Dr. Phillip J. Koo, MD, presented “Imaging of Prostate Cancer Patients Following Definitive Therapy” at the 26th Annual Perspectives in Urology: Point-Counterpoint, November 12, 2017 in Scottsdale, AZ

How to cite: Koo, Phillip J. “Imaging of Prostate Cancer Patients Following Definitive Therapy” November 12, 2017. Accessed Jul 2024. https://grandroundsinurology.com/Imaging-of-Prostate-Cancer-Patients-Following-Definitive-Therapy/


Dr. Phillip J. Koo, MD, discusses advanced tools such as multi-parametric MRI and PET/CT for prostate cancer imaging, and how they outperform traditional, previously available methods.

Imaging of Prostate Cancer Patients Following Definitive Therapy


So today I was asked to talk about imaging prostate cancer in patients following definitive therapy, so a little different.  

These are my disclosures.  

Oh, first the ARS question.  The best FDA-approved tool for detecting metastatic disease is, A, Prostascint; B, PSMA PET/CT; fluciclovine PET/CT, Sodium Chloride PET/CT, or C-11 Choline PET/CT.

All right, great, so let’s–all right, hopefully we’ll clear this up at the end of this talk.  

Great, so the objectives are to talk about the detection of recurrent disease and metastatic disease, and we really divide this into two areas.  For local disease we’re going to talk briefly about multi-parametric MRI, and for advance disease we’re going to talk about PET/CT.  

In reality, I think there is a little bit of overlap, and I think that overlap is going to grow as we learn more about this different imaging agents, but we’re just going to spend a little bit of time on prostate MRI because it was covered so well this morning.  

So for local disease we’re really talking about multiparametric MRI, and just as a side note, I think Robert and Michael did an amazing job talking about rarely the whole spectrum of the use of prostate MRI upfront and how it sort of integrates with this idea of MR-guided biopsies.  I think the data is always so heterogeneous, and I think the problem that we have with a lot of these imagine studies is they’re retrospective, their single-site.  

Heterogenous patient populations, protocols are different, and just endless amounts of errors that really just doesn’t create pure data that I think is difficult, and that’s why we see a lot of different papers that talk about sensitivities in 60% range and then another paper that talks about in the 90-something percent range.  I think that’s a huge problem.  I think quality MRI, not just in the community, in academic settings, too, I think is difficult.  

So PI-RADS was created to help solve this, and I just recently gave a talk about controversies in prostate cancer imaging, and I think that’s a huge take-home message, is that we talk about prostate MRI like it’s a great tool, and it is, but there are real issues with quality.  

So I think PI-RADS is a great tool to help standardize the technical protocols and standardize the interpretation of these.  But in reality, I think probably 80% of the people in this room, when you order a prostate MRI I don’t think you’re getting PI-RADS, you know?  

And I think that’s a problem, and I think we need to demand that and expect that from our various radiology groups that give us a product that I think is a higher quality product.  

So anyway, we’re going to talk about multiparametric MRI in the setting of local recurrence, and that’s for the use of MRI in patients who have biochemical failure after radical prostatectomy and radiation.  It could evaluate for local recurrence versus normal tissue versus scar, and it could help guide biopsies.  

So this is a meta-analysis, and this is about prostate MRI upfront, and since we’ve heard about this already, I just wanted to just make the statement that it is a good tool.  This is an ROC curve and it’s clearly not along this line, because if it’s along this line that means the test is useless.  So I think it is a good tool, we just need to find the right place for it.  

In local recurrence, which is part of the AUA guidelines, it is indicated to get a multiparametric MRI.  So this is one study that looked at 187 men who underwent mpMRI followed by TRUS-guided biopsies of the prostatic fossa.  Local recurrence was detected in 132 of those 187 patients.  The median PSA in those patients was 0.59.  Median lesion size was 1, and I agree, small lesions are difficult to detect because of the resolution capabilities.  

But overall, the sensitivity specificity, positive predictive value, negative predictive value were good, but in this case the specificity wasn’t quite as good, and I think the reason why is because sometimes you get scarring, you post-surgical changes, inflammation and things like that, that could be interpreted local recurrence, but may not prove to be.  For local recurrence, pirads is not the way they’re read because it’s just not meant for the use in local recurrence.

So the conclusion is it helps target biopsies to more accurately diagnose radiation failure and to possibly determine who may benefit from more local or focal salvage therapy.

All right, so we’ll sort of fast forward to talk about metastatic disease.  

When we’re talking about detecting metastatic disease we’re really focusing on two things.  We want to detect disease in bones and we want to detect disease in bones and we want to detect disease in soft tissues.  This is the chart from the RADAR group recommendations which was published in Urology, chaired by Dr. Crawford in March of 2014.  

It doesn’t really project well, but it basically creates three buckets with regards to the patients in different stages, so newly diagnosed patients, biochemically recurrent patients, and M0CRPC patients, and it gives recommendations with regards to how you should image them and when you should image them.  This was based on using bone scans and CTs.  

In that article that we published in 2014, we met I think in 2013, we discussed the advance pet imaging tools that were available, but they just weren’t ready for prime time.  So we’re just going to go over a little bit about the history of imaging and talk about what’s available today and how we could sort of take that home and practice a little differently in the future.  

So bones scans are great; they’ve been around for decades.  So they were first FDA-approved in the 1970s and we’ve been using them every day, every day, for decades.  I think we saw a real change in practice in 2012 when C-11 choline PET/CT was approved.  The biggest problem, though, was that no one in this room had access to this.  That approval was basically limited to the Mayo Clinic in Rochester because they had a cyclotron [phonetic] that NDA was specific for their facility.  

So though we read about it, there was a lot of buzz about this, you couldn’t get it, and I think that was extremely frustrating.  Fast forward four years, 2016, we have a new radiopharmaceutical called F18 fluciclovine.  It’s called axumen [phonetic].  It now is available in a lot of markets.  I think right now probably 17 or 18 different markets have this available, and we’ll talk about that a little later.  

So bone scans, they’re planar images, they’re 2-D.  You just take a picture from the front and a picture from the back.  Again, 1970s, it uses diphosphonates.  The nice advantage to this is it can image the whole body, but the downfall is it only images the bone, it doesn’t image the soft tissues.  

The other big disadvantage of bone scan is that it’s imaging a bony response that your bones have, not the disease itself.  So when a metastatic lesion lodges itself somewhere it creates this bony response, this reparative response, and that’s what it’s capturing, so that’s why we talk a lot about flare with these bone scans.

Then CT came along in 1971, and this was great as it developed because it could detect soft tissue disease and lytic bone mets and sclerotic bone mets.  But typically we would see bony lesions a little later than we would with a bone scan.

Sodium fluoride PET/CT, though, we’ve talked about it a little more over the past five years.  It was actually first approved in 1972, and the advantage was that it was whole-body imaging, it used PET/CT, which gives you better resolution, but the disadvantage was it was bone-only.  But when it came to imaging the bone it was a bone scan on steroids because it was just really, really much better.

So this was a case of a bone scan that was read as negative, and I agree, but the patient had high-grade, high-risk prostate cancer, elevated PSA, sodium fluoride PET/CT just showed numerous number of mets in the axial and proximal appendicular skeleton that the bone scan missed.

Whole-body MRI, this is something that we read a little about in the US, it’s not really available, but in Europe it is included in the guidelines, the European guidelines.  The data with whole-body MRI for detecting bony disease is actually very good.  I think the biggest problem is the protocols vary from site to site.  

Granted, there are papers out there that have talked about a standardized protocol, but they’re just not implemented.  In the US there’s no code for this, so if you imagine doing a protocol on a patient, whole body in an MRI scanner, as an administrator I have to tie up that scanner for an hour, an hour and a half, two hours.  Granted, the – – is something you could tailor to protocols, but that’s a lot of lost revenue.  

And the US, again, it’s just not something that we have a code for, it’s not reimbursable for, so maybe there might be a handful, not even a handful, maybe one or two trials that might be using this in the US.  It’s possible for us to tailor this to detect soft tissue disease, but again, our protocol then is going to get even longer and it’s just not feasible at the moment.

So Version 2.0; this is when we had technologies to image bone and soft tissues in one exam.  

The first test that actually did this, which was the first PSMA imaging agent, was processing first approved in 1996.  It uses the PSMA molecule and it images bone and soft tissues.  The problem was it didn’t use PET/CT, it used a gamma camera, and it was removed from the NCCN guidelines recently.  

I think this was just not the best test because you had the image over several days, so it wasn’t convenient for patients, and I think the biggest problem was the PSMA molecule, the monoclonal antibody attached to an internal epitope PSMA, not an external epitope.  So basically the cell had to be dying or necrotic in order for this – – radar tracer to see the internal epitope that it needed to attach to.  

So I think for a variety of reasons it just sort of fell out of favor, and I would argue in 2017 we should no longer be discussion Prostascint, so hopefully can put that to bed.

But we’ve made significant advancements, and this is what I like to call Version 2.1.  It’s imagine bone and soft tissues in one exam using PET/CT, radiopharmaceuticals and PET/CT.  

So when we talk about PET/CT we’re really talking about a modality, and then the app that we inserted to it is the radiopharmaceutical, and that is what provides the functionality of the output of any PET/CT exam.  So similar to what an iPhone and app, the iPhone is the computer, the machine that runs it, and you can do whatever you want on that iPhone as long as you have a proper or applicable application.

So the first PET radiopharmaceutical that was developed clinically was FDG; that was a glucose analogue, it was FDA-approved in 1997.  It had real limited use in prostate cancer, because what we knew though in prostate cancer that initial diagnosis, and even in biochemical recurrence, prostate cancer cells has relatively lower use of glucose utilization.  

At the University of Colorado some of my research – – with multiple people there was to somehow alter the metabolism of these prostate cancer cells to make them use more glucose, but we weren’t able to do that successfully.  One of the limitations also is the fact that the FDG gets excreted by the genital urinary system, so you get a lot of activity in the bladder and that bladder could obscure the prostate and what not.  

Another disadvantage of FDG is that inflammation, infection, those white blood cells use a lot of glucose, because oftentimes you get a lot of false positives because of infection or inflammation, and your body actually has a lot of inflammation at various times.

So Version 2.2, C-ll Choline.

This was a game changer, again; FDA approves in 2012, the use of PET/CT, and it was indicated in patients who were biochemically recurrent who had non-additive or inconclusive imaging with bone scans and PET/CTs and other imaging tests.  It was incorporated into the NCCN guidelines a couple years ago, but the biggest problem here was limited availability.  

That C-11 isotope only has a half-life of 20 minutes, so if it only has a half-life of 20 minutes, you basically need an onsite cyclotron in order to produce it, transport it, inject it, and get the patient image.  That’s why Mayo in Rochester was the first and only to have this in that clinical setting.  It was available at other places under 90, cost recovery, but it was expensive.

I think there was good data regarding the use of C-11 choline.  The sensitivities and specificities, these don’t project well, but multiple studies showing that it had improved detection compared to other bone scans, and this is a list of a bunch of other single-site studies that talk about the use of C-11 choline.  

A lot of this comes out of Europe and again, the Mayo Clinic, but again, single-site, the data isn’t quite the level that we have for a lot of our other clinical trials.

Again, we talked about the limitations, but because of that 20-minutes half-life it really did not become–it limited its commercialization.  So at the University of Colorado we actually had a cyclotron and we tried to produce this and commercialize this ourselves, just internally, but the cost upfront were just too much to overcome.

So 2.3; this is where we actually have something that is commercializeable [phonetic].  All right, show me the money.

This is fluciclovine, – -.  It was FDA-approved in 2016 and its label, this comes directly from the FDA label, suspected prostate cancer recurrence based on elevated PSA levels following prior treatment.  I think when we zero in on that we think biochemical recurrence, but I think we can expand as you see fit, because it’s actually pretty broad, in my opinion.

It’s a synthetic amino acid analogue actively transported into the cell via these transporters.  Doesn’t get metabolized, and it stays in the cell.  So once that F18 isotope gets it way into the cell it starts emitting this positron which undergoes this annihilation reaction, which then our cameras can capture and create these real pretty images.

So this comes from Emory University, and they did some head-to-head trials of fluciclovine versus Prostascint, and you can see that the sensitivity was better than Prostascint for local disease and for extra-prostatic disease.  Specificity was better here.  I guess it was a little worse compared to Prostascint locally, and basically the positive predictive value and accuracy was better.  Overall, just a good test.  

When you compared it to CT scans at varying PSA levels, it shows here that the fluciclovine was much better than the CT scan and varies all PSA levels.  This isn’t surprising because a PET/CT using fluciclovine is basically a CT scan with a radiopharmaceutical as opposed to just a CT scan.

So this comes from Italy, from Nanni [phonetic] and colleagues.  This compared fluciclovine versus choline head-to-head in a prospective study with 50 patients, single site, and basically fluciclovine picked up disease in more patients than choline did, and it picked up more lesions in those patients than choline did.  Then when we look distributed by PSA levels, it also performed a little better than choline, so the basic conclusion here is that it’s as good as choline for detecting recurrent disease.  

Another study that was published in the Journal of Urology recently talks about the detection of disease in the prostate bed outside of the prostate, the whole body at very PSA levels, and at a PSA level of less than 0.79 it shows its detection rate in the whole body just over 41% using fluciclovine.  So I think it’s a very–in terms of performance, it actually performs very well, and I don’t think we’ve ever had a diagnostic agent, imaging agent, that could detect recurrent disease at those low PSA levels.

So I think the big question though is, is this just all lead time bias?  Who cares?  We have these tools that can detect disease; who cares?  Is it going to change outcomes?  And I’ll tell you, that’s a hard question, because this is a diagnostic test, and what happened after that, it’s hard to track all that and really attribute a certain benefit to a diagnostic tool that occurred upstream. 

So granted, this slide was for screening, but what we really need to show is that we detect disease sooner and death occurs later.  We’re not there yet, but I think that’s the direction we need to be head, but again, I don’t know if that’s realistic for an imaging test.  So I’ll show you some data that shows its ability to maybe have some value in our clinics.  

This is a study that was published in JAM in 2017.  It talks about the impact of fluciclovine PET on target – – definition for post-prostatectomy salvage-rated therapy.  Their end conclusion was that there was significant difference in target volume, higher doses to the penile bed.  Dr. Beyer shows here the information from the Falcon trial that was just released at Astro that talked about, again, changes in that radiation planning.  

I believe a secondary endpoint of that Falcon trial is looking at response to salvage radiation, and those have not been released yet, but I think that’ll be interesting to see as well.

This is change in salvage radiotherapy, 87 patients, 42 were randomized to fluciclovine and 34 of 42 has positive findings, similar to what Dr. Beyer showed.  Radiotherapy changes, decisions changed in 17 those of those 42 patients.  Two did not get radiation because they had extra-pelvic disease, and the radiation field changed in 15 of those patients.

So this idea of change in management I think is real.  So that Falcon trial showed 61.5% change in management, a significant change in management.  When we look at data from Galleon 68 PSMA it showed a change in management of 53% of patients, and another PSMA agent using a small molecule showed change in management in 51% of the patients.  So I think these are clearly change in management, but again, what does that do to outcomes?  And I don’t think we know that.

So this is sort of a table that I put together that talks about the imaging tools that we have available and the Medicare coverage.  So TEK-99 [phonetic] bone scans, they’re good for bony disease, they’re not good for soft tissue disease, but they are covered by Medicare.  CT scans can detect bone disease; they’re not as good as a bone scan because we’re waiting for that sclerotic reaction, but they do detect soft tissue disease and visceral metastatic disease.  They are covered by Medicare.  

Whole-body MRI, very good for detecting bony disease.  Soft tissue disease, we’re not sure, it depends on the protocol, but in the US it is not covered by Medicare.  Sodium fluoride PET/CT is great for bones, doesn’t really detect soft tissue disease.  It’s been covered under the National Oncologic PET Registry for

the past three years.  So this is sort of breaking news.  If you have been ordering sodium fluoride PET/CT you have until December 14th, because after December 14th the NLPR registry is going to close.  CMS indicated that they’re not planning on renewing it, so what will happen is the data will be analyzed and they’ll make a final decision whether they cover it or not. Personally, I don’t think it’s going to be covered, so I think you have it until three or four weeks to get your last sodium fluoride PET/CTs in.

FDG PET/CT, bone, soft tissue, not great.  Patients with later-stage disease, more advanced disease, I think FDG’s PET/CT gets better.  It is covered by Medicare, but it’s only covered for patients who are getting subsequent treatment strategy scans, so patients where you’re looking for a response to therapy.

C-11 choline is great for bone, great for soft tissues.  It is paid for by Medicare but it’s limited.  F18 fluciclovine PET/CT is great for bone, great for soft tissues; it is covered by Medicare.  PSMA, put four plusses for this because based on the data thus far it’s probably the best for bone and soft tissues, but it is not covered by Medicare, it is not FDA-approved.  That being said, it’s widely used in Australia and widely used in Europe.  

In the US there are a few, mostly university medical centers that now have INDs for this, that will be offering this under an IND cost recovery program.  When I met with some people with UCSF this past week, the number I was being quoted was around $1,500 to get the scanner.  So if you have any patients who are interested, you can refer them to UCSF, the University of Wisconsin, I think maybe Sloan-Kettering or Cornell to get these scans for some out-of-pocket cost.

So if we go back to the RADAR guidelines, for newly diagnosed patients I guess we could use sodium fluoride but we only have three weeks left to do use it.  PSMA is not covered by Medicare, it’s not FDA-approved yet.  

However, if you have a patient with high-risk disease and you really think they might have metastatic disease and you have a negative bone scan and CT and they have the means to pay for it, maybe it might be worth paying to get a PSMA scan upfront. And this is sort of the next area of research for a lot of the companies, is figuring out how these advanced PET pharmaceuticals perform in the initial diagnosis setting.

For patients with biochemical occurrence, I think our best bet right now is going to be fluciclovine PET/CT versus choline PET/CT, and those are approved by the FDA and available on reimbursed.  For patients with N0CRPC, I think it’s sort of yet to be seen.  FDG could be used.  Fluciclovine, I don’t think we talk about much in that space, but I think it does have a role there.  PSMA PET/CT could use there, sodium fluoride, again, limited, and choline.  

So the conclusion; advanced imaging tools such as multi-parametric MRI and PET/CT, they’re far better than the traditional methods that we’ve had available to us.  The conclusion here is they’re here to stay, you know?  So it’s really up to all of us to figure out how best to use them and incorporate them into our practices.  And the last conclusion is I think it’s important that you get to know your radiologist, to get to know who they are, what their specialties are, what their sub-specialties are, and expect and demand a little more from them.  

I think the radiologists are here to support your practices, urologists and oncologists, so I think it’s important to develop that working relationship and that communication.  If you actually read the AUA guidelines about multiparametric MRI, it actually includes a paragraph about that communication circle and that feedback circle, how it benefits both sides, and Dr. Cookson alluded to that as well.  

So I’m turning out of time.  I’ll make up time on the other lectures, but just quick pieces.

Here’s an 80-year-old man, Gleason 3+3 treated and he had a PSA of 6.5.  His PSA rose a few years later.  

Patient got an MRI, it was read as indeterminate.  They talk about this little area here.  They didn’t use PI-RADS, and again, this happens all the time.  

We got an adjuvant PET/CT, so this focal area right there, biopsy-proven to be recurrent disease.  

Another case, Gleason 3+4, treated with EBRT in 2012.  PSA started to rise, the bone scan was negative.

They got a prostate MRI, showed this odd lesion here on the right pubic bone, sort of hedged whether it was met or not; I don’t think I was really buying it.  

We got actually a PET/CT.  What we did find was there was a left external iliac node that was lighting up internal – -, which on CT you could actually see, but without the contrast it’s sort of hard to detect that.  And that bone lesion didn’t have any uptake, so I did not think it was metastatic disease.  

This was another patient, bone scan and CT were negative.  

Got a fluciclovine PET/CT showed this new focus of uptake here on the right pubic synthesis.  There was something–there was a little sclerotic focus here on CT, which I thought to myself I tried to give the patient the benefit of the doubt, is that just some inflammation of degenerative disease?  Went back in the patient’s jacket.

Two years ago had a CT scan, there was nothing there, so I think this ended up being metastatic, at least we treated it, we couldn’t biopsy it, and we also had a focus of recurrent disease in the right seminal vesicle.

So I think this is going to bring up a lot of questions in the future that we can talk about with regards to how you treat – – metastatic disease and what you do, especially since you won’t be able to biopsy a lot of these lesions, and we don’t have those answers yet.  

Anecdotally I know that in Australia and some of these places that are using PSMA are doing extended lymphadenectomies; their increasing their radiation fields.  With those extended lymphadenectomies you might see more lymphedema that we haven’t seen in the past.  So I think these are all questions that we need to work on together.  But again, I think this tool is here to stay.


Phillip J. Koo, MD, is the Chief of Diagnostic Imaging at the Banner MD Anderson Cancer Center in Phoenix, Arizona. Prior to this, he was Chief of Nuclear Medicine and Associate Professor of Radiology at the University of Colorado School of Medicine in Aurora, Colorado. He completed his fellowship at the Harvard Medical School Joint Program in Nuclear Medicine in Boston, Massachusetts. Dr. Koo is a diplomate of both the American Board of Radiology (ABR) and American Board of Nuclear Medicine(ABNM).