A Systematic Review of Federated Learning in the Healthcare Area: From the Perspective of Data Properties and Applications

Prayitno Prayitno,Karisma Trinanda Putra,K S M Tozammel Hossain,Chi-Ren Shyu,Zon-Yin Shae,Hsing-Chung Chen,Yuan-Yu Tsai,Wei Jiang

doi:10.3390/app112311191

Abstract

Recent advances in deep learning have shown many successful stories in smart healthcare applications with data-driven insight into improving clinical institutions’ quality of care. Excellent deep learning models are heavily data-driven. The more data trained, the more robust and more generalizable the performance of the deep learning model. However, pooling the medical data into centralized storage to train a robust deep learning model faces privacy, ownership, and strict regulation challenges. Federated learning resolves the previous challenges with a shared global deep learning model using a central aggregator server. At the same time, patient data remain with the local party, maintaining data anonymity and security. In this study, first, we provide a comprehensive, up-to-date review of research employing federated learning in healthcare applications. Second, we evaluate a set of recent challenges from a data-centric perspective in federated learning, such as data partitioning characteristics, data distributions, data protection mechanisms, and benchmark datasets. Finally, we point out several potential challenges and future research directions in healthcare applications.

Highlights

Deep learning technology has shown promising results in smart healthcare applications to assist medical diagnosis and treatment based on clinical data
To increase both the diversity and quantity of training data is through the collaboration of several healthcare institution to create a single deep learning model while maintaining patient privacy and confidentially
We review the current progress on federated learning in the healthcare field

Summary

Introduction

Deep learning technology has shown promising results in smart healthcare applications to assist medical diagnosis and treatment based on clinical data. Good performance of the deep learning model on smart healthcare applications highly depends on a diverse and vast amount of training data [9]. These training data were obtained from various clinical observations such as biomedical sensors, individual patients, clinical institutions, hospitals, pharmaceutical industries, and health insurance companies. This institutional data bias has been shown to have high accuracy when evaluated on the same clinical institution’s data. To increase both the diversity and quantity of training data is through the collaboration of several healthcare institution to create a single deep learning model while maintaining patient privacy and confidentially

Objectives

Results

Discussion

Conclusion