SEATTLE – For the last 40 years, doctors have treated patients with acute myeloid leukemia, or AML, with a few chemotherapy drugs. Only 25 percent of them survive for five years. Now, researchers at the University of Washington have brought computers and artificial intelligence into the mix.
Candice Ferro was diagnosed with AML about a year ago. She had chemo, radiation and a bone marrow transplant.
“I feel so much better, like healthier,” Ferro said.
Her doctor chose Ferro’s drugs because of a gene mutation she has and may soon get to use even more precise information. Professor Su-In Lee, Ph.D, professor of computer science and engineering at University of Washington, and her team of researchers are developing a cancer algorithm called Merge.
“Merge is an artificial intelligence algorithm that can automatically learn from large amounts of data and complex biological knowledge how to choose the best drug for an individual cancer patient,” said Lee.
Merge gets its data from Lee’s study of 42 AML patients and from many large-scale studies funded by the NIH. Candice’s doctor then tests leukemia cells against 150 drugs and drug combinations.
“After 72 hours, we would check if the cells survived those drugs or didn’t survive those drugs, and then we would find out which drugs might work best for that patient,” said Pamela Becker, MD, Ph.D, professor of medicine at University of Washington.
Lee says Merge is a successful algorithm, partly because there’s so much patient information to use.
“Data are becoming more and more available,” Lee said.
Ferro has been in three clinical trials that could bring precision medicine for AML even more quickly.
Now, she feels free to make plans with her son.
“I feel like I got a lot of time to make up for!” Ferro said.
Dr. Becker and Professor Lee’s collaboration is only in the lab right now. They are looking forward to new trials, both to find out why the algorithm chooses certain drugs, and to eventually test it on patients who are newly diagnosed.