Read the Full Video Transcript

Louise Emmett: It's great to have the opportunity to talk about SPECT because it's something that I think is going to... They're not rotating. That's such a shame. Okay, normally these rotate, but Jeremie made a comment before about how we were improving our cameras. But I'd like to say that these two images, you can see the one on the left and then the one on the right, this is a patient who had a marked response to Lutetium PSMA therapy with two doses. They were done on the worst possible SPECT camera that you can imagine. It had no quantitation, but post-quantitation analysis has really helped that. Actually, the images are not... I'm missing the top of the slide. Maybe try the next slide. We'll see.

These are my disclosures, and biomarkers to guide treatment, what do we actually need? And as Jeremie's gone through at the moment, what we have at the moment is we have PSA, it's an amazing biomarker, and we have RECIST and PCWG3 with bone scan and CT scan, particularly in the registration trials. But PSA progression within 12 weeks is often considered unreliable because of the possibility of flare. It's not used early. And PCWG3 requires confirmation with two plus two for disease progression and can take months. Some patients will actually die without radiographic progression. There's really a need for better imaging if we can get the evidence for it to better guide our patient's treatment, particularly with Lutetium PSMA. When we look at how patients progress on the registration trials, 55% of the patients on therapy demonstrated PSA progression prior to dose three, week 12.

A proportion of our patients actually progress really early. In VISION 30% of patients didn't have a significant response to treatment. We need to be able to identify those patients if we want a better personalized treatment. Better dynamic biomarkers will allow better personalization of treatment and hopefully improved patient outcomes.

This really beautiful slide was put together by James Buteau and Michael Hofman and it's the screening imaging for SUVmean from the TheraP trial. And we're getting really good data about screening imaging. We know that patients who had very high SUVmean have a good response to treatment. 89% of those patients had a 50% response, and patients in the lowest quartile with a low SUVmean have a much lower 50% response rate at 29%, but that's still 29%. 29% of those patients actually have a 50% response rate with a low SUVmean.

We need to be better. We shouldn't be excluding patients just on SUVmean, we need to have better biomarkers to help decide. Why is there that disparity? Why do 29% of patients with low SUVmean, when we know they don't have particularly high PSMA expression, we know they've got heterogeneous disease, why do they still respond? It's because we're missing half of the puzzle. We can talk about PSMA expression, we can talk about PSMA density, but we don't have any information on radiation sensitivity or resistance. The screening imaging tells us nothing about that. In order to know more about that, we need to know about mechanisms of radiation resistance and we need to know a lot more about the genetics of the tumors that we're treating. We do have this great nomogram. This nomogram gives us a very good idea based on clinical parameters. It will tell us whether a patient will or will not respond.

That's clinical and imaging. The imaging parameters we've talked about and then whether a patient's had chemotherapy or not, whether they've got liver metastases or not, what their hemoglobin is. But I would say those things are fairly prognostic markers for any treatment failure. It's not specifically to Lutetium PSMA that these will predict whether they're not going to respond or not. And I wouldn't choose not to put a patient on treatment because they've had chemotherapy or not.

When we are looking at how we can identify whether a patient actually has radiation-resistant or sensitive disease, because I think that's super important in the future in terms of when we're going to choose whether a patient has chemotherapy or whether they have targeted radiation treatment. We know that radiation treatment induces DNA damage, single-strand, double-strand DNA breaks, potentially bystander effects.

But when you get to the mechanism of cell death, it gets much more complicated. Most cells that have radiation damage die from apoptosis, but we know of multiple other mechanisms, some of which are immune-mediated. We know some genetics are associated with that. TP53 induces the inability to apoptose, but it's likely many other genetic factors are involved. And we have many cells that forget how to die. I think of it as radiation-resistant clones as zombie cells with multiple DNA damage, lots of mitotic problems, but they continue to limp along anyway. How can we identify those cells?

Predicting radiation sensitivity and resistance is super, super complex. But I will put to you that there's a very easy way that we can actually measure that. We can measure that by looking at the response to the first dose. We're going to talk about interim response treatment using multiple biomarkers, but also early response treatments so that we have the opportunity to intervene because we know we're going to have agents coming along that will help us do that. RECIP is really the first part of that, and Jeremie's spoken about that. That is using PET, using volume assessment with new lesions to look at disease progression. And what they did was they did a baseline and a week 12 PSMA PET in patients undergoing Lutetium PSMA, and they found that if you had an increase in volume plus new lesions, those patients with progressive disease on imaging had an 8.3-month overall survival compared to 21.7-month overall survival if they had a partial response, as in a reduction in volume on their serial PET imaging. And then if you did PSA plus RECIP, two biomarkers, PSMA and PSA, you actually got a nice stratification of responders and progressors. Super interesting, really nice data.

The question is, can we do it in SPECT? And this was the subject of a PhD that Dr. Sarennya Pathmanandavel has been doing within the prospective LuPIN study. LuPIN was a 56-patient prospective trial, six doses, six weekly that we undertook at St. Vincent's. Each patient had a SPECT CT undertaken after each dose of Lutetium PSMA done at the 24-hour time point. Each of those images were quantified. And what she found was in our patients' 56-patient study, spectrum of volume increased in 32%. And when we compared that to PSA progression-free survival, patients with any increase in total tumor volume had a progression-free survival of 4.5 months compared to 7.1 months in patients who had either stable volume or reduced volume. What's interesting was volume was very important in predicting whether a patient was responding or not responding on PSA PFS, but change in SUVmax or SUVmean was not important. And that's also been found in the PET studies with serial imaging. A change in intensity is less important than a change in volume.

And that's the Kaplan-Meier curve of progression-free survival for any increase in total tumor volume. And the other thing was, can we use PET, can we use SPECT? One thing that Sarennya did was she did a comparison of the volumes between the dose one SPECT and the baseline PET images. And there was a very nice correlation between the volumes on the two. You definitely don't see smaller lesions on SPECT compared to PET. The spatial resolution's not as good, but we don't make patient decisions in terms of stopping or starting treatment on very small lesions. We do it on significant increases in tumor volume.

Then the question was, can we actually take it earlier, not use it as an interim response biomarker, but an early response biomarker? Because if we're going to consider changing treatment in these patients who are not responding to treatment or were primarily resistant, we need to do it pretty early before they enter marrow failure or have other problems such that they can't actually change treatment. This is 127 men from the ReSPECT study who were treated with Lutetium PSMA I&T; they received a median of eight gigabecquerels and Lutetium PSMA SPECT was once again undertaken, vertex to mid-thighs. It's a 25-minute image on a standard camera that you would use for a bone scan, 24 hours following treatment. And then the patients had quantified analysis of the SPECT images at baseline, dose one, and then dose two. What this is really doing is it's measuring the response to the first dose of treatment.

And what the study showed once again was an increase in SPECT total tumor volume was associated with a shorter PSA progression-free survival, 3.7 versus 6.7 months. And then once again, if you added PSA response to that at six weeks, increased PSA and increased total tumor volume, patients had a 2.8-month progression-free survival compared to nine months if a patient had any reduction in PSA and reduction in SPECT total tumor volume. You could use the two biomarkers together. This also small study but looks promising for using SPECT for treatment response.

Just last month, another study came out from a Swiss registry, once again using early SPECT. Baseline and six weeks in 73 patients treated with lutetium PSMA I&T, they also used quantitative analysis of the SPECT of cycle one and two. And all of this was done actually with MIM software for the SPECT analysis. They found that 75% of patients had stable or reduced total tumor volume at cycle two. And then median PSA progression-free survival for any increase in total tumor volume was 5.1 months compared to 9.2 months in patients who had any reduction in volume or stable volume. Similarly, median overall survival for any increase in total tumor volume was 10.3 months versus not reached.

And when they looked at SUVmax or SUVmean, as in all the other previous studies, a change in SUVmax or a change in SUVmean is not predictive of progression-free or overall survival. When we look and we're using the same agent that we're using for treatment, we're treating the PSMA, we're imaging the PSMA receptor, change in intensity should not be used to decide whether a patient is or is not responding. Just looking at the hazard ratios, response to total tumor volume for longer overall survival had a hazard ratio of 0.28. If they looked at patients who had a response in PSA at six weeks for longer overall survival, the hazard ratio was 0.21. And when they looked together, looked at the change in total tumor volume and the change in PSA, the hazard ratio for overall survival was 0.09 as a composite and progression-free survival 0.11. It looks like if you use the two biomarkers together, your ability to stratify progression-free survival and overall survival actually improves nicely. Two separate studies now that have shown SPECT at baseline and six weeks looks like it might be very nice at stratifying patients into responders, non-responders.

When should we be doing a SPECT if we're going to be doing a SPECT? It's very inconvenient to get patients to come back at 24 hours, and it's quite difficult. This is a study that we've been doing in 30 patients where we've been doing the baseline PET, we've been doing a post-therapy, four-hour SPECT, and a 24-hour SPECT. And that's what we've got here. This is the four-hour SPECT on the same patient, the PET, and the 24-hour. And they look pretty similar, certainly on the four-hour SPECT, the tumor-to-background ratio is not as good. The images when you're looking at them on their own are not as nice. But there's a very nice correlation between the volumes with a correlation of 0.995 between the four-hour and the 24-hour. Based on this information, we're actually moving to the four-hour SPECT imaging after therapy because it's much more convenient for our patients. They don't have to stay overnight.

I talk about total tumor volume, and I think we have to be very, very clear that when we use it clinically, we are not using just total tumor volume. We actually look at the CT as well. We do a diagnostic CT in our patients each time they come for their SPECT because of the heterogeneity you get, particularly in liver, and that's very hard to see on a low-dose CT. We've just moved clinically to embedding our diagnostic contrast-enhanced CT into our SPECT images.

But prior to that, we were using them, just doing them separately and looking at them together. We never don't look at the diagnostic CT at the same time that we look at the SPECT. And this is a really great case for that. This is a patient who's actually on the LuPOP trial. This is their dose one high-volume disease, PSA 248. This is dose three, nice reduction. They've got a marked reduction in total tumor volume, PSA 56, but they have a new liver lesion, that wouldn't have been visible on the low-dose CT. Definitely visible on the contrast-enhanced CT and definitely means that patient needs management of that liver disease before it progresses too much.

On the Enza-P trial, already spoken about Enza-P, but basically, Professor Narjess Ayati is managing doing the harmonization of these SPECT cameras. What we've actually done is we've done cylindrical, and IEC, and CT phantoms across Australia at all 15 sites. We've developed sensitivity factors for every single one of the cameras involved. We've stripped the filtration from each of the site's SPECT CT and we've refiltered and reanalyzed so that we have identical SPECT CT from all sites across Australia. And you admit the images just look beautiful and we have great data coming out of all of these sites. It has been a piece of work learning how to do this, but we're pretty on top of it now. If anyone needs any help harmonizing across the US, or Canada, or anything like that, we would be happy to help.

How can you actually use early response biomarkers? In Australia, unfortunately, we don't have funding for lutetium PSMA, and my patients have to pay. And that really led to desperation or innovation in terms of how we treated those patients. They couldn't afford all six doses in all cases. Really, it became those patients who really needed to continue treatment, who did continue treatment. And we used the baseline and six weeks SPECT and PSA to try and stratify our patients. It was a lutetium PSMA clinical program. We did PSA and lutetium total tumor volume at six weeks. And we divided patients into response groups; I was going to tell them their tumor dog type. Response group one, actually fluffy dogs, tumors. And these patients have a marked reduction in SPECT volume and a marked reduction in PSA at the six-week mark. As you can see with this patient, you can see it hasn't actually all gone, but it's dramatically dropped down. The trajectory in terms of the reduction of disease is marked, and that was what I would call a highly radiation-sensitive clone.

And then response group two, we call them kelpies, and this is where the SPECT volume's gone down, the PSA has gone down, but it's not marked. And you can see this in this patient here. If you were looking at this just without the quantitation, it'd be quite hard to see what's happening. With the quantitation, you can see they've gone from about 4,500 to 3,700 mils. Down, and the PSA's also down. Those patients we kept going six doses, six weekly, just as in the registration trials.

And then response group three, we call them pit bulls or rottweilers, it's a little bit different between the two. But essentially, those patients have got an increase in SPECT volume, or an increase in PSA, or both, or they have new lesions on their CT that are PSMA negative, like the patient I showed you with the CT. Those patients did like this and Jeremie's already shown you these Kaplan-Meier curves. If you're a fluffy white dog with a good response, your progression-free survival is 12 months.

Overall survival, 22 months. And those patients had a median break in treatment after dose two of six months before we started again at PSA rise. Response group two, or kelpies, had a progression-free survival of 6.2 months with an overall survival of 15 months with no break in treatment. And then the rottweilers or the pit bulls, 2.8-month progression-free survival, 11-month overall survival. We tried to move those patients to early treatment whenever we could if there wasn't another treatment available, often there wasn't. We continued treatment for palliative purposes. I can tell you those patients who had progressed on SPECT and PSA at dose two, if they did go on to get further treatment with dose three and four, we did not have a single patient who had a subsequent response to treatment. I think the dose two time point is quite good at identifying radiation-sensitive and radiation-resistant clones.

Treatment response at dose two, I believe, is a proxy marker of radiation sensitivity or resistance and does predict progression-free and overall survival. Certainly easier than any of the genetics that we've got at the moment, which is extremely complex. It appears to have equivalent prognostic value to interim PET, but we certainly need to look at that in more detail. It's definitely more cost-effective and easier to obtain. I think the identification of early progression, doing early SPECT rather than interim SPECT, gives us the opportunity to intervene early. If we are failing lutetium, can we add alpha? If we are failing lutetium, do those patients do better with cabazitaxel? What opportunities have we got that we can add to improve these patients' outcomes? I think these are all trials that we need to be doing. Multiple biomarkers are potentially better than one. At the moment, we have PSMA, we have PSA, what else can we add? And I really want to express that tumor volume is important in looking at progression with lutetium PSMA, not change in intensity, which is what I see a lot of people doing. And thank you.