Forecasting children’s tumor with an unprecedented accuracy by artificial intelligence

American scientists have developed an artificial intelligence -based system that, by analyzing consecutive MRI scans, predicts the possibility of a brain tumor return to children.

Designed in collaboration with Massachusetts Hospital, Boston Children’s Hospital, and Dana-Farber Cancer Center, the system uses a new “time learning” approach.

This method identifies the initial symptoms of tumor recurrence with a accuracy of 1 to 2 percent to one year after surgery, which can reduce the need for recurrent imaging and family anxiety.

This technology detects hidden patterns that may stay away from ophthalmologists by analyzing 2 MRI scans of 5 children. Unlike traditional methods that examine individual scans, time learning observes gradual changes over time and increases prediction accuracy.

The researchers hope that with clinical tests, the effectiveness of the system will be approved to improve medical care by reducing imaging for children and early intervention for high -risk cases.

This progress promises to wider applications of artificial intelligence in medical imaging.