Mersana Therapeutics, Inc. (0M40.F) Earnings Call Transcript & Summary

Good morning, and welcome to the Mersana Therapeutics conference call. [Operator Instructions] I would now like to turn the call over to Sarah Carmody, Executive Director, Investor Relations and Corporate Communications. Please proceed.

Good morning. Welcome to Mersana's conference call and webcast to discuss the selection of the NaPi2b biomarker cutoff and our commercial diagnostic development path. A replay of today's call will be available on the Investors and Media section of our website. After our prepared remarks, we will open the call for Q&A. Before we begin, I'd like to mention that our call will contain forward-looking statements within the meaning of federal securities laws. These are not statements of historical facts and are based on management's beliefs and assumptions and on information currently available. They are subject to risks and uncertainties that could cause the actual results and the implementation of the company's plans to vary materially, including the risk that preclinical testing or early clinical results may not be predictive of the results or success of our ongoing or later preclinical or clinical studies; that the identification, development and testing of the company's product candidates and new platforms will take longer and/or cost more than planned; and that our clinical studies may not be initiated or completed on schedule, if at all. These risks are discussed in the company's SEC filings, including, without limitation, the company's annual report on Form 10-K filed on February 26, 2021, and subsequent filings. In addition, while we expect that the COVID-19 pandemic might adversely affect the company's preclinical and clinical development efforts, business operations and financial results, the extent of the impact on the company's operations and the value of end market for the company's common stock will depend on future developments that are highly uncertain and cannot be predicted with confidence at this time, such as the ultimate duration of the pandemic, travel restrictions, quarantine, physical distancing and business closure requirements in the U.S. and in other countries, and the effectiveness of actions taken globally to contain and treat the disease. Except as required by law, the company assumes no obligation to update these forward-looking statements publicly, even if new information becomes available in the future. With that, I'll turn the call over to Anna Protopapas, Mersana's President and Chief Executive Officer.

Thank you, Sarah. Good morning, everyone, and welcome to today's call where we will outline our expected commercial diagnostic development path, including the NaPi2b biomarker cutoff as part of the ongoing UPLIFT registration strategy with the continued evaluation of upifitamab rilsodotin, or UpRi for short, in platinum-resistant ovarian cancer. The recent initiation of UPLIFT is an important milestone in our efforts to bring UpRi to heavily pretreated ovarian cancer patients who have an acute need for new therapeutic options. The initiation of UPLIFT is also an important milestone for Mersana, as it represents our first ever registration-enabling strategy as we strive to achieve our vision of bringing life-changing ADCs to patients in need. As a reminder, the design of UPLIFT is informed by both FDA feedback and the proof-of-concept we demonstrated in the expansion cohort where UpRi showed clinically meaningful activity in heavily pretreated ovarian cancer patients with an objective response rate of around 30%, including confirmed complete responses in patients who have failed prior bevacizumab and PARP inhibitors. We demonstrated this activity without the severe neutropenia, peripheral neuropathy or ocular toxicity that limit other ADC platforms, and we observed a biomarker response relationship on which we will elaborate today. The UPLIFT cohort is implemented as an amendment to the Phase I expansion study, allowing for significant operational efficiencies and leveraging our current momentum in patient enrollment. UPLIFT is designed to enroll a broader patient population than other studies in this space, consistent with where we observed activity in the expansion cohort. More prior lines, more flexible inclusion of prior bevacizumab treatment consistent with the bevacizumab label and no exclusion for baseline peripheral neuropathy. UPLIFT also offers 2 shots on goal with the primary endpoint in the high NaPi2b population and a secondary endpoint in the overall population, allowing us to fully evaluate the role of the biomarker in enriching for patient outcome. We believe that the data to date for the -- from the expansion cohort, the robustness of our diagnostic development plan we have undertaken and will describe here today as well as the design-only UPLIFT gives us a high probability of success. To that end, today's presentation is an informational session intended to outline the very systematic approach we have taken to the diagnostic development. With me today are Arvin Yang, our Chief Medical Officer; and Becky Mosher, our Vice President of Translational Medicine, who will walk you through our diagnostic development road map, starting with the research stage and clinical stage evaluation and the path to validation of a robust, predictive and reproducible commercial diagnostic. Note that the data that Arvin and Becky will present represent years of work because the development plan for the diagnostic started with UpRi, then known as XMT-1536, was still at the research stage and advanced in tandem with UpRi's critical development. Today's presentation will cover the following topics. First, Arvin will provide a brief description of what we believe to be the characteristics of an optimal IHC assay, namely its robustness, predictiveness and reproducibility, and then we will review the progress to date. Following that, Becky will review the scoring and cut-off decisions we've made based on the totality of the data, including preclinical, translational and clinical data. And finally, Arvin will discuss the path forward to validate and commercialize the diagnostic. With that, I will now turn the call over to Arvin.

Thank you, Anna. Let me first start on Slide 4 by describing the 3 key factors on which we focused in developing an optimal diagnostic assay; its robustness, its predictive power and its reproducibility. Robustness refers to the ability of an assay to distinguish low, medium and high expressors of our wide dynamic range. This requires staining that spans a broad spectrum of color. Illustrated on the bottom left panel of this slide is an assay that has narrow range. It's stain dark, which makes it difficult to discern between the low, the medium and the high expressors. Recall that the lifastuzumab assay scored 93% of all ovarian cancer patients as medium or high. In the bar above, you can see an assay with a broader dynamic range or the gradient of color, which therefore allows the pathologist to discern differences in expression level. This is a critical aspect of an assay. And as Becky will explain, we optimized this early in our development through extensive preclinical work using patient-derived xenografts and tissue bank samples. A second key attribute of an optimal assay is its predictiveness. This refers to the ability of an assay to enrich outcomes. We've demonstrated this in our expansion cohort data where higher objective response rates are seen in the higher NaPi2b patients relative to those observed in the lower NaPi2b patients. And finally, reproducibly, refers to the ability of a commercial assay to be reliably reproduced in any lab around the world. In other words, the assay results should be consistent irrespective of where the test is run. We will touch on each one of these key areas in this presentation. As Anna mentioned, the development of our diagnostic assay started several years ago when the UpRi program was still at the research stage and was advanced in tandem with the clinical development of UpRi. This slide shows a flow diagram describing our road map for the development of a robust, predictive and reproducible commercial diagnostic. A successful diagnostic requires 3 components: the assay, the scoring methodology and the cutoff. We've been evaluating these 3 components at the research stage and in the clinical trial to date. Now the next stage is to select and then validate these 3 components in UPLIFT, as we will describe here today. We started with a research assay, a manual bench-top assay that we fine-tuned before converting to a GLP qualified clinical research lab assay. Our clinical research lab was QualTek/DLS, an experienced provider of such clinical research assays. We use this assay to collect patient data throughout our dose escalation and expansion studies to understand the relationship between NaPi2b expression and outcomes. We've disclosed these data in terms of H-score. But in parallel, we've also evaluated other scoring methods, which we will describe in the subsequent slides today. This work allowed us to arrive at our proposed clinical trial biomarker cutoff and scoring methodology. Today, we will discuss next steps, which will include taking our proposed cutoff and applying it in the commercial assay we are developing through our diagnostic development partner, Leica Biosystems, to make a single robust, reproducible assay, which will allow us to seek to validate our proposed diagnostic in the UPLIFT study. I will now turn it over to Becky.

Thank you, Arvin. I'm very pleased to join today's call to help describe the extensive work we have done to date with the goal of developing a robust, predictive and reproducible commercial diagnostic. Let me start with Slide 7 and briefly review the composition of an immunohistochemistry assay, or what we refer to as an IHC assay, which includes a tissue sample, a platform, a stain and a scoring method. The middle panel shows an example of ovarian cancer tissue stained with the clinical research assay. The blue hematoxylin counterstain helps you locate the regions of interest in the sample, for example, the tumor cells. The brown DAB shows you the presence of the target. The pathologist reading the assay will determine how much of the brown is present in the tumor areas by looking at the percentage of cells that are light, medium or dark brown and can then combine the percentage and intensity values into a score. Now there are various ways to score the combination of percentage and intensity. For example, H-score or by tumor proportion score, or TPS. Here on Slide 8, you see 3 graphical representations of tumors with different proportions of light, medium and dark brown cells, which are typically scored as 1+, 2+ or 3+, respectively. Using the same counting methods, we can arrive at H-scores or TPS scores by applying the equation shown. The H-score method weighs the darker brown cells more than the medium or lighter brown cells. The TPS is a straight percentage of all staining cells at any intensity of brown. As a result, the H-score will range from 0 to 300, which is percentage times an intensity factor of 0 to 3. And the TPS, which is a simple percentage, will range from 0 to 100. In the example on the slide, all of the tumors have an H-score of around 100 but the TPS varies. For purposes of illustration, we have also provided how these tumors would be scored if a PS2 method was used, which only counts cells that have an intensity of 2+ or 3+. We have worked to select a scoring method that reflects what we see in our clinical data and to choose the method that we believe will be most tractable when this assay is used in the real-world and in different lab environments, which I will discuss on Slide 9. Here, we are showing 3 different scoring methods. I'll begin with H-score. In performing an H-score, the pathologist distinguishes levels of brown, 1+, 2+ and 3+ or light, medium and dark, and the associated percent of cells with each intensity. The H-score provides more weight depending on the intensity of stain and can range from 0 to 300. To determine an H-score, the pathologist must distinguish between blue, beige, tan and brown. The H-score has the most information in terms of percentage of cells at any given intensity, but is most complex to score consistently and may be best suited for the clinical research setting and centralized reading. It is useful for information gathering, but will not be the most straightforward for commercial use. Now contrast this with the tumor proportion score, TPS, where pathologist must distinguish between brown of any intensity and blue. The pathologist must make only a single distinction, is it brown or is it blue, making the exercise more straightforward. Lastly, as we mentioned earlier, there are hybrid methods such as PS2. This method asks the pathologist to find all 2+ and 3+ cells or medium and dark brown cells, which would be positive, and distinguish them from the 1+ or light brown cells, which would be called negative, as were the blue cells. This is the equivalent of distinguishing tan from beige. Moving to Slide 10, it is important to remember that IHC companion diagnostic assays are common, and methods incorporating intensity and proportion have been incorporated into approved companion diagnostics, although the proportion used for the cutoff may be dependent on the tumor type or line of therapy. For example, TPS is used in the PD-L1 assay. HER2/neu was an example of an assay where intensity is taken into account. However, since the HER2/neu assay is really an assay for gene amplification, the evaluation may be more straightforward in this application. Now I'll turn to Slide 11 and discuss the concept of robustness for an optimal diagnostic assay. The robustness refers to the consistent ability of an assay to distinguish low, medium and high expressors over a wide dynamic range. Robustness allows pathologist to consistently make the distinction between higher and lower levels of expression. The lower left graphic shows a spectrum from blue to brown, but the brown is very saturated, making it harder to discern between higher and lower levels of expression, which may prevent the correlation of expression and response. In the graphic above, there was a broader range in the intensity of staining, so it is easier to see the distinction and shape of brown. This allows for clearer distinctions between expression levels and is a great starting point to make those fine distinctions and where to cut the data. I will now show you how we employed an assay to try to get to a place where we could discern differences as a first step. These differences could be related back to outcome using preclinical models called a mouse clinical trial. The mouse clinical trial represented on Slide 12 tested the efficacy of UpRi in multiple patient-derived tumor models. The waterfall plot shows best response in a series of 20 human primary xenografts of ovarian cancer. The in vivo portion of the study was run not knowing the expression level of NaPi2b in advance. Once the response in these models was known, xenograft material from this study was tested by IHC to develop an assay that would most broadly discern responsiveness from nonresponsiveness. This helped us calibrate our assay prior to employing it in the clinical research lab for use in our clinical trials. Not shown here is additional work we performed on patient tissue banks to further validate the broad dynamic range of our assay. On Slide 13, in the left hand panel, we are showing you the H-scores determined by evaluating tissue samples from each ovarian cancer patient enrolled in the dose escalation and expansion portions of our study, where each circle represents an individual patient. You can see that we have an assay that can discern the entire range of expression in this population. Note that few people have 0 expression, which is consistent with NaPi2b being a lineage marker for ovarian cancer. The right-hand panel shows the same patients using the components of H-score, but recalculated as a TPS score. Again, you can see that we get the full range from 0 to 100, so we do not lose the ability to capture the full dynamic range of expression. H-score and TPS are related because they are determined from the same underlying data, proportion and intensity, just calculated in a different way. Here on Slide 14, I'm showing you the same data as was discussed on the prior slide, now graphed as TPS on the Y-axis and H-score on the X-axis. We have drawn a vertical line at an H-score of 110, which is the empiric value that we've determined to be the lowest level of expression at which we observed response in dose escalation. When we draw the line at 110, we see 68% of samples appear above the line. The second horizontal line indicates TPS scores of 75 and above. And here, you see that this method captures 62% of patients, very similar to the proportion of population captured with H-score. We also capture the majority of the same patients using either method, which makes sense because these values are derived from the same numbers. We've confirmed the prevalence with characterization and additional tissues for tissue banks. I will now turn the presentation back over to Arvin to discuss the predictability and reproducibility aspects of an optimal diagnostic assay and to share with you our proposed cutoff for UPLIFT.

Thank you, Becky. Now we will move on to the predictability of the assay. The primary purpose of a diagnostic is to enrich for positive outcomes in patients. We will show you how using a TPS of 75 and above allows us to identify an enriched population with the same predictive value as the H-score cutoff we have used to bifurcate our data in public disclosures to date. Slide 16 shows the efficacy data from the 47 evaluable ovarian cancer patients treated with UpRi in our expansion cohort, evaluated by H-score as of the December 3, 2020, cutoff date. We see in the overall population, the objective response rate was 28%, but in the higher NaPi2b population, there's an enrichment with a 32% objective response rate, including both confirmed complete responses, relative to the lower expression at 15% objective response rate, indicating the potential goal of the biomarker and enriching for overall response. Turning to Slide 17. Here, you see that when we looked at the same patients, but evaluated by the TPS methodology, which is the proportion of cells staying positive regardless of intensity, we see that the TPS methodology also enriches for response, similar to the H-score. Based on the 75% cutoff, the higher NaPi2b patients have an enriched objective response rate of 39%, relative to the lower NaPi2b patients, where we saw 11% objective response rates, results that are comparable to the H-score results that we have demonstrated to date. TPS of greater than or equal to 75 is the predefined threshold that we will use to define our higher NaPi2b population in the UPLIFT study. So why is TPS equally as predicted to H-score and enriching for response? We hypothesized that based on our ADC technology, having a higher DAR, each antibody internalization delivers more payload and the intensity of expression may be less important. So using a TPS, which assesses the proportion of positive cells, regardless of intensity, is equally predictive of an enriched response. Meaning, we may not need a 3+ intensity to bind sufficient antibody and a 1+ may be sufficient to have clinical impact. This is further supported by the controlled bystander effect built into our Dolaflexin platform, which impacts a broader range of cells that are nearby, and then becomes locked in and does not appear to diffuse more broadly to cause toxicity seen with other ADC technologies. We don't require that 100% of cells expressed NaPi2b to have clinical effect. Our data to date indicates that a TPS of 75% is sufficient to show an enriched response. Finally, on Slide 19, I will discuss the last component of an optimized diagnostic, ensuring that the assay can be put on a platform that can be run reproducibly anywhere in the world. These are the aspects that make an assay run reliably anywhere. One, clear guidelines on how the assay will be read; two, an IHC platform that can be run in a variety of settings, such as a hospital lab or a central lab; and three, an assay that produces results that can be interpreted consistently regardless of the lab. Slide 20 recaps our decision to move to TPS from H-score, a more straightforward methodology that calculates the proportion of any cell staying positive. We've shown that TPS at the prespecified 75% cutoff has approximately the same predictive ability as H-score in enriching for benefit. As I mentioned earlier, Leica Biosystems has been working closely with us for several years to develop our commercial diagnostic assay. Leica Biosystems is a worldwide supplier of equipment for IHC assays, including instrumentation and reagents. They've been working to prepare the assay that will be tested in the UPLIFT study. Slide 22 summarizes the design of our UPLIFT, single-arm registrational strategy in platinum-resistant ovarian cancer patients. As a reminder, the design of UPLIFT enrolls all patients regardless of NaPi2b status, as long as they fit the inclusion criteria. The study allows for 2 shots on goal with the primary endpoint of objective response rate in the higher NaPi2b patients and the secondary endpoint of objective response rate in the overall population. Based upon the prespecified cutoff of TPS 75, patients will be bifurcated into a high or low NaPi2b population after they've come on study. And enrollment for the primary endpoint is expected to be reached, but we enroll approximately 100 higher NaPi2b patients. This is a great opportunity in a prospectively defined retrospective analysis for us to validate the biomarker cutoff as well as the proposed commercial assay, either as a complementary or a companion diagnostic. In summary, we set out to establish a robust, predictive and reproducible diagnostic assay. We showed you that we established a clinical research assay with a robust dynamic range that allowed us to see the full spectrum of expression and, therefore, to be able to enrich a population more likely to derive benefit. We had shown previously that an H-score of 110 will enrich for a response, but now show that TPS also predictively selects just as well and can be operationally more straightforward to perform. Finally, to support commercialization, we are working with Leica Biosystems, who will deploy the assay using the TPS methodology as a reproducible assay that we will be able to run in hospitals and reference labs. The UPLIFT study will be used to validate the predictiveness of the proposed commercial assay. With that, I'll turn it back to Anna.

Thank you, Arvin and Becky. As you can see, a tremendous amount of work has gone into developing what we believe is an optimal commercial diagnostic to identify ovarian cancer patients most likely to benefit from UpRi. Before we turn the call over to Q&A, I'd like to remind you of our anticipated goals and milestones for this year, as shown here on Slide 24. We have another busy but potentially transformational year ahead of us. A year from now, we could have UPLIFT substantially enrolled, an important step towards bringing UpRi to late-stage ovarian cancer patients in need. We also expect to have UPGRADE, the combination dose escalation umbrella study underway, allowing the potential to bring UpRi to patients in earlier lines of therapy, an opportunity to expand the number of patients that could benefit from UpRi. Additionally, a year from now, we could have significantly advanced the pipeline. In the second half of the year, we anticipate having the clinical data necessary from the ongoing UpRi lung expansion cohort, the XMT-1592 dose escalation study, to better understand the potential of targeting NaPi2b to treat lung adenocarcinoma patients and make decisions as to which molecule we could take forward. In addition, we're advancing 1660, our first-in-class Dolasynthen ADC, and XMT-2056, our first Immunosynthen ADC, through IND-enabling studies and plan for both molecules to be in the clinic a year from now. We look forward to updating you on our progress in building UpRi as a foundational medicine in the treatment of ovarian cancer and building out a robust and maturing pipeline to demonstrate our position as leaders in ADC innovation. With that, I will turn the call over to the operator for Q&A.

[Operator Instructions] Our first question will come from the line of Jonathan Chang from SVB Leerink.

First question, how does your biomarker strategy in diagnostic test compare to how other companies are measuring their targets for ADC programs and platinum-resistant ovarian cancer?

Jonathan, this is Arvin. I can take that. So let me first start by saying that we've had a comprehensive plan for several years, really with the objective of defining this robust, predictive and reproducible assay that ultimately can be a commercial assay when validated through our registration strategy. Remember, we've been collecting this TPS data right alongside with our H-score data. And it's just a matter of actually calculating that data differently. And so we do think actually this transparent disclosure in regards to our comprehensive plan, the stepwise approach is a particular differentiator. Now specifically in regards to other assays and so forth, one thing I'll point out in relationship to some of the assays that have been described previously is, obviously, the PS2 assay, when utilized for different targets, would have a different antibody as well as a different platform and so forth. And as a reminder, the PS2, it sort of sits between an H-score and a TPS approach, whereby you have to determine 2 different shades of the same color, tan versus gray-ish in order to determine a positive or negative. In contrast, for the TPS, we're looking actually at all of the brown stain regardless of that intensity, 1, 2 or 3, in order to determine whether or not it's positive or negative, meaning the pathologist needs to look at basically whether it's a brown or blue stain.

Got it. That's helpful. And just second question, how should we be thinking about a potential biomarker strategy for the lung cancer opportunity?

Thanks, Jonathan. So in regards to lung cancer, we continue to evaluate that data on similarly in a very robust and stepwise approach, just in relationship to what we've described as far as first evaluating by H-score in order to determine what might be the best cutoff, just in relationship to maximizing the potential for benefit.

Our next question will come from the line of Tom Shrader from BTIG.

It's a wonderfully intuitive result, so it's kind of nice. I have a question about the overall trial design. It's kind of based on this idea that if you have a big, very positive subgroup, you can drag the whole ITT population into positivity. My question is -- I mean -- and MacroGenics did the same thing, even though the subgroup was actually less good than the comparator. But the question is, is that a done deal in terms of the FDA? Or to get the whole ITT, would you need some ORR and some decent DOR? Or is it just pure statistics? Just whatever comments you'd make would be great.

Thank you for the question. So let me first start just in relationship with the study design to level set. And so we have 2 shots on goals. As the slide described in relationship to, we're enrolling all patients and then retrospectively assessing in order to bifurcate between the high and the low. And we have a primary endpoint in the high, but the key secondary is an overall population. The key here is that the benefit/risk will be assessed in both of these populations with recognition that if there's sufficient benefit/risk in, obviously, the higher population as the primary endpoint, then that would be the indication. If the benefit/risk is also sufficient in the ITT, because remember, there were patients that also benefit in the lower population, recognizing that it was a smaller sample set, right? So ultimately, it'll be a regulatory discussion just in relationship to whether or not the activity on the benefit/risk is sufficient, including both the high and the low, in order to support that ITT population, because recognizing that one would want to ensure that one can maximally benefit, again, this high unmet need population where that they've essentially failed many more active agents in the earlier line therapy and single-agent chemotherapy with a response rate only upwards of 12% is the current standard of care in that population.

Got it. That's exactly what I wanted to know. And then did you have any really significant responses below TPS of 75 in your data so far?

Right. So we showed that slide actually on the slide just in relationships with the lower patient, right? So we did see 2 patients based upon the lower 75, that we [ spend ] in the denominator of the 18. So that's an 11% response rate in the lower population. So again, it is supporting our strategy, now exactly in relationship to evaluating this in a larger data set.

Our next question will come from the line of Brad Canino from Crédit Suisse.

This has been a very clear and helpful presentation. I just want to ask, if and when this diagnostic is commercially available, do you envision a requirement for fresh tissue or archival tissue would be okay?

Yes. Let me actually pass that over to Dr. Mosher.

Brad, so we've explored the relative use of fresh and archival tissue early in our program because that's a very interesting question in terms of clinical applicability. And it appears that archived tissue is going to be informative. And so that's the strategy that we're going to take.

And I would just add to that. Actually, that's a significant advantage from the standpoint of being able to use the archival tissue. These patients can actually utilize tissue that was obtained, actually, obviously, at an earlier time point, hence, the definition of archival, in relationship to determining the status.

Great. And then I want to follow up on the last question as well because I'm thinking about what these new diagnostic data are telling me. And to me, it increases the probability of success for UPLIFT. But it also, in my mind, increases the likelihood that you have a biomarker restricted label, given you're seeing an 11% ORR, like you said in the NaPi2b low. And that's relatively in line with the systemic therapy benchmarks that you spoke about before. So if you could just expand on your confidence in maybe getting an ITT on the label and what specific designs in the study are giving you that confidence?

Well, thank you for the question. And so the benefit/risk is ultimately the parameters by which FDA will make discussions and decisions in collaboration with the sponsor. And so the point of that is not just the activity of the anti-tumor effects, but it's also the tolerability profile that will be taken into context. And so UpRi, obviously, offers the advantage of an ADC in regards to having the safety profile that's been reproducible in a way with no evidence in relationships with a severe peripheral neuropathy, the ocular toxicities and the neutropenia. And so it affords a potentially different benefit/risk relative to systemic chemotherapies that otherwise are the standard of care at this time. And so the totality of the information on really what we'll be taking into account when one is considering whether it be a specific NaPi2b selected population or the overall population. One additional piece I mentioned is that we would also have the potential to have a complementary diagnostic in order to then have a broad indication, but a complementary diagnostic that the physicians could utilize in order to determine if a patient had a potential differential likelihood of benefit. And this is another vehicle by which physicians and patients as well as regulators could utilize this information to make determinations on benefit/risk.

Our next question will come from the line of Colleen Kusy from Baird.

Does the change in assay, does that change the cutoff at all that you see for the bar for approval in the high NaPi2b patients?

Thanks, Colleen, for the question. So just in regards to the bar for approval, this -- it does not change the bar for approval in relationship to the assessment. The bar will be essentially defined by the unmet need in this heavily refractory population which has, as we've described earlier, within the 12% objective response rate that we defined from the 3 pivotal studies, including the failed Forward I, with a 12% objective response rate in the single-agent chemotherapy that was described there as well as the other 2 studies. And so that is the bar, as you're describing it, in relationship to -- of the level of benefit/risk that UpRi would need to demonstrate the differentiation from.

Great. That's helpful. And in terms of the Phase III study for the operational standpoint from this, do you expect the TPS to be kind of similar to the process for the H-score in terms of ease and the time? And I guess how soon will you be able to know when you hit the roughly 100 threshold of patients in the high NaPi2b cohort?

Right. So the pivotal phase -- the pivotal amendment and relationships of the UPLIFT strategy to describe that, because as mentioned, it's an amendment to an ongoing study. And so we've been able to get sites to rapidly amend and then start initiation on to the UPLIFT study itself. The patients are enrolled without the knowledge of their NaPi2b status, meaning that when they enroll, assuming they're eligible, they're then tested. So the testing itself is not impacting their ability in order to be enrolled onto the study, just to clarify that aspect of it. We've been working very closely with our partners, including Leica, in order to ensure that, that testing result is available as -- after they've come on study already. And it's done on a very frequent basis by which we can then assess and enumerate the number of higher NaPi2b patients.

[Operator Instructions] Our next question will come from the line of Boris Peaker from Cowen.

Just quickly on the perspective from the pathologists, have you asked whether they have a strong preference for TPS versus H-score? Whether it's time that it takes to complete it? Any kind of resource utilization, reimbursement, whatever that may be?

Thanks for the question. So just to clarify your question so we can make sure we address it accurately. You're referring to pathologists that would be the end customers?

Yes. That would ultimately be doing this test, whether they have a preference for TPS versus H-score?

Yes. So actually, let me pass that over to Becky. She has significant knowledge in relationship to the pathologists, in relationship to what makes the most sense for them in utilizing as a commercial diagnostic assay.

Yes. So thanks a lot for asking that question. That's certainly a question that we asked ourselves as we started to think about, what method would convey the appropriate amount of information but would also be most clinically tractable? As we described, performing TPS really only requires one distinction to be made by the pathologist, the distinction of brown and blue. And so it's our belief and the belief of Leica, our partner -- our diagnostic partner, that this is going to be the most clinically tractable method of scoring when this is broadly deployed in general practice.

I'm just curious from the test they do right now, I mean how many of them are TPS versus H-score? What's more popular?

Yes. So I think some of it is context dependent. But as we described, the PD-L1 assay is a broadly deployed assay, which uses a TPS cut point. So that's a very familiar method of reading for pathologists. It's not anything new.

And our next question will come from the line of Jessica Fye from JPMorgan.

It looks like within the data so far that moving to a TPS score led to 55% of patients being NaPi2b high. So do you anticipate potentially enrolling a few more patients in the overall study to ensure you're getting to 100 NaPi2b high patients?

Jessica, good hearing from you. This is Arvin. So when you calculate the 55%, I believe you're calculating it from the efficacy data set. And so when you look at the broader data set, from -- it was Slide 14, what we see is essentially 62% of patients scoring higher when using the TPS of 75. So with this larger data set, we are seeing approximately 62%. Again, within the study, we will be able to complete the study when we achieve essentially 100 higher NaPi2b patients.

Arvin, to add to that. We have analyzed a substantial amount of tissue from tissue banks as well. And I did feel fairly confident that allowed 60% prevalence, it would represent a TPS over 75.

Okay. Got it. I think also, as on the January update, you were maybe waiting for some more scans from a number of patients who just haven't been scanned yet as of the December 3 cut. Now that you have scans for those folks, did that reinforce the 110 H-score as the optimal cutoff, off of which you're basing this TPS cutoff of 75?

So just in relationship to the additional updates as far as the clinical data set, we're still enrolling the expansion cohort and converting them into -- over to UPLIFT at this time. So we're really focused on the activation of the UPLIFT study right now in relationship to -- in relationship to our execution. And so at an appropriate time, when we've completed the expansion cohort, we'll discuss data releases just in relationship to the expansion cohort itself.

Okay. Got it. And maybe lastly, now that you're kicking off UPLIFT here, any kind of better insight into when you expect to complete enrollment?

I'll take this, Jess. I think our current best estimate is that a year from now we should be substantially enrolled. I think as we've said in an earlier call, would like to be further ahead in enrollment before we are in a position to get more definitive time line. But I think based on the path that this study is an amendment to the existing expansion cohort and there was robust enrollment there, the fact that we have strong support from the corporate groups in the U.S. and our expanding in sites in Europe with support from ENGOT, the European Corporative Group, we feel we're on track to achieve our enrollment goals. But again, I think we'll be able -- be in a position to give a more definitive time line once we are a bit further ahead.

Our next question will come from the line of David Nierengarten from Wedbush Securities.

I was just curious if there are any thoughts about the dynamics of testing patients for either NaPi2b or folate or both? And kind of how that decision tree might work for the oncologist and pathologist when seeing a patient, maybe not in the clinical trial, but in the future, a potential commercial setting?

So let me take that and I can see if others want to address it also. But from the standpoint of -- in relationship to correlation between -- we haven't seen correlation between NaPi2b and folate in relationship to expressions. So recognizing that, I think that from the standpoint of the high unmet need, one will be taking into account the fact that the benefit/risk of the agents available in this refractory setting. Recognizing that approximately 60-or-so percent of the patients that we've seen through our tumor banks and so forth are potentially positive or the higher NaPi2b. So just in thinking about the potential that one performs a test, it has a potential outcome that would make them eligible. Certainly, that there would be more than a 50% or the 60% chance that they could be positive in order to obtain that will receive UpRi, assuming the benefit/risk is supportive of being available to them.

I guess, maybe to put a finer point on it when you think about the doctor -- again, the doctor and pathologist, is it more normal or logical when you have a couple of different markers potentially to go from the least exclusionary, for example, the NaPi2b, to the more exclusionary, for example, high folate expressors as kind of like thinking about estrogen receptor positive breast cancer versus HER2-positive breast cancer? Is there any other nuance or thought that we should be thinking about on the application of the test in the future?

Sure. I mean ultimately, the physicians are taking into account the idea that the benefit/risk for these patients, I think, is the critical aspect. And so those are the parameters by which I would anticipate physicians would be making choices. So to your point, it's going to be the test results, but also the available drug that would be available to them. And keep in mind, I think these are really high unmet need patients with significant unmet need. It's never clear whether or not they will be eligible in relationship to receiving additional therapy. So one has to really maximize the shot that one is getting the best therapy upfront. And so that will be a combination of the toxicity profile as well as the level of benefit as well as what's available to them potentially from a biomarker status perspective. Those would just be what I would speculate as the variables that the physicians themselves would be using as criteria. I don't know if that sort of helps to address the thinking in relationship to what you're describing.

The other thing I would add is that we're really aiming to study UpRi in the broadest possible population, as Arvin has described. We have up to 4 lines of therapy. We are not excluding patients who have underlying neuropathy. So I think at the end of the day, the data we will collect from UpRi is really going to drive physician's understanding of the risk/benefit, and I think will have an impact on the selection of agent in addition to the prevalence of the biomarker.

I'm not showing any further questions in the queue. I'd like to turn the call back over to Anna for any closing remarks.

So I want to thank all of you for joining the call today. The initiation of UPLIFT is obviously a very important milestone for Mersana and for our efforts to bring UpRi to heavily pretreated platinum ovarian cancer patients. I think we believe that the robust proof-of-concept data from the expansion cohort and the feedback we've received from the FDA has allowed us to design UPLIFT as a study that has a high probability of success. The robust, predictable, reproducible assay we just described today, I think, further increases the probability of success because it ensures we can deliver UpRi to patients that benefit the most. And as Arvin and Becky have described, the development of our diagnostic has really been a multiyear program, encompassing research, translational and clinical work, a program that has advanced at each stage in tandem with the development of UpRi. So thank you, again, for joining the call. And as always, if you have any follow-up questions, we're available and you should reach out. Thanks.

This concludes today's conference call. Thank you for participating. You may now disconnect.