Deep learning model effective at predicting survival in patients with multiple brain metastases

A deep learning model – using a machine learning method to model non-linear relationships between pixel-level imaging predictors and survival data – has been shown to be effective at predicting survival in patients with multiple brain metastases, researchers have reported.

The model outperformed traditional Cox proportional hazard (CPH) models based on linear relationships between clinical factors and survival.

Dr Enoch Chang from the Department of Therapeutic Radiology at Yale School of Medicine, in New Haven, Connecticut, USA, presented the new findings at the SNO 2020 virtual conference.