The rise of AI in precision medicine
Personalized health has to be one of the most exciting and promising areas in my industry. For those unfamiliar with the term, it broadly refers to individualizing care based on information we know about the patient. It’s a change in paradigm from the prevailing medical approach, where we treat patients based on how the population average responds. Under that regime, you might be broadly effective across a population, but the chance that any individual may have no response (or even a negative response) is a very real possibility.
Enter precision medicine. Precision medicine is a subset of this personalized movement, and generally refers to individualizing the therapy (often pharmaceutical) based on the genomic profile of the patient and our understanding of how that will modulate the response of the patient. We’ve of course been aware of the importance of genomics in treatment for decades, the classic case being how people of different race respond to antihypertensive treatment. The difference now however is the relative affordability of genomic sequencing, and the skyrocketing cost of certain drug therapies.
We see a lot of the early applications of precision medicine surfacing in modern oncology, but the potential for wholesale system level change is clear. It was only relatively recently that we first sequenced the genome for hundreds of millions of dollars in 2000. Now genomic sequencing is readily available for about a thousand dollars, with projections suggesting that will further drop to a hundred dollars in fifteen years. When the cost of chemotherapy regularly totals tens of thousands of dollars for any given patient, doesn’t it make sense to find out ahead of time whether the drug will be effective?
So, it was with excitement that I accepted the invitation to come speak on artificial intelligence at the Precision Medicine Leaders Summit in San Diego recently. I was most excited about the learning opportunity, to understand more about this rapidly progressing field, and perhaps meet some of the luminaries. San Diego is a pretty fabulous place to visit in summer as well. As I sat though sessions leading up to the panel I was increasingly humbled by just how much I didn’t know about this burgeoning industry, and just how smart the people attending the conference were. I’m familiar with events that dumb down the subject matter and therefore never explore the thorny issues, and also with other events that entertain without much substance. This conference was a standout for me. Issues were 
debated robustly in panels of highly qualified individuals that blended commercial, technical, clinical and academic perspectives. It’s clear that precision medicine is here to stay, growing fast, and there are plenty of issues to sort through – regulatory, business model, ethical and technical.
I’m relieved to report that the panel not only addressed some of the aspects of AI for this field that really need further public debate, but that it was actually fun. Kudos to my fellow panelists Dr. Atul Butte, Andy Bartley and Dr. Nick van Terheyden for a lively discussion, our excellent moderator Dr. Thomas Wilckens. A special thanks to Adrijana Kekic from the Mayo Clinic for the photo!
I leave with a new appreciation for the importance of the work we do within the Genomics group at Microsoft. Their focus on secondary and tertiary analysis of genomics is but a small part of the larger personalized health movement that some of our Microsoft partners are increasingly making a focus of their product strategy.
If you’d like to learn more about the work that Microsoft is doing in the genomics space, visit this link and download the free eBook.
