Ascelia Pharma AB (publ) (ACE) Earnings Call Transcript & Summary

Good morning, everyone. Thank you for joining us this early morning to get an update on the press release that we sent out last night. First, I'll spend about 5 minutes going through the key messages of the press release and our perspectives on this, and then we'll open up to questions. You can ask questions by writing them in the window to the right in the viewer to access this event. So as usual, I'll be making a number of forward-looking statements. Before going into the content of the press release, I'll just do a quick reminder to everybody about the structure of the SPARKLE clinical study. So the 85 patients who have completed SPARKLE will be -- are evaluated by 3 independent radiologists, reader 1, 2 and 3. And for the primary endpoint, they are evaluating -- providing a score for the border delineation and the lesion contrast. For a reader to be successful and reach the endpoint of the study, they need to score statistically significant improvement on both the border delineation and the lesion contrast. For the study to be successful, 2 out of the 3 readers need to be successful. Now I'll come to the press release. So in the press release, we communicated that a reevaluation is required to reach a conclusion. I'll talk a bit more and explain why that is. So it's a well-known complication that radiologists are, like human beings, will rate an image in a somewhat different way. That is why a study design has 2 out of 3 readers to be successful because that is understood. The surprising issue that we have run into with SPARKLE and that we announced yesterday is high level of intra-reader variability. So what is that? Intra-reader availability is a measure of the difference in evaluation of an image from the first time to the second time a reader sees this. For example, if intra-reader variability is high, the radiologist may score an image with a low score the first time. When they see it the next time, they will score the image very high, or the other way around. But in reality, that means that the high level of inconsistency and it's -- you cannot conclude on the read and discourse from such a reader. This issue is not unique to Orviglance, unfortunately, and this study. It's something that is well known to the FDA, and that's why the FDA has provided guidance to the industry in terms of managing this issue. We have adhered to this guidance from the FDA and our dialogue with the FDA. And therefore, in the design of SPARKLE, we took a subset, a specific number of patients from the SPARKLE study, whose images would go into the database twice. That means for these images, the radiologists would all rate these images twice. When we analyzed the data, we saw for some of the readers that the intra-reader variability was high. That means they scored low and high or high or low when they looked at these images that were in the database twice. This is -- as mentioned, this makes it impossible to make a conclusion on the efficacy. And this is very unfortunate and it's very surprising. We will look into to understand better why that happens so we can take every precaution possible to make sure that does not happen in the reevaluation. So for us to get conclusive results from the SPARKLE study, we need to do the reevaluation. We need to do it as quickly as possible, and we will align with the FDA during the process to make sure we get their input as well. But we don't know the exact time line for when the reevaluation is completed. All our efforts and activities in Ascelia Pharma will be focused on the reevaluation. It also means that we will take cost-saving initiatives to preserve the cash to extend our runway as much as possible. This is completely, you could say, new situation. So we don't have any planned time line or know the exact restructuring and cost-saving initiatives. So we expect to have that ready to announce mid-September. Just to finish up before going into the questions. The intra-reader variability issue doesn't change our confidence in Orviglance. It doesn't change the unmet medical need. The unmet medical need Orviglance is addressing is the same today as it was yesterday and will be the same tomorrow. It doesn't change the addressable market opportunity of $800 million globally. The issue doesn't change the consistent positive results we've seen in Phase I and Phase II. It doesn't change the positive anecdotal feedback we got from the investigators in SPARKLE. But it does change the time line. As frustrating and annoying that, that is, it does change the time line, and we'll come back with an updated time line in mid-September. What we also noted in the press release yesterday was that we saw the common -- the adverse effects that we saw on SPARKLE are in line with the previous studies that we have completed. We think that is very good news, that in this more sick patient population that we studied previously, we see a consistent pattern. We have all the images for the reading evaluation in the database. We need to align with the FDA. We need to identify new readers and do a reading process, including training of the readers, and that is what we are focused on at this point. We continue to be very optimistic about Orviglance and the potential and how it can help patients. That was a brief explanation and summary of our press release. I'd like to open up to questions now. And I'd also like to invite our Chief Scientific Officer, Andreas Norlin, to this webcast. Morning, Andreas.

Good morning, yes.

So we have the first question here. Here's the question. How could this happen? Is this a common issue with contrast agent clinical studies?

Well, the answer is it's not unheard of because as you explained earlier, Magnus, the images are assessed by readers, radiologists. And they are of course experienced, they are trained for the task, but they are human beings. So there might be some variation, and that's quite normal. And as per the guidance, we have implemented this -- the test to measure this variability. And when that came out, we saw that it was larger or bigger variability than what we think is acceptable. And it's not unheard of. It has happened in other studies before. And that's the reason for why this guidance exists, on why doing this. It's annoying but this is the fact.

There's another question. How long will the reevaluation take and how will it impact the overall time line? We don't know exactly how long it will be, and we also appreciate that people want to have an understanding of what this time line is. We will communicate sort of detailed time line based on our planning -- detailed planning in mid-September. To give some kind of, you could say, ballpark understanding of this, our -- we want to do it as fast as possible. Our current expectation is that it will certainly be less than a year but probably more than 6 months. That's what we expect. But we will come back with a detailed time line in September. Here's another question. Please remind us, how many readers was included in the study? Did the smaller study size of the study contribute to the risk? Andreas, maybe you can answer that one.

Yes, for sure. Yes, there were 3 readers. They were all independent, doing their own assessments on the images. And no, the size of the study is not related to this issue. It's entirely related to the reader and how they perform in their job.

Yes. Another question here. Are there any measures that can be taken in order to reduce the risk for the same thing happening again with a new group of readers?

Well, I think there's a lot to learn from this, and we are doing that analysis as we speak or that will be very high up in our focus when planning for the reread. There are certainly many things we can do to minimize this risk. I will not -- I cannot say exactly what the root cause was at this time. So...

But I think it's important also to note that this is not -- we knew this was a potential issue. We have not seen it previously.

Yes.

So we have been diligent in selecting experienced radiologists. They have reader training. They are, I would say, training processes in place. So we've been diligent in understanding and taking every step on the way. But we still see this outcome, which is what -- as Andreas said, we need to look further into this to see how we can improve it to certainly avoid this again. Another question here. Did you discuss this potential issue with the FDA ahead of the Phase III design?

That was part of the discussions because the protocol was discussed and presented to the FDA. And again, it is part of the guidance to industry from the FDA. So yes, it has been addressed in the interactions with FDA.

Yes. And just to put in a bit more flavor on that, the reason why this is in the guidance, why the FDA is having that conversation with drug development companies is, of course, if you have high level of intra-reader variability, you cannot firmly make a conclusion on the efficacy. And then it would not be able to go through a regulatory process. That's why this quality check is incorporated into the study design and why we can deal with this issue now and not in a review process. Another question is, did the smaller study size of the study contribute to this risk?

No. That's our firm understanding, that it has nothing to do with the study size.

It's an issue related to radiologists having too high level of variability, too low consistency. So a larger study would not have any impact.

Yes, exactly.

There's a question here. Can you elaborate on how this does not change your confidence? So the confidence that we have in Orviglance is based on the clinical studies that we have conducted to date, about the available literature showing that hepatobiliary phase imaging is very important for liver imaging, the anecdotal positive feedback that we have received from the investigators in the SPARKLE clinical study. And that does not change our, you could say, belief and confidence in Orviglance that it should be successful. We have no reason to believe that Orviglance should not be as efficacious as we thought before getting this kind of data readout. A question here. Will your current financing be sufficient to reach a readout from the revaluation? The answer is that, that is our ambition but we do not know yet. So as mentioned, we will be looking at cost-saving initiatives in the company to make sure that we have funding until we get the results from the reevaluation. That's our ambition, and we will communicate once we have the plans in place and we understand the cost implications in mid-September. There's a question here. Is this issue caused by a too low efficacy?

We cannot conclude. That's the whole issue we have here. Because of the high variability, we cannot conclude on the data. So unfortunately, we cannot answer that question.

Yes. And you could say, when it's variable, you don't know whether the evaluation -- when the variability is very high, you don't know whether it's actually a true high score or a true low score. It becomes more like guesswork with high level of intra-reader variability. Again, super -- very unfortunate, but that's how it is. And we want to communicate in a data-driven way. That's how you want to have a process with the medical community and the regulators, and that's what we do. A question here. If you remove the radiologist with high intra-reader variability, how does the data set look like then in terms of trends? Is there any sign of efficacy?

I think, again, we need to conclude -- we cannot conclude and we refrain from concluding because we need to have the full set of data in order to evaluate the study. It would be very dangerous to go out and speculate on what 1 reader or 2 readers might mean. So we'll come back when we have done the reevaluation.

And the follow-up question here is, is there any -- are there any signs of efficacy? I would turn it around. To me, when I look at the data and the -- with a very substantial significant limitations on what it can conclude, to me, there's nothing in the data that changes my belief in Orviglance having the potential to show a very strong significant improvement. But it's too difficult to look at the -- or conclude anything from the data. But there's nothing that disproves that we can have very strong efficacy for Orviglance. I think this was the final question that we have received. Oh, here's one question that comes up here. Was the variability present in both enhanced and unenhanced images?

So the evaluation takes all of the images into account. So the question is, yes, across -- when comparing reads from the first read and the second read, you can see it all over. And that's what's adding to this high uncertainty.

And it was both in the unenhanced and it was in the enhanced.

Yes, exactly. Yes.

So there's a question here. Can you be certain that FDA will not require additional studies or more patients to be included?

Yes, who knows what FDA thinks.

Yes. I think coming back to the guidance to the industry. So the guidance to the industry that FDA has provided that we have discussed with the FDA is to implement a subset of patients whose images will be evaluated twice to evaluate the intra-reader variability. So this has nothing to do with the integrity of the study, the images that are in the database. This is a quality check for the reader consistency. And what we want to do is analyze the data, having readers analyze the images, scoring the images and having a statistically sound analysis of that. We think that is also the interest of the FDA. They want to have low reader intra-variability, and that is what we intend to obtain with the reevaluation that we are starting the process for now. So we don't think that we need another clinical study. We think if we had to go with a clinical study, then we would not -- then we basically would be throwing out this clinical study and not concluding anything. We don't think that is very meaningful, not from an ethical perspective where the patients have been involved in the trial. We need to conclude on the results from the trial, and that's what we want to do with the reevaluation. A question here. Could it be an issue of image quality?

There is no signs of that. The readers assess the overall readability and quality of the images, and we didn't see anything there that points to that, that would be the issue.

Another follow-up question here. What type of equipment has been used in terms of MRI? And here, I think they're thinking about tesla?

Yes. So the protocol allows both 1.5 and 3 tesla machines. And most of the images have been conducted with 1.5, yes.

Another question here. Could this be related to the slightly different patient population, now also including suspected tumors?

No. I would say suspected tumors is, of course, allowed by the protocol. But it has not affected the overall assessment, to our understanding. We could not see any signs of that when analyzing the data.

Is the high level of variability related to 1 or several readers?

Several readers.

I think that was the final question. Thank you so much for joining us this morning, and thank you very much for the questions. This is a surprising and regrettable outcome from the read. We're very surprised. But we remain confident in Orviglance, and we're already starting to take steps for getting the reevaluation underway and coming into a dialogue with the FDA. So we will keep you updated and certainly no later than mid-September with an updated time line and financial implications. So thank you for joining us.