Abstract:We develop a novel harmonization approach for T1‐weighted magnetic resonance imaging using a style‐encoding generative adversarial network that can be used to harmonize entire images for a variety of international, multi‐cohort, neuroimaging collaborations. Results demonstrated that this model avoids the need to control for clinical or demographic information. We showed that our harmonization removed the cross‐site variances, while preserving the anatomical information and clinical meaningful patterns. Recent work within neuroimaging consortia have aimed to identify reproducible, and often subtle, brain signatures of psychiatric or neurological conditions. To allow for high‐powered brain imaging analyses, it is often necessary to pool MR images that were acquired with different protocols across multiple scanners. Current retrospective harmonization techniques have shown promise in removing site‐related image variation. However, most statistical approaches may over‐correct for technical, scanning‐related, variation as they cannot distinguish between confounded image‐acquisition based variability and site‐related population variability. Such statistical methods often require that datasets contain subjects or patient groups with similar clinical or demographic information to isolate the acquisition‐based variability. To overcome this limitation, we consider site‐related magnetic resonance (MR) imaging harmonization as a style transfer problem rather than a domain transfer problem. Using a fully unsupervised deep‐learning framework based on a generative adversarial network (GAN), we show that MR images can be harmonized by inserting the style information encoded from a single reference image, without knowing their site/scanner labels a priori. We trained our model using data from five large‐scale multisite datasets with varied demographics. Results demonstrated that our style‐encoding model can harmonize MR images, and match intensity profiles, without relying on traveling subjects. This model also avoids the need to control for clinical, diagnostic, or demographic information. We highlight the effectiveness of our method for clinical research by comparing extracted cortical and subcortical features, brain‐age estimates, and case–control effect sizes before and after the harmonization. We showed that our harmonization removed the site‐related variances, while preserving the anatomical information and clinical meaningful patterns. We further demonstrated that with a diverse training set, our method successfully harmonized MR images collected from unseen scanners and protocols, suggesting a promising tool for ongoing collaborative studies. Source code is released in USC‐IGC/style_transfer_harmonization (github.com).

Segmentation-Renormalized Deep Feature Modulation for Unpaired Image Harmonization

Shape-Consistent Generative Adversarial Networks for Multi-Modal Medical Segmentation Maps

Towards cross-modal organ translation and segmentation: A cycle- and shape-consistent generative adversarial network

Self-Attentive Spatial Adaptive Normalization for Cross-Modality Domain Adaptation

Harmonization Across Imaging Locations(HAIL): One-Shot Learning for Brain MRI

A Deep Learning Harmonization of Multi-Vendor MRI for Robust Intervertebral Disc Segmentation

Harmonized neonatal brain MR image segmentation model for cross-site datasets

Normalization of breast MRIs using cycle-consistent generative adversarial networks

Style transfer generative adversarial networks to harmonize multisite MRI to a single reference image to avoid overcorrection

IGUANe: a 3D generalizable CycleGAN for multicenter harmonization of brain MR images

ReeGAN: MRI image edge-preserving synthesis based on GANs trained with misaligned data

MedGAN: Medical Image Translation using GANs

Cross-Domain Medical Image Translation by Shared Latent Gaussian Mixture Model

Towards annotation-efficient segmentation via image-to-image translation

Diffusion based multi-domain neuroimaging harmonization method with preservation of anatomical details

Frequency Domain Decomposition Translation for Enhanced Medical Image Translation Using GANs

Harmonizing Flows: Leveraging normalizing flows for unsupervised and source-free MRI harmonization

EnGAN: Enhancement Generative Adversarial Network in Medical Image Segmentation

Multi-modality MRI Arbitrary Transformation Using Unified Generative Adversarial Networks

Multi-Texture GAN: Exploring the Multi-Scale Texture Translation for Brain MR Images

SUSAN: Segment Unannotated image Structure using Adversarial Network