VS-DDPM: Efficient Diffusion Model for Medical Imaging

Researchers propose 3D Variable-Step Denoising Diffusion Probabilistic Model (VS-DDPM), a framework that accelerates diffusion model inference while maintaining generative quality. Tested on four medical imaging tasks—missing MRI, tumor removal, MRI-to-sCT, and CBCT-to-sCT—within BraTS2025 and SynthRAD2025 challenges, VS-DDPM achieved state-of-the-art performance in missing MRI synthesis with Dice scores of 0.80, 0.83, and 0.88 for enhancing tumor, tumor core, and whole tumor regions, and an SSIM of 0.95. For MRI tumor removal, it attained RMSE of 0.053, PSNR of 26.77, and SSIM of 0.918. The model is designed for high efficiency under hardware and time constraints.