Distinguishing True and False Positives on PSMA PET Scans "Presentation" - Courtney Lawhn Heath

February 5, 2024

At the 2024 UCSF-UCLA PSMA Conference, Courtney Lawhn Heath emphasizes the complexity of interpreting PSMA PET scans, highlighting the need to recognize common false positives and the limitations of treating PSMA PET as a straightforward diagnostic tool for prostate cancer. Dr. Lawhn Heath advocates for a nuanced approach that integrates anatomical imaging and clinical context to accurately interpret findings and improve prostate cancer management.

UCLAUCSF video logo stroke


Courtney Lawhn Heath, MD, UCSF, San Francisco, CA

Read the Full Video Transcript

Courtney Lawhn Heath: Hi, everyone. I guess I'll take my mask off. Please excuse my voice. I'm getting over a cold. But so, I'm Courtney Lawhn Heath. I'm a radiologist and nuclear medicine physician practicing clinically here at UCSF, and I'm here to give a practical talk about false positives on PSMA PET. Now that we've found out that the actual compound may not be that important, there are some generalities that we can talk about across the spectrum here. Now, I'm honored to be here because I became a nuclear medicine physician largely because of PSMA PET, and as I got to watch as a trainee, watch this compound revolutionize the care of prostate cancer patients.

That being said, it's not simply a magic tool that will give you a binary answer of, "Yes, this is prostate cancer. No, this is not prostate cancer." I think that a lot of problems with PSMA PET interpretation come from thinking of it this way, and it kind of reminds me of this... I don't know if you guys ever saw the HBO show, Silicon Valley, but there was a fictional app on that show called SeeFood. It was an image recognition app, and the only thing it could tell you was whether or not an image was a hot dog. I think sometimes people treat PSMA PET this way, like, "Okay, great, it's prostate cancer or it's not."

Unfortunately, that is not how PSMA works, but I think we can actually get kind of close if we keep in mind some of the common and important false positives with this. I don't know if you guys have heard people make the joke, it's kind of old and tired at this point, but people have said, "Hey, they should take the P and the S out of PSMA." Prostate-specific, because there's so much false-positive uptake. Ugh, LOL.

Anyway, it's not actually that bad, but there's a fair amount to be aware of. So, let's get started. I don't have any disclosures here. This talk is super simple. I'm just going to discuss a basic framework for evaluating incidental findings on PSMA PET and then just walk us through, by body region, those main false-positive findings. Now, of course, as we've seen, here's the normal biodistribution of PSMA. This is Gallium 68 PSMA-11, I'm sure very familiar to all of us at this point. Also, there is a little bit of, in addition to all the major players that have been talked about, there's also some uptake along the tracheobronchial epithelium and some other areas that are a little bit more subtle. But anyway, when we see uptake that doesn't fit this typical biodistribution, we should talk about a basic way of approaching that.

And so basically, there are benign and malignant findings that can appear on PSMA PET that often, I'll say, with benign lesions, they often, though not always, have pretty low-level uptake, benign findings tend to, and can typically be ill-defined. But really if we're calling something benign, the location and the anatomical imaging are really going to be the key, the location and the anatomical imaging appearance. Malignant lesions, on the other hand, are much more variable in their uptake characteristics and for distinguishing exactly what malignancy we're dealing with, it's often not just about the appearance of the lesion on anatomic imaging, but actually the distribution of the lesions within the body that might be the most suggestive of the underlying malignancy. So, let's now step through the body from head to toe looking at the most common benign and malignant incidental findings on PSMA PET.

And then after that, we're going to look at findings that can potentially occur anywhere in the body. Okay, so let's start from the top with the head and neck. Now, interestingly, you may be aware that because of PSMA PET, we've literally learned about the existence of accessory salivary tissue in areas that people didn't know existed before. And so accessory salivary tissue is a major source in the head and neck of incidental PSMA uptake. One area that trips up my trainees a lot is the infamous cheek met. It can be very hot and asymmetric like you see here, but generally, it's just benign accessory salivary tissue. Also, the major salivary glands can be strikingly asymmetric in some patients. For instance, this is a patient who had had a prior right submandibular gland transfer in order to get radiation to a right-sided floor of mouth tumor.

So, they had their left submandibular gland removed and relocated. So, at first glance, you might look at this and be tempted to call the left side a submandibular nodal met, and I've seen that read before. I've seen that done before. But really, that's normal. That's the normal submandibular gland, and it's the right side that's abnormal, although benign, and the anatomic imaging and obviously history really help a lot here. We also see uptake in meningiomas. Now, the classic teaching here for our boards is that meningiomas have uptake on DOTATATE PETs and somatostatin receptor imaging, but they also show uptake on PSMA PET. Now, thankfully, usually MRI is sufficient to confirm whether this is a meningioma or a dural met, but just be sure not to over-call a met without additional imaging characterization. When in doubt, more imaging is always helpful.

In the neck, thyroid adenomas can have uptake, as can parathyroid adenomas actually, giving fluorocholine a run for its money there. So, we need to recommend dedicated imaging for these findings in the neck as clinically indicated, usually starting with a thyroid ultrasound. For malignant entities in the head and neck, we have head and neck squamous cell carcinoma. It can be quite PSMA avid. And this particular case was one that we reported that was the first that we knew about this. The patient didn't know about it. Nobody knew about it until he got his PSMA PET scan for staging his high-risk prostate cancer. Adenoid cystic carcinoma can also have uptake, as well as a lot of other cancers of the salivary glands and ducts.

There's also thyroid cancer. The uptake is mainly seen in differentiated subtypes, but we found that it's really quite variable actually across subtypes. And then gliomas also have PSMA uptake. It turns out PSMA is expressed in glial tissue. Now, moving down into the chest, we have pretty much any cause of inflammation, like organizing pneumonia. In this case, the patient had just had radiation to the overlying rib there, which was very effective. We don't see the uptake in that rib anymore, but we do see a little uptake in this radiation pneumonitis subjacent to that rib.

Reactive lymph nodes, especially in the hila and mediastinum. We see these all the time. But notice, again, generally pretty low level. Sarcoidosis, several different pneumoconioses, gynecomastia, and even elastofibromas can also have PSMA uptake. So, a lot to be aware of there. Now, on the malignant side in the chest, the most common cause of abnormal malignant uptake in the chest is actually primary lung cancer. But you can also see uptake in thymomas, breast cancer, and mesotheliomas.

We actually had a patient a few months ago who has both prostate cancer and breast cancer. Now, some benign things that you might come across in the abdomen. Inguinal lymph nodes. Now, this is something that we actually see all the time, low-level uptake in inguinal nodes. And when that's isolated, it's very unlikely to represent metastatic prostate cancer. Because as we'll recall, the inguinal region doesn't directly drain the deep organs of the pelvis, so it'd be very unusual for there to be nodal mets there in the absence of closer-in nodal mets.

Liver and spleen hemangiomas can also look kind of crazy and concerning at times on PSMA PET. And here's where contrast-enhanced CT or MRI really will come in handy and make the difference. You can also see inflammatory uptake in diverticuli, and not just colonic diverticuli but Meckel's diverticuli too. This is an acute case report of PSMA uptake in a Meckel's diverticulum. Also, you can see uptake in adrenal adenomas and renal oncocytomas. So, multiple benign causes of uptake. Also, unfortunately, you can sometimes see PSMA uptake in the prostate in non-prostate cancer conditions, like prostatitis and BPH, although I've found this to usually be pretty low-level uptake in most cases.

In fact, having a sense of more and less suspicious uptake patterns in the prostate can be helpful, and it can help with our confidence, especially when we're looking for suspected but not yet confirmed disease in the prostate. So, I want to just very quickly go over some of those common patterns that we might see. So, diffuse uptake in the transitional zone is generally the least suspicious distribution, followed by central zone uptake. Now, I find the transitional zone and central zone to be almost impossible to distinguish on PET, but maybe that's just me. But the key is that the uptake is sparing the periphery of the gland. When the focal uptake is quite asymmetric, that's more suspicious even if the uptake is still fairly central, and the most suspicious appearance is asymmetric uptake at the periphery of the gland.

Now, like all guidelines, this is not perfectly predictive one way or the other. It's actually based on real-world data, which, as we know, is never completely perfect. It's from the PRIMARY trial, which looked at PSMA PET-CT and multiparametric MRI in men with suspected prostate cancer. They compared imaging to the gold standard of biopsy and derived these patterns from that. And as you can see, the division there is neither perfectly sensitive nor specific for all cases, but it's still a pretty helpful general rubric, and certainly, you can see the likelihood really increasing as we get more peripheral and asymmetric in the distribution.

Additionally, something else to keep in mind that trumps all of this is the intensity of uptake. In this PRIMARY trial, they found that for any prostate focus of uptake with an SUVmax of more than 12, 100% of them turned out to be prostate cancer at biopsy. Now, of course, this only applies to the prostate gland itself and this one trial, and generally, using SUVs is not considered a clinically valid practice. This is not a clinically validated thing yet. But it's just something to keep in mind, that very intense uptake in the prostate should definitely increase your index of suspicion. Again, usually, there will be MRI and pathology to help, but we can still make some educated conjectures on PSMA PET.

Okay, so going back to false positives in the abdomen and pelvis, it turns out there are a boatload of abdominal-pelvic malignancies that take up PSMA but are not prostate cancer, and these include hepatocellular carcinoma, renal cell carcinoma. In fact, some have suggested that PSMA may actually be superior to good old FDG PET for staging these cancers because the uptake tends to be so high, although that's not mainstream practice yet. And really, uptake also varies by subtype, with clear cell RCCs being more avid than other subtypes. So, it's certainly not a perfect solution. But there can also be uptake in all of these other malignancies, and this is not the complete list either.

So, quite a lot of other things can take up PSMA, other cancers in the abdomen and pelvis. Now, we have those entities that transcend individual regions, and I group those into three categories: nerve stuff, bone stuff, and infection and inflammation. Now, nerve stuff includes benign tumors and, of course, autonomic ganglia, as we've heard about, the classic PSMA incidentaloma that trips up all the new residents. Recall that these can be unilateral or bilateral.

We've talked about these, the stellate, the celiac, the hypogastric, and the sacral. And of course, hypogastric and sacral are particularly challenging given their location near some of the areas where we really worry about nodal metastasis. But generally, a lower level of uptake and symmetry can be very helpful. For the bone category, it's pretty much you name it, there's uptake in it. Lots of benign lesions and malignant lesions have uptake in the bones. Anatomic imaging is typically critical to being completely confident in calling these things benign. We've got degenerative changes. This is an uptake in a degenerative osteophyte in the lumbar spine.

We have benign fibro-osseous lesions like enchondromas and fibrous dysplasia. Those are most classically in the ribs, as we've heard, where they often look like a lucent lesion with a well-corticated rim, although we don't always have a CT correlate. Fractures can take up PSMA. Hemangiomas with their typically benign corduroy or salt and pepper CT appearance. And Paget's disease can also take up PSMA, and that classically involves entire bones or the entire hemipelvis. So, morphologic imaging is helpful, and distribution's helpful, but it's not always super classic.

I find this to be a helpful rule of thumb for solitary bone lesions. This is from the PSMA PET Procedure Guideline, and I'm just going to quote it. This is from Dr. Fendler, who just spoke. "In the case of single lesions, especially in the ribs, and the absence of a definite morphological correlate typical for malignancy, interpretation of mets should be cautious to avoid over-staging." So, I'm not saying not to call it, but we're not not saying it. No, it's a little complex though. Just don't jump right into calling a met necessarily. Now, luckily, there is a framework in the literature that could actually support us in this a little bit more, and it's called PROMISE.

It's a molecular T&M staging system that we use for our reporting. But even if you don't use it for reporting, it has all these helpful flowcharts in it to help us decide the probability of prostate cancer based on morphologic imaging. So, just a quick example of that. Here's a patient with locally recurrent disease in the prostate, but also a single focus of faint uptake in a left-sided rib with no CT correlate. So, you can actually go through the little bone flowchart in PROMISE, and we said there's no lesion on morphologic imaging. There's a single focus of uptake. The intensity is quite low. There's an intensity rating called the expression score, which is also in PROMISE, which is very simple. That means per PROMISE criteria, this lesion is likely negative. So, you can help give us confidence not to over-call every little ditzel, but in a systematic and informed way. So, I encourage you to check that out.

The next category is infection and inflammation, which of course can happen anywhere in the body. This is an example. This is osteomyelitis of the left pubic bone, and you can see a large amount of phlegmon also extending into the soft tissues there with PSMA uptake. Of course, morphologic imaging, and definitely clinical history, very helpful in these cases. Also, malignancies that can turn up anywhere include all of these entities: bone, skin, lymph. And this is a case of Hodgkin's lymphoma here. Note the low-level uptake. I feel like that's pretty typical for lymphoma.

Here's a myeloma case. Now, that's tricky. We might be tempted to say, "Well, hey, it's a lytic lesion. It can't be prostate cancer because prostate cancer is sclerotic." But actually, I'm sure if you've read enough of these PSMA imaging or these just prostate cancer imaging cases, early prostate cancer mets can be lytic. They don't always read the textbook. So, it can be tricky. Sometimes pathology is all that we can rely on. Now, before I finish up, I just want to also remind us, we've been talking about false positives, but PSMA-negative prostate cancers also exist, false negatives.

And you can see that with this patient whose prostate cancer is really only avid on FDG-PET. Definitely an exception to the rule there. This is a close-up of his disease in the liver where you can see it's basically photopenic. Luckily, it's quite rare, only about 5% of primary prostate cancers, and it tends to be seen more in longstanding, castration-resistant metastatic disease, in which case it has de-differentiated over time. Typically, these cancers have a neuroendocrine phenotype and are much more aggressive. But just be aware that these do exist, although they're much more rare.

So, in conclusion, it's important to be aware of the many causes of benign and malignant uptake on PSMA, not just prostate cancer, and of course caution with over-calling isolated bone uptake, especially in the ribs. But I know, I acknowledge, the struggle is real, but frameworks like PROMISE can really help us with equivocal findings. And I want to thank the incredible faculty, staff, and trainees here at UCSF in the section of Molecular Imaging & Therapeutics, and to the organizers for having me, and to all of you for your attention. Thank you so much.