PSMA PET/CT- Struggling with Increased Sensitivity
Presentation from the Society of Nuclear Medicine and Molecular Imaging (SNMMI) Annual Meeting. (Length of video ~ 23 minutes)
Michael Hofman discusses PSMA as the new “gold standard” and how it is significantly superior to existing imaging modalities through case studies from the Peter MacCallum Cancer Centre and a review of abstracts.
- Systemic Radioligand Therapy with (177)Lu Labeled Prostate Specific Membrane Antigen Ligand for Imaging and Therapy in Patients with Metastatic Castration-Resistant Prostate Cancer.
- 68Ga-PSMA has high detection rate of prostate cancer recurrence outside the prostatic fossa in patients being considered for salvage radiation treatment.
- AUA 2017: 11C-Choline versus 68ga-PSMA PET/CT scan for the detection of nodal recurrence from prostate cancer: results from a large, multi-institutional salvage lymph node dissection series
Professor Michael Hofman is a nuclear medicine physician in the Centre for Cancer Imaging at the Peter MacCallum Cancer in Melbourne. He has a broad interest in positron emission tomography (PET), molecular imaging applications in oncology and radionuclide therapy.
Professor Hofman has a vision of improving patient outcomes by using molecular imaging to non-invasively characterize disease thereby enabling improved selection of the most appropriate therapy for an individual patient and better assessment of therapeutic response. He has a particular interest in novel PET radiotracers, including F-18 and Ga-68, and theranostic applications including neuroendocrine and prostate cancers. Currently, he is the principal investigator of a randomized multi-center trial of PSMA PET/CT in PET/CT, a prospective trial of Lu-177 PSMA therapy and a prospective trial investigating the role of PET V/Q scanning in patients with lung cancer.
Professor Hofman previously completed a clinical and research fellowship at Guys and St Thomas in London. He is a fellow of the International Cancer Imaging Society (ICIS) and serves on the editorial board of Cancer Imaging, Leukaemia Lymphoma and the Journal of Medical Imaging and Radiation Oncology (JMIRO). He is an inaugural scientific member of the Australasian Radiopharmaceutical Trials Network (ARTnet).
v2 Updated August 2019
- What is PSMA?
- Chemical Agents
- Gallium Hbed PSMA
- Gallium vs. Choline
- Rapid Growth in Adoption
- What is the Gold Standard in Imaging?
- How to Determine Veracity of Results?
- Randomized Trial of Flouride PET/CT Versus MDP
- What if the Histopathology is Wrong?
- How Does Better Imaging Improve Patient Management?
- Just Because You Can Now See It, Doesn't Mean the Disease is New
- No Good Evidence Intervention Will Improve Patient Outcomes
- Potential for Stage Migration and Lead Time Bias
- Need for More Prospective Studies to Demonstrate Improvement in Patient Outcomes
- PSMA Provides a New Modality
We've had an interest in our center for imaging patients with prostate cancer for some time and particularly with novel traces. You can see our volumes of PET scans per quarter and we had several studies involving fluorocholine which peaked at around 60 PET scans per quarter and we also were involved in some trials of fluorine 18 as a PET scan bone imaging agent. And you can see what's happened in since the middle of 2014, is that the use of both these agents have really dropped right off and that's because of the availability of gallium 68 PSMA PET/CT, which has really overtaken it. And this is not because we are driving it through studies, but it is actually because our referral population, both the radiation oncologists and surgeons and the medical oncologists, are really finding this modality very useful and they are voting with their feet and ordering lots of studies.
So despite our growing interest in choline imaging, which we invested significant resources in building the program, PSMA really came along and all that hard work was gone overnight, and we have adapted by switching to the gallium PSMA.
Google's always correct so if we look at the Google trends, if you do a search for gallium 68 PSMA, you can see that there's a exponential growing interest in the modality. So what is it?
PSMA is prostate specific membrane antigen. This is actually a full light hydrolyzed receptor that sits on the cell surface of prostate cancer cells and tumors. It's density increases with higher grade tumors, metastatic disease, hormone refractory, prostate cancer. It is a full light hydrolyzed receptor. It's precise function is not really well known. This is not a new target. People have been trying to use this to image prostate cancers for several decades.
The most prominently one known, Indium 111, labeled to an antibody called ProstaScint. So ProstaScint was an antibody that targeted the intracellular domain of this receptor. It's performance was suboptimal, one, because it was limited by SPECT CT and resolution of but mainly because it was an antibody targeting the intracellular domain, which is not really available unless the cell is breaking down. More recently, there is J591 which is, again, an antibody which is better because it targets the extracellular epitope, but both of these are still limited by being an antibody, which means it's got slow recognition of the target and high background clearance. And the real way forward has come through these small molecule inhibitors. So these are more like peptides like Dotatate from your endocrine tumors rather than antibodies. So there are a very small structure which binds with very high affinity to the extra cellular domain of PSMA.
We have a number of agents for gallium 68, technetium, and also fluoride and the one which has probably used most around the world is PSMA 11 which is the Hbed CC. There's also a PSMA INT which seems to have similar performance and there's growing interest in the fluorinated compounds, which also have, the second generation ones, appear to have similar performance as well. The gallium Hbed PSMA was first done by the Heidelberg group in Germany and it's really rapidly proliferated thanks to their efforts in developing this compound. Now this is the first study they published back in 2013 and not that long ago, and since then there's been really rapid proliferation.
Now this is the first study we performed in our center with the gallium Hbed PSMA back in 2014 and you can see very high uptake in the prostate primary. This is a patient with a Gleason eight tumor, very high uptake in a very small lymph node, which is very hard to see on the CT scan and an osseous metastasis. So in the very first patient, this patient was scheduled for a prostatectomy at the age of 77 and on the basis of normal conventional imaging, CT, MRI and bone scan, he was going to proceed with surgery. This scan changed that course of management to systemic therapy. You can see on the MRI scan just how small this lymph node is.
How does it compare to choline? Well, it's really a lot, lot better than choline. Not a little bit better but a lot better, and that's our experience and the experience of others that are using gallium PSMA. So this is a patient that in our center had both scans and you can see the SUVmax on the PSMA on the left of 25 compared to 3.5 on choline.
So really and almost an order of magnitude difference in uptake and in the same patient, if we go a little bit higher up, there's a very small presacral node which can not be seen on the flourocholine, but is seen with confidence with the PSMA PET. So the Heidelberg group in Germany confirmed this in a study of 37 patients that both had choline PET and PSMA PET and in almost every single patient, there was a substantially higher uptake by SUV on the PSMA compared to the choline pit. You can see in many patients, 10 to 20 fold higher had tumor to background contrast. And I think the images I've seen with the FACBC are not too dissimilar than the choline, so I think we can expect that the gallium PSMA really has much higher uptake than amino acid imaging.
This is another patient, a 65 year old with external beam radiotherapy back in 2006, now has rising PSA. When we look at the gallium PSMA pit, there's a lot of abnormalities. It's hard to believe that this patient has normal conventional imaging, but when you look at the CT scan, these abnormalities are really very small. Some of these nodes as small as two to three millimeters yet the target to background contrast is very high, so there's just very high affinity binding.
So PSMA PET/CT is rapidly emerging as a potential new gold standard. It may supersede other imaging as a one stop shop. It's got potential for wide clinical availability at relatively low cost. In Australia for example, there are now, at last count, 22 sites performing gallium PSMA PET/CT. Two years ago they would have been two sites and so it's really rapidly adopted. If the US followed that course, at the moment there would be over 300 centers performing gallium PSMA if you did a population comparison. So that may be the way things progress.
We get images with very high tumor to background contrast, but it's fair to say that we have very little prospective data on improvement of patient outcomes with PSMA PET/CT. So the talk was titled struggling with increased sensitivity and the rest of the talk is going to really focus more on some principles of imaging rather than going through the evidence based for PSMA PET in greater detail, which is really a topic for another talk.
In the proceeding talks, we've heard a lot of numbers being presented, lots of sensitivities, specificities, but how do we know that these numbers are correct? We want to know accuracy, which is the receiver operating curve, a measure of sensitivity and specificity, but we need to have a gold standard against what to measure it and which of these is best? Is it CT, MRI, bone scanning, choline, FSABC, PSMA?
So what is the gold standard? And it can be quite hard to find a gold standard. And there's three types of golds and not all golds are equal. Rose gold is currently a tipping plain yellow or white gold as the metal of choice. So what is our gold standard in imaging?
The Paradox of the Best Performing Test
If we look at this study, a prospective study of a fluoride PET bone scanning versus conventional bone scanning, 44 patients prospective so potentially high quality. The ROC curve for fluoride PET is 0.99 which is almost perfect. It's a perfect test according to this study, yet, and compared to the bone scan of 0.64. So we have what we call the paradox of the gold standard, which is by definition, it's the best performing test available and there's no criteria and standard against which it can be compared.
So every time we get a better test, the sensitivity and specificity of all the tests that came before it suddenly drop lower. So if we take this example, a patient with a normal bone scan, normal fluoride, PET bone scan, if you have no better test, you'll say that the PET, that the fluoride PET was accurate. You'll say this was 100% accurate. Maybe over the subsequent years you'd be proven wrong, but when you see the PSMA PET in the same patient, you realize very quickly that this is far from 100% accurate and the accuracy of your test that was previously 99% accurate may only be 60% accurate. If we look a bit closer at this patient, these are all bony metasticies and there's really no change discernible on the CT component of the scan, and that's why they're not lighting up on the bone scan because they're not yet sclerotic. These are marrow based metastases, they may become sclerotic over time and there'll be seen after some treatment.
So if we have this PSMA PET/CT, which looks fairly convincing, how do we know that this result is true? Could this just be false positive results? And the way to determine that would be to stick needles in the bone and biopsy them, but that's really not very feasible. And even if you put a needle into a deep structure like this and it came back normal, you would not know if that was sampling error. So if we're going to do high quality prospective studies, how do we validate some of these new very high sensitivity traces? It poses us with a problem.
I was involved in a randomized trial of fluoride PET/CT versus MDP. This was a randomized prospective study involving over 500 patients and biopsy is clearly not achievable in everyone. So this criteria system was developed, which I won't go through in detail, but a hard criteria was biopsy confirming metastasis, but then you had other criteria which you could add up the score such as typical appearance of multifocal metastatic disease or increase in number over time.
The six months follow up was defined at a point where you looked back at the data to determine whether your scan was correct or not. This is not a bad way to compare modality since you're defining upfront a very strict and rigid criteria for positivity. Certainly with PSMA PET/CT, we're starting to identify micro metastatic disease that we could never really do with PET imaging before.
This is a patient with a Gleason four plus four, has had a radical prostatectomy two years ago, rising PSA, normal CT, normal fluoride bone scan, and the PSMA PET scan is abnormal. You can see very high uptake in lymph nodes that on the CT scan are really barely visible. If we go a little bit higher, we see a, what looks like a supraclavicular nodal metastases.
We do try to biopsy wherever possible, and when we first started seeing high uptake in such small nodes, our referral base didn't really believe these results so we did as much biopsy as we could. In this case an ultrasound guided biopsy of this seven millimeter supraclavicular node that confirmed prostate cancer adenocarcinoma. So one way where we can use histopathology is in patients undergoing pelvic nodal dissection. This is not a bad standard of reference because the surgeon's cutting out all the lymph nodes in a chain and then you can compare that to the imaging findings. And this has been done by the group from Munich where they took 130 patients that were undergoing either PET/CT or PET/MRI and they compared the sensitivity and specificity of CT or MRI, which is the green compared to just the PSMA PET in the other color. And you can see they find a sensitivity of about 66% for gallium PSMA and the specificity of at 99%.
So there are almost no false positive results in this series, but the sensitivity was not perfect. Obviously there is micro metastatic disease and in three cases this was because of disease smaller than three millimeters in size and in five patients, there was actually no uptake in the primary site. So we do have some instances of PSMA negative prostate cancers where other imaging modalities may still be useful.
But what about this case from our center? This is a patient who had a gallium PSMA staging PET scan and there's very high uptake in a external iliac near lymph node, which is barely visible on the CT measuring about two to three millimeters in short axis. Well, this patient proceeded to a pelvic nodal detection and the pathologic stage was N0. So if we believe the histopathology is the gold standard, we say that the PSMA PET scan is wrong, but could it be that the surgeon just missed this small lymph node and in fact the histopathology is wrong?
It's certainly possible, and so histopathology can be false negative and has false positive rights just like imaging as well. So the true gold standard can be hard to find and really follow up in this case will be critical and I'm sure it'll confirm that the PSMA PET scan was indeed correct. We know we can do better because the group from Munich now use radio guided surgery where they inject technetium 99 M labeled PSMA prior to the surgery and then the surgeon goes in with the probe. And when he does this surgery, he frequently finds tiny lymph nodes with the probe that he tells me he would missed without using that technique. So without this, your histopathology from the pelvis can't really be a true gold standard.
It's called prostate-specific membrane antigen, which sounds very good, but with increasing experience, we know that it's certainly not specific. We see uptake in a whole range of tumors, both benign and malignant. Here's a few examples. Renal cell carcinoma, very high uptake in hepatocellular carcinoma, the case of Paget's disease and a schwannoma. There's many others, but this is another limitation of the technique.
So we're going to move on from, in the theme of struggling with increased sensitivity from accuracy over to management. So I've got no doubt in my mind that from all the modalities that we've seen to date that the PSMA PET, whether it's the gallium or the fluorinated is by far the more accurate of all the tests that we have, but how does that confer? How does that change patient management? We assume that if we have a more accurate test that that will improve health outcomes but it may not necessarily be the case. A lot of treatment is done with good intentions, but there can be unintended consequences. And particularly in the prostate cancer world, there's a lot of logos like this. Fight like a warrior, I fought hard, I won. So there's a lot of incentives for the patients to fight their disease hard and attack their disease and take on this warlike mentality.
What we see with gallium PSMA is just because you see new disease doesn't mean that it has in fact progressed, because we now have a new more sensitive modality. It may be that the disease that we see was there before, we just couldn't see it before. Oligometastatic disease is a concept that I personally struggle with because I think we still often just see the tip of the iceberg with imaging and I think it's important to keep in mind that PSMA PET/CT provides a powerful new means to monitor disease. Just like people use PSA for active surveillance, you might have people with small volume metastatic disease where your best doing nothing but monitoring disease and the PSMA PET provides you with a method to do it.
So this is a patient who's eight years after a prostatectomy with rising PSA and when the gallium PSMA PET scan is done, we see focal intense uptake in a solitary mesorectal node, which measures two to three millimeters and we're really seeing micro metastatic disease. And I think the title of the slide is just because you can see it, doesn't mean you should treat it because we don't know how long that lymph nodes been there for. This is not in the classical nodal dissection. This lymph node could have been there five years ago and maybe it hasn't changed and we don't know that. So it's easy now to say let's cut it out because we can see it or let's give it stereotactic radiotherapy, but I look at an image like this and think if it's taken eight years for this lymph node to get to two to three millimeters, this is extremely indolent disease and perhaps it's best left alone.
So the test certainly enables us to localize where the PSA is coming from with a very high sensitivity and specificity, but we don't have good evidence that early intervention is actually going to improve patient outcomes, and the interventions can be quite morbid. So doing a big operation to cut this lymph node out may leave the patient much more symptomatic than if you left it alone. So there is potential to cause more harm than good. And I've certainly seen a lot of examples of that in our experience with the PSMA PET in our institution. This is a patient had rising PSA with normal conventional imaging and we can see a chain of lymph nodes on the gallium PSMA PET scan. This is the largest lymph node. It probably measures maybe a little bit more than four millimeters, maybe it's more like five or six millimeters.
And in the pre PSMA PET era, when you couldn't see this disease and you just had a rising PSA, you had really two options, which was to do nothing, observe the patient and wait for imaging to become positive. Or you could say the PSA is rising, let's start hormonal treatment. In the PSMA era, since we can see the disease, we've got several new options. We can say, "Let's cut these out." We could say, "Let's give them radiotherapy," or now that we see disease, some of the oncologists say, "Oh, let's treat this with chemotherapy." So there's additional management options, and this patient, and I'm not here to debate which is best, but in the real world, this patient after being referred back, underwent a pelvic and retroperitoneal nodal dissection and that did result in excision of all these nodes. But three months later there was biochemical recurrence and now we see supraclavicular nodal and mediastinal node and a boney metastasis. So I don't think surgery really improved the quality of life of this patient, at all the ultimate outcome of this patient. In fact, I would say that the gallium PSMA PET, by identifying disease that was perhaps inappropriately acted on, actually made this patient worse.
Here's another case of oligometastatic disease rising PSA. Conventional imaging normal, PSMA PET/CT demonstrating a very high uptake in this really tiny, four millimeter external iliac nodal metastasis. Again, before PSMA PET, the option would have been observation or maybe hormonal therapy. Now we've got additional options. This patient chose to have stereotactic ablative radiotherapy to this lymph node with the aim of curing this patient because this was a solitary side of disease. But when a PSMA PET scan was done six months later, there was recurrence in a common iliac node, slightly higher up the chain. And in the era of pre PSMA pet, the radiation oncologist, if they were going to give radiotherapy, may have chosen just to treat the entire nodal field. Now that they can see the disease, they're narrowing down their fields. And as they do that, we're just seeing out of treatment failures down the track and clearly there's micro metastatic disease that we miss.
So there's potential for a stage migration and lead time bias where we see disease earlier and institute treatments. And we have adverse effects from the early interventions and symptoms from that without changing the ultimate course in this patient. So unlike FDG PET, which are often upstaged patients and resulted in deescalation of therapy. You had someone with lung cancer, the surgeon was going to do a big thoracotomy, but you do a staging PET and you see bony metastasis, you appropriately deescalate the therapy. With some of these more advanced prostate cancer imaging, we're doing the opposite by escalating therapies and we can say that that's a problem for the referrers and not for us as imaging doctors, but I think if we are doing these more advanced tests, we need to take some responsibility for what happens downstream. It's important not just to treat the PSA or the PSMA PET scan, but we need to use the scan rationally and accumulate the evidence that changes in management really improve patient outcomes.
Fortunately, we do have some really good prospective data emerging for PSMA PET/CT. This is a good example. This is a study from Sydney which looked at patients with biochemical recurrence PSA between 0.05 and 1. This is a group that conventionally would have just got radio therapy to the prostate bed blindly and as our first speaker said, when you do that, 50% of patients will have biochemical recurrence despite a curative intent procedure. And these patients with normal conventional imaging were then subjected to PSMA PET/CT and there was a 30% high management impact by identifying a disease outside the prostate bed. And it's hard to argue that this is not inappropriate management change because if you see a bony metastasis or pelvic nodal disease, you know these are patients that would have failed if you just irradiated the prostate bed. So in some cases the radiotherapy field was in large. In the other cases, treatment was modified in a very logical and convincing way. So we need to do more prospective studies to show that we really are improving outcomes in this group of patients.
So we're increasingly recognizing PSMA as the new gold standard, as the most accurate test. In our experience, it's significantly superior to existing modalities. We definitely have a need to generate high level evidence. Now, some prostate cancers do not express PSMA, so it's not perfect, and some other cancers express PSMA. We're now imaging micro metastatic disease for one of the first times, but it's important not to treat everything that we see, and it does provide a new modality to just simply monitor disease and PSMA theranostics with lutetium PSMA is also a game changer, but not for discussion today. So thank you very much and happy to participate in any questions for the audience.