Helping healthcare professionals find better ways to treat cancer

Landing a spot on the roster of 2016 IBM Fellows is the latest in a lifetime of achievements in biological research for Ajay Royyuru, director of Healthcare & Life Science Research for IBM.

“In high school, and even as an undergraduate in college, I knew computers existed, but I had no access to them,” Royyuru says. “Then, in my second year, my biophysics professor, Dr. Vidya Kothekar, taught by having us do research,” he continues. “We would be in her office, and she would say, ‘I want you to think about this,’ and we were to go research it in the library and come back ready to discuss it. I was like a kid in a candy store, with this instructor who pushed us to experience things instead of just reading about them. It was a dramatic learning experience.”

Kothekar taught her students the rudiments of computational biology research, setting Royyuru on a path to marry science and technology, and laying the groundwork for the work he’s doing with IBM now.

Royyuru, who grew up in India, says he was always a biology buff and planned early for a career in research biology. He attended the All India Institute of Medical Sciences for his Bachelor of Science degree in human biology — a rare opportunity at a time when most biology programs focused on general botany and zoology. “It was an awesome experience, to be able to immerse myself that way in the biology that really matters, at that formative period of my life,” he says.

Portrait of Ajay Royyuru


I was like a kid in a candy store, with this instructor who pushed us to experience things instead of just reading about them.

Treat patients, not body parts

For as long as doctors have been treating cancer patients, the strategy has been to focus on the tissue where the cancer resides. So, for example, a chemotherapy drug approved to treat breast cancer would be different than a drug approved to treat liver cancer.

Today, the focus is shifting to treatment based on the genetic characteristics of the cancer cells themselves, instead of the body part where they reside.

Royyuru has been leading a team for three years dedicated to helping healthcare professionals find better ways to treat cancers based on their genomic characteristics.

He explains that cancer develops when the DNA in normal cells becomes corrupt, usually after many replications. “Imagine that you have this sheet of exquisite text,” he says. “And you make a copy of it on a very high-quality copier. Then you make a second copy from that copy, and then a third from the second. And so on. By the time you get to the millionth copy, that text is going to start to look slightly flawed.

“Similarly, that normal replication process would accumulate a lot of errors over the generations of cell replication, but cells have correcting mechanisms that act to proofread and correct those errors, ensuring that each cell is a high-fidelity copy of the original, which is why most cells do not become cancer cells. But if the proofreading mechanism itself develops an error — like a scratch on the glass of the photocopy machine — then every copy is going to have errors.

The focus is shifting to treatment based on the genetic characteristics of the cancer cells themselves, instead of the body part where they reside. 

“Cancer cells are marked by these genomic variations. In the beginning there may be one or two variations, but once there are errors in the copy machine itself, there may be explosive growth in the genetic errors that are produced. Very late-stage cancer cells look so bizarre they’re unrecognizable. They’re just off on their own and they become very hard to control. But if you can find a variation early, you have an opportunity to produce an exceptional response.”

Image of microbes magnified by electron microscope

How Watson fits in

“Using Watson for Genomics, we’re building a system of insight to translate genomic variation into cancer treatment decisions,” Royyuru explains. “Today, 15 leading medical institutions use it as a research tool to inform deliberations on new cancer patients. Clinicians can look at the patient’s raw genetic evidence to determine personalized treatment options.”

What that means for clinicians and patients is treatment targeted to the specific genetic anomaly present in a patient. “You hear about miraculous cures where a patient’s cancer just disappears,” he says. “We thinks of those patients as exceptional responders, and that response happens when the treatment is exactly right for the cancer. For example, there’s a famous variant called BRAF-V600E. It typically occurs in melanomas, and there’s an approved treatment that hits it bang-on and pretty much knocks it out. But that mutation can occur in any cell at any point in time. If it occurs in a liver cell, for example, it will still respond to the same treatment.

“Precision medicine is moving to a place where literally every cancer patient is going to say ‘Look at my cancer, not at my tissues’,” adds Royyuru. Corresponding to this shift in treatment approach is an explosion of knowledge about those genomic aberrations and which treatments work on each of them.

“Fifty years into the war on cancer, our knowledge base is exploding. Tens of thousands of new journal papers are being published every year, and the result is cognitive overload. If you are a practicing oncologist, you have this flood of new knowledge coming at you, and you have to make a quick decision — what of this knowledge is applicable to this variation and how do I treat it?” he says. “It’s a very difficult task, matching the cancer with the best treatment, and doing it quickly. Cancer doesn’t wait.”

“With Watson for Genomics, we’ve trained Watson to do the matching between the genomic data for a specific patient and the knowledge base. We’ve created a solid solution that takes the genomic data from the patient’s sample, analyzes it against the entirety of available knowledge and gives you treatment options in five minutes.

“Watson is giving doctors a whole slew of options with evidence mined from the scientific literature. Think of Watson as a very capable research assistant. Watson reads and doesn’t forget.”

Watson for Genomics is available to institutes via the IBM Cloud, and its knowledge base is regularly updated with the newest scientific literature to ensure doctors and their patients the benefit of the latest findings in scientific literature.

Image of full length portrait of Ajay holding a model of a DNA helix


Think of Watson as a very capable research assistant. Watson reads and doesn’t forget.

More on the intersection of healthcare and technology

Since joining IBM nearly 20 years ago, Royyuru has worked with institutions around the world on research to advance personalized, information-based medicine. He previously led the life sciences research portfolio through the Computational Biology Center and led the IBM Research team collaboration with National Geographic Society on the Genographic Project.

Royyuru has authored numerous research publications and several patents in structural and computational biology. His work has been featured in FortuneThe New York TimesThe Washington Post, BBC, Forbes, Scientific American, Nature Medicine and Nature.

As director of Healthcare & Life Sciences Research for IBM, Royyuru oversees multiple teams investigating the opportunities that lie at the intersection of technology and healthcare. One exciting new project is a partnership with Pfizer that uses wearable technology to continually monitor the symptoms of Parkinson’s patients. “The problem in the field of chronic neurologic diseases like Parkinson’s is that the symptoms really vary minute by minute,” he says. “Through this collaboration, we’re looking to be able to fine-tune treatment as symptoms wax and wane.”