Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

MedEdit: Counterfactual Diffusion-based Image Editing on Brain MRI

Malek Ben Alaya, Daniel M. Lang, Benedikt Wiestler, Julia A. Schnabel, Cosmin I. Bercea

TL;DR

This paper addresses the challenge of generating realistic counterfactual brain MRI edits to simulate diseases like stroke while keeping the original scan faithful. It introduces MedEdit, a conditional diffusion-based editing method that extends RePaint with mask-conditioned inpainting and a mask-dilution strategy to capture indirect pathological effects such as brain atrophy. In evaluations on Atlas v2.0 stroke data, MedEdit achieves the lowest FID and strong Dice performance, surpassing Palette and naïve RePaint and approaching or outperforming SDEdit, with clinical neuroradiologist assessments confirming realism. The work highlights the need for clinically grounded evaluation metrics and suggests that MedEdit can enable realistic counterfactual imaging for broader pathologies and modalities, improving data augmentation and disease progression analysis.

Abstract

Denoising diffusion probabilistic models enable high-fidelity image synthesis and editing. In biomedicine, these models facilitate counterfactual image editing, producing pairs of images where one is edited to simulate hypothetical conditions. For example, they can model the progression of specific diseases, such as stroke lesions. However, current image editing techniques often fail to generate realistic biomedical counterfactuals, either by inadequately modeling indirect pathological effects like brain atrophy or by excessively altering the scan, which disrupts correspondence to the original images. Here, we propose MedEdit, a conditional diffusion model for medical image editing. MedEdit induces pathology in specific areas while balancing the modeling of disease effects and preserving the integrity of the original scan. We evaluated MedEdit on the Atlas v2.0 stroke dataset using Frechet Inception Distance and Dice scores, outperforming state-of-the-art diffusion-based methods such as Palette (by 45%) and SDEdit (by 61%). Additionally, clinical evaluations by a board-certified neuroradiologist confirmed that MedEdit generated realistic stroke scans indistinguishable from real ones. We believe this work will enable counterfactual image editing research to further advance the development of realistic and clinically useful imaging tools.

MedEdit: Counterfactual Diffusion-based Image Editing on Brain MRI

TL;DR

This paper addresses the challenge of generating realistic counterfactual brain MRI edits to simulate diseases like stroke while keeping the original scan faithful. It introduces MedEdit, a conditional diffusion-based editing method that extends RePaint with mask-conditioned inpainting and a mask-dilution strategy to capture indirect pathological effects such as brain atrophy. In evaluations on Atlas v2.0 stroke data, MedEdit achieves the lowest FID and strong Dice performance, surpassing Palette and naïve RePaint and approaching or outperforming SDEdit, with clinical neuroradiologist assessments confirming realism. The work highlights the need for clinically grounded evaluation metrics and suggests that MedEdit can enable realistic counterfactual imaging for broader pathologies and modalities, improving data augmentation and disease progression analysis.

Abstract

Denoising diffusion probabilistic models enable high-fidelity image synthesis and editing. In biomedicine, these models facilitate counterfactual image editing, producing pairs of images where one is edited to simulate hypothetical conditions. For example, they can model the progression of specific diseases, such as stroke lesions. However, current image editing techniques often fail to generate realistic biomedical counterfactuals, either by inadequately modeling indirect pathological effects like brain atrophy or by excessively altering the scan, which disrupts correspondence to the original images. Here, we propose MedEdit, a conditional diffusion model for medical image editing. MedEdit induces pathology in specific areas while balancing the modeling of disease effects and preserving the integrity of the original scan. We evaluated MedEdit on the Atlas v2.0 stroke dataset using Frechet Inception Distance and Dice scores, outperforming state-of-the-art diffusion-based methods such as Palette (by 45%) and SDEdit (by 61%). Additionally, clinical evaluations by a board-certified neuroradiologist confirmed that MedEdit generated realistic stroke scans indistinguishable from real ones. We believe this work will enable counterfactual image editing research to further advance the development of realistic and clinically useful imaging tools.
Paper Structure (8 sections, 7 equations, 1 figure, 1 table, 1 algorithm)

This paper contains 8 sections, 7 equations, 1 figure, 1 table, 1 algorithm.

Table of Contents

  1. Introduction
  2. Background
  3. Method
  4. Experiments
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments

Figures (1)

  • Figure 2: Examples of counterfactuals obtained with Palette, Naïve RePaint, SDEdit and MedEdit. All methods model the pathology well for the last case in the bottom row (shown in purple difference maps). Additionaly, MedEdit also precisely models indirect pathological changes induced by the pathology, as shown in turquoise. In this case the stroke lesions caused the ventricle on the same side to enlarge.