PI-RADS

Introduction to PI-RADS

The purpose of this talk is to just give you the background and evolution of Pi Rads.

And then in the rest of the talks, we'll start getting into really how to use it.

And this is gonna be an introduction to Pi Rads.

So, PI Rads is a collaboration between the American College of Radiology, the AME Tech Foundation's International Prostate, MRI Working Group, and the European Society of Radiology.

This is PY Rads version two. That's what V two stands for.

And RAD stands for Prostate Imaging Reporting and Data System.

So let's talk about where PI Rads is coming from and what we're trying to accomplish.

Then we'll start getting into how it works.

Background and Evolution

The AME tech foundation organized something called the International Prostate, MRI Working Group.

And in 2010, there was a meeting in Stockholm and there was discussion about how can we move prostate MRI forward.

And one thing that became readily apparent is that the field had reached a certain level of maturity, but that everybody was doing their own thing.

And in order for it to move forward, there had to be some kind of standardization.

So I suggested the idea of PY rads, which is along the lines of Birads that has been very successful in standardizing how breast, mammography and breast MR is done and reported.

Yellow Barons, who is a member of this team, is from Akon in the Netherlands.

And he went back after this meeting, and along with his colleagues at the European Society of Euro Radiology organized a working group for prostate mr.

And together in 2011, they created the first version of PY rads.

And this was published in 2012, although it was the elements of it were known in 2011.

Also following that initial meeting in 2010, I started lobbying the American College of Radiology to start putting together a Py Rads effort.

It took until the end of 2011 to get this together, and we decided to do this as a collaboration between the A CR, the European Society of Euro Radiology, and the AME Tech Foundation, because it didn't really seem to be very wouldn't have been great if the different places in the world doing prostate. Mr started doing their own thing.

It would just add to the confusion.

So we put together a committee of international radiologists, and in this case, international means us, Canada, and Europe, which you see here is the steering committee.

People at the bottom were I'd say in a advisory capacity on the steering committee.

And then we also formulated we also formed a working groups on a number of different areas to get the work done.

This was financed by the American College of Radiology and staffed by the American College of Radiology.

And we also have a number of informal advisors from other areas, including urologists and radiation therapists, pathologists, and medical oncologists.

There's also a large number of radiologists from the United States and Europe who served on these various working groups and have also helped been very helpful in reviewing the document as it's been developed.

So there's been a lot of people involved in this, and this is a list of the aims of PI rads.

But without going into this in detail, really the aims come down to trying to standardize the way that MRI is performed, how it's displayed, and how it's interpreted and how it's reported.

And you can imagine that not only doing this across institutions in the United States, but doing this across all the strong personalities and all of the countries involved in this.

It was a formidable undertaking, but the ultimate goal of this, of course, was to improve patient outcomes.

PI-RADS Version 2

So PY Rads version two refers to this A-C-R-E-S-U-R and AME tech collaboration as opposed to PY Rads version one, which was done, which was created by the ESUR.

And what we realized when the first version of PY RADS came out was that it was very useful in that it did seek to standardize things, and there were a lot of very good useful information in that document, but that there was also some limitations to it, which is understandable 'cause it was the first version.

The instructions were not very clear.

The way they the scoring system worked was just very difficult for some people to use.

And although it has been validated in some studies, in other studies, it hasn't been validated.

So that was the reason for going on with Pi Ed's version two.

Now, we started meeting in early 2012 via conference calls and emails.

And what became apparent very quickly with this international group is that everybody on the committee is an expert.

That's why they were on the committee.

Everyone on the group knew what they were doing.

Everyone was doing it correctly, but everyone was doing it differently.

And that's the truth.

And of course, this, if everybody continues to do what they're doing their own way, this would be an impediment to dissemination, acceptance, teaching, research, et cetera.

So one of the first big major steps we had after over a year of discussion and debate, and actually very, very useful discussion, I think we all learned from these discussions, is we got together for two days at a CR headquarters in Reston, Virginia.

And we reviewed and assessed cases together, and it's amazing how a bunch of experts looking at the same case, saw things completely differently and came up with a completely different diagnoses.

And at this meeting, we came up with a some consensus on many of the toughest concepts and terminology.

That was a big major step.

The other major step is for everybody to realize that none of us had the answer and everybody was willing to give and was willing to be flexible in order to get to standardization across different institutions and different countries.

Assessment Principles

So PI Rads version two, the assessment, how when we look at the MR scans, these are some of the things we decided.

The assessment and the that you come up with from the MR scan is gonna be based on the MRI only.

So it doesn't take into account the clinical history or other factors such as PSA or even the prior biopsy.

Now what the recommendation that's made on the basis of the MRI, of course it's gonna take those things into consideration.

But when you're scoring or assessing the scan, you are first looking at the MRI in and of itself, and that's the the assessment that you're gonna come up with, the PY rads that you're gonna come up with.

As you already heard, we decided not to include in this version Mr. Spectroscopic imaging or any of the other techniques that are under develop MR under development.

MRSI has actually been around for quite some time.

There are quite a few groups around the world who find it very useful.

The reason we decided not to include it in this version of PI rads, it's very hard to make it reproducible across institutions very technically challenging.

So the intention is not to discourage the use of spectroscopic imaging, it's just that we decided at this point it wasn't ready to be put in this document.

And instead of it taking us close to three years to produce higher ED'S version two, it probably would've taken us twice as long.

Scoring System

The assessment is gonna be a fi is a five point scale, based on T two weighted imaging and a five point scale for T two weighted imaging and a five point scale for diffusion weighted imaging for each lesion.

And then for the a dynamic contrast enhanced scan, you already heard from at least one of the speakers that the importance of this seems to be diminishing.

Since we could not get agreement on how to assess this with a five point scale, the decision was made to use a very simple two point scale for DCE.

It's either positive or negative.

I'll get into the details of that later.

For each lesion, once you assess the T two, the DWI and the DCE, you come up with a final assessment, and that's also based on a five point scale, which indicates the likelihood or probability of clinically significant cancer answer.

And we'll talk about what that means also in a while.

So you're not adding these scores from T 2D, WI and DCE, rather, there's a waiting.

And again, you already heard about this for the peripheral zone, the assessment is weighted heavily for the on the diffusion weighted imagings, so DWI and uh A DC.

A DC stands for apparent diffusion coefficient, which you derive from DW, a number of different DWI acquisitions.

The transition zone cancers or the transition zone assessment is based mainly on T two weighted images.

So this is quite different than the first version of PY Raz.

And this is something that frankly evolved to a large extent as we were developing Pi Rads version two.

But there's now I think good consensus on this.

Assessment Categories

So here are the assessment categories.

So the assessment for each lesion is based on the likelihood or probability of clinically significant cancer.

And a score of one doesn't mean that there's definitely not cancer, it just means that the likelihood is is very, very unlikely.

And at the other end of the extreme, a score of a five means a assessment.

Category five means that clinically significant cancer is highly likely.

Now we when we were developing this, we debated should we put percentages here for each of these different assessment categories one through five.

And at one point we did that, and then we backed off of it because we felt we don't have data to support it.

So it is a little bit ambiguous.

A three simply means when you look at it on MRI, you don't know if it's mal clinically significant cancer or not.

What's important here is that ones and twos mean it's probably not something you want a biopsy, and fours and fives means it probably is something you want a biopsy.

But these kinds of recommendations are gonna based on the clinical situation and other factors.

So this was progress, just this defining what clinically significant cancer is.

Coming up with these categories took us quite some time when we finally sat down to actually look at images at studies and decide how we wanted, what assessment category we wanted to come up with.

We were in some cases consistent.

So this is a effort that Daniel Margolis and led, I think it was Mao's idea, ha his idea.

And these were sort of mind experiments, not quite along the same lines as Einstein's mind experiments, but basically taking every combination of T two, DWI and DCE and see what assessment category we came up with.

And the good news is was that at the extremes for the ones and twos or the fours and fives, we were actually quite consistent.

Where we ran into some issues was in the middle deciding what's a three and differentiating between a two and a three and a three and a four.

Now, we also asked Daniel to take our what we were doing at this time and put it into a simple flow chart that any radiologist or urologist could use to figure out how to integrate T 2D WI and DCE.

And this is what he came up with.

So he said, Hmm, we better go back to the drawing board because this is not gonna work.

Okay. And in fact, if we have something like that, it might as well be written in a foreign language.

Here's the Chinese version of prostate guidelines, which somebody sent to me recently.

I don't know what it says, but my guess is it's probably pretty close to PI RADS version one.

Okay, so I think one of the reasons it took us a long time to produce this document is we had to, through trial and error, come to the agreement that we need to make this simple and it's just not gonna be perfect.

Okay? So we'll tell you right off the bat, A CR PY RADS version two is not perfect.

It's a always gonna be a work in progress until we get to version three, but we decided to keep it simple so that people could actually use it.

And so here it is, it's really pretty simple peripheral zone scoring, one to five, okay, I'm sorry, the assessment final assessment is one to five and it's mainly based on DWI transition zone.

It's mainly the final assessment is one to five.

It's mainly based on T two for one, DWIs one and two in the peripheral zone.

It really doesn't matter what T two and DCER, you're still gonna be, still gonna be probably benign.

And if there's a four and five on DWI, no matter what T two and DCE are, it's probably gonna be clinically significant cancer transition zone.

It's gonna depend on T two where these we differ in the between the transition zone and the peripheral zone is in this three category.

And frankly, we went back and forth trying to figure out what a three was in the European document.

The definition of A three is anything that's not a two or a four.

Well, we frankly didn't find that particularly useful.

And we went back and forth about what should constitute a three for the peripheral zone and transition zone.

And I'll tell you, it's very difficult and this is an area that's just gonna have to evolve over time as we get more experience as everybody starts using a CR, I'm sorry, Py RADS version two, and we start testing it and subjecting it to research, this will become better defined as time goes on.

Right now, it's a bit problematic.

Oh, I don't know why the colors aren't showing up here.

That's this is problematic.

Sector Map

So one of the other things we did with the document is everybody's been using different maps for prostate cancer.

And what we did is we looked at these the maps that are out there that some urologists use, that pathologists use, that various radiologists use.

And based on what we know and what we thought would be useful, we came up with a sector map, which 36 sectors plus two for the seminal vecal and one for the membrane, this urethra.

So for a total of 39.

And I assure you that there are more colors here than are displayed there.

And a group at Hopkins put this together based on our recommendations.

Now let me tell you a couple of things about this sector map.

So when I first heard about 16 sectors or 24 sectors and then finally 39 sectors, I said, the radiologists are gonna shoot us because nobody wants to sit there and fill out forms with 39 sectors in there.

But in fact, it's actually quite simple.

The only thing you're using this sector map for is to draw either on a piece of paper or this will be incorporated into workstations or just click on a sector in there to show where a finding is.

And the reason we got to this level of detail is this is gonna be important as we try to correlate our findings with pathology and with the, and also as a to assist urologists doing biopsies using MR guidance.

You heard earlier about the central zone as one of the key zone, one of the zones in the prostate.

Interestingly enough, on none of the other sector maps that we could find that radiologists were using or that urologists were using, the central zone didn't appear.

So the central zone is now there.

And it's really important as you heard earlier, because this, at least in my experience, has been one of the most common reasons for false positives on MRI is mistaking this central zone for a big cancer.

You'll see, I'm sure more examples of this later.

Reporting Guidelines

So here's some other things that you need to know about py rads.

In the report, you should describe up to four findings with pi Rads three, four, or five.

And each should be assigned to sector.

If there are more than four suspicious findings and only the ones with the highest likelihood of clinically significant cancer should be reported.

There may be instances where you wanna report more than four.

Maybe that's your practice or you're trying to do research or maybe for biopsy purposes.

Measurements

And for measurements, this is the last thing I'm gonna talk about in this talk.

Since we're using the diffusion to do the primary is the primary determinant in the peripheral zone, we decided that's the sequence that you should use to measure suspected cancers in the peripheral zone.

Since we're using T two as the primary determinant in the transition zone, that's the image you should measure the lesion on.

In the transition zone, we don't know, really know if this is the best way to do things.

All we're trying to do right now is standardized if the lesion is bigger in the cranial cord, dead dimension on your coronal images or sagal images, that's the you should report that information as well.

And in your report, you should always indicate which image you measured it on and where you see the lesion, the image number, and the series.

Thank you.