Dataset Information

An international study presenting a federated learning AI platform for pediatric brain tumors.

ABSTRACT: While multiple factors impact disease, artificial intelligence (AI) studies in medicine often use small, non-diverse patient cohorts due to data sharing and privacy issues. Federated learning (FL) has emerged as a solution, enabling training across hospitals without direct data sharing. Here, we present FL-PedBrain, an FL platform for pediatric posterior fossa brain tumors, and evaluate its performance on a diverse, realistic, multi-center cohort. Pediatric brain tumors were targeted due to the scarcity of such datasets, even in tertiary care hospitals. Our platform orchestrates federated training for joint tumor classification and segmentation across 19 international sites. FL-PedBrain exhibits less than a 1.5% decrease in classification and a 3% reduction in segmentation performance compared to centralized data training. FL boosts segmentation performance by 20 to 30% on three external, out-of-network sites. Finally, we explore the sources of data heterogeneity and examine FL robustness in real-world scenarios with data imbalances.

SUBMITTER: Lee EH

PROVIDER: S-EPMC11368946 | biostudies-literature | 2024 Sep

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

An international study presenting a federated learning AI platform for pediatric brain tumors.

Lee Edward H EH Han Michelle M Wright Jason J Kuwabara Michael M Mevorach Jacob J Fu Gang G Choudhury Olivia O Ratan Ujjwal U Zhang Michael M Wagner Matthias W MW Goetti Robert R Toescu Sebastian S Perreault Sebastien S Dogan Hakan H Altinmakas Emre E Mohammadzadeh Maryam M Szymanski Kathryn A KA Campen Cynthia J CJ Lai Hollie H Eghbal Azam A Radmanesh Alireza A Mankad Kshitij K Aquilina Kristian K Said Mourad M Vossough Arastoo A Oztekin Ozgur O Ertl-Wagner Birgit B Poussaint Tina T Thompson Eric M EM Ho Chang Y CY Jaju Alok A Curran John J Ramaswamy Vijay V Cheshier Samuel H SH Grant Gerald A GA Wong S Simon SS Moseley Michael E ME Lober Robert M RM Wilms Mattias M Forkert Nils D ND Vitanza Nicholas A NA Miller Jeffrey H JH Prolo Laura M LM Yeom Kristen W KW

Nature communications 20240902 1

While multiple factors impact disease, artificial intelligence (AI) studies in medicine often use small, non-diverse patient cohorts due to data sharing and privacy issues. Federated learning (FL) has emerged as a solution, enabling training across hospitals without direct data sharing. Here, we present FL-PedBrain, an FL platform for pediatric posterior fossa brain tumors, and evaluate its performance on a diverse, realistic, multi-center cohort. Pediatric brain tumors were targeted due to the ...[more]

PMID: 39223133

Similar Datasets

Project description:BackgroundThe natural history of pediatric patients with metastasis of primary brain tumors within and outside the central nervous system is poorly understood, as too are possible clinical correlates with outcome. Correspondingly, the aim of this study was to interrogate a national database to characterize this diagnosis and its clinical course in pediatric patients.MethodsThe U.S. National Cancer Database (NCDB) was interrogated between the years 2005-2016 for all patients aged 18 years and younger with a primary brain tumor diagnosis, as well as evidence of disease metastasis at initial diagnosis. Data were summarized and overall survival (OS) was modeled using Kaplan-Meier and Cox regression analyses.ResultsOut of a total of 8615 pediatric brain tumor patients, 356 (4%) had evidence of metastasis at initial diagnosis. Compared to patients without metastasis, patients with metastasis were statistically younger, more often male, and less likely to have private health insurance (all p < 0.050). With respect to clinical characteristics, the primary tumors of patients with metastasis were statistically more likely to be located in the cerebellum; be of higher histologic grading, with a higher proportion of medulloblastoma diagnoses and lower proportion of malignant glioma and pilocytic astrocytoma diagnoses; and were more likely to be treated by subtotal surgical resection, chemotherapy and radiation therapy (all p < 0.050) when compared to patients without metastasis. Five-year OS for those with metastasis was significantly lower than those without (48% vs. 75%, p < 0.001), with the median overall survival for patients with metastasis being 53 months (95% CI 29-86). Multivariate analysis indicated that a shorter OS was independently associated with the primary diagnoses of malignant glioma (HR 27.7, p = 0.020) and Atypical Teratoid/Rhabdoid Tumor (ATRT, HR 41.1, p = 0.041) and with WHO grades 3 (HR 20.1, p = 0.012) and 4 (HR 11.5, p < 0.001). Longer OS was significantly and independently associated with surgery (HR 0.49, p < 0.001), chemotherapy (HR 0.53, p = 0.041), and radiation therapy (HR 0.57, p = 0.026).ConclusionsAlthough uncommon, pediatric brain tumors with evidence of metastasis at initial diagnosis will present with a distinct socioeconomic and clinical profile compared to patients without metastasis. Multiple predictors are independently associated with overall prognosis, and understanding these features should be validated in prospective efforts to identify vulnerable patients earlier in order to maximize the impact of treatment.

Project description:BackgroundMachine learning and artificial intelligence have shown promising results in many areas and are driven by the increasing amount of available data. However, these data are often distributed across different institutions and cannot be easily shared owing to strict privacy regulations. Federated learning (FL) allows the training of distributed machine learning models without sharing sensitive data. In addition, the implementation is time-consuming and requires advanced programming skills and complex technical infrastructures.ObjectiveVarious tools and frameworks have been developed to simplify the development of FL algorithms and provide the necessary technical infrastructure. Although there are many high-quality frameworks, most focus only on a single application case or method. To our knowledge, there are no generic frameworks, meaning that the existing solutions are restricted to a particular type of algorithm or application field. Furthermore, most of these frameworks provide an application programming interface that needs programming knowledge. There is no collection of ready-to-use FL algorithms that are extendable and allow users (eg, researchers) without programming knowledge to apply FL. A central FL platform for both FL algorithm developers and users does not exist. This study aimed to address this gap and make FL available to everyone by developing FeatureCloud, an all-in-one platform for FL in biomedicine and beyond.MethodsThe FeatureCloud platform consists of 3 main components: a global frontend, a global backend, and a local controller. Our platform uses a Docker to separate the local acting components of the platform from the sensitive data systems. We evaluated our platform using 4 different algorithms on 5 data sets for both accuracy and runtime.ResultsFeatureCloud removes the complexity of distributed systems for developers and end users by providing a comprehensive platform for executing multi-institutional FL analyses and implementing FL algorithms. Through its integrated artificial intelligence store, federated algorithms can easily be published and reused by the community. To secure sensitive raw data, FeatureCloud supports privacy-enhancing technologies to secure the shared local models and assures high standards in data privacy to comply with the strict General Data Protection Regulation. Our evaluation shows that applications developed in FeatureCloud can produce highly similar results compared with centralized approaches and scale well for an increasing number of participating sites.ConclusionsFeatureCloud provides a ready-to-use platform that integrates the development and execution of FL algorithms while reducing the complexity to a minimum and removing the hurdles of federated infrastructure. Thus, we believe that it has the potential to greatly increase the accessibility of privacy-preserving and distributed data analyses in biomedicine and beyond.

Dataset Information

An international study presenting a federated learning AI platform for pediatric brain tumors.

Publications

An international study presenting a federated learning AI platform for pediatric brain tumors.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets