Redefining the MedTech Landscape - AI Technology

Interview Transcript

Article | AI Corrections
24th January 2023 Atheneum Team

Expert Profile






Naheed Kurji is a Co-Founder, and the President and CEO of Cyclica, a Toronto-based venture-backed biotechnology company that leverages its integrated AI-augmented drug discovery platform to decentralize the discovery of new and better medicines. Naheed is also a co-founder, Board member, and Chair of The Alliance for Artificial Intelligence in Healthcare (AAIH).

Section 1: Why is AI redefining MedTech?

1.1. To what extent has AI disrupted the way we deliver healthcare?

It’s undeniable that AI is having a meaningful impact in the way in which healthcare innovation and healthcare outcomes are being driven. Looking at the present, there are four big buckets in which AI is having an impact across healthcare.
The first AI use-case is in understanding biological drivers of diseases, making new hypotheses on which biological targets the pharmaceutical industry should focus on designing medicines for. New AI technologies are taking a wealth of data and creating new hypotheses on which target(s) to focus on.

The second bucket are companies who design or discover molecules and the medicines. A vast majority of capital has gone into organizations that built new methods to design modalities to interact with the biological targets that are linked to the disease. There are hundreds of companies that have now emerged and distinguishing between them is oftentimes hard.

The third bucket is on clinical trial improvement, clinical trial efficiencies. Companies are taking clinical trial data to improve the way in which medicines are brought to the right patients in the right clinical trial. Clinical trial data has oftentimes been disparate, hard to access, hard to standardize. For example, in the Alzheimer’s space, there’s over 150 failed clinical trials.

The last bucket, are companies that are building AI methods to improve how physicians manage patient care. I’m really excited by some of the companies that have raised a lot of money and advanced interesting methods. For example, AI methods have emerge to help radiologists improve the way they their business to ensure that patients who require a radiologist to look at a scan, don’t get put at the bottom of the file and in a queue of a few weeks before radiologists can get there. Can AI improve the way in which patients who desperately need readouts, get those readouts? And support how radiologists do their business?

1.2. Why should healthcare providers use this technology more?

Artificial Intelligence is a fairly esoteric term that people are trying to understand. They are trying to understand the true application of what otherwise at a surface levels can seem very dehumanized. So, why should physicians use it?

If you want to make better decisions, use AI, but it’s not just about trusting that AI is going to solve it. AI is going to open up new hypotheses, things that we would not have otherwise been able to consider in the speed at which we would need to consider it.

Ultimately, I believe that whether it’s physicians using a technology to support a diagnosis, looking at a radiological scan, or it’s a pharmaceutical company leveraging AI to design new chemistry for a biological target, you have a human on the other end that needs to make the final decision.

The historical paradigm was that it used to be humans who guided technology. Today and going forward, I believe that technology should guide humans, but ultimately the two things are necessary.

1.3. Are there any barriers to successfully implement this technology?

It’s not just about the scientific problem, and it’s not just about the technology. If we don’t have the mindset to affect behavioral change on adoption then we will make minimal progress. I therefore spend a lot of my time within the AAIH and within Cyclica not just optimizing technology, not just optimizing on this application scientifically, but ensuring that we bring on the right people who want to be guided by this.

I believe that the mindset of people is the biggest barrier to adoption. It takes time for adoption because mindsets generally, especially in the healthcare space where patients are the beneficiaries and the risk is very high of getting it wrong are much more skeptical. It’s very important to have proof points and success stories. Scientifically, success stories are about development, understandability and interpretability.

One of the limitations of AI companies is that we have glamorized the idea that you press a button and you can solve a problem. Yet, today it’s not that simple and I constantly fight against that. People need to understand how it works so that they can trust the outcomes.
So, in our space, there has to be a healthy dose of skepticism. However, the mindset is going to shift, especially as we show more proof points in the space.

1.3.1. Have you faced any resistance from patients or healthcare staff?

All the time! In AI there is artificial and intelligence, which makes this seemingly dehumanized. Healthcare is already a very difficult area. The people, the beneficiaries don’t necessarily understand what they have, or it is being treated or managed. That lack of understanding for any human being is scary to begin with. Now, you put your trust into somebody else, that’s already a stretch and now, you’re trusting this AI that is giving this person that you’ve just put your trust to advise them. It just becomes a cycle of challenging human behavior and we face this all the time.

Ultimately it’s going to take effort. So, the question is always the same. How do we humanize what is otherwise perceived as an esoteric, poorly understood, but very exciting opportunity.

1.4. How does this technology reshape what the future of healthcare will look like?

On a scale of 1 to 10, 1 being nothing, 10 being absolutely important. It’s going to be 10.

I believe AI technologies will be the equivalent of electricity in an operating room or the next best technology that has ever been implemented to do something that’s going to be AI. It’s going to be sustainable and it’s going to be prevailing for the next wave.

The next big step is about how to improve on it? Can we do it faster? What about Quantum AI? I think quantum computing is still 5-20 years away on a meaningful basis. But now you map in AI on GPUs and CPU and you think about scalability. Can we do this on Quantum? Maybe. Either way that’s a really exciting space.

We’re in a space that has been on a decline in healthcare for the past 60 years. You’re not going to change it in 5 or 10 years. I think it’s going to be a steady growth, but it’s going to be on the back of AI/ML technologies.

Section 2: The Future of Digital Disruption in MedTech

2.1. Besides AI, what other MedTech innovations are you most excited about?

I am really excited about AI for robotics in our healthcare space, how to improve lab work, as it is very manual. Hence, how can one make that more autonomous, is something I am really excited about.

Robots are increasingly making that much more efficient. As a result, we believe that the discovery of medicine is going to be demonstrably impacted by that. So, that’s number one.

I’ve gone to some pharma facilities where manufacturing is all robots that are moving things around. I think that’s really exciting in the healthcare space and what’s the potential to box and ship more medicines to more patients phenomenally exciting. It’s still fairly in its nascency in terms of adoption and only available to those companies that have the billions of dollars to create these facilities. The question is how do you then bring it to an industry of biotech and earlier stage companies to have access to that?

The other space I’m particularly interested in is the clinical trial space, clinical trials are fraught with failure. It’s really difficult to engage patient groups to jump in and get into a clinical trial. As a result, oftentimes a molecule will go into a Phase Two clinical trial for efficacy, and it will fail. Not because the drug is bad but largely because the readout or the end point just did not come out. And it’s because the patient population was the wrong patient population.

Last one is also the component of how do we change the business model that has prevailed because of the infrastructure that has existed? It’s been costly and it has to be a recoup of money. Now, how can we go tackle problems with AI that have been unavailable to the existing infrastructure because of a poor return on investment? How do we now go after these diseases in Sub-Saharan Africa that nobody wants to touch because frankly, nobody can pay for the medicines. They’re just too costly to go after, or there’s not a market for it. How can we personalized medicine on a global scale?

I’m really excited about all that and I firmly believe AI will open many doors.

2.2. In which area do you expect the biggest breakthrough/ innovation in the next 5-10 years?

I think foremost, all hospitals are going to be re-instituted with an AI first strategy. I don’t think that’s even going to be called AI. It’s just going to be a data centric approach where the entire infrastructure is going to be focused on aggregating, organizing, standardizing its data collection approach.

Currently, going to a hospital and getting data is hard so we’re strongly limited. There’s going to be a huge policy push to get this data approach, standardized, harmonize and accessible.

I think there’s going to be a huge effort on how regulatory agencies, governments, hospital systems and patients think about data. As a result, AI is going to be integral part on how decisions are made from operations to patient adherence, to inpatient surgeries to outpatients.

The other part that I’m really excited about is my industry. I think it desperately needs to change and society is demanding that it changes. 52 drugs are brought to the market every single year on average, yet there are 10,000 untreated diseases. Did you know that the technologies that exist today, AI or other are focused in only on 7% of the total addressable target space?

93% of protein targets are generally not being drugged or not being focused on because the technologies are simply incompatible for going after those targets. As a result, there are millions of patients suffering from those 10,000 plus diseases that are simply not getting their drugs. AI is going after that.

We’re going to see so much more impact and especially on the personalized medicine front. Especially after the human genome project and data started to become available 20 years ago. I think in the next 10 years, truly globally, personalized medicine is going to be achieved and AI is going to be at the center of that.

2.3. If MedTech is set to grow significantly, how do you think MedTech providers can create a competitive advantage?

There are 400 in the AI drug discovery and development space. In 10 years’ time, there aren’t going to be 500 companies. At some point in time, there’s going to be consolidation, there’s going to be a vertical and horizontal integration of technologies and methods. My vision is you can’t be a point solution to the space.

Service models are not going to work that there’s going to be an integration. That integration is going to be a combination of science technology and the right people to use that technology on end-to-end basis.

I believe there’s going to be a lot coming together instead of separate companies trying to do it themselves. And we’re already seeing that. So, in the next 10 years you’re going to see a new type of biopharma company with a new data first mindset.

2.3.1. Are there any security concerns in terms of data protection in regard to AI?

Yes absolutely. This is a big question. There are definitely security concerns and there ought to be because ultimately at some point in the value chain, we’re talking about patient data.

In Canada it’s really hard to touch patient data. In Europe, it’s very hard to touch patient data without the right approvals. Now, you have to tie that to the government policies and the frameworks that are being created.

Currently, there’s a lot of effort going towards implementing blockchain technologies to access that data on a more secure and private basis. Patient privacy, data privacy, data security should be always front and center, but we have to overcome it. We have to be willing to figure out a solution because that’s a huge unmined opportunity.

Eventually, it comes back to security they are inextricably linked. We need to keep it front and center. I want to evangelize, I think the Canadian government is doing a good job with this. I think we can do more. I think the European governments can be doing more. I certainly think the US government is doing a lot. And I know it’s in Congress.

They’re talking a lot about this could be doing more to fund innovation, to enable data harmonization, data security of patient level data, implement blockchain technologies, get that data into the hands of innovative companies and then support those companies to innovate technologies, to have impact.