

She was then recommended for biopsy that revealed invasive ductal carcinoma.

In this premenopausal woman with invasive breast cancer, she was referred for genetic counseling and for an MRI for extensive disease evaluation.

This is her MRI.

This is a single image from the sagittal reconstruction, showing that there's the MRI correlate to her ultrasound and mammographically apparent mass.

But then there are these additional, and there turned out to be actually about six discrete masses in this woman's right breast all biopsy proven, ultimately invasive ductal carcinoma.

This is how MRI was incorporated into her workup, but it was very important that this step was taken at that point.

Case 2: 65-Year-Old Woman with New Palpable Abnormality Near Lumpectomy Scar

Next case, 65-year-old woman, new palpable abnormality near the lumpectomy scar.

Here we have a patient with heterogeneously dense or extremely dense breasts, and we see her scar marker over her lumpectomy scar.

We see a skin BB placed at the site of her palpable area of concern.

Immediately adjacent to this skin BB is a very coarse calcification, kind of looks like a bird.

Here's a magnified image.

It really looks like it's close to her skin.

To confirm that it was actually in her skin, I went in and talked to the patient and examined her and saw that in fact, she had some redness there.

She said that over the last six months she had lost some weight, and so started feeling this even more.

We happened to have a CT scan that she had had for another reason, and saw that this is in fact very coarse calcification within her skin.

No ultrasound would be indicated in this particular case because we have a classically benign finding of a coarse dystrophic calcification within the skin that explains her palpable abnormality.

Case 3: 48-Year-Old Woman with Itchy Nipples

Next case, 48-year-old woman who came into our clinic with itchy nipples.

Because she's over screening age, we started with a mammogram.

We weren't expecting to find any mammographic finding specifically to correlate with her presenting symptom.

She has scattered fibroglandular densities, and this abnormality was identified in the left breast.

We had no priors for comparison. What is the abnormality?

Is it a focal asymmetry, a mass, calcifications, architectural distortion, or there's no abnormality?

The vast majority of people identify this mass, which is exactly what it was.

Now we can go to our BI-RADS lexicon and characterize this mass that we see.

We use shape descriptors, margin descriptors, density descriptors, and the presence or absence of associated findings to characterize this mass and our level of suspicion.

We can see this oval circumscribed mass with a lot of low density and even radiolucent areas with no associated calcifications.

This mass does contain fat, which is very reassuring.

We were ready to stop there.

She happened to have an ultrasound that was done at an outside facility, and she submitted those images for review.

The ultrasound also shows similar features.

Again, we have an oval circumscribed mass with an abrupt interface with the surrounding breast parenchyma, a lot of hyperechoic components and no calcifications.

What would your most likely diagnosis be?

A hundred percent great work.

This in fact was a hamartoma.

No ultrasound was indicated at our visit.

The ultrasound that was performed at that outside facility was not needed.

This is an example of a classically benign fat containing mammographic mass, which happened to be an incidental finding in this patient with itchy nipples.

It's very important to recommend clinical follow-up for that kind of presenting symptom, especially because that can be the presentation for Paget's disease.

Always important to include that clinical information in your impression.

Case 4: 44-Year-Old Woman with Superficial Palpable Abnormality

Next case, 44-year-old woman with a superficial palpable abnormality in the upper outer right breast.

She's over screening age.

We started with a mammogram.

We have a BB placed at the site of her palpable area of concern.

She has scattered fibroglandular densities.

There's no real mass underlying that BB, no real distortion, but we did see some calcifications.

What would your next step be? Negative routine screening in a year, six month follow up, additional mammographic views and ultrasound, recommend biopsy or an MRI?

We're in the diagnostic clinic, we have the benefit of real time workup.

Additional mammographic views and ultrasound is a correct answer.

On spot magnification views where we're trying to better characterize these calcifications, we confirm they're seen on two views.

There's nothing really classically benign about them.

They couldn't be called benign.

The patient moved on to ultrasound in the clinic, and the sonographer was unable to show any discrete mass that corresponded to the patient's palpable abnormality.

Now what's your next step? Same options.

Just because her presenting symptom was this palpable abnormality and we had no basically imaging correlate for her palpable abnormality, we did see calcifications and the calcifications on their own were considered suspicious.

Biopsy of the calcifications was recommended.

These are the stereotactic core biopsy pre-procedural images, and you see the skin BB very close to this cluster of calcifications.

At the very beginning of the biopsy, the radiologist thought, these are pretty close to the skin.

She was taken out of the stereo unit and additional views were obtained, including this tangential view with magnification technique confirming that this target that was recommended for biopsy was actually in the skin.

The biopsy was terminated.

The reason I bring this case up is that incomplete workups can lead to unnecessary headaches.

Imagine yourself in the clinic and you see the patient who's presenting with a palpable, the ultrasound is negative, you find this incidental cluster, but those views that were obtained may not have been sufficient at the time of the diagnostic.

Tangential views can always be obtained if you're about to recommend a stereo biopsy in the diagnostic clinic to confirm that in fact, they warrant biopsy.

Always confirm that calcifications are amenable to stereo before recommending biopsy.

They should not be in the skin.

They should be seen on two orthogonal views and accessible with the patient in the prone position.

Case 5: 67-Year-Old Man with Right Breast Tenderness

Next case, 67-year-old man with a three month history of right breast tenderness.

Here we have a BB placed at the site of the patient's tenderness and CC and MLO views.

What is the finding on mammogram? Is it gynecomastia? Malignancy, negative, lipoma?

Very good. That's correct. Three quarters of you correctly identified gynecomastia.

What is your next step? You see classic gynecomastia.

Do you move on to ultrasound, short term follow-up, clinical follow-up, or refer for biopsy?

In this particular case, the next step should be clinical follow-up only.

No ultrasound is indicated in this particular case.

There was a classic flame shaped retroareolar tissue that can have a variety of patterns.

It can also be asymmetric, but if the appearance of gynecomastia is classic on mammography, there's no need for ultrasound.

Ultrasound can be very confusing and lead to biopsy recommendations.

Classic gynecomastia should be managed with reassurance and clinical follow-up as necessary.

Case 6: 49-Year-Old Man with Left Nipple Pain

Next case, 49-year-old man with left nipple pain for two months.

This is a very different presentation.

We have the BB on the left side, CC and MLO views.

What is the finding on this mammogram?

Is it gynecomastia, a retroareolar abnormality, no findings or a lipoma?

Retroareolar abnormality is the correct answer.

Given the fact that there was this retroareolar abnormality, we did not see gynecomastia, we did move on to an ultrasound.

The ultrasound shows this almost two centimeter irregular hypoechoic mass hypervascular on Doppler interrogation, and it's centered in the skin.

The differential diagnosis of this finding would include all of the following, except your answer is correct.

The differential would include all of those possible diagnoses except gynecomastia.

Exactly.

What is your next step?

Now that you have this clinical symptom, mammographic abnormality that is not gynecomastia, a focal finding that is suspicious on ultrasound, your next step should be, this is a spread, and I'm glad we brought this up.

40% of you thought that the mammo service should biopsy this, about 40% thought it should be referred to surgery or derm. 20% had a different response. Clinical follow-up only.

This ended up being recommended for biopsy by surgery, but could be referred to dermatology or surgery.

The pathology of this finding was actually invasive ductal carcinoma.

At our institution, anything that's in the skin is referred for clinical evaluation by either dermatologist or surgeon.

Those are the folks who end up doing biopsy.

We reserve biopsy of breast lesions in the mammo section.

That's in part because there's an increased risk of infection in findings that are in the skin.

Ideally, you wouldn't need to biopsy anything in the skin if it has classically benign features.

This clearly did not have classically benign features.

Biopsy was warranted, but it was done by surgery and it was invasive ductal.

In this particular case, both mammography and ultrasound were indicated.

The initial mammographic workup failed to reveal gynecomastia to explain the palpable. Ultrasound was an appropriate supplemental exam.

In general, anything hypervascular should raise level of suspicion and should be considered for biopsy.

But even sebaceous cyst should be considered for referral to derm or surgery for appropriate management that could include surgical excision.

Incidental Findings

In the remainder of the talk, I'm gonna focus on incidental findings.

Incidental findings can present themselves either as breast findings detected on chest or body CT or MRI, or they can be extramammary findings, usually seen on a breast MRI.

For extramammary findings on breast MRI, you can see them on any sequence really, including the localizer, the T1, the T2, post contrast sequences.

I just wanted to give a list of the regional structures that may be seen on a breast MRI.

You start coming up with a checklist for the structures that you should be checking.

Besides the breasts on an MRI, so thyroid and cervical lymph nodes in the neck, in the chest, everything from the lungs to the great vessels, all the bones that are visualized, including the spine, ribs and clavicles, and even abdominal structures, including the gallbladder, liver, and some of these kidney and adrenal findings you see on the localizer.

You're responsible for everything that is in the field of view.

It's just important to have a checklist.

I just wanted to first show you a couple of examples of extramammary findings on breast MRI and then we'll finish this section with incidental breast lesions.

This is a sternal metastasis that was detected on a breast MRI as an enhancing kind of infiltrative abnormality.

It was confirmed on CT with this destructive appearance on CT and bone scan with increased uptake.

On the MRI, it's actually best seen on the coronal views.

Low on T1 and enhancing, sternal metastasis.

Anyone see the finding here outside of the breast?

This lung mass almost looks like a PE but this was an MRI that was ordered for extensive disease evaluation in a woman who had a right breast cancer.

I don't show you her cancer, but it's on this side.

This was not expected, not previously known and actually not reported on the outside study.

This was sent in for review.

The patient was being seen for second opinion at Stanford at our tumor board, and we didn't have access to the images until later.

We were reviewing the report to give our recommendations.

But then when the images came in, we saw this mass and we made sure to let everyone know.

These are both post contrast and T2 weighted images showing what turned out to be a metastatic lesion in the rib.

I could go on and on.

But the point is that there are extramammary structures that are covered on breast MRI, and so it's important to have a standard search pattern so you don't miss those.

In general, if an incidental finding is detected in the breast, and I'll show you some cases at the end of this section of this type of situation, if a breast lesion is found on, then the finding should be correlated with mammography, ultrasound or MRI and characterized as benign, indeterminate, or suspicious enough to warrant further workup.

Detecting a finding on CT that you know is obtained for some other reason isn't enough.

Once you detect the finding, you need to make a recommendation for correlation and make sure that recommendation is followed closely.

The features most predictive of malignancy include irregular margins, irregular shape, and rim enhancement.

Similarly, if incidental extramammary abnormalities are identified, they should be correlated with any other available cross-sectional imaging.

So CT, MRI or ultrasound and again, characterized as benign, indeterminate or suspicious enough to warrant further workup.

Incidental Breast Lesions

Here's an example of a 47-year-old woman with a history of E coli UTI who was decompensating.

She was an inpatient, she was on antibiotics, she had ARDS, so she's a very sick patient and she had this CT to exclude abscess.

We see her consolidation in her lungs and her pleural effusions, but anyone see anything else?

Excellent. This is a rim enhancing mass in the right breast, not the cause of her problems as an inpatient, but certainly an issue that needs to be pursued.

The recommendation was made to correlate this with mammography.

She didn't have a recent mammogram because of all of her other health issues.

Here we have her MLO views and we see a correlate.

We see at least a one view finding on the right MLO view to correspond to that mass on CT.

Here again in the outer breast we have the mammographic correlate.

She went on to ultrasound and we see this heterogeneous mass with some irregular margins and peripheral vascularity recommended for biopsy and confirmed to be invasive ductal.

Incidental finding with an important diagnosis made in this very sick patient.

Next case, 75-year-old woman with HCC post TACE.

CT to evaluate for recurrence.

Anyone see anything? We're in the lungs. Soft tissue windows. Exactly.

This is a post contrast image.

We see that it's an enhancing mass.

Also given her other medical conditions, she hadn't had a recent mammogram, but we see these CC views.

She has scattered fibroglandular densities.

We know where we're looking because we saw a finding in the medial breast.

There is the correlate at this point.

It looks just a little asymmetric on the MLO view.

It's much more evident in the upper breast.

We get our additional mammographic views that confirm the presence of fine pleomorphic calcifications and spiculation within this mass.

We can appropriately direct an ultrasound.

Then subsequent biopsy also confirmed invasive ductal and DCIS.

Again, incidental to her hepatic pathology, but very important.

Next case is a 50-year-old woman.

She was referred for an MR venogram because she had right arm swelling, but she has this known right upper extremity congenital vascular malformation.

Really a prominent abnormality.

She had an MR venogram and there was this speculated mass in the right outer breast.

Everybody see it.

This appropriately was referred for correlation with mammography.

She didn't have any recent prior mammograms for us to use for comparison.

We did our standard workup with CC and MLO views.

We see this finding in the upper right breast.

We don't really see a correlate on the CC view, but we also happen to see this set of abnormalities in the far upper part of the right breast.

Some of the irregularity extends into the skin.

Luckily we had the information regarding her congenital vascular malformation.

We thought, is there any way that this could be contributing to this abnormality in the breast?

We weren't sure. We had no priors.

Because it was a one view finding, we only saw it on the MLO view.

We wanted to make sure we could confirm that it explained the MRI abnormality.

We got additional mammographic views, an exaggerated CC lateral, a tangential, and a full lateral.

If you would click the circle, then you see that this tracks appropriately.

That in fact corresponds to the finding on MRI.

We have an idea of what we think it could be, but we need to go on to ultrasound in this particular case.

We see this hyperechoic irregular structure that corresponds in location and size and shape to the mammographic finding in the MRI finding.

In this particular patient, no vascularity, at least demonstrable on our Doppler.

What would your BI-RADS assessment be? BI-RADS one negative, BI-RADS two benign, probably benign, suspicious or highly suggestive of malignancy.

Okay, that's an interesting spread.

I can understand why many people were concerned.

There were some irregular features.

However, we then had three pieces of imaging information to suggest that this could be part of the same process that she had already presented with.

We considered this a benign finding, an extension of her upper extremity congenital vascular malformation into the right breast.

This would be something that one should be able to consider benign and move on, and a biopsy would not necessarily be a great idea in this particular patient.

Next case, 75-year-old woman with a renal cell carcinoma.

Staging CT, anyone see the finding? Excellent.

Already you're thinking of hypervascular abnormalities because she has renal cell carcinoma.

This is alarming even though breast isn't necessarily a typical site for a metastasis in renal cell.

Recommendation was made to correlate with mammography.

We have three views of the left breast CC, kind of an exaggerated CC view.

Very hard to see any discrete abnormality, MLO view.

But then finally on the lateral we see something that could correlate and it looks like at least a one view asymmetry.

We know there's a mass there based on CT, and so we can direct an ultrasound.

Here again, this is a hyperechoic mass has some irregularity, but because of its echogenicity is very reassuring.

We were prepared to leave it alone.

Hemangioma can be seen in the breast.

But because the patient had coexistent renal cell carcinoma, the treating team requested a biopsy regardless.

A biopsy was performed with vacuum assistance.

There's our post biopsy metal marker confirming that yes, in fact, that was the mammographic correlate to the finding on CT.

The pathology was a bland vascular lesion without atypia.

Although quite alarming on CT, the reassuring features on ultrasound specifically could have obviated a biopsy in this case.

We did the biopsy because of the clinical concern of the team.

Summary

As a summary for the diagnostic pitfalls that I've reviewed today, appropriate image sequencing can facilitate accurate diagnosis and limit unnecessary utilization.

That's specifically referring to the use of mammography and ultrasound in the diagnostic setting and incidental findings in the breast and extramammary regional structures should be correlated with other imaging modalities and fully characterized, because they may warrant further workup.

It's not enough to see the finding, although it's very important to see the finding, but it's very important to also make specific recommendations for how to deal with the finding.

That's the end of my talk.