14-15 Overview & In Bore

Introduction to Targeted Biopsies

My name is Claire Tempi.

We're going to start the session on targeted biopsies.

I'm gonna do an overview
and then follow up with
our experience from the Brigham Hospital
of in-bore prostate biopsies.

Traditional Detection Methods

As we all know, the ability
to detect prostate cancer has been based on palpation
by digital rectal examination,
which is obviously palpating the tissue's
elastic properties.
Focal prostate cancer is firmer to the finger.

I'm not sure if the surgeons still agree
that the finger doesn't matter,
but tactile sensation on the digital rectal exam is
apparently the best way to, one of the ways to try to find
and to detect prostate cancer.

Prostate Specific Antigen (PSA)

The prostate specific antigen.
We've spoken a little bit about it already today.

There are some subgroups
and some numbers in PSA
that we should talk a little bit about.

Remember that zero to four is nominally the normal scales.
Up to four is normal, but some people would say two.
It's really age and race specific.

So there are nomograms that will tell you
for your given patient, what is the upper limit
of normal in PSA, we know we've spoken already
that PSA is made by both cancer and hyperplastic tissues.

And a way to try to differentiate that,
if I can skip down here to density, is to put the prostate
specific antigen level
and the prostate gland volume together in the,
in a specific individual.

And we can calculate
and we do in our reports, calculate the 3D volume
of all prostate prostates
and allow calculation of PSA density, which will then,
if it's of a ratio of greater than 0.1, is concerning.

Prostate specific antigen velocity.
PSA velocity rate of change is also important
and a velocity of baseline going up to 0.35 is thought to be significant.

And there are a lot of numbers here that can be analyzed.

And the free PSA has Pete mentioned already less than 10% is
obviously concerning for doing a biopsy.

The data and the information
that's gleaned from PSA is really irreplaceable.

So to say that PSA should not be done in anybody anymore is
obviously a mistake, I think.

And there's a lot of information in PSA
and tracking family history, of course,
as we've mentioned already as well is an important one.

Standard Diagnosis with Ultrasound

So let's just look at the way the standard prostate cancer
was diagnosed up until recently.

Ultrasound of course,
is the standard technique which is used
to image the prostate, primarily used to show
where the prostate gland is
and how to subdivide it for biopsies.

It's also used to guide brachytherapy.

It really is not a good tool for detecting prostate cancer.
At least 40% are isoechoic.

There are ways, however,
that prostate ultrasound could be made better.

And there are folks like Barry Goldberger
and Ethan Halper in the group at Jefferson
and folks in Europe and other places
who really are doing very nice work with ultrasound.

And I think it's a very underrated modality.

There are extremely exciting things with strain
and shear wave imaging that are coming out in other organs
that could be applied to the prostate,

but be that as it may today, most ultrasound
of the prostate is based on gray scale ultrasound,
which is not good for detecting focal prostate cancer.

Traditional Biopsy Approaches

So what are the biopsy approaches that have been in
the diagnostic workup of men with prostate cancer being used
for the last 30, 40, 50 years?

Transrectal ultrasound guided using the ultrasound probe
in the rectum and a sort of systematic sampling
of the prostate, dividing it into right left base apex.

And the number of cores taken per session will vary from
urologist to urologist
and practice to practice in general, I think it's around 12
to 15 of a local urologist in Boston does
20 cores all the time.

So it varies. There is a theory that more cores more likely
to hit a prostate cancer
that hasn't really been borne out in the literature,
but still the number of cores varies,
but it is a significant number.

And just remember those numbers when we come talking to,
talking about targeted biopsy later, right, left base.

Now the other approach which can be used usually in planning
of treatment, if somebody's interested in focal therapy
and they really want to know about the focal vocality
of a tumor, is to do what's called the template
saturation or mapping biopsy.

They've all got the same name, same terms,
this is a grid placed up against the perineum.

Samples are taken at five millimeter increments
across the whole prostate.

You can see the diagram here.

Some people use ultrasound to guide that procedure.
Some people do not. And it can be, again, anything of 60
to 80 cores per session
and usually it's done in an operating room with anesthesia.

These are all non-targeted biopsy procedures.

Biopsy Reporting and Pathology

So what do we learn once we've done a
biopsy of prostate cancer?

Just before we get into the targeting part,
a cancer diagnosis, the typical prostate biopsy report,
we'll tell you if it's on the right
or if it's on the left versus a mapped biopsy report.

It's very rare to get mapped once when we, we are trying
to correlate our MRI to biopsy findings.

We're probably lucky if we know if the biopsy's
positive on the right or the left.

Sometimes you'll be told if it's in the apex
or mid gland, but the reporting is very, very random
and very different depending on where it's done.

Some places use the BAREL mapping technique,
which is very nice, and you'll get a report
that shows similar to the locations
and sectors that we're using in PI-RADS.

You'll see the biopsy report mapped out that way,
but that's really unusual.

Then you get the Gleason grade and then if there's presence
or absence of PIN or intraepithelial neoplasia or ASAP
or any of the other subcategories
of non prostate cancer diagnoses, the number
of positive cores is used by a lot of analysis.

And then the percent positivity of a core.

Now percent positivity is used in our institution versus
cancer core length or total core length.

And I think there's a big difference here,
and this is somewhere where the pathologists are going
to have to get standardized.

I think cancer core length
or CCL is really
where we should be at measuring it millimeter
by millimeter cut percents can be really confusing.

You can have 10% at each end of the tip,
and that's called 20%.
Or you can have 20 at one end of a core,
and that's the same as the person I just
described with 10 and 10.

They're not the same. They could be two
different prostate cancers.

So there's a little bit of an area there.

I think that's really important for us
as we move into this realm of doing biopsies
to understand much more clearly
what our pathologists are doing.

Gleason Grading Changes

So we've talked a little bit about the Gleason grading,
and I'm gonna go through this really fast
because we've already heard about it today,

but clearly there's been significant changes
and the overall changes have led
to overall in all men being diagnosed with prostate cancer
to higher scores being assigned.

And the big issue is
that the small angular cells are now called fours.

So what used to be a three is now a four.

And so it's increased the high grade rate.

The goal of it, a lot of it was,
and this was led by Jonathan Epstein
and a bunch of expert pathologists, was really to try
to improve the concordance
between the biopsy Gleason
and the final radical prostatectomy
Gleason, as you've already heard.

And so this is an example of that.

Clinically Significant Disease

Let's just talk now again about the clinically significant
disease 'cause this is relevant
to this paper which came out in Radiology last year.

We all know that clinically significant should mean disease
that matters, disease that needs to be treated,
or biopsied, and then
of course focus on cancer that matters.

There are many different definitions of
what is clinically significant disease.

Some of them are listed here.

Most of them will agree that it's a Gleason four.

And what percentage of length
and percentage of core involvement is
debated across the groups.

Flaws in Ultrasound-Guided Biopsy

Why are we changing? Why are we even considering
a change in the standard of care?

The ultrasound guided biopsy really is a flawed approach.
Unfortunately, it's non-targeted.

It's blind biopsy, essentially it's systematic sampling,
random sampling, and it has been shown
that it misses cancers
and unfortunately misses a lot
of clinically significant cancers.

And we know that many of these men have been upgraded at
subsequent radical prostatectomy, not just
because the pathologists are unable to determine the,
the agreement across the Gleason, but really
because the sampling is picking the smaller lesions
or not finding the big lesions.

Evidence for Targeted Biopsies

There are many, many papers
that have appeared over the last five years I suppose,
which really illustrate the value
of the targeted versus the TRUS.

First of all, written in this paper for localization,
where you can see that by localization.
MRI was more accurate across the board here,
particularly in the transitional zone
and in the mid gland regions.

You can see here that the MRI correctly localized the
cancers better than TRUS.

And then if we look at it in a pre radical prostatectomy
stratification, MR guided prostate biopsies here compared
to its systematic 10 core biopsies,
again found the truer representation
of the higher Gleason grade in these patients
who had both MR guided and TRUS guided biopsies.

This is coming again from the European groups.

Illustrative Case Example

So this is a kind of an anecdotal case upon which we should
never base our practices,
but this is a case that clearly,
I think illustrates one of the problems.

Here's a 62-year-old man with a PSA of 15.

He had one core positive on the left side of his prostate
that demonstrated at three plus three.

And you can all see, and we've been looking at these cases
all day, there's a huge lesion on the right side here,
low on T2 subtraction, a DCE, you can see the PK analysis across the bottom here.

This is a PI-RADS five without a doubt,
with extracapsular extension.

And this can be missed on
transrectal ultrasound guided biopsy.

This is not an anterior lesion.

This is clearly a very posterior lesion
millimeters from the rectal wall.

So unfortunately it is again, a random sampling
and I think these are the illustrative examples as
to why things need to change.

Meta-Analysis and Systematic Review

Now, I'm supposed to be doing an overview
and then I'll talk about in-bore
but this is a very nice meta-analysis
or sort of an overview paper,
really a systematic review from the use from,
from a combination of folks.

Actually Caroline Moore is the first author of this,
but it represents a lot of people looking at a lot
of the different papers and really has a nice
summary of what's going on.

Showing targeted biopsy is more efficient,
and that's one of the most important points to drive home,
is that we're going to be taking,
we're gonna be diagnosing clinically significant cancer
with less cores, so it's a less morbid procedure.

So the mean biopsy target cores are about 3.7.

And this is true in our study, which is coming out soon
with 3.3 0.8, ours was 3.7, this is 3.8.

Anyway, it's less, much less compared to the 12 cores.

The targeted approach found avoids the clinically
insignificant cancers in up to 10%
of patients in these papers that were reviewed.

So Mark Emberton and Hamma
and the group at UCL have been pioneering the sort
of concept of where imaging should fit in before a biopsy or
after a biopsy or where.

And really, I think we're gonna talk more about this a
little bit later on, whether MRI will play a role
as sort of a screening test to get to do a biopsy.

Methods for Targeted Biopsies

So how do we do these targeted biopsies?

We've got choices. We can sort
of obviously remain in the ultrasound domain using cognitive
approaches where one would have an MRI
and then use the cognitive ability of a radiologist
or urologist to sample.

We can do MRI in-bore or we can try to fuse the ultrasound
and the MRI together we've been doing in-bore biopsies
for many years, and I'll share our experience in a minute.

And clearly that's a very nice way to do it.

It's very clean and simple.

You target using the modality that you use to find it,
but there are obviously serious cost-effective,
cost effective, cost-effective issues and access.

Of course, not everyone has the ability to do these in MRI.

So one of the other ways to do this is to really concept
of fusion biopsies, which my colleagues will speak about
after me, which I think is a very pragmatic approach,
to taking the best of both worlds.

The efficiency of ultrasound with the,
with the accuracy data of MRI.

So in-bore biopsies can be targeted where MRI will target
and guide, they can be done either transperineal
or transrectally.

I'll talk to you about transperineal
and I think Satna will talk
to you about transrectal in-bore biopsies.

And then out-bore, as I said, there's the fusion ones
and then the cognitive biopsies.

And then of course, we've already discussed these as well.

Cognitive Approach Example

So let's just look at an example of a cognitive approach.

So here's a patient who came to us from Cincinnati.

He'd had, I guess he didn't see you Satna, sorry.

He probably should have, he had this PSA history here,
five going up to six, and then it went up to 7.9,
and then he had a biopsy and his first biopsy showed Gleason
six three plus three 5% one core positive.

So he was put in active surveillance.

And then within less than a year practically,
or a little over a year, his PSA went up to 12.

And so the question was, should we repeat the TRUS biopsy
or do an MRI scan?

And he actually was coming to Boston for his,
he was coming to Boston for his lymphoma follow-up,
or sorry, his leukemia, CLL.

And so we did an MRI.

So here's the MRI scan,
and you can see, I've got another slide.

We'll show you our abnormal lesion.

So here's really a very ill-defined focal area
of abnormal T2 signal, not very good.

Probably like a PI-RADS three level, I think
that one is, which is really non circumscribed rounded
and only mildly hypointense on T2.

But when we go to diffusion,
which remember dominates in the peripheral zone,
you can see on his 1400
and corresponding ADCs that there's a clear
and specific mass right there that has restricted diffusion.

And this is a PI-RADS four lesion.

So the urologist came and looked at these images
and said, oh, that's very easy.

I should be able to sample that
directly under ultrasound guidance.

After he had looked at these images with me, he went
and did a cognitive sampling.

And of course, let me
before, I'm running ahead of myself.

So the dynamic contrast imaging also correlated.

You can see the early contrast enhancement there.

And this is a PI-RADS positive, a PI-RADS,
DCE positive lesion with a type three curve.

So the cognitive biopsy,
did indeed find significant disease.

Gleason nine was present there on the biopsy, Gleason seven
as well in another core.

And he went on to have surgery
and was found to have a Gleason seven at final pathology.

This man traveled the country, didn't he?

Cincinnati, Boston, New York.

But that's somewhat typical sometimes.

But anyway, you can see here that
that works when you have, you know, somebody
who really can look at those images and do it.

This is what people do in breast taking breast MRI data into
breast ultrasound sampling.

Cognitive can work in-bore biopsies.

In-Bore Biopsy Devices

There's been a fair amount of development of devices
and robots and enabling tools
to allow physicians access the prostate in a
narrow bore space.

These are just an overview of some
of them which have come to the market
through the Invivo one, which is Transrectal,
which is IGO group and DynaCad.

You can see here the fusion systems,
which have we've talked a little bit about already
and we'll talk more in the next part of this session.

Many, many different varieties using different forms
of registration, different forms of display.

Fusion Systems and Yield Comparison

And an important thing
yet another paper here from the Arnold Villas Group in France.

And while I mentioned this, there's a very nice, if any
of you has, any of you have iPads, there's a wonderful app on the iPad that,
it's not Arnold, I'm trying
to remember the name of the Philipp Poosh,
I think is his name, who developed this free app
that you can look at for prostate MRI on your iPad.

So just file that thought for later.

But anyway, so they've done,
they've compared the yield from targeted and systematic here,
and you can see again, of course, targeted
is higher yield.

And again, repeating the same results
that many people have shown.

Now, targeted biopsies show the higher grade disease,
which is much more clinically significant.

Role of PI-RADS in Targeted Biopsies

So again, why are we using PI-RADS in this group?

Of course, this is obviously the motivation a lot
behind PI-RADS was to help define which lesions
need targeted sampling.

And I think most of us agree that really PI-RADS three,
four and five will be the ones
that we want to target for sampling.

There'll be debate about the threes,
because this is of course, as you've heard,
a challenging area for us to define.

It's what's not a two and not a four.

But most of us for sure will agree that any lesion
that hasn't been sampled
or a patient who has pathology,
that looks like only three plus three.

And if they have a PI-RADS four
or five, that lesion needs to be sampled to make sure
that they diagnose and detect any presence
of Gleason four disease.

UCL Group Paper on Yield and Definitions

So this is the paper. This is a paper
that came again from the UCL group last year,
comparing the yield based on various different definitions
for clinically significant disease.

And if you look at definition number two,
which was a Gleason greater than three four in any core
length greater than four millimeters,
there are negative predictive values were in the high
eighties when you went to a different definition of clinically significant disease, which is more, you know,
more greater tumor burden here with.

Now the Gleason four is the predominant pattern
and the tumor length is longer.

The negative predictive value goes way up
to high 90, 96, 93
and 94 across three different readers.

So this study was nice. And how did they do it?

I should have said this in the beginning,
what they did was 1.5 T non endorectal coil imaging.

The image analysis was performed by the radiologist,
and then separately, of course, a transperineal templated biopsy at five millimeter
increments was used as sort of the gold standard
to validate and to compare the results.

And this is the way, this is the accuracy results for the MRI.

So I think that's useful data,
and that was not done using PI-RADS,
and it's important to know that.

But clearly the results are very compelling.

Brigham Hospital Experience with In-Bore Biopsies

So we've been doing prostate biopsies
for quite some time at the Brigham,
and we've been able to use this technique
because we have the advantage of having some NIH funding
to do this, and computer scientists who help me do that
with software developed through the 3D Slicer
and ability to allow for direct non rigid registration
of pre-acquired multiparametric MRI.

And then subsequently a week or so
or two weeks later, the in-bore MRI examinations to register those, that those data points.

So the advantage of course, as I've already said, is
that you can do it either one way or the other.

The transperineal is as an advantage in so far,
there's really no infection risk.

You sterilize the skin, the needle goes straight in.

And one of the most important aspects
of the transperineal is
that you can access any part of the prostate.

You can sample the anterior aspect in a huge prostate gland.

You can go in to those 12 o'clock locations with ease.

There's no difficulty in getting to any part as opposed
to the transrectal approach, which in many
of the big glands, you will not be able to reach that far.

The disadvantages, as I've already said, of course,
is magnet time inconvenience.

It's not that inconvenience.

It's an hour to an hour and a half procedure.

It's not that bad. Obviously cost
and it's not universally available.

Case Examples from Brigham Experience

This is an old case
and it just shows you going back
to the very beginning when we did this at 0.5 Tesla,
this was the classic story back in the day
and still is the day where you see patients who present
for these biopsies repeatedly rising PSAs
and negative biopsies.

This man had a significant two centimeter lesion that was,
even after all that time, was still a localized cancer.

So targeting sampling is very important.

And here's another case to illustrate this rationale because you've got a 58-year-old man here whose
PSA, he's on an active surveillance protocol.

His PSA is going up and up and up
and he's keeps getting biopsied
and being told he's a Gleason six disease,
one would suggest he shouldn't be an
active surveillance anymore.

The MRI kept showing a two centimeter lesion,
and this was sort of before the urologist
and many people who really had come to the agreement
that targeted biopsies were necessary.

They were doing TRUS guided biopsies,
which were not cognitive, not informed by the MRI,
and they were really kept missing this lesion,
which was an anterior lesion.

And this is the classic story,
and here's the images to go with the story.

You can see in the T2 images, the DCE images and the ADC
and the diffusion, this is DCE here.

At 12 o'clock you can see a clear lesion
that's very hard to get at TRUS guided biopsy.

Workflow and Setup

So in our system,
we do our patients in a wide bore Siemens magnet,
which is 70 centimeters across.

We use a transperineal approach.

We put a template up against the perineum.

We acquired the images, as I said,
several weeks prior to biopsy.

And then on the day of the biopsy, we sample directly into
that using the registration tool in the 3D Slicer.

And when you can obviously sample the lesion here,
you can sample a T2 abnormality, a diffusion abnormality,
and then the PK abnormalities,
which may show focally different regions
within the same lesion.

And so it allows us to get site specific pathology
and correlate those backwards.

And here's some of the workflow.

My colleague, Kamal and colleague does a lot
of these biopsies now, and you can see the setup.

I'm gonna, I'm running out of time,
so I'm gonna go a little bit faster through this.

Most of this is all homemade,
sort of fairly simplified wood based frames
and templates that are standard used for transperineal brachytherapy.

It's the same plastic plexiglass templates.

Results and Analysis

We just wanna show you the results of a study
where we had three radiologists review all images prior
to biopsy and identify all targets.

And we pre, we were publishing this,
this will come out in Radiology next month, I hope,
90 patients were reported here.

Our patient population divides into three groups.

We have men who've never had a prostate cancer diagnosis.

We have men who are on active surveillance,
obviously have a prior diagnosis,
and men who have had prior treatment
and are thought to have failed locally

overall in our patient population, 3.7 targets per patient,
56% or so are positive.

And you see in the subgroups as you'd expect in men
with prior diagnosis of cancer,
the targeted biopsies were much more commonly positive,
but pretty good in the patients who'd
never received a diagnosis.

These are typically men who've had negative TRUS
guided biopsies in the past.

The anterior lesions are very commonly cancerous.

And then looking at the pulse sequences in which one had the
highest yield, just like the remainder of the literature,
the ADC is the most common.

An example of a nice case here you can see in the DCE
and then here, the lesion
and then the images are showing the needle going
to the tip where you see the lesion.

The needle tip inside the lesion are very important
to really validate where you're sampling from.

And we do gradient echo images,
and now we've actually shifted to using T2 weighted
images that are acquired very fast to show the tip
of the needle inside the lesion to really prove
that we've sampled that There's a lot of work coming out,
really looking at decision analysis now
to see the value added.

And is this really going to be a pragmatic approach
to biopsying patients?

And really showing, and I don't wanna go into this,
not enough time to go through all this,
but it's in your handout material.

The analysis really showed that you get fewer
and better, again, as I've already said.

And the future research will be really looking into the
diagnostic parameters.

The benefit being the high negative predictive value,
I think of the MRI, and we're going
to be hopefully reducing the number of over diagnoses
and targeted biopsies.

Future Directions and Guidelines

The English have really been moving forward on this
with some of the guidelines here, which are now from Bupa.

For example, one of the private insurers in the UK saying, consider MRI
for all men for biopsy for suspect cancer
and appropriate biopsy strategy depending
on the MRI findings.

So they're taking it into sort of guidelines
and more national approaches for looking at these, the,
the application of targeted biopsies,
which I think is so important.

So again, I think, like Pete, as well said earlier,
we think our predictions for the future really are
that the blind biopsy is going
to become unacceptable if it's not informed
by imaging to show where you're going.

The MRI does detect clinically significant disease,
and the new standard will be for clinically significant,
either probably in a pragmatic world,
will be the ultrasound fused systems.

There are some concerns about registration
and accuracy, which we'll talk about,
and our whole gland approaches are going away.

So instead of doing this from now on,
we're obviously doing this and then correlating
with pathology and immunohistochemistry here.

So thank you very much for your attention.