8 MRI Technical Considerations

Introduction to the Speaker

The next speaker is Daniel Margolis,
who is the co-director of Prostate MRI,
and assistant professor of Radiology at UCLA.
And Daniel is going to talk to you about MRI technical considerations.

Technical Considerations for Prostate MRI

I'm gonna talk today about technical considerations.
The first consideration is why would we use MRI.
As we all know,
CT does not delineate prostate anatomy,
and whereas ultrasound can identify features within the prostate,
it can't distinguish between benign and malignant tissue.
And as we very elegantly heard today,
prostate MRI allows us for both anatomic and functional characterization.
But it's imperative that we have adequate technical quality to do so.
There's a lot that we can do beyond anatomic T2WI imaging.
T2WI imaging delineates normal and abnormal tissue,
but just like ultrasound is nonspecific.
Old inflammation, BPH and hemorrhage can appear similar to cancer.
Spectroscopy is useful for identifying abnormal chemicals.
Diffusion weighted imaging can detect cellular packing,
and dynamic contrast imaging maps blood flow to the prostate.
These are the different components that we want to look at:
T2 weighted imaging, spectroscopic imaging,
diffusion weighted imaging,
and dynamic contrast enhanced imaging.
And all those spectroscopy is not a component of the upcoming PI-RADS classification.
Some centers still use it.
And so I'll briefly touch on the technical considerations.

Quality Assurance

In terms of quality assurance,
nearly all scanners are sufficient for whole pelvis imaging.
Looking at the bones and lymph nodes,
but because of the small size of the prostate,
a small field of view image is required for adequate characterization.
And you want to have adequate SNR,
especially for diffusion weighted imaging.
So we recommend at least a 1.5 Tesla magnet.
And to ensure adequate resolution of the prostate capsule or pseudo capsule,
you would consider using an endorectal coil.
Qualitative assessment requires calibration.
So if you're going to look at quantitative parameters,
you wanna make certain that all of the magnets in your system are using the same protocols
and are calibrated.

Diffusion Weighted Imaging

In terms of diffusion weighted imaging,
I'm just going to cover the technical considerations.
The actual review of images will be covered in a separate talk.
But we need at least two b values.
The low b value can be zero,
but we recommend at least 100 square seconds per millimeter.
Because this obviates the perfusion component.
The high b value should be at least 800.
However,
we want to review images with a b value of ideally over a thousand.
1400 is suggested.
And this can be natively acquired or calculated.
The advantage to a calculated high b value is it exploits the improved SNR for lower b values
while still providing a diagnostic image.

Dynamic Contrast Enhancement

In terms of considerations for dynamic contrast enhancement,
the pharmacokinetic map can be helpful for region of interest choice and for follow up.
It may not add much to the PI-RADS overall assessment,
but it can be very useful for identifying suspicious areas.
However,
you can use a simple subtraction map for detection,
and the calculated parameters are multifactorial.
Temporal resolution is ideally less than 15 seconds,
and over at least four minutes if you are going to calculate pharmacokinetic parameters.
And you may consider using a view sharing technique such as Siemens TWIST or GE DISCO
to achieve both high temporal and spatial resolution.
And if you are going to perform spectroscopy,
some contrast agents will actually alter the measured spectra,
and therefore,
you must do spectroscopy before dynamic contrast.

Spectroscopic Imaging

In terms of spectroscopic imaging,
I'll cover the basics of the technique.
We generally recommend use of endorectal coil because of the SNR requirements.
And the spectra vary based on field strength.
TR and TE are usually fairly T2 weighted,
but can be chosen to isolate specific J couplings.
Choline, creatine and citrate peaks are measured,
and we look at the choline plus creatine to citrate ratio per voxel to evaluate the suspicion.
Values are generally expressed in parts per million,
which obviates the field strength dependence.
And this is the general formula.
Here's some examples of normal and abnormal spectra.
The choline, creatine and citrate peaks are at 3.2, 3.0, and 2.6 parts per million respectively
with water normalized at 4.7.
And because choline is used in cell replication, it's elevated.
Citrate, a normal component of the glandular tissue is utilized in the tricarboxylic acid cycle.
And when prescribing the MRS technique,
you want to prescribe a rectangular voxel around the prostate itself,
and then use very selective saturation bands to exclude the fat signal.
Here would be an example of a narrow box around the prostate,
and here's how you would wanna place the saturation bands.

Value of Spectroscopy

Does it add value?
Well,
Jeff very elegantly prescribed an evaluation.
It was a multicenter trial comparing MRI versus MRI plus spectroscopy for sextant localization.
There was no significant difference in AUC.
However,
when looking at the various components of the analysis,
quality was non-uniform,
and there was some difference in experience in the readers.
And probably most importantly,
some cancers were below the detection level for spectroscopic imaging.
And so many single center trials have shown value of spectroscopy.

Need for Endorectal Coil

Do we really need an endorectal coil?
When comparing an endorectal coil at 1.5 Tesla versus a 3.0 Tesla magnet without,
the first publication said no,
you no longer need the endorectal coil at three point zero Tesla.
However,
a year later, the group from with Bedo published in AJR that actually it does matter.
A few years after that,
we had a group publishing that no,
you don't need the coil,
but yellow barrens a year after that said yes,
you do.
Why is there this discrepancy?
It depends on what they were looking at.
If you're just looking at a gross estimate of burden and grade of disease,
in fact, you don't need the endorectal coil.
The endorectal coil is most useful for the detection and characterization of extracapsular extension or seminal vesicle invasion.
In other words,
it's most advantageous for staging but not for grading.
And this is supported by a number of different publications.
One consideration is DCE is robust to field strength.
Here's a rather egregious example.
You can see that on this image at 1.5 Tesla with an endorectal coil,
there's irregularity of this lesion so that it looks like there's capsular invasion
and there's obvious asymmetry on the ADC map,
whereas at three Tesla without the coil,
you can see that there's a low signal focus here,
but the borders are obscured.
That would change the overall PI-RADS assessment.
Additionally,
although there is a low ADC here,
you see that this image is rather noisy,
and it's hard to confirm that this is actually the lowest signal on the image.
Why this is probably has most to do with the gaseous distention of the rectum.
It's a component of both field strength and signal and image noise.
And here's another example of hemorrhage.
This is a T1 weighted image.
On the T1 weighted image, you can see the high signal from hemorrhage in the left prostate.
And you can see that on this image, there's no endorectal coil.
On the T2WI image with an endorectal coil, you see that our features are much more distinct,
and that there's low signal here.
Although this may be a PI-RADS three in the peripheral gland in terms of T2 assessment,
you would have to take into account the fact that there's hemorrhage.
And there is a concept of the hemorrhage exclusion sign.
Because of the enrichment of citrate in normal tissue, this contributes to an anticoagulant effect.
And the feeling is that if you see hemorrhage,
that decreases the likelihood that this represents a tumor.
Also, you can use your pre contrast T1 weighted image for the detection of hemorrhage.
T1 weighted images can be used both for the detection of hemorrhage and for dynamic contrast.
That may come up a little later in the questions.
And again,
you get better contrast of background with the coil.

Endorectal Coil Technique

In terms of coil technique, patients with their own sedation may do better.
We want to place the coil with the patient in the fetal position.
You can consider using a perfluorocarbon to distend the balloon rather than gas to decrease susceptibility artifact.
There's a bumper that can be used for optimal positioning,
and you want to insufflate between 45 and 90 milliliters.
The coil will reposition itself as the balloon is insufflated.
And then you want to readjust the position of the coil when the patient is supine.
The coil should be placed facing anteriorly.
You want the coil elements at two and 10 o'clock on the axial images,
and you want to cover from the prostatic apex to the seminal vesicles.
You can see this coil is riding a little high.
We applied the bumper to help adjust the position of the coil.
And there are relative contraindications: pelvic surgery,
radiation therapy, severe hemorrhoids or anal fissure.
But the only absolute contraindication would be absence of an anus either surgical or developmental.
Although if you can defer imaging for resolution of something that's time limited like hemorrhoids or fissure,
you would want to do so.

Coil Placement Considerations

In terms of coil placement considerations, lidocaine gel may be considered,
but it can also be irritating.
Most of the discomfort that occurs during coil placement is from the coil distension.
And that's not obviated by the lidocaine gel.
We generally only use it for patients with hemorrhoids or who have severe discomfort.
We want to make certain that you aspirate all of the gas from the balloon before placing it.
You can use the patient's gown to help optimize the position of the gluteal tissue.
And sometimes placement with the coil at 90 degrees can be very useful.
You want to reinforce the patient to breathe slowly and to maintain the fetal position.
This seems self-evident, but during placement, the patients often forget.
And some patients will have a very vigorous sphincter reflex,
but this being a skeletal muscle can be fatigued.
And so if you try repeated gentle attempts,
it usually obviates any difficulty.
And then you want to make certain that you or your technologist in my case
use a small towel during removal 'cause you don't know what's coming out.

Additional Technical Considerations

A few other technical considerations.
Again,
you want to use small field of view T2 weighted imaging.
We're not covering T2WI imaging as a separate component here.
Generally the stock pulse sequences are adequate,
but you want to make certain that you're getting adequate resolution or adequate SNR
by optimizing your pulse sequence parameters like your matrix and number of excitations
to achieve the level of resolution sufficient for the PI-RADS descriptor category categorization.
And you can consider using a 3D technique,
which can be very useful if you're going to register your T2 dataset with ultrasound.
You want to consider using an antiperistaltic agent,
and the IV glucagon is the only one that we commonly use.
And this will reduce motion artifact.
You want to make certain that all patients evacuate before imaging.
This is imperative for coil placement.
But as we saw from the example a little earlier,
it's also very important that there's no gas in the rectum,
because this will cause significant susceptibility artifact
and can also contribute to a peristaltic motion artifact.
And there are some experts that will actually even without the use of an endorectal coil
place a small catheter to aspirate the gas in the rectum.
I have the techs do that work for me, so I don't ask them to do that,
'cause I ask enough of them already.
But one of the counter considerations is that sometimes there's just a lot of gas
and when do you stop?
And then there are some experts that recommend that the patients refrain from sexual activity for three days ahead of time
because you want to make certain that the seminal vesicles are well distended for optimal detection of seminal vesicle invasion.
We actually have our schedulers recommend our patients refrain from any activity.
Generally it's well adhered to, although it does cause a significant amount of stress in a minority.

Software and Reporting

And then one other consideration is do you need specific software?
There is no specific recommendation,
and the use of CAD has not shown superiority to general review.
However,
there are some systems that help standardize reporting
and using a standardized reporting system such as PI-RADS has been shown to improve the performance,
especially in naive readers.
CAD can also improve detection potentially,
and improve throughput by providing a multiparametric review.
I think the throughput is maybe the most important consideration
because it really helps me speed through the cases when I'm using CAD.
And then obviously if you're going to use image co-registration,
you need to use the CAD that corresponds to the fusion system.

Patient's Point of View

And then one last important point is what is the patient's point of view?
It's important to reinforce with your schedulers and your front desk staff
what the patient is going to go through so that they can coach the patient to reduce patient dissatisfaction.
You want to do at least three weeks and ideally six to eight weeks after biopsy
to reduce hemorrhage artifact.
And for patients that are going to have a coil,
you want to consider using an enema the night before and the morning of,
but all patients should evacuate completely.
And in fact,
even for patients without a coil, we suggest that they consider either an enema or a suppository the night before.
A mild sedative also helps with coil tolerance.
One consideration though is that an enema relatively soon before imaging can increase peristaltic motion.
And again,
this brings us to an antiperistaltic agent.
It's important that patients know that in addition to an IV for IV contrast,
they will receive an injection of an antiperistaltic agent,
and that the scan can take anywhere from a little less than half an hour
if you're doing just a focused evaluation to one and a half hours
if you're doing abdomen and pelvis and spectroscopy and coil placement.
And then the post-processing can take up to an hour.
This is also important because some patients will wanna know right away when they get off the scanner
and it's important that their expectations are appropriately managed.
And you also wanna make certain that the tech QC the images before sending them for review.
In some cases,
some of the series might initially fail and need to be reprocessed.

Acknowledgment

And I'd like to acknowledge my collaborators.
In order to establish a healthy and high quality service,
it takes input from all of the radiologists that are going to be reading the study,
from the technologists that are going to be doing the actual scans,
from these schedulers.
We've had a lot of input from our physicists,
and our allied support to ensure adequate quality.
It also takes buy-in from your urologists and pathologists
that they understand what you can and more importantly,
what you can't provide.
And also that you have some way to follow up on the results in pathology
because this will help gauge how high quality your studies are.
And obviously, we've had some support from our external collaborators.

Illustrative Case

And if I have time,
I'd like to show one case that sort of illustrates the importance of image quality.
Let's go to hang.
I wanted to show this because it has an example of in this case the perfusion images were very useful
because here you can see that towards the apex there's a low signal area with restricted diffusion on the ADC map.
And on the perfusion images,
on the pre contrast T1, it shows that there's hemorrhage on the contralateral side.
This would suggest that you can use the hemorrhage exclusion sign that this low signal here on the right is less suspicious.
However,
the presence of hemorrhage also can alter your perceived diffusion and measured perfusion values.
And in fact,
this patient had four plus three disease.
It was much worse on the left, but there was some disease on the right.
The hemorrhage exclusion sign is unfortunately not absolute.
And you can still get cancer where there's hemorrhage.
When you see hemorrhage,
it's important that you recognize that this will be a technical limitation of your study.

Thank you very much.