Hello, everyone. My name is Dr. April Armstrong. I am professor and chief of dermatology at UCLA. And today I'm going to talk to you about practical points in psoriasis studies. This is an area that's very near and dear to my heart. I have been involved with psoriasis clinical trials for over 20 years and have really seen the evolution of the different endpoints and the various considerations in terms of what goes into these studies. So I'm very excited and happy to share these insights and pearls with you so that hopefully next time when you look at a psoriasis clinical trial, read the results, you'll have some renewed understanding of what goes into the planning of these studies and also can interpret the results with this new knowledge in mind. Okay, so the objectives of our talk today are a few. First, we're going to talk about some of the common inclusion and exclusion criteria for psoriasis studies. Then we're going to talk about considerations of outcomes and endpoints. Which outcomes and endpoints do we select? And also talk about length of time period at which you are looking at these endpoints because they do matter. And then finally, we're going to talk about some of the analysis methods that you've probably seen in more recent years, which include indirect comparisons or network meta-analyses, essentially analysis that allows us to potentially indirect compare drugs when there are no direct head-to-head clinical trials available for these medications. Okay, let's get started. So first, let's talk about some of the common inclusion and exclusion criteria for systemic medications for moderate to severe plaque psoriasis. When you look at moderate to severe plaque psoriasis clinical trials, there is a great sense of uniformity, actually, in terms of the patient population that are typically enrolled into clinical trials. In fact, when you take a look at the specific criteria, most of them haven't changed over time. And when we're talking about moderate to severe plaque psoriasis studies, when they're considering inclusion criteria, have nearly always considered a POSI score of 12 or greater, BSA score of 10% or more, and IGA of 3 or more. So these are pretty much uniform across the studies, which allows us later to actually compare across the different studies. Here are some of the important exclusion criteria that apply to most studies. Most studies will exclude patients with acute or chronic viral hepatitis B or C or HIV infection. They will exclude patients with active infection, including active TB. They may allow patients with a history of latent and treated TB into clinical trials. They will also exclude patients with any active suicidal ideation or positive suicidal behavior, even with major depression. Now the next point, malignancy, is interesting because the earlier clinical trials in moderate to severe plaque psoriasis have excluded patients with any history of malignancy, excluding non-melanoma skin cancer. However, the recent trend in the past 10 years or so have allowed patients who may have a history of malignancy, but it has been in remission for five years or greater. And I think we are seeing that because malignancy is actually quite a common comorbidity. And what we're seeing is that clinical trials are trying to enroll more real-world patients. And therefore, patients with internal malignancies who have been deemed in remission for five years or longer, the more recent trials have allowed them to be enrolled in the clinical studies. All of the clinical trials to date in moderate to severe plaque psoriasis have allowed patients with non-melanoma skin cancers to be a part of the trial. Okay, so let's talk about then topical medications and how their inclusion and exclusion criteria may be different. I just talked about moderate to severe plaque psoriasis for systemic medication trials. So all the inclusion and exclusion criteria I talked about pertain to systemic medications that are approved for moderate to severe plaque psoriasis. But let's talk about topical medications. Now, topical medications for plaque psoriasis can be used in patients with potentially any disease severity. And depending on the particular agent that's used, the particular agent that's studied could try to look at all disease severity, mild, moderate to severe, or they can look at potentially a subset, such as mild and moderate patients only. So I'll give you some examples. So examples of inclusion for topical medications, as you may suspect, for typically for these medications, there is a limit to body surface area involvement. There's a lower limit and as well as an upper limit. So for example, when we look at one of the recently approved medications, BSA involvement is on the lower end defined as 3% body surface area and the upper end as 20%. The reason they may have an upper end for topical therapies for plaque psoriasis, as you may imagine, is because when people have very severe disease, they are probably not a great candidate for topical use. However, 20% is still pretty severe. And in clinical studies, sometimes the investigators wanted to see how a topical medication may do as a monotherapy in those with higher body surface area involvement. So other examples of inclusion criteria include IGA score of 2, 3, or 4 as screening and baseline. The key thing to look out for in terms of the baseline table, actually, is look at the distribution of mild, moderate, and severe patients. You may think that a clinical trial that enrolls mild, moderate, or severe patients may be dividing those patients in terms of the population into a third, a third, and third. However, in fact, this is not the case. In fact, when we look at most topical therapies, moderate consists of the majority of the patient population. So in the example I gave, typically they would do 10% mild, 80% moderate, and 10% severe. Now, you may be asking the question, why that's the case? Well, when we're thinking about why patient population may be enriched more towards the moderate population, we typically have to tie this conversation with the endpoint that they're looking at. So typically, when we're evaluating endpoints for topical medications, the endpoint is a static endpoint of clear, almost clear, with at least a two-point reduction in the IGA score. So if you have a lot of severe patients, even if they move from IGA of four to IGA of two, they will not be considered a responder, even though they have improved quite a bit. Therefore, most clinical trials will try to, for topical therapies, will try to enrich the moderate population, meaning including more of the moderate population, because those are the patients that you're able to see the most response or movement along the endpoint that you're looking at. Okay, so perfect segue to go into the next section, which is consideration of endpoints in psoriasis clinical trials. First, let's take a look at endpoints for systemic medications in moderate to severe plaque psoriasis. And this is one example, for example, and I provide the reference at the bottom of the slide. So when we're talking about endpoints, these days, typically, there will be co-primary endpoints for systemic medications, looking at moderate to severe plaque psoriasis. Here is an example from a recently approved IL-17 medication. It's a highly efficacious medication. And therefore, they select, for example, POSI-90, so at least 90% improvement in a person's psoriasis severity as compared to baseline as measured by psoriasis area and severity index, POSI. So POSI-90 is at least 90% improvement, so 90% improvement or more at week 16 compared to baseline. What you'll notice is that in the previous biologic studies with our first generation biologics or oral therapies, oftentimes, POSI-75 is actually considered a co-primary endpoint for those clinical trials. And I think the reason for this is that first generation biologics or oral therapies to date may have slightly more muted response as compared to, let's say, later generation IL-17 or IL-23s. And therefore, POSI-75 may be a good measurement for them. But for a very highly efficacious later generation IL-17 medication, for example, POSI-90 at week 16 is now considered in this clinical trial as a co-primary endpoint. Also, clear or almost clear, IgA0 or 1, which is clear or almost clear with at least a two point improvement at week 16 is also typically considered as a co-primary endpoint. What about all the other endpoints? Now, we group those under secondary endpoints. Now, primary endpoints are especially important because they are used to determine sample size, how large one should power the study, how many patients to recruit. Secondary endpoints are also very important, clinically important endpoints. And they can look at other aspects of the response. For example, how quickly does the patient respond? So you can look at POSI-75 at week 4, what proportion of patients achieve POSI-75 at week 4. Here, I list some other endpoints, POSI-90, IgA0 or 1 at different time points than what was assessed in the co-primary endpoints. POSI-100 at different endpoints. And of course, safety assessment are almost always a part of the important list of endpoints. OK, so raising the bar. And when we especially here comparing highly efficacious medications, what you will notice that when you're comparing these extremely effective systemic medications such as biologics, later generation biologics, POSI-75 won't cut it in terms of distinguishing the performance between two highly efficacious systemic medications. So what you will see in the literature is higher bars. So POSI-90 or POSI-100. So let's take a look at this particular study. This is our first IL-23 versus IL-17 inhibitor head-to-head study. And you will notice is that what they use in terms of the endpoint is POSI-90. And the reason is because POSI-90 is a stringent endpoint and it can distinguish between two highly efficacious medications. And here we see an example of POSI-100 used as an outcome measure for evaluating differences between two highly performing biologics, one IL-17 against another IL-17. Now, if you use POSI-75, it's a lower bar. It's less able to distinguish among two highly efficacious medications. Okay, now I'm going to talk a little bit about dynamic versus static endpoints. And what do they mean? So let's first take a look at dynamic endpoints. Dynamic endpoints are endpoints that assess change. Static endpoints, as its name suggests, are looking at a immovable target. So let's take a look. So when we talk about dynamic endpoints, it's always the comparison between one time point versus another time point to see what change has happened. When we are talking about POSI-75, POSI-90, and POSI-100, as you probably know, these are talking about percent improvement. So that is, for example, at week 16, compared to baseline, if we're looking at POSI-75, if we're looking at POSI-75, has a person improved at least 75% since the baseline. And then clinical trial count of everyone who has achieved that or higher, and they report it as a proportion of patients who have improved 75 percent at week 16 compared to baseline or more. And then we have PASI 90, and then we have 100 percent improvement. So those are dynamic endpoints. You will also see change in PASI BSA from baseline as another dynamic endpoint, or change in dermatology-like quality index, a DLQI from baseline. And what's the implication of using dynamic endpoints? Well, a lot of these dynamic endpoints, and especially PASI, when we're looking at percent changing PASI, is that PASI is not a linear scale when we look at it. And so the patients with greater disease are going to more likely show greater change. So populations with greater severity are more able to show a change. So therefore, when we're talking about moderate to severe plaque psoriasis, oftentimes you want to use a dynamic change as a scale to evaluate, because you're more likely to see a change there. Now let's look at the right-hand side, static endpoints. Static endpoints are endpoints that don't move. You'll notice in the example I gave previously, the FDA also wants us to look at static endpoints, proportion of patients that achieve clear or almost clear, for example. So examples include that, and also those who achieve PASI 2. This is not 2 percent change, but PASI score of 2 or less. So this is considered a static endpoint, or BSA 1 percent or less. And why do we use static endpoints? Because static endpoints oftentimes are a bit more, I would say, interpretable, thinking about, we can all imagine what a person with clear, almost clear skin looks like, and therefore could imagine and evaluate patients, the proportion of patients that achieve that. Now, but what's the problem of using static endpoints? Now, static endpoints is especially a particular problem if you have a very severe disease population. Why? Because if you set the static endpoints at very stringent or low, let's say clear or almost clear, if you have a population that's mostly severe or very severe, that's a huge movement to move from severe to very severe, all the way to the endpoint of a very stringent of static endpoint. Okay, so hopefully I talked about the dynamic and static endpoint, and that gave you some insights into these two different types of endpoints. Next, I'm going to talk about time frame. When we assess things is very, very important. Now, in plaques rises trials, typically the placebo control period is three or four months, and four months will generally give you a higher response rate, because that's typically around the time that the medication is still continued to effect change, and you will see an increase in the response rate. So the more recent studies use four months rather than three months, and also there's a trend in looking at earlier points, for example, as a secondary measure, not as a primary endpoint, but a secondary endpoint, looking at earlier time points, four weeks, for example, and that can evaluate speed of onset. Okay, so let's take a look at this one, and then it's an example of why time points matter. So this is Ixikizumab versus Gacelkumab, two very good medications, and we're using a very stringent endpoint, POSI 100, to compare the two. So for this particular trial, the primary endpoint is set at week 12, and what you see is that by week 12, Ixikizumab is superior to Gacelkumab. However, if you extend out to time frame, if you look at week 24, you will see that the two medications are very similar in terms of their response. So what you learn from this is that Ixikizumab is faster in onset in this particular case, as shown in this clinical trial. You also learn that when you evaluate the endpoint is important because you have a different story and different results depending on when you evaluate the endpoint. You saw this earlier in terms of Gacelkumab versus Zikukinumab. Again, when you evaluate endpoint is important. This study, the endpoint is at one year where Ixikizumab is superior to Zikukinumab. However, if you evaluate the two medications at 16 weeks, they look very similar. Okay, one of the things that we are faced with in terms of challenge is when we try to choose a systemic medication, we may not have all the head-to-head trial information to inform us of our decision. And this is where methods such as indirect comparisons can come into play and help us in terms of clinical decision-making as another tool. So let's take a look at this. So indirect comparisons is when no head-to-head study is available. However, you may have drug A versus drug B. You may have another trial, drug C versus drug B. So what you'll notice is that they have a common anchor. And in most RISA studies, that's a placebo. That will actually, because they have a common anchor, granted they're in different populations, you can potentially make some indirect comparisons between A and C, but knowing what are some of the caveats in the indirect comparisons. So there are different types of indirect comparisons that you can make. You can compare aggregate data of drug A versus drug B in what's considered network meta-analysis. You can have individual level data from drug A indirectly compared to aggregate data from drug B, which is a type of mix of individual data, patient data, and summary data level comparison. This type of match-adjusted indirect comparison oftentimes. And then very rarely you'll have individual patient data from one trial of one drug versus individual patient data from another medication. The reason is because typically manufacturers have their own data and ownership. And so the third type is actually very unlikely to see in publications. All right. So let's take a look at the overview of this and especially the network meta-analysis you'll see quite a bit. So network meta-analysis, as I described earlier, they can take advantage of the common comparator of the placebo and then indirectly compare it to medications. And network meta-analysis has been used by Institute for Clinical and Economic Review for evaluating different medications simultaneously. And also SPDEA considers this type of data as an adjunct to the head-to-head clinical trials as well. So let's take a look at network meta-analysis often expressed as number needed to treat or the number of patients that will need to be treated with the intervention as compared to oftentimes the placebo to gain one additional responder. So if your drug A has number needed to treat to gain one additional responder as four patients versus drug B number needed to treat to gain one additional responder as two patients, then a therapy is more effective if it has a small number needed to treat to gain one additional responder. So let's take a look at this. One thing you will notice that in this NMA, network meta-analysis, this is a bit older one. So it matters if you're using POSI-75 as response as the endpoint to look at or a different, more stringent endpoint. So POSI-75, my goodness. So take a look. For most of our biologic medications, you only need to treat a little over one person to have that response, which means that nearly every person that comes to the door, you're able to probably get the POSI-75 response for most of our biologics. You will see that the confidence interval overlap a lot among the different biologics when we're looking at POSI-75. Now I advanced the slide and here you're looking at POSI-100. So here, number needed to treat, you need to treat a bit more patients. So every 2.5 patients, then you'll get a POSI-100 response with some of our highly performing medications when we're talking about POSI-100. You will also notice there are a bit more separation now between the different medications and you can see a bit of separation in the performance among the medications. You may also see a different type of analysis that's done where you look at area under the curve. So not only look at one endpoint at a one particular time, but also look at the totality of performance across a certain time period. We call this area under the curve. This particular analysis looked at area under the curve and have shown, for example, there's still a lot of overlap, I will say, among the top performing medications. But overall, this is helpful because it takes into account the totality of experience, for example, during this first 16 weeks rather than just one point in time at 16 weeks. These are examples of Cochrane Networkman Analysis Review. Here is looking at not only the response at the x-axis, but also taking into account safety and tolerability dimensions on the y-axis. It can put the different medications on different parts of this graph. So one of the things that we've faced in our field is looking at different network meta-analyses and everyone come up with the different results. How is that possible? Why is that the case? Well, a few things. Most network meta-analyses are using publicly available published information. What that means is that if I do a network meta-analysis and you do a network meta-analysis using the same studies and same method, we should get the exact same result. So the debate really should be about selection of studies and the methods that one uses to conduct the study. So when we're looking at why different NMAs have different findings, well, maybe the studies included are different. Some studies may have included more studies versus others. The outcome that's deemed most important by the investigators may be different. This is actually really important because the results could look very similar if you choose POSI-75, but if you choose POSI-100, they may look quite different. Time period. This is very, very key. What time period are you looking at? And I showed you some examples of how your conclusions can be vastly different depending on the time period that you're looking at. And then finally, analytical methods. There are different ways in which data can be analyzed in network meta-analysis, and there may be differences among the different studies. And then the debate among which method could be best is often something that I think is a worthwhile debate when we look at the results and if the results show different things. I know I went over a lot of information. I just wanted to highlight a few takeaways for the audience. First of all, the inclusion criteria is fairly uniform across the moderate to severe plaque psoriasis studies, and they're important because they not only define the patient population, but they can also define the likelihood effect size, so the difference between the placebo and inactive arms. The exclusion criteria impact mostly safety data. And when we talk about clinical trials, especially I would say phase two or phase three studies, pre-approval, the exclusion criteria there do select a patient population probably less likely to have a lot of comorbidities different from real-world studies, most of the time post-approval, where they will have a more heterogeneous patient population. Stringent endpoints are needed to distinguish among highly performing medications, and the choice of endpoints are closely tied to the time of assessment. And then finally, indirect comparisons can be helpful in terms of offering insights when we do not have head-to-head comparisons. And with that, I want to thank you very much for your attention for the past half an hour, and I hope that next time when you pick up a journal and read an article in plaque psoriasis studies, that these different key points will come to your mind and help inform you with your data interpretation. Thank you very much, and see you next time.

psoriasis clinical trials

inclusion criteria

exclusion criteria

endpoints

systemic medications

network meta-analysis