Photo by Darren Baker
An international competition has produced models that can help predict outcomes in patients with acute myeloid leukemia (AML), according to researchers.
For the competition, known as the DREAM 9 challenge, 31 teams of computational researchers attempted to predict outcomes using data from hundreds of patients with AML.
DREAM, which stands for Dialogue for Reverse Engineering Assessment and Methods, is a platform for crowd-sourced studies that focus on developing computational tools to solve biomedical problems.
Essentially, it’s a competition that serves as a large, long-standing, international scientific collaboration.
“We used DREAM as a way to get general insight into making more accurate predictive models of clinical outcomes,” said Amina Qutub, PhD, of Rice University in Houston, Texas.
She and her colleagues described this effort in PLOS Computational Biology.
For the DREAM 9 challenge, each team was presented with training data from 191 AML patients, which included demographic information, such as age and gender, and more complex proteomic and phosphoprotein data that describes signaling protein pathways believed to play a role in AML.
The teams were also presented with a test set of 100 AML patients and were asked to predict response to therapy, remission duration, or overall survival for these patients.
The top-performing models—by Team EvoMed of Arizona State University and Team Chipmunks of the Ontario Institute for Cancer Research—were able to predict patient response to therapy with an accuracy of close to 80%.
Both of these models were impacted by the perturbation of PIK3CA and NPM1, which singles out these proteins as candidates for further study, according to researchers.
Another discovery resulting from this competition was that, overall, the 31 models were not as effective for predicting outcomes in patients classified as “resistant to therapy” than for responsive patients.
The median model prediction accuracy was 42% for resistant patients and 73% for responsive patients.
