Federated learning enables big data for rare cancer boundary detection
Федеративное обучение обеспечивает использование больших данных для выявления границ редких опухолей
2022-12-05
SCID: 54.1/tbbeescd
Discuss with AI
Federated learningGlioblastomaMulti-site dataTumor boundary detectionTumor delineation
Figures from the paper
Abstract (AI)
Although machine learning (ML) has shown promise across disciplines, out-of-sample generalizability is concerning. This is currently addressed by sharing multi-site data, but such centralization is challenging/infeasible to scale due to various limitations. Federated ML (FL) provides an alternative paradigm for accurate and generalizable ML, by only sharing numerical model updates. Here we present the largest FL study to-date, involving data from 71 sites across 6 continents, to generate an automatic tumor boundary detector for the rare disease of glioblastoma, reporting the largest such dataset in the literature (n = 6, 314). We demonstrate a 33% delineation improvement for the surgically targetable tumor, and 23% for the complete tumor extent, over a publicly trained model. We anticipate our study to: 1) enable more healthcare studies informed by large diverse data, ensuring meaningful results for rare diseases and underrepresented populations, 2) facilitate further analyses for glioblastoma by releasing our consensus model, and 3) demonstrate the FL effectiveness at such scale and task-complexity as a paradigm shift for multi-site collaborations, alleviating the need for data-sharing.
Key Findings
1
A federated learning study across 71 sites on six continents created an automatic glioblastoma tumor-boundary detector using the largest reported dataset, comprising 6,314 cases.
2
Delineation of the complete tumor extent improved by 23% relative to a publicly trained model.
3
Federated learning enabled multi-site model training by sharing numerical model updates rather than centralized patient data.
4
The federated model improved delineation of surgically targetable tumor by 33% compared with a publicly trained model.
5
The study demonstrates that federated learning can support large-scale, complex collaborations for rare diseases while reducing the need for direct data sharing.
Research Object
glioblastoma tumor boundaries, including the surgically targetable tumor and complete tumor extent, across multi-site imaging data
Research Subject
the accuracy and out-of-sample generalizability of automatic tumor boundary delineation using federated learning across 71 sites
Publication Details
Publication Date
2022-12-05
Journal
Publisher
ISSN
Cited by
385
Open access PDF
Access Type
Author Information
Download PDF
Subscribe to digest