Before we begin, I would like to remind you that this conference call contains forward-looking statements within the meanings of Section 21E of the Securities Exchange Act of 1934, as amended and as defined in the US Private Securities Litigation Reform Act of 1995. These forward-looking statements can be identified by terminology such as will, expect, anticipate, future, intends, plans, believes, estimates, target, confident, and similar statements.
Statements that are not historical facts, including statements about Burning Rock's beliefs and expectations, are forward-looking statements. Such statements are based upon management's current expectations and current market operating conditions and relate to events that involve known or unknown risks, uncertainties and other factors, all of which are difficult to predict and many of which are beyond Burning Rock's control. Statements.
This is my turn? Hello? Hey, this is Yusheng Han from Burning Rock. I'm the CEO and founder, and today you also have our CFO, Leo Li, and our CTO, Joe Zhang, online. Today we have a brief introduction of recent progress, and then we'll -- I will hand out to Leo, talking about the financials, and then Joe talking about our pipeline progress. So, let's turn to page three, which shows that what Burning Rock is doing. We started from therapy selection and expect expanded to early detection, MRD, and biopharma business. So far, that is, that's our business construction.
Let's turn to page 4, which is the page that I think most of the investors, care about most, that's about breakeven. We set a goal to breakeven in terms of non-GAAP profit minus SG&A. We say that in second quarter this year, that's the first time in Burning Rock to reach the goal. Well, in third quarter, actually, there is some interesting industry disruption. You know that even most of the medical conferences or meetings were held abnormally. So, in third quarter, we were a little bit impacted by this event.
The profit dropped to the negative part, which is a negative, sorry, 8.9 million CNY. But we are still moving to the breakeven level. Especially, I think that this kind of volatility will pass by the end of this year. So let's turn to page 5, that we set a goal to breakeven sometime this year. And we think that we are moving toward that direction very firmly and without the disruption.
In terms of the therapy selection, despite of the industry disruption, we still continue to growth in for the in-hospital model, which means that the in-hospital revenues has a 10% year-over-year growth. And the part that has been impacted was the central lab model. In MRD, we have a strong clinical validation publications, which is with the MEDAL study for lung cancer published in Cancer Cell, which is a big milestone for our MRD study. And we know that the other studies, for example, like, like colon cancer and esophageal cancer, are on the way.
For biopharma, despite of the, you know, the struggling time of capital market for biopharmas, we still have a strong growth showing our systematic value of Burning Rock, with revenue growth of 31 year-on-year. And our backlogs still keeps growing. And one thing to mention is that we just signed a CDx contract with BI. And then in terms of early detection, there is a big step, or I mean, a big stone - a milestone, for that is that we got a breakthrough device designation granted by China National Medical Products Administration, which is NMPA.
And we are the only, the, our multi-cancer early detection product is the only test that has received a BDD from both FDA and NMPA. And, let's turn to page six. That explain that, how the industry volatility impact our volume. You can see that the central lab model has been impacted negative 31%, while for for in-hospital in-hospital model is still growing in terms of volume. So, we think that, yeah, and that actually impacts our competitors much more than Burning Rock, because Burning Rock is the only company that in-hospital model represents more than half of our revenues. And then I'll turn to Leo about our financials.
Move on to page 7. As you mentioned earlier, the whole China healthcare industry had a disruption during the quarter. And what that meant for us was actually two divergent trends, which actually accelerated the path that we were already on.
So if you look at central lab, that was down heavily, down 41% on a year-over-year basis in the third quarter. That channel continued to grow, and it grew 10% year-over-year, which is rare in the diagnostics industry or at least in our specialty industry in China. So I think that continues to prove the value of the in-hospital segment, where the real value or the profit is. And central lab, as we mentioned earlier, is being shifted more towards in-hospital.
So I think the events in the third quarter only accelerated that, and that moved us even more on the right track in terms of moving the mainstream of our businesses towards the in-hospital segment. So at some point last year, we were in more in-hospital segments by volume, and you can see in the third quarter, we are actually in-hospital represented the largest segment, overtaking central lab.
So I think with that transition complete, at some point down the road, our volume and revenue trends will match again as we complete this transition. Then moving on, we can see that pharma services still had a strong growth quarter in the third quarter. Revenues grew 31% on a year-over-year basis, so we're very pleased with that result.
Our backlog continued to grow, particularly from multinational companies, and Yusheng gave an example of a recent win in his remarks. So we're very pleased with the progress and the outlook that we have in the sector.
Overall, because of the industry impacts and a drop in central lab, our revenue was down 17% on a year-over-year basis, and we're, we're very conscious of this trend, and we are managing our expense, or our cost base, appropriately in accordance to the new industry setup. So I think we had a, we had a temporary dip in our operating margins for the third quarter, and that should turn for the better in the quarters down the road.
So measured on a Non-GAAP basis, by gross profit, minus SG&A, I mean, we hit a positive territory in the second quarter, but we didn't in the third quarter, and we're looking to turn this to a positive number again at some quarter down the road. So we're working towards that, and we have done some rejigging in the recent time period to make sure that we're on the right track. So our operating loss did widen a little bit in the third quarter compared to the second quarter, and we're aware of that. We've made the adjustments towards the end of September, then we hope to be on a better track going forward.
Notably, we are also managing our operating cash flows, we're managing our expenses, our cash expenses very carefully. And you can see the net operating cash outflow this quarter has dropped to CNY 47 million per quarter.
And if you look at our operating margins, if you look at our sales and marketing expenses as a percent of revenue, despite the drop in revenues, I mean, our sales and marketing expenses were 45% of our revenues, similar to 44% that we achieved in the second quarter. And as we think industry volume normalizes down the road, our operating margins should improve, and we hope to get back to the positive Non-GAAP territory in the quarters down the road.
So that's a quick recap of the PNL for the quarter. Then I think we should also mention our cash position, as that's been a focus, if we go to page 8. So we had cash outflow of about CNY 532 million in the year of 2022. That's not our run rate for 2023. Our guidance at the start of this year was about CNY 400 million outflow, and you can see in the three quarters so far to this year, we had an outflow of about CNY 249 million. So significantly below the outflow guidance that we set out at the start of this year.
If you look at our third quarter quarterly cash operating cash outflow, 47 million CNY, that's even close to the run rate that we set for the year over 2024. So on that run rate, and looking at the cash balance at the end of this period of 637 million CNY, we're sitting on more than three years of cash runway. So we are in no rush to do anything.
We're sitting on ample cash balance, and I just want to reiterate our strong cash position relative to our cash outflow on this page. And the reasons for the reduced cash outflow, I think we have mentioned that before, and I'll just reiterate those here again. So first, our commercial operations is coming towards profitability.
We were even slightly positive in the second quarter. Our R&D spend, specifically our early cancer detection programs, are maturing. As these programs complete, the expenses associated with those programs will run off, and that will help reduce cash spend naturally.
So, you know, as we look to complete our spend on early cancer detection and as we move towards better profitability, then we'll look to improve our operating cash outflows, and we make sure that we're sitting on ample cash balance. The next, I'd like to turn to Joe, who has some important pipeline and clinical publication data to share with you all.
Thanks, Leo. So, so, I'm going to present, a little bit, give you guys an introduction about the, the pipeline update. Majorly focus on the, MRD, which is minimal residual disease or molecular residual disease. So let's turn to, page 10. So page 10, basically, representing, presenting the Burning Rock MRD clinical publication. So basically, MRD has a lot of utilities, as shown in the picture showing here, basically, it can be done, like, before the new adjuvant to look at the, baseline ctDNA level. Then also we can do, like, after new adjuvant therapy to look at the treatment effectiveness.
Or like, more commonly, going to be used as, like, a post-surgical, after resection, to look at a landmark MRD to get a prognosis or, either some kind of a prediction value on the MR, based on the MRD status. There also could be a lot of these, a lot of treatment effectiveness assessment after that adjuvant therapy, and also the longitudinal monitoring for the surveillance.
So, Burning Rock has a bunch of different kind of publication, related to different kind of cancer type. That's including the non-small cell lung cancer, as well as the colorectal cancer, gastric cancer, pancreatic cancer, and BTC, biliary tract cancer.
So, all the publication, most of them are actually showing in the poster format, being presented in a different kind of academic meeting or clinical meeting, happened like within two years. So the major one, I just want to give you guys a little bit more introduction about the Cancer Cell publication, which is related to the non-small cell lung cancer. Let's turn to page 11.
So, the technology we've been using actually called brPROPHET. In brief, they call Prophet. So basically, it's a based on a whole exome sequencing. We got the tumor whole exome sequencing, get up to 50 mutation, which is a tumor specific.
Then we design the personalized panel, MRD panel, and use this personalized panel, trying to capture a potential ctDNA fragment from the plasma collected from the same patient. Based on the proprietary ultra deep sequencing and also, excuse me, the proprietary sequencing results, as well as the MRD coding algorithm, then we can determine the MRD status of that patient for that time point of blood collection.
So, on the right page, basically showing the analytical performance of this brPROPHET assay. So, it's including the two type of... So, on the top right panel, basically talking about uncontrived cell line sample diluted down to 8 ppm.
As you can see here, we still can see, out of 50 loci , there's quite a bit different, significant difference compared to the baseline, which is uncontrived, which is background cell line. So this give us some kind of confidence showing this assay is sensitive enough to detect a very low allele frequency, very rare tumor fraction, based on the on the ctDNA from the patient.
On the on the top right bottom right panel, that's showing the quantitative property of this assay. Based on the the algorithm we used, we can estimate, based on the detection sensitivity, detect capability, as well as the allele frequency of each locus , we can assess estimate what's the ctDNA fraction from that patient.
As you can see here, this is contrived data, but it definitely give us a lot of confidence showing the expectation and the estimation, showing very good correlation. So based on this technology, we move to page 12. Basically, we work with a top-tier hospital in Beijing People's Hospital, and we published this technology utilization on non-small cell lung cancer in Cancer Cell, which is a top journal, top-tier journal, for translational medicine.
As you can see here, so this is being published in October 9. And there's a couple of highlights. I don't want to read it one by one, but you can look at it. Mostly, it's just showing the brPROPHET outperformed the fixed-panel MRD assay in a head-to-head comparison in non-small cell lung cancer.
Move to page 13, basically give you an overview of this study. So, the cohort is that we enrolled about 181 patients non-small cell lung cancer. But as you can see here, most of them are actually 63% are Stage 1 patients, very early-stage patients after surgery. And also, there's a few Stage 2 and Stage 3. And the sampling time is, we basically collect the tumor, adjacent pair of tissue, normal tissue, collect at surgery, and we do the whole exome sequencing on those. As well as we collect a blood sample, collect the preoperative, and three days and 30 days after surgery.
Also, we do a follow-up time, so at every time one patient go back to see the doctor, after, like, six months after or a year after, we, if possible, collect the follow-up blood sample. Then we using three different kind of approach to look at MRD status. The first one is basically showing on the top right. It's a whole exome sequencing based personalized panel design, as well as doing this kind of MRD assay we call brPROPHET assay. As comparison, we're also using also a fixed panel target sequencing to do the tumor either tumor-agnostic or tumor-informed calling to determine the MRD assay.
But you can think about it this way, so basically, whole exome sequencing give us many more potential mutation loci compared to relatively smaller fixed panel target sequencing. So by using this data, let's move to page 14. Basically, there are several major conclusion observation we have seen. So the page 14 showing actually, if you look at the preoperative plasma sample, since you know, this sample is basically blood coming from untreated patient.
And that's why, ideally, if your assay is perfect, you should be able to see every patient blood will get, like, close to very high percentage of detection sensitivity. As you can see here, the sensitivity definitely grow up from Stage 1A till Stage 3.
At Stage 3, you get a higher sensitivity, detection sensitivity. But, as you can see here, the orange-ish color, which is representing the profit assay. So, it outperforms the tumor-agnostic panel sequencing, as well as the tumor-informed fixed panel sequencing. So, and totally, basically, at the Stage 2B, we already see 40% detectability, and also Stage 2, we see 75%, Stage 3 we see 83% positivity.
As you can see here in the panel B, showing on the right page, as you can see here, it's for all three MRD positive patient sample, the cfDNA, ctDNA fraction showing the highest one, which is showing in the box plot with a red background.
But if you look at the orange color, which means on the right, there's a 30 patient, preoperative patient, only showing a positive in brPROPHET assay. But the ctDNA fraction is way lower. It's actually two magnitudes lower than the all three positive patient.
This is just to reflect the detection sensitivity importance, trying to get those kind of, you know, very low allele frequency, very low tumor fraction patient, trying to confirm the MRD status. Of course, this is a preoperative. So let's move to the page 15. Basically, we use a postoperative blood sample to determine the real MRD status and also come associate this kind of a status with the relapse potential, to look at the disease-free survival.
On the left page, basically, as we are using the landmark time point, which is three days or three... Which is time point B, or 30 days after surgery, which is time point C, to look at the survival curve. As you can see, the DFS survival percentage, if your MRD negative, the patient usually showing the even from one time point, landmark time point to track. And the follow-up up to, like, 1,200 days, you still show, show very way higher disease-free survival compared to MRD positive patient, which that's showing the HR ratio reach to 5, basically, for time point C, the HR hazard ratio reach to 16.4, which is a pretty significant difference.
On the right side of the panel, we're basically utilizing longitudinal MRD analysis. This is basically for multiple point, time point assessment on the post-surgery patient plasma sample. If there's any MRD positive signal showing any time of the blood collection, we deem as MRD positive. But if all the sample collected from our patient is MRD negative, then we deem it MRD negative. As you can see here, the separation will be even better. That just give us a lot of confidence, also reflect a lot of other publications.
The continuous surveillance require multiple time point collection usually give us better separation of the disease-free survival. So basically, it's also now related to the Stage 1 or Stage 2 and 3, basically showing very similar trend.
So, in summary, basically, this paper published in Cancer Cell gave us very good example, showing how the personalized WES-based MRD assay can give a very good prognosis value on the non-small cell lung cancer, even at the Stage 1 and also 2 and 3.
And, so, of course, we are still keep working on this assay and trying to working on multiple different kind of cancer type, and also trying to, with other top-tier hospital, on the not only prognosis value, we also want to look at the predictive value and to see any kind of clinical utility related to drug selection or any kind of treatment effectiveness assessment. So that conclude my part. Thank you. It's back to the moderator?
Yeah. Thank you for participating in today's call. You may now disconnect. Everyone, have a great day.
Thank you.