Federated learning enables big data for rare cancer boundary detection

Sarthak Pati, Ujjwal Baid, Brandon Edwards, Micah Sheller, Shih‐Han Wang, G. Anthony Reina, Patrick Foley, Alexey D. Gruzdev, Deepthi Karkada, Christos A. Davatzikos, Chiharu Sako, Satyam Ghodasara · 2022

Read Federated learning enables big data for rare cancer boundary detection by Sarthak Pati, Ujjwal Baid, Brandon Edwards, Micah Sheller, Shih‐Han Wang, G. Anthony Reina, Patrick Foley, Alexey D. Gruzdev, Deepthi Karkada, Christos A. Davatzikos, Chiharu Sako, Satyam Ghodasara (2022) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/openalex-w4310852955.txt.

Description

Sarthak Pati, Ujjwal Baid, Brandon Edwards, and colleagues show how federated learning can pool brain tumour data without moving it. Their study appeared in Nature Communications in 2022. Rare cancers give each hospital few cases to train detection software. Pooling scans centrally is hard: privacy law, ownership, and sheer file size stand in the way. Federated learning instead trains models at each site and shares only model updates.

The team applied the method to glioblastoma, an aggressive brain cancer, focusing on detecting the tumour's extent in scans. Models trained this way can match centrally trained ones, the authors note from prior work. Learning from larger, more varied data should sharpen boundary detection. The approach also opens collaboration to hospitals that could never share raw patient data.

Catalog record

CREATORSarthak Pati, Ujjwal Baid, Brandon Edwards, Micah Sheller, Shih‐Han Wang, G. Anthony Reina, Patrick Foley, Alexey D. Gruzdev, Deepthi Karkada, Christos A. Davatzikos, Chiharu Sako, Satyam Ghodasara
YEAR2022
COLLECTIONPAPERS
FORMATTXT
FILE SIZE115 KB
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEOpenAlex / Nature Communications

Source: https://doi.org/10.1038/s41467-022-33407-5

Browse the full library · About the Tower of Babel