Interview: Sham Sokka, Chief Operating and Technology Officer, DeepHealth
This article was originally published on April 5, 2024. The version you are seeing is an updated version which has had a statement that DeepHealth’s cloud-native operating system has already launched removed as this is incorrect as of the time of publishing. This version has also been made open access at the request of DeepHealth.
Sham Sokka is Chief Operations and Technology Officer for DeepHealth, a digital technology subsidiary of RadNet, which is the largest provider of outpatient imaging services in the US, with almost 370 centres across the country.
DeepHealth technology is already used in over 300 radiology departments and imaging centres across the world. Over 15 million exams each year are performed using DeepHealth solutions with more than two million AI-informed clinical decisions.
Sham Sokka, Chief Operations and Technology Officer, DeepHealth
“In 2012 we [RadNet] were considering whether to buy our platform products or build our own. We decided to build our own. Being such a large imaging provider, we wanted to have that as a control point. So we acquired a radiology information systems (RIS) business (RIS is like EMR for radiology), and a picture archiving and communications system (PACS) business. Together, these formed the core IT platforms for radiology, until 2020 when AI became a huge part of the story,” Sokka tells us.
To leverage the benefits of revolutionary AI innovations, in February DeepHealth announced a strategic partnership with Incepto, Europe’s leading platform for artificial intelligence solutions applied to medical imaging. The aim is to provide access to a wider AI-powered portfolio with integrated AI clinical solutions and workflow tools to address multiple challenges faced by radiologists and staff.
“Practice patterns are very local in Europe,” Sokka says. “Spain differs from France, which differs from the UK. In part this is because of differences in how the government pays for it, but it’s also because of different clinical guidelines. So we’ve partnered with Incepto, which is very strong in France and the UK, because they have an existing footprint. We’re talking to local radiologists, with people that know the local workflow.
“There’s a huge amount of legacy products in Europe. Many of our customers are asking for a modernised platform that can read remotely, and bring different elements of the workflow together. We’re in the process of upgrading customers over the next three to five years. Less than 2% of the radiology industry is on the cloud. In the next 5-10 years the sector is going to undergo a major paradigm shift. All kinds of capabilities are going to be unlocked. It’s not a question of if the sector adopts the cloud, it’s about when.
“Over the past few years, we acquired a couple of AI companies to bolster our platform. The AI we use breaks into two areas: operational and clinical. The operational side covers things like getting the patient to come to the centre, making sure the right test is ordered, the right image protocol is done etc. This can account for 60-70% of the cost of doing an image. There tends to be so much focus on making radiologists better, and this operational side often ends up getting ignored. For example we do 2m mammograms a year, we have DeepHealth’s infrastructure to manage that. One of the hardest parts is bringing patients back every year (although image interpretation is hard too). Patients may skip a year, and if they do the risk goes up. So much of the work we’re doing is around how we bring patients back, what messaging works, how to educate patients about the risk profile they’re in so they come back at the right time. We also manage all patient engagement platforms, to determine which patients to follow and when, consolidating clinical info from previous studies so the doctor looking at the current study has context. All that is really critical.
“And then we have image streaming and AI on images, which enables us to say ‘we think cancer is there’, or not, or to look at an area more carefully. Imaging is leading the health care AI revolution, with 100s of clinically approved products.
“Diagnostic specialities are very different in imaging, and different expertise and AI models are required. We’ve acquired all those over the years in DeepHealth, but also have RadNet as a place to deploy and validate solutions. This is of huge value for us because we have to do it to make the business more effective, so we’re now productising those for external consumption.
“We feel we’re now uniquely positioned to be able to bring the operational piece, plus the clinical piece, and then the AI layer across. In the market there are specialists in each area. But that means a provider has to buy multiple solutions and make them interoperable. But we’re offering a way to bring it all together in one environment. It’s crazy to think it doesn’t exist already.
“The idea is as you start to modernise systems, you can start to move to a more efficient radiology network enabling radiologists to increase their value, efficiency, accuracy and personal satisfaction. Think of a cardiologist or primary care physician: right now, when they need a diagnostic test, they have to schedule an appointment for imaging, and the radiologist has to interpret it. Imagine if aspects of that process were automated from assisted image acquisition (protocol automation), postprocessing, automated diagnosis, or report generation. That’s the next wave. Wave one was getting the images digitised. Wave two is bringing in AI to processes and connecting siloed data and workforces – what we’re doing now.
“To summarise: We’re bringing an integrated radiology platform for operations, clinical and AI driven automation, with a cloud-based infrastructure, allowing remote access. This also brings a very robust security infrastructure, modernised for scale, modernised for AI development, to leverage all tech, including generative AI. We’re putting it all together to have an impact in the diagnostics space that’s exploding. We’re having an impact on multiple clinical specialisms, in particular cancer and neurology. That’s why we call ourselves ‘DeepHealth’ and not ‘Deepradiology’.”
Read more articles from Healthcare Business International by signing up for the HBI Headline News newsletter. Get exclusive content delivered straight to your inbox every week. Don’t miss out – register for free now!
We would welcome your thoughts on this story. Email your views to Martin De Benito Gellner or call 0207 183 3779.
