Welcome to Limitless Medical Imaging Podcast, the podcast that connects healthcare leaders of the future with innovative solutions. Limitless Imaging Podcast is an audio experience by Medicai, which is a suite of systems that allows healthcare organizations to create next-generation medical imaging workflows and innovate on patient experience.
In this podcast, we'll feature a variety of conversations with the purpose of providing you with interesting insights and perspectives from medical imaging industry experts. You can subscribe to our podcast on Apple Podcast, Spotify, and YouTube, and follow us on Medicai.io.
- Hi, my name is Andra and I am a marketing manager at Medicai. Today I would like to introduce you to my guest, Andrei Blaj, who is the COO and co-founder of Medicai. Hello Andrei and thank you for taking the time to discuss a little bit about some of the ideas that we build around. This time we'll take a deeper dive into how we as healthcare organization leaders working with medical imaging can better prepare to take full advantage of the future.
To get started, I would really love for you to share a little bit about some of the values and ideas that are important to Medicai and how these values fit into today's developments in healthcare and, more specifically, in digital healthcare.
- Hi, Andra. First of all, thank you very much for inviting me to talk with you about what we do and about how the future of Medicai is looking from our perspective. I would like to point out that Medicai is trying to position itself in the market as a company that is forward-looking. When I say forward-looking it mean means that we look at how the world will be in a couple of years from now. In terms of regulation, we expect that the regulators are pushing for a future where the healthcare providers exchange data between them easily and also share the data with the patient and empower them to access the right healthcare services.
We put a great emphasis on creating the infrastructure that enables healthcare providers to interconnect their organizations and their software system systems and also to enable their patients to access their data in a digital way.
- Okay, so if we're talking about data access and data exchange, maybe we could share a little bit about the concept of interoperability.
- So traditional medical image viewing methods often require doctors to log into several applications to view the images associated with the patient. This can create clinical blind spots and ultimately have a negative impact on patient diagnosis. But some of today's medical imaging visualization solutions, allow clinicians to access, view and collaborate on images without being tied to a specific system or workstation.
- I would like for you to share with us a little bit about why is this concept of interoperability important, and what should healthcare leaders be mindful of when optimizing for interoperability and accessibility.
- So the systems that are in hospitals right now, in the vast majority of hospitals, are systems that keep the data in silos, data silos. I wouldn't say that the systems have been created like this by design, I think that it's just because they were developed 20 years ago, 30 years ago, with the technology that was available at that moment. And also probably, most of those systems were a big investment at that time. So changing those systems, I think it is not as simple or easy to make a decision.
But the future is going to be a future for healthcare providers that can communicate with other systems easily. Now the question is how do you get from a point where you have all the data in silos to a point where you have your systems open in a secure and compliant way?
I think it's like there are two options, right? So option one is to change what you have right now with the modern system, modern technology that is interoperable by default. Or option two is to add a layer, a smart layer on top of what you have, that enables this kind of interoperable future.
But I think that most organizations have to plan to get to a point where they are interoperable with other systems. And they are interoperable and are able to send data to the patient in a digital format.
- So how can we plan better? How can we construct a coherent and cohesive imaging strategy?
Specifically talking about medical imaging.
- I think when you say strategy, it's a very big word. Creating a strategy is always a mind-bending exercise, especially for leaders in the healthcare system. I would argue that you need to take it easier on yourself, right? So you need to break down what you need to do in small bite-sized steps that are easier to manage. I would start by asking myself some simple questions, right? So what kind of data does my organization generate? Where is that data generated? Then what do I wanna do with that data? So do I wanna have it to available on premises for my radiologist to see that and that's it?
Or do I want that data to be available on-premise for my radiologist and then, out of the premises of the hospital, for the doctors that are on my team? Do I need to send that data to doctors that collaborate with my organization? Do I need to send that data to patients? So, what exactly is the life cycle of those pieces of data?
And after this question, you will realize where you need to store that data. So for example, if you only use it on-prem, maybe the solution will be to hold it on-prem. If you need it on-prem and then you also need to give access to it to some other people, then maybe storing it in the cloud is a good solution.
And then the next question would be for how long do I need that data? So probably every data that I generate in the hospital, I would need it for a couple of days, weeks, maybe months. And then you have the regulations that tell you to keep that data for a couple of years after you generated that data.
In the US depends from state to state, and in the EU also depends from country to country. Also security plays an important role when you define your medical imaging strategy. I would also take that into consideration.
- Okay. So when we're thinking about the management of imaging data and the life cycle of data, you mentioned a few aspects. So you mentioned access, you mentioned storage, you mentioned security. What other things do we need to take into consideration?
- So another important thing that you need to define would be the users. Who accesses that data? And then with users, you also have to define user roles and accessibility roles.
So, for example, 20 years ago, probably your PACS system had only two roles, no access or full access. In today's world where you have HIPAA, where you have GDPR, which regulate the access to protected healthcare information of the patients, you need to think about who has access to what data, and very importantly why. So most of the regulations say that doctors or people from your organization should access the data if they need to, right? So you shouldn't access more data than you need to do your job.
So to summarize:
What type of data do you access? Where do you keep it? The security layers, cybersecurity layers, who has access to that data and, then, define roles and access controls on that data.
- What would be the steps for healthcare organizations when it comes to choosing the right technology? What, healthcare leaders should be mindful of?
- We're talking about technology specifically for medical imaging?
- Yes.
- Okay, so there are a lot of options when it comes to software for medical imaging out there. The most popular systems that are on the market right now are systems that are deployed on the premises of the hospital. So mostly on-prem PACS systems, and on-prem VNAs. Then there's a new category that is starting to emerge, and those are cloud-based solutions. So cloud-based PACS, cloud-based VNA, cloud-based image exchange solutions. Then there are solutions that blend the two, that make the on-prem solutions and cloud solutions work together, to create different workflows and different services for patients, doctors, referring doctors and other stakeholders that want to interact with medical imaging. For example, research companies or drug development companies that are using medical imaging for new drug development or for clinical studies, clinical trials; imaging data is a product or a piece of data that those companies, that research companies use, that develop different new methods to diagnose better or to support the radiologists better.
They need access to medical imaging data, usually anonymized medical imaging to develop their algorithms. So for example, if you are an innovative company that develops an AI that does early detection of breast cancer, to be able to train your AI, you would need access to probably tens of thousands of mammographies so that the algorithm can learn how an early cancer looks like.
You would also probably need data on different stages of that cancer throughout the early stage, then later stage, to enable the algorithm to detect the possible problem.
- So alongside access, what other things are important when choosing the right technology for medical imaging?
- In the current times when you have a lot of pressure put on investment budgets on all healthcare organizations, I would be very mindful of a couple of things. One of them is how big the investment that I need to make. And then the other thing is, what changes if I bring in this new software or this new technology?
So if I, for example, if I need to completely change the system and then completely change a couple of workflows, and then I need to go through a security audit again, I would think twice or three times before changing any piece of software. So I think it's very important for new technologies, especially in this space, to be easy to deploy, to connect to what I already have and to enhance what I already have in my healthcare organization. So I would definitely choose technologies that lower the risk of things going wrong for me.
- As a decision maker, how important is it for these technologies to be customized?
- It's definitely something that's more important right now than it was probably 10, 15 years ago. So right now healthcare providers differentiate themselves from others by having different workflows, and I think the best providers are the ones that also embrace the new digital workflows in their healthcare services. This means that they provide their patients with the ability to do, online consultations. They provide their patients with the ability to do online evaluations. So they don't ask for their potential patients to come into the clinic, run a couple of tests, and then see whether they are eligible for a certain treatment or not. So taking this into consideration, that healthcare providers want to create their own workflows, then I think it's getting more and more important to have ways to personalize your technologies to meet your needs and to map your workflows properly.
- So talking about personalization, this year only FDA approved around 79 algorithms for radiology.
What do healthcare organizations need in order to leverage artificial intelligence specifically in medical imaging?
- Yeah, this is a great question. I think this is a question that outlines how can you prepare for the future. Right? I think the key is to work on your interoperability, so on your ability to connect your organization in a secure way and compliant way, and scalable way with other organizations, including, innovative organizations like companies that develop new AI methods in radiology or applied in any other healthcare specializations. So, interoperability means that you would like to open up a pipeline that connects your organization database to the AI development company database. You set the access control on your side. So you say, what type of data do I allow this organization to access from my database? You also say whether you want to anonymize the data before sending it to them or not. And then you send the data to that organization and that organization looks at the data and sends back the results to you, ideally directly in your PACS system or right next to medical images. So this would be a flow that enables you to collaborate in a meaningful way with these companies. And the step that you need to do is to work on the interoperability of your technology stack.
- And if we're talking about AI and the technologies that are changing the medical imaging industry now and in the future as well, how, how else do you think medical imaging will change, taking into consideration developments in technology? Or what developments in our society do you see that can impact the transformation of medical imaging?
- The biggest factor for any big transformation is a big usage of that technology. So if you talk about medical imaging, we see two trends. One trend is that diagnostic is getting more and more popular.
In healthcare, you want to measure 100 times and cut once. Meaning that you wanna do all the medical imaging that you can before going for a treatment or going for an operation, right? So this is one factor that will greatly increase the volume of medical imaging being generated in total.
More access to more and more diagnostics. And then the second thing is that the quality of the imaging is increasing. So the machines that are developed, MRI machines, CT machines, PET scan machines generate images at a much higher resolution than before. So putting these two together, you will have a much bigger volume of data being generated.
Then you also add in the mix the fact that companies will want to hold to the data longer for two reasons. One reason is regulation. So the government or the state will make you hold the data for longer. Second reason is for research purposes. So you would want to hold that data for longer so that you can use it to develop new methods to diagnose, to train AIs, to learn from that data in the future.
So I think these are the factors that will push this industry towards more and more innovation. I think the innovation will happen, especially in the connectivity of one organization with the other. So interconnectivity of organizations, and also in automatic or semi-automatic methods, including based on AI to do certain repetitive tasks.
Why? Because it's easier to train technology than it is to train people. So to train one radiologist, you need to have him go through school for probably 15 to 20 years. We hope that we can train AI faster to manage high volumes of data in the future.
- So looking at this future with better equipment and more accessible machines, with personalization and AI, what advice would you give healthcare leaders to achieve these goals?
- I think the best approach is to take it one step at a time. So what I would recommend to do in the next one or two years is to make your infrastructure interoperable.
There are different pathways that you can take to achieve this. One of them is to change what you have with something that is better and has interconnectivity built in. Another one is a more agile approach, to use a technology like the one that we developed, that connects into the already existing systems, on-prem or in the cloud, and adds this layer of interconnectivity on top of what you already have.
Of course, there are other solutions too, so there's always a solution to build your own custom software that makes what you have interconnected with other systems. And there are companies that have the capability to do this. So if you wanna take one idea after this talk, this podcast, is that you want to become interoperable with other systems because the future will belong to organizations that can connect with other organizations. And that can provide patients with the right treatment, at the right location.
Okay. Thank you, Andrei, for this conversation and thank you to our audience for listening to this episode in its entirety. If you have any questions and would like to follow our projects, please join us on LinkedIn, Facebook or Instagram at medicai.io, or contact andrei@medicai.org.
- Thank you very much, Andra. It was a pleasure to talk.
Unlock the power of seamless healthcare collaboration and efficiency. Take a step towards the future of healthcare interoperability by downloading our whitepaper today.
