Matthew Smith: The next part of this session will focus on the clinical scenario of when low-volume disease on conventional imaging goes to high-volume disease using next-generation imaging, kind of a continuation of what we discussed earlier this morning. We've had an audience request that they want the bells put to use. I have the goat bell. [inaudible] has the cow bell. So feel free to speak as long as you want. We'll put the bells to use. Michael, go right ahead.

Michael Hofman: Thank you. I might slow down for my talk then. When low-volume on conventional imaging becomes high-volume on our next-generation imaging... How frequent is this? What should we do? None of my disclosures are really relevant for this talk.

So on the left, we have a whole-body bone scan, and on the right a whole-body MRI. On the left, it's a radiolabeled bisphosphonate. It doesn't image tumor. This is my bread and butter nuclear medicine, and you are imaging a surrogate endpoint, the bone reaction to the tumor. On the right, beautiful images from Anwar Padhani. I don't do whole-body MRI myself, and even a pep like image. So if you are the patient, you might wonder why is my doctor subjecting me to this insensitive kind of archaic imaging on the left? My interest is in PSMA PET, and there's been an explosion of research in the last five to 10 years.

On the left, we have a bone scan this time with SPECT CT. And I think if you are doing a bone scan, you ought to do it with SPECT CT. We see three bone metastases. One day later, the PSMA PET shows more than 20 bone metastases. The charted criteria defines high burden disease as more than four bone metastases, one outside the axial skeleton or a visceral metastasis on the CT. So we have this disconnect between conventional imaging and next generation imaging. Three patients with bone scans, all symptomatic with back pain, newly diagnosed, hormone sensitive prostate cancer, and many medical oncologists are not interested in what the PSMA PET shows because they want to guide their management using phase one randomized controlled data. And we know from the STAMPEDE trial that all these three patients, despite the presence of metastatic disease should have radiotherapy to the prostate bed. And if you do some next generation imaging, it might misguide you. You See all these other lesions and then you change your management but the patient outcomes are worse.

Hazard ratio was 0.68 and 42% of patients in the STAMPEDE trial had low burden metastatic disease. And, they looked at this further. And what they found was that as the number of bone metastases increased, there was a relationship and the overall survival benefit was quantitatively beneficial in those with three or less metastases. And it didn't matter where the location was. Let's pause and rewind for a minute. Let's re-look at those bone scans. In fact, none of these patients have metastatic disease. If you do a SPECT CT, which they should have had, but this was not performed in the STAMPEDE trial. This is all benign degenerative findings. And this is more common than we might think. There's data showing that as you go from planar two dimensional scans to three dimensional SPECT bone scans to adding in SPECT plus side by side CT to a SPECT CT, which is the current standard of care, you improve accuracy. And more accurate imaging is not stage migration.

Stage migration refers to the concept where you have more sensitive imaging. So we're moving patients to seeing more disease. But if we're increasing specificity decreasing false positive results, I can tell you, this is not stage migration. And if we're increasing reporter agreement, this is also not stage migration. So, how much more accurate is PSMA PET CT? Well, we have great data from the November funded pro PSMA study, 27% greater accuracy, 92% for PSMA compared to 65% for the combined findings of a contrast enhanced CT and a bone scan with SPECT CT, two scans versus one scan also much better for the patient experience to have a PSMA PET scan in 60 minutes, rather than a CT bone scan on separate days.

Let's deep dive a little bit further because there were much more equivocal studies in the conventional imaging arm, randomized in the pro PSMA study. For the primary endpoint, we considered equivocal studies negative and using that the false positive rate for conventional imaging was 6% compared to 1% for PSMA PET, but perhaps more reflective of clinical practice is if you consider these equivocal findings positive, because in my experience, radiologists have a fear of missing out. They do not want to miss a bone metastasis or a liver metastasis, no matter what. And particularly in non-cancer centers, radiologists have a tendency to over call findings because they don't want to miss that metastasis. So we did a sensitivity analysis. This is in the supplementary appendix on the Lancet publication. And if you consider equivocal findings positive for metastatic disease, there was a 23% false positive rate for CT bone scan combined compared to 7% for PSMA. And this is where the SPECT CT. If you took just a plain R bone scan, it would be even higher. So that's one in four patients being erroneously called metastatic when they don't actually have disease.

Here is one such patient on the left bone scan showing unequivocal bone metastasis, CT showing liver metastasis. The PSMA PET showed no metastatic disease. Well in the trial, we had six month follow up. And the piece of my PET was proven correct. This patient proceeded to a radical prostatectomy and had undetectable PSA postoperatively. Management would've been misguided on the basis of standard imaging. So, in this analysis, it wasn't really number of bone metastases. It was actually the number of osteoblastic lesions. And my theory is that as the number increases the likelihood of false positives on conventional imaging decreases. So in fact, the 1, 2, 3 lesions, they may actually be M0 disease. Whereas the true positive disease, if you have 10 lesions on a bone scan, no doubt you've got metastatic disease.

Now, I do believe that the tumor burden may be important. I do believe in the up scope of effect. So part of the effect in the STAMPEDE trial probably is real. And certainly if you're considering systemic treatment intensification, double therapy, triplet therapy, you may want to know is that high burden or low burden, but do you want to sit there trying to count lesions? Because I could tell you, if you show the same scans to two radiologists, one will see five lesions. One will see seven lesions. We want to do it more accurately, and we can do this with PSMA PET because we can automatically quantitate the tumor volume using automated tools. These will increasingly become available over the next three to five years. And I think they will be validated and we will get a tumor volume in mills, which is accurate. Actually imaging tumor, not a surrogate outcome like osteoblastic lesions. Not progressing to the next slide. Someone might need to help me push to the next slide.

Speaker 3: Sorry that you can't ring the bell.

Michael Hofman: No, you can't ring the bell. We can. I think I've only got two or three slides left, but for some reason, presentation has paused on this slide. But it's important to note really that our imaging technology just continues to improve. So when a referral orders a bone scan, they don't specify, I just want to planar two dimensional bone scan. They just say, I want a bone scan. And our bone scanning technology has improved. So even within a defined imaging modality, our technology continues to improve.

This is a nice manuscript from a German group published recently in the Journal of Nuclear Medicine, showing that perhaps a 40 mil cutoff using those quantitative pep parameters translated to a low burden on conventional imaging compared to high burden. We suddenly need more evidence to know what PSMA quantitative thresholds and the best way to get this data is to incorporate PSMA PET into all our prospective clinical trials. And then we can look back.

Don't worry, two more slides. So imaging is progressed. Look, it's an honor to be the first to have the cow bell.

So total body PET is coming. Even PET technology continues to improve, continues to get more sensitive with this technology. You can do a whole body PSMA PET in two or three minutes. Quite remarkable.

And there we go. Paused again.

Speaker 3: Sorry, Michael.

Michael Hofman: That's all right. I think there's one slide left. There is only one slide.

It's might be my fancy videos.

Speaker 3: Yeah, I guess you were too interesting.

Michael Hofman: A picture tells a thousand words.

Last slide. So really our technology advances outpace our evidence base, but I don't think we want to get caught behind using inferior technology to stage our patients less accurately. To me, it's somewhat implausible that less accurate imaging will better inform patient management and outcomes. We don't want to make wrong decisions using 2D planar bone scans, more accurate imaging is not stage migration that refers to an increased sensitivity. And a lot of this advanced technology is decreasing false positive results. Lastly, please don't extrapolate this presentation to metastatic castration resistant prostate cancer. That's another talk. Because I think serial imaging in a metastatic castration resistance setting with prostate cancer working group is really well validated. And we ought to be cautious with our new techniques like PSMA PET CT.

So with that, I'd like to thank everyone and apologize for a few of the pictures.