Privacy-preserving harmonization via distributed ComBat

Abstract

Challenges in clinical data sharing and the need to protect data privacy have led to the development and popularization of methods that do not require directly transferring patient data. In neuroimaging, integration of data across multiple institutions also introduces unwanted biases driven by scanner differences. These scanner effects have been shown by several research groups to severely affect downstream analyses. To facilitate the need of removing scanner effects in a distributed data setting, we introduce distributed ComBat, an adaptation of a popular harmonization method for multivariate data that borrows information across features. We present our fast and simple distributed algorithm and show that it yields equivalent results using data from the Alzheimer's Disease Neuroimaging Initiative. Our method enables harmonization while ensuring maximal privacy protection, thus facilitating a broad range of downstream analyses in functional and structural imaging studies.

Keywords: ComBat; Distributed analysis; Harmonization; Privacy-preserving; Site effect.

Fig. 1.. Distributed ComBat illustration.

The procedure to perform distributed ComBat harmonization is outlined as follows. a, Each site sends its deidentified summary statistics to a central site for estimation of regression coefficients which are then passed back to the sites. b, Each site sends summary statistics to a central site for estimation of the population variance which is then passed back to the sites. c, The sites can then use the global regression coefficients and variance estimates to perform the remaining ComBat steps and obtain harmonized data.

Fig. 2.. Distributed ComBat parameter estimates.

Scatter plots compare parameter estimates from distributed ComBat versus those obtained from ComBat with all data at one location. a and b show empirical Bayes point estimates for location and scale respectively. c displays the regression coefficients obtained from each method.