NantHealth Inc (NH) 2017 Q2 法說會逐字稿

Good day, ladies and gentlemen, and welcome to the NantHealth Second Quarter 2017 Financial Results Conference Call. (Operator Instructions)

As a reminder, this conference is being recorded. I would now like to turn the conference over to Robert Jaffe, Investor Relations for NantHealth. Please begin.

Welcome, everyone, and thank you for joining us today to discuss NantHealth's 2017 Second Quarter Financial Results. On the call today are: Dr. Patrick Soon-Shiong, Chief Executive Officer; Ron Louks, Chief Operating Officer; and Paul Holt, Chief Financial Officer.

This call is being broadcast live at www.nanthealth.com. A playback will be available for 3 months on NantHealth's website. I'd like to remind -- I'd like to make a cautionary statement and remind everyone that all of the information discussed on today's call is covered under the safe harbor provisions of the Litigation Reform Act. The company's discussion today will include forward-looking information reflecting management's current forecast of certain aspects of the company's future and actual results could differ materially from those stated or implied.

In addition, during the course of this call, we may refer to non-GAAP financial measures that are not prepared in accordance with U.S. generally accepted accounting principles, and may be different from non-GAAP financial measures used by other companies. Investors are encouraged to review NantHealth's press release announcing its full 2017 second quarter financial results for the company's reasons for including those non-GAAP financial measures in its financial results announcement.

The reconciliation of non-GAAP financial measures to the most directly comparable GAAP financial measures is also contained in the company's earnings press release issued earlier today.

Today, Dr. Soon-Shiong will provide a brief overview of the quarter and discuss the business lines, followed by Paul, who will discuss the financial results in more detail. We will then open the call for questions.

With that said, I will now turn the call over to Dr. Patrick Soon-Shiong.

Thanks, Robert. Good afternoon, everyone, and welcome to NantHealth 2017 second quarter results. Today we have much to report including a number of very positive business and operational developments and to remind everybody we are a very young company, only 13 months since our IPO.

So let me begin with our plan to focus on our core competencies and enhanced profitability and position the company for growth. As many of you know, prior to our IPO, we completed a number of acquisitions to establish this integrated software system to enable our leading clinical -- learning system to evolve. These acquisitions, including, Net. Orange, which is the clinical operating system or cOS software; iSirona, the device connectivity solutions; Harris, the middle way into connecting various multiple EMRs; Eviti, the decision support engine for cancer. And finally, NaviNet, the payer portal. In addition, we are entered into a relationship with NantOmics to deliver GPS.

With all these acquisitions, NantHealth ballooned close to 1,000 employees. And over the last 6 to 9 months, (inaudible) took the decision to methodically analyze and carefully identify synergies amongst the many skilled employees that were acquired through these acquisitions.

With the hundreds of software engineers and architects of these disperse systems, this was an enormous task, which we now have finally completed.

The company is now focused being the leading cancer diagnostic outcomes driven solutions provider, and we are now positioned for growth of our artificial intelligence platform for cancer care.

We have now implemented this restructuring, which is now included integrating these multiple acquisitions made prior to our IPO. And this includes now, the integration also of GPS Cancer ProteoGenomic Molecular Profiling Solution, which involves large machine learning by [providers].

And we're also integrating the Eviti Clinical Decision Support which currently covers over 23 million lives with Connected Care and Provider/Payer Engagement services.

This integration allows our company now to be focused on cancer and become a company leading artificial intelligence and the delivery of 21st century care. And as a consequence, we identified non-core assets, which included the Provider/Patient Portal.

We have now signed an agreement with Allscripts to sell this Provider/Patient Portal assets, and allow the remaining engineering teams to focus on the artificial intelligence and decision support agents. The sale of the assets and restructuring limitation will see a significant reduction of operating losses. The consideration for the sale of assets include the conveyance of 15 million NantHealth shares previously purchased by Allscripts.

So between these synergies and the transfer of some number of personnel to Allscripts, together with NantHealth's internal consolidation, our headcount will be reduced by approximately 300 staff.

This sale and the other steps the company is now taking, including the integration of the acquired businesses is expected to result in an annualized cost savings of more than $70 million.

Now let's turn our attention to GPS Cancer. I believe we now really have reached an inflection point and turned the corner. Persistence in believing that the science, although complex, provides the most comprehensive information to practicing clinicians and point of care is beginning to bear fruit. July has been the largest month of orders to date in the history of the company. And this month also, we have now completed the documentation and validation studies of GPS Cancer for filing to CMS for coverage.

With regard to the number of GPS users with the release spelled out that there were 379 commercial tests ordered in Q2, with 264 commercial tests delivered. And the number of oncologists that have ordered the test has grown to 432. We remain focused on extending coverage and driving physician engagement for this solution around the world. And we strongly believe that GPS will result in extended and improved quality of life internationally as well as in the United States. And we are pursuing opportunities through partnerships with locally-based resellers and we've now added a seasoned sales executive to aid in our efforts. As we've just recently reported, Sistemas Medicos Nacionales became the first international payer to cover GPS Cancer for patients bringing a new standard of care to Mexico.

On the domestic front, our efforts include introducing pilot programs with commercial insurance and self-insured payers to accelerate coverage adoption. In addition, we are continuing to add experienced professionals to our sales and clinical support teams to further drive GPS physician adoption nationwide.

I'm pleased to say that our team has developed a deep, late-stage pipeline that includes national and regional health plans, and new orders are trending favorably, as a result of these increased resources. With this continued adoption and understanding of the value that GPS brings to the prescribing physician, I am now beginning to see how [cancer care and the payment of] cancer care is changing in this nation. I fully believe that the future is value-based care, which is outcomes divided by cost. In order for both the physician and the payer to understand the outcomes and cost in real-time, an integrated platform of Cancer Care is needed. This is NantHealth's mission, to provide a platform that combines machine learning, diagnostics at the most sophisticated level -- down to the level of protein, where drugs interact. To provide the clinical decision support engine at point of care called Eviti, with which the doctor and patient could interact and then interact with the payer through the NaviNet system.

This is our goal. This is our mission. And we have now executed that plan. So driving through the next element is Eviti, which is a clinical decision support engine, and we are pleased to announce that these covered lives have now increased approximately $23.4 million at the end of Q2 from $22.5 million, the end of Q1.

Significantly, the company has now signed a 3-year contract extension with Blue Cross and Blue Shield of Nebraska, for Payor engagement as a SaaS service line. This agreement also adds NaviNet Open Document Exchange solution, which is a solution set that now allows document exchange between the provider and the payer. We've also added a new NaviNet Open customer, with the execution agreement with the Medical Mutual of Omaha, and then finally, the company has now executed an agreement with a new payer customer, which is intended for nationwide deployment of this Eviti NantHealth Clinical Decision Support solutions.

So with that, I would like to now turn over the call over to Paul to discuss our financial results in more detail. Paul?

Thank you, Patrick, and hello, everyone. I want to first review our second quarter financial results, and then review the impacts of our restructuring plan including the sale of assets to Allscripts.

Second quarter 2017 revenue of about $26.2 million declined by approximately $5.3 million or 17% from the $31.5 million reported in the same quarter of the prior year. The decline was primarily the result of the prior year revenue, including a significant amount of revenue from the completion of a large services arrangement. This is partially offset by growth in our recurring revenue lines as well as GPS revenue.

I'm pleased to report that revenue grew on a sequential basis. The second quarter revenue rose 17% compared to the first quarter revenue of $22.5 million. The increase was primarily in our software maintenance revenue and resulted from increase in Connected Care revenue from completed projects.

GPS delayed revenue of $415,000 was up 900% compared to $45,000 reported in the prior year second quarter. GPS revenue recognize in the quarter primarily comes from 3 sources. Our international distributors, cash received for payers who don't have agreements with us, and profile delivered to patients who are insured by payers with whom we have agreements. A substantial portion of the 264 profiles delivered in the quarter or with insured by payers that we do not have agreements with, resulting in cash basis treatment.

We expect that proportion of revenue being recognized on a cash basis to come down, by the time as we gain experience with the growing number of payers who are covering the GPS profile. I would note, we'll regain to see an increase in reimbursements received from payers who have yet to sign formal coverage agreements. This is due to our ongoing collections efforts including utilization of pre-authorizations, and supporting documentation and the overall building and appeal process. Collections in July have exceeded our internal expectations and have provided a significant head start for third quarter GPS revenue.

Going forward, our focus is on expanding both our volume profiles and a portion of GPS profiles that are reimbursed, as well as increasing reimbursement from payers who are not yet formally covering GPS.

Turning to other revenue lines. Software and hardware revenue was $3.4 million versus $4.1 million in the year-ago quarter. As I have mentioned before, revenue from this line is often a bit choppy, given the timing and completion of device connection implementations. SaaS revenue was $15.3 million compared to $15.2 million a year ago. The increase was primarily driven by growth in our Clinical Decision Support revenue offset by a small amount of sales incentive deferral related to NaviNet in the year-ago quarter.

Other services revenue declined to $2.4 million versus $7.6 million a year ago. The decrease is primarily due to the prior year quarter, including the recognition of a large services -- software services arrangement with a strategic customer of our patient and provider engagement software.

Gross profit was $9.6 million or 36% of revenue compared to $9.3 million or 29% of revenue in the same quarter a year ago. The increase in gross profit was tied to several factors, including a lower amount of stock compensation expense, decline in recognition of previously deferred customer implementation software services cost, partially offset by higher GPS-related cost.

On a sequential basis, gross profit increased by $6.1 million or 175% from $3.5 million in Q1.

Selling, general and administrative expenses decreased substantially to $22.9 million from $47.2 million in last year's second quarter, largely due to lower stock compensation expense. Research and development expenses decreased $12.5 million or 51.3% to $11.8 million from $24.3 million in the prior year second quarter. The primary drivers, lower stock compensation expense, R&D expense as a percentage of total revenue was 45% this quarter compared to 77% a year ago. From the 2017 second quarter, we recorded an income tax expense of $0.2 million versus an income tax benefit of $14.2 million in the prior year second quarter. Our prior year tax benefit was mostly attributed to the reduction of the deferred tax liability as a result of amortization of NaviNet's purchase accounting intangibles and the conversion of NantHealth from an LLC to a C corporation, just prior to the IPO.

We reported a loss from related party equity investment of $38.9 million. Our loss from equity method investment for the 3 months ended June 30, included a $36 million noncash impairment charge as a result of our termination of the fair value of our investment, NantOmics, and declines to load our carrying value as of June 30. Or, I'm sorry, as of March 31. And that this decline in value was other than temporary.

The decline in the fair value of our investment in NantOmics was primarily caused by delays in our GPS revenue growth from earlier expectations.

On a GAAP basis, net loss was $70.1 million or $0.58 per share compared to $54.1 million or $0.54 per share for the 2016 second quarter.

On a non-GAAP basis, adjusted loss per share was $0.18 compared to $0.15 for the second quarter of last year.

Turning now to the restructuring plan and sale of assets that Patrick mentioned earlier. A combination of the sale of assets and cost reductions in our restructuring plan is expected to have a significant impact on our financial performance, moving forward. As mentioned, annualized cost savings are expected to be in excess of $70 million, inclusive of headcount and the spend reductions. These reductions will be spread across cost of revenue, SG&A and R&D, with a larger impact on R&D and SG&A. Headcount reductions and vendor spend reductions are in process. Regarding our status in sale to Allscripts, we will provide additional information regarding the impact on our operations and balance sheet after the close of the transaction, which is expected in the current quarter.

The consideration we're expecting to receive from this transaction of 15 million shares, represents approximately 12% of our total shares outstanding. Upon completion of the transition, our shares outstanding will be reduced to approximately $106.6 million from $121.9 million.

Turning to the balance sheet. At June 30, 2017, our cash was $92.7 million versus $124.9 million at the end of last quarter. We expect that our restructuring plan will significantly reduce our cash burn going forward, and significantly accelerate our path to profitability.

With that, I will now turn the call over back to Robert.

Thanks, Paul. Operator, we've completed our prepared remarks, and we'd now like to open the call to questions. Operator?

(Operator Instructions) The first question is from Brandon Couillard of Jefferies.

Patrick, with respect to your comments about GPS orders in July, can you give us a sense of what may have changed? Whether you had any single relationship like BofA or cancer treatment centers start to come online? And ramp up orders? And any sense of the relative magnitude of July versus kind of the prior run rate would be helpful.

Yes. Well, I don't think anything specifically other than really the understanding of the people using the test, right? And I think what's exciting to me is, I am getting e-mails directly from the doctors who are using the test, as it affects the outcomes in the decision-making process for the patients. And I think the -- for the first time, people are understanding the word ProteoGenomics, which we put in our press release, I think for the first time there. I think what's happened, even the NCI's now released a new web portal called ProteoGenomics. I think people for the first time understand that, that's where the drugs work and this is where they get actionable information. At ASCO, I think the impact that was made by the doctors at the University of Indiana who presented that in fact, there was an 80% actionable response -- actionable treatment that they could have from the GPS Cancer test. And the remaining 20% was the patients progressed rapidly for them to actually intervene. Really made a real impact, and one people could now begin to understand what's this test is all about. So I can't really point to anyone specific. I mean, that's a good -- there's not one specific, unusual purchase order in the month of July. I can't say however that the month of July, it was bigger than everything we've had, ever, and it speaks well to the run rate. I don't want to really give the number of the run rate at this point, because clearly, I don't want to give forward-looking expectations that we don't really know. I hope that helps, Brandon, but I think the opportunity for people to really understand this test, unlike any other test, speaks to actionable data. The other thing that just come up, and we've not just done a good job in explaining this to doctors. We have a tiny little statement at the bottom of our report that says [provenance]. And I think Illumina has now begun to highlight the concern that, the biopsy at the molecular level result that you may receive as a doctor, actually may in error and inadvertently be of the wrong patient. And I think that has not been fully appreciated, that this issue of -- is this tissue from the right patient? And that has plagued even those to chemistry, where the tissue, because it's sliced across many samples in these high-throughput labs, could be mislabeled and be read for the wrong patient. In fact, that article's been written that 1%, 2.1%, 2% of samples of prostate and breast cancer actually read the wrong patient. But this is even accentuated in the world of molecular contamination. We overcome that uniquely in GPS because we measure the tumor normal, meaning we have the actual patient's normal blood. So the validation that actually this sample is from this patient, is uniquely spelt out through the GPS test. I don't know if all these factors, or it's just a matter of time, because we've not had significant change in reverse, but in clearly, once CMS covers us, which is our hope, that will also change. Does that cover --

You mentioned, the -- no, that's helpful. You mentioned, I think, the number of docs, ordering doctors, 432 in the quarter. I think it's the first time you disclosed that number. Can you give us what that number was in the first quarter, so we have a relative sense of the growth of the user base?

I don't fully recall. It was -- it went from 350 -- approximately 350 to 430.

Again and lastly, I don't think you mentioned the Horizon Blue Cross pilot. What's the -- can you give us an update on the status of that? And when we might see some type of data readout or results from that pilot?

Yes, that's ongoing. And maybe -- let me introduce Sandeep Reddy -- Dr. Sandeep Reddy, who is our Chief Medical Officer, that maybe can speak a little bit to within -- to with the status of that, but that's ongoing right now.

Thank you, Patrick. So Brandon, the reality is that is an ongoing, actively accruing protocol, whereas the patients are actively added week-by-week. There is a large number that we were trying to accrue, about 104 patients. So we're early in the process, it's too early to try to make any claims on what the outcomes will be, but we're very optimistic.

The next question is from Charles Rhyee of Cowen & Company.

Patrick, when we look at the number of tests, sequentially you talk about the ramp up in July. But given that you're about a year in, and I think, in prior quarters you guys have spoken about sort of this order of velocity. And we haven't really quite have seen that yet here. Is there anything to suggest that, why often we should [expect] to see this ramp up? And I guess, the second part of that question is, we have reimbursement from a number of payers. What does the order volume look like in those payer groups? Basically, people that are -- can be reimbursed. Can you talk about sort of what the ordering pattern in that kind of area is?

Right. Well, as you know, in the business, Blue Cross, for example in Philadelphia, it was the first. And I must say I'm disappointed, because it's tough because within that environment, the 4 major academic centers, or the 3 major, depends on how you look at academic centers, have within their own organization, they're own (inaudible). And that's just the reality of where that is. But this is ramping up, and then again, I'll ask Sandeep maybe to give you this -- some indication because he's closer to this than I am, in terms of the actual numbers. But let me give you my take on all of this, in terms of this ordering philosophy. I think there's 3 elements that can affect this. Clearly, and the biggest and most complex element, is people understanding what they're even getting. And as I've said, what is helpful to us now, for the first time the National Cancer Institute has woken up to this concept. It's, oh my goodness, it's not Genomics it's ProteoGenomics. And they called it ProteoGenomics. Secondly is, I'm now through this Department of Defense. There's this thing called the APOLLO program, which is the ProteoGenomics learning system, which we are part. And thirdly, the fact that was presented at ASCO, that is close to 100%, (inaudible) at 20% that didn't progress so rapidly, in 100 consecutive tests, actual data. That fact is not lost, has not been lost, and is now taken, hopefully taken hold. The second fact, obviously, is the reimbursement. That one of the problems with reimbursement, even in the face of reimbursement, it is completing revenue streams that affect the actual academic sentence, albeit on gene panels, without human normal, albeit it's not for the active data. We have to break through that information, all that knowledge of, to -- all the way to the patient, and to all the providers. And then the third thing that I think is going to really help us to get educational is the fact that this test now we've launched in our clinical trial program. As you know, on NantKwest side of it, we've launched the NANT Cancer Vaccine, pancreatic cancer vaccine, the study, combination studies including for the QUILT trial. So if you go on to clinicaltrials.gov and you punch in search Q-U-I-L-T, which is quantitative lifelong trial, you'll begin to see all of a sudden, a [indiscernible] amount of trials pop-up. But what's important isn't -- is these newer trials, has evolved now, will require GPS as a complementary diagnostic. It will serve as an educational tool, as we begin use that. So I think the 3 things is, one deduction of actual data, of getting reimbursed. And once we get ubiquitous investment, hopefully through CMS, that may change the picture. And the third one is, just getting used to seeing the data and using the test, even in the phase of clinical trial. Does that help you, Charles?

Yes it does. As you know, in the release, you also talked about having the discussions with CMS. Can you give us a sense on, so what the time line at this point is, that you'd expect to be able to submit your data and you get a decision?

Yes. So as you know, the CMS has put out not any guidelines, but as you know, they also covered for foundation, lung, et cetera. We've gone into them, met with them face-to-face, had very good conversations. And we think that we took, no not the guidelines, but took the way they did this, so they took, they have what they call these cell lines for which they've done the analysis for which there are absolutely no certainties, the mutations, limitations calling the (inaudible), et cetera. We put it through our process. And most importantly, I don't think people realize that our process is completely machine-read. So post-sequencing, the biosomatics is completely artificially intelligent through machine read. So we then use that process -- that process results into a cap level, but we've never submitted the entire process at the basis of -- for CMS coverage. We've written that up, and then we've also do something different. We do tumor, against normal, unlike any other thing at CMS have seen, and we're putting that all together. We have completed that very detailed, not only validation, documentation, but we put it together at (inaudible). We anticipate that, hopefully, by first quarter 2018, [indiscernible] in response.

Okay. And just one question on the technology side. This asset sale to Allscripts, if I look at it kind of like you're selling the what used to be the CareFx solution, I guess that Harris brought a while back. But if I remember what CareFx was, there's kind of, sort of an interoperability layer. If the reason for the sale that is kind of redundant now, just looking at the other assets that you have? And if not, just curious as to why it makes sense for you to sell that piece?

Yes. Well, good question. So as you saw, when we bought all these different companies, literally, in almost in the year that we were all saying, planning the IPO, we ended up with literally a thousand, close to a thousand engineers across this disparate system. And clearly, it was necessary, first step to see which elements were important, as it relates to a doctor to make a decision and for a doctor to get paid based on that decision, and for us to be able to measure outcomes. And what we ended up with was -- what we just presented GPS Cancer as a test, the Eviti tool as a decision support engine, and the payer providing engagement to enable that, provided to ensure that he's going to be paid, based on that clinical decision. And if we can create this document exchange, and (inaudible) now, maybe information on the outcome. We've since said, the interaction, which is a different interaction of trying to (inaudible) EMR and which is really basically the business of dbMotion. Which is inside the business of Allscripts. And we had a sum total of maybe 11 salespeople, which doesn't make any sense for us to be able to support that business. And since that, therefore, it became very clear to us that, that was what we call non-core to this vision of really pursuing this artificial intelligence layer. We have not presented to Wall Street how deep and how far we are, with regard to what we talk about predictive modeling and artificial intelligence. We have 5 computer years of outcomes data based on the Genomics Proteomics analysis. We've built thousands and thousands of predictive models based on outcomes. And we thought this is the best use, not only of our resources, but allows them -- the company to be focused purely on software as it relates to a decision support, it relates to supporting the artificial intelligence platform, and this little ware, that actually talks to EMRs was merely a tool and not really [core]. And it was dead in the hands of an organization with hundreds of salespeople. Calling on customers during the [indiscernible] implementations.

The next question is from Joe Munda of First Analysis.

Can you hear me, okay? Patrick, just real quick couple of questions. I'm just taking a look at the last couple of quarters. Here's the number of GPS Cancers delivered. First quarter, roughly 266, this quarter 264. My question really centers on the capacity to deliver more tests, and as well as, I know why your fleet of sequencers is built around the high seek action. I was wondering if you had converted any of those over to [de novo seq] to expand capacity. Any thoughts there? Can you deliver more than 260 essentially a quarter?

The answer to that -- first of all, Joe, thanks for the question, good question. We actually have the capacity not only to deliver, we actually have the [de novo seq] running, validated already. And we have, more importantly, the bioinformatic infrastructure. And I think this is what was underestimation, if that's an English word, by that many people, of the capacity to deliver the bioinformatic sense of scale. We quite literally now have the capacity, and are actually moving something like, petabytes, maybe 1 to 2 petabytes of data a week from our system into our supercomputing network. So to give you some insight into scale, I don't know if you see or saw the recent report by IBM, where they compared themselves to New York Genome Center. And what they did was they took one glioblastoma, and the results of that glioblastoma was already sequenced, and the mutation was already called, and then they took those data and used that data to integrate it, and match it and map it, with drugs related to the whole genomic mutation calling, on 1 patient, and they were proud to say, it took them, IBM, 10 minutes. And I think Human Genome, this Genomic Group, the human portion of it, 160 hours. That is 1 patient, which doesn't even involve the actual computation analysis of mutation calls itself. I would probably (inaudible). What we are able to do is to take directly out of the sequencing machine, the whole Genome sequence, and the tumor and the norm and machinery not only mutation calls (inaudible) but all the way to the mapping, to the drug. And we have the capabilities of doing this to the tunes of hundreds a month. So I think that was the infrastructure, bioinformatic infrastructure, we spent 5 to 8 years building. And I don't need to recall, many years ago, we said, we could do that in a single patient, when you do thousands of patients a week. We could do that, if you did the mathematics, it turns out to 42 seconds, per patient, bioinformatically. So that's where we are. Now we're closed to 20 seconds, bioinformatically. If you've gotten the same data, the simple data of the tumor normal, and map it to the drugs. I was joking, but maybe not half joking, that would be a wonderful Genomics jeopardy, if we took 10 tumor samples, tumor normals, of the sequence data directly out of the machine, and put a side-by-side, ourselves against what's and against any human analysis. First, just to get the imitation callings, and then second to the mapping, to the drugs. And I think then, maybe, people will understand the enormity of the bioinformatic and artificial intelligence platform that was needed to accomplish this kind of scale. We plan on this scale, recognizing this 1.6 million used diagnosis cancer patients a year, just in the United States. And 13 million survivors. Search 15 million people. And unless you actually plan for that kind of scale, I think, you would not really be possible to say that we want to provide this information at point of care. So thank you for that question, because not only have we planned for that scale bioinformatically, our next challenge is to plan for that scale on the basis of sequencing. So that's why we brought in the novo machine, and are already are running validated samples.

Okay. And then, in terms of -- that's helpful. In terms of the sales force devoted to GPS, can you give us some sense of where or what number that is today? Or at the end of the quarter, I'm sorry?

Well, we're increasing it, it's not large when we started. I can tell you, we more than doubled it now. And we're clearly going to win the teams now. We clearly need to increase it. I think as I said, I wanted to introduce Dr. Reddy, who's really going to be instrumental and helping us now grow specific to the GPS business. As you know, he was the Chief of Staff at Los Alamitos and he was adjunct professor of UCLA and formerly of Caris, and now he's here. And I think he will really help to give run locks, to drive this business, by increasing our sales force and then increasing the quality of the sales team, who have real knowledge about cancer and cancer doctors.

Okay. And then, I guess, one follow-up question on the asset sale. You said, it's non-core. I'm just curious, how much of that sale or how much of the cost savings of the $70 million annualized is devoted to -- from that sale?

I don't think we have -- we can break this out because Joe, it was all integrated with regard to some of the synergies. And we also, obviously, did a bit of headcount reduction as a consequence of not having to support that critical product. But I think, in aggregate, maybe a little later on, Paul until the deal closes, would really have a good idea, can break out how this $70 million savings will have cross R&D operations and G&A. But it's hard for me to break that out specifically for you now.

Thank you. And at this time, there's no further questions. I turn the call back over to Robert Jaffe for closing remarks.

Thanks, Operator. We look forward to sharing our progress on our next scheduled conference call in November. Thank you all for joining us today.

Thank you. Ladies and gentlemen, this concludes today's conference. You may now disconnect. Good day.