Clinical features outperform MRI radiomics for brain metastasis survival prediction, except in melanoma

In a study of 422 brain metastasis patients, clinical model best predicted survival (accuracy 0.72). Radiomic MRI model underperformed overall (accuracy 0.52) but outperformed in 54 melanoma patients (accuracy 0.77, 80% low-risk and 67% high-risk correctly classified).

• Overall, clinical-only bootstrap aggregating model achieved test set macro F-1 0.62 and accuracy 0.72 for survival risk stratification in brain metastases.
• In melanoma brain metastases (n=54), radiomic features from MRI predicted survival better, with macro F-1 0.71 and accuracy 0.77; 80% low-risk and 67% high-risk correctly classified.

Automated summary

RadPigeon summaries are original and for information only. They are not clinical advice.