An artificial intelligence (AI) model built based on DNA methylation profiling provided a look at outcomes for children with medulloblastoma and ependymoma, according to study findings published in Nature Genetics. The model may be used to guide personalized treatment decisions, including treatment intensity, to balance neurocognitive effects with survival goals.
“Our goal is to move beyond broad risk categories and toward a more individualized understanding of each child’s tumor,” said senior study author and corresponding author Volker Hovestadt, PhD, a computational biologist at Dana-Farber Cancer Institute. “For children with brain tumors, that could help us better match treatment intensity to the biology of the disease and, ultimately, reduce long-term side effects for some patients.”
Background and Study Methods
Researchers developed a precision AI tool called MANTIS (Methylation profiling and Artificial Neural networks for Time-dependent Individualized Survival predictions) for predicting outcomes in pediatric patients with medulloblastoma and ependymoma.
The tool was trained on genome-wide DNA methylation array data, including both DNA methylation patterns and copy number changes, from 2,786 patients to generate individualized survival probabilities over a 10-year period.
Key Findings
MANTIS achieved concordance indices up to 0.80 ± 0.07, which was above the predictive accuracy of current clinical indicators.
MANTIS was also adapted to test radiation therapy treatment intensity to predict the effects across different dose levels. In this case, it found that about about 60% of patients with average-risk medulloblastoma would not have a significant survival disadvantage (P = .683) when treated with less-intensive, lower-dose craniospinal irradiation, which could potentially result in better neurocognitive outcomes.
“One exciting aspect of this work is that it shows how AI can learn from molecular tests that are already part of care,” said first author Sabina Stefan, PhD, a postdoctoral researcher at Dana-Farber Cancer Institute. “With further clinical study, this type of approach could help support more personalized treatment decisions not only for pediatric brain tumors, but potentially for many other cancers as well.”
The study authors suggested that MANTIS supplies a foundation for AI-based precision prognostics and therapy personalization in pediatric brain tumors and provides a blueprint for doing the same for other cancer types.
DISCLOSURES: The study was supported by the St. Jude Research Collaboratives program. Further support was provided by the Charles H. Hood Foundation, the Children’s Cancer Research Fund and the V Foundation. For full disclosures of the study authors, visit nature.com.
