Embracing the Disharmony in Medical Imaging: A Simple and Effective Framework for Domain Adaptation

TL;DR

This paper proposes a simple framework that embraces data disharmony in medical imaging, using pretrained classifiers and auxiliary tasks to improve domain adaptation and generalization across diverse datasets.

Contribution

It introduces a novel approach that leverages pretrained models and auxiliary covariates for effective domain adaptation without relying solely on harmonization techniques.

Findings

Significant improvement in intra-site domain adaptation.

Enhanced inter-site generalization on large-scale MRI datasets.

Effective use of auxiliary tasks with covariates for adaptation.

Abstract

Domain shift, the mismatch between training and testing data characteristics, causes significant degradation in the predictive performance in multi-source imaging scenarios. In medical imaging, the heterogeneity of population, scanners and acquisition protocols at different sites presents a significant domain shift challenge and has limited the widespread clinical adoption of machine learning models. Harmonization methods which aim to learn a representation of data invariant to these differences are the prevalent tools to address domain shift, but they typically result in degradation of predictive accuracy. This paper takes a different perspective of the problem: we embrace this disharmony in data and design a simple but effective framework for tackling domain shift. The key idea, based on our theoretical arguments, is to build a pretrained classifier on the source data and adapt this model to new data. The classifier can be fine-tuned for intra-site domain adaptation. We can also tackle situations where we do not have access to ground-truth labels on target data; we show how one can use auxiliary tasks for adaptation; these tasks employ covariates such as age, gender and race which are easy to obtain but nevertheless correlated to the main task. We demonstrate substantial improvements in both intra-site domain adaptation and inter-site domain generalization on large-scale real-world 3D brain MRI datasets for classifying Alzheimer's disease and schizophrenia.