RDNT (2025 - Q2)

Good day, and welcome to RadNet, Inc. Second Quarter 2025 Financial Results Conference Call. [Operator Instructions] Please note this event is being recorded. I would now like to turn the conference over to Mr. Mark Stolper, Executive Vice President and Chief Financial Officer of RadNet. Please go ahead. Mark D.

Thank you. Good morning, ladies and gentlemen, and thank you for joining Dr. Howard Berger and me today to discuss RadNet's second quarter 2025 financial results. Before we begin today, we'd like to remind everyone of the safe harbor statement under the Private Securities Litigation Reform Act of 1995. This presentation contains forward-looking statements within the meaning of the U.S. Private Securities Litigation Reform Act of 1995. Specifically, statements concerning anticipated future financial and operating performance, RadNet's ability to continue to grow the business by generating patient referrals and contracts with radiology practices, recruiting and retaining technologists, receiving third-party reimbursement for diagnostic imaging services, successfully integrating acquired operations, generating revenue and adjusted EBITDA for the acquired operations as estimated, among others, are forward-looking statements within the meaning of the safe harbor. Forward-looking statements are based on management's current preliminary expectations and are subject to risks and uncertainties, which may cause RadNet's actual results to differ materially from the statements contained herein. These risks and uncertainties include those risks set forth in RadNet's reports filed with the SEC from time to time, including RadNet's annual report on Form 10- K for the year ended December 31, 2024. Undue reliance should not be placed on forward-looking statements, especially guidance on future financial performance, which speaks only as of the date it is made. RadNet undertakes no obligation to update publicly any forward-looking statements to reflect new information, events or circumstances after the date they were made or to reflect the occurrence of unanticipated events. And with that, I'd like to turn the call over to Dr. Berger.

Thank you, Mark. Good morning, everyone, and thank you for joining us today. On today's call, Mark and I plan to provide you with highlights from our second quarter 2025 results, give you more insight into factors which affected this performance and discuss our future strategy. After our prepared remarks, we will open the call to your questions. I'd like to thank all of you for your interest in our company and for dedicating a portion of your day to participate in our conference call this morning. Let's begin. I am very pleased with the performance in the second quarter. After experiencing significant disruption in the business during the first quarter of this year as a result of the California wildfires, and severe winter weather conditions in the Northeast and Houston markets, the second -- in the second quarter, business rebounded, and we achieved record quarterly revenue and adjusted EBITDA. Relative to last year's second quarter, total company revenue increased 8.4% to a quarterly record $498.2 million, and Digital Health segment revenue increased 30.9% to a quarterly record of $20.7 million. Contributing to core imaging center revenue growth, for a variety of factors. First, industry trends continue to provide tailwinds. Imaging technology advances in equipment, post-processing software, artificial intelligence, contrast materials and nuclear isotopes continue to drive increased utilization of diagnostic imaging within health care in general. Furthermore, within this growing industry, the shift of procedural volumes away from the more expensive hospital alternatives, to more cost-effective ambulatory freestanding centers continues. Second, improvement continues in reimbursement rates with commercial and capitated payers that recognize the position RadNet offers as a lower-priced alternative to hospital-based imaging. To this end, we have been successful in receiving rate increases from many of the larger commercial payors and several capitated contracts have been converted to higher paying fee-for-service relationships. Lastly and perhaps most importantly, the focus has been on driving more advanced imaging procedures. MRI, CT and PET/CT, and increasing advanced imaging capacity through a variety of initiatives. Advanced imaging has a higher per procedure pricing and typically better margins. During the second quarter of this year, advanced imaging as a percentage of total procedures increased to 27.5%, from 26.5% in last year's same quarter, an improvement of 102 basis points. This increase is due in part to initiatives that have been identified within each of these modalities. For example, within MRI, the 9% aggregate and 6.6% same-center growth in the second quarter, as compared with last year's second quarter, is partially the result of capacity created from investments made in MRI software upgrades and operating protocols, which enable shorter scan times. The shorter scan times allow for the scheduling of more patients in the same hours of operation. Within CT, programs have been expanded on both coasts to offer more complex procedures. An example of this is cardiac CT angiography, which is growing rapidly and which, in some cases, are enhanced with reimbursed artificial intelligence-assisted analytics. Within PET/CT, emphasis has been on newer diagnostic and screening offerings for prostate cancer, Alzheimer's disease, dementia and new procedures with leading-edge tumor-specific radioactive tracer. PET/CT has been the fastest-growing procedure. This quarter, PET/CT increased 22.4% on an aggregate basis, and 16.2% on a same-center basis as compared with last year's second quarter. The increase in advanced imaging, particularly MRI, has also been driven by the implementation of TechLive, our remote screening technology recently cleared by the FDA. TechLive is a vendor-agnostic integrated solution, enabling remote scanning of MRI, CT, PET/CT and ultrasound procedures. Amidst tech labor shortages and inflationary wage pressure, TechLive empowers technologists to scan for multiple locations, enables improved operational efficiency, extends center operating hours and enhances access to complex procedures. The most significant impact we are experiencing with TechLive is its ability to expand hours of operation by staffing exam rooms, which previously would have been closed. As an example, in a pilot deployment at 64 locations inside of RadNet's New York area facilities, TechLive significantly contributed to a 42% decrease in MRI room closures and during the second quarter of 2025, as compared with the same period of 2024. Currently, more than 300 of RadNet's MR, CT, PET/CT, and ultrasound systems are connected with DeepHealth's Tech Live solution, and are targeting to substantially all of RadNet's advanced imaging equipment to be connected with TechLive in early 2026. The strong revenue growth from all the factors just discussed and in particular, the initiatives driving more advanced imaging, along with cost-effective management, contributed to the record adjusted EBITDA and margin expansion in the quarter. Adjusted EBITDA during the second quarter of 2025 increased by 12.3% to a quarterly record of $81.2 million, up from $72.3 million in last year's second quarter, and adjusted EBITDA margin increased to 16.3% during the second quarter of 2025, which compares with 15.7% in last year's second quarter, an improvement of almost 60 basis points. The strong operating results in the second quarter relative to our internal budget resulted in the decision to increase 2025 full year guidance ranges for revenue and adjusted EBITDA. Mark will discuss this in more detail in his prepared remarks. Steady progress also continues in the Digital Health operating segment. The EBCD DeepHealth, AI-powered breast cancer screening program continues to expand. Currently, we are experiencing a blended adoption rate nationally, approaching 45%, with more cancers being found across RadNet centers, which otherwise might have been detected at a much later date. On July 17, the previously announced acquisition of iCAD, a global leader in clinically proven AI-powered breast health solutions was completed. iCAD's ProFound Breast Health Suite, and RadNet's DeepHealth AI-powered screening solutions, together can materially expand and improve patient diagnosis and outcomes on a global basis through further enabling accuracy and early detection. With over 1,500 health care provider locations, facilitating over 8 million annual mammograms in 50 countries, iCAD's installed base and strong sales, engineering and marketing capabilities will provide immediate broad and valuable customer relationships and commercialization capabilities that can accelerate DeepHealth's objectives. On June 4, the acquisition of See-Mode Technologies, a global innovator in AI for ultrasound imaging was completed. See-Mode's initial applications to detect and characterize thyroid nodules and breast lesions in ultrasound imaging, improve diagnostic accuracy and enhance clinical workflows. With the inherent complexity of ultrasound imaging and its dependency on the individual capabilities of the technologists and radiologists, the opportunity to improve care through AI is significant. With demand exceeding available appointment slots for many of the 900 ultrasound units in 326 of our locations, the increase in capacity created by See-Mode's technology should improve our ability to drive better access and more revenue through RadNet's existing centers. Early deployment of See-Mode's FDA-approved thyroid ultrasound artificial intelligence across 83 of the imaging centers has demonstrated up to 30% reduction in scan time and it is anticipated that See-Mode will be fully implemented in the remaining centers by the end of the first quarter of 2026. Furthermore, our reimbursement code already exists that makes a portion of our approximately 250,000 annual thyroid ultrasounds eligible for additional reimbursement. An initiative is ongoing to pursue FDA approval for See- Mode's next application in breast AI ultrasound, which constitutes over 600,000 of RadNet's approximately 2.7 annually ultrasound exams performed. While initial focus will be on the implementation within RadNet, these technologies will also be sold and marketed by the Digital Health division to third parties as the offerings are further commercialized. Finally, financial liquidity and leverage continues to be carefully managed. As of June 30, 2025, our cash balance was $833 million, and net debt to adjusted EBITDA ratio was 0.96. An attractive pipeline of acquisition opportunities are being evaluated for both the core imaging services division and for Digital Health, and we have confidence in our ability to invest the cash balance over time and opportunities that advance RadNet's strategic objectives. At this time, I'd like to turn the call back over to Mark to discuss some of the highlights of our second quarter 2025 performance. When he is finished, I will make some closing remarks.

Thank you, Howard. I'm now going to briefly review our second quarter 2025 performance and attempt to highlight what I believe to be some material items. I will also give some further explanation of certain items in our financial statements as well as provide some insights in some of the metrics that drove our second quarter performance. I will also provide an update to 2025 financial guidance levels, which were released in conjunction with our 2024 year-end results in February, and amended following our first quarter financial results in May. In my discussion, I will use the term adjusted EBITDA, which is a non-GAAP financial measure. The company defines adjusted EBITDA as earnings before interest, taxes, depreciation and amortization, and excludes losses or gains on the disposal of equipment, other income or loss, loss on debt extinguishments and noncash equity compensation. Adjusted EBITDA includes equity and earnings, and unconsolidated operations, and subtracts allocations of earnings to noncontrolling interest in subsidiaries, and is adjusted for noncash or extraordinary and onetime items taking place during the period. A full quantitative reconciliation of adjusted EBITDA to net income or loss attributable to RadNet income and shareholders is included in our earnings release. With that said, I'd now like to review our second quarter 2025 results. As Dr. Berger highlighted in his remarks, our business bounced back in the second quarter nicely, recovering to more anticipated levels following the California wildfires and severe winter weather conditions in the Northeast that significantly impacted us in the first quarter of this year. In the second quarter, we returned to the type of growth we have been demonstrating over the last several years. Our results were highlighted by strong performance and growth in advanced imaging, which is being driven by many of the initiatives that Dr. Berger discussed. Advanced imaging grew 9% in aggregate, and 6.6% on a same-center basis relative to last year's second quarter. PET/CT, which continues to be our fastest-growing modality grew 22.4% in aggregate, and 16.2% on a same-center basis, predominantly on the growth of PSMA and amyloid brain studies. The disproportional growth in advanced imaging relative to routine imaging has been a steady trend and continues to help us absorb the inflationary pressure we and the rest of the industry have been feeling with respect to the availability and rising cost of labor, especially as it pertains to radiology technologists. We believe there is continued opportunities for margin improvement as a result of driving more advanced imaging in our centers, and through the implementation of many of the software tools offered by the Digital Health division. During the quarter, we opened one new facility in New Brunswick, New Jersey and have 9 additional facilities that we are targeting to open by the end of the year, which includes 3 joint venture facilities and 6 wholly owned locations. These de novo openings will give us the necessary capacity to support the heavy diagnostic imaging demand in our markets and should enable us to continue similar organic growth into the future. For the second quarter of 2025, RadNet reported total company revenue of $498.2 million and adjusted EBITDA of $81.2 million, both quarterly records. Revenue increased $38.5 million, or 8.4%, and adjusted EBITDA increased $8.9 million, or 12.3%, as compared with the second quarter of 2024. The combination of the strong top line growth and our ability to manage operating costs effectively caused EBITDA margins to improve by 57 basis points relative to last year's second quarter. The Digital Health segment reported revenue of $20.7 million, a 30.9% increase from last year's second quarter. Breaking this down further, AI revenue within Digital Health increased 21.6%, from a combination of growing the EBCD program revenue, and through expanded licensing of DeepHealth lung, prostate and neuro solutions, primarily in Europe. Radiology software revenue within Digital Health grew 36.1% relative to last year's second quarter, from a combination of internal revenue from RadNet's imaging centers adopting further elements of DeepHealth OS, including certain contact center software and TechLive, as well as from external sales of workflow software solutions. We continue to make important investments in sales, marketing, development, customer support and implementation teams necessary to support anticipated growth over the next 5 years. Despite these investments in infrastructure related operating expenses EBITDA for Digital Health grew 4.1% over last year's second quarter. As Dr. Berger mentioned, we finished the second quarter of 2025 with a strong cash and liquidity position. At quarter end, we had $833 million of cash on the balance sheet, full availability of a $282 million credit facility, and a term loan that is priced at SOFR plus 225 basis points, reflective of the refinancing transaction we completed last April, and the repricing transaction we completed last November. Continued improvements in revenue cycle have driven down DSOs, or days sales outstanding, to 32.4 days, slightly lower than where we were at this time last year. With regards to our financial leverage as of March 31, 2025, unadjusted for bond and term loan discounts, we had $264.6 million of net debt which is our total debt at par value less our cash balance. Note that this debt includes RadNet's ownership percentage of New Jersey Imaging Network's net debt of $36.2 million for which RadNet is neither a borrower nor a guarantor. At quarter end, our net debt to adjusted EBITDA leverage was slightly less than 1x. Given the strong second quarter results and the positive trends we continue to experience, we elected to increase revenue and adjusted EBITDA guidance for our Imaging Center business. We increased revenue by $15 million at the low and high ends of the guidance ranges, and increased adjusted EBITDA by $3 million at the low and high end of the ranges. We also increased our capital expenditure guidance range by $7 million, which is reflective of additional growth investment opportunities we plan to pursue in the second half of the year. Otherwise, all guidance ranges for Imaging Center segment remain unchanged. For Digital Health, we plan to update guidance ranges upon announcing our third quarter results in November. This update will reflect contributions from the iCAD and See-Mode transactions, and incorporate any other information about the operating segment we have at that time. At that time, we should have a much better sense as to how the integration of iCAD and See-Mode is progressing as well as each's revenue contribution for the second half of the year. I'll now take a few minutes to give you an update on 2026 anticipated Medicare reimbursement rates. As a reminder, Medicare represents about 23% of our business mix. With respect to Medicare reimbursement, several weeks ago, we received a matrix for proposed rates by CPT code, which is typically part of the physician fee schedule proposal that is released about this time every year. We have completed an initial analysis and compared those proposed rates to our current 2025 rates, and we volume weighted our analysis using expected 2026 procedure volumes. In the proposed rule, Medicare is proposing increasing the conversion factor in the Medicare fee schedule by 3.3%, from $32.35 to $33.42, along with certain changes to the RVUs, or relative value units, of specific radiology CPT procedure codes, and to the Medicare geographic practice cost indices for GPCIs. Our initial analysis of all these moving parts of the proposal indicates that RadNet on roughly $1.9 billion in revenue will benefit from an approximately $4 million to $5 million Medicare revenue uplift in 2026. We are very pleased that we will likely be getting this modest increase next year after about 5 years of annual cuts to Medicare reimbursement, including in 2025, when we have absorbed and are continuing to absorb about a $7 million to $8 million cut. We hope that this is a recognition from CMS that it must be -- must compensate providers appropriately, and that its reimbursement should be commensurate with the rising cost of providing services. The Medicare physician fee final rule is expected to be released on or about November 1, 2025. There is no assurance that the final rule will be consistent with this proposal. On our third quarter financial results call in November, we hope to be able to provide more certainty around 2026 Medicare rates. I'd now like to turn the call back over to Dr. Berger, who will make some closing remarks.

Thank you, Mark. The diagnostic imaging industry has entered a period of transformation. Traditionally, diagnostic imaging has relied primarily on manual processes and labor to complete most aspects of services. The manual nature of the industry has historically leaned heavily on the availability of skilled labor for all clinical aspects of the patient journey, as well as for performing all the requisite non-patient facing support functions, including scheduling, pre-authorization, insurance verification, revenue cycle, coding, just to name a few. More recently, quantum leaps in computing power, machine learning, and artificial intelligence have demonstrated the possibilities for the future of the industry. Having under one roof both the largest scale imaging services business in the United States, and a leading radiology-focused digital health segment that is building innovative workflow and clinical solutions, RadNet is at the nexus of this industry's transformation. Over the past 5 years, RadNet has been investing in both clinical and operational artificial intelligence. Within clinical, or predictive AI, the focus of our investments and development efforts are in 2 areas. The first is in the development of interpretive AI solutions focused on population health screening. Today, the diagnostic imaging industry principally performs procedures on patients who are already symptomatic, sick or injured. In contrast, the solutions being developed in breast, lung and prostate are designed to lower the cost and increase the effectiveness of diagnostic imaging-based screening tools to detect diseases earlier when better patient outcomes are achievable. Launching the EBCD AI-powered breast screening program, working with the NHS in the U.K. to implement the targeted lung health check program, and beginning an MRI-based AI-powered prostate screening program, are examples of what is possible in diagnostic imaging-based population health. The second area of RadNet's clinical AI and technology investment is in AI tools that can make the clinical staff, technologists and radiologists more productive. Two examples of these are the recent purchase of See-Mode and the recently FDA-cleared TechLive solution. In regards to See-Mode, the technology eliminates certain manual processes of technologists and radiologists in conjunction with commonly performed thyroid and breast ultrasound exam, thereby increasing center-level productivity by reducing scanning times and creating valuable scanning capacity. TechLive, as we discussed earlier, reduces exam room closures from insufficient staffing, and allows for the expansion of hours of operations, thereby also helping to create screening capacity -- excuse me, scanning capacity. Within operational or generative AI, the focus has been on developing DeepHealth's OS, a comprehensive end-to-end cloud- native operations and image management solution designed to automate many of the manual functions that are currently performed throughout the diagnostic imaging workflow. RadNet's performance in the second quarter reflected some of the early benefits from the strategy -- from these strategic technological initiatives and investment in both clinical and operational AI. During the quarter, revenue was enhanced from the clinical investment in screening programs from breast, lung and prostate, and from the early deployment of TechLive. In addition, costs were reduced and margins were improved from some of the early implementation of certain DeepHealth OS modules at selected RadNet centers. In the coming quarters, we will continue to expand all of these programs within RadNet centers, and we'll be more actively marketing the Digital Health solutions through external customers, including some of our existing health system joint venture partners, and current customers of our legacy eRAD software solutions. We look forward to updating you on our progress in the coming months. Operator, we are now ready for the question-and-answer portion of the call.

[Operator Instructions] The first question comes from David MacDonald with Truist.

Just a couple of questions. First, can you guys just talk a little bit about -- once all of this is kind of fully rolled out in 2026 between the capacity you're adding with de novos, with the TechLives, and See-Modes of the world, and also just the improvement in the underlying technology, quicker throughput. Can you give us some sense, or just talk in more detail about either how much incremental capacity you anticipate having, or just the incremental leverage around your in-place labor force?

David, thanks for joining us. Well, I think I can perhaps, at this time, answer by example, some of which I gave in our prepared remarks. If we take a look, for example, at our New York market, where we have probably the greatest demand and some of the greatest challenges from a staffing standpoint, we were able to reduce the number of hours that our centers were forced to close -- not close, but certain modalities were unable to have scheduled patients by 40%. I'm talking primarily about MRI at this time since that's the most complete, or larger amount of -- the modality with the largest amount of TechLive adoption. That is a substantial improvement that has created capacity by expanding hours. And so expanding hours and creating that capacity can only -- you can only realize the benefit if you have the demand for procedures that will allow you to fill those now available slots. That's why we chose to introduce this into the Greater New York city marketplace because of the enormous number of -- the volume of procedures that we do and the demand that we have. So I think that's a good measure for us on looking at just the impact of TechLive, which is primarily one that allows us to open rooms that would otherwise have been closed. And if we use that as a benchmark for what we would anticipate elsewhere. I think that that's a reasonable assumption. So that in and of itself was creating 40% improvement in capacity for those centers that had rooms closed. If we were to take a particular day that we might have closed 2 hours, or may have been capable of opening 2 hours more, and being able to do 3 patients on average per hour, that would be 6 per day. So if you have the demand to fill those slots, I think you can do some of the math to see what the potential improvement there. And another example that I'll give you, which is more related to equipment upgrades, we have been able to take some of our MRI scanners and upgrade the software that allows for faster processing of the imaging, and therefore reducing the amount of time that the patient has to spend on the scanner. In our West Coast marketplace, where we did an extensive amount of these upgrades and looked at those centers, we were able to create just through the investment which is relatively nominal. If you look at the cost of an MRI scan, we're talking about $150,000 to $200,000 for the upgrade, which is all software, versus a price point of closer to $1.5 million for an MRI scanner, we were able to create another 4 to 5 scans per day in the same hours of operation. And again, you could do the math and see what kind of potential improvement in revenue that could generate. So we think that, that is a very good benchmark for us to use to explain some of the margin improvement. While it's nice to generate more revenue, the goals that we have from what we're already doing, as well as what we have planned which is more difficult to measure at this point until more of the DeepHealth OS system is actually tested, implemented and operationalized, could have the kind of margin improvement that we've already demonstrated and then some. But we're talking about something that is going to take another 12 months to become fully implemented. But suffice it to say, 2 important things from -- maybe a little longer answer, David, than you expected. But number one is increasing capacity with limited investment is clearly a driver of margin improvement. The other is really completing the process of something that would be transformative for the entire patient journey, which could have even more profound impact once we're able to fully implement it and make certain that the tools give us the results that we anticipate.

And then, guys, there a couple -- go ahead. I'm sorry.

No, I was going to just add one more thing to Dr. Berger's answer, which is to say that this year in 2025, we are targeting to build 11 facilities, and we've opened 2 thus far. We've got 9 that are on the docket for the second half of the year to open. And we have 11 projects in the works for additional de novo facilities in 2026. So if you add those 2 together, you're talking about 22 additional centers on 405 current locations. So that's about a 5% increase in centers, or additional potential capacity to be filled, going forward. So the combination of what we're trying to do within the existing footprint, as well as expanding the existing footprint pretty substantially, I think, should give us the ability to continue to drive similar types of same-center procedure volume growth into the future.

And then, guys, just a couple of other quick ones. One, can you just give us a sense of initial feedback from the iCAD customer base, how you think about the cross-selling opportunities and just as you start to put the 2 companies together?

Well, I think we're a little bit early in that journey, David, given that we just closed and are starting to do a deeper dive into iCAD as of July 17. So we're barely 3 weeks into that process. What we are pleased with is the quality of the sales and marketing team, the engineers and the management team, to help embrace both cross-selling and most importantly, take the iCAD products along with the DeepHealth's breast AI products and put them into a more comprehensive offering. So I think it's a little early to give you more specifics other than we're very pleased with what we see so far. And I think I'll be able to make more comments on that at the end of our third quarter financial reporting.

Okay. And just last one for me. Just curious, additional conversations, or anything that's been prompted by some of the EBCD reimbursement announcement that you guys put out. Just what you're hearing from other payers in terms of potential further expansion of coverage of that product?

Yes. That's another one of those journeys that we're going to have to be patient with. But I think the announcement that we did at one of our capitated groups has now endorsed and is paying for EBCD for all their membership at no cost to their patients. Should be some impetus for other payers to recognize that the value proposition that we're talking about is rather profound. And I think putting pressure to have these kind of screening tools, particularly, I think, at what I believe are very modest prices and recognize the opportunity for better outcomes is extraordinary. And I think we are having conversations with a number of the payors. It's just a matter of time until this becomes fully recognized as the state-of-the-art here in the U.S., and it's adopted. But I think we're just going to have to go through the efforts of having the consumer, meaning our patients, recognize it ahead of perhaps the payors that are going to adopt it, put pressure themselves on those systems to do the right thing and that is to make good medicine, good business.

The next question comes from Brian Tanquilut with Jefferies.

Congrats on the quarter. Maybe, Howard, just -- Mark, just to follow up, as I think about some of these acquisitions that you've done, I mean See-Mode comes up as a technology that supports the other side of the business, right, even though it runs under the Digital Health side. I mean Howard, you talked about the -- what technology is doing to open up capacity. So if I'm just trying to think of quantifying these things. I mean, number one, is there -- is this one of the things where there's enough backlog in demand that as long as you add capacity, you feel like you can add the revenue? And then when I think of See-Mode, I think in the past, you've said it adds like 2 scans a day or something like that. And running the math, I mean it's a pretty big number. And if you can walk us through how you're thinking about that and kind of like what kind of margins should we be thinking about what these technology additions add to the business?

Brian, nice talk to you as always. Let me couch that, Brian, using some of the examples that you've given. There's actually 3 ways that we see See-Mode in particular with their thyroid ultrasound AI benefiting us and which really was the decision to acquire See-Mode as opposed to just license it. The first is looking at the radiologists, thyroid ultrasound is perhaps one of the most difficult and consuming from a time standpoint, diagnostic procedure that is done in all of radiology, and it's extraordinarily manually intensive for the radiologists as well as the technologists. We're quite confident that the early results of improved interpretation for the thyroid ultrasound can be reduced perhaps as much by 50%. And that's after a pretty substantial trial that we've done in the initial implementation. So that will create a lot of additional capacity for our radiologies to help manage the turnaround time for reporting, and for us to have to use less outside teleradiology services to satisfy the demand that we have. On the technology side of it, and the equipment side of it, the early reporting is that we can reduce at least 30% of the scan time, which means that the normal amount of time that we -- or the traditional amount of time that we allocate for doing an ultrasound of any kind is about 30 minutes. If we can reduce that by 30%, which is what we're showing initially, that's a pretty substantial reduction, and potentially could give us 1 to 2 more slots per hour per system. Now again, much like I was describing with MRI, number one, do you have that capacity, meaning the demand to fill the capacity if you create it? And number two, will that happen on every ultrasound unit? And the answer to that is probably no, because not every piece of ultrasound equipment that we have functions at that same level of intensity like our busiest centers. But in those busiest -- busier centers, we have -- I'll go back. We have over 900 ultrasound in 300 -- and I think the number was 28 centers, and about 30% of those -- 25% or 30% now have the See-Mode. And we're putting those into the centers where we have the biggest challenges. So what this will allow us to do in those centers that have the demand, and we have extraordinary demand. Ultrasound has always been one of the most sought after imaging modalities and it's growing perhaps more rapidly than almost any other procedure that we have, is the one that became a major focus for us as we went to the next imaging modality after mammography to try to create improvements in AI. So I think we'll have better statistical information on this, Brian, even very modest additional capacity. And again, the issue is creating the capacity and then make certain that you adjust your scheduling to fit those -- that capacity needs with the demand in a particular location. Our teams on both coasts are quite confident that they will be able to do that. But like everything else, there will be a period of adjustment because it's not just build it, now come. It's a matter of kind of integrating it into the entire workflow program to create that kind of success. But the math on the numbers is quite overwhelming. And we think in the long run, not just what RadNet can do, but what this can do for the entire industry will help not only drive this as a sales opportunity, but will enhance our ability overall to externalize commercially our entire platform because all the tools that we want to build on that platform, hopefully will have similar benefits either for the radiologists, the technologists or both. So it might be a little bit early for me to give more specifics. I'd rather do that when I feel we've had a better chance to do this. We just completed the See-Mode acquisition on June 1. So after doing some test work, we're now in full deployment to get all of the rest of the either 70% or 75% of our ultrasound scanners upgraded to the See-Mode technology, and it can't happen fast enough. I have often said in the past, if I had -- if I could get enough rooms to put ultrasounds in, we could fill them. Now at least we feel that we don't need more ultrasound systems, and we don't need more techs. We need more capacity on the existing systems.

And I'll add one other thing to your question about the flow-through and the margin. The flow-through will be very substantial because we're talking about just doing more scans in the same facilities, in the same hours of operations. Because essentially, what See-Mode does is allow for shorter scanning times. And so theoretically, the principal cost that would come from the incremental scans that go through the same centers, in the same business hours, is really the cost of the radiologist read or interpret the incremental scans, which usually runs in the 20%-ish range, give or take a few points on that revenue. So it's -- as we've always said, the most profitable growth we can have is driving more revenue through the same fixed cost base of centers.

No, it's awesome. It sounds really good. Mark, maybe as I think about M&A, capital deployment, I know you added, or you tapped a credit facility during the quarter. So your cash balance is pretty -- is even more significant now. Just curious how you're thinking about capital deployment on the M&A front more specifically? And then I know you're building 11 de novos. So just curious how you're thinking about the cost of that, and what the de novo kind of outlook is as we think about the growth algorithm going forward?

Sure. So our pipeline of opportunities has grown, and I think we're more confident today than we have been in past quarters to say that we should be putting some of that capital -- that cash balance to work in the second half of this year. So we're excited about that. I think predominantly, most of that is earmarked towards Imaging Center, traditional imaging center acquisitions, as well as some new joint venture acquisitions where -- joint venture opportunities where we'll be putting some capital to work. And then there are a couple of Digital Health acquisitions that we are evaluating as well. So stay tuned. Hopefully, we'll be in a position to talk about some of these things publicly in the coming months. On the acquisition -- on the de novo side, as I mentioned, we're targeting 9 new facility openings by year-end. Hopefully, we can get that done. And then we've got 11 additional facilities that are in various stages of development and construction that should open in 2026. And beyond that, it remains to be seen what is the opportunity? We tend to allocate our capital based upon the opportunities that exist at the time, whether it's earmarked for M&A, de novos, hospital joint ventures, etc. It just so happens that we see a lot of opportunity on the de novo side today where -- to build substantial capacity where we've got a lot of demand in the local markets. Or we're not -- or we need points of access to service patient populations that we currently don't have today. We're constantly evaluating and reevaluating what the best use of our capital is.

The next question comes from John Ransom with Raymond James.

A couple for me. Mark, as we think about M&A, and I know you haven't announced anything, but just directionally, how should we think about the next year, currently considering in digital assets versus traditional imaging assets?

Sure. I think more of the capital is going to be allocated at this point, given what we see and what we have in the pipeline towards the Imaging Center, the traditional services business. There are some acquisitions we are looking at on the Digital Health side. I think most of them are smaller than the iCAD opportunity. So I don't necessarily see us doing something with a lot of our capital towards imaging -- towards Digital Health at this point. Obviously, that could change if something were to come to the table that's so compelling. But right now, the majority of our pipeline, or actionable pipeline, is on the imaging services side.

Great. Second one for me. Look, Mark, for a state school grad, this Digital Health story has gotten complicated. So we're still working it, figuring it out. But let's just kind of focus on for now the technology as it exists in your core imaging segment. So there are a few things going on, right? There's the conversion to DeepHealth iOS. There's remote tech, which you certainly talked a lot more about this call. I mean there's -- you've got a call center, I know that you spend a lot of money on. But if we were just to dream a little bit and look out a few years, and I'm not asking you for '26 guidance, but just directionally, what kind of margin could all of this unlock when you kind of fully deploy it?

Well, I think the best way to answer that question is to look at some of the other radiology software providers, or just SaaS-based software businesses out there that where we think, and we hope, and we aspire to grow our Digital Health software business in the range of 30% per year for many years to come. That doesn't mean that every year might be 30%. It could be a little lower, a little higher in any given year, but that's the trajectory that we're looking for, for that business. And we think once mature, if you look at other SaaS-based software businesses, we should have EBITDA margins approaching 50% with little capital intensivity certainly as compared to our core Imaging Center business.

That was a different question than I asked, but I appreciate that color. Now I'm asking like just on the core imaging, once you fully deploy the DeepHealth iOS, once you fully deploy remote tech, once you maybe look at some opportunities in your call center, what's the imaging EBITDA margin unlock percentage once you just start using -- once you fully deploy all these tools in the out years. Not with Digital Health segment, along -- what could this do with your core business?

Now I understand your question. Yes. I think -- Howard, do you want to take that? Or do you want to me to give a shot at it?

I'll give a shot at it, and then you can amplify on it Mark. You have to, John, look at not how the total would look, but what are the sum of the parts look like that will give you the total savings? And what I mean by that is our platform is a modular platform, and it allows us to look at one function, if you will, at a time, multiple if we want, but one at a time, to see what kind of a potential impact going from a more manual labor intense process to a short -- to one that is more tech enabled, if you will, I don't want to just use the word AI, but tech-enabled. And for example, in the contact centers, we look at our current typical call handling time. And right now, we measure on a blended basis, the average call in minutes. Obviously, some calls are shorter than others, but even if I were to use 5 minutes as a typical call, and I actually think it's longer than that, if we were able to get a modest 10% reduction on every call that is logged into our contact centers, that improvement is enormous when you consider that we do about 11 million procedures annually. But each procedure is associated with multiple calls. Some of those calls are not to schedule a patient but to ask for direction or to give insurance information or to ask about directions on the closest center. All of those are things that we feel we can have impact on. And if you do simple math, and something that I think is easily achievable, a modest 10% improvement on this, the numbers are just staggering. And for us, while it may not necessarily lead to less people we employ, we want to make the people that we do employ that much more productive so that we can grow the volume without the dependency of continuing to have to be challenged with the labor cost, and availability of labor that everybody in the industry is facing. So I think we could do some statistics around that. We haven't as yet. But everything that we do now will have either a full, or a partial component that is addressed by artificial intelligence. And I named a few of those a couple of times in my prepared remarks.

Maybe you can tell, what do you spend on your call center today? What's that cost?

Yes. Our overall call center cost for the company is north of $60 million a year. That includes the people, the systems, the facility rent, services that we currently use offshore. So it's very substantial.

Yes. I've covered you guys for a long time. I don't think I've ever discussed the cost center. So that's -- maybe I'm the last guy of the party here, but that's pretty interesting. And then lastly, Mark, you've said before putting your centers on DeepHealth iOS as an 8-figure, good guy. I know you're just starting that process. Is the goal still to have all these done by the end of the year? And do you have any -- obviously, 8 figures could be $10 million, or $99 million. So have you thought any more about narrowing that good guy as well?

Yes. I mean, it's going to be hard to come up with a number today. I mean we're blazing a new path with some of these details or Digital Health initiatives. I think we'll have a better feeling as we start -- just like with TechLive as we start rolling that out, and we began the pilot. Now we have a little bit better feeling for that. I think we'll -- as we get into our budgeting process towards the end of this year, which starts looking at what 2026 will look like, I think we'll have a better feeling for some of that. But some of these modules aren't going to be being -- aren't being adopted until 2026, or throughout 2026. So even if we were to come up with a big number, it's not going to happen overnight. It's going to be gradual for the next few years.

The next question comes from Andrew Mok with Barclays.

Just one question for me. Mark, I think you called out a $4 million to $5 million benefit from the physician fee schedule next year. I think this year's rate headwind without the doc fix was $7 million to $8 million, and the headline rate update next year would suggest at least 3% or 4%-plus and maybe a larger benefit than what you called out. So can you help us understand the items weighing on that benefit when you did the CPT analysis?

Yes, sure. We always caution every year people to just extrapolate the headline that you might read in news stories, or even CMS when they give their little tables in there because you have to go CPT code by CPT code. You have to then also incorporate the GPCIs, or the geographic codes based upon where in the United States you operate. And so it's often that when it looks like a big impact one way or another, traditionally, it's been a cut for us. Our cut is not quite as big as the impact it shows. While it's the same thing that's happening next year on the increase, where our increase is not quite as big as what the headline number says. Because even though the conversion factor increased by 3 -- or is increasing by 3.3% next year for non-APM facility-based Medicare fee schedule billers. When you look at the individual CPT codes, they've made a number of adjustments, particularly to the practice expense portion of the procedures where some of these RVUs are going down, and some of these GPCIs are going down that therefore, mitigate some of that otherwise headline number benefit to us. So it's -- we've done an exhaustive analysis that we do every year, and we're confident in the 4% to 5% uplift in our reimbursement next year.

The next question comes from Larry Solow with CJS Securities.

Question just on, Howard you gave some really nice detail on the See-Mode kind of the strategy, and it really makes sense. I'm just curious on the -- with iCAD, is the strategy from a high level? Clearly, your exposure to mammography procedures is going up like four-, fivefold. Is the strategy just to upsell your AI equipment to these customers? Or is it a combination of the ProFound breast health suite that iCAD brings along with your technologies. Can you just give us a little more color just from a high level there?

Sure, Larry. The iCAD products, which is their ProFound suite, and the DeepHealth product, which is primarily our Saige-Dx platform, or Smart Mammography, are really very different in some respects. And well, what our teams are working on already is how do we blend those into a single set of offerings to both the current RadNet customers, as well as iCAD customers. But I would also like to point out that iCAD also has a couple of other products that they either have, or will be getting, FDA approval that we can add to the existing EBCD program that will expand the value proposition for early breast care detection. So we're still working out some of those details. And again, this has only been the last 3 weeks that we've been able to speak directly to the various people and departments inside of iCAD. But I think you can expect both customer bases, if you will, that which iCAD brought to the table, and which RadNet is bringing to the table, to be beneficiaries of a more comprehensive solution, not only for the mammography that we're doing here in the United States. But iCAD, I want to point out, has a substantial number of customers outside the United States, primarily in Europe that we think some of our other products in the DeepHealth program, or just in mammography will gain a lot of traction. So we're very pleased that we've done the iCAD transaction. And I think the marketplace should be pleased with that, too, because I think putting these 2 brands together is going to have a far greater impact on the overall value proposition of artificial intelligence and early disease detection. And I'm specifically using the word disease detection and not breast cancer detection because some of the tools that we're looking at, things like risk assessment and breast arterial calcification are all things that fall into the category of screening, but not necessarily just for breast cancer.

Got you. I appreciate all the color. And Mark, I know you're not ready to give guidance, but just from a high level, I think when you first announced the acquisition, I think you mentioned that there was some spending that you were going to do organically, that you may not have to do now as you acquire iCAD? Is that kind of still -- as you had a couple of more months at least not under the hood, but at least to plan your own internal strategy. Is that still kind of the case?

Yes. So before we announced iCAD, we had earmarked up to about $20 million of additional infrastructure investments within Digital Health for 2025, which was the reason why we weren't going to -- we weren't anticipating in '25 to see a lot of flow-through on the profitability side of that business. With the iCAD acquisition, some of that investment in infrastructure which was earmarked for sales and marketing teams, commercialization teams, deployment, customer support, implementation, that we're getting from the iCAD transaction and some of the talented team members that have come on board. So I think what we're synthesizing that this quarter and in November when we release our third quarter results, we'll update the guidance for Digital Health, which will incorporate the revenue contributions that we see for the second half of the year from both iCAD and See-Mode, as well as any adjustments that we might make on the profitability side, which will have this benefit that we just talked about as incorporated into the results.

The next question comes from Yuan Zhi with B. Riley.

Congrats on another strong quarter. With increased utilization mentioned by multiple payers, how did that impact your conversations on capitated contracts?

Are you saying -- I just want to make sure I understand the question. Are you asking about with the improved pricing that we're getting from some of the commercial insurance companies, is that affecting our viewpoint or our outlook on capitation? Is that it?

So my question is the multiple health care insurance companies have mentioned that they have increased utilization, including imaging and diagnostic imaging. I wonder, did they try to hedge those risks by signing more capitated contracts? And how does that impact your conversation of capitated contracts?

Got it. Got it. Understand. Well, look, I think we've been in the capitation business for close to 30 years, and we enjoy that business. It's predominantly a book of business that's here in California. We've managed up to almost about 2 million lives at the peak of our capitation. We don't mind taking risk, with taking risk and managing risk appropriately. We've had the -- these contracts have been very profitable for us over a number of decades. However, as you mentioned, utilization does continue to go up in diagnostic imaging. We see that and feel that within our capitation contracts and upon renewal, we have to make sure that we're getting compensated commensurate with the efforts that -- and the number of procedures that were contributing to these contracts. And in situations where we can't -- or we don't feel like we're being paid adequately, and we can't get the increases that I think would allow us to continue these relationships, we flip some of these relationships over to fee-for-service relationships where they're no longer obligated to send us all of these patients like they were when we were capitating for them. But because of our strong position in many of these local markets, we end up seeing most, if not all, of these patients anyway under fee-for-service arrangements where we can get significantly better pricing. So what you've seen in our revenue over the last, I'd say, 4 to 6 quarters is that our capitation revenue has come down. But our overall revenue has gone up, and that's because we're taking some -- we're getting increased reimbursement when we flip these contracts to fee-for-service relationships. And the capitation arrangements that we currently still perform are being performed at rates that we think are appropriate given the other books of business that we have.

Got it. And then for this remote scanning solution, how quickly do you think you will be able to deploy it outside of the New York facilities? Any regulatory requirement for someone, let's say, in California to operate a machine in New York?

Well, we've actually begun that process already. So we have technologists in Florida that are operating scanners in New York. We have technologists in Arizona that are assisting scanning in California. So we've already gone beyond the New York market. I only used the New York market because it's one of the more mature ones, and it's the one where we -- where the demand issues are particularly profound. So we hope to have the rest of our MRI fleet enabled with TechLive by the early part of 2026 in that first quarter. So we're rapidly moving along that path right now.

We're also -- yes. I was going to just add one other thing. We're also testing the ability for technologists, or better techs or more capable techs to control multiple machines at one time, which -- while I don't know that all of our techs have the capabilities of doing this, but it will give us some additional leverage with our labor force and will allow some of our technologists to make more money, and be more productive.

Got it. And one last question from us. Considering the rapid increase of Alzheimer's related imaging, where are we in terms of the current capacity utilization for PET scan and MRI to have the remaining capacity for more backlogs or for utilization?

Yes. Go ahead, Howard.

I was going to say the answer to that is yes. And it's probable that the increased demand that we're seeing for MRI is, in fact, coming from Alzheimer's. Because once you do the PET/CT and identify a candidate that might be -- or a patient that might be a candidate for a drug treatment, you then wind up doing, I believe it is either 3 or 4 MRIs in that first year on the drug treatment. So we're getting a double bang for the buck, if you will, that our MRI scanners are getting more challenged by the success that we're having in the amyloid and Alzheimer's program, which has grown dramatically here since the beginning of the year.

And to your question directly on the PET/CT capacity. We have a fair bit of PET/CT capacity. Today, we're running at a run rate, if you just multiply the PET/CTs we completed this quarter, times 4, close to about 90,000 scans per year. We're doing that on about 68 or 69 PET machines nationwide, if you assume that they're open 255 workdays. We're talking about doing an average, today, of about 5 PET/CTs per machine per day. A very busy PET/CT scanner can do upwards of 15 scans per day. So we have a fair bit of capacity on these PET/CTs to continue to grow. Whether it's the PSMA business, the amyloid brain studies, or some of the other tumor-specific, or more complex exams that we're doing with some of these new radioactive tracers. Where we'll run into some capacity issues is around the CT modality because many of -- most all of our PET/CT scanners today do double duty in the sense that when they're not doing the PET/CT work, they're doing traditional CTs. And so to the extent that our PET/CTs get busy and busier, which we hope they will with the PET/CT portion of that, we'll have to offload the CT volume to additional scanners or additional capacity.

The next question comes from Jim Sidoti with Sidoti & Co.

Can you give us some sense on how many centers or what percentage of your 405 centers are now using DeepHealth?

Well, Jim, nice to hear from you. The centers -- none of the centers are really using the DeepHealth operating system platform with anything more than 1 or 2 modules. One of those modules is, for example, TechLive. Another module that we're beginning to deploy is our new reporting tools that significantly shortened the interpretation time for our radiologists and move on to an own platform versus the platform that we license right now. And so we also are working on our contact centers, our kiosks in our imaging centers. All of these are in some way being tested, but on a case-by-case basis. So nobody is using the full range of the potential of the DeepHealth OS in the outpatient space, even in RadNet at this point in time, because we're still tweaking some of the algorithms and protocols we used for it. But slowly but surely, every facet of what we do is being tested to improve virtually every component of, as we describe it, the patient journey. Whether it's scheduling, or whether it's the call-ins, whether it's notifying patients of the prep that they might need for a particular exam, directions, a number of things that can be answered by chat bots and not necessarily live people. So when I was giving the example, of really the scale that we're talking about that even a 10% improvement of reducing the call time, given 11 million procedures and probably, I'm just guessing, each procedure has, on average, maybe 3 or 4 calls, so we may be handling in the neighborhood of 40 million to 50 million calls a year. And as Mark pointed out, the cost of our contact centers is quite substantial, but they can do additional capacity if we create these kind of efficiency tools.

So how long do you expect it will be before you are fully implemented in your centers? Are we talking about 12 months, 18 months? How long do you expect this process to...

Yes. We're probably talking about towards the end of 2026 to have it fully implemented with all of these tools. So I think 15 to 18 months for all of the centers. There will be some of the centers, and some regions, that will go ahead of others, obviously. But our timetable is to not rush this because it's a heavy lift number one. And number two, it is the platform that we think will be a one of a kind in the industry that we're going to get a lot of traction externally for. So we want to make certain that we test this out in an environment at scale that can truly make certain that it achieves all of the operational performance tools that we know are capable.

Okay. And then for the second quarter, what was the same-center volume increase just on an overall procedure basis? I know you called it out for some of the different modalities, but overall, what was the same-center volume increase?

Yes. So what we started doing, Jim, is we started kind of bifurcating the reporting between advanced imaging and routine imaging. And the reason is because we do far more routine imaging by the number of procedure volumes than we do advanced imaging. It's almost a 75-25 in favor of routine imaging. But really advanced imaging, the 27.5% of the advanced imaging that we do by procedure volume, provides more -- between 60% and 65% of our revenue. So it would be misleading let's say, if advanced imaging were expanding very rapidly and routine imaging warrants, the routine would skew the numbers if we did a blended rate. To answer your question, MRI, CT and PET/CT advanced imaging, together increased about 6.6% on a blended basis for those 3. MRI itself on a same-center basis was up 6.6%. CT on a blend on a same-center basis was up 5.9%. PET/CT was up 16.2%. And routine imaging, which is ultrasound, mammo, X-ray, all other exams, that was only up, I think, about 1.25% -- 1.4%, it was up. So if you blend all of that together, which we're no longer doing because of this distortion, it would be about a 2.7% increase on all procedure volumes.

Okay. All right. And then the last one for me. When you talked about the iCAD acquisition initially, you indicated they were going to have about 25 or 26 sales reps marketing your DeepHealth. Now that the deal is closed, have you been able to keep all those folks? And is that a good number for us to consider for your DeepHealth sales force?

Yes, I think that's a good number for us to go with at this point in time. Again, it's only been about 3 weeks, and we're trying to be as efficient as possible with getting to all of the people. But we think we're going to keep that entire sales force.

Yes, all of those team members have come on board, and we hope, and anticipate, that we'll be able to keep them.

This concludes our question-and-answer session. I would like to turn the conference back over to Dr. Howard G. Berger for closing remarks.

Thank you, operator. I again would like to take the opportunity to thank all of our shareholders for their continued support and, in particular, the employees of RadNet for their dedication and hard work to help having produced these extraordinary results in Q2. Management will continue its endeavor to be a market leader that provides great services with an appropriate return on investment for all shareholders. We are looking forward to your next call, and thank you very much for your time today.

Thank you. The conference has now concluded.Thank you for attending today's presentation. You may now disconnect.