FibroGen Inc (FGEN) 2017 Q4 法說會逐字稿

Welcome to the FibroGen, Inc. Fourth Quarter and Year-end 2017 Financial Results Conference Call. My name is Adrienne, and I'll be your operator for today's call.

(Operator Instructions) Please note, this conference is being recorded. A webcast of this call will be available on the company website for 2 weeks from today's date.

For opening remarks and introduction, I'll now turn the call over to Karen Bergman, Vice President, Investor Relations and Corporate Communications. Please go ahead.

Thank you, Adrienne. Good afternoon, everyone. Thank you for joining our call today. We're reporting financial results and corporate update for the fourth quarter and year-end 2017.

Joining our call today will be Tom Neff, our Chief Executive Officer, who will begin by reviewing our late-stage product development programs; Dr. Seth Porter, Vice President of Fibrosis Therapeutics, who will provide updates on our pamrevlumab program; Dr. Peony Yu, Chief Medical Officer, who will discuss updates on our anemia program; and Mr. Pat Cotroneo, Chief Financial Officer, and he will review financial performance for the quarter and 2017. Tom will then wrap up with discussion of our goals for 2018 through 2019.

Following our prepared remarks, we'll open the call for Q&A where we'll also be joined by Dr. Elias Kouchakji, Vice President, Clinical Development, Drug Safety and Pharmacovigilance; and Ms. Chris Chung, Vice President of our China operations.

On this call, we do expect to make forward-looking statements regarding our business, including our collaborations with AstraZeneca and Astellas; financial guidance; the initiation, enrollment, design, conduct and results of clinical trials; our regulatory strategies and potential regulatory results; our research and development activities and certain other business matters. For risks and uncertainties regarding our business and statements made on the call today as well as factors beyond our control that may cause differences between current expectations and actual results, we refer you to our Form 10-K for the fiscal year ended December 31, 2017, filed with the Securities and Exchange Commission. Copies of these filings can be found in the Investors section of our website. We undertake no obligation to update any forward-looking statement, whether as a result of new information, future developments or otherwise.

Once again, the format for today's call will include remarks from our management team, and then we'll open the lines up to take your questions.

A webcast of this conference call will be available for 2 weeks on the Investors Page of our website, www.fibrogen.com.

And with that, it's my pleasure to turn the call over to our CEO, Tom Neff.

Thank you, Karen. 2017 was a fantastic year for both of our lead programs, pamrevlumab in fibrosis and roxadustat in anemia. We achieved multiple clinical and regulatory milestones.

Starting with our pamrevlumab program from our Phase II double-blind placebo-controlled study, PRAISE, in patients with idiopathic pulmonary fibrosis, significant treatment effect of pamrevlumab has been shown in multiple parameters clinically important in IPF. These include lung function, proportion of patients suffering disease progression, CT assessment of lung fibrosis and health-related quality-of-life measures.

In August 2017, we first reported meeting primary efficacy endpoint of change in forced vital capacity percent predicted with statistically significant results as well as via volumetric changes in FVC measured in milliliters. We also reported the lowering of the rate of disease progression to 10% in the pamrevlumab arm versus 31.4% in the placebo arm over 48 weeks of treatment with a p-value of approximately 0.01. Disease progression was defined as FVC percent predicted minus 10 or greater or death.

These efficacy results and safety data were reported at the European Respiratory Society International Congress in September. High-resolution computed tomography, or HRCT, was used to assess the progression of fibrosis from baseline to week 24 and 48. Quantitative lung fibrosis assessment in collaboration with our MedQIA group shows statistically significant attenuation of fibrosis at 24 and 48 weeks. While the methods have advanced for analysis of HRCT measures over the past half decade, the results from Study 067 are consistent with our first open-label IPF Study 049.

Pamrevlumab is the first IPF drug candidate reported to have statistically significant effect on fibrosis. To evaluate the impact of pamrevlumab on IPF patients, how they feel and their activities, we used 2 health-related quality-of-life measures for respiratory disease. There is a statistically significant (inaudible) impact between pamrevlumab and placebo-treated patients, and the UCSD shortness of breath questionnaire suggesting reduction of dyspnea by pamrevlumab. We also saw compelling results from the St. George's Respiratory Questionnaire that is validated for COPD and has been broadly applied in IPF. We plan to present the HRCT data, the USCD shortness of breath questionnaire and the St. George's Respiratory Questionnaire results at upcoming respiratory conferences likely to be either ATS or ERS later this year.

Finally, it is important to know pamrevlumab's safety profile and Study 067 death rate was -- with half of placebo, with the rate of treatment (inaudible) leading to discontinuation only 43% of placebo arm. It was well tolerated without safety signals.

In conclusion, the totality of data from the Phase II IPF studies, FVC disease progression, quantitative HRCT, health-related quality of life and an acceptable safety profile suggesting that pamrevlumab has distinctive potential to be a beneficial treatment for patients with IPF.

We are making a concerted effort in our Phase III Clinical Development plan on IPF that we would like to share with you by midyear. In pancreatic cancer, we are developing pamrevlumab for patients in LAPC who are not eligible for local -- for a reception. LAPC refers to locally advanced pancreatic cancer. There is no standard treatment for this patient population based in short-life expectancy, and pamrevlumab may represent an entirely new approach for this serious and fatal condition. In our Phase II trial, we observed that treatment with combination of chemotherapy, plus pamrevlumab changed eligibility for surgical resection of a majority of the treated patients from nonresectable to resectable, a clinically important result that had not been previously reported with other drug candidates. While there is no standard of care for locally advanced pancreatic cancer, the control arm in this study received chemotherapy -- while the treatment arm received chemotherapy plus pamrevlumab. The combination with pamrevlumab yielded a rate of resectability and resection several fold greater than chemotherapy alone. Successful resection is expected to provide a better outlook for survival as compared to nonresected patients.

Moving to our anemia program. 2017 was an equally exciting year for roxadustat. Back in January 2017, we reported positive safety and efficacy results from 2 China Phase III pivotal studies of roxadustat for the treatment of anemia associated with chronic kidney disease. These results supported our roxadustat NDA for treatment of CKD, which was accepted for review by the China Food and Drug Administration, or CFDA, on October -- in October '17. That filing triggered payment of $15 million milestone to FibroGen by our partner, AstraZeneca. In China, roxadustat has received priority review designation from the CFDA. Certain activities in the NDA review process, such as clinical trial site inspection, manufacturing site inspections have been initiated or are partially completed on schedule with projected approval timing. We expect to remain actively engaged in managing review and inspection activities in the next few months. We are on track in the establishment of our commercial API manufacturing facility in (inaudible). Construction is substantially completed as is equipment installation. We will be focusing on plant commissioning and process transfer and qualification in the upcoming months.

The FibroGen AstraZeneca China commercialization team is in launch-preparation mode. The top launch priority is addressing critical market access constraints in China, such as gaining hospital listings and gaining entry into various reimbursement lists. Both companies are staffing up to ensure we are launch-ready when the NDA is approved and the first batch of commercial product is released.

For our U.S. program, we and our partner AstraZeneca have agreed that continued projections of adjudicated MACE events in various scenarios are no longer needed and instead now, we have set study completion timelines as follows: to complete enrollment in the second quarter of 2018; to report top line results in the fourth quarter of 2018; and to file the NDA for roxadustat and CKD anemia during the first half of 2019. We are also working with Astellas to discuss corresponding timetables for Europe and Japan and expect to have a plan soon.

In other therapeutic indications, we recently achieved first inpatient for study 082 -- first patient in study for patient 082, this is the MDS study in the U.S. evaluating roxadustat for treatment in anemia and patients who are on transfusion, so to reduce or eliminate transfusion dependence.

I would also like to provide an update on our plan for expanding our roxadustat indications jointly with AstraZeneca. We have committed to progressing further evaluation of chemotherapy-induced anemia in the U.S. and in China. CIA is a very large market opportunity where the use of ESA therapy has dropped from more than $5 billion annually to virtually nil due to the well-documented safety problems associated with ESA products. After intensive study over the past few years, we believe roxadustat's profile differs in numerous respects as the HIF mechanism address all aspects of retrogenesis without exposure to excessive levels of EPO. As such, we believe that roxadustat can provide a much-needed therapy that can come -- overcome the inflammation, a common hurdle for EPO and avoid the documented safety problems of ESAs.

Moving on to corporate matters. FibroGen is focused entirely on executing the necessary steps that will realize the potential in our product pipeline and build value for our shareholders. In April, we raised $115 million, supporting the expansion of roxadustat into new indications in China. And in August, we raised $356 million, supporting the advancement of pamrevlumab into Phase III in pivotal studies in IPF in pancreatic cancer. We continue to be most grateful and appreciative for the support we receive from our investors and our analysts as we advance our primary objective of delivering first-in-class therapies to patients with limited or no treatment offerings.

We concluded 2017 with $762.2 million in cash and cash equivalents.

And with that, I'd like to ask Seth Porter to discuss the status of our pamrevlumab product platform in more detail. Seth, please go ahead.

Thank you, Tom. As Tom mentioned, in August 2017, we reported positive top line results from our randomized double-blind placebo-controlled Phase II clinical trial, Study 067, designed to evaluate the safety and efficacy of pamrevlumab in patients with mild to moderate IPF. We also presented top line results from 2 substudies that were added to evaluate the safety of combining pamrevlumab with approved IPF therapies. In the double-blind placebo-controlled 48-week portion of the study, 103 patients were randomized 1:1 to receive either pamrevlumab or placebo every 3 weeks. Lung function was assessed at baseline and at weeks 12, 24, 36 and 48. Quantitative HRCT assessments were performed at baseline, week 24 and week 48. Pamrevlumab met the primary efficacy endpoint of change in FVC, the same primary endpoint used in pivotal trials for approved IPF therapies. The average decline in FVC percent predicted from baseline to week 48 was 2.85% in the pamrevlumab arm as compared to an average decline of 7.17% in the placebo arm, a statistically significant difference of 4.33% with a p-value of 0.0331. Similar results were obtained for FVC values. Pamrevlumab-treated patients had an average decrease in FVC of 129 milliliters at week 48 compared to an average decrease of 308 ml in patients receiving placebo, a statistically significant difference of 178 milliliters, a relative difference of 57.9%, with a p-value of 0.0249.

In addition, the proportion of patients with disease progression, defined as FVC percent predicted decline greater than or equal to 10% or death was lower in the pamrevlumab group compared to placebo at each assessment. The difference between treatments was statistically significant at weeks 36 and 48. Specifically, at week 48, in the pamrevlumab-treated arm, only 10% of subjects experienced disease progression or death compared to 31.4% in the placebo arm, with a p-value of 0.0103. The percent of pamrevlumab patients who experienced disease progression and also discontinued therapy was less than 15% of that in the placebo arm.

With regard to high-resolution CT assessments, we measured quantitative lung fibrosis using HRCT at baseline, week 24 and week 48 using computer algorithms developed at UCLA to specifically assess the forms of lung fibrosis that are characteristic of IPF and to measure changes in them over time. We found a statistically significant attenuation of fibrosis progression from baseline in the pamrevlumab-treated arm versus the placebo arm at both week 24 and week 48. Furthermore, a correlation between changes in FVC percent predicted in quantitative lung fibrosis was confirmed at both weeks 24 and 48. The correlation between changes in FVC and quantitative lung fibrosis is consistent with the results obtained in our open-label Phase II study, Study 049, and strengthens our belief that pamrevlumab is truly anti-fibrotic and can lead to benefits in both lung structure and function. We are not aware of any therapy that has shown a statistically significant effect on lung fibrosis that's detected by similar quantitative measures of lung fibrosis.

Pamrevlumab-treated patients also achieved a clinically meaningful improvement in health-related quality of life using this St. George's Respiratory Questionnaire, or SGRQ, comparing active to placebo during the 40-week treatment period. The SGRQ is a respiratory quality-of-life instrument that is validated in chronic obstructive pulmonary disease. Additionally, a subset of 43 clinical trials subjects were assessed using the UCSD shortness of breath questionnaire that yielded a statistically significant difference between active and placebo arms, and indicated pamrevlumab may attenuate the progressive worsening of dyspnea in IPF patients in comparison to placebo. Consistent with previous studies, pamrevlumab was well tolerated. TEAEs were mild or moderate in severity and considered typical of the patient's underlying medical conditions. Infusion-related TEAEs were generally mild. Compared with placebo, fewer pamrevlumab patients were hospitalized following an IPF-related or respiratory TEAE or died for any reason. The reduction in mortality over placebo was an important and encouraging clinical finding. The comparisons of pulmonary function outcomes, disease progression, safety and mortality discussed above were largely presented in September at the ERS meeting. We plan to present the SGRQ, UCSD shortness of breath questionnaire and HRCT results in upcoming medical conferences. Our IPF team is focused on reviewing data with thought leaders and investigators to develop our Phase III trial program.

Switching to pancreatic cancer. In January 2017, we reported interim results from a randomized active control neoadjuvant Phase II trial combining pamrevlumab with nab-paclitaxel plus gemcitabine in 37 patients with locally advanced pancreatic cancer at the ASCO GI Symposium. We completed enrollment in this trial in the first half of 2017 and completed the 6-month treatment period and surgical assessment at the end of 2017. A greater proportion of subjects treated on the combination arm were converted from nonresectable to resectable status, consistent with our data reported at the ASCO GI meeting.

Based on these promising results, the pancreatic cancer team is working to prepare an FDA briefing and finalize a pivotal trial protocol. If pamrevlumab can be proven to enhance tumor resection in patients with otherwise nonresectable tumors, it would be expected to double their survival time based on post-hoc analyses of resection benefit.

Switching now into Duchenne muscular dystrophy. With regard to our ongoing Phase II trial evaluating pamrevlumab in the treatment of Duchenne muscular dystrophy, the trial is focused on nonambulatory DMD patients, and we are making good progress toward our enrollment goal of 22 patients. Based on what we believe to be the anti-fibrotic mechanism of action in DMD, the trial is not specific for any particular dystrophy mutation, which may allow pamrevlumab to benefit a broad DMD patient population.

Thank you for your time today, and I'll turn the call back over to Tom.

Thank you, Seth. Dr. Peony Yu, our Chief Medical Officer, will now talk about our global Phase III program for roxadustat in addition to providing comments on activity in China and Japan and other opportunities for this program. Peony?

Thank you, Tom. As many of you know, we and our partners, AstraZeneca and Astellas, are conducting a global Phase III program designed to support regulatory approvals of roxadustat for treatment of CKD anemia in dialysis and non-dialysis-dependent patients in the United States, the European Union, Japan and China. Our U.S. and EU Phase III program has an aggregate target enrollment of more than 8,000 patients and is the largest Phase III clinical program conducted to date for a development of an anemia drug.

As previously described, the Phase III program is powered to demonstrate cardiovascular safety using MACE composite endpoint at primary safety endpoint. MACE is being evaluated in 2 separate study pools: dialysis-dependent CKD patients and non-dialysis-dependent CKD patients. In the non-dialysis Phase III study, roxadustat is being compared to placebo. The regulatory requirement is to demonstrate superiority on efficacy and noninferiority on safety. For dialysis-dependent patients, roxadustat is compared to a epoetin alfa, and the regulatory requirement is to show noninferiority in efficacy and in safety.

Safety data from roxadustat Phase II and Phase III studies have undergone a number of periodic reviews by an independent Data Safety Monitoring Board, or DSMB, on a regular basis. The DSMB has consistently recommended, most recently in December of last year, study continuation under the current protocol with no changes.

In support of the U.S. NDA submission, we and AstraZeneca are completing Phase III program enrollment in the second quarter of 2018. Data readout from our Phase III study is planned for the fourth quarter of 2018, and we are targeting submission of U.S. NDA during the first half of 2019.

Turning to China. CFDA accepted our China NDA filing of roxadustat for treatment of CKD anemia in Q4 2017. This NDA represents the first filing of any HIF-PHI worldwide. China could be the first approval country for a first-in-class drug, and we would be privileged to be part of that.

In Japan, our partner, Astellas, has announced the completion of 3 of its 6 Phase III trials, top line results from a study in peritoneal dialysis patients were reported last year, and Astellas has stated that data readout from the other 2 completed studies in dialysis patients are planned for first quarter of 2018.

2018 is an exciting year for roxadustat as we look forward to the results from a number of Phase III studies we're preparing and supporting NDAs of this first-in-class drug around the world. We will update you as appropriate during the coming year.

A brief update on other roxadustat programs. We are also pursuing other opportunities in anemia therapy and have launched development of roxadustat in myelodysplastic syndrome, or MDS. MDS is a serious disorder that impairs the ability of the bone marrow to produce healthy red blood cells and is, in most cases, associated with anemia. There is no approved treatment for anemia in MDS patients in the U.S. or in China. We started patient dosing in our Phase III global study in transfusion-dependent MDS patients and will start enrolling nontransfusion-dependent MDS patients into a separate Phase II/III study in China first half of 2018.

We are actively evaluating opportunities to expand roxadustat indications beyond anemia of CKD and MDS. Among the various types of anemia, we and AstraZeneca identified chemotherapy-induced anemia, or CIA, in cancer patients to be most compelling in light of the unmet medical need, the large market potential and the extensive preclinical efficacy and de-risking safety result in support of this indication. At present, we are making plans for the optimal development path in both the U.S. and China. I'd like to now turn the call back to Tom.

Thank you, Peony. Pat Cotroneo, our Chief Financial Officer, will walk us through the financial highlights for the fourth quarter and year-end 2017. Pat?

Thank you, Tom. As announced today, total revenue for the quarter ended December 31, 2017, was $42.5 million. For the same period, operating expenses were $66.3 million, and net loss was $22.1 million or $0.27 per basic and diluted share. Included in operating expenses for the quarter ended December 31, 2017, was an aggregate noncash portion totaling $11.8 million, of which $9.9 million was a result of stock-based compensation expense. Total revenue for the year ended December 31, 2017, was $125.7 million. For the same period, operating expenses were $248.3 million, and net loss was $126.2 million, or $1.73 per basic and diluted share. Included in operating expenses for the year ended December 31, 2017, was an aggregate noncash portion totaling $45.3 million, of which $37.5 million was a result of stock-based compensation expense. As of December 31, 2017, we had $762.2 million in cash as compared to $342.2 million at the end of 2016, reflecting the net returns of our 2 recent financing. For these purposes, total cash refers to cash, including cash, cash equivalents, receivables, investments consisting primarily of investment-grade corporate debt and restricted time deposits relating to our building lease. On our balance sheet, the category of long-term investments consists entirely of investment-grade corporate debt with remaining maturities of fewer than 2 years.

With the submission of our NDA filing for roxadustat in China, we received a $15 million milestone from our partner, AstraZeneca, during the fourth quarter of 2017, which was fully recognized as licensed and milestone revenue in the same quarter. We are currently projecting 2018 year-end cash balance in the range of $600 million to $620 million.

Thank you, everyone, and I would like to turn the call back over to Tom.

Thank you, Pat. 2018 looks to be an exciting year for both pamrevlumab and roxadustat. Relating to the most advanced indications in pamrevlumab and IPF, we expect to report data supporting benefits of pamrevlumab in disease progression, fibrosis and patient-reported outcomes at upcoming medical conferences, either ATS or ERS. Our near-term focus is on designing an optimal Phase III program for supporting product registration.

In pancreatic cancer, we plan to report results from our Phase II trial pamrevlumab and locally advanced pancreatic cancer at an upcoming medical conference. We hope to very shortly have pivotal trial design agreed with FDA. For roxadustat, with our partner AstraZeneca, we plan to report results from our Phase III roxadustat studies in CKD anemia in the fourth quarter of 2018. We and AstraZeneca expect to file a U.S. NDA for roxadustat and anemia CKD in the first half of 2019. We are working with Astellas to finalize plans for filing in the EU and Japan. In Japan, we expect data readout from the long-term human dialysis ESA conversion study and hemodialysis correction studies during the first quarter of 2018. In China, we anticipate we may receive a market approval decision for roxadustat NDA for CKD anemia by the end of 2018. We expect further work with both our partners on a strategic plan for new indications, particularly in CIA.

With that, I would like to turn the call back over to Karen for question-and-answer. Karen?

Thanks, Tom. Operator or Adrienne, if you could kindly open up the line for questions.

(Operator Instructions) And our first question comes from Michael Yee from Jefferies.

Two questions for you. One for Tom. Appreciate the details around the timing for data this year. I look forward to that. Maybe you can comment around what changed in terms of the analysis around the adjudicated events? Was it just hitting the number of events? Or was there great clarity on how that was progressing and perhaps the blinded event rate? In other words, just comment on what you saw there and why you feel perhaps good about that? And then my other question is on pamrevlumab. Presumably the pancreatic data should be finished or near finished. Maybe just comment about when we could see that data and how that plays into when you will just announce your regulatory update.

So for roxadustat and AstraZeneca, Michael, I think really what happened was, we've had an ongoing effort to count adjudicated unique MACE events for a couple of years with efforts to project in various scenarios when we would pass the statistically required numbers. And as you look at the totality of those efforts of counting and projecting, we got to the point that we felt like we were adequately covered in all the scenarios we could think of as being reasonable. And as a result, we said we now need to really focus on execution, and that therein, there really was a decision to move to a timetable-based approach to get to the NDA. Elias, I think maybe you would like to take the pancreatic cancer question, please.

Mike, for the question on the pancreatic cancer, as you heard from Seth, the study has completed the enrollment and the resection although the patient is still on the study. So the analysis is still ongoing, and that patient is being followed. We hope to have this data available by midyear.

Okay, and just if I could have one follow up for Tom on what you said about roxa. Do you -- can you extract any information or any information about the blinded event rates that make you feel good about the study or make you feel good that, that's -- that the event rates should hit noninferiority?

What we could do about this is look at the pooled event rates in the non-dialysis and the dialysis area, and what I'll do is let Peony comment on exactly what we've seen to date with these pooled analyses so that you get the details accurate from the horse's mouth. Go ahead with that, Peony.

Mike, as you know that since the studies are -- the program is still ongoing, although we are seeing that we are near completion, the assessments that we can make by following the MACE event rate alone is the combined rate of the -- of both treatment arms. And we are, mainly in the non-dialysis of those that -- the patients in the study have more events, kidney disease than previous than reported in previous trial in non-dialysis patients. The event rate from -- based on per patient exposure year, we have observed that to be lower than the previous reported studies. Similarly, in the dialysis pool, the combined rate also appears to be lower than previously reported studies in EPO. And at this time, this is about all we can comment on the rate, and we have already made adjustments along the way using sample size, so we are very comfortable with where we are in this program as well as the time line for completing this program.

And the next question comes from Terence Flynn from Goldman Sachs.

This is actually Holly on for Terence. When you release the data for roxa in CKD, will you plan to report top line data from the non-dialysis and dialysis patients together or sequentially?

So I think the thinking in that area is that we will have top line data to report in the fourth quarter. What we won't be able to report in the fourth quarter is the pool MACE analysis. And so I think we are looking at this as getting the studies done, the 5 between FibroGen, which has 3, and AstraZeneca which has 2. Again, these studies done as promptly as is possible with no intent to sequence or couple or anything, just get them done. The Astellas studies, I have to indicate here that we haven't had a chance yet to meet with the Astellas' upper management on the scheduling, so we'll know in 1 week or 2 about the plan there, with potentially 2 more studies with Astellas. And the core activity for the remainder of the NDA is really very, very close analysis to the pooled MACE and in each non-dialysis versus placebo and in dialysis versus epoetin alfa So that's the current thinking.

And our next question comes from Andy Hsieh from William Blair.

I just wanted to follow up on Michael's question. Is it fair to characterize the change for the MACE study to be from an event-driven study to a landmark study? Is that a fair characterization?

Peony, do you want to take that one on?

Sure. Andy, the nature of the study has not changed. What has changed is that we have much more data based on what we have been tracking during the course of the Phase III, and we are at the point that we are confident that we will have sufficient data to meet our safety data requirements for the program. Does that help?

So right. So I guess the question has to do with the analysis plan. So it -- my interpretation, and please correct me if I'm wrong, it sounds to me that what Tom described, the change can be characterized as basically going from event-driven, right, you need to hit certain number of events, to a landmark, which means, once you hit a certain time, you stop the trial then do the analysis. Is that a fair characterization about the change?

Andy, let me try this. We have looked at scenarios for cumulative MACE count of a variety scenarios, I'm sure you can think of a few. And we have come to the conclusion that we are likely to exceed the necessary MACE numbers in the non-dialysis in the scenarios we know about there, and in dialysis in the scenarios we think about there, and the eventualities that we think are likely to happen. And as a result, there's a judgment made that we are adequately covered now across the range of possibilities. And so instead of having things focused on the forecasting exercise and accounting which, of course, involves an education committee and there's a little bit of a lag effect, everyone's agreed we don't need to do that anymore, we just need to focus on getting done. And so it's really an agreement with our partner, AstraZeneca. We are now executing against a timetable given that we have met the MACE requirements across the categories that we think are likely to happen. Is that clear enough?

Yes, that's very clear.

And our next question comes from Geoffrey Porges from Leerink.

Just wondering if you could give us a little sense of where you are with pamrevlumab now in terms of the dose. You've been starting 30, 35 milligrams per kilogram. I'm wondering if you're taking that dose higher. And secondly, if you've received any feedback on resectability, a survival pivotal trial endpoint in pancreatic cancer?

Elias, go ahead.

Geoff, first of all, for resectability, we have not yet spoken with the agency, but this is as planned, so we are in the process of preparing for this discussion. For the dose, as you mentioned, we -- some of our studies in pancreatic cancer is 35 mg per kg every 2 weeks, while in IPF, we study this at 30 mg per kg every 3 weeks. We are doing some work in simulations. You have seen some of this as presented, and we think that there is still an opportunity to modify this dose. So the final dose have not yet been defined for the IPF, and we will take this again as a -- on our follow-up discussion with the agency with a very convincing argument how we will need to update the dose to support a wider therapeutic margin. I'd like to emphasize the therapeutic margin we are in now is very good, but if we can go above and beyond this as we should to make our best effort for the patient's benefit.

And our next question comes from Joel Beatty from Citi.

First one's on roxadustat, could you discuss, is the trial powered to show superiority? Did you see that as a potential for either the non-dialysis or dialysis patients?

Peony, you can answer that.

Okay. Joel, as mentioned for non-dialysis, for comparison with placebo, efficacy target is for superiority, and for non, for safety comparison, the regulatory requirement is noninferiority. As you know, that placebo is the gold standard for safety. And so we are powered to show that when the drug is non inferior to placebo is absolutely safe. Now if we could further derive benefit of -- safety benefit from anemia treatment, then we see that as an icing on the cake. As for the dialysis program, the regulatory requirement is the noninferiority in the efficacy and noninferiority in safety endpoint. And we would also perceive an opportunity to be superior in safety to be a plus. In the China study, also we originally powered for noninferiority, we ended up demonstrating superiority in efficacy. That is Study 806. So until we see the data, we will hopefully have a more direct answer at that time.

Okay. And then one other question, regarding the cash balance guidance, does that include any Phase III cost for pamrevlumab?

Pat, you want to take that one?

No, it doesn't.

And the next question comes from Difei Yang from Mizuho Securities.

This is actually Alex on for Difei. You mentioned earlier you expect a decision in China for CKD anemia by the end of 2018. Are there any steps leading into final approval? Like for example, the AdCom meeting in the U.S., anything similar there?

Chris, do you want to take that on?

Sure. Absolutely. Alex, so China, there are (inaudible) very similar to the U.S. FDA. However, it's highly up to the judgment of the CFDA as to where they're needed. At this point in time, we have not received any indication as to whether the CFDA plans to cross (inaudible).

Okay, got it. That's useful. And then just one on IPF. Could you share at this point any color on the Phase III trial and what that might look like? Will it have an active control arm?

Sure. This is Elias. At this time, we still have not made the final decisions. So still, this will be up to the discussion with the agency. We already developed a multitude of scenarios, and that all these will be presented hopefully in the near future.

And our next question comes from Alex Schwartz from Stifel.

This question, maybe for Seth. So with the IPF data you released in your January corporate presentations, you're going to HRCT data that shows pamrevlumab slowing fibrosis first placebo, and you talked about subgroup of 90 patients that had more significant attenuation in their fibrosis. So can you go more into these patients? Are there any identifiable characteristics by baseline or biomarkers that might be predictive of who might respond? And then, secondly, can you maybe help me with the context of the data a little bit more? Was it a few patients who had a more significant attenuation or reversal in their fibrosis? Or is it that most patients had some attenuation in their fibrosis or both?

So Alex, this is Tom Neff. I think you have a misconception here that I have to correct. The data we referred to today, the HRCT data that we worked on with MedQIA, involves all patients, not a subgroup, and it is done in the same rigorous manner that the FTC data was done. So what you saw in January was a comparison at that time to something reported by a company that has an approved product in the area, and we were not yet quite finished with the MedQIA analysis that will be presented either at ATS or ERS later in the year. So just to frame it properly, that's the situation. Do you want to reframe the question for Seth?

So maybe just help me a little bit more. So the 103 patients in the 48 weeks, there's a 90-patient subgroup analysis. Can you just kind of -- or maybe help me understand it better, kind of what's the difference between those 2 populations?

So as Tom mentioned, that 90-patient analysis was done earlier, and we hadn't completed our analysis. We've now completed our analysis and we've looked at the entire intent-to-treat population. And so the data that I reported -- that we reported today, showing statistical significance at week 24 and week 48, that's for the entire population and then we have HRCT data. So I would really focus on that. The data that we reported during JPMorgan in terms of 90 patients, that was sort of an interim assessment. And for that, a number of patients were taken out of the analysis pool because they were largely from one country in which we had sort of on unexplainable results. We've now moved beyond that, and so I don't think we need to focus on that subset analysis.

Part of the rigor and looking at the data was looking for any unusual subgroups, and there was one particular country in which the treated did worsen and the placebo did better kind of thing, and you just sort of stare at it and say, "Is that really real?" And that investigation is ongoing actually. But the results we're reporting for HRCT that we discussed in this call is on the full 103, and it is statistically significant both at 24 and 48 weeks.

And this concludes the Q&A session. We'll now turn the call back over to Tom for final remarks.

Thank you all for joining the call today. I would like to express my thanks to our FibroGen teams in the U.S. and in China for an incredible year and my gratitude to our partners and our investors for their support. We look forward to updating you on our progress in the coming quarters and year. I'd like to wish everyone a good afternoon and evening. Thank you very much for being here.

Thank you, ladies and gentlemen. This concludes today's conference. Thank you for participating, and you may now disconnect.