Dr. Lynda Chin, Co-founder, President and CEO of Apricity Health on how she became a pioneer in precision oncology, the data sharing challenge, and how not to build an ugly bookcase!

What initially sparked your interest in medicine?

I was into mathematics and physics when I was in high school, and I wasn’t sure whether I should go to business school or into medicine. I ended up studying a combination of computer science, engineering, and biology in college, taking full advantage of Brown’s open curriculum. Those years were amazing. I liked doing things across different domains, such as catching and training bats in behavioral experiments to record their sounds with hardware I built myself, then analyzing the sound with software I wrote myself…  the experience sparked my interest in human physiology more broadly, so I decided to go to a medical school rather than PhD training in graduate school.

It’s interesting that you didn’t go on to pursue neuroscience in medical school but ended up in dermatology and subsequently, oncology. How did that happen?

I continued doing research during medical school, working first in an infectious disease lab, then a molecular biology lab. I seriously thought about going into surgery or emergency medicine, but in the end, decided on dermatology because I know I’d enjoy the intellectual challenges of making a clinical diagnosis and love the opportunity to do office-based surgery; moreover, I wanted to select a discipline that would give me time to do research.

As a board certified dermatologist in practice, I started a research program studying melanoma in mouse models.  Cancer genomics opened the door to studying cancer on a completely different level across all tumor types, starting with GBM, the first project in the Cancer Genome Atlas (TCGA) project.  So I became the biologist in the genomic community and the genomic scientist in the biology community – it was most interesting to straddle across these domains.

How do you balance your work and life as a scientist, a wife, a mother, a businesswoman and a healthcare leader?

I don’t like the term balance… Balance suggests that one can have it all, doing it equally well on all fronts, and not feeling as if having missed out on anything. In reality, it’s a tradeoff because we only have 24 hours a day, like everyone else. So when we choose to do something, whether to attend a school parent-teacher conference or a committee meeting, we are choosing not to do something else.  As women at work, we choose, we make tradeoff. There’s no magic here, just hard work and discipline to choose or prioritize!  This is very individual as it is no one-size fits all.

I don’t like the word “balance”, as I have seen many women who regard themselves as failure because they don’t feel “balance”, as they may feel guilty about missing out on family time, or left out of a promotion opportunity for missing key meetings at work.   I prefer “trade-off” or “joggle”, because that means it is OK to feel guilty sometimes for not being able to attend a school concert your child is playing at, or OK to feel disappointed to let go of a work project that could be an advancement opportunity because you want to spend time with family.  Acknowledging that these are difficult decisions helps one joggles the demand from work and life in a way that is right for that person.

How do you relax?

Recently, I’ve just been watching a lot of TV and I also enjoy working out. I was trained in martial arts when I was in medical school so I may take that up again one day, together with the piano. They could be my new (old) hobbies.

When did you become involved in medical AI?

I don’t like labeling myself as being in Medical AI.  Advanced analytics, data and technologies, or EKG or a CT scanner… they are all tools that empower us to study human pathology and practice medicine.

As genomic technology matures to generate larger and larger datasets with increasing resolution, as digital and cloud technologies enable tracking of human behaviors, such as activity or sleep, we need new tools to help make sense and make use of these data.  That is where advanced analytics or medical AI comes in.  It is a set of tools, necessary and powerful.  But in isolation, it is no more useful than hammers and nails.  If I do not know what the purpose of a bookcase is, or I do not have a clue as to how someone would use a bookcase, I would not build great bookcase, even though I could be trained to use a hammer and nails..

The same for AI. We need more than efficient AI tools, as merely having medical AI tools will not improve human health.  We need experts who know how to use these tools, in context of all other tools, and to know what outcome they are looking for.

That’s why I think it’s so important that AIMed is growing a community of physicians who understand both medicine, what data can tell them about their patients and how AI can help them use data to better care for their patients. They are the ones who know how to design and build a beautiful bookcase!

A growing number of researchers rely on synthetic data to overcome the data challenges you mention. Thoughts?

In science, we are constantly developing and using model systems to explore, learn and elucidate.  From cell lines, mice to human on clinical trials, they represent model systems that we use.  Each model has pro and con, because they do not recapitulate fully the diversity and complexity of real world patients.  I think of data in the same way.  Data sets are useful, but each has strength and limitation.  As scientists, we must always be mindful of such, and should never rely on a single model or dataset to draw conclusions, because they would likely be biased and non-generalizable.  So, synthetic data can be very useful, yet I would never consider it the same or equivalent to real world patient data.  That is not to say that we should always use real world data for every analysis, because there are limitation to real world patient data that make them unsuitable for a certain types of analyses.  Thus, we need to have quality and diverse datasets, including synthetic ones, to enable and empower analytic development and insight discovery.

How do you think we can overcome the reluctance to share data?

There is a big challenging problem with no single solution.   On one end, there are security, privacy and trust concern; on the other end, value and incentive (mis)alignment.  On top of this, there remains a paucity of robust evidence of scientific validity and patient benefit derived from shared data.  So, lots more to do.

To accelerate progress in this area, I do believe that the field would benefit greatly from having a digital highway system as a foundational infrastructure that can enable data access and data sharing.  Just like an interstate highway that anyone can drive on, no matter one is driving a diesel truck, gasoline SUV or electric care, a universally accessible digital highway would enable equitable access to data, facilitate discovery of unbiased insights from big data, accelerate AI medicine.   Sean Khozin and I have written a perspective on this topic.  I welcome all of you to read and send us comments on this.  The link is https://pubmed.ncbi.nlm.nih.gov/34062153/

What advice would you give to someone starting their career in medical AI?

I give the same advice to all the postdoctoral trainees who come to my lab. Do not tell me what technology you want to use or what model you want to build. Tell me what problem you need to solve. AI is part of the tool kit, a powerful tool, but nonetheless, one of the many tools.  Focus on the problem, and use the best tools (which may or may not be AI), to solve it.

What’s the best piece of advice you ever received?

Well, there have been many great advice from friends, mentors, peers, and colleagues…. I would cite the one from Sir Professor Mike Stratton, who led the Cancer Genome Project at the Wellcome Sanger Institute.  When I was uncertain about getting into cancer genomics and joining the TCGA consortium, Mike said to me, “Lynda, the cancer genome will only be sequenced once”. It was a piece of advice but also an encouragement. It stays with me as it reminds me that, opportunity is not just for the prepared, it’s for the willing and the brave.  You would miss the leading edge if you are waiting till you feel fully prepared before you innovate…

What’s left to conquer?

There is no end… right now, I am learning how to be an entrepreneur. In another five years I will be ready for something else.