Lesion-Aware Post-Training of Latent Diffusion Models for Synthesizing Diffusion MRI from CT Perfusion

2025-10-10

AI Summary

Medical image translation often suffers from poor diagnostic reliability due to severe distortion and loss of small but clinically critical lesions—e.g., acute ischemic stroke lesions—owing to their low spatial prevalence. To address this, we propose a lesion-aware post-training framework for latent diffusion models (LDMs). Our method fine-tunes a pre-trained LDM in the latent space using a weighted pixel-space loss that explicitly prioritizes reconstruction fidelity in lesion regions. Gradient optimization is further guided by lesion masks to jointly preserve global anatomical structure and recover local lesion details. Evaluated on CT perfusion-to-DWI/ADC synthesis across 817 patient cases, our approach significantly improves PSNR, SSIM, and lesion boundary sharpness (p < 0.01) over state-of-the-art diffusion-based and GAN-based baselines. The framework demonstrates strong clinical interpretability and practical deployability.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Large language models for search
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Image-to-Image translation models can help mitigate various challenges inherent to medical image acquisition. Latent diffusion models (LDMs) leverage efficient learning in compressed latent space and constitute the core of state-of-the-art generative image models. However, this efficiency comes with a trade-off, potentially compromising crucial pixel-level detail essential for high-fidelity medical images. This limitation becomes particularly critical when generating clinically significant structures, such as lesions, which often occupy only a small portion of the image. Failure to accurately reconstruct these regions can severely impact diagnostic reliability and clinical decision-making. To overcome this limitation, we propose a novel post-training framework for LDMs in medical image-to-image translation by incorporating lesion-aware medical pixel space objectives. This approach is essential, as it not only enhances overall image quality but also improves the precision of lesion delineation. We evaluate our framework on brain CT-to-MRI translation in acute ischemic stroke patients, where early and accurate diagnosis is critical for optimal treatment selection and improved patient outcomes. While diffusion MRI is the gold standard for stroke diagnosis, its clinical utility is often constrained by high costs and low accessibility. Using a dataset of 817 patients, we demonstrate that our framework improves overall image quality and enhances lesion delineation when synthesizing DWI and ADC images from CT perfusion scans, outperforming existing image-to-image translation models. Furthermore, our post-training strategy is easily adaptable to pre-trained LDMs and exhibits substantial potential for broader applications across diverse medical image translation tasks.

Problem

• Enhancing lesion reconstruction in medical image translation
• Improving diagnostic reliability through pixel-level detail preservation
• Overcoming limitations of latent diffusion models for clinical structures

Innovation

• Post-training framework enhances latent diffusion models
• Incorporates lesion-aware medical pixel space objectives
• Improves lesion delineation in CT-to-MRI translation