back arrow
June 25, 2019

Mapping tumor evolution: an interview with Cambridge Cancer Genomics

#
biotech
#
life science
Book lab instruments

By Akash Patel, Outreach Manager and Writer, Science Entrepreneur Club

I sat down with John Cassidy, CEO of Cambridge Cancer Genomics (CCG.ai), to talk about his journey. I heard about CCG.ai a while ago but really started to pay attention when they hit the Forbes ‘30 under 30’ list for Europe. They’re an exciting startup building a Precision Oncology platform using AI and Machine Learning that can be integrated into clinical workflow. With this, they can precisely map tumor evolution and tailor cancer therapies in accordance with the stage of evolution. Tumors are incredibly heterogeneous, genetically unstable and genetically change in response to therapy - hence CCG.ai's ambition to map their evolution. By carrying out DNA and RNA sequencing they unpick the tumors' biology and then correlate it with the therapeutics most likely to be effective. Finally, they take liquid biopsies to measure treatment response and predict the patients’ response to treatment going forward.

What made you decide to start Cambridge Cancer Genomics?

We started the company about three years ago when, both my co-founder and I, were at the University of Cambridge − me finishing my PhD and my co-founder doing a postdoc. During this time we were working on a virtual reality tumor project, building models and spending a lot of time at the hospital where most of the computers were running Windows 95-XP. There was no way that a virtual reality tumor model was going to make its way from the lab into the clinic, especially with those computers. So we started to work out which types of tests that we had developed in the lab could be translated into the clinic and what kind of fundamental understanding we had as scientists which wasn’t being represented in cancer medicine. One of these understandings was tumor evolution and tumor heterogeneity, which is broadly the origin of our current work.

We’re seeing an increasing number of AI cancer startups in Europe and the US. What makes you different?

Lab booking system

The fundamental thing to understand is that most startups claiming to use AI aren't using AI. Amongst the companies that are using AI are some that focus on drug discovery and modeling how drugs bind to their targets. We’re more based on: once we have a drug, how do we use it in the best way? It's all well and good to have a new drug but it may only be effective for 5% of the population, so how do we find that 5%? How do we ensure that the tumors are in the right kind of state to benefit from these new drugs? For leukemias, for example, there are ten treatment options and the doctor has about 20 minutes to decide which drug to give to the patient for the following six months. How do we use AI to determine the best drug choice? How do we develop this drug so that − say in 15-years time when there are 30 drugs for leukemia − we can decide which drug is best to give to which patient? That’s where we focus.

So you’re building this platform for precision oncology and looking to get it into hospitals, but how have you tackled selling to the NHS?

Yeah, so this is progressing and what we’re doing is being less worried around monetization and more focused on achieving product-market fit − understanding what users want. The most important thing is finding organizations and people that share our values and want to champion our cause − Genomics England, for instance, and the PRECISION-Panc program in Glasgow. These partnerships have helped to push our product and it's the way we get our software directly into the NHS and impacting patients.

But, having said that, adoption of new tech into the NHS can take decades. So there is definitely room to work faster and people are realizing that this disruption is coming anyway. Healthcare can embrace it and work with it, or ignore it, but it’s still going to be affected. For example, if you look at Babylon Health in London, without really working with the NHS, they came in and built an online GP practice and took away all the young patients who were really putting a huge strain on the NHS. The reality is that startups will just come in and disrupt, whether the NHS is willing to accept it or not. I think if you want to manage this process there needs to be more engagement between startups and the NHS.

Having been through the Silicon Valley-based startup accelerator, Y Combinator, what did you make of the experience?

We joined Y Combinator about six months after forming CCG.ai and it was a fantastic experience. Accelerators are almost essential unless you have done it before or you’re some well-known professor who’s been running a lab for years. Y Combinator was perfect for us who were working on products with data and building our partnerships, and we also learned how to talk about our science in a way that investors could understand it.

In the US, investors have a bigger appetite to put more money into considerably riskier investments than in the UK. So traditionally ventures in the UK go in and say “we are going to double your investment in three-years time.” Startups say this as if it’s a good thing, but for a US investor, that’s terrible. They want a one in ten chance of your company turning into a billion-dollar one. If not they don’t care. A one- or two-times return on the original investment barely makes a difference to their portfolio. What makes a difference is a 100-times return. That kind of mentality makes the US investors much more ambitious about what they do.

As a researcher turned entrepreneur, what was that journey like?

Lab equipment scheduling software

I was a wet lab biologist, I did my PhD on drug-resistant populations in breast cancer. The moment when I realized that academic focus was really different from entrepreneurial thinking, was when someone presented a paper on a new drug molecule that could inhibit metastasis of prostate cancers. But I realized that it was never going to get to market as there was no patent on it and no way of commercially exploiting it. It was just another discovery that was going to be published and that was it. It led me to understand that sometimes if you want your science to get out there and have the best impact don’t publish a paper but start a commercial endeavor instead.

Forming a startup is really not that different from being a postdoc. People think that working in a startup is really risky, but the reality is that it’s no riskier than working in a lab where you’ve got only two years of funding. In fact, my job right now is pretty similar to that of a research group leader. I spend a lot of my time speaking to investors or writing grants, and I sit in a lot of meetings trying to guide people in the right directions throughout the business.

What elements of the entrepreneurial ecosystem in Cambridge were important to your journey?

It’s the network. We’re building an ecosystem focused around AI called camb.ai and we have monthly dinners for startups and MBAs, which are very free and collaborative. You can have a pint with someone who does laser research and perhaps you come up with an idea such as “let's use lasers to target cancer!” You make these serendipitous connections, and that's the key to it all.

To find out more take a look at www.ccg.ai

Related posts

References