Maintenance Notice

Due to necessary scheduled maintenance, the JMIR Publications website will be unavailable from Wednesday, July 01, 2020 at 8:00 PM to 10:00 PM EST. We apologize in advance for any inconvenience this may cause you.

Who will be affected?

Accepted for/Published in: Journal of Medical Internet Research

Date Submitted: Jul 16, 2020
Date Accepted: Nov 6, 2020

The final, peer-reviewed published version of this preprint can be found here:

Web-Based Privacy-Preserving Multicenter Medical Data Analysis Tools Via Threshold Homomorphic Encryption: Design and Development Study

Lu Y, Zhou T, Tian Y, Zhu S, Li J

Web-Based Privacy-Preserving Multicenter Medical Data Analysis Tools Via Threshold Homomorphic Encryption: Design and Development Study

J Med Internet Res 2020;22(12):e22555

DOI: 10.2196/22555

PMID: 33289676

PMCID: 7755539

Web-Based Privacy-Preserving Multicenter Medical Data Analysis Tools via Threshold Homomorphic Encryption

  • Yao Lu; 
  • Tianshu Zhou; 
  • Yu Tian; 
  • Shiqiang Zhu; 
  • Jingsong Li

ABSTRACT

Background:

Data sharing in multicenter medical research can improve the generalizability of research, accelerate progress, enhance collaborations among institutions, and lead to new discoveries in pooled data from multiple data sources. Despite the above benefits, many medical institutions are unwilling to share their data, as sharing may cause sensitive information leakage to researchers, other institutions and unauthorized users. Secure machine learning frameworks based on homomorphic encryption have made great progress in recent years, but nearly all of them use a single secret key and do not mention how to securely evaluate the trained model, which makes them impractical in multicenter medical applications.

Objective:

The goal of this study is to provide a privacy-preserving machine learning protocol for multiple data providers and researchers (e.g., logistic regression). This protocol allows researchers to train models and evaluate their trained models on medical data from multiple data sources while providing privacy protection for both the sensitive data and the learned model.

Methods:

We adapted a novel threshold homomorphic encryption scheme to guarantee privacy requirements. We devised (1) new relinearization key generation techniques for more scalability and multiplicative depth and (2) new model training strategies for simultaneously training multiple models in x-fold-cross-validation.

Results:

Using a client/server architecture, we evaluated the performance of our protocol. Experimental results demonstrate that in 10-fold-cross-validation, our privacy-preserving logistic regression model training and evaluation over 10 attributes in a dataset of 49,152 samples take approximately 7 min and 20 min, respectively.

Conclusions:

We present the first privacy-preserving multiparty logistic regression model training and evaluation protocol based on threshold homomorphic encryption. Our protocol is practical for real-world use and may promote multicenter medical research to some extent.


 Citation

Please cite as:

Lu Y, Zhou T, Tian Y, Zhu S, Li J

Web-Based Privacy-Preserving Multicenter Medical Data Analysis Tools Via Threshold Homomorphic Encryption: Design and Development Study

J Med Internet Res 2020;22(12):e22555

DOI: 10.2196/22555

PMID: 33289676

PMCID: 7755539

Download PDF


Request queued. Please wait while the file is being generated. It may take some time.

© The authors. All rights reserved. This is a privileged document currently under peer-review/community review (or an accepted/rejected manuscript). Authors have provided JMIR Publications with an exclusive license to publish this preprint on it's website for review and ahead-of-print citation purposes only. While the final peer-reviewed paper may be licensed under a cc-by license on publication, at this stage authors and publisher expressively prohibit redistribution of this draft paper other than for review purposes.