Generalizable Single-Source Cross-modality Medical Image Segmentation via Invariant Causal Mechanisms

TL;DR

This paper introduces a causality-inspired approach using diffusion models to improve single-source domain generalization in cross-modality medical image segmentation, achieving superior results across multiple anatomies and imaging modalities.

Contribution

It combines causality principles with diffusion-based augmentation to enhance domain-invariant feature learning in medical image segmentation from a single source.

Findings

Outperforms state-of-the-art SDG methods on multiple tasks

Effectively simulates diverse imaging styles while preserving content

Demonstrates robustness across different anatomies and modalities

Abstract

Single-source domain generalization (SDG) aims to learn a model from a single source domain that can generalize well on unseen target domains. This is an important task in computer vision, particularly relevant to medical imaging where domain shifts are common. In this work, we consider a challenging yet practical setting: SDG for cross-modality medical image segmentation. We combine causality-inspired theoretical insights on learning domain-invariant representations with recent advancements in diffusion-based augmentation to improve generalization across diverse imaging modalities. Guided by the ``intervention-augmentation equivariant'' principle, we use controlled diffusion models (DMs) to simulate diverse imaging styles while preserving the content, leveraging rich generative priors in large-scale pretrained DMs to comprehensively perturb the multidimensional style variable. Extensive experiments on challenging cross-modality segmentation tasks demonstrate that our approach consistently outperforms state-of-the-art SDG methods across three distinct anatomies and imaging modalities. The source code is available at \href{https://github.com/ratschlab/ICMSeg}{https://github.com/ratschlab/ICMSeg}.